id sid tid token lemma pos bracis-33661 1 1 modeling model VERB bracis-33661 1 2 eeg eeg PROPN bracis-33661 1 3 data datum NOUN bracis-33661 1 4 into into ADP bracis-33661 1 5 graphs graph NOUN bracis-33661 1 6 for for ADP bracis-33661 1 7     SPACE bracis-33661 1 8 the the DET bracis-33661 1 9     SPACE bracis-33661 1 10 prognostic prognostic NOUN bracis-33661 1 11 of of ADP bracis-33661 1 12     SPACE bracis-33661 1 13 patients patient NOUN bracis-33661 1 14 in in ADP bracis-33661 1 15     SPACE bracis-33661 1 16 coma coma NOUN bracis-33661 1 17 using use VERB bracis-33661 1 18 graph graph NOUN bracis-33661 1 19 neural neural ADJ bracis-33661 1 20 networks network NOUN bracis-33661 1 21 | | ADV bracis-33661 1 22 springer springer NOUN bracis-33661 1 23 nature nature PROPN bracis-33661 1 24 link link PROPN bracis-33661 1 25 ( ( PUNCT bracis-33661 1 26 formerly formerly ADV bracis-33661 1 27 springerlink springerlink NOUN bracis-33661 1 28 ) ) PUNCT bracis-33661 1 29 skip skip VERB bracis-33661 1 30 to to ADP bracis-33661 1 31 main main ADJ bracis-33661 1 32 content content NOUN bracis-33661 1 33 advertisement advertisement NOUN bracis-33661 1 34 log log NOUN bracis-33661 1 35 in in ADP bracis-33661 1 36 menu menu NOUN bracis-33661 1 37 find find VERB bracis-33661 1 38 a a DET bracis-33661 1 39 journal journal NOUN bracis-33661 1 40 publish publish VERB bracis-33661 1 41 with with ADP bracis-33661 1 42 us we PRON bracis-33661 1 43 track track VERB bracis-33661 1 44 your your PRON bracis-33661 1 45 research research NOUN bracis-33661 1 46 search search NOUN bracis-33661 1 47 cart cart NOUN bracis-33661 1 48 home home NOUN bracis-33661 1 49 intelligent intelligent ADJ bracis-33661 1 50 systems system NOUN bracis-33661 1 51 conference conference NOUN bracis-33661 1 52 paper paper NOUN bracis-33661 1 53 modeling model VERB bracis-33661 1 54 eeg eeg PROPN bracis-33661 1 55 data datum NOUN bracis-33661 1 56 into into ADP bracis-33661 1 57 graphs graph NOUN bracis-33661 1 58 for for ADP bracis-33661 1 59     SPACE bracis-33661 1 60 the the DET bracis-33661 1 61     SPACE bracis-33661 1 62 prognostic prognostic NOUN bracis-33661 1 63 of of ADP bracis-33661 1 64     SPACE bracis-33661 1 65 patients patient NOUN bracis-33661 1 66 in in ADP bracis-33661 1 67     SPACE bracis-33661 1 68 coma coma NOUN bracis-33661 1 69 using use VERB bracis-33661 1 70 graph graph NOUN bracis-33661 1 71 neural neural ADJ bracis-33661 1 72 networks network NOUN bracis-33661 1 73 conference conference NOUN bracis-33661 1 74 paper paper NOUN bracis-33661 1 75 first first ADV bracis-33661 1 76 online online ADV bracis-33661 1 77 : : PUNCT bracis-33661 1 78 31 31 NUM bracis-33661 1 79 january january NOUN bracis-33661 1 80 2025 2025 NUM bracis-33661 1 81 pp pp PROPN bracis-33661 1 82 223–237 223–237 NUM bracis-33661 1 83 cite cite VERB bracis-33661 1 84 this this DET bracis-33661 1 85 conference conference NOUN bracis-33661 1 86 paper paper NOUN bracis-33661 1 87 access access NOUN bracis-33661 1 88 provided provide VERB bracis-33661 1 89 by by ADP bracis-33661 1 90 university university PROPN bracis-33661 1 91 of of ADP bracis-33661 1 92 notre notre PROPN bracis-33661 1 93 dame dame PROPN bracis-33661 1 94 hesburgh hesburgh PROPN bracis-33661 1 95 library library PROPN bracis-33661 1 96 download download PROPN bracis-33661 1 97 book book NOUN bracis-33661 1 98 pdf pdf PROPN bracis-33661 1 99 download download NOUN bracis-33661 1 100 book book NOUN bracis-33661 1 101 epub epub PROPN bracis-33661 1 102 intelligent intelligent ADJ bracis-33661 1 103 systems system NOUN bracis-33661 1 104 ( ( PUNCT bracis-33661 1 105 bracis bracis NOUN bracis-33661 1 106 2024 2024 NUM bracis-33661 1 107 ) ) PUNCT bracis-33661 1 108 modeling model VERB bracis-33661 1 109 eeg eeg NOUN bracis-33661 1 110 data datum NOUN bracis-33661 1 111 into into ADP bracis-33661 1 112 graphs graph NOUN bracis-33661 1 113 for for ADP bracis-33661 1 114     SPACE bracis-33661 1 115 the the DET bracis-33661 1 116     SPACE bracis-33661 1 117 prognostic prognostic NOUN bracis-33661 1 118 of of ADP bracis-33661 1 119     SPACE bracis-33661 1 120 patients patient NOUN bracis-33661 1 121 in in ADP bracis-33661 1 122     SPACE bracis-33661 1 123 coma coma NOUN bracis-33661 1 124 using use VERB bracis-33661 1 125 graph graph NOUN bracis-33661 1 126 neural neural ADJ bracis-33661 1 127 networks network NOUN bracis-33661 1 128 download download NOUN bracis-33661 1 129 book book NOUN bracis-33661 1 130 pdf pdf PROPN bracis-33661 1 131 download download NOUN bracis-33661 1 132 book book NOUN bracis-33661 2 1 epub epub NOUN bracis-33661 2 2 odelmo odelmo PROPN bracis-33661 2 3 o. o. PROPN bracis-33661 2 4 nascimento nascimento PROPN bracis-33661 2 5 junior junior PROPN bracis-33661 2 6     SPACE bracis-33661 2 7 orcid orcid NOUN bracis-33661 2 8 : : PUNCT bracis-33661 2 9 orcid.org/0000-0001-8360-56009 orcid.org/0000-0001-8360-56009 ADJ bracis-33661 2 10 , , PUNCT bracis-33661 2 11 dhara dhara PROPN bracis-33661 2 12 l. l. PROPN bracis-33661 2 13 c. c. PROPN bracis-33661 2 14 assis assis PROPN bracis-33661 2 15     SPACE bracis-33661 2 16 orcid orcid NOUN bracis-33661 2 17 : : PUNCT bracis-33661 2 18 orcid.org/0000-0002-3840-55949 orcid.org/0000-0002-3840-55949 NOUN bracis-33661 2 19 , , PUNCT bracis-33661 2 20 joão joão PROPN bracis-33661 2 21 b. b. PROPN bracis-33661 2 22 destro destro PROPN bracis-33661 2 23 filho filho PROPN bracis-33661 2 24     SPACE bracis-33661 2 25 orcid orcid NOUN bracis-33661 2 26 : : PUNCT bracis-33661 2 27 orcid.org/0000-0001-5306-80589 orcid.org/0000-0001-5306-80589 PROPN bracis-33661 2 28 , , PUNCT bracis-33661 2 29 liang liang PROPN bracis-33661 2 30 zhao zhao PROPN bracis-33661 2 31     SPACE bracis-33661 2 32 orcid orcid PROPN bracis-33661 2 33 : : PUNCT bracis-33661 2 34 orcid.org/0000-0002-1502-660410 orcid.org/0000-0002-1502-660410 PROPN bracis-33661 2 35 & & CCONJ bracis-33661 2 36 … … PUNCT bracis-33661 2 37 murillo murillo PROPN bracis-33661 2 38 g. g. PROPN bracis-33661 2 39 carneiro carneiro PROPN bracis-33661 2 40     SPACE bracis-33661 2 41 orcid orcid PROPN bracis-33661 2 42 : : PUNCT bracis-33661 2 43 orcid.org/0000-0002-2915-89909,10 orcid.org/0000-0002-2915-89909,10 NOUN bracis-33661 2 44     SPACE bracis-33661 2 45 show show VERB bracis-33661 2 46 authors author NOUN bracis-33661 2 47 part part NOUN bracis-33661 2 48 of of ADP bracis-33661 2 49 the the DET bracis-33661 2 50 book book NOUN bracis-33661 2 51 series series NOUN bracis-33661 2 52 : : PUNCT bracis-33661 2 53 lecture lecture NOUN bracis-33661 2 54 notes note NOUN bracis-33661 2 55 in in ADP bracis-33661 2 56 computer computer NOUN bracis-33661 2 57 science science NOUN bracis-33661 2 58 ( ( PUNCT bracis-33661 2 59 ( ( PUNCT bracis-33661 2 60 lnai lnai ADJ bracis-33661 2 61 , , PUNCT bracis-33661 2 62 volume volume NOUN bracis-33661 2 63 15415 15415 NUM bracis-33661 2 64 ) ) PUNCT bracis-33661 2 65 ) ) PUNCT bracis-33661 2 66 included include VERB bracis-33661 2 67 in in ADP bracis-33661 2 68 the the DET bracis-33661 2 69 following follow VERB bracis-33661 2 70 conference conference NOUN bracis-33661 2 71 series series NOUN bracis-33661 2 72 : : PUNCT bracis-33661 2 73 brazilian brazilian ADJ bracis-33661 2 74 conference conference NOUN bracis-33661 2 75 on on ADP bracis-33661 2 76 intelligent intelligent ADJ bracis-33661 2 77 systems system NOUN bracis-33661 2 78 232 232 NUM bracis-33661 2 79 accesses access NOUN bracis-33661 2 80 abstract abstract ADJ bracis-33661 2 81 in in ADP bracis-33661 2 82 this this DET bracis-33661 2 83 work work NOUN bracis-33661 2 84 we we PRON bracis-33661 2 85 present present VERB bracis-33661 2 86 a a DET bracis-33661 2 87 modeling modeling NOUN bracis-33661 2 88 approach approach NOUN bracis-33661 2 89 using use VERB bracis-33661 2 90 graph graph NOUN bracis-33661 2 91 neural neural ADJ bracis-33661 2 92 networks network NOUN bracis-33661 2 93 ( ( PUNCT bracis-33661 2 94 gnns gnns NOUN bracis-33661 2 95 ) ) PUNCT bracis-33661 2 96 to to PART bracis-33661 2 97 support support VERB bracis-33661 2 98 the the DET bracis-33661 2 99 prognosis prognosis NOUN bracis-33661 2 100 of of ADP bracis-33661 2 101 patients patient NOUN bracis-33661 2 102 in in ADP bracis-33661 2 103 coma coma NOUN bracis-33661 2 104 ( ( PUNCT bracis-33661 2 105 ppc ppc NOUN bracis-33661 2 106 ) ) PUNCT bracis-33661 2 107 based base VERB bracis-33661 2 108 on on ADP bracis-33661 2 109 electroencephalogram electroencephalogram NOUN bracis-33661 2 110 ( ( PUNCT bracis-33661 2 111 eeg eeg NOUN bracis-33661 2 112 ) ) PUNCT bracis-33661 2 113 exams exam NOUN bracis-33661 2 114 . . PUNCT bracis-33661 3 1 coma coma NOUN bracis-33661 3 2 is be AUX bracis-33661 3 3 a a DET bracis-33661 3 4 critical critical ADJ bracis-33661 3 5 medical medical ADJ bracis-33661 3 6 condition condition NOUN bracis-33661 3 7 characterized characterize VERB bracis-33661 3 8 by by ADP bracis-33661 3 9 an an DET bracis-33661 3 10 extended extended ADJ bracis-33661 3 11 state state NOUN bracis-33661 3 12 of of ADP bracis-33661 3 13 unconsciousness unconsciousness NOUN bracis-33661 3 14 . . PUNCT bracis-33661 4 1 eeg eeg PROPN bracis-33661 4 2 is be AUX bracis-33661 4 3 an an DET bracis-33661 4 4 exam exam NOUN bracis-33661 4 5 that that PRON bracis-33661 4 6 captures capture VERB bracis-33661 4 7 the the DET bracis-33661 4 8 electrical electrical ADJ bracis-33661 4 9 activity activity NOUN bracis-33661 4 10 of of ADP bracis-33661 4 11 the the DET bracis-33661 4 12 brain brain NOUN bracis-33661 4 13 through through ADP bracis-33661 4 14 electrodes electrode NOUN bracis-33661 4 15 positioned position VERB bracis-33661 4 16 on on ADP bracis-33661 4 17 the the DET bracis-33661 4 18 patient patient NOUN bracis-33661 4 19 ’s ’s PART bracis-33661 4 20 scalp scalp NOUN bracis-33661 4 21 . . PUNCT bracis-33661 5 1 it it PRON bracis-33661 5 2 provides provide VERB bracis-33661 5 3 valuable valuable ADJ bracis-33661 5 4 information information NOUN bracis-33661 5 5 about about ADP bracis-33661 5 6 cerebral cerebral ADJ bracis-33661 5 7 functions function NOUN bracis-33661 5 8 and and CCONJ bracis-33661 5 9 can can AUX bracis-33661 5 10 also also ADV bracis-33661 5 11 be be AUX bracis-33661 5 12 used use VERB bracis-33661 5 13 for for ADP bracis-33661 5 14 the the DET bracis-33661 5 15 monitoring monitoring NOUN bracis-33661 5 16 and and CCONJ bracis-33661 5 17 diagnosis diagnosis NOUN bracis-33661 5 18 of of ADP bracis-33661 5 19 several several ADJ bracis-33661 5 20 diseases disease NOUN bracis-33661 5 21 . . PUNCT bracis-33661 6 1 in in ADP bracis-33661 6 2 this this DET bracis-33661 6 3 study study NOUN bracis-33661 6 4 , , PUNCT bracis-33661 6 5 we we PRON bracis-33661 6 6 investigate investigate VERB bracis-33661 6 7 four four NUM bracis-33661 6 8 gnn gnn PROPN bracis-33661 6 9 architectures architecture NOUN bracis-33661 6 10 specifically specifically ADV bracis-33661 6 11 designed design VERB bracis-33661 6 12 for for ADP bracis-33661 6 13 ppc ppc ADJ bracis-33661 6 14 analysis analysis NOUN bracis-33661 6 15 , , PUNCT bracis-33661 6 16 leveraging leverage VERB bracis-33661 6 17 the the DET bracis-33661 6 18 power power NOUN bracis-33661 6 19 of of ADP bracis-33661 6 20 convolutional convolutional ADJ bracis-33661 6 21 - - PUNCT bracis-33661 6 22 based base VERB bracis-33661 6 23 layers layer NOUN bracis-33661 6 24 . . PUNCT bracis-33661 7 1 in in ADP bracis-33661 7 2 addition addition NOUN bracis-33661 7 3 , , PUNCT bracis-33661 7 4 we we PRON bracis-33661 7 5 also also ADV bracis-33661 7 6 develop develop VERB bracis-33661 7 7 a a DET bracis-33661 7 8 modeling modeling NOUN bracis-33661 7 9 strategy strategy NOUN bracis-33661 7 10 to to PART bracis-33661 7 11 represent represent VERB bracis-33661 7 12 eeg eeg PROPN bracis-33661 7 13 sequential sequential ADJ bracis-33661 7 14 data datum NOUN bracis-33661 7 15 into into ADP bracis-33661 7 16 a a DET bracis-33661 7 17 graph graph NOUN bracis-33661 7 18 , , PUNCT bracis-33661 7 19 named name VERB bracis-33661 7 20 eegraph eegraph NOUN bracis-33661 7 21 , , PUNCT bracis-33661 7 22 effectively effectively ADV bracis-33661 7 23 harnessing harness VERB bracis-33661 7 24 the the DET bracis-33661 7 25 spatial spatial ADJ bracis-33661 7 26 positions position NOUN bracis-33661 7 27 of of ADP bracis-33661 7 28 electrodes electrode NOUN bracis-33661 7 29 . . PUNCT bracis-33661 8 1 the the DET bracis-33661 8 2 main main ADJ bracis-33661 8 3 contribution contribution NOUN bracis-33661 8 4 of of ADP bracis-33661 8 5 representing represent VERB bracis-33661 8 6 the the DET bracis-33661 8 7 eeg eeg NOUN bracis-33661 8 8 exam exam NOUN bracis-33661 8 9 in in ADP bracis-33661 8 10 a a DET bracis-33661 8 11 graph graph NOUN bracis-33661 8 12 structure structure NOUN bracis-33661 8 13 is be AUX bracis-33661 8 14 also also ADV bracis-33661 8 15 to to PART bracis-33661 8 16 explore explore VERB bracis-33661 8 17 all all DET bracis-33661 8 18 the the DET bracis-33661 8 19 inherent inherent ADJ bracis-33661 8 20 spatial spatial ADJ bracis-33661 8 21 information information NOUN bracis-33661 8 22 the the DET bracis-33661 8 23 exam exam NOUN bracis-33661 8 24 has have VERB bracis-33661 8 25 . . PUNCT bracis-33661 9 1 as as SCONJ bracis-33661 9 2 it it PRON bracis-33661 9 3 is be AUX bracis-33661 9 4 a a DET bracis-33661 9 5 highly highly ADV bracis-33661 9 6 sensitive sensitive ADJ bracis-33661 9 7 exam exam NOUN bracis-33661 9 8 with with ADP bracis-33661 9 9 generally generally ADV bracis-33661 9 10 noisy noisy ADJ bracis-33661 9 11 results result NOUN bracis-33661 9 12 on on ADP bracis-33661 9 13 the the DET bracis-33661 9 14 micro micro ADJ bracis-33661 9 15 - - NOUN bracis-33661 9 16 volt volt NOUN bracis-33661 9 17 scale scale NOUN bracis-33661 9 18 , , PUNCT bracis-33661 9 19 its its PRON bracis-33661 9 20 spatial spatial ADJ bracis-33661 9 21 exploration exploration NOUN bracis-33661 9 22 may may AUX bracis-33661 9 23 reveal reveal VERB bracis-33661 9 24 the the DET bracis-33661 9 25 potential potential NOUN bracis-33661 9 26 to to PART bracis-33661 9 27 improve improve VERB bracis-33661 9 28 the the DET bracis-33661 9 29 ppc ppc ADJ bracis-33661 9 30 task task NOUN bracis-33661 9 31 . . PUNCT bracis-33661 10 1 experimental experimental ADJ bracis-33661 10 2 results result NOUN bracis-33661 10 3 considering consider VERB bracis-33661 10 4 real real ADJ bracis-33661 10 5 data datum NOUN bracis-33661 10 6 showed show VERB bracis-33661 10 7 that that SCONJ bracis-33661 10 8 our our PRON bracis-33661 10 9 proposed propose VERB bracis-33661 10 10 architecture architecture NOUN bracis-33661 10 11 outperformed outperform VERB bracis-33661 10 12 the the DET bracis-33661 10 13 lstm lstm PROPN bracis-33661 10 14 model model NOUN bracis-33661 10 15 in in ADP bracis-33661 10 16 terms term NOUN bracis-33661 10 17 of of ADP bracis-33661 10 18 f1 f1 NOUN bracis-33661 10 19 - - PUNCT bracis-33661 10 20 score score NOUN bracis-33661 10 21 and and CCONJ bracis-33661 10 22 accuracy accuracy NOUN bracis-33661 10 23 . . PUNCT bracis-33661 11 1 statistical statistical ADJ bracis-33661 11 2 tests test NOUN bracis-33661 11 3 evidenced evidence VERB bracis-33661 11 4 the the DET bracis-33661 11 5 enhanced enhanced ADJ bracis-33661 11 6 predictive predictive ADJ bracis-33661 11 7 performance performance NOUN bracis-33661 11 8 achieved achieve VERB bracis-33661 11 9 by by ADP bracis-33661 11 10 exploiting exploit VERB bracis-33661 11 11 the the DET bracis-33661 11 12 spatial spatial ADJ bracis-33661 11 13 aspects aspect NOUN bracis-33661 11 14 of of ADP bracis-33661 11 15 eeg eeg NOUN bracis-33661 11 16 data datum NOUN bracis-33661 11 17 using use VERB bracis-33661 11 18 gnns gnn NOUN bracis-33661 11 19 . . PUNCT bracis-33661 12 1 moreover moreover ADV bracis-33661 12 2 , , PUNCT bracis-33661 12 3 the the DET bracis-33661 12 4 findings finding NOUN bracis-33661 12 5 highlight highlight VERB bracis-33661 12 6 the the DET bracis-33661 12 7 significant significant ADJ bracis-33661 12 8 potential potential NOUN bracis-33661 12 9 of of ADP bracis-33661 12 10 gnns gnns NOUN bracis-33661 12 11 to to PART bracis-33661 12 12 contribute contribute VERB bracis-33661 12 13 to to ADP bracis-33661 12 14 the the DET bracis-33661 12 15 ppc ppc ADJ bracis-33661 12 16 task task NOUN bracis-33661 12 17 from from ADP bracis-33661 12 18 eeg eeg PROPN bracis-33661 12 19 data data PROPN bracis-33661 12 20 . . PUNCT bracis-33661 13 1 access access NOUN bracis-33661 13 2 provided provide VERB bracis-33661 13 3 by by ADP bracis-33661 13 4 university university PROPN bracis-33661 13 5 of of ADP bracis-33661 13 6 notre notre PROPN bracis-33661 13 7 dame dame PROPN bracis-33661 13 8 hesburgh hesburgh PROPN bracis-33661 13 9 library library PROPN bracis-33661 13 10 . . PUNCT bracis-33661 14 1 download download PROPN bracis-33661 14 2 conference conference NOUN bracis-33661 14 3 paper paper NOUN bracis-33661 14 4 pdf pdf NOUN bracis-33661 14 5 similar similar ADJ bracis-33661 14 6 content content NOUN bracis-33661 14 7 being be AUX bracis-33661 14 8 viewed view VERB bracis-33661 14 9 by by ADP bracis-33661 14 10 others other NOUN bracis-33661 14 11 a a DET bracis-33661 14 12 graph graph NOUN bracis-33661 14 13 theory theory NOUN bracis-33661 14 14 analysis analysis NOUN bracis-33661 14 15 on on ADP bracis-33661 14 16 distinguishing distinguish VERB bracis-33661 14 17 eeg eeg NOUN bracis-33661 14 18 - - PUNCT bracis-33661 14 19 based base VERB bracis-33661 14 20 brain brain NOUN bracis-33661 14 21 death death NOUN bracis-33661 14 22 and and CCONJ bracis-33661 14 23 coma coma NOUN bracis-33661 14 24 chapter chapter NOUN bracis-33661 14 25 © © PROPN bracis-33661 14 26 2017 2017 NUM bracis-33661 14 27 eeg eeg NOUN bracis-33661 14 28 models model NOUN bracis-33661 14 29 and and CCONJ bracis-33661 14 30 analysis analysis NOUN bracis-33661 14 31 chapter chapter NOUN bracis-33661 14 32 © © NOUN bracis-33661 14 33 2021 2021 NUM bracis-33661 14 34 eeg eeg NOUN bracis-33661 14 35 models model NOUN bracis-33661 14 36 and and CCONJ bracis-33661 14 37 analysis analysis NOUN bracis-33661 14 38 chapter chapter NOUN bracis-33661 14 39 © © PROPN bracis-33661 14 40 2023 2023 NUM bracis-33661 14 41 1 1 NUM bracis-33661 14 42 introduction introduction NOUN bracis-33661 14 43 coma coma NOUN bracis-33661 14 44 is be AUX bracis-33661 14 45 a a DET bracis-33661 14 46 serious serious ADJ bracis-33661 14 47 medical medical ADJ bracis-33661 14 48 condition condition NOUN bracis-33661 14 49 characterized characterize VERB bracis-33661 14 50 by by ADP bracis-33661 14 51 a a DET bracis-33661 14 52 prolonged prolonged ADJ bracis-33661 14 53 state state NOUN bracis-33661 14 54 of of ADP bracis-33661 14 55 unconsciousness unconsciousness NOUN bracis-33661 14 56 in in ADP bracis-33661 14 57 which which PRON bracis-33661 14 58 an an DET bracis-33661 14 59 individual individual NOUN bracis-33661 14 60 has have VERB bracis-33661 14 61 little little ADJ bracis-33661 14 62 or or CCONJ bracis-33661 14 63 no no PRON bracis-33661 14 64 level level NOUN bracis-33661 14 65 of of ADP bracis-33661 14 66 response response NOUN bracis-33661 14 67 [ [ X bracis-33661 14 68 9 9 NUM bracis-33661 14 69 , , PUNCT bracis-33661 14 70 10 10 NUM bracis-33661 14 71 ] ] PUNCT bracis-33661 14 72 . . PUNCT bracis-33661 15 1 the the DET bracis-33661 15 2 electroencephalogram electroencephalogram NOUN bracis-33661 15 3 ( ( PUNCT bracis-33661 15 4 eeg eeg NOUN bracis-33661 15 5 ) ) PUNCT bracis-33661 15 6 is be AUX bracis-33661 15 7 a a DET bracis-33661 15 8 medical medical ADJ bracis-33661 15 9 exam exam NOUN bracis-33661 15 10 that that PRON bracis-33661 15 11 can can AUX bracis-33661 15 12 record record VERB bracis-33661 15 13 the the DET bracis-33661 15 14 brain brain NOUN bracis-33661 15 15 ’s ’s PART bracis-33661 15 16 electrical electrical ADJ bracis-33661 15 17 activity activity NOUN bracis-33661 15 18 . . PUNCT bracis-33661 16 1 it it PRON bracis-33661 16 2 represents represent VERB bracis-33661 16 3 the the DET bracis-33661 16 4 summation summation NOUN bracis-33661 16 5 effect effect NOUN bracis-33661 16 6 of of ADP bracis-33661 16 7 many many ADJ bracis-33661 16 8 excitatory excitatory NOUN bracis-33661 16 9 and and CCONJ bracis-33661 16 10 inhibitory inhibitory ADJ bracis-33661 16 11 postsynaptic postsynaptic ADJ bracis-33661 16 12 potentials potential NOUN bracis-33661 16 13 produced produce VERB bracis-33661 16 14 in in ADP bracis-33661 16 15 the the DET bracis-33661 16 16 pyramidal pyramidal ADJ bracis-33661 16 17 layer layer NOUN bracis-33661 16 18 of of ADP bracis-33661 16 19 the the DET bracis-33661 16 20 cerebral cerebral ADJ bracis-33661 16 21 cortex cortex NOUN bracis-33661 16 22 [ [ X bracis-33661 16 23 29 29 NUM bracis-33661 16 24 ] ] PUNCT bracis-33661 16 25 . . PUNCT bracis-33661 17 1 data datum NOUN bracis-33661 17 2 from from ADP bracis-33661 17 3 this this DET bracis-33661 17 4 exam exam NOUN bracis-33661 17 5 are be AUX bracis-33661 17 6 commonly commonly ADV bracis-33661 17 7 used use VERB bracis-33661 17 8 for for ADP bracis-33661 17 9 monitoring monitoring NOUN bracis-33661 17 10 and and CCONJ bracis-33661 17 11 diagnosing diagnose VERB bracis-33661 17 12 various various ADJ bracis-33661 17 13 diseases disease NOUN bracis-33661 17 14 , , PUNCT bracis-33661 17 15 some some PRON bracis-33661 17 16 of of ADP bracis-33661 17 17 which which PRON bracis-33661 17 18 are be AUX bracis-33661 17 19 schizophrenia schizophrenia NOUN bracis-33661 18 1 [ [ X bracis-33661 18 2 24 24 NUM bracis-33661 18 3 ] ] PUNCT bracis-33661 19 1 and and CCONJ bracis-33661 19 2 epilepsy epilepsy VERB bracis-33661 19 3 [ [ PUNCT bracis-33661 19 4 26 26 NUM bracis-33661 19 5 ] ] PUNCT bracis-33661 19 6 . . PUNCT bracis-33661 20 1 the the DET bracis-33661 20 2 adoption adoption NOUN bracis-33661 20 3 of of ADP bracis-33661 20 4 eeg eeg PROPN bracis-33661 20 5 data datum NOUN bracis-33661 20 6 for for ADP bracis-33661 20 7 disease disease NOUN bracis-33661 20 8 detection detection NOUN bracis-33661 20 9 tasks task NOUN bracis-33661 20 10 often often ADV bracis-33661 20 11 considers consider VERB bracis-33661 20 12 well well ADV bracis-33661 20 13 - - PUNCT bracis-33661 20 14 known know VERB bracis-33661 20 15 machine machine NOUN bracis-33661 20 16 learning learning NOUN bracis-33661 20 17 ( ( PUNCT bracis-33661 20 18 ml ml NOUN bracis-33661 20 19 ) ) PUNCT bracis-33661 20 20 algorithms algorithm NOUN bracis-33661 20 21 such such ADJ bracis-33661 20 22 as as ADP bracis-33661 20 23 adaptive adaptive ADJ bracis-33661 20 24 boosting boosting NOUN bracis-33661 20 25 , , PUNCT bracis-33661 20 26 support support NOUN bracis-33661 20 27 vector vector NOUN bracis-33661 20 28 machine machine NOUN bracis-33661 20 29 and and CCONJ bracis-33661 20 30 tree tree NOUN bracis-33661 20 31 - - PUNCT bracis-33661 20 32 based base VERB bracis-33661 20 33 algorithms algorithm NOUN bracis-33661 20 34 that that PRON bracis-33661 20 35 compare compare VERB bracis-33661 20 36 portions portion NOUN bracis-33661 20 37 of of ADP bracis-33661 20 38 the the DET bracis-33661 20 39 exam exam NOUN bracis-33661 20 40 for for ADP bracis-33661 20 41 one one NUM bracis-33661 20 42 or or CCONJ bracis-33661 20 43 more more ADJ bracis-33661 20 44 electrodes electrode NOUN bracis-33661 20 45 [ [ X bracis-33661 20 46 17 17 NUM bracis-33661 20 47 , , PUNCT bracis-33661 20 48 22 22 NUM bracis-33661 20 49 , , PUNCT bracis-33661 20 50 28 28 NUM bracis-33661 20 51 ] ] PUNCT bracis-33661 20 52 . . PUNCT bracis-33661 21 1 to to PART bracis-33661 21 2 enable enable VERB bracis-33661 21 3 the the DET bracis-33661 21 4 application application NOUN bracis-33661 21 5 of of ADP bracis-33661 21 6 ml ml NOUN bracis-33661 21 7 models model NOUN bracis-33661 21 8 to to ADP bracis-33661 21 9 examination examination NOUN bracis-33661 21 10 data datum NOUN bracis-33661 21 11 , , PUNCT bracis-33661 21 12 traditional traditional ADJ bracis-33661 21 13 approaches approach NOUN bracis-33661 21 14 involve involve VERB bracis-33661 21 15 the the DET bracis-33661 21 16 application application NOUN bracis-33661 21 17 of of ADP bracis-33661 21 18 various various ADJ bracis-33661 21 19 filters filter NOUN bracis-33661 21 20 to to ADP bracis-33661 21 21 all all DET bracis-33661 21 22 channels channel NOUN bracis-33661 21 23 to to PART bracis-33661 21 24 extract extract VERB bracis-33661 21 25 spectral spectral ADJ bracis-33661 21 26 attributes attribute NOUN bracis-33661 21 27 [ [ X bracis-33661 21 28 5 5 NUM bracis-33661 21 29 ] ] PUNCT bracis-33661 21 30 . . PUNCT bracis-33661 22 1 subsequently subsequently ADV bracis-33661 22 2 , , PUNCT bracis-33661 22 3 these these DET bracis-33661 22 4 attribute attribute NOUN bracis-33661 22 5 vectors vector NOUN bracis-33661 22 6 are be AUX bracis-33661 22 7 formed form VERB bracis-33661 22 8 and and CCONJ bracis-33661 22 9 processed process VERB bracis-33661 22 10 through through ADP bracis-33661 22 11 the the DET bracis-33661 22 12 cited cite VERB bracis-33661 22 13 algorithms algorithm NOUN bracis-33661 22 14 . . PUNCT bracis-33661 23 1 in in ADP bracis-33661 23 2 addition addition NOUN bracis-33661 23 3 to to ADP bracis-33661 23 4 disease disease NOUN bracis-33661 23 5 - - PUNCT bracis-33661 23 6 related relate VERB bracis-33661 23 7 monitoring monitoring NOUN bracis-33661 23 8 and and CCONJ bracis-33661 23 9 diagnosis diagnosis NOUN bracis-33661 23 10 , , PUNCT bracis-33661 23 11 eeg eeg NOUN bracis-33661 23 12 data datum NOUN bracis-33661 23 13 can can AUX bracis-33661 23 14 be be AUX bracis-33661 23 15 used use VERB bracis-33661 23 16 for for ADP bracis-33661 23 17 behavioral behavioral ADJ bracis-33661 23 18 and and CCONJ bracis-33661 23 19 emotional emotional ADJ bracis-33661 23 20 analysis analysis NOUN bracis-33661 23 21 of of ADP bracis-33661 23 22 patients patient NOUN bracis-33661 23 23 . . PUNCT bracis-33661 24 1 some some DET bracis-33661 24 2 studies study NOUN bracis-33661 24 3 show show VERB bracis-33661 24 4 that that SCONJ bracis-33661 24 5 it it PRON bracis-33661 24 6 is be AUX bracis-33661 24 7 possible possible ADJ bracis-33661 24 8 to to PART bracis-33661 24 9 obtain obtain VERB bracis-33661 24 10 good good ADJ bracis-33661 24 11 results result NOUN bracis-33661 24 12 in in ADP bracis-33661 24 13 predicting predict VERB bracis-33661 24 14 emotions emotion NOUN bracis-33661 24 15 applying apply VERB bracis-33661 24 16 feature feature NOUN bracis-33661 24 17 extraction extraction NOUN bracis-33661 24 18 methods method NOUN bracis-33661 24 19 by by ADP bracis-33661 24 20 examining examine VERB bracis-33661 24 21 time time NOUN bracis-33661 24 22 series series NOUN bracis-33661 24 23 followed follow VERB bracis-33661 24 24 by by ADP bracis-33661 24 25 the the DET bracis-33661 24 26 application application NOUN bracis-33661 24 27 of of ADP bracis-33661 24 28 ml ml NOUN bracis-33661 24 29 models model NOUN bracis-33661 24 30 [ [ X bracis-33661 24 31 20 20 NUM bracis-33661 24 32 , , PUNCT bracis-33661 24 33 27 27 NUM bracis-33661 24 34 ] ] PUNCT bracis-33661 24 35 . . PUNCT bracis-33661 25 1 these these DET bracis-33661 25 2 studies study NOUN bracis-33661 25 3 performed perform VERB bracis-33661 25 4 good good ADJ bracis-33661 25 5 results result NOUN bracis-33661 25 6 in in ADP bracis-33661 25 7 predicting predict VERB bracis-33661 25 8 emotions emotion NOUN bracis-33661 25 9 by by ADP bracis-33661 25 10 combining combine VERB bracis-33661 25 11 a a DET bracis-33661 25 12 series series NOUN bracis-33661 25 13 of of ADP bracis-33661 25 14 attributes attribute NOUN bracis-33661 25 15 extracted extract VERB bracis-33661 25 16 from from ADP bracis-33661 25 17 the the DET bracis-33661 25 18 time time NOUN bracis-33661 25 19 series series NOUN bracis-33661 25 20 resulting result VERB bracis-33661 25 21 from from ADP bracis-33661 25 22 the the DET bracis-33661 25 23 eeg eeg NOUN bracis-33661 25 24 examination examination NOUN bracis-33661 25 25 , , PUNCT bracis-33661 25 26 most most ADJ bracis-33661 25 27 of of ADP bracis-33661 25 28 the the DET bracis-33661 25 29 time time NOUN bracis-33661 25 30 combining combine VERB bracis-33661 25 31 attributes attribute NOUN bracis-33661 25 32 from from ADP bracis-33661 25 33 the the DET bracis-33661 25 34 time time NOUN bracis-33661 25 35 and and CCONJ bracis-33661 25 36 frequency frequency NOUN bracis-33661 25 37 domains domain NOUN bracis-33661 25 38 . . PUNCT bracis-33661 26 1 to to PART bracis-33661 26 2 improve improve VERB bracis-33661 26 3 the the DET bracis-33661 26 4 results result NOUN bracis-33661 26 5 , , PUNCT bracis-33661 26 6 other other ADJ bracis-33661 26 7 studies study NOUN bracis-33661 26 8 combine combine VERB bracis-33661 26 9 time time PROPN bracis-33661 26 10 series series PROPN bracis-33661 26 11 attributes attribute VERB bracis-33661 26 12 with with ADP bracis-33661 26 13 the the DET bracis-33661 26 14 result result NOUN bracis-33661 26 15 of of ADP bracis-33661 26 16 facial facial ADJ bracis-33661 26 17 expression expression NOUN bracis-33661 26 18 detection detection NOUN bracis-33661 27 1 [ [ X bracis-33661 27 2 14 14 NUM bracis-33661 27 3 , , PUNCT bracis-33661 27 4 30 30 NUM bracis-33661 27 5 ] ] PUNCT bracis-33661 27 6 . . PUNCT bracis-33661 28 1 other other ADJ bracis-33661 28 2 studies study NOUN bracis-33661 28 3 show show VERB bracis-33661 28 4 the the DET bracis-33661 28 5 impact impact NOUN bracis-33661 28 6 that that PRON bracis-33661 28 7 a a DET bracis-33661 28 8 musical musical ADJ bracis-33661 28 9 stimulus stimulus NOUN bracis-33661 28 10 can can AUX bracis-33661 28 11 bring bring VERB bracis-33661 28 12 to to ADP bracis-33661 28 13 the the DET bracis-33661 28 14 brain brain NOUN bracis-33661 28 15 activity activity NOUN bracis-33661 28 16 behavior behavior NOUN bracis-33661 28 17 [ [ X bracis-33661 28 18 15 15 NUM bracis-33661 28 19 , , PUNCT bracis-33661 28 20 25 25 NUM bracis-33661 28 21 ] ] PUNCT bracis-33661 28 22 , , PUNCT bracis-33661 28 23 and and CCONJ bracis-33661 28 24 even even ADV bracis-33661 28 25 the the DET bracis-33661 28 26 relationship relationship NOUN bracis-33661 28 27 between between ADP bracis-33661 28 28 these these DET bracis-33661 28 29 musical musical ADJ bracis-33661 28 30 stimuli stimulus NOUN bracis-33661 28 31 with with ADP bracis-33661 28 32 the the DET bracis-33661 28 33 behavioral behavioral ADJ bracis-33661 28 34 recovery recovery NOUN bracis-33661 28 35 of of ADP bracis-33661 28 36 comatose comatose ADJ bracis-33661 28 37 patients patient NOUN bracis-33661 28 38 [ [ X bracis-33661 28 39 7 7 NUM bracis-33661 28 40 ] ] PUNCT bracis-33661 28 41 . . PUNCT bracis-33661 29 1 research research NOUN bracis-33661 29 2 on on ADP bracis-33661 29 3 the the DET bracis-33661 29 4 therapeutic therapeutic ADJ bracis-33661 29 5 potential potential NOUN bracis-33661 29 6 of of ADP bracis-33661 29 7 musical musical ADJ bracis-33661 29 8 stimulation stimulation NOUN bracis-33661 29 9 in in ADP bracis-33661 29 10 coma coma NOUN bracis-33661 29 11 patient patient NOUN bracis-33661 29 12 recovery recovery NOUN bracis-33661 29 13 often often ADV bracis-33661 29 14 employs employ VERB bracis-33661 29 15 a a DET bracis-33661 29 16 comparative comparative ADJ bracis-33661 29 17 approach approach NOUN bracis-33661 29 18 by by ADP bracis-33661 29 19 analyzing analyze VERB bracis-33661 29 20 the the DET bracis-33661 29 21 temporal temporal ADJ bracis-33661 29 22 signal signal NOUN bracis-33661 29 23 from from ADP bracis-33661 29 24 each each DET bracis-33661 29 25 electrode electrode NOUN bracis-33661 29 26 before before ADP bracis-33661 29 27 and and CCONJ bracis-33661 29 28 after after ADP bracis-33661 29 29 the the DET bracis-33661 29 30 stimulus stimulus NOUN bracis-33661 29 31 . . PUNCT bracis-33661 30 1 these these DET bracis-33661 30 2 comparisons comparison NOUN bracis-33661 30 3 involve involve VERB bracis-33661 30 4 the the DET bracis-33661 30 5 extraction extraction NOUN bracis-33661 30 6 of of ADP bracis-33661 30 7 statistical statistical ADJ bracis-33661 30 8 variables variable NOUN bracis-33661 30 9 from from ADP bracis-33661 30 10 the the DET bracis-33661 30 11 time time NOUN bracis-33661 30 12 series series PROPN bracis-33661 30 13 data data PROPN bracis-33661 30 14 . . PUNCT bracis-33661 31 1 additionally additionally ADV bracis-33661 31 2 , , PUNCT bracis-33661 31 3 there there PRON bracis-33661 31 4 are be VERB bracis-33661 31 5 applications application NOUN bracis-33661 31 6 of of ADP bracis-33661 31 7 ml ml NOUN bracis-33661 31 8 models model NOUN bracis-33661 31 9 to to PART bracis-33661 31 10 comprehensively comprehensively ADV bracis-33661 31 11 assess assess VERB bracis-33661 31 12 the the DET bracis-33661 31 13 impact impact NOUN bracis-33661 31 14 generated generate VERB bracis-33661 31 15 by by ADP bracis-33661 31 16 such such ADJ bracis-33661 31 17 stimuli stimulus NOUN bracis-33661 31 18 on on ADP bracis-33661 31 19 patients patient NOUN bracis-33661 31 20 ’ ' PUNCT bracis-33661 31 21 neurological neurological ADJ bracis-33661 31 22 responses response NOUN bracis-33661 31 23 and and CCONJ bracis-33661 31 24 recovery recovery NOUN bracis-33661 31 25 trajectories trajectory NOUN bracis-33661 31 26 [ [ X bracis-33661 31 27 23 23 NUM bracis-33661 31 28 ] ] PUNCT bracis-33661 31 29 . . PUNCT bracis-33661 32 1 most most ADJ bracis-33661 32 2 approaches approach NOUN bracis-33661 32 3 to to ADP bracis-33661 32 4 eeg eeg NOUN bracis-33661 32 5 analysis analysis NOUN bracis-33661 32 6 involve involve VERB bracis-33661 32 7 extracting extract VERB bracis-33661 32 8 information information NOUN bracis-33661 32 9 from from ADP bracis-33661 32 10 descriptive descriptive ADJ bracis-33661 32 11 statistical statistical ADJ bracis-33661 32 12 variables variable NOUN bracis-33661 32 13 or or CCONJ bracis-33661 32 14 using use VERB bracis-33661 32 15 simple simple ADJ bracis-33661 32 16 signal signal NOUN bracis-33661 32 17 pre pre ADJ bracis-33661 32 18 - - ADJ bracis-33661 32 19 processing processing ADJ bracis-33661 32 20 methods method NOUN bracis-33661 32 21 [ [ X bracis-33661 32 22 18 18 NUM bracis-33661 32 23 , , PUNCT bracis-33661 32 24 19 19 NUM bracis-33661 32 25 ] ] PUNCT bracis-33661 32 26 . . PUNCT bracis-33661 33 1 another another DET bracis-33661 33 2 well well ADV bracis-33661 33 3 - - PUNCT bracis-33661 33 4 known know VERB bracis-33661 33 5 approaches approach NOUN bracis-33661 33 6 include include VERB bracis-33661 33 7 recurrent recurrent ADJ bracis-33661 33 8 neural neural ADJ bracis-33661 33 9 networks network NOUN bracis-33661 33 10 ( ( PUNCT bracis-33661 33 11 rnn rnn PROPN bracis-33661 33 12 ) ) PUNCT bracis-33661 33 13 , , PUNCT bracis-33661 33 14 such such ADJ bracis-33661 33 15 as as ADP bracis-33661 33 16 echo echo NOUN bracis-33661 33 17 state state NOUN bracis-33661 33 18 networks network NOUN bracis-33661 33 19 and and CCONJ bracis-33661 33 20 long long ADJ bracis-33661 33 21 short short ADJ bracis-33661 33 22 - - PUNCT bracis-33661 33 23 term term NOUN bracis-33661 33 24 memory memory NOUN bracis-33661 33 25 ( ( PUNCT bracis-33661 33 26 lstm lstm NOUN bracis-33661 33 27 ) ) PUNCT bracis-33661 34 1 [ [ X bracis-33661 34 2 1 1 NUM bracis-33661 34 3 , , PUNCT bracis-33661 34 4 3 3 NUM bracis-33661 34 5 , , PUNCT bracis-33661 34 6 4 4 NUM bracis-33661 34 7 , , PUNCT bracis-33661 34 8 21 21 NUM bracis-33661 34 9 , , PUNCT bracis-33661 34 10 25 25 NUM bracis-33661 34 11 , , PUNCT bracis-33661 34 12 32 32 NUM bracis-33661 34 13 , , PUNCT bracis-33661 34 14 33 33 NUM bracis-33661 34 15 ] ] PUNCT bracis-33661 34 16 . . PUNCT bracis-33661 35 1 lstm lstm NOUN bracis-33661 35 2 is be AUX bracis-33661 35 3 particularly particularly ADV bracis-33661 35 4 effective effective ADJ bracis-33661 35 5 for for ADP bracis-33661 35 6 handling handle VERB bracis-33661 35 7 temporal temporal ADJ bracis-33661 35 8 data datum NOUN bracis-33661 35 9 and and CCONJ bracis-33661 35 10 managing manage VERB bracis-33661 35 11 relationships relationship NOUN bracis-33661 35 12 between between ADP bracis-33661 35 13 states state NOUN bracis-33661 35 14 . . PUNCT bracis-33661 36 1 in in ADP bracis-33661 36 2 eeg eeg NOUN bracis-33661 36 3 analysis analysis NOUN bracis-33661 36 4 tasks task NOUN bracis-33661 36 5 , , PUNCT bracis-33661 36 6 lstm lstm NOUN bracis-33661 36 7 often often ADV bracis-33661 36 8 includes include VERB bracis-33661 36 9 a a DET bracis-33661 36 10 regularization regularization NOUN bracis-33661 36 11 layer layer NOUN bracis-33661 36 12 during during ADP bracis-33661 36 13 the the DET bracis-33661 36 14 training training NOUN bracis-33661 36 15 process process NOUN bracis-33661 36 16 . . PUNCT bracis-33661 37 1 this this DET bracis-33661 37 2 architectural architectural ADJ bracis-33661 37 3 choice choice NOUN bracis-33661 37 4 enhances enhance VERB bracis-33661 37 5 the the DET bracis-33661 37 6 model model NOUN bracis-33661 37 7 ’s ’s PART bracis-33661 37 8 ability ability NOUN bracis-33661 37 9 to to PART bracis-33661 37 10 accurately accurately ADV bracis-33661 37 11 interpret interpret VERB bracis-33661 37 12 and and CCONJ bracis-33661 37 13 analyze analyze VERB bracis-33661 37 14 the the DET bracis-33661 37 15 examination examination NOUN bracis-33661 37 16 results result NOUN bracis-33661 37 17 . . PUNCT bracis-33661 38 1 lstms lstms ADJ bracis-33661 38 2 yield yield NOUN bracis-33661 38 3 satisfactory satisfactory ADJ bracis-33661 38 4 results result NOUN bracis-33661 38 5 when when SCONJ bracis-33661 38 6 applied apply VERB bracis-33661 38 7 to to ADP bracis-33661 38 8 eeg eeg NOUN bracis-33661 38 9 exams exam NOUN bracis-33661 38 10 for for ADP bracis-33661 38 11 the the DET bracis-33661 38 12 aforementioned aforementioned ADJ bracis-33661 38 13 applications application NOUN bracis-33661 38 14 . . PUNCT bracis-33661 39 1 however however ADV bracis-33661 39 2 , , PUNCT bracis-33661 39 3 they they PRON bracis-33661 39 4 miss miss VERB bracis-33661 39 5 a a DET bracis-33661 39 6 significant significant ADJ bracis-33661 39 7 opportunity opportunity NOUN bracis-33661 39 8 to to PART bracis-33661 39 9 explore explore VERB bracis-33661 39 10 the the DET bracis-33661 39 11 scalp scalp NOUN bracis-33661 39 12 position position NOUN bracis-33661 39 13 of of ADP bracis-33661 39 14 each each DET bracis-33661 39 15 electrode electrode NOUN bracis-33661 39 16 and and CCONJ bracis-33661 39 17 the the DET bracis-33661 39 18 relationships relationship NOUN bracis-33661 39 19 between between ADP bracis-33661 39 20 channels channel NOUN bracis-33661 39 21 . . PUNCT bracis-33661 40 1 graph graph VERB bracis-33661 40 2 neural neural ADJ bracis-33661 40 3 networks network NOUN bracis-33661 40 4 ( ( PUNCT bracis-33661 40 5 gnns gnns NOUN bracis-33661 40 6 ) ) PUNCT bracis-33661 40 7 have have VERB bracis-33661 40 8 the the DET bracis-33661 40 9 potential potential NOUN bracis-33661 40 10 to to PART bracis-33661 40 11 work work VERB bracis-33661 40 12 with with ADP bracis-33661 40 13 graph graph NOUN bracis-33661 40 14 - - PUNCT bracis-33661 40 15 structured structure VERB bracis-33661 40 16 data datum NOUN bracis-33661 40 17 . . PUNCT bracis-33661 41 1 recent recent ADJ bracis-33661 41 2 works work NOUN bracis-33661 41 3 on on ADP bracis-33661 41 4 applying apply VERB bracis-33661 41 5 gnns gnn NOUN bracis-33661 41 6 to to ADP bracis-33661 41 7 emotion emotion NOUN bracis-33661 41 8 prediction prediction NOUN bracis-33661 41 9 and and CCONJ bracis-33661 41 10 other other ADJ bracis-33661 41 11 general general ADJ bracis-33661 41 12 eeg eeg NOUN bracis-33661 41 13 classification classification NOUN bracis-33661 41 14 tasks task NOUN bracis-33661 41 15 [ [ X bracis-33661 41 16 8 8 NUM bracis-33661 41 17 , , PUNCT bracis-33661 41 18 34 34 NUM bracis-33661 41 19 ] ] PUNCT bracis-33661 41 20 have have AUX bracis-33661 41 21 obtained obtain VERB bracis-33661 41 22 promising promising ADJ bracis-33661 41 23 results result NOUN bracis-33661 41 24 . . PUNCT bracis-33661 42 1 besides besides SCONJ bracis-33661 42 2 such such ADJ bracis-33661 42 3 positive positive ADJ bracis-33661 42 4 outcomes outcome NOUN bracis-33661 42 5 , , PUNCT bracis-33661 42 6 gnns gnns PROPN bracis-33661 42 7 can can AUX bracis-33661 42 8 also also ADV bracis-33661 42 9 explore explore VERB bracis-33661 42 10 connections connection NOUN bracis-33661 42 11 involving involve VERB bracis-33661 42 12 the the DET bracis-33661 42 13 spatial spatial ADJ bracis-33661 42 14 positioning positioning NOUN bracis-33661 42 15 of of ADP bracis-33661 42 16 the the DET bracis-33661 42 17 electrodes electrode NOUN bracis-33661 42 18 ( ( PUNCT bracis-33661 42 19 nodes node NOUN bracis-33661 42 20 ) ) PUNCT bracis-33661 42 21 and and CCONJ bracis-33661 42 22 the the DET bracis-33661 42 23 correlations correlation NOUN bracis-33661 42 24 between between ADP bracis-33661 42 25 the the DET bracis-33661 42 26 graph graph NOUN bracis-33661 42 27 data data NOUN bracis-33661 42 28 features feature NOUN bracis-33661 42 29 , , PUNCT bracis-33661 42 30 which which PRON bracis-33661 42 31 is be AUX bracis-33661 42 32 a a DET bracis-33661 42 33 few few ADJ bracis-33661 42 34 explored explore VERB bracis-33661 42 35 topic topic NOUN bracis-33661 42 36 in in ADP bracis-33661 42 37 the the DET bracis-33661 42 38 literature literature NOUN bracis-33661 42 39 . . PUNCT bracis-33661 43 1 in in ADP bracis-33661 43 2 this this DET bracis-33661 43 3 work work NOUN bracis-33661 43 4 , , PUNCT bracis-33661 43 5 we we PRON bracis-33661 43 6 investigate investigate VERB bracis-33661 43 7 four four NUM bracis-33661 43 8 gnns gnns ADJ bracis-33661 43 9 architectures architecture NOUN bracis-33661 43 10 for for ADP bracis-33661 43 11 the the DET bracis-33661 43 12 prognosis prognosis NOUN bracis-33661 43 13 of of ADP bracis-33661 43 14 patients patient NOUN bracis-33661 43 15 in in ADP bracis-33661 43 16 coma coma NOUN bracis-33661 43 17 ( ( PUNCT bracis-33661 43 18 ppc ppc NOUN bracis-33661 43 19 ) ) PUNCT bracis-33661 43 20 using use VERB bracis-33661 43 21 eeg eeg PROPN bracis-33661 43 22 data datum NOUN bracis-33661 43 23 . . PUNCT bracis-33661 44 1 our our PRON bracis-33661 44 2 graph graph NOUN bracis-33661 44 3 classification classification NOUN bracis-33661 44 4 approach approach NOUN bracis-33661 44 5 incorporates incorporate VERB bracis-33661 44 6 convolutional convolutional ADJ bracis-33661 44 7 - - PUNCT bracis-33661 44 8 based base VERB bracis-33661 44 9 layers layer NOUN bracis-33661 44 10 specifically specifically ADV bracis-33661 44 11 designed design VERB bracis-33661 44 12 for for ADP bracis-33661 44 13 graph graph NOUN bracis-33661 44 14 data datum NOUN bracis-33661 44 15 , , PUNCT bracis-33661 44 16 allowing allow VERB bracis-33661 44 17 the the DET bracis-33661 44 18 exploration exploration NOUN bracis-33661 44 19     SPACE bracis-33661 44 20 of of ADP bracis-33661 44 21 spatial spatial ADJ bracis-33661 44 22 relationships relationship NOUN bracis-33661 44 23 between between ADP bracis-33661 44 24 electrodes electrode NOUN bracis-33661 44 25 due due ADJ bracis-33661 44 26 to to ADP bracis-33661 44 27 the the DET bracis-33661 44 28 representation representation NOUN bracis-33661 44 29 of of ADP bracis-33661 44 30 the the DET bracis-33661 44 31 eeg eeg NOUN bracis-33661 44 32 examination examination NOUN bracis-33661 44 33 in in ADP bracis-33661 44 34 the the DET bracis-33661 44 35 form form NOUN bracis-33661 44 36 of of ADP bracis-33661 44 37 a a DET bracis-33661 44 38 graph graph NOUN bracis-33661 44 39 structure structure NOUN bracis-33661 44 40 . . PUNCT bracis-33661 45 1 in in ADP bracis-33661 45 2 order order NOUN bracis-33661 45 3 to to PART bracis-33661 45 4 represent represent VERB bracis-33661 45 5 eeg eeg PROPN bracis-33661 45 6 sequential sequential ADJ bracis-33661 45 7 data datum NOUN bracis-33661 45 8 into into ADP bracis-33661 45 9 graphs graph NOUN bracis-33661 45 10 , , PUNCT bracis-33661 45 11 we we PRON bracis-33661 45 12 also also ADV bracis-33661 45 13 develop develop VERB bracis-33661 45 14 eegraph eegraph NOUN bracis-33661 45 15 , , PUNCT bracis-33661 45 16 a a DET bracis-33661 45 17 modeling modeling NOUN bracis-33661 45 18 strategy strategy NOUN bracis-33661 45 19 in in ADP bracis-33661 45 20 which which PRON bracis-33661 45 21 each each DET bracis-33661 45 22 vertex vertex NOUN bracis-33661 45 23 in in ADP bracis-33661 45 24 the the DET bracis-33661 45 25 graph graph NOUN bracis-33661 45 26 corresponds correspond VERB bracis-33661 45 27 to to ADP bracis-33661 45 28 an an DET bracis-33661 45 29 electrode electrode NOUN bracis-33661 45 30 , , PUNCT bracis-33661 45 31 and and CCONJ bracis-33661 45 32 the the DET bracis-33661 45 33 edges edge NOUN bracis-33661 45 34 are be AUX bracis-33661 45 35 constructed construct VERB bracis-33661 45 36 from from ADP bracis-33661 45 37 the the DET bracis-33661 45 38 electrodes electrode NOUN bracis-33661 45 39 position position NOUN bracis-33661 45 40 in in ADP bracis-33661 45 41 the the DET bracis-33661 45 42 scalp scalp NOUN bracis-33661 45 43 . . PUNCT bracis-33661 46 1 while while SCONJ bracis-33661 46 2 the the DET bracis-33661 46 3 ppc ppc ADJ bracis-33661 46 4 task task NOUN bracis-33661 46 5 is be AUX bracis-33661 46 6 continuously continuously ADV bracis-33661 46 7 evolving evolve VERB bracis-33661 46 8 , , PUNCT bracis-33661 46 9 we we PRON bracis-33661 46 10 recognize recognize VERB bracis-33661 46 11 certain certain ADJ bracis-33661 46 12 opportunities opportunity NOUN bracis-33661 46 13 for for ADP bracis-33661 46 14 improvement improvement NOUN bracis-33661 46 15 . . PUNCT bracis-33661 47 1 for for ADP bracis-33661 47 2 example example NOUN bracis-33661 47 3 , , PUNCT bracis-33661 47 4 the the DET bracis-33661 47 5 topological topological ADJ bracis-33661 47 6 structure structure NOUN bracis-33661 47 7 of of ADP bracis-33661 47 8 eeg eeg NOUN bracis-33661 47 9 channels channel NOUN bracis-33661 47 10 is be AUX bracis-33661 47 11 often often ADV bracis-33661 47 12 underutilized underutilize VERB bracis-33661 47 13 in in ADP bracis-33661 47 14 the the DET bracis-33661 47 15 formulation formulation NOUN bracis-33661 47 16 of of ADP bracis-33661 47 17 eeg eeg NOUN bracis-33661 47 18 representations representation NOUN bracis-33661 47 19 [ [ X bracis-33661 47 20 2 2 NUM bracis-33661 47 21 ] ] PUNCT bracis-33661 47 22 . . PUNCT bracis-33661 48 1 to to PART bracis-33661 48 2 address address VERB bracis-33661 48 3 this this PRON bracis-33661 48 4 , , PUNCT bracis-33661 48 5 our our PRON bracis-33661 48 6 study study NOUN bracis-33661 48 7 focuses focus VERB bracis-33661 48 8 on on ADP bracis-33661 48 9 leveraging leverage VERB bracis-33661 48 10 the the DET bracis-33661 48 11 graph graph NOUN bracis-33661 48 12 structure structure NOUN bracis-33661 48 13 to to PART bracis-33661 48 14 uncover uncover VERB bracis-33661 48 15 more more ADV bracis-33661 48 16 informative informative ADJ bracis-33661 48 17 patterns pattern NOUN bracis-33661 48 18 and and CCONJ bracis-33661 48 19 relationships relationship NOUN bracis-33661 48 20 among among ADP bracis-33661 48 21 the the DET bracis-33661 48 22 electrodes electrode NOUN bracis-33661 48 23 , , PUNCT bracis-33661 48 24 leading lead VERB bracis-33661 48 25 to to ADP bracis-33661 48 26 enhanced enhance VERB bracis-33661 48 27 prognostic prognostic ADJ bracis-33661 48 28 accuracy accuracy NOUN bracis-33661 48 29 . . PUNCT bracis-33661 49 1 by by ADP bracis-33661 49 2 addressing address VERB bracis-33661 49 3 such such DET bracis-33661 49 4 a a DET bracis-33661 49 5 challenge challenge NOUN bracis-33661 49 6 , , PUNCT bracis-33661 49 7 we we PRON bracis-33661 49 8 aim aim VERB bracis-33661 49 9 to to PART bracis-33661 49 10 contribute contribute VERB bracis-33661 49 11 in in ADP bracis-33661 49 12 the the DET bracis-33661 49 13 ppc ppc ADJ bracis-33661 49 14 task task NOUN bracis-33661 49 15 from from ADP bracis-33661 49 16 eeg eeg NOUN bracis-33661 49 17 signals signal NOUN bracis-33661 49 18 , , PUNCT bracis-33661 49 19 an an DET bracis-33661 49 20 approach approach NOUN bracis-33661 49 21 compliant compliant NOUN bracis-33661 49 22 with with ADP bracis-33661 49 23 most most ADJ bracis-33661 49 24 of of ADP bracis-33661 49 25 the the DET bracis-33661 49 26 brazilian brazilian ADJ bracis-33661 49 27 public public ADJ bracis-33661 49 28 hospitals hospital NOUN bracis-33661 49 29 in in ADP bracis-33661 49 30 terms term NOUN bracis-33661 49 31 of of ADP bracis-33661 49 32 cost cost NOUN bracis-33661 49 33 and and CCONJ bracis-33661 49 34 equipment equipment NOUN bracis-33661 49 35 availability availability NOUN bracis-33661 49 36 . . PUNCT bracis-33661 50 1 our our PRON bracis-33661 50 2 experiments experiment NOUN bracis-33661 50 3 with with ADP bracis-33661 50 4 real real ADJ bracis-33661 50 5 data datum NOUN bracis-33661 50 6 shows show VERB bracis-33661 50 7 that that SCONJ bracis-33661 50 8 both both CCONJ bracis-33661 50 9 the the DET bracis-33661 50 10 proposed propose VERB bracis-33661 50 11 gnns gnns PROPN bracis-33661 50 12 architectures architecture NOUN bracis-33661 50 13 and and CCONJ bracis-33661 50 14 the the DET bracis-33661 50 15 graph graph NOUN bracis-33661 50 16 modeling model VERB bracis-33661 50 17 strategy strategy NOUN bracis-33661 50 18 improve improve VERB bracis-33661 50 19 the the DET bracis-33661 50 20 predictive predictive ADJ bracis-33661 50 21 results result NOUN bracis-33661 50 22 regarding regard VERB bracis-33661 50 23 the the DET bracis-33661 50 24 outcome outcome NOUN bracis-33661 50 25 of of ADP bracis-33661 50 26 comatose comatose ADJ bracis-33661 50 27 patients patient NOUN bracis-33661 50 28 . . PUNCT bracis-33661 51 1 in in ADP bracis-33661 51 2 summary summary NOUN bracis-33661 51 3 , , PUNCT bracis-33661 51 4 our our PRON bracis-33661 51 5 main main ADJ bracis-33661 51 6 contributions contribution NOUN bracis-33661 51 7 can can AUX bracis-33661 51 8 be be AUX bracis-33661 51 9 stated state VERB bracis-33661 51 10 as as SCONJ bracis-33661 51 11 follows follow VERB bracis-33661 51 12 : : PUNCT bracis-33661 51 13 development development NOUN bracis-33661 51 14 of of ADP bracis-33661 51 15 four four NUM bracis-33661 51 16 gnns gnns ADJ bracis-33661 51 17 architectures architecture NOUN bracis-33661 51 18 based base VERB bracis-33661 51 19 on on ADP bracis-33661 51 20 convolutional convolutional ADJ bracis-33661 51 21 dynamics dynamic NOUN bracis-33661 51 22 to to PART bracis-33661 51 23 predict predict VERB bracis-33661 51 24 ppc ppc ADJ bracis-33661 51 25 outcomes outcome NOUN bracis-33661 51 26 . . PUNCT bracis-33661 52 1 the the DET bracis-33661 52 2 proposed propose VERB bracis-33661 52 3 architectures architecture NOUN bracis-33661 52 4 aim aim VERB bracis-33661 52 5 to to PART bracis-33661 52 6 explore explore VERB bracis-33661 52 7 both both DET bracis-33661 52 8 nonlinear nonlinear ADJ bracis-33661 52 9 transformations transformation NOUN bracis-33661 52 10 and and CCONJ bracis-33661 52 11 local local ADJ bracis-33661 52 12 connections connection NOUN bracis-33661 52 13 of of ADP bracis-33661 52 14 the the DET bracis-33661 52 15 vertices vertex NOUN bracis-33661 52 16 . . PUNCT bracis-33661 53 1 design design NOUN bracis-33661 53 2 of of ADP bracis-33661 53 3 eegraph eegraph NOUN bracis-33661 53 4 , , PUNCT bracis-33661 53 5 a a DET bracis-33661 53 6 modeling modeling NOUN bracis-33661 53 7 strategy strategy NOUN bracis-33661 53 8 to to PART bracis-33661 53 9 transform transform VERB bracis-33661 53 10 raw raw ADJ bracis-33661 53 11 eeg eeg NOUN bracis-33661 53 12 data datum NOUN bracis-33661 53 13 into into ADP bracis-33661 53 14 graph graph NOUN bracis-33661 53 15 structures structure NOUN bracis-33661 53 16 . . PUNCT bracis-33661 54 1 eegraph eegraph PROPN bracis-33661 54 2 seeks seek VERB bracis-33661 54 3 to to PART bracis-33661 54 4 exploit exploit VERB bracis-33661 54 5 the the DET bracis-33661 54 6 connections connection NOUN bracis-33661 54 7 between between ADP bracis-33661 54 8 the the DET bracis-33661 54 9 vertices vertex NOUN bracis-33661 54 10 ( ( PUNCT bracis-33661 54 11 electrodes electrode NOUN bracis-33661 54 12 ) ) PUNCT bracis-33661 54 13 of of ADP bracis-33661 54 14 graph graph NOUN bracis-33661 54 15 structures structure NOUN bracis-33661 54 16 as as ADV bracis-33661 54 17 much much ADV bracis-33661 54 18 as as ADP bracis-33661 54 19 possible possible ADJ bracis-33661 54 20 . . PUNCT bracis-33661 55 1 development development NOUN bracis-33661 55 2 of of ADP bracis-33661 55 3 an an DET bracis-33661 55 4 experimental experimental ADJ bracis-33661 55 5 setup setup NOUN bracis-33661 55 6 to to PART bracis-33661 55 7 compare compare VERB bracis-33661 55 8 the the DET bracis-33661 55 9 proposed propose VERB bracis-33661 55 10 architectures architecture NOUN bracis-33661 55 11 against against ADP bracis-33661 55 12 lstm lstm NOUN bracis-33661 55 13 , , PUNCT bracis-33661 55 14 and and CCONJ bracis-33661 55 15 provide provide VERB bracis-33661 55 16 some some DET bracis-33661 55 17 evidence evidence NOUN bracis-33661 55 18 about about ADP bracis-33661 55 19 their their PRON bracis-33661 55 20 satisfactory satisfactory ADJ bracis-33661 55 21 results result NOUN bracis-33661 55 22 . . PUNCT bracis-33661 56 1 the the DET bracis-33661 56 2 remainder remainder NOUN bracis-33661 56 3 work work NOUN bracis-33661 56 4 is be AUX bracis-33661 56 5 divided divide VERB bracis-33661 56 6 as as SCONJ bracis-33661 56 7 follows follow VERB bracis-33661 56 8 : : PUNCT bracis-33661 56 9 sect sect NOUN bracis-33661 56 10 . . PUNCT bracis-33661 56 11     SPACE bracis-33661 57 1 2 2 NUM bracis-33661 57 2 presents present VERB bracis-33661 57 3 the the DET bracis-33661 57 4 background background NOUN bracis-33661 57 5 for for ADP bracis-33661 57 6 the the DET bracis-33661 57 7 understanding understanding NOUN bracis-33661 57 8 of of ADP bracis-33661 57 9 the the DET bracis-33661 57 10 proposed propose VERB bracis-33661 57 11 architectures architecture NOUN bracis-33661 57 12 . . PUNCT bracis-33661 58 1 section section NOUN bracis-33661 58 2     SPACE bracis-33661 58 3 3 3 NUM bracis-33661 58 4 describes describe VERB bracis-33661 58 5 the the DET bracis-33661 58 6 proposed propose VERB bracis-33661 58 7 approaches approach NOUN bracis-33661 58 8 . . PUNCT bracis-33661 59 1 section section NOUN bracis-33661 59 2     SPACE bracis-33661 59 3 4 4 NUM bracis-33661 59 4 shows show VERB bracis-33661 59 5 the the DET bracis-33661 59 6 experimental experimental ADJ bracis-33661 59 7 results result NOUN bracis-33661 59 8 , , PUNCT bracis-33661 59 9 and and CCONJ bracis-33661 59 10 sect sect NOUN bracis-33661 59 11 . . PUNCT bracis-33661 59 12     SPACE bracis-33661 60 1 5 5 NUM bracis-33661 60 2 brings bring VERB bracis-33661 60 3 the the DET bracis-33661 60 4 conclusion conclusion NOUN bracis-33661 60 5 of of ADP bracis-33661 60 6 the the DET bracis-33661 60 7 work work NOUN bracis-33661 60 8 as as ADV bracis-33661 60 9 well well ADV bracis-33661 60 10 as as ADP bracis-33661 60 11 forthcoming forthcoming ADJ bracis-33661 60 12 investigations investigation NOUN bracis-33661 60 13 . . PUNCT bracis-33661 61 1 2 2 NUM bracis-33661 61 2 background background NOUN bracis-33661 61 3 this this DET bracis-33661 61 4 section section NOUN bracis-33661 61 5 presents present VERB bracis-33661 61 6 a a DET bracis-33661 61 7 brief brief ADJ bracis-33661 61 8 description description NOUN bracis-33661 61 9 of of ADP bracis-33661 61 10 the the DET bracis-33661 61 11 mathematical mathematical ADJ bracis-33661 61 12 concepts concept NOUN bracis-33661 61 13 related relate VERB bracis-33661 61 14 to to ADP bracis-33661 61 15 the the DET bracis-33661 61 16 gnn gnn PROPN bracis-33661 61 17 and and CCONJ bracis-33661 61 18 lstm lstm NOUN bracis-33661 61 19 layers layer NOUN bracis-33661 61 20 investigated investigate VERB bracis-33661 61 21 in in ADP bracis-33661 61 22 this this DET bracis-33661 61 23 paper paper NOUN bracis-33661 61 24 . . PUNCT bracis-33661 62 1 2.1 2.1 NUM bracis-33661 62 2 graph graph NOUN bracis-33661 62 3 neural neural ADJ bracis-33661 62 4 networks network NOUN bracis-33661 62 5 for for ADP bracis-33661 62 6 a a DET bracis-33661 62 7 better well ADJ bracis-33661 62 8 understanding understanding NOUN bracis-33661 62 9 of of ADP bracis-33661 62 10 gnns gnns NOUN bracis-33661 62 11 and and CCONJ bracis-33661 62 12 their their PRON bracis-33661 62 13 respective respective ADJ bracis-33661 62 14 convolutional convolutional ADJ bracis-33661 62 15 layers layer NOUN bracis-33661 62 16 used use VERB bracis-33661 62 17 in in ADP bracis-33661 62 18 this this DET bracis-33661 62 19 work work NOUN bracis-33661 62 20 , , PUNCT bracis-33661 62 21 it it PRON bracis-33661 62 22 is be AUX bracis-33661 62 23 important important ADJ bracis-33661 62 24 to to PART bracis-33661 62 25 reiterate reiterate VERB bracis-33661 62 26 the the DET bracis-33661 62 27 graph graph NOUN bracis-33661 62 28 data data NOUN bracis-33661 62 29 structure structure NOUN bracis-33661 62 30 definition definition NOUN bracis-33661 62 31 in in ADP bracis-33661 62 32 which which PRON bracis-33661 62 33 these these DET bracis-33661 62 34 networks network NOUN bracis-33661 62 35 operate operate VERB bracis-33661 62 36 . . PUNCT bracis-33661 63 1 a a DET bracis-33661 63 2 graph graph NOUN bracis-33661 63 3 \ \ PROPN bracis-33661 63 4 ( ( PUNCT bracis-33661 63 5 \mathcal \mathcal ADJ bracis-33661 63 6 { { PUNCT bracis-33661 63 7 g g NOUN bracis-33661 63 8 = = SYM bracis-33661 63 9 ( ( PUNCT bracis-33661 63 10 v v NOUN bracis-33661 63 11 , , PUNCT bracis-33661 63 12 e e NOUN bracis-33661 63 13 ) ) PUNCT bracis-33661 63 14 } } PUNCT bracis-33661 63 15 \ \ NOUN bracis-33661 63 16 ) ) PUNCT bracis-33661 63 17 is be AUX bracis-33661 63 18 the the DET bracis-33661 63 19 data data NOUN bracis-33661 63 20 representation representation NOUN bracis-33661 63 21 formed form VERB bracis-33661 63 22 by by ADP bracis-33661 63 23 connecting connect VERB bracis-33661 63 24 vertices vertex NOUN bracis-33661 63 25 through through ADP bracis-33661 63 26 edges edge NOUN bracis-33661 63 27 , , PUNCT bracis-33661 63 28 which which PRON bracis-33661 63 29 are be AUX bracis-33661 63 30 denoted denote VERB bracis-33661 63 31 respectively respectively ADV bracis-33661 63 32 by by ADP bracis-33661 63 33 \ \ PROPN bracis-33661 63 34 ( ( PUNCT bracis-33661 63 35 \mathcal \mathcal PROPN bracis-33661 63 36 { { PUNCT bracis-33661 63 37 v v NOUN bracis-33661 63 38 } } PUNCT bracis-33661 63 39 \ \ NOUN bracis-33661 63 40 ) ) PUNCT bracis-33661 63 41 and and CCONJ bracis-33661 63 42 \ \ PROPN bracis-33661 63 43 ( ( PUNCT bracis-33661 63 44 \mathcal \mathcal X bracis-33661 63 45 { { PUNCT bracis-33661 63 46 e e NOUN bracis-33661 63 47 } } PUNCT bracis-33661 63 48 \ \ NOUN bracis-33661 63 49 ) ) PUNCT bracis-33661 63 50 . . PUNCT bracis-33661 64 1 each each DET bracis-33661 64 2 edge edge NOUN bracis-33661 64 3 can can AUX bracis-33661 64 4 contain contain VERB bracis-33661 64 5 a a DET bracis-33661 64 6 weight weight NOUN bracis-33661 64 7 value value NOUN bracis-33661 64 8 \ \ PROPN bracis-33661 64 9 ( ( PUNCT bracis-33661 64 10 \mathcal \mathcal PROPN bracis-33661 64 11 { { PUNCT bracis-33661 64 12 w w PROPN bracis-33661 64 13 } } PUNCT bracis-33661 64 14 \ \ NOUN bracis-33661 64 15 ) ) PUNCT bracis-33661 64 16 . . PUNCT bracis-33661 65 1 an an DET bracis-33661 65 2 adjacency adjacency NOUN bracis-33661 65 3 matrix matrix NOUN bracis-33661 65 4 \ \ PROPN bracis-33661 65 5 ( ( PUNCT bracis-33661 65 6 \mathcal \mathcal PROPN bracis-33661 65 7 { { PUNCT bracis-33661 65 8 a a PRON bracis-33661 65 9 } } PUNCT bracis-33661 65 10 \ \ NOUN bracis-33661 65 11 ) ) PUNCT bracis-33661 65 12 can can AUX bracis-33661 65 13 be be AUX bracis-33661 65 14 used use VERB bracis-33661 65 15 to to PART bracis-33661 65 16 represent represent VERB bracis-33661 65 17 an an DET bracis-33661 65 18 overview overview NOUN bracis-33661 65 19 of of ADP bracis-33661 65 20 the the DET bracis-33661 65 21 final final ADJ bracis-33661 65 22 structuring structuring NOUN bracis-33661 65 23 of of ADP bracis-33661 65 24 the the DET bracis-33661 65 25 graph graph NOUN bracis-33661 65 26 . . PUNCT bracis-33661 66 1 graph graph VERB bracis-33661 66 2 convolutional convolutional ADJ bracis-33661 66 3 network network NOUN bracis-33661 66 4 ( ( PUNCT bracis-33661 66 5 gcn gcn NOUN bracis-33661 66 6 ) ) PUNCT bracis-33661 66 7 have have VERB bracis-33661 66 8 their their PRON bracis-33661 66 9 layer layer NOUN bracis-33661 66 10 propagation propagation NOUN bracis-33661 66 11 behavior behavior NOUN bracis-33661 66 12 governed govern VERB bracis-33661 66 13 by by ADP bracis-33661 66 14 the the DET bracis-33661 66 15 following follow VERB bracis-33661 66 16 equation equation NOUN bracis-33661 66 17 [ [ X bracis-33661 66 18 16 16 NUM bracis-33661 66 19 ] ] X bracis-33661 66 20 : : PUNCT bracis-33661 66 21 $ $ SYM bracis-33661 66 22 $ $ SYM bracis-33661 66 23 \begin{aligned \begin{aligne VERB bracis-33661 66 24 } } PUNCT bracis-33661 66 25 h^{(l h^{(l NOUN bracis-33661 66 26 + + X bracis-33661 66 27 1 1 NUM bracis-33661 66 28 ) ) PUNCT bracis-33661 66 29 } } PUNCT bracis-33661 67 1 = = SYM bracis-33661 67 2 \sigma \sigma PROPN bracis-33661 67 3 ( ( PUNCT bracis-33661 67 4 \tilde{d}^{-\frac{1}{2 \tilde{d}^{-\frac{1}{2 PROPN bracis-33661 67 5 } } PUNCT bracis-33661 67 6 } } PUNCT bracis-33661 67 7 \tilde{a \tilde{a NOUN bracis-33661 67 8 } } PUNCT bracis-33661 67 9 \tilde{d}^{-\frac{1}{2 \tilde{d}^{-\frac{1}{2 PROPN bracis-33661 67 10 } } PUNCT bracis-33661 67 11 } } PUNCT bracis-33661 67 12 h^{(l h^{(l NOUN bracis-33661 67 13 ) ) PUNCT bracis-33661 67 14 } } SYM bracis-33661 67 15 w^{(l)})~ w^{(l)})~ PROPN bracis-33661 67 16 , , PUNCT bracis-33661 67 17 \end{aligned}$$ \end{aligned}$$ X bracis-33661 67 18 ( ( PUNCT bracis-33661 67 19 1 1 NUM bracis-33661 67 20 ) ) PUNCT bracis-33661 67 21 with with ADP bracis-33661 67 22 the the DET bracis-33661 67 23 variables variable NOUN bracis-33661 67 24 related relate VERB bracis-33661 67 25 to to ADP bracis-33661 67 26 a a DET bracis-33661 67 27 certain certain ADJ bracis-33661 67 28 undirected undirected ADJ bracis-33661 67 29 graph graph NOUN bracis-33661 67 30 \ \ PROPN bracis-33661 67 31 ( ( PUNCT bracis-33661 67 32 \mathcal \mathcal ADJ bracis-33661 67 33 { { PUNCT bracis-33661 67 34 g g NOUN bracis-33661 67 35 } } PUNCT bracis-33661 67 36 \ \ NOUN bracis-33661 67 37 ) ) PUNCT bracis-33661 67 38 , , PUNCT bracis-33661 67 39 where where SCONJ bracis-33661 67 40 \ \ PROPN bracis-33661 67 41 ( ( PUNCT bracis-33661 67 42 \tilde{a \tilde{a NOUN bracis-33661 67 43 } } PUNCT bracis-33661 67 44 \ \ NOUN bracis-33661 67 45 ) ) PUNCT bracis-33661 67 46 is be AUX bracis-33661 67 47 the the DET bracis-33661 67 48 adjacency adjacency NOUN bracis-33661 67 49 matrix matrix NOUN bracis-33661 67 50 , , PUNCT bracis-33661 67 51 \ \ PROPN bracis-33661 67 52 ( ( PUNCT bracis-33661 67 53 \tilde{d \tilde{d ADJ bracis-33661 67 54 } } PUNCT bracis-33661 67 55 \ \ NOUN bracis-33661 67 56 ) ) PUNCT bracis-33661 67 57 is be AUX bracis-33661 67 58 the the DET bracis-33661 67 59 diagonal diagonal ADJ bracis-33661 67 60 matrix matrix NOUN bracis-33661 67 61 of of ADP bracis-33661 67 62 \ \ PROPN bracis-33661 67 63 ( ( PUNCT bracis-33661 67 64 \tilde{a}\ \tilde{a}\ PROPN bracis-33661 67 65 ) ) PUNCT bracis-33661 67 66 governed govern VERB bracis-33661 67 67 by by ADP bracis-33661 67 68 \(\tilde{d}_{ii \(\tilde{d}_{ii PROPN bracis-33661 67 69 } } PUNCT bracis-33661 67 70 = = SYM bracis-33661 67 71 \sum \sum NOUN bracis-33661 67 72 _ _ PUNCT bracis-33661 67 73 { { PUNCT bracis-33661 67 74 j j NOUN bracis-33661 67 75 } } PUNCT bracis-33661 67 76 \tilde{a}_{ij}\ \tilde{a}_{ij}\ PROPN bracis-33661 67 77 ) ) PUNCT bracis-33661 67 78 , , PUNCT bracis-33661 67 79 \ \ PROPN bracis-33661 67 80 ( ( PUNCT bracis-33661 67 81 w w NOUN bracis-33661 67 82 ^l\ ^l\ PUNCT bracis-33661 67 83 ) ) PUNCT bracis-33661 67 84 is be AUX bracis-33661 67 85 the the DET bracis-33661 67 86 matrix matrix NOUN bracis-33661 67 87 of of ADP bracis-33661 67 88 weights weight NOUN bracis-33661 67 89 of of ADP bracis-33661 67 90 neural neural ADJ bracis-33661 67 91 network network NOUN bracis-33661 67 92 in in ADP bracis-33661 67 93 layer layer NOUN bracis-33661 67 94 l l NOUN bracis-33661 67 95 , , PUNCT bracis-33661 67 96 \(\sigma \(\sigma NOUN bracis-33661 67 97 ( ( PUNCT bracis-33661 67 98 ) ) PUNCT bracis-33661 68 1 \ \ NOUN bracis-33661 68 2 ) ) PUNCT bracis-33661 68 3 function function NOUN bracis-33661 68 4 represents represent VERB bracis-33661 68 5 the the DET bracis-33661 68 6 activation activation NOUN bracis-33661 68 7 functions function NOUN bracis-33661 68 8 at at ADP bracis-33661 68 9 \(l^{th}\ \(l^{th}\ ADJ bracis-33661 68 10 ) ) PUNCT bracis-33661 68 11 layer layer NOUN bracis-33661 68 12 , , PUNCT bracis-33661 68 13 and and CCONJ bracis-33661 68 14 finally finally ADV bracis-33661 68 15 \(h^{l \(h^{l PROPN bracis-33661 68 16 + + CCONJ bracis-33661 68 17 1}\ 1}\ NUM bracis-33661 68 18 ) ) PUNCT bracis-33661 68 19 is be AUX bracis-33661 68 20 the the DET bracis-33661 68 21 resulting result VERB bracis-33661 68 22 matrix matrix NOUN bracis-33661 68 23 at at ADP bracis-33661 68 24 the the DET bracis-33661 68 25 end end NOUN bracis-33661 68 26 of of ADP bracis-33661 68 27 \ \ PROPN bracis-33661 68 28 ( ( PUNCT bracis-33661 68 29 l^{th}\ l^{th}\ NOUN bracis-33661 68 30 ) ) PUNCT bracis-33661 68 31 layer layer NOUN bracis-33661 68 32 , , PUNCT bracis-33661 68 33 where where SCONJ bracis-33661 68 34 \(h^{0}\ \(h^{0}\ NOUN bracis-33661 68 35 ) ) PUNCT bracis-33661 68 36 is be AUX bracis-33661 68 37 the the DET bracis-33661 68 38 initial initial ADJ bracis-33661 68 39 inputs input NOUN bracis-33661 68 40 , , PUNCT bracis-33661 68 41 called call VERB bracis-33661 68 42 x x NOUN bracis-33661 68 43 , , PUNCT bracis-33661 68 44 and and CCONJ bracis-33661 68 45 \(h^{n}\ \(h^{n}\ ADJ bracis-33661 68 46 ) ) PUNCT bracis-33661 68 47 , , PUNCT bracis-33661 68 48 with with ADP bracis-33661 68 49 n n PRON bracis-33661 68 50 as as ADP bracis-33661 68 51 the the DET bracis-33661 68 52 last last ADJ bracis-33661 68 53 layer layer NOUN bracis-33661 68 54 , , PUNCT bracis-33661 68 55 contains contain VERB bracis-33661 68 56 the the DET bracis-33661 68 57 results result NOUN bracis-33661 68 58 in in ADP bracis-33661 68 59 the the DET bracis-33661 68 60 final final ADJ bracis-33661 68 61 step step NOUN bracis-33661 68 62 of of ADP bracis-33661 68 63 our our PRON bracis-33661 68 64 learning learning NOUN bracis-33661 68 65 process process NOUN bracis-33661 68 66 , , PUNCT bracis-33661 68 67 called call VERB bracis-33661 68 68 y. y. NOUN bracis-33661 68 69 equation equation NOUN bracis-33661 68 70 ( ( PUNCT bracis-33661 68 71 1 1 X bracis-33661 68 72 ) ) PUNCT bracis-33661 68 73 is be AUX bracis-33661 68 74 the the DET bracis-33661 68 75 final final ADJ bracis-33661 68 76 representation representation NOUN bracis-33661 68 77 of of ADP bracis-33661 68 78 the the DET bracis-33661 68 79 summary summary NOUN bracis-33661 68 80 equations equation NOUN bracis-33661 68 81 of of ADP bracis-33661 68 82 the the DET bracis-33661 68 83 three three NUM bracis-33661 68 84 stages stage NOUN bracis-33661 68 85 of of ADP bracis-33661 68 86 gcn gcn NOUN bracis-33661 68 87 before before ADP bracis-33661 68 88 the the DET bracis-33661 68 89 prediction prediction NOUN bracis-33661 68 90 , , PUNCT bracis-33661 68 91 which which PRON bracis-33661 68 92 are be AUX bracis-33661 68 93 : : PUNCT bracis-33661 68 94 feature feature NOUN bracis-33661 68 95 propagation propagation NOUN bracis-33661 68 96 , , PUNCT bracis-33661 68 97 linear linear ADJ bracis-33661 68 98 transformation transformation NOUN bracis-33661 68 99 and and CCONJ bracis-33661 68 100 nonlinear nonlinear ADJ bracis-33661 68 101 activation activation NOUN bracis-33661 68 102 . . PUNCT bracis-33661 69 1 the the DET bracis-33661 69 2 feature feature NOUN bracis-33661 69 3 propagation propagation NOUN bracis-33661 69 4 stage stage NOUN bracis-33661 69 5 can can AUX bracis-33661 69 6 be be AUX bracis-33661 69 7 updated update VERB bracis-33661 69 8 for for ADP bracis-33661 69 9 the the DET bracis-33661 69 10 entire entire ADJ bracis-33661 69 11 graph graph NOUN bracis-33661 69 12 represented represent VERB bracis-33661 69 13 by by ADP bracis-33661 69 14 : : PUNCT bracis-33661 69 15 $ $ SYM bracis-33661 69 16 $ $ SYM bracis-33661 69 17 \begin{aligned \begin{aligne VERB bracis-33661 69 18 } } PUNCT bracis-33661 69 19 \tilde{h}^{(l \tilde{h}^{(l NOUN bracis-33661 69 20 ) ) PUNCT bracis-33661 69 21 } } PUNCT bracis-33661 69 22 \leftarrow \leftarrow VERB bracis-33661 69 23 \tilde{d}^{-\frac{1}{2 \tilde{d}^{-\frac{1}{2 PROPN bracis-33661 69 24 } } PUNCT bracis-33661 69 25 } } PUNCT bracis-33661 69 26 \tilde{a \tilde{a NOUN bracis-33661 69 27 } } PUNCT bracis-33661 69 28 \tilde{d}^{-\frac{1}{2 \tilde{d}^{-\frac{1}{2 PROPN bracis-33661 69 29 } } PUNCT bracis-33661 69 30 } } PUNCT bracis-33661 69 31 h^{(l h^{(l NOUN bracis-33661 69 32 1)}~. 1)}~. NUM bracis-33661 69 33 \end{aligned}$$ \end{aligned}$$ NOUN bracis-33661 69 34 ( ( PUNCT bracis-33661 69 35 2 2 NUM bracis-33661 69 36 ) ) PUNCT bracis-33661 69 37 after after ADP bracis-33661 69 38 the the DET bracis-33661 69 39 first first ADJ bracis-33661 69 40 stage stage NOUN bracis-33661 69 41 , , PUNCT bracis-33661 69 42 eq eq NOUN bracis-33661 69 43 . . PUNCT bracis-33661 70 1 ( ( PUNCT bracis-33661 70 2 2 2 X bracis-33661 70 3 ) ) PUNCT bracis-33661 70 4 will will AUX bracis-33661 70 5 be be AUX bracis-33661 70 6 combined combine VERB bracis-33661 70 7 with with ADP bracis-33661 70 8 the the DET bracis-33661 70 9 weight weight NOUN bracis-33661 70 10 matrix matrix NOUN bracis-33661 70 11 \ \ PROPN bracis-33661 70 12 ( ( PUNCT bracis-33661 70 13 w w NOUN bracis-33661 70 14 ^l\ ^l\ PUNCT bracis-33661 70 15 ) ) PUNCT bracis-33661 70 16 for for ADP bracis-33661 70 17 each each DET bracis-33661 70 18 iteration iteration NOUN bracis-33661 70 19 , , PUNCT bracis-33661 70 20 resulting result VERB bracis-33661 70 21 in in ADP bracis-33661 70 22 the the DET bracis-33661 70 23 linear linear ADJ bracis-33661 70 24 transformation transformation NOUN bracis-33661 70 25 stage stage NOUN bracis-33661 70 26 . . PUNCT bracis-33661 71 1 finally finally ADV bracis-33661 71 2 , , PUNCT bracis-33661 71 3 the the DET bracis-33661 71 4 function function NOUN bracis-33661 71 5 \(\sigma \(\sigma NOUN bracis-33661 71 6 ( ( PUNCT bracis-33661 71 7 ) ) PUNCT bracis-33661 71 8 \ \ NOUN bracis-33661 71 9 ) ) PUNCT bracis-33661 71 10 can can AUX bracis-33661 71 11 be be AUX bracis-33661 71 12 applied apply VERB bracis-33661 71 13 to to PART bracis-33661 71 14 complete complete VERB bracis-33661 71 15 the the DET bracis-33661 71 16 process process NOUN bracis-33661 71 17 : : PUNCT bracis-33661 71 18 $ $ SYM bracis-33661 71 19 $ $ SYM bracis-33661 71 20 \begin{aligned \begin{aligne VERB bracis-33661 71 21 } } PUNCT bracis-33661 71 22 h^{(l h^{(l NOUN bracis-33661 71 23 ) ) PUNCT bracis-33661 71 24 } } PUNCT bracis-33661 71 25 \leftarrow \leftarrow VERB bracis-33661 71 26 \sigma \sigma PROPN bracis-33661 71 27 ( ( PUNCT bracis-33661 71 28 \tilde{d}^{-\frac{1}{2 \tilde{d}^{-\frac{1}{2 PROPN bracis-33661 71 29 } } PUNCT bracis-33661 71 30 } } PUNCT bracis-33661 71 31 \tilde{a \tilde{a NOUN bracis-33661 71 32 } } PUNCT bracis-33661 71 33 \tilde{d}^{-\frac{1}{2 \tilde{d}^{-\frac{1}{2 PROPN bracis-33661 71 34 } } PUNCT bracis-33661 71 35 } } PUNCT bracis-33661 71 36 h^{(l-1 h^{(l-1 NUM bracis-33661 71 37 ) ) PUNCT bracis-33661 71 38 } } PUNCT bracis-33661 72 1 w^{(l)})~. w^{(l)})~. PROPN bracis-33661 72 2 \end{aligned}$$ \end{aligned}$$ X bracis-33661 73 1 ( ( PUNCT bracis-33661 73 2 3 3 NUM bracis-33661 73 3 ) ) PUNCT bracis-33661 73 4 at at ADP bracis-33661 73 5 the the DET bracis-33661 73 6 end end NOUN bracis-33661 73 7 , , PUNCT bracis-33661 73 8 gcn gcn ADJ bracis-33661 73 9 prediction prediction NOUN bracis-33661 73 10 is be AUX bracis-33661 73 11 performed perform VERB bracis-33661 73 12 by by ADP bracis-33661 73 13 applying apply VERB bracis-33661 73 14 a a DET bracis-33661 73 15 softmax softmax NOUN bracis-33661 73 16 layer layer NOUN bracis-33661 73 17 in in ADP bracis-33661 73 18 order order NOUN bracis-33661 73 19 to to PART bracis-33661 73 20 normalize normalize VERB bracis-33661 73 21 the the DET bracis-33661 73 22 results result NOUN bracis-33661 73 23 and and CCONJ bracis-33661 73 24 weights weight NOUN bracis-33661 73 25 obtained obtain VERB bracis-33661 73 26 during during ADP bracis-33661 73 27 the the DET bracis-33661 73 28 update update NOUN bracis-33661 73 29 process process NOUN bracis-33661 73 30 . . PUNCT bracis-33661 74 1 simple simple ADJ bracis-33661 74 2 graph graph NOUN bracis-33661 74 3 convolution convolution NOUN bracis-33661 74 4 ( ( PUNCT bracis-33661 74 5 sgc sgc PROPN bracis-33661 74 6 ) ) PUNCT bracis-33661 74 7 differs differ VERB bracis-33661 74 8 from from ADP bracis-33661 74 9 gcn gcn NOUN bracis-33661 74 10 by by ADP bracis-33661 74 11 the the DET bracis-33661 74 12 removal removal NOUN bracis-33661 74 13 of of ADP bracis-33661 74 14 the the DET bracis-33661 74 15 non non ADJ bracis-33661 74 16 - - ADJ bracis-33661 74 17 linear linear ADJ bracis-33661 74 18 transformations transformation NOUN bracis-33661 74 19 sequence sequence NOUN bracis-33661 74 20 between between ADP bracis-33661 74 21 the the DET bracis-33661 74 22 layers layer NOUN bracis-33661 74 23 of of ADP bracis-33661 74 24 the the DET bracis-33661 74 25 gcn gcn ADJ bracis-33661 74 26 structure structure NOUN bracis-33661 74 27 and and CCONJ bracis-33661 74 28 by by ADP bracis-33661 74 29 exploring explore VERB bracis-33661 74 30 more more ADJ bracis-33661 74 31 the the DET bracis-33661 74 32 feature feature NOUN bracis-33661 74 33 propagation propagation NOUN bracis-33661 74 34 stage stage NOUN bracis-33661 74 35 [ [ X bracis-33661 74 36 31 31 NUM bracis-33661 74 37 ] ] PUNCT bracis-33661 74 38 . . PUNCT bracis-33661 75 1 the the DET bracis-33661 75 2 feature feature NOUN bracis-33661 75 3 propagation propagation NOUN bracis-33661 75 4 stage stage NOUN bracis-33661 75 5 for for ADP bracis-33661 75 6 the the DET bracis-33661 75 7 sgc sgc PROPN bracis-33661 75 8 has have VERB bracis-33661 75 9 as as ADP bracis-33661 75 10 main main ADJ bracis-33661 75 11 objective objective NOUN bracis-33661 75 12 to to PART bracis-33661 75 13 explore explore VERB bracis-33661 75 14 the the DET bracis-33661 75 15 potential potential NOUN bracis-33661 75 16 of of ADP bracis-33661 75 17 the the DET bracis-33661 75 18 relations relation NOUN bracis-33661 75 19 between between ADP bracis-33661 75 20 the the DET bracis-33661 75 21 neighboring neighboring NOUN bracis-33661 75 22 vertices vertex NOUN bracis-33661 75 23 via via ADP bracis-33661 75 24 configuration configuration NOUN bracis-33661 75 25 of of ADP bracis-33661 75 26 the the DET bracis-33661 75 27 adjacency adjacency NOUN bracis-33661 75 28 matrix matrix NOUN bracis-33661 75 29 . . PUNCT bracis-33661 76 1 therefore therefore ADV bracis-33661 76 2 , , PUNCT bracis-33661 76 3 eq eq ADJ bracis-33661 76 4 . . PUNCT bracis-33661 77 1 ( ( PUNCT bracis-33661 77 2 2 2 X bracis-33661 77 3 ) ) PUNCT bracis-33661 77 4 previously previously ADV bracis-33661 77 5 used use VERB bracis-33661 77 6 in in ADP bracis-33661 77 7 gcn gcn ADJ bracis-33661 77 8 layers layer NOUN bracis-33661 77 9 can can AUX bracis-33661 77 10 be be AUX bracis-33661 77 11 rewritten rewrite VERB bracis-33661 77 12 for for ADP bracis-33661 77 13 use use NOUN bracis-33661 77 14 in in ADP bracis-33661 77 15 sgc sgc PROPN bracis-33661 77 16 architectures architecture NOUN bracis-33661 77 17 as as SCONJ bracis-33661 77 18 follows follow VERB bracis-33661 77 19 : : PUNCT bracis-33661 77 20 $ $ SYM bracis-33661 77 21 $ $ SYM bracis-33661 77 22 \begin{aligned \begin{aligne VERB bracis-33661 77 23 } } PUNCT bracis-33661 77 24 \tilde{h}^{(l \tilde{h}^{(l NOUN bracis-33661 77 25 ) ) PUNCT bracis-33661 77 26 } } PUNCT bracis-33661 77 27 \leftarrow \leftarrow PROPN bracis-33661 77 28 ( ( PUNCT bracis-33661 77 29 \tilde{d}^{-\frac{1}{2 \tilde{d}^{-\frac{1}{2 PROPN bracis-33661 77 30 } } PUNCT bracis-33661 77 31 } } PUNCT bracis-33661 77 32 \tilde{a \tilde{a NOUN bracis-33661 77 33 } } PUNCT bracis-33661 77 34 \tilde{d}^{-\frac{1}{2}})^{l \tilde{d}^{-\frac{1}{2}})^{l PROPN bracis-33661 77 35 } } PUNCT bracis-33661 77 36 h^{(0)}~. h^{(0)}~. PROPN bracis-33661 77 37 \end{aligned}$$ \end{aligned}$$ X bracis-33661 77 38 ( ( PUNCT bracis-33661 77 39 4 4 NUM bracis-33661 77 40 ) ) PUNCT bracis-33661 77 41 from from ADP bracis-33661 77 42 eq eq ADP bracis-33661 77 43 . . PUNCT bracis-33661 77 44 ( ( PUNCT bracis-33661 77 45 4 4 NUM bracis-33661 77 46 ) ) PUNCT bracis-33661 77 47 , , PUNCT bracis-33661 77 48 the the DET bracis-33661 77 49 product product NOUN bracis-33661 77 50 by by ADP bracis-33661 77 51 the the DET bracis-33661 77 52 matrix matrix NOUN bracis-33661 77 53 of of ADP bracis-33661 77 54 weights weight NOUN bracis-33661 77 55 \ \ PROPN bracis-33661 77 56 ( ( PUNCT bracis-33661 77 57 w w NOUN bracis-33661 77 58 ^l\ ^l\ NUM bracis-33661 77 59 ) ) PUNCT bracis-33661 77 60 results result NOUN bracis-33661 77 61 is be AUX bracis-33661 77 62 given give VERB bracis-33661 77 63 by by ADP bracis-33661 77 64 : : PUNCT bracis-33661 77 65 $ $ SYM bracis-33661 77 66 $ $ SYM bracis-33661 77 67 \begin{aligned \begin{aligne VERB bracis-33661 77 68 } } PUNCT bracis-33661 77 69 \tilde{h}^{(l \tilde{h}^{(l NOUN bracis-33661 77 70 ) ) PUNCT bracis-33661 77 71 } } PUNCT bracis-33661 77 72 \leftarrow \leftarrow PROPN bracis-33661 77 73 ( ( PUNCT bracis-33661 77 74 \tilde{d}^{-\frac{1}{2 \tilde{d}^{-\frac{1}{2 PROPN bracis-33661 77 75 } } PUNCT bracis-33661 77 76 } } PUNCT bracis-33661 77 77 \tilde{a \tilde{a NOUN bracis-33661 77 78 } } PUNCT bracis-33661 77 79 \tilde{d}^{-\frac{1}{2}})^{l \tilde{d}^{-\frac{1}{2}})^{l ADJ bracis-33661 77 80 } } PUNCT bracis-33661 77 81 h^{(0 h^{(0 ADJ bracis-33661 77 82 ) ) PUNCT bracis-33661 77 83 } } PUNCT bracis-33661 77 84 w^{(l)}~. w^{(l)}~. VERB bracis-33661 77 85 \end{aligned}$$ \end{aligned}$$ NOUN bracis-33661 77 86 ( ( PUNCT bracis-33661 77 87 5 5 NUM bracis-33661 77 88 ) ) PUNCT bracis-33661 77 89 to to PART bracis-33661 77 90 obtain obtain VERB bracis-33661 77 91 the the DET bracis-33661 77 92 normalized normalize VERB bracis-33661 77 93 value value NOUN bracis-33661 77 94 of of ADP bracis-33661 77 95 the the DET bracis-33661 77 96 sgc sgc PROPN bracis-33661 77 97 predictions prediction NOUN bracis-33661 77 98 for for ADP bracis-33661 77 99 the the DET bracis-33661 77 100 classes class NOUN bracis-33661 77 101 , , PUNCT bracis-33661 77 102 a a DET bracis-33661 77 103 softmax softmax NOUN bracis-33661 77 104 function function NOUN bracis-33661 77 105 is be AUX bracis-33661 77 106 applied apply VERB bracis-33661 77 107 over over ADP bracis-33661 77 108 eq eq ADP bracis-33661 77 109 . . PUNCT bracis-33661 78 1 ( ( PUNCT bracis-33661 78 2 5 5 NUM bracis-33661 78 3 ) ) PUNCT bracis-33661 78 4 . . PUNCT bracis-33661 79 1 sample sample NOUN bracis-33661 79 2 and and CCONJ bracis-33661 79 3 aggregate aggregate VERB bracis-33661 79 4 convolutional convolutional ADJ bracis-33661 79 5 network network NOUN bracis-33661 79 6 ( ( PUNCT bracis-33661 79 7 sage sage NOUN bracis-33661 79 8 ) ) PUNCT bracis-33661 79 9 introduces introduce VERB bracis-33661 79 10 an an DET bracis-33661 79 11 approach approach NOUN bracis-33661 79 12 for for ADP bracis-33661 79 13 aggregating aggregate VERB bracis-33661 79 14 feature feature NOUN bracis-33661 79 15 information information NOUN bracis-33661 79 16 from from ADP bracis-33661 79 17 a a DET bracis-33661 79 18 node node NOUN bracis-33661 79 19 ’s ’s PART bracis-33661 79 20 local local ADJ bracis-33661 79 21 neighborhood neighborhood NOUN bracis-33661 79 22 , , PUNCT bracis-33661 79 23 enhancing enhance VERB bracis-33661 79 24 the the DET bracis-33661 79 25 representation representation NOUN bracis-33661 79 26 learning learn VERB bracis-33661 79 27 for for ADP bracis-33661 79 28 graph graph NOUN bracis-33661 79 29 - - PUNCT bracis-33661 79 30 structured structure VERB bracis-33661 79 31 data datum NOUN bracis-33661 79 32 . . PUNCT bracis-33661 80 1 the the DET bracis-33661 80 2 propagation propagation NOUN bracis-33661 80 3 behavior behavior NOUN bracis-33661 80 4 of of ADP bracis-33661 80 5 the the DET bracis-33661 80 6 sage sage NOUN bracis-33661 80 7 layer layer NOUN bracis-33661 80 8 is be AUX bracis-33661 80 9 given give VERB bracis-33661 80 10 by by ADP bracis-33661 80 11 eqs eqs PROPN bracis-33661 80 12 . . PUNCT bracis-33661 81 1 ( ( PUNCT bracis-33661 81 2 6 6 NUM bracis-33661 81 3 ) ) PUNCT bracis-33661 81 4 and and CCONJ bracis-33661 81 5 ( ( PUNCT bracis-33661 81 6 7 7 NUM bracis-33661 81 7 ) ) PUNCT bracis-33661 81 8 , , PUNCT bracis-33661 81 9 which which PRON bracis-33661 81 10 collectively collectively ADV bracis-33661 81 11 describe describe VERB bracis-33661 81 12 the the DET bracis-33661 81 13 sampling sampling NOUN bracis-33661 81 14 and and CCONJ bracis-33661 81 15 aggregation aggregation NOUN bracis-33661 81 16 processes process NOUN bracis-33661 81 17 integral integral ADJ bracis-33661 81 18 to to ADP bracis-33661 81 19 the the DET bracis-33661 81 20 model model NOUN bracis-33661 82 1 [ [ X bracis-33661 82 2 12 12 NUM bracis-33661 82 3 ] ] PUNCT bracis-33661 82 4 . . PUNCT bracis-33661 83 1 the the DET bracis-33661 83 2 first first ADJ bracis-33661 83 3 step step NOUN bracis-33661 83 4 in in ADP bracis-33661 83 5 the the DET bracis-33661 83 6 sage sage NOUN bracis-33661 83 7 layer layer NOUN bracis-33661 83 8 involves involve VERB bracis-33661 83 9 sampling sample VERB bracis-33661 83 10 a a DET bracis-33661 83 11 fixed fix VERB bracis-33661 83 12 - - PUNCT bracis-33661 83 13 size size NOUN bracis-33661 83 14 set set NOUN bracis-33661 83 15 of of ADP bracis-33661 83 16 neighbors neighbor NOUN bracis-33661 83 17 for for ADP bracis-33661 83 18 each each DET bracis-33661 83 19 node node NOUN bracis-33661 83 20 , , PUNCT bracis-33661 83 21 which which PRON bracis-33661 83 22 can can AUX bracis-33661 83 23 be be AUX bracis-33661 83 24 described describe VERB bracis-33661 83 25 by by ADP bracis-33661 83 26 : : PUNCT bracis-33661 83 27 $ $ SYM bracis-33661 83 28 $ $ SYM bracis-33661 83 29 \begin{aligned \begin{aligne VERB bracis-33661 83 30 } } PUNCT bracis-33661 83 31 \mathcal \mathcal ADJ bracis-33661 83 32 { { PUNCT bracis-33661 83 33 n}(v n}(v PROPN bracis-33661 83 34 ) ) PUNCT bracis-33661 84 1 = = PUNCT bracis-33661 85 1 \text \text ADJ bracis-33661 85 2 { { PUNCT bracis-33661 85 3 sample}(\tilde{\mathcal sample}(\tilde{\mathcal ADJ bracis-33661 85 4 { { PUNCT bracis-33661 85 5 n}}(v n}}(v NOUN bracis-33661 85 6 ) ) PUNCT bracis-33661 85 7 , , PUNCT bracis-33661 85 8 k)~ k)~ PROPN bracis-33661 85 9 , , PUNCT bracis-33661 85 10 \end{aligned}$$ \end{aligned}$$ X bracis-33661 85 11 ( ( PUNCT bracis-33661 85 12 6 6 NUM bracis-33661 85 13 ) ) PUNCT bracis-33661 85 14 where where SCONJ bracis-33661 85 15 \(\mathcal \(\mathcal ADJ bracis-33661 85 16 { { PUNCT bracis-33661 85 17 n}(v)\ n}(v)\ NOUN bracis-33661 85 18 ) ) PUNCT bracis-33661 85 19 represents represent VERB bracis-33661 85 20 the the DET bracis-33661 85 21 set set NOUN bracis-33661 85 22 of of ADP bracis-33661 85 23 neighbors neighbor NOUN bracis-33661 85 24 sampled sample VERB bracis-33661 85 25 for for ADP bracis-33661 85 26 node node NOUN bracis-33661 85 27 \ \ PROPN bracis-33661 85 28 ( ( PUNCT bracis-33661 85 29 v v ADP bracis-33661 85 30 \ \ PROPN bracis-33661 85 31 ) ) PUNCT bracis-33661 85 32 , , PUNCT bracis-33661 85 33 \(\tilde{\mathcal \(\tilde{\mathcal ADJ bracis-33661 85 34 { { PUNCT bracis-33661 85 35 n}}(v)\ n}}(v)\ NOUN bracis-33661 85 36 ) ) PUNCT bracis-33661 85 37 is be AUX bracis-33661 85 38 the the DET bracis-33661 85 39 full full ADJ bracis-33661 85 40 set set NOUN bracis-33661 85 41 of of ADP bracis-33661 85 42 neighbors neighbor NOUN bracis-33661 85 43 of of ADP bracis-33661 85 44 \ \ PROPN bracis-33661 85 45 ( ( PUNCT bracis-33661 85 46 v v NOUN bracis-33661 85 47 \ \ PROPN bracis-33661 85 48 ) ) PUNCT bracis-33661 85 49 , , PUNCT bracis-33661 85 50 and and CCONJ bracis-33661 85 51 \ \ PROPN bracis-33661 85 52 ( ( PUNCT bracis-33661 85 53 k k PROPN bracis-33661 85 54 \ \ PROPN bracis-33661 85 55 ) ) PUNCT bracis-33661 85 56 is be AUX bracis-33661 85 57 the the DET bracis-33661 85 58 number number NOUN bracis-33661 85 59 of of ADP bracis-33661 85 60 neighbors neighbor NOUN bracis-33661 85 61 to to ADP bracis-33661 85 62 sample sample NOUN bracis-33661 85 63 . . PUNCT bracis-33661 86 1 once once SCONJ bracis-33661 86 2 the the DET bracis-33661 86 3 neighbors neighbor NOUN bracis-33661 86 4 are be AUX bracis-33661 86 5 sampled sample VERB bracis-33661 86 6 , , PUNCT bracis-33661 86 7 the the DET bracis-33661 86 8 next next ADJ bracis-33661 86 9 step step NOUN bracis-33661 86 10 is be AUX bracis-33661 86 11 to to PART bracis-33661 86 12 aggregate aggregate VERB bracis-33661 86 13 their their PRON bracis-33661 86 14 features feature NOUN bracis-33661 86 15 using use VERB bracis-33661 86 16 an an DET bracis-33661 86 17 aggregation aggregation NOUN bracis-33661 86 18 function function NOUN bracis-33661 86 19 \ \ PROPN bracis-33661 87 1 ( ( PUNCT bracis-33661 87 2 \text \text PROPN bracis-33661 87 3 { { PUNCT bracis-33661 87 4 agg agg PROPN bracis-33661 87 5 } } PUNCT bracis-33661 87 6 \ \ PROPN bracis-33661 87 7 ) ) PUNCT bracis-33661 87 8 , , PUNCT bracis-33661 87 9 which which PRON bracis-33661 87 10 can can AUX bracis-33661 87 11 be be AUX bracis-33661 87 12 mean mean ADJ bracis-33661 87 13 , , PUNCT bracis-33661 87 14 lstm lstm NOUN bracis-33661 87 15 - - PUNCT bracis-33661 87 16 based base VERB bracis-33661 87 17 , , PUNCT bracis-33661 87 18 or or CCONJ bracis-33661 87 19 pooling pooling NOUN bracis-33661 87 20 - - PUNCT bracis-33661 87 21 based base VERB bracis-33661 87 22 . . PUNCT bracis-33661 88 1 the the DET bracis-33661 88 2 aggregated aggregate VERB bracis-33661 88 3 feature feature NOUN bracis-33661 88 4 vector vector NOUN bracis-33661 88 5 is be AUX bracis-33661 88 6 then then ADV bracis-33661 88 7 concatenated concatenate VERB bracis-33661 88 8 with with ADP bracis-33661 88 9 the the DET bracis-33661 88 10 node node NOUN bracis-33661 88 11 ’s ’s PART bracis-33661 88 12 own own ADJ bracis-33661 88 13 feature feature NOUN bracis-33661 88 14 vector vector NOUN bracis-33661 88 15 and and CCONJ bracis-33661 88 16 transformed transform VERB bracis-33661 88 17 through through ADP bracis-33661 88 18 a a DET bracis-33661 88 19 weight weight NOUN bracis-33661 88 20 matrix matrix NOUN bracis-33661 88 21 and and CCONJ bracis-33661 88 22 non non ADJ bracis-33661 88 23 - - ADJ bracis-33661 88 24 linear linear ADJ bracis-33661 88 25 activation activation NOUN bracis-33661 88 26 function function NOUN bracis-33661 88 27 , , PUNCT bracis-33661 88 28 as as SCONJ bracis-33661 88 29 follows follow VERB bracis-33661 88 30 : : PUNCT bracis-33661 89 1 $ $ SYM bracis-33661 89 2 $ $ SYM bracis-33661 89 3 \begin{aligned \begin{aligne VERB bracis-33661 89 4 } } PUNCT bracis-33661 89 5 h_v^{(l+1 h_v^{(l+1 NOUN bracis-33661 89 6 ) ) PUNCT bracis-33661 89 7 } } PUNCT bracis-33661 90 1 = = SYM bracis-33661 90 2 \sigma \sigma PROPN bracis-33661 90 3 \left \left PROPN bracis-33661 90 4 ( ( PUNCT bracis-33661 90 5 w^{(l w^{(l NOUN bracis-33661 90 6 ) ) PUNCT bracis-33661 90 7 } } PUNCT bracis-33661 90 8 \cdot \cdot PROPN bracis-33661 90 9 \text \text NOUN bracis-33661 90 10 { { PUNCT bracis-33661 90 11 concat concat NOUN bracis-33661 90 12 } } PUNCT bracis-33661 90 13 \left \left PROPN bracis-33661 90 14 ( ( PUNCT bracis-33661 90 15 h_v^{(l h_v^{(l PROPN bracis-33661 90 16 ) ) PUNCT bracis-33661 90 17 } } PUNCT bracis-33661 90 18 , , PUNCT bracis-33661 90 19 \text \text ADP bracis-33661 90 20 { { PUNCT bracis-33661 90 21 agg agg PROPN bracis-33661 90 22 } } PUNCT bracis-33661 90 23 \left \left PROPN bracis-33661 90 24 ( ( PUNCT bracis-33661 90 25 \ \ PROPN bracis-33661 90 26 { { PUNCT bracis-33661 90 27 h_u^{(l h_u^{(l PROPN bracis-33661 90 28 ) ) PUNCT bracis-33661 90 29 } } PUNCT bracis-33661 90 30 , , PUNCT bracis-33661 90 31 \forall \forall DET bracis-33661 90 32 u u PROPN bracis-33661 90 33 \in \in PROPN bracis-33661 90 34 \mathcal \mathcal PROPN bracis-33661 90 35 { { PUNCT bracis-33661 90 36 n}(v n}(v PROPN bracis-33661 90 37 ) ) PUNCT bracis-33661 90 38 \ \ ADJ bracis-33661 90 39 } } PUNCT bracis-33661 90 40 \right \right NOUN bracis-33661 90 41 ) ) PUNCT bracis-33661 90 42 \right \right NOUN bracis-33661 90 43 ) ) PUNCT bracis-33661 90 44 \right \right NOUN bracis-33661 90 45 ) ) PUNCT bracis-33661 90 46 ~ ~ NOUN bracis-33661 90 47 , , PUNCT bracis-33661 90 48 \end{aligned}$$ \end{aligned}$$ X bracis-33661 90 49 ( ( PUNCT bracis-33661 90 50 7 7 NUM bracis-33661 90 51 ) ) PUNCT bracis-33661 90 52 in in ADP bracis-33661 90 53 which which PRON bracis-33661 90 54 \ \ PROPN bracis-33661 90 55 ( ( PUNCT bracis-33661 90 56 h_v^{(l h_v^{(l PROPN bracis-33661 90 57 ) ) PUNCT bracis-33661 90 58 } } PUNCT bracis-33661 90 59 \ \ NOUN bracis-33661 90 60 ) ) PUNCT bracis-33661 90 61 denotes denote VERB bracis-33661 90 62 the the DET bracis-33661 90 63 feature feature NOUN bracis-33661 90 64 vector vector NOUN bracis-33661 90 65 of of ADP bracis-33661 90 66 node node PROPN bracis-33661 90 67 \ \ PROPN bracis-33661 90 68 ( ( PUNCT bracis-33661 90 69 v v NOUN bracis-33661 90 70 \ \ NOUN bracis-33661 90 71 ) ) PUNCT bracis-33661 90 72 at at ADP bracis-33661 90 73 layer layer NOUN bracis-33661 90 74 \ \ PROPN bracis-33661 90 75 ( ( PUNCT bracis-33661 90 76 l l NOUN bracis-33661 90 77 \ \ PROPN bracis-33661 90 78 ) ) PUNCT bracis-33661 90 79 , , PUNCT bracis-33661 90 80 \ \ PROPN bracis-33661 90 81 ( ( PUNCT bracis-33661 90 82 \text \text X bracis-33661 90 83 { { PUNCT bracis-33661 90 84 concat concat NOUN bracis-33661 90 85 } } PUNCT bracis-33661 90 86 \ \ NOUN bracis-33661 90 87 ) ) PUNCT bracis-33661 90 88 is be AUX bracis-33661 90 89 the the DET bracis-33661 90 90 concatenation concatenation NOUN bracis-33661 90 91 operation operation NOUN bracis-33661 90 92 , , PUNCT bracis-33661 90 93 \ \ PROPN bracis-33661 90 94 ( ( PUNCT bracis-33661 90 95 \text \text PROPN bracis-33661 90 96 { { PUNCT bracis-33661 90 97 agg agg PROPN bracis-33661 90 98 } } PUNCT bracis-33661 90 99 \ \ NOUN bracis-33661 90 100 ) ) PUNCT bracis-33661 90 101 represents represent VERB bracis-33661 90 102 the the DET bracis-33661 90 103 chosen choose VERB bracis-33661 90 104 aggregation aggregation NOUN bracis-33661 90 105 function function NOUN bracis-33661 90 106 , , PUNCT bracis-33661 90 107 \ \ PROPN bracis-33661 90 108 ( ( PUNCT bracis-33661 90 109 w^{(l w^{(l NOUN bracis-33661 90 110 ) ) PUNCT bracis-33661 90 111 } } PUNCT bracis-33661 90 112 \ \ NOUN bracis-33661 90 113 ) ) PUNCT bracis-33661 91 1 is be AUX bracis-33661 91 2 the the DET bracis-33661 91 3 weight weight NOUN bracis-33661 91 4 matrix matrix NOUN bracis-33661 91 5 for for ADP bracis-33661 91 6 layer layer NOUN bracis-33661 91 7 \ \ NOUN bracis-33661 91 8 ( ( PUNCT bracis-33661 91 9 l l NOUN bracis-33661 91 10 \ \ PROPN bracis-33661 91 11 ) ) PUNCT bracis-33661 91 12 , , PUNCT bracis-33661 91 13 and and CCONJ bracis-33661 91 14 \ \ PROPN bracis-33661 91 15 ( ( PUNCT bracis-33661 91 16 \sigma \sigma PROPN bracis-33661 91 17 \ \ PROPN bracis-33661 91 18 ) ) PUNCT bracis-33661 91 19 is be AUX bracis-33661 91 20 the the DET bracis-33661 91 21 activation activation NOUN bracis-33661 91 22 function function NOUN bracis-33661 91 23 . . PUNCT bracis-33661 92 1 the the DET bracis-33661 92 2 final final ADJ bracis-33661 92 3 node node ADJ bracis-33661 92 4 representations representation NOUN bracis-33661 92 5 can can AUX bracis-33661 92 6 be be AUX bracis-33661 92 7 used use VERB bracis-33661 92 8 for for ADP bracis-33661 92 9 downstream downstream ADJ bracis-33661 92 10 tasks task NOUN bracis-33661 92 11 such such ADJ bracis-33661 92 12 as as ADP bracis-33661 92 13 node node ADJ bracis-33661 92 14 classification classification NOUN bracis-33661 92 15 or or CCONJ bracis-33661 92 16 link link NOUN bracis-33661 92 17 prediction prediction NOUN bracis-33661 92 18 , , PUNCT bracis-33661 92 19 often often ADV bracis-33661 92 20 incorporating incorporate VERB bracis-33661 92 21 a a DET bracis-33661 92 22 softmax softmax NOUN bracis-33661 92 23 layer layer NOUN bracis-33661 92 24 for for ADP bracis-33661 92 25 normalization normalization NOUN bracis-33661 92 26 and and CCONJ bracis-33661 92 27 classification classification NOUN bracis-33661 92 28 purposes purpose NOUN bracis-33661 92 29 . . PUNCT bracis-33661 93 1 graph graph VERB bracis-33661 93 2 convolutional convolutional ADJ bracis-33661 93 3 network network NOUN bracis-33661 93 4 ii ii PROPN bracis-33661 93 5 ( ( PUNCT bracis-33661 93 6 gcnii gcnii PROPN bracis-33661 93 7 ) ) PUNCT bracis-33661 93 8 extends extend VERB bracis-33661 93 9 the the DET bracis-33661 93 10 traditional traditional ADJ bracis-33661 93 11 gcn gcn NOUN bracis-33661 93 12 by by ADP bracis-33661 93 13 introducing introduce VERB bracis-33661 93 14 initial initial ADJ bracis-33661 93 15 residual residual ADJ bracis-33661 93 16 and and CCONJ bracis-33661 93 17 identity identity NOUN bracis-33661 93 18 mapping mapping NOUN bracis-33661 93 19 , , PUNCT bracis-33661 93 20 which which PRON bracis-33661 93 21 address address VERB bracis-33661 93 22 the the DET bracis-33661 93 23 limitations limitation NOUN bracis-33661 93 24 of of ADP bracis-33661 93 25 deeper deep ADJ bracis-33661 93 26 gnns gnns NOUN bracis-33661 93 27 . . PUNCT bracis-33661 94 1 the the DET bracis-33661 94 2 propagation propagation NOUN bracis-33661 94 3 behavior behavior NOUN bracis-33661 94 4 of of ADP bracis-33661 94 5 gcnii gcnii NOUN bracis-33661 94 6 is be AUX bracis-33661 94 7 governed govern VERB bracis-33661 94 8 by by ADP bracis-33661 94 9 eqs eqs PROPN bracis-33661 94 10 . . PUNCT bracis-33661 95 1 ( ( PUNCT bracis-33661 95 2 8) 8) NUM bracis-33661 95 3 and and CCONJ bracis-33661 95 4 ( ( PUNCT bracis-33661 95 5 9 9 NUM bracis-33661 95 6 ) ) PUNCT bracis-33661 95 7 , , PUNCT bracis-33661 95 8 which which PRON bracis-33661 95 9 incorporate incorporate VERB bracis-33661 95 10 these these DET bracis-33661 95 11 novel novel ADJ bracis-33661 95 12 characteristics characteristic NOUN bracis-33661 95 13 [ [ X bracis-33661 95 14 6 6 NUM bracis-33661 95 15 ] ] PUNCT bracis-33661 95 16 . . PUNCT bracis-33661 96 1 the the DET bracis-33661 96 2 initial initial ADJ bracis-33661 96 3 residual residual ADJ bracis-33661 96 4 connection connection NOUN bracis-33661 96 5 allows allow VERB bracis-33661 96 6 each each DET bracis-33661 96 7 layer layer NOUN bracis-33661 96 8 to to PART bracis-33661 96 9 directly directly ADV bracis-33661 96 10 utilize utilize VERB bracis-33661 96 11 the the DET bracis-33661 96 12 input input NOUN bracis-33661 96 13 features feature NOUN bracis-33661 96 14 of of ADP bracis-33661 96 15 the the DET bracis-33661 96 16 first first ADJ bracis-33661 96 17 layer layer NOUN bracis-33661 96 18 , , PUNCT bracis-33661 96 19 enhancing enhance VERB bracis-33661 96 20 information information NOUN bracis-33661 96 21 flow flow NOUN bracis-33661 96 22 and and CCONJ bracis-33661 96 23 mitigating mitigate VERB bracis-33661 96 24 the the DET bracis-33661 96 25 vanishing vanish VERB bracis-33661 96 26 gradient gradient NOUN bracis-33661 96 27 problem problem NOUN bracis-33661 96 28 . . PUNCT bracis-33661 97 1 from from ADP bracis-33661 97 2 eq eq ADP bracis-33661 97 3 . . PUNCT bracis-33661 97 4     SPACE bracis-33661 97 5 1 1 NUM bracis-33661 97 6 , , PUNCT bracis-33661 97 7 this this PRON bracis-33661 97 8 can can AUX bracis-33661 97 9 be be AUX bracis-33661 97 10 defined define VERB bracis-33661 97 11 as as SCONJ bracis-33661 97 12 follows follow VERB bracis-33661 97 13 : : PUNCT bracis-33661 97 14 $ $ SYM bracis-33661 97 15 $ $ SYM bracis-33661 97 16 \begin{aligned \begin{aligne VERB bracis-33661 97 17 } } PUNCT bracis-33661 97 18 h^{(l+1 h^{(l+1 NOUN bracis-33661 97 19 ) ) PUNCT bracis-33661 97 20 } } PUNCT bracis-33661 98 1 = = SYM bracis-33661 98 2 \sigma \sigma PROPN bracis-33661 98 3 \left \left PROPN bracis-33661 98 4 ( ( PUNCT bracis-33661 98 5 ( ( PUNCT bracis-33661 98 6 1 1 NUM bracis-33661 98 7 \alpha \alpha ADJ bracis-33661 98 8 ) ) PUNCT bracis-33661 98 9 \left \left PROPN bracis-33661 98 10 ( ( PUNCT bracis-33661 98 11 \tilde{d}^{-\frac{1}{2 \tilde{d}^{-\frac{1}{2 PROPN bracis-33661 98 12 } } PUNCT bracis-33661 98 13 } } PUNCT bracis-33661 98 14 \tilde{a \tilde{a NOUN bracis-33661 98 15 } } PUNCT bracis-33661 98 16 \tilde{d}^{-\frac{1}{2 \tilde{d}^{-\frac{1}{2 PROPN bracis-33661 98 17 } } PUNCT bracis-33661 98 18 } } PUNCT bracis-33661 98 19 h^{(l h^{(l NOUN bracis-33661 98 20 ) ) PUNCT bracis-33661 98 21 } } PUNCT bracis-33661 98 22 w^{(l w^{(l NOUN bracis-33661 98 23 ) ) PUNCT bracis-33661 98 24 } } PUNCT bracis-33661 98 25 \right \right PROPN bracis-33661 98 26 ) ) PUNCT bracis-33661 98 27 + + CCONJ bracis-33661 98 28 \alpha \alpha ADJ bracis-33661 98 29 h^{(0 h^{(0 ADJ bracis-33661 98 30 ) ) PUNCT bracis-33661 98 31 } } PUNCT bracis-33661 98 32 \right \right PROPN bracis-33661 98 33 ) ) PUNCT bracis-33661 98 34 ~ ~ NOUN bracis-33661 98 35 , , PUNCT bracis-33661 98 36 \end{aligned}$$ \end{aligned}$$ X bracis-33661 98 37 ( ( PUNCT bracis-33661 98 38 8) 8) NUM bracis-33661 98 39 where where SCONJ bracis-33661 98 40 \(\alpha \(\alpha ADV bracis-33661 98 41 \ \ NOUN bracis-33661 98 42 ) ) PUNCT bracis-33661 98 43 is be AUX bracis-33661 98 44 a a DET bracis-33661 98 45 hyperparameter hyperparameter NOUN bracis-33661 98 46 controlling control VERB bracis-33661 98 47 the the DET bracis-33661 98 48 balance balance NOUN bracis-33661 98 49 between between ADP bracis-33661 98 50 the the DET bracis-33661 98 51 initial initial ADJ bracis-33661 98 52 residual residual ADJ bracis-33661 98 53 connection connection NOUN bracis-33661 98 54 and and CCONJ bracis-33661 98 55 the the DET bracis-33661 98 56 current current ADJ bracis-33661 98 57 layer layer NOUN bracis-33661 98 58 ’s ’s PART bracis-33661 98 59 transformation transformation NOUN bracis-33661 98 60 . . PUNCT bracis-33661 99 1 the the DET bracis-33661 99 2 identity identity NOUN bracis-33661 99 3 mapping mapping NOUN bracis-33661 99 4 characteristic characteristic NOUN bracis-33661 99 5 further far ADV bracis-33661 99 6 stabilizes stabilize VERB bracis-33661 99 7 the the DET bracis-33661 99 8 training training NOUN bracis-33661 99 9 of of ADP bracis-33661 99 10 deep deep ADJ bracis-33661 99 11 graph graph NOUN bracis-33661 99 12 networks network NOUN bracis-33661 99 13 by by ADP bracis-33661 99 14 incorporating incorporate VERB bracis-33661 99 15 identity identity NOUN bracis-33661 99 16 mappings mapping NOUN bracis-33661 99 17 , , PUNCT bracis-33661 99 18 which which PRON bracis-33661 99 19 preserve preserve VERB bracis-33661 99 20 the the DET bracis-33661 99 21 information information NOUN bracis-33661 99 22 from from ADP bracis-33661 99 23 the the DET bracis-33661 99 24 previous previous ADJ bracis-33661 99 25 layer layer NOUN bracis-33661 99 26 , , PUNCT bracis-33661 99 27 such such ADJ bracis-33661 99 28 as as SCONJ bracis-33661 99 29 follows follow VERB bracis-33661 99 30 : : PUNCT bracis-33661 99 31 $ $ SYM bracis-33661 99 32 $ $ SYM bracis-33661 99 33 \begin{aligned \begin{aligne VERB bracis-33661 99 34 } } PUNCT bracis-33661 99 35 h^{(l+1 h^{(l+1 NOUN bracis-33661 99 36 ) ) PUNCT bracis-33661 99 37 } } PUNCT bracis-33661 100 1 = = SYM bracis-33661 100 2 \sigma \sigma PROPN bracis-33661 100 3 \left \left PROPN bracis-33661 100 4 ( ( PUNCT bracis-33661 100 5 ( ( PUNCT bracis-33661 100 6 1 1 NUM bracis-33661 100 7 \alpha \alpha ADJ bracis-33661 100 8 ) ) PUNCT bracis-33661 100 9 \left \left PROPN bracis-33661 100 10 ( ( PUNCT bracis-33661 100 11 \tilde{d}^{-\frac{1}{2 \tilde{d}^{-\frac{1}{2 PROPN bracis-33661 100 12 } } PUNCT bracis-33661 100 13 } } PUNCT bracis-33661 100 14 \tilde{a \tilde{a NOUN bracis-33661 100 15 } } PUNCT bracis-33661 100 16 \tilde{d}^{-\frac{1}{2 \tilde{d}^{-\frac{1}{2 PROPN bracis-33661 100 17 } } PUNCT bracis-33661 100 18 } } PUNCT bracis-33661 100 19 h^{(l h^{(l NOUN bracis-33661 100 20 ) ) PUNCT bracis-33661 100 21 } } PUNCT bracis-33661 100 22 w^{(l w^{(l NOUN bracis-33661 100 23 ) ) PUNCT bracis-33661 100 24 } } PUNCT bracis-33661 100 25 \right \right PROPN bracis-33661 100 26 ) ) PUNCT bracis-33661 100 27 + + CCONJ bracis-33661 100 28 \alpha \alpha ADJ bracis-33661 100 29 h^{(0 h^{(0 ADJ bracis-33661 100 30 ) ) PUNCT bracis-33661 100 31 } } PUNCT bracis-33661 101 1 + + CCONJ bracis-33661 101 2 \beta \beta PROPN bracis-33661 101 3 h^{(l h^{(l NOUN bracis-33661 101 4 ) ) PUNCT bracis-33661 101 5 } } PUNCT bracis-33661 101 6 \right \right PROPN bracis-33661 101 7 ) ) PUNCT bracis-33661 101 8 ~ ~ NOUN bracis-33661 101 9 , , PUNCT bracis-33661 101 10 \end{aligned}$$ \end{aligned}$$ X bracis-33661 101 11 ( ( PUNCT bracis-33661 101 12 9 9 NUM bracis-33661 101 13 ) ) PUNCT bracis-33661 101 14 where where SCONJ bracis-33661 101 15 \(\beta \(\beta PROPN bracis-33661 101 16 \ \ PROPN bracis-33661 101 17 ) ) PUNCT bracis-33661 101 18 is be AUX bracis-33661 101 19 another another DET bracis-33661 101 20 hyperparameter hyperparameter NOUN bracis-33661 101 21 that that PRON bracis-33661 101 22 controls control VERB bracis-33661 101 23 the the DET bracis-33661 101 24 contribution contribution NOUN bracis-33661 101 25 of of ADP bracis-33661 101 26 the the DET bracis-33661 101 27 identity identity NOUN bracis-33661 101 28 mapping mapping NOUN bracis-33661 101 29 term term NOUN bracis-33661 101 30 \(h^{(l)}\ \(h^{(l)}\ PROPN bracis-33661 101 31 ) ) PUNCT bracis-33661 101 32 , , PUNCT bracis-33661 101 33 further far ADV bracis-33661 101 34 enhancing enhance VERB bracis-33661 101 35 the the DET bracis-33661 101 36 model model NOUN bracis-33661 101 37 ’s ’s PART bracis-33661 101 38 capacity capacity NOUN bracis-33661 101 39 to to PART bracis-33661 101 40 maintain maintain VERB bracis-33661 101 41 feature feature NOUN bracis-33661 101 42 representation representation NOUN bracis-33661 101 43 throughout throughout ADP bracis-33661 101 44 layers layer NOUN bracis-33661 101 45 . . PUNCT bracis-33661 102 1 the the DET bracis-33661 102 2 final final ADJ bracis-33661 102 3 node node ADJ bracis-33661 102 4 representations representation NOUN bracis-33661 102 5 generated generate VERB bracis-33661 102 6 by by ADP bracis-33661 102 7 gcnii gcnii NOUN bracis-33661 102 8 typically typically ADV bracis-33661 102 9 involves involve VERB bracis-33661 102 10 a a DET bracis-33661 102 11 softmax softmax NOUN bracis-33661 102 12 layer layer NOUN bracis-33661 102 13 for for ADP bracis-33661 102 14 normalization normalization NOUN bracis-33661 102 15 and and CCONJ bracis-33661 102 16 classification classification NOUN bracis-33661 102 17 purposes purpose NOUN bracis-33661 102 18 . . PUNCT bracis-33661 103 1 2.2 2.2 NUM bracis-33661 103 2 long long ADJ bracis-33661 103 3 short short ADJ bracis-33661 103 4 - - PUNCT bracis-33661 103 5 term term NOUN bracis-33661 103 6 memory memory NOUN bracis-33661 103 7 lstm lstm NOUN bracis-33661 103 8 is be AUX bracis-33661 103 9 a a DET bracis-33661 103 10 type type NOUN bracis-33661 103 11 of of ADP bracis-33661 103 12 rnn rnn NOUN bracis-33661 103 13 architecture architecture NOUN bracis-33661 103 14 designed design VERB bracis-33661 103 15 to to PART bracis-33661 103 16 address address VERB bracis-33661 103 17 the the DET bracis-33661 103 18 vanishing vanish VERB bracis-33661 103 19 gradient gradient NOUN bracis-33661 103 20 problem problem NOUN bracis-33661 103 21 and and CCONJ bracis-33661 103 22 capture capture VERB bracis-33661 103 23 long long ADJ bracis-33661 103 24 - - PUNCT bracis-33661 103 25 term term NOUN bracis-33661 103 26 dependencies dependency NOUN bracis-33661 103 27 in in ADP bracis-33661 103 28 sequential sequential ADJ bracis-33661 103 29 data datum NOUN bracis-33661 103 30 [ [ X bracis-33661 103 31 11 11 NUM bracis-33661 103 32 , , PUNCT bracis-33661 103 33 13 13 NUM bracis-33661 103 34 ] ] PUNCT bracis-33661 103 35 . . PUNCT bracis-33661 104 1 lstm lstm PROPN bracis-33661 104 2 uses use VERB bracis-33661 104 3 specialized specialized ADJ bracis-33661 104 4 memory memory NOUN bracis-33661 104 5 units unit NOUN bracis-33661 104 6 to to PART bracis-33661 104 7 store store VERB bracis-33661 104 8 and and CCONJ bracis-33661 104 9 retrieve retrieve VERB bracis-33661 104 10 information information NOUN bracis-33661 104 11 selectively selectively ADV bracis-33661 104 12 . . PUNCT bracis-33661 105 1 unlike unlike ADP bracis-33661 105 2 traditional traditional ADJ bracis-33661 105 3 rnns rnn NOUN bracis-33661 105 4 , , PUNCT bracis-33661 105 5 lstm lstm NOUN bracis-33661 105 6 has have VERB bracis-33661 105 7 a a DET bracis-33661 105 8 stronger strong ADJ bracis-33661 105 9 memory memory NOUN bracis-33661 105 10 mechanism mechanism NOUN bracis-33661 105 11 that that PRON bracis-33661 105 12 allows allow VERB bracis-33661 105 13 it it PRON bracis-33661 105 14 to to PART bracis-33661 105 15 retain retain VERB bracis-33661 105 16 information information NOUN bracis-33661 105 17 for for ADP bracis-33661 105 18 longer long ADJ bracis-33661 105 19 periods period NOUN bracis-33661 105 20 and and CCONJ bracis-33661 105 21 overcome overcome VERB bracis-33661 105 22 the the DET bracis-33661 105 23 vanishing vanish VERB bracis-33661 105 24 gradient gradient NOUN bracis-33661 105 25 problem problem NOUN bracis-33661 105 26 . . PUNCT bracis-33661 106 1 the the DET bracis-33661 106 2 lstm lstm PROPN bracis-33661 106 3 model model NOUN bracis-33661 106 4 adopted adopt VERB bracis-33661 106 5 for for ADP bracis-33661 106 6 comparison comparison NOUN bracis-33661 106 7 purposes purpose NOUN bracis-33661 106 8 in in ADP bracis-33661 106 9 the the DET bracis-33661 106 10 present present ADJ bracis-33661 106 11 work work NOUN bracis-33661 106 12 consists consist VERB bracis-33661 106 13 of of ADP bracis-33661 106 14 five five NUM bracis-33661 106 15 key key ADJ bracis-33661 106 16 functions function NOUN bracis-33661 106 17 [ [ X bracis-33661 106 18 11 11 NUM bracis-33661 106 19 ] ] SYM bracis-33661 106 20 : : PUNCT bracis-33661 106 21 1 1 X bracis-33661 106 22 . . X bracis-33661 106 23 forget forget VERB bracis-33661 106 24 gate gate NOUN bracis-33661 106 25 : : PUNCT bracis-33661 106 26 denoted denote VERB bracis-33661 106 27 by by ADP bracis-33661 106 28 \(f_t\ \(f_t\ ADJ bracis-33661 106 29 ) ) PUNCT bracis-33661 106 30 , , PUNCT bracis-33661 106 31 it it PRON bracis-33661 106 32 computes compute VERB bracis-33661 106 33 the the DET bracis-33661 106 34 amount amount NOUN bracis-33661 106 35 of of ADP bracis-33661 106 36 previously previously ADV bracis-33661 106 37 stored store VERB bracis-33661 106 38 information information NOUN bracis-33661 106 39 to to PART bracis-33661 106 40 forget forget VERB bracis-33661 106 41 . . PUNCT bracis-33661 107 1 it it PRON bracis-33661 107 2 is be AUX bracis-33661 107 3 calculated calculate VERB bracis-33661 107 4 by by ADP bracis-33661 107 5 applying apply VERB bracis-33661 107 6 a a DET bracis-33661 107 7 sigmoid sigmoid NOUN bracis-33661 107 8 activation activation NOUN bracis-33661 107 9 function function NOUN bracis-33661 107 10 ( ( PUNCT bracis-33661 107 11 \(\sigma \(\sigma NOUN bracis-33661 107 12 \ \ NOUN bracis-33661 107 13 ) ) PUNCT bracis-33661 107 14 ) ) PUNCT bracis-33661 107 15 to to ADP bracis-33661 107 16 the the DET bracis-33661 107 17 linear linear ADJ bracis-33661 107 18 transformation transformation NOUN bracis-33661 107 19 of of ADP bracis-33661 107 20 the the DET bracis-33661 107 21 concatenated concatenate VERB bracis-33661 107 22 input input NOUN bracis-33661 107 23 ( ( PUNCT bracis-33661 107 24 \([h_{t-1 \([h_{t-1 NOUN bracis-33661 107 25 } } PUNCT bracis-33661 107 26 , , PUNCT bracis-33661 107 27 x_t]\ x_t]\ PROPN bracis-33661 107 28 ) ) PUNCT bracis-33661 107 29 ) ) PUNCT bracis-33661 107 30 using use VERB bracis-33661 107 31 the the DET bracis-33661 107 32 weight weight NOUN bracis-33661 107 33 matrix matrix NOUN bracis-33661 107 34 ( ( PUNCT bracis-33661 107 35 \(w_f\ \(w_f\ NOUN bracis-33661 107 36 ) ) PUNCT bracis-33661 107 37 ) ) PUNCT bracis-33661 107 38 and and CCONJ bracis-33661 107 39 bias bias NOUN bracis-33661 107 40 vector vector NOUN bracis-33661 107 41 ( ( PUNCT bracis-33661 107 42 \(b_f\ \(b_f\ NOUN bracis-33661 107 43 ) ) PUNCT bracis-33661 107 44 ) ) PUNCT bracis-33661 107 45 . . PUNCT bracis-33661 108 1 $ $ SYM bracis-33661 108 2 $ $ SYM bracis-33661 108 3 \begin{aligned \begin{aligne VERB bracis-33661 108 4 } } PUNCT bracis-33661 108 5 f_t f_t PUNCT bracis-33661 109 1 = = SYM bracis-33661 109 2 \sigma \sigma PROPN bracis-33661 109 3 ( ( PUNCT bracis-33661 109 4 w_{xf}x_t w_{xf}x_t X bracis-33661 109 5 + + CCONJ bracis-33661 109 6 w_{hf}h_{t-1 w_{hf}h_{t-1 NOUN bracis-33661 109 7 } } PUNCT bracis-33661 109 8 + + CCONJ bracis-33661 109 9 w_{cf}c_{t-1 w_{cf}c_{t-1 PROPN bracis-33661 109 10 } } PUNCT bracis-33661 109 11 + + ADJ bracis-33661 109 12 b_f b_f SYM bracis-33661 109 13 ) ) PUNCT bracis-33661 109 14 \end{aligned}$$ \end{aligned}$$ X bracis-33661 109 15 ( ( PUNCT bracis-33661 109 16 10 10 NUM bracis-33661 109 17 ) ) SYM bracis-33661 109 18 2 2 NUM bracis-33661 109 19 . . PUNCT bracis-33661 109 20 input input NOUN bracis-33661 109 21 gate gate NOUN bracis-33661 109 22 : : PUNCT bracis-33661 109 23 denoted denote VERB bracis-33661 109 24 by by ADP bracis-33661 109 25 \(i_t\ \(i_t\ NOUN bracis-33661 109 26 ) ) PUNCT bracis-33661 109 27 , , PUNCT bracis-33661 109 28 it it PRON bracis-33661 109 29 determines determine VERB bracis-33661 109 30 the the DET bracis-33661 109 31 amount amount NOUN bracis-33661 109 32 of of ADP bracis-33661 109 33 new new ADJ bracis-33661 109 34 information information NOUN bracis-33661 109 35 to to PART bracis-33661 109 36 be be AUX bracis-33661 109 37 stored store VERB bracis-33661 109 38 in in ADP bracis-33661 109 39 the the DET bracis-33661 109 40 current current ADJ bracis-33661 109 41 memory memory NOUN bracis-33661 109 42 state state NOUN bracis-33661 109 43 . . PUNCT bracis-33661 110 1 it it PRON bracis-33661 110 2 is be AUX bracis-33661 110 3 calculated calculate VERB bracis-33661 110 4 by by ADP bracis-33661 110 5 applying apply VERB bracis-33661 110 6 a a DET bracis-33661 110 7 sigmoid sigmoid NOUN bracis-33661 110 8 activation activation NOUN bracis-33661 110 9 function function NOUN bracis-33661 110 10 ( ( PUNCT bracis-33661 110 11 \(\sigma \(\sigma NOUN bracis-33661 110 12 \ \ NOUN bracis-33661 110 13 ) ) PUNCT bracis-33661 110 14 ) ) PUNCT bracis-33661 110 15 to to ADP bracis-33661 110 16 the the DET bracis-33661 110 17 linear linear ADJ bracis-33661 110 18 transformation transformation NOUN bracis-33661 110 19 of of ADP bracis-33661 110 20 the the DET bracis-33661 110 21 concatenated concatenate VERB bracis-33661 110 22 input input NOUN bracis-33661 110 23 ( ( PUNCT bracis-33661 110 24 \([h_{t-1 \([h_{t-1 NOUN bracis-33661 110 25 } } PUNCT bracis-33661 110 26 , , PUNCT bracis-33661 110 27 x_t]\ x_t]\ PROPN bracis-33661 110 28 ) ) PUNCT bracis-33661 110 29 ) ) PUNCT bracis-33661 110 30 using use VERB bracis-33661 110 31 the the DET bracis-33661 110 32 weight weight NOUN bracis-33661 110 33 matrix matrix NOUN bracis-33661 110 34 ( ( PUNCT bracis-33661 110 35 \(w_i\ \(w_i\ PROPN bracis-33661 110 36 ) ) PUNCT bracis-33661 110 37 ) ) PUNCT bracis-33661 110 38 and and CCONJ bracis-33661 110 39 bias bias NOUN bracis-33661 110 40 vector vector NOUN bracis-33661 110 41 ( ( PUNCT bracis-33661 110 42 \(b_i\ \(b_i\ NOUN bracis-33661 110 43 ) ) PUNCT bracis-33661 110 44 ) ) PUNCT bracis-33661 110 45 . . PUNCT bracis-33661 111 1 $ $ SYM bracis-33661 111 2 $ $ SYM bracis-33661 111 3 \begin{aligned \begin{aligne VERB bracis-33661 111 4 } } PUNCT bracis-33661 111 5 i_t i_t PUNCT bracis-33661 111 6 = = SYM bracis-33661 111 7 \sigma \sigma PROPN bracis-33661 111 8 ( ( PUNCT bracis-33661 111 9 w_{xi}x_t w_{xi}x_t NOUN bracis-33661 111 10 + + CCONJ bracis-33661 111 11 w_{hi}h_{t-1 w_{hi}h_{t-1 ADV bracis-33661 111 12 } } PUNCT bracis-33661 111 13 + + CCONJ bracis-33661 111 14 w_{ci}c_{t-1 w_{ci}c_{t-1 ADJ bracis-33661 111 15 } } PUNCT bracis-33661 111 16 + + CCONJ bracis-33661 111 17 b_i b_i NUM bracis-33661 111 18 ) ) PUNCT bracis-33661 111 19 \end{aligned}$$ \end{aligned}$$ X bracis-33661 111 20 ( ( PUNCT bracis-33661 111 21 11 11 NUM bracis-33661 111 22 ) ) SYM bracis-33661 111 23 3 3 NUM bracis-33661 111 24 . . PUNCT bracis-33661 111 25 candidate candidate NOUN bracis-33661 111 26 cell cell NOUN bracis-33661 111 27 state state NOUN bracis-33661 111 28 : : PUNCT bracis-33661 111 29 denoted denote VERB bracis-33661 111 30 by by ADP bracis-33661 111 31 \(\widetilde{c}t\ \(\widetilde{c}t\ PROPN bracis-33661 111 32 ) ) PUNCT bracis-33661 111 33 , , PUNCT bracis-33661 111 34 it it PRON bracis-33661 111 35 represents represent VERB bracis-33661 111 36 the the DET bracis-33661 111 37 new new ADJ bracis-33661 111 38 information information NOUN bracis-33661 111 39 that that PRON bracis-33661 111 40 can can AUX bracis-33661 111 41 be be AUX bracis-33661 111 42 added add VERB bracis-33661 111 43 to to ADP bracis-33661 111 44 the the DET bracis-33661 111 45 memory memory NOUN bracis-33661 111 46 . . PUNCT bracis-33661 112 1 it it PRON bracis-33661 112 2 is be AUX bracis-33661 112 3 calculated calculate VERB bracis-33661 112 4 by by ADP bracis-33661 112 5 applying apply VERB bracis-33661 112 6 a a DET bracis-33661 112 7 hyperbolic hyperbolic ADJ bracis-33661 112 8 tangent tangent NOUN bracis-33661 112 9 activation activation NOUN bracis-33661 112 10 function function VERB bracis-33661 112 11 to to ADP bracis-33661 112 12 the the DET bracis-33661 112 13 linear linear ADJ bracis-33661 112 14 transformation transformation NOUN bracis-33661 112 15 of of ADP bracis-33661 112 16 the the DET bracis-33661 112 17 concatenated concatenate VERB bracis-33661 112 18 input input NOUN bracis-33661 112 19 ( ( PUNCT bracis-33661 112 20 \([h{t-1 \([h{t-1 ADV bracis-33661 112 21 } } PUNCT bracis-33661 112 22 , , PUNCT bracis-33661 112 23 x_t]\ x_t]\ PROPN bracis-33661 112 24 ) ) PUNCT bracis-33661 112 25 ) ) PUNCT bracis-33661 112 26 using use VERB bracis-33661 112 27 the the DET bracis-33661 112 28 weight weight NOUN bracis-33661 112 29 matrix matrix NOUN bracis-33661 112 30 ( ( PUNCT bracis-33661 112 31 \(w_c\ \(w_c\ NOUN bracis-33661 112 32 ) ) PUNCT bracis-33661 112 33 ) ) PUNCT bracis-33661 112 34 and and CCONJ bracis-33661 112 35 bias bias NOUN bracis-33661 112 36 vector vector NOUN bracis-33661 112 37 ( ( PUNCT bracis-33661 112 38 \(b_c\ \(b_c\ PROPN bracis-33661 112 39 ) ) PUNCT bracis-33661 112 40 ) ) PUNCT bracis-33661 112 41 . . PUNCT bracis-33661 113 1 $ $ SYM bracis-33661 113 2 $ $ SYM bracis-33661 113 3 \begin{aligned \begin{aligne VERB bracis-33661 113 4 } } PUNCT bracis-33661 113 5 \widetilde{c}t \widetilde{c}t X bracis-33661 113 6 = = PUNCT bracis-33661 113 7 \tanh \tanh PROPN bracis-33661 113 8 ( ( PUNCT bracis-33661 113 9 w_{xc}x_t w_{xc}x_t PROPN bracis-33661 113 10 + + CCONJ bracis-33661 113 11 w_{hc}h_{t-1 w_{hc}h_{t-1 VERB bracis-33661 113 12 } } PUNCT bracis-33661 113 13 + + CCONJ bracis-33661 113 14 b_c b_c X bracis-33661 113 15 ) ) PUNCT bracis-33661 113 16 \end{aligned}$$ \end{aligned}$$ X bracis-33661 113 17 ( ( PUNCT bracis-33661 113 18 12 12 NUM bracis-33661 113 19 ) ) PUNCT bracis-33661 113 20 4 4 NUM bracis-33661 113 21 . . X bracis-33661 113 22 cell cell NOUN bracis-33661 113 23 state state NOUN bracis-33661 113 24 update update NOUN bracis-33661 113 25 : : PUNCT bracis-33661 113 26 it it PRON bracis-33661 113 27 combines combine VERB bracis-33661 113 28 the the DET bracis-33661 113 29 forget forget NOUN bracis-33661 113 30 gate gate NOUN bracis-33661 113 31 activation activation NOUN bracis-33661 113 32 vector vector NOUN bracis-33661 113 33 ( ( PUNCT bracis-33661 113 34 \(f_t\ \(f_t\ ADJ bracis-33661 113 35 ) ) PUNCT bracis-33661 113 36 ) ) PUNCT bracis-33661 113 37 , , PUNCT bracis-33661 113 38 the the DET bracis-33661 113 39 input input NOUN bracis-33661 113 40 gate gate NOUN bracis-33661 113 41 activation activation NOUN bracis-33661 113 42 vector vector NOUN bracis-33661 113 43 ( ( PUNCT bracis-33661 113 44 \(i_t\ \(i_t\ NOUN bracis-33661 113 45 ) ) PUNCT bracis-33661 113 46 ) ) PUNCT bracis-33661 113 47 , , PUNCT bracis-33661 113 48 and and CCONJ bracis-33661 113 49 the the DET bracis-33661 113 50 candidate candidate NOUN bracis-33661 113 51 cell cell NOUN bracis-33661 113 52 state state NOUN bracis-33661 113 53 ( ( PUNCT bracis-33661 113 54 \(\widetilde{c}_t\ \(\widetilde{c}_t\ NOUN bracis-33661 113 55 ) ) PUNCT bracis-33661 113 56 ) ) PUNCT bracis-33661 113 57 to to PART bracis-33661 113 58 update update VERB bracis-33661 113 59 the the DET bracis-33661 113 60 current current ADJ bracis-33661 113 61 memory memory NOUN bracis-33661 113 62 state state NOUN bracis-33661 113 63 ( ( PUNCT bracis-33661 113 64 \(c_t\ \(c_t\ NOUN bracis-33661 113 65 ) ) PUNCT bracis-33661 113 66 ) ) PUNCT bracis-33661 113 67 . . PUNCT bracis-33661 114 1 element element ADJ bracis-33661 114 2 - - ADJ bracis-33661 114 3 wise wise ADJ bracis-33661 114 4 multiplication multiplication NOUN bracis-33661 114 5 ( ( PUNCT bracis-33661 114 6 \(\odot \(\odot NOUN bracis-33661 114 7 \ \ NOUN bracis-33661 114 8 ) ) PUNCT bracis-33661 114 9 ) ) PUNCT bracis-33661 114 10 is be AUX bracis-33661 114 11 performed perform VERB bracis-33661 114 12 between between ADP bracis-33661 114 13 the the DET bracis-33661 114 14 forget forget NOUN bracis-33661 114 15 gate gate NOUN bracis-33661 114 16 and and CCONJ bracis-33661 114 17 the the DET bracis-33661 114 18 previous previous ADJ bracis-33661 114 19 memory memory NOUN bracis-33661 114 20 state state NOUN bracis-33661 114 21 , , PUNCT bracis-33661 114 22 and and CCONJ bracis-33661 114 23 between between ADP bracis-33661 114 24 the the DET bracis-33661 114 25 input input NOUN bracis-33661 114 26 gate gate NOUN bracis-33661 114 27 and and CCONJ bracis-33661 114 28 the the DET bracis-33661 114 29 candidate candidate NOUN bracis-33661 114 30 cell cell NOUN bracis-33661 114 31 state state NOUN bracis-33661 114 32 . . PUNCT bracis-33661 115 1 $ $ SYM bracis-33661 115 2 $ $ SYM bracis-33661 115 3 \begin{aligned \begin{aligne VERB bracis-33661 115 4 } } PUNCT bracis-33661 115 5 c_t c_t PUNCT bracis-33661 115 6 = = PUNCT bracis-33661 116 1 f_tc_{t-1 f_tc_{t-1 PROPN bracis-33661 116 2 } } PUNCT bracis-33661 116 3 + + CCONJ bracis-33661 116 4 i_t\widetilde{c}_t i_t\widetilde{c}_t NUM bracis-33661 116 5 \end{aligned}$$ \end{aligned}$$ X bracis-33661 116 6 ( ( PUNCT bracis-33661 116 7 13 13 NUM bracis-33661 116 8 ) ) SYM bracis-33661 116 9 5 5 NUM bracis-33661 116 10 . . PUNCT bracis-33661 116 11 output output PROPN bracis-33661 116 12 gate gate PROPN bracis-33661 116 13 : : PUNCT bracis-33661 116 14 denoted denote VERB bracis-33661 116 15 by by ADP bracis-33661 116 16 \(o_t\ \(o_t\ PROPN bracis-33661 116 17 ) ) PUNCT bracis-33661 116 18 , , PUNCT bracis-33661 116 19 it it PRON bracis-33661 116 20 controls control VERB bracis-33661 116 21 the the DET bracis-33661 116 22 amount amount NOUN bracis-33661 116 23 of of ADP bracis-33661 116 24 information information NOUN bracis-33661 116 25 to to PART bracis-33661 116 26 be be AUX bracis-33661 116 27 output output NOUN bracis-33661 116 28 from from ADP bracis-33661 116 29 the the DET bracis-33661 116 30 memory memory NOUN bracis-33661 116 31 state state NOUN bracis-33661 116 32 . . PUNCT bracis-33661 117 1 it it PRON bracis-33661 117 2 is be AUX bracis-33661 117 3 calculated calculate VERB bracis-33661 117 4 by by ADP bracis-33661 117 5 applying apply VERB bracis-33661 117 6 a a DET bracis-33661 117 7 sigmoid sigmoid NOUN bracis-33661 117 8 activation activation NOUN bracis-33661 117 9 function function NOUN bracis-33661 117 10 ( ( PUNCT bracis-33661 117 11 \(\sigma \(\sigma NOUN bracis-33661 117 12 \ \ NOUN bracis-33661 117 13 ) ) PUNCT bracis-33661 117 14 ) ) PUNCT bracis-33661 117 15 to to ADP bracis-33661 117 16 the the DET bracis-33661 117 17 linear linear ADJ bracis-33661 117 18 transformation transformation NOUN bracis-33661 117 19 of of ADP bracis-33661 117 20 the the DET bracis-33661 117 21 concatenated concatenate VERB bracis-33661 117 22 input input NOUN bracis-33661 117 23 ( ( PUNCT bracis-33661 117 24 \([h_{t-1 \([h_{t-1 NOUN bracis-33661 117 25 } } PUNCT bracis-33661 117 26 , , PUNCT bracis-33661 117 27 x_t]\ x_t]\ PROPN bracis-33661 117 28 ) ) PUNCT bracis-33661 117 29 ) ) PUNCT bracis-33661 117 30 using use VERB bracis-33661 117 31 the the DET bracis-33661 117 32 weight weight NOUN bracis-33661 117 33 matrix matrix NOUN bracis-33661 117 34 ( ( PUNCT bracis-33661 117 35 \(w_o\ \(w_o\ NOUN bracis-33661 117 36 ) ) PUNCT bracis-33661 117 37 ) ) PUNCT bracis-33661 117 38 and and CCONJ bracis-33661 117 39 bias bias NOUN bracis-33661 117 40 vector vector NOUN bracis-33661 117 41 ( ( PUNCT bracis-33661 117 42 \(b_o\ \(b_o\ ADJ bracis-33661 117 43 ) ) PUNCT bracis-33661 117 44 ) ) PUNCT bracis-33661 117 45 . . PUNCT bracis-33661 118 1 $ $ SYM bracis-33661 118 2 $ $ SYM bracis-33661 118 3 \begin{aligned \begin{aligne VERB bracis-33661 118 4 } } PUNCT bracis-33661 118 5 o_t o_t PUNCT bracis-33661 119 1 = = SYM bracis-33661 119 2 \sigma \sigma PROPN bracis-33661 119 3 ( ( PUNCT bracis-33661 119 4 w_{xo}x_t w_{xo}x_t NOUN bracis-33661 119 5 + + CCONJ bracis-33661 119 6 w_{ho}h_{t-1 w_{ho}h_{t-1 ADV bracis-33661 119 7 } } PUNCT bracis-33661 119 8 + + CCONJ bracis-33661 119 9 w_{co}c_t w_{co}c_t PROPN bracis-33661 119 10 + + CCONJ bracis-33661 119 11 b_o b_o NOUN bracis-33661 119 12 ) ) PUNCT bracis-33661 119 13 \end{aligned}$$ \end{aligned}$$ X bracis-33661 119 14 ( ( PUNCT bracis-33661 119 15 14 14 NUM bracis-33661 119 16 ) ) PUNCT bracis-33661 119 17 the the DET bracis-33661 119 18 lstm lstm PROPN bracis-33661 119 19 model model NOUN bracis-33661 119 20 involves involve VERB bracis-33661 119 21 various various ADJ bracis-33661 119 22 parameters parameter NOUN bracis-33661 119 23 , , PUNCT bracis-33661 119 24 including include VERB bracis-33661 119 25 weight weight NOUN bracis-33661 119 26 matrices matrix NOUN bracis-33661 119 27 ( ( PUNCT bracis-33661 119 28 \(w_f\ \(w_f\ NOUN bracis-33661 119 29 ) ) PUNCT bracis-33661 119 30 , , PUNCT bracis-33661 119 31 \(w_i\ \(w_i\ PROPN bracis-33661 119 32 ) ) PUNCT bracis-33661 119 33 , , PUNCT bracis-33661 119 34 \(w_c\ \(w_c\ NOUN bracis-33661 119 35 ) ) PUNCT bracis-33661 119 36 , , PUNCT bracis-33661 119 37 \(w_o\ \(w_o\ NOUN bracis-33661 119 38 ) ) PUNCT bracis-33661 119 39 ) ) PUNCT bracis-33661 119 40 , , PUNCT bracis-33661 119 41 bias bias NOUN bracis-33661 119 42 vectors vector NOUN bracis-33661 119 43 ( ( PUNCT bracis-33661 119 44 \(b_f\ \(b_f\ NOUN bracis-33661 119 45 ) ) PUNCT bracis-33661 119 46 , , PUNCT bracis-33661 119 47 \(b_i\ \(b_i\ NOUN bracis-33661 119 48 ) ) PUNCT bracis-33661 119 49 , , PUNCT bracis-33661 119 50 \(b_c\ \(b_c\ PROPN bracis-33661 119 51 ) ) PUNCT bracis-33661 119 52 , , PUNCT bracis-33661 119 53 \(b_o\ \(b_o\ NOUN bracis-33661 119 54 ) ) PUNCT bracis-33661 119 55 ) ) PUNCT bracis-33661 119 56 , , PUNCT bracis-33661 119 57 and and CCONJ bracis-33661 119 58 the the DET bracis-33661 119 59 hidden hidden ADJ bracis-33661 119 60 state state NOUN bracis-33661 119 61 \(h_t\ \(h_t\ NOUN bracis-33661 119 62 ) ) PUNCT bracis-33661 119 63 . . PUNCT bracis-33661 120 1 these these DET bracis-33661 120 2 parameters parameter NOUN bracis-33661 120 3 are be AUX bracis-33661 120 4 learned learn VERB bracis-33661 120 5 during during ADP bracis-33661 120 6 the the DET bracis-33661 120 7 training training NOUN bracis-33661 120 8 process process NOUN bracis-33661 120 9 and and CCONJ bracis-33661 120 10 play play VERB bracis-33661 120 11 a a DET bracis-33661 120 12 crucial crucial ADJ bracis-33661 120 13 role role NOUN bracis-33661 120 14 in in ADP bracis-33661 120 15 the the DET bracis-33661 120 16 network network NOUN bracis-33661 120 17 ’s ’s PART bracis-33661 120 18 ability ability NOUN bracis-33661 120 19 to to PART bracis-33661 120 20 capture capture VERB bracis-33661 120 21 and and CCONJ bracis-33661 120 22 model model VERB bracis-33661 120 23 complex complex ADJ bracis-33661 120 24 temporal temporal ADJ bracis-33661 120 25 dependencies dependency NOUN bracis-33661 120 26 . . PUNCT bracis-33661 121 1 3 3 NUM bracis-33661 121 2 model model NOUN bracis-33661 121 3 description description NOUN bracis-33661 121 4 this this DET bracis-33661 121 5 section section NOUN bracis-33661 121 6 provides provide VERB bracis-33661 121 7 a a DET bracis-33661 121 8 detailed detailed ADJ bracis-33661 121 9 description description NOUN bracis-33661 121 10 of of ADP bracis-33661 121 11 all all DET bracis-33661 121 12 the the DET bracis-33661 121 13 components component NOUN bracis-33661 121 14 developed develop VERB bracis-33661 121 15 to to PART bracis-33661 121 16 enable enable VERB bracis-33661 121 17 the the DET bracis-33661 121 18 gnn gnn PROPN bracis-33661 121 19 learning learning NOUN bracis-33661 121 20 over over ADP bracis-33661 121 21 eeg eeg PROPN bracis-33661 121 22 data datum NOUN bracis-33661 121 23 . . PUNCT bracis-33661 122 1 we we PRON bracis-33661 122 2 also also ADV bracis-33661 122 3 present present VERB bracis-33661 122 4 the the DET bracis-33661 122 5 dataset dataset NOUN bracis-33661 122 6 adopted adopt VERB bracis-33661 122 7 in in ADP bracis-33661 122 8 this this DET bracis-33661 122 9 study study NOUN bracis-33661 122 10 , , PUNCT bracis-33661 122 11 the the DET bracis-33661 122 12 formulation formulation NOUN bracis-33661 122 13 and and CCONJ bracis-33661 122 14 pseudocode pseudocode NOUN bracis-33661 122 15 of of ADP bracis-33661 122 16 the the DET bracis-33661 122 17 modeling modeling NOUN bracis-33661 122 18 approach approach NOUN bracis-33661 122 19 to to PART bracis-33661 122 20 represent represent VERB bracis-33661 122 21 eeg eeg NOUN bracis-33661 122 22 data datum NOUN bracis-33661 122 23 into into ADP bracis-33661 122 24 graphs graph NOUN bracis-33661 122 25 , , PUNCT bracis-33661 122 26 the the DET bracis-33661 122 27 selection selection NOUN bracis-33661 122 28 of of ADP bracis-33661 122 29 hyper hyper NOUN bracis-33661 122 30 - - NOUN bracis-33661 122 31 parameters parameter NOUN bracis-33661 122 32 for for ADP bracis-33661 122 33 the the DET bracis-33661 122 34 gnn gnn PROPN bracis-33661 122 35 networks network NOUN bracis-33661 122 36 , , PUNCT bracis-33661 122 37 and and CCONJ bracis-33661 122 38 the the DET bracis-33661 122 39 validation validation NOUN bracis-33661 122 40 process process NOUN bracis-33661 122 41 employed employ VERB bracis-33661 122 42 . . PUNCT bracis-33661 123 1 an an DET bracis-33661 123 2 overview overview NOUN bracis-33661 123 3 about about ADP bracis-33661 123 4 the the DET bracis-33661 123 5 complete complete ADJ bracis-33661 123 6 pipeline pipeline NOUN bracis-33661 123 7 of of ADP bracis-33661 123 8 the the DET bracis-33661 123 9 model model NOUN bracis-33661 123 10 can can AUX bracis-33661 123 11 be be AUX bracis-33661 123 12 see see VERB bracis-33661 123 13 in in ADP bracis-33661 123 14 the the DET bracis-33661 123 15 fig fig NOUN bracis-33661 123 16 . . PUNCT bracis-33661 123 17     SPACE bracis-33661 124 1 1 1 NUM bracis-33661 124 2 , , PUNCT bracis-33661 124 3 and and CCONJ bracis-33661 124 4 all all DET bracis-33661 124 5 its its PRON bracis-33661 124 6 respective respective ADJ bracis-33661 124 7 steps step NOUN bracis-33661 124 8 will will AUX bracis-33661 124 9 be be AUX bracis-33661 124 10 detailed detail VERB bracis-33661 124 11 in in ADP bracis-33661 124 12 this this DET bracis-33661 124 13 section section NOUN bracis-33661 124 14 . . PUNCT bracis-33661 125 1 fig fig NOUN bracis-33661 125 2 . . PUNCT bracis-33661 126 1 1 1 X bracis-33661 126 2 . . X bracis-33661 126 3 complete complete ADJ bracis-33661 126 4 pipeline pipeline NOUN bracis-33661 126 5 fluxogram fluxogram NOUN bracis-33661 126 6 . . PUNCT bracis-33661 127 1 ( ( PUNCT bracis-33661 127 2 a a X bracis-33661 127 3 ) ) PUNCT bracis-33661 127 4 electroencephalogram electroencephalogram NOUN bracis-33661 127 5 before before ADP bracis-33661 127 6 preprocessing preprocesse VERB bracis-33661 127 7 step step NOUN bracis-33661 127 8 ; ; PUNCT bracis-33661 127 9 ( ( PUNCT bracis-33661 127 10 b b X bracis-33661 127 11 ) ) PUNCT bracis-33661 127 12 eeg eeg NOUN bracis-33661 127 13 exam exam NOUN bracis-33661 127 14 as as ADP bracis-33661 127 15 a a DET bracis-33661 127 16 graph graph NOUN bracis-33661 127 17 , , PUNCT bracis-33661 127 18 with with ADP bracis-33661 127 19 each each DET bracis-33661 127 20 electrode electrode NOUN bracis-33661 127 21 ’s ’s PART bracis-33661 127 22 respective respective ADJ bracis-33661 127 23 temporal temporal ADJ bracis-33661 127 24 values value NOUN bracis-33661 127 25 as as ADP bracis-33661 127 26 vertices vertex NOUN bracis-33661 127 27 , , PUNCT bracis-33661 127 28 and and CCONJ bracis-33661 127 29 the the DET bracis-33661 127 30 patient patient NOUN bracis-33661 127 31 ’s ’s PART bracis-33661 127 32 actual actual ADJ bracis-33661 127 33 outcome outcome NOUN bracis-33661 127 34 stored store VERB bracis-33661 127 35 as as ADP bracis-33661 127 36 a a DET bracis-33661 127 37 target target NOUN bracis-33661 127 38 class class NOUN bracis-33661 127 39 of of ADP bracis-33661 127 40 the the DET bracis-33661 127 41 graph graph NOUN bracis-33661 127 42 ; ; PUNCT bracis-33661 127 43 ( ( PUNCT bracis-33661 127 44 c c X bracis-33661 127 45 ) ) PUNCT bracis-33661 127 46 graph graph NOUN bracis-33661 127 47 structure structure NOUN bracis-33661 127 48 after after ADP bracis-33661 127 49 modeling model VERB bracis-33661 127 50 the the DET bracis-33661 127 51 exam exam NOUN bracis-33661 127 52 data datum NOUN bracis-33661 127 53 : : PUNCT bracis-33661 127 54 each each DET bracis-33661 127 55 electrode electrode NOUN bracis-33661 127 56 is be AUX bracis-33661 127 57 a a DET bracis-33661 127 58 node node NOUN bracis-33661 127 59 , , PUNCT bracis-33661 127 60 and and CCONJ bracis-33661 127 61 all all DET bracis-33661 127 62 the the DET bracis-33661 127 63 edges edge NOUN bracis-33661 127 64 have have VERB bracis-33661 127 65 the the DET bracis-33661 127 66 same same ADJ bracis-33661 127 67 weights weight NOUN bracis-33661 127 68 ( ( PUNCT bracis-33661 127 69 w w NOUN bracis-33661 127 70 = = NOUN bracis-33661 127 71 1 1 NUM bracis-33661 127 72 ) ) PUNCT bracis-33661 127 73 ; ; PUNCT bracis-33661 127 74 ( ( PUNCT bracis-33661 127 75 d d X bracis-33661 127 76 ) ) PUNCT bracis-33661 127 77 graph graph NOUN bracis-33661 127 78 neural neural ADJ bracis-33661 127 79 network network NOUN bracis-33661 127 80 representation representation NOUN bracis-33661 127 81 , , PUNCT bracis-33661 127 82 created create VERB bracis-33661 127 83 based base VERB bracis-33661 127 84 on on ADP bracis-33661 127 85 the the DET bracis-33661 127 86 convolutions convolution NOUN bracis-33661 127 87 principles principle NOUN bracis-33661 127 88 ; ; PUNCT bracis-33661 127 89 and and CCONJ bracis-33661 127 90 ( ( PUNCT bracis-33661 127 91 e e X bracis-33661 127 92 ) ) PUNCT bracis-33661 127 93 the the DET bracis-33661 127 94 final final ADJ bracis-33661 127 95 result result NOUN bracis-33661 127 96 ( ( PUNCT bracis-33661 127 97 outcome outcome NOUN bracis-33661 127 98 probability probability NOUN bracis-33661 127 99 ) ) PUNCT bracis-33661 127 100 of of ADP bracis-33661 127 101 the the DET bracis-33661 127 102 pipeline pipeline NOUN bracis-33661 127 103 . . PUNCT bracis-33661 128 1 full full ADJ bracis-33661 128 2 size size NOUN bracis-33661 128 3 image image NOUN bracis-33661 128 4 3.1 3.1 NUM bracis-33661 128 5 eeg eeg NOUN bracis-33661 128 6 dataset dataset VERB bracis-33661 128 7 the the DET bracis-33661 128 8 dataset dataset NOUN bracis-33661 128 9 used use VERB bracis-33661 128 10 in in ADP bracis-33661 128 11 this this DET bracis-33661 128 12 study study NOUN bracis-33661 128 13 is be AUX bracis-33661 128 14 private private ADJ bracis-33661 128 15 . . PUNCT bracis-33661 129 1 with with ADP bracis-33661 129 2 approval approval NOUN bracis-33661 129 3 of of ADP bracis-33661 129 4 the the DET bracis-33661 129 5 research research NOUN bracis-33661 129 6 ethics ethic NOUN bracis-33661 129 7 committee committee PROPN bracis-33661 129 8 of of ADP bracis-33661 129 9 the the DET bracis-33661 129 10 federal federal ADJ bracis-33661 129 11 university university PROPN bracis-33661 129 12 of of ADP bracis-33661 129 13 uberlândia uberlândia PROPN bracis-33661 129 14 ( ( PUNCT bracis-33661 129 15 ufu ufu PROPN bracis-33661 129 16 ) ) PUNCT bracis-33661 129 17 , , PUNCT bracis-33661 129 18 eeg eeg NOUN bracis-33661 129 19 exams exam NOUN bracis-33661 129 20 recorded record VERB bracis-33661 129 21 from from ADP bracis-33661 129 22 comatose comatose ADJ bracis-33661 129 23 patients patient NOUN bracis-33661 129 24 in in ADP bracis-33661 129 25 the the DET bracis-33661 129 26 intensive intensive ADJ bracis-33661 129 27 care care NOUN bracis-33661 129 28 unit unit NOUN bracis-33661 129 29 of of ADP bracis-33661 129 30 the the DET bracis-33661 129 31 ufu ufu PROPN bracis-33661 129 32 clinical clinical ADJ bracis-33661 129 33 hospital hospital NOUN bracis-33661 129 34 were be AUX bracis-33661 129 35 collected collect VERB bracis-33661 129 36 . . PUNCT bracis-33661 130 1 table table NOUN bracis-33661 130 2     SPACE bracis-33661 130 3 1 1 NUM bracis-33661 130 4 shows show VERB bracis-33661 130 5 the the DET bracis-33661 130 6 class class NOUN bracis-33661 130 7 distribution distribution NOUN bracis-33661 130 8 . . PUNCT bracis-33661 131 1 eeg eeg NOUN bracis-33661 131 2 exams exam NOUN bracis-33661 131 3 considered consider VERB bracis-33661 131 4 the the DET bracis-33661 131 5 10–20 10–20 NUM bracis-33661 131 6 electrode electrode NOUN bracis-33661 131 7 system system NOUN bracis-33661 131 8 , , PUNCT bracis-33661 131 9 which which PRON bracis-33661 131 10 the the DET bracis-33661 131 11 disposition disposition NOUN bracis-33661 131 12 along along ADP bracis-33661 131 13 the the DET bracis-33661 131 14 scalp scalp NOUN bracis-33661 131 15 is be AUX bracis-33661 131 16 shown show VERB bracis-33661 131 17 by by ADP bracis-33661 131 18 fig fig NOUN bracis-33661 131 19 . . PUNCT bracis-33661 131 20     SPACE bracis-33661 132 1 2 2 NUM bracis-33661 132 2 . . PUNCT bracis-33661 133 1 after after ADP bracis-33661 133 2 the the DET bracis-33661 133 3 preprocessing preprocessing NOUN bracis-33661 133 4 , , PUNCT bracis-33661 133 5 each each DET bracis-33661 133 6 eeg eeg NOUN bracis-33661 133 7 exam exam NOUN bracis-33661 133 8 is be AUX bracis-33661 133 9 denoted denote VERB bracis-33661 133 10 by by ADP bracis-33661 133 11 10 10 NUM bracis-33661 133 12 two two NUM bracis-33661 133 13 - - PUNCT bracis-33661 133 14 second second NOUN bracis-33661 133 15 segments segment NOUN bracis-33661 133 16 selected select VERB bracis-33661 133 17 through through ADP bracis-33661 133 18 visual visual ADJ bracis-33661 133 19 inspection inspection NOUN bracis-33661 133 20 by by ADP bracis-33661 133 21 a a DET bracis-33661 133 22 specialized specialized ADJ bracis-33661 133 23 physician physician NOUN bracis-33661 133 24 in in ADP bracis-33661 133 25 the the DET bracis-33661 133 26 field field NOUN bracis-33661 133 27 . . PUNCT bracis-33661 134 1 along along ADP bracis-33661 134 2 with with ADP bracis-33661 134 3 the the DET bracis-33661 134 4 results result NOUN bracis-33661 134 5 , , PUNCT bracis-33661 134 6 the the DET bracis-33661 134 7 datasets dataset NOUN bracis-33661 134 8 were be AUX bracis-33661 134 9 labeled label VERB bracis-33661 134 10 regarding regard VERB bracis-33661 134 11 the the DET bracis-33661 134 12 outcome outcome NOUN bracis-33661 134 13 of of ADP bracis-33661 134 14 the the DET bracis-33661 134 15 comatose comatose ADJ bracis-33661 134 16 patient patient NOUN bracis-33661 134 17 , , PUNCT bracis-33661 134 18 considering consider VERB bracis-33661 134 19 the the DET bracis-33661 134 20 following follow VERB bracis-33661 134 21 groups group NOUN bracis-33661 134 22 : : PUNCT bracis-33661 134 23 favorable favorable ADJ bracis-33661 134 24 outcome outcome NOUN bracis-33661 134 25 is be AUX bracis-33661 134 26 related relate VERB bracis-33661 134 27 to to ADP bracis-33661 134 28 patients patient NOUN bracis-33661 134 29 with with ADP bracis-33661 134 30 positive positive ADJ bracis-33661 134 31 outcome outcome NOUN bracis-33661 134 32 . . PUNCT bracis-33661 135 1 unfavorable unfavorable ADJ bracis-33661 135 2 outcome outcome NOUN bracis-33661 135 3 is be AUX bracis-33661 135 4 related relate VERB bracis-33661 135 5 to to ADP bracis-33661 135 6 patients patient NOUN bracis-33661 135 7 with with ADP bracis-33661 135 8 clinical clinical ADJ bracis-33661 135 9 or or CCONJ bracis-33661 135 10 brain brain NOUN bracis-33661 135 11 death death NOUN bracis-33661 135 12 ; ; PUNCT bracis-33661 135 13 table table NOUN bracis-33661 135 14 1 1 NUM bracis-33661 135 15 . . PUNCT bracis-33661 136 1 outcomes outcome NOUN bracis-33661 136 2 final final ADJ bracis-33661 136 3 class class NOUN bracis-33661 136 4 distributionfull distributionfull NOUN bracis-33661 136 5 size size NOUN bracis-33661 136 6 table table NOUN bracis-33661 136 7 fig fig NOUN bracis-33661 136 8 . . PUNCT bracis-33661 137 1 2 2 X bracis-33661 137 2 . . NUM bracis-33661 137 3 electrodes electrode VERB bracis-33661 137 4 position position NOUN bracis-33661 137 5 on on ADP bracis-33661 137 6 patients patient NOUN bracis-33661 137 7 ’ ' PUNCT bracis-33661 137 8 cortex cortex NOUN bracis-33661 137 9 . . PUNCT bracis-33661 138 1 full full ADJ bracis-33661 138 2 size size NOUN bracis-33661 138 3 image image NOUN bracis-33661 138 4 eeg eeg NOUN bracis-33661 138 5 data datum NOUN bracis-33661 138 6 were be AUX bracis-33661 138 7 collected collect VERB bracis-33661 138 8 at at ADP bracis-33661 138 9 a a DET bracis-33661 138 10 frequency frequency NOUN bracis-33661 138 11 of of ADP bracis-33661 138 12 100 100 NUM bracis-33661 138 13 to to PART bracis-33661 138 14 400 400 NUM bracis-33661 138 15     SPACE bracis-33661 138 16 hz hz VERB bracis-33661 138 17 for for ADP bracis-33661 138 18 the the DET bracis-33661 138 19 given give VERB bracis-33661 138 20 electrodes electrode NOUN bracis-33661 138 21 [ [ X bracis-33661 138 22 fp1 fp1 NOUN bracis-33661 138 23 , , PUNCT bracis-33661 138 24 fp2 fp2 PROPN bracis-33661 138 25 , , PUNCT bracis-33661 138 26 f7 f7 PROPN bracis-33661 138 27 , , PUNCT bracis-33661 138 28 f3 f3 PROPN bracis-33661 138 29 , , PUNCT bracis-33661 138 30 fz fz NOUN bracis-33661 138 31 , , PUNCT bracis-33661 138 32 f4 f4 PROPN bracis-33661 138 33 , , PUNCT bracis-33661 138 34 f8 f8 PROPN bracis-33661 138 35 , , PUNCT bracis-33661 138 36 t3 t3 PROPN bracis-33661 138 37 , , PUNCT bracis-33661 138 38 c3 c3 PROPN bracis-33661 138 39 , , PUNCT bracis-33661 138 40 cz cz NOUN bracis-33661 138 41 , , PUNCT bracis-33661 138 42 c4 c4 NOUN bracis-33661 138 43 , , PUNCT bracis-33661 138 44 t4 t4 PROPN bracis-33661 138 45 , , PUNCT bracis-33661 138 46 t5 t5 PROPN bracis-33661 138 47 , , PUNCT bracis-33661 138 48 p3 p3 PROPN bracis-33661 138 49 , , PUNCT bracis-33661 138 50 pz pz NOUN bracis-33661 138 51 , , PUNCT bracis-33661 138 52 p4 p4 ADJ bracis-33661 138 53 , , PUNCT bracis-33661 138 54 t6 t6 PROPN bracis-33661 138 55 , , PUNCT bracis-33661 138 56 o1 o1 PROPN bracis-33661 138 57 , , PUNCT bracis-33661 138 58 o2 o2 PROPN bracis-33661 138 59 ] ] PUNCT bracis-33661 138 60 . . PUNCT bracis-33661 139 1 to to PART bracis-33661 139 2 address address VERB bracis-33661 139 3 the the DET bracis-33661 139 4 possible possible ADJ bracis-33661 139 5 difference difference NOUN bracis-33661 139 6 in in ADP bracis-33661 139 7 the the DET bracis-33661 139 8 frequency frequency NOUN bracis-33661 139 9 at at ADP bracis-33661 139 10 which which PRON bracis-33661 139 11 data datum NOUN bracis-33661 139 12 were be AUX bracis-33661 139 13 collected collect VERB bracis-33661 139 14 , , PUNCT bracis-33661 139 15 the the DET bracis-33661 139 16 amount amount NOUN bracis-33661 139 17 of of ADP bracis-33661 139 18 data datum NOUN bracis-33661 139 19 collected collect VERB bracis-33661 139 20 for for ADP bracis-33661 139 21 400 400 NUM bracis-33661 139 22     SPACE bracis-33661 139 23 hz hz VERB bracis-33661 139 24 was be AUX bracis-33661 139 25 used use VERB bracis-33661 139 26 as as ADP bracis-33661 139 27 reference reference NOUN bracis-33661 139 28 . . PUNCT bracis-33661 140 1 3.2 3.2 NUM bracis-33661 140 2 eegraph eegraph NOUN bracis-33661 140 3 : : PUNCT bracis-33661 140 4 modeling model VERB bracis-33661 140 5 eeg eeg PROPN bracis-33661 140 6 data datum NOUN bracis-33661 140 7 into into ADP bracis-33661 140 8 graphs graph NOUN bracis-33661 140 9 as as SCONJ bracis-33661 140 10 illustrated illustrate VERB bracis-33661 140 11 by by ADP bracis-33661 140 12 steps step NOUN bracis-33661 140 13 figs fig NOUN bracis-33661 140 14 . . PUNCT bracis-33661 140 15     SPACE bracis-33661 140 16 1(b 1(b NUM bracis-33661 140 17 ) ) PUNCT bracis-33661 140 18 and and CCONJ bracis-33661 140 19 1(c 1(c NUM bracis-33661 140 20 ) ) PUNCT bracis-33661 140 21 , , PUNCT bracis-33661 140 22 we we PRON bracis-33661 140 23 develop develop VERB bracis-33661 140 24 eegraph eegraph NOUN bracis-33661 140 25 , , PUNCT bracis-33661 140 26 a a DET bracis-33661 140 27 modeling modeling NOUN bracis-33661 140 28 strategy strategy NOUN bracis-33661 140 29 to to PART bracis-33661 140 30 represent represent VERB bracis-33661 140 31 eeg eeg PROPN bracis-33661 140 32 exam exam NOUN bracis-33661 140 33 data datum NOUN bracis-33661 140 34 into into ADP bracis-33661 140 35 graphs graph NOUN bracis-33661 140 36 . . PUNCT bracis-33661 141 1 the the DET bracis-33661 141 2 graph graph NOUN bracis-33661 141 3 ’s ’s PART bracis-33661 141 4 data datum NOUN bracis-33661 141 5 structure structure NOUN bracis-33661 141 6 , , PUNCT bracis-33661 141 7 as as SCONJ bracis-33661 141 8 seen see VERB bracis-33661 141 9 in in ADP bracis-33661 141 10 the the DET bracis-33661 141 11 background background NOUN bracis-33661 141 12 section section NOUN bracis-33661 141 13 , , PUNCT bracis-33661 141 14 requires require VERB bracis-33661 141 15 the the DET bracis-33661 141 16 definition definition NOUN bracis-33661 141 17 of of ADP bracis-33661 141 18 vertices vertex NOUN bracis-33661 141 19 and and CCONJ bracis-33661 141 20 edges edge NOUN bracis-33661 141 21 in in ADP bracis-33661 141 22 order order NOUN bracis-33661 141 23 to to PART bracis-33661 141 24 leverage leverage VERB bracis-33661 141 25 the the DET bracis-33661 141 26 power power NOUN bracis-33661 141 27 of of ADP bracis-33661 141 28 gnns gnns ADJ bracis-33661 141 29 frameworks framework NOUN bracis-33661 141 30 . . PUNCT bracis-33661 142 1 therefore therefore ADV bracis-33661 142 2 , , PUNCT bracis-33661 142 3 eegraph eegraph PROPN bracis-33661 142 4 is be AUX bracis-33661 142 5 divided divide VERB bracis-33661 142 6 in in ADP bracis-33661 142 7 two two NUM bracis-33661 142 8 main main ADJ bracis-33661 142 9 steps step NOUN bracis-33661 142 10 : : PUNCT bracis-33661 142 11 node node NOUN bracis-33661 142 12 and and CCONJ bracis-33661 142 13 edge edge NOUN bracis-33661 142 14 construction construction NOUN bracis-33661 142 15 . . PUNCT bracis-33661 143 1 in in ADP bracis-33661 143 2 the the DET bracis-33661 143 3 node node ADJ bracis-33661 143 4 construction construction NOUN bracis-33661 143 5 step step NOUN bracis-33661 143 6 , , PUNCT bracis-33661 143 7 each each DET bracis-33661 143 8 electrode electrode NOUN bracis-33661 143 9 is be AUX bracis-33661 143 10 denoted denote VERB bracis-33661 143 11 by by ADP bracis-33661 143 12 a a DET bracis-33661 143 13 graph graph NOUN bracis-33661 143 14 vertex vertex NOUN bracis-33661 143 15 ( ( PUNCT bracis-33661 143 16 \ \ PROPN bracis-33661 143 17 ( ( PUNCT bracis-33661 143 18 \mathcal \mathcal PROPN bracis-33661 143 19 { { PUNCT bracis-33661 143 20 v v NOUN bracis-33661 143 21 } } PUNCT bracis-33661 143 22 \ \ NOUN bracis-33661 143 23 ) ) PUNCT bracis-33661 143 24 ) ) PUNCT bracis-33661 143 25 , , PUNCT bracis-33661 143 26 whose whose DET bracis-33661 143 27 attributes attribute NOUN bracis-33661 143 28 contain contain VERB bracis-33661 143 29 their their PRON bracis-33661 143 30 respective respective ADJ bracis-33661 143 31 temporal temporal ADJ bracis-33661 143 32 eeg eeg NOUN bracis-33661 143 33 records record NOUN bracis-33661 143 34 . . PUNCT bracis-33661 144 1 figure figure VERB bracis-33661 144 2     SPACE bracis-33661 144 3 1(b 1(b NUM bracis-33661 144 4 ) ) PUNCT bracis-33661 144 5 illustrates illustrate VERB bracis-33661 144 6 the the DET bracis-33661 144 7 step step NOUN bracis-33661 144 8 in in ADP bracis-33661 144 9 eegraph eegraph PROPN bracis-33661 144 10 responsible responsible ADJ bracis-33661 144 11 for for ADP bracis-33661 144 12 modeling model VERB bracis-33661 144 13 the the DET bracis-33661 144 14 eeg eeg NOUN bracis-33661 144 15 records record NOUN bracis-33661 144 16 for for ADP bracis-33661 144 17 each each DET bracis-33661 144 18 electrode electrode NOUN bracis-33661 144 19 to to PART bracis-33661 144 20 vertices vertex NOUN bracis-33661 144 21 in in ADP bracis-33661 144 22 the the DET bracis-33661 144 23 graph graph NOUN bracis-33661 144 24 structure structure NOUN bracis-33661 144 25 . . PUNCT bracis-33661 145 1 at at ADP bracis-33661 145 2 this this DET bracis-33661 145 3 stage stage NOUN bracis-33661 145 4 , , PUNCT bracis-33661 145 5 the the DET bracis-33661 145 6 temporal temporal ADJ bracis-33661 145 7 values value NOUN bracis-33661 145 8 for for ADP bracis-33661 145 9 each each DET bracis-33661 145 10 electrode electrode NOUN bracis-33661 145 11 are be AUX bracis-33661 145 12 assigned assign VERB bracis-33661 145 13 to to ADP bracis-33661 145 14 the the DET bracis-33661 145 15 respective respective ADJ bracis-33661 145 16 vertices vertex NOUN bracis-33661 145 17 in in ADP bracis-33661 145 18 the the DET bracis-33661 145 19 form form NOUN bracis-33661 145 20 of of ADP bracis-33661 145 21 an an DET bracis-33661 145 22 attribute attribute NOUN bracis-33661 145 23 vector vector NOUN bracis-33661 145 24 , , PUNCT bracis-33661 145 25 and and CCONJ bracis-33661 145 26 simultaneously simultaneously ADV bracis-33661 145 27 , , PUNCT bracis-33661 145 28 the the DET bracis-33661 145 29 respective respective ADJ bracis-33661 145 30 patient patient NOUN bracis-33661 145 31 outcomes outcome NOUN bracis-33661 145 32 are be AUX bracis-33661 145 33 also also ADV bracis-33661 145 34 assigned assign VERB bracis-33661 145 35 as as ADP bracis-33661 145 36 the the DET bracis-33661 145 37 target target NOUN bracis-33661 145 38 class class NOUN bracis-33661 145 39 of of ADP bracis-33661 145 40 the the DET bracis-33661 145 41 final final ADJ bracis-33661 145 42 graph graph NOUN bracis-33661 145 43 data datum NOUN bracis-33661 145 44 structure structure NOUN bracis-33661 145 45 . . PUNCT bracis-33661 146 1 in in ADP bracis-33661 146 2 the the DET bracis-33661 146 3 edge edge NOUN bracis-33661 146 4 construction construction NOUN bracis-33661 146 5 step step NOUN bracis-33661 146 6 , , PUNCT bracis-33661 146 7 a a DET bracis-33661 146 8 matrix matrix NOUN bracis-33661 146 9 p p NOUN bracis-33661 146 10 , , PUNCT bracis-33661 146 11 defined define VERB bracis-33661 146 12 by by ADP bracis-33661 146 13 ( ( PUNCT bracis-33661 146 14 15 15 NUM bracis-33661 146 15 ) ) PUNCT bracis-33661 146 16 , , PUNCT bracis-33661 146 17 is be AUX bracis-33661 146 18 built build VERB bracis-33661 146 19 as as ADP bracis-33661 146 20 a a DET bracis-33661 146 21 reference reference NOUN bracis-33661 146 22 for for ADP bracis-33661 146 23 the the DET bracis-33661 146 24 creation creation NOUN bracis-33661 146 25 of of ADP bracis-33661 146 26 the the DET bracis-33661 146 27 edges edge NOUN bracis-33661 146 28 between between ADP bracis-33661 146 29 nodes node NOUN bracis-33661 146 30 ( ( PUNCT bracis-33661 146 31 electrodes electrode NOUN bracis-33661 146 32 ) ) PUNCT bracis-33661 146 33 , , PUNCT bracis-33661 146 34 as as SCONJ bracis-33661 146 35 exemplified exemplify VERB bracis-33661 146 36 in in ADP bracis-33661 146 37 fig fig NOUN bracis-33661 146 38 . . PUNCT bracis-33661 146 39     SPACE bracis-33661 146 40 1(c 1(c NUM bracis-33661 146 41 ) ) PUNCT bracis-33661 146 42 . . PUNCT bracis-33661 147 1 all all DET bracis-33661 147 2 links link NOUN bracis-33661 147 3 were be AUX bracis-33661 147 4 created create VERB bracis-33661 147 5 by by ADP bracis-33661 147 6 considering consider VERB bracis-33661 147 7 the the DET bracis-33661 147 8 electrodes electrode NOUN bracis-33661 147 9 positions position NOUN bracis-33661 147 10 along along ADP bracis-33661 147 11 the the DET bracis-33661 147 12 scalp scalp NOUN bracis-33661 147 13 , , PUNCT bracis-33661 147 14 as as SCONJ bracis-33661 147 15 shown show VERB bracis-33661 147 16 in in ADP bracis-33661 147 17 fig fig NOUN bracis-33661 147 18 . . PUNCT bracis-33661 147 19     SPACE bracis-33661 148 1 2 2 NUM bracis-33661 148 2 , , PUNCT bracis-33661 148 3 to to PART bracis-33661 148 4 effectively effectively ADV bracis-33661 148 5 leverage leverage VERB bracis-33661 148 6 the the DET bracis-33661 148 7 spatial spatial ADJ bracis-33661 148 8 relationships relationship NOUN bracis-33661 148 9 between between ADP bracis-33661 148 10 the the DET bracis-33661 148 11 nodes node NOUN bracis-33661 148 12 and and CCONJ bracis-33661 148 13 represent represent VERB bracis-33661 148 14 them they PRON bracis-33661 148 15 as as ADP bracis-33661 148 16 connections connection NOUN bracis-33661 148 17 . . PUNCT bracis-33661 149 1 the the DET bracis-33661 149 2 padding padding NOUN bracis-33661 149 3 elements element NOUN bracis-33661 149 4 , , PUNCT bracis-33661 149 5 denoted denote VERB bracis-33661 149 6 by by ADP bracis-33661 149 7 \(\times \(\time NOUN bracis-33661 149 8 \ \ NOUN bracis-33661 149 9 ) ) PUNCT bracis-33661 149 10 in in ADP bracis-33661 149 11 ( ( PUNCT bracis-33661 149 12 15 15 NUM bracis-33661 149 13 ) ) PUNCT bracis-33661 149 14 , , PUNCT bracis-33661 149 15 were be AUX bracis-33661 149 16 used use VERB bracis-33661 149 17 only only ADV bracis-33661 149 18 as as ADP bracis-33661 149 19 a a DET bracis-33661 149 20 facilitator facilitator NOUN bracis-33661 149 21 in in ADP bracis-33661 149 22 the the DET bracis-33661 149 23 visual visual ADJ bracis-33661 149 24 interpretation interpretation NOUN bracis-33661 149 25 and and CCONJ bracis-33661 149 26 also also ADV bracis-33661 149 27 for for ADP bracis-33661 149 28 the the DET bracis-33661 149 29 implementation implementation NOUN bracis-33661 149 30 of of ADP bracis-33661 149 31 a a DET bracis-33661 149 32 neighboring neighboring NOUN bracis-33661 149 33 heuristic heuristic NOUN bracis-33661 149 34 described describe VERB bracis-33661 149 35 next next ADV bracis-33661 149 36 . . PUNCT bracis-33661 150 1 although although SCONJ bracis-33661 150 2 they they PRON bracis-33661 150 3 are be AUX bracis-33661 150 4 not not PART bracis-33661 150 5 directly directly ADV bracis-33661 150 6 present present ADJ bracis-33661 150 7 in in ADP bracis-33661 150 8 the the DET bracis-33661 150 9 construction construction NOUN bracis-33661 150 10 of of ADP bracis-33661 150 11 the the DET bracis-33661 150 12 graph graph NOUN bracis-33661 150 13 structures structure NOUN bracis-33661 150 14 , , PUNCT bracis-33661 150 15 they they PRON bracis-33661 150 16 have have VERB bracis-33661 150 17 a a DET bracis-33661 150 18 great great ADJ bracis-33661 150 19 importance importance NOUN bracis-33661 150 20 enabling enable VERB bracis-33661 150 21 the the DET bracis-33661 150 22 eegraph eegraph NOUN bracis-33661 150 23 strategy strategy NOUN bracis-33661 150 24 not not PART bracis-33661 150 25 only only ADV bracis-33661 150 26 to to PART bracis-33661 150 27 have have VERB bracis-33661 150 28 a a DET bracis-33661 150 29 positional positional ADJ bracis-33661 150 30 reference reference NOUN bracis-33661 150 31 for for ADP bracis-33661 150 32 each each DET bracis-33661 150 33 vertex vertex NOUN bracis-33661 150 34 in in ADP bracis-33661 150 35 the the DET bracis-33661 150 36 edge edge NOUN bracis-33661 150 37 construction construction NOUN bracis-33661 150 38 process process NOUN bracis-33661 150 39 , , PUNCT bracis-33661 150 40 but but CCONJ bracis-33661 150 41 also also ADV bracis-33661 150 42 to to PART bracis-33661 150 43 define define VERB bracis-33661 150 44 different different ADJ bracis-33661 150 45 levels level NOUN bracis-33661 150 46 of of ADP bracis-33661 150 47 neighboring neighboring NOUN bracis-33661 150 48 relationships relationship NOUN bracis-33661 150 49 . . PUNCT bracis-33661 151 1 $ $ SYM bracis-33661 151 2 $ $ SYM bracis-33661 151 3 \begin{aligned \begin{aligne VERB bracis-33661 151 4 } } PUNCT bracis-33661 151 5 p p NOUN bracis-33661 151 6 = = PUNCT bracis-33661 151 7 \begin{bmatrix \begin{bmatrix NUM bracis-33661 151 8 } } PUNCT bracis-33661 151 9 \times \times PROPN bracis-33661 151 10 & & CCONJ bracis-33661 151 11 f_{p1 f_{p1 PROPN bracis-33661 151 12 } } PUNCT bracis-33661 151 13 & & CCONJ bracis-33661 151 14 \times \times PROPN bracis-33661 151 15 & & CCONJ bracis-33661 151 16 f_{p2 f_{p2 PROPN bracis-33661 151 17 } } PUNCT bracis-33661 151 18 & & CCONJ bracis-33661 151 19 \times \times ADP bracis-33661 151 20 \\ \\ NOUN bracis-33661 151 21 f_{7 f_{7 PROPN bracis-33661 151 22 } } PUNCT bracis-33661 151 23 & & CCONJ bracis-33661 151 24 f_{3 f_{3 PROPN bracis-33661 151 25 } } PUNCT bracis-33661 151 26 & & CCONJ bracis-33661 151 27 f_{z f_{z PROPN bracis-33661 151 28 } } PUNCT bracis-33661 151 29 & & CCONJ bracis-33661 151 30 f_{4 f_{4 PROPN bracis-33661 151 31 } } PUNCT bracis-33661 151 32 & & CCONJ bracis-33661 151 33 f_{8 f_{8 PROPN bracis-33661 151 34 } } PUNCT bracis-33661 151 35 \\ \\ NOUN bracis-33661 151 36 t_{3 t_{3 PROPN bracis-33661 151 37 } } PUNCT bracis-33661 151 38 & & CCONJ bracis-33661 151 39 c_{3 c_{3 PROPN bracis-33661 151 40 } } PUNCT bracis-33661 151 41 & & CCONJ bracis-33661 151 42 c_{z c_{z PROPN bracis-33661 151 43 } } PUNCT bracis-33661 151 44 & & CCONJ bracis-33661 151 45 c_{4 c_{4 PROPN bracis-33661 151 46 } } PUNCT bracis-33661 151 47 & & CCONJ bracis-33661 151 48 t_{4 t_{4 PROPN bracis-33661 151 49 } } PUNCT bracis-33661 151 50 \\ \\ NOUN bracis-33661 151 51 t_{5 t_{5 NOUN bracis-33661 151 52 } } PUNCT bracis-33661 151 53 & & CCONJ bracis-33661 151 54 p_{3 p_{3 PROPN bracis-33661 151 55 } } PUNCT bracis-33661 151 56 & & CCONJ bracis-33661 151 57 p_{z p_{z PROPN bracis-33661 151 58 } } PUNCT bracis-33661 151 59 & & CCONJ bracis-33661 151 60 p_{4 p_{4 VERB bracis-33661 151 61 } } PUNCT bracis-33661 151 62 & & CCONJ bracis-33661 151 63 t_{6 t_{6 NOUN bracis-33661 151 64 } } PUNCT bracis-33661 151 65 \\ \\ NOUN bracis-33661 151 66 \times \times PROPN bracis-33661 151 67 & & CCONJ bracis-33661 151 68 o_{1 o_{1 NOUN bracis-33661 151 69 } } PUNCT bracis-33661 151 70 & & CCONJ bracis-33661 151 71 \times \times PROPN bracis-33661 151 72 & & CCONJ bracis-33661 151 73 o_{2 o_{2 NOUN bracis-33661 151 74 } } PUNCT bracis-33661 151 75 & & CCONJ bracis-33661 151 76 \times \times ADP bracis-33661 151 77 \\ \\ NOUN bracis-33661 151 78 \end{bmatrix \end{bmatrix VERB bracis-33661 151 79 } } PUNCT bracis-33661 151 80 \end{aligned}$$ \end{aligned}$$ X bracis-33661 151 81 ( ( PUNCT bracis-33661 151 82 15 15 NUM bracis-33661 151 83 ) ) PUNCT bracis-33661 151 84 we we PRON bracis-33661 151 85 enhance enhance VERB bracis-33661 151 86 our our PRON bracis-33661 151 87 eegraph eegraph NOUN bracis-33661 151 88 modeling modeling NOUN bracis-33661 151 89 strategy strategy NOUN bracis-33661 151 90 by by ADP bracis-33661 151 91 designing design VERB bracis-33661 151 92 an an DET bracis-33661 151 93 edge edge NOUN bracis-33661 151 94 construction construction NOUN bracis-33661 151 95 heuristic heuristic ADJ bracis-33661 151 96 able able ADJ bracis-33661 151 97 take take VERB bracis-33661 151 98 advantage advantage NOUN bracis-33661 151 99 of of ADP bracis-33661 151 100 different different ADJ bracis-33661 151 101 levels level NOUN bracis-33661 151 102 of of ADP bracis-33661 151 103 neighboring neighboring NOUN bracis-33661 151 104 relationships relationship NOUN bracis-33661 151 105 among among ADP bracis-33661 151 106 the the DET bracis-33661 151 107 nodes node NOUN bracis-33661 151 108 . . PUNCT bracis-33661 152 1 such such DET bracis-33661 152 2 a a DET bracis-33661 152 3 heuristic heuristic NOUN bracis-33661 152 4 is be AUX bracis-33661 152 5 named name VERB bracis-33661 152 6 h h NOUN bracis-33661 152 7 - - NOUN bracis-33661 152 8 eegraph eegraph NOUN bracis-33661 152 9 , , PUNCT bracis-33661 152 10 in in ADP bracis-33661 152 11 which which PRON bracis-33661 152 12 h h NOUN bracis-33661 152 13 defines define VERB bracis-33661 152 14 the the DET bracis-33661 152 15 maximum maximum ADJ bracis-33661 152 16 geodesic geodesic ADJ bracis-33661 152 17 distance distance NOUN bracis-33661 152 18 to to PART bracis-33661 152 19 be be AUX bracis-33661 152 20 considered consider VERB bracis-33661 152 21 in in ADP bracis-33661 152 22 the the DET bracis-33661 152 23 creation creation NOUN bracis-33661 152 24 of of ADP bracis-33661 152 25 adjacent adjacent ADJ bracis-33661 152 26 edges edge NOUN bracis-33661 152 27 . . PUNCT bracis-33661 153 1 figure figure NOUN bracis-33661 153 2     SPACE bracis-33661 153 3 3 3 NUM bracis-33661 153 4 shows show VERB bracis-33661 153 5 the the DET bracis-33661 153 6 graph graph NOUN bracis-33661 153 7 constructed construct VERB bracis-33661 153 8 for for ADP bracis-33661 153 9 1 1 NUM bracis-33661 153 10 - - PUNCT bracis-33661 153 11 eegraph eegraph NOUN bracis-33661 153 12 , , PUNCT bracis-33661 153 13 in in ADP bracis-33661 153 14 which which PRON bracis-33661 153 15 only only ADV bracis-33661 153 16 adjacent adjacent ADJ bracis-33661 153 17 vertices vertex NOUN bracis-33661 153 18 are be AUX bracis-33661 153 19 connected connect VERB bracis-33661 153 20 , , PUNCT bracis-33661 153 21 ignoring ignore VERB bracis-33661 153 22 the the DET bracis-33661 153 23 padding padding NOUN bracis-33661 153 24 elements element NOUN bracis-33661 153 25 in in ADP bracis-33661 153 26 matrix matrix NOUN bracis-33661 153 27 p. p. NOUN bracis-33661 153 28 figure figure NOUN bracis-33661 153 29     SPACE bracis-33661 153 30 4 4 NUM bracis-33661 153 31 shows show VERB bracis-33661 153 32 the the DET bracis-33661 153 33 connection connection NOUN bracis-33661 153 34 of of ADP bracis-33661 153 35 the the DET bracis-33661 153 36 \(f_{p1}\ \(f_{p1}\ PROPN bracis-33661 153 37 ) ) PUNCT bracis-33661 153 38 electrode electrode PROPN bracis-33661 153 39 / / SYM bracis-33661 153 40 node node NOUN bracis-33661 153 41 considering consider VERB bracis-33661 153 42 one one NUM bracis-33661 153 43 , , PUNCT bracis-33661 153 44 two two NUM bracis-33661 153 45 and and CCONJ bracis-33661 153 46 three three NUM bracis-33661 153 47 levels level NOUN bracis-33661 153 48 of of ADP bracis-33661 153 49 neighborhood neighborhood NOUN bracis-33661 153 50 by by ADP bracis-33661 153 51 setting set VERB bracis-33661 153 52 maximum maximum ADJ bracis-33661 153 53 geodesic geodesic ADJ bracis-33661 153 54 distance distance NOUN bracis-33661 153 55 h h NOUN bracis-33661 153 56 respectively respectively ADV bracis-33661 153 57 equal equal ADJ bracis-33661 153 58 to to ADP bracis-33661 153 59 1 1 NUM bracis-33661 153 60 , , PUNCT bracis-33661 153 61 2 2 NUM bracis-33661 153 62 and and CCONJ bracis-33661 153 63 3 3 NUM bracis-33661 153 64 . . PUNCT bracis-33661 154 1 formally formally ADV bracis-33661 154 2 , , PUNCT bracis-33661 154 3 the the DET bracis-33661 154 4 edge edge NOUN bracis-33661 154 5 construction construction NOUN bracis-33661 154 6 step step NOUN bracis-33661 154 7 is be AUX bracis-33661 154 8 given give VERB bracis-33661 154 9 by by ADP bracis-33661 154 10 : : PUNCT bracis-33661 154 11 $ $ SYM bracis-33661 154 12 $ $ SYM bracis-33661 154 13 \begin{aligned \begin{aligne VERB bracis-33661 154 14 } } PUNCT bracis-33661 154 15 a(v_i a(v_i PROPN bracis-33661 154 16 , , PUNCT bracis-33661 154 17 v_j v_j NUM bracis-33661 154 18 ) ) PUNCT bracis-33661 155 1 = = PRON bracis-33661 155 2 { { PUNCT bracis-33661 155 3 \left\ \left\ PROPN bracis-33661 155 4 { { PUNCT bracis-33661 155 5 \begin{array}{ll \begin{array}{ll PROPN bracis-33661 155 6 } } PUNCT bracis-33661 155 7 1 1 NUM bracis-33661 155 8 , , PUNCT bracis-33661 155 9 & & CCONJ bracis-33661 155 10 \text \text PROPN bracis-33661 155 11 { { PUNCT bracis-33661 155 12 if if SCONJ bracis-33661 155 13 } } PUNCT bracis-33661 155 14 d(v_i d(v_i PROPN bracis-33661 155 15 , , PUNCT bracis-33661 155 16 v_j v_j NUM bracis-33661 155 17 ) ) PUNCT bracis-33661 155 18 \le \le PROPN bracis-33661 155 19 h~,\\ h~,\\ NOUN bracis-33661 155 20 0 0 NUM bracis-33661 155 21 , , PUNCT bracis-33661 155 22 & & CCONJ bracis-33661 155 23 \text \text PROPN bracis-33661 155 24 { { PUNCT bracis-33661 155 25 otherwise otherwise ADV bracis-33661 155 26 , , PUNCT bracis-33661 155 27 } } PUNCT bracis-33661 155 28 \end{array}\right \end{array}\right PROPN bracis-33661 155 29 . . PUNCT bracis-33661 155 30 } } PUNCT bracis-33661 155 31 \end{aligned}$$ \end{aligned}$$ X bracis-33661 155 32 ( ( PUNCT bracis-33661 155 33 16 16 NUM bracis-33661 155 34 ) ) PUNCT bracis-33661 155 35 in in ADP bracis-33661 155 36 which which PRON bracis-33661 155 37 a a DET bracis-33661 155 38 refers refer VERB bracis-33661 155 39 to to ADP bracis-33661 155 40 the the DET bracis-33661 155 41 adjacency adjacency NOUN bracis-33661 155 42 matrix matrix NOUN bracis-33661 155 43 obtained obtain VERB bracis-33661 155 44 from from ADP bracis-33661 155 45 our our PRON bracis-33661 155 46 h h NOUN bracis-33661 155 47 - - PUNCT bracis-33661 155 48 eegraph eegraph NOUN bracis-33661 155 49 modeling modeling NOUN bracis-33661 155 50 , , PUNCT bracis-33661 155 51 and and CCONJ bracis-33661 155 52 \(d(v_i \(d(v_i NUM bracis-33661 155 53 , , PUNCT bracis-33661 155 54 v_j)\ v_j)\ NOUN bracis-33661 155 55 ) ) PUNCT bracis-33661 155 56 is be AUX bracis-33661 155 57 the the DET bracis-33661 155 58 geodesic geodesic ADJ bracis-33661 155 59 distance distance NOUN bracis-33661 155 60 between between ADP bracis-33661 155 61 nodes node NOUN bracis-33661 155 62 \(v_i\ \(v_i\ NOUN bracis-33661 155 63 ) ) PUNCT bracis-33661 155 64 and and CCONJ bracis-33661 155 65 \(v_j\ \(v_j\ PROPN bracis-33661 155 66 ) ) PUNCT bracis-33661 155 67 . . PUNCT bracis-33661 156 1 fig fig NOUN bracis-33661 156 2 . . PUNCT bracis-33661 157 1 3 3 X bracis-33661 157 2 . . NOUN bracis-33661 157 3 electrodes electrode NOUN bracis-33661 157 4 structured structure VERB bracis-33661 157 5 as as ADP bracis-33661 157 6 a a DET bracis-33661 157 7 graph graph NOUN bracis-33661 157 8 . . PUNCT bracis-33661 158 1 full full ADJ bracis-33661 158 2 size size NOUN bracis-33661 158 3 image image NOUN bracis-33661 158 4 fig fig NOUN bracis-33661 158 5 . . PUNCT bracis-33661 159 1 4 4 X bracis-33661 159 2 . . NOUN bracis-33661 159 3 example example NOUN bracis-33661 159 4 of of ADP bracis-33661 159 5 the the DET bracis-33661 159 6 fp1 fp1 PROPN bracis-33661 159 7 channel channel NOUN bracis-33661 159 8 method method NOUN bracis-33661 159 9 connections connection NOUN bracis-33661 159 10 . . PUNCT bracis-33661 160 1 full full ADJ bracis-33661 160 2 size size NOUN bracis-33661 160 3 image image NOUN bracis-33661 160 4 3.3 3.3 NUM bracis-33661 160 5 neural neural ADJ bracis-33661 160 6 networks network NOUN bracis-33661 160 7 techniques technique VERB bracis-33661 160 8 the the DET bracis-33661 160 9 gnns gnns ADJ bracis-33661 160 10 frameworks framework NOUN bracis-33661 160 11 investigated investigate VERB bracis-33661 160 12 in in ADP bracis-33661 160 13 this this DET bracis-33661 160 14 study study NOUN bracis-33661 160 15 are be AUX bracis-33661 160 16 gcn gcn NOUN bracis-33661 161 1 [ [ X bracis-33661 161 2 16 16 NUM bracis-33661 161 3 ] ] PUNCT bracis-33661 161 4 , , PUNCT bracis-33661 161 5 sgc sgc PROPN bracis-33661 161 6 [ [ X bracis-33661 161 7 31 31 NUM bracis-33661 161 8 ] ] PUNCT bracis-33661 161 9 , , PUNCT bracis-33661 161 10 sage sage NOUN bracis-33661 161 11 [ [ X bracis-33661 161 12 12 12 NUM bracis-33661 161 13 ] ] PUNCT bracis-33661 161 14 and and CCONJ bracis-33661 161 15 gcnii gcnii NOUN bracis-33661 162 1 [ [ X bracis-33661 162 2 6 6 NUM bracis-33661 162 3 ] ] PUNCT bracis-33661 162 4 . . PUNCT bracis-33661 163 1 for for ADP bracis-33661 163 2 comparative comparative ADJ bracis-33661 163 3 purposes purpose NOUN bracis-33661 163 4 , , PUNCT bracis-33661 163 5 we we PRON bracis-33661 163 6 considered consider VERB bracis-33661 163 7 a a DET bracis-33661 163 8 lstm lstm ADJ bracis-33661 163 9 network network NOUN bracis-33661 163 10 [ [ X bracis-33661 163 11 33 33 NUM bracis-33661 163 12 ] ] PUNCT bracis-33661 163 13 , , PUNCT bracis-33661 163 14 a a DET bracis-33661 163 15 widely widely ADV bracis-33661 163 16 adopted adopt VERB bracis-33661 163 17 technique technique NOUN bracis-33661 163 18 for for ADP bracis-33661 163 19 sequential sequential ADJ bracis-33661 163 20 data datum NOUN bracis-33661 163 21 analysis analysis NOUN bracis-33661 163 22 . . PUNCT bracis-33661 164 1 section section NOUN bracis-33661 164 2     SPACE bracis-33661 164 3 2 2 NUM bracis-33661 164 4 presents present VERB bracis-33661 164 5 a a DET bracis-33661 164 6 brief brief ADJ bracis-33661 164 7 description description NOUN bracis-33661 164 8 about about ADP bracis-33661 164 9 each each DET bracis-33661 164 10 one one NUM bracis-33661 164 11 of of ADP bracis-33661 164 12 the the DET bracis-33661 164 13 techniques technique NOUN bracis-33661 164 14 under under ADP bracis-33661 164 15 comparison comparison NOUN bracis-33661 164 16 . . PUNCT bracis-33661 165 1 to to PART bracis-33661 165 2 be be AUX bracis-33661 165 3 specific specific ADJ bracis-33661 165 4 , , PUNCT bracis-33661 165 5 all all DET bracis-33661 165 6 gnns gnns ADJ bracis-33661 165 7 models model NOUN bracis-33661 165 8 considered consider VERB bracis-33661 165 9 three three NUM bracis-33661 165 10 convolutional convolutional ADJ bracis-33661 165 11 layers layer NOUN bracis-33661 165 12 . . PUNCT bracis-33661 166 1 in in ADP bracis-33661 166 2 the the DET bracis-33661 166 3 sage sage NOUN bracis-33661 166 4 layers layer NOUN bracis-33661 166 5 architecture architecture NOUN bracis-33661 166 6 , , PUNCT bracis-33661 166 7 we we PRON bracis-33661 166 8 adopted adopt VERB bracis-33661 166 9 the the DET bracis-33661 166 10 mean mean ADJ bracis-33661 166 11 aggregation aggregation NOUN bracis-33661 166 12 function function NOUN bracis-33661 166 13 , , PUNCT bracis-33661 166 14 and and CCONJ bracis-33661 166 15 for for ADP bracis-33661 166 16 the the DET bracis-33661 166 17 gcnii gcnii ADJ bracis-33661 166 18 layers layer NOUN bracis-33661 166 19 architecture architecture NOUN bracis-33661 166 20 , , PUNCT bracis-33661 166 21 we we PRON bracis-33661 166 22 adopted adopt VERB bracis-33661 166 23 0.1 0.1 NUM bracis-33661 166 24 and and CCONJ bracis-33661 166 25 0.5 0.5 NUM bracis-33661 166 26 values value NOUN bracis-33661 166 27 for for ADP bracis-33661 166 28 the the DET bracis-33661 166 29 \(\alpha \(\alpha NOUN bracis-33661 166 30 \ \ NOUN bracis-33661 166 31 ) ) PUNCT bracis-33661 166 32 and and CCONJ bracis-33661 166 33 \(\beta \(\beta PRON bracis-33661 166 34 \ \ NOUN bracis-33661 166 35 ) ) PUNCT bracis-33661 166 36 hyperparameters hyperparameter NOUN bracis-33661 166 37 , , PUNCT bracis-33661 166 38 respectively respectively ADV bracis-33661 166 39 . . PUNCT bracis-33661 167 1 they they PRON bracis-33661 167 2 were be AUX bracis-33661 167 3 trained train VERB bracis-33661 167 4 with with ADP bracis-33661 167 5 a a DET bracis-33661 167 6 number number NOUN bracis-33661 167 7 of of ADP bracis-33661 167 8 100 100 NUM bracis-33661 167 9 epochs epoch NOUN bracis-33661 167 10 , , PUNCT bracis-33661 167 11 a a DET bracis-33661 167 12 batch batch NOUN bracis-33661 167 13 size size NOUN bracis-33661 167 14 of of ADP bracis-33661 167 15 64 64 NUM bracis-33661 167 16 and and CCONJ bracis-33661 167 17 adaptive adaptive ADJ bracis-33661 167 18 movement movement NOUN bracis-33661 167 19 estimation estimation NOUN bracis-33661 167 20 ( ( PUNCT bracis-33661 167 21 adam adam PROPN bracis-33661 167 22 ) ) PUNCT bracis-33661 167 23 optimizer optimizer NOUN bracis-33661 167 24 with with ADP bracis-33661 167 25 a a DET bracis-33661 167 26 learning learn VERB bracis-33661 167 27 rate rate NOUN bracis-33661 167 28 of of ADP bracis-33661 167 29 \(10^{-3}\ \(10^{-3}\ NOUN bracis-33661 167 30 ) ) PUNCT bracis-33661 167 31 . . PUNCT bracis-33661 168 1 the the DET bracis-33661 168 2 parameters parameter NOUN bracis-33661 168 3 and and CCONJ bracis-33661 168 4 structures structure NOUN bracis-33661 168 5 of of ADP bracis-33661 168 6 the the DET bracis-33661 168 7 created create VERB bracis-33661 168 8 gnns gnns NOUN bracis-33661 168 9 were be AUX bracis-33661 168 10 similarly similarly ADV bracis-33661 168 11 designed design VERB bracis-33661 168 12 to to PART bracis-33661 168 13 ensure ensure VERB bracis-33661 168 14 a a DET bracis-33661 168 15 fair fair ADJ bracis-33661 168 16 comparison comparison NOUN bracis-33661 168 17 between between ADP bracis-33661 168 18 them they PRON bracis-33661 168 19 . . PUNCT bracis-33661 169 1 regarding regard VERB bracis-33661 169 2 lstm lstm ADJ bracis-33661 169 3 architecture architecture NOUN bracis-33661 169 4 parameters parameter NOUN bracis-33661 169 5 , , PUNCT bracis-33661 169 6 we we PRON bracis-33661 169 7 identified identify VERB bracis-33661 169 8 the the DET bracis-33661 169 9 optimal optimal ADJ bracis-33661 169 10 configuration configuration NOUN bracis-33661 169 11 , , PUNCT bracis-33661 169 12 which which PRON bracis-33661 169 13 consisted consist VERB bracis-33661 169 14 of of ADP bracis-33661 169 15 a a DET bracis-33661 169 16 single single ADJ bracis-33661 169 17 lstm lstm NOUN bracis-33661 169 18 layer layer NOUN bracis-33661 169 19 with with ADP bracis-33661 169 20 20 20 NUM bracis-33661 169 21 neurons neuron NOUN bracis-33661 169 22 , , PUNCT bracis-33661 169 23 trained train VERB bracis-33661 169 24 over over ADP bracis-33661 169 25 100 100 NUM bracis-33661 169 26 epochs epoch NOUN bracis-33661 169 27 , , PUNCT bracis-33661 169 28 yielding yield VERB bracis-33661 169 29 the the DET bracis-33661 169 30 best good ADJ bracis-33661 169 31 predictive predictive ADJ bracis-33661 169 32 performance performance NOUN bracis-33661 169 33 . . PUNCT bracis-33661 170 1 table table NOUN bracis-33661 170 2 2 2 NUM bracis-33661 170 3 . . PUNCT bracis-33661 171 1 performance performance NOUN bracis-33661 171 2 comparison comparison NOUN bracis-33661 171 3 : : PUNCT bracis-33661 171 4 metric metric ADJ bracis-33661 171 5 value value NOUN bracis-33661 171 6 / / SYM bracis-33661 171 7 standard standard ADJ bracis-33661 171 8 deviationfull deviationfull NOUN bracis-33661 171 9 size size NOUN bracis-33661 171 10 table table NOUN bracis-33661 171 11 4 4 NUM bracis-33661 171 12 experimental experimental ADJ bracis-33661 171 13 results result NOUN bracis-33661 171 14 in in ADP bracis-33661 171 15 order order NOUN bracis-33661 171 16 to to PART bracis-33661 171 17 ensure ensure VERB bracis-33661 171 18 a a DET bracis-33661 171 19 reliable reliable ADJ bracis-33661 171 20 and and CCONJ bracis-33661 171 21 fair fair ADJ bracis-33661 171 22 comparison comparison NOUN bracis-33661 171 23 between between ADP bracis-33661 171 24 the the DET bracis-33661 171 25 results result NOUN bracis-33661 171 26 generated generate VERB bracis-33661 171 27 by by ADP bracis-33661 171 28 the the DET bracis-33661 171 29 network network NOUN bracis-33661 171 30 architectures architecture NOUN bracis-33661 171 31 used use VERB bracis-33661 171 32 , , PUNCT bracis-33661 171 33 the the DET bracis-33661 171 34 k k ADJ bracis-33661 171 35 - - ADJ bracis-33661 171 36 fold fold ADJ bracis-33661 171 37 stratified stratified ADJ bracis-33661 171 38 method method NOUN bracis-33661 171 39 was be AUX bracis-33661 171 40 applied apply VERB bracis-33661 171 41 . . PUNCT bracis-33661 172 1 the the DET bracis-33661 172 2 dataset dataset NOUN bracis-33661 172 3 was be AUX bracis-33661 172 4 partitioned partition VERB bracis-33661 172 5 into into ADP bracis-33661 172 6 10 10 NUM bracis-33661 172 7 folds fold NOUN bracis-33661 172 8 maintaining maintain VERB bracis-33661 172 9 proportionality proportionality NOUN bracis-33661 172 10 between between ADP bracis-33661 172 11 the the DET bracis-33661 172 12 classes class NOUN bracis-33661 172 13 , , PUNCT bracis-33661 172 14 using use VERB bracis-33661 172 15 8 8 NUM bracis-33661 172 16 partitions partition NOUN bracis-33661 172 17 for for ADP bracis-33661 172 18 training training NOUN bracis-33661 172 19 , , PUNCT bracis-33661 172 20 1 1 NUM bracis-33661 172 21 for for ADP bracis-33661 172 22 validation validation NOUN bracis-33661 172 23 and and CCONJ bracis-33661 172 24 1 1 NUM bracis-33661 172 25 for for ADP bracis-33661 172 26 test test NOUN bracis-33661 172 27 . . PUNCT bracis-33661 173 1 table table NOUN bracis-33661 173 2     SPACE bracis-33661 173 3 1 1 NUM bracis-33661 173 4 made make VERB bracis-33661 173 5 explicit explicit ADJ bracis-33661 173 6 a a DET bracis-33661 173 7 considerable considerable ADJ bracis-33661 173 8 imbalance imbalance NOUN bracis-33661 173 9 between between ADP bracis-33661 173 10 the the DET bracis-33661 173 11 number number NOUN bracis-33661 173 12 of of ADP bracis-33661 173 13 registers register NOUN bracis-33661 173 14 for for ADP bracis-33661 173 15 each each DET bracis-33661 173 16 outcome outcome NOUN bracis-33661 173 17 , , PUNCT bracis-33661 173 18 a a DET bracis-33661 173 19 factor factor NOUN bracis-33661 173 20 of of ADP bracis-33661 173 21 great great ADJ bracis-33661 173 22 weight weight NOUN bracis-33661 173 23 for for ADP bracis-33661 173 24 the the DET bracis-33661 173 25 choice choice NOUN bracis-33661 173 26 of of ADP bracis-33661 173 27 metrics metric NOUN bracis-33661 173 28 presented present VERB bracis-33661 173 29 in in ADP bracis-33661 173 30 the the DET bracis-33661 173 31 comparison comparison NOUN bracis-33661 173 32 of of ADP bracis-33661 173 33 table table NOUN bracis-33661 173 34     SPACE bracis-33661 173 35 2 2 NUM bracis-33661 173 36 . . PUNCT bracis-33661 173 37 to to PART bracis-33661 173 38 deal deal VERB bracis-33661 173 39 with with ADP bracis-33661 173 40 this this DET bracis-33661 173 41 imbalance imbalance NOUN bracis-33661 173 42 , , PUNCT bracis-33661 173 43 f1 f1 NOUN bracis-33661 173 44 - - PUNCT bracis-33661 173 45 score score NOUN bracis-33661 173 46 was be AUX bracis-33661 173 47 used use VERB bracis-33661 173 48 as as ADP bracis-33661 173 49 the the DET bracis-33661 173 50 most most ADV bracis-33661 173 51 important important ADJ bracis-33661 173 52 metric metric ADJ bracis-33661 173 53 to to ADP bracis-33661 173 54 both both DET bracis-33661 173 55 selection selection NOUN bracis-33661 173 56 of of ADP bracis-33661 173 57 hyper hyper NOUN bracis-33661 173 58 - - NOUN bracis-33661 173 59 parameters parameter NOUN bracis-33661 173 60 and and CCONJ bracis-33661 173 61 evaluation evaluation NOUN bracis-33661 173 62 of of ADP bracis-33661 173 63 the the DET bracis-33661 173 64 predictive predictive ADJ bracis-33661 173 65 performance performance NOUN bracis-33661 173 66 of of ADP bracis-33661 173 67 the the DET bracis-33661 173 68 neural neural ADJ bracis-33661 173 69 networks network NOUN bracis-33661 173 70 ’ ' PUNCT bracis-33661 173 71 ( ( PUNCT bracis-33661 173 72 nn nn X bracis-33661 173 73 for for ADP bracis-33661 173 74 short short ADJ bracis-33661 173 75 in in ADP bracis-33661 173 76 the the DET bracis-33661 173 77 table table NOUN bracis-33661 173 78 ) ) PUNCT bracis-33661 173 79 architectures architecture NOUN bracis-33661 173 80 . . PUNCT bracis-33661 174 1 the the DET bracis-33661 174 2 indicated indicate VERB bracis-33661 174 3 recall recall NOUN bracis-33661 174 4 and and CCONJ bracis-33661 174 5 precision precision NOUN bracis-33661 174 6 values value NOUN bracis-33661 174 7 were be AUX bracis-33661 174 8 chosen choose VERB bracis-33661 174 9 to to PART bracis-33661 174 10 facilitate facilitate VERB bracis-33661 174 11 the the DET bracis-33661 174 12 understanding understanding NOUN bracis-33661 174 13 of of ADP bracis-33661 174 14 the the DET bracis-33661 174 15 existing exist VERB bracis-33661 174 16 trade trade NOUN bracis-33661 174 17 - - PUNCT bracis-33661 174 18 off off NOUN bracis-33661 174 19 in in ADP bracis-33661 174 20 each each DET bracis-33661 174 21 architecture architecture NOUN bracis-33661 174 22 , , PUNCT bracis-33661 174 23 making make VERB bracis-33661 174 24 evident evident ADJ bracis-33661 174 25 any any DET bracis-33661 174 26 model model NOUN bracis-33661 174 27 fit fit ADJ bracis-33661 174 28 problems problem NOUN bracis-33661 174 29 with with ADP bracis-33661 174 30 the the DET bracis-33661 174 31 data datum NOUN bracis-33661 174 32 . . PUNCT bracis-33661 175 1 and and CCONJ bracis-33661 175 2 finally finally ADV bracis-33661 175 3 , , PUNCT bracis-33661 175 4 the the DET bracis-33661 175 5 choice choice NOUN bracis-33661 175 6 of of ADP bracis-33661 175 7 accuracy accuracy NOUN bracis-33661 175 8 was be AUX bracis-33661 175 9 made make VERB bracis-33661 175 10 so so SCONJ bracis-33661 175 11 that that SCONJ bracis-33661 175 12 we we PRON bracis-33661 175 13 could could AUX bracis-33661 175 14 always always ADV bracis-33661 175 15 compare compare VERB bracis-33661 175 16 with with ADP bracis-33661 175 17 the the DET bracis-33661 175 18 reference reference NOUN bracis-33661 175 19 f1 f1 NOUN bracis-33661 175 20 - - PUNCT bracis-33661 175 21 score score NOUN bracis-33661 175 22 , , PUNCT bracis-33661 175 23 ensuring ensure VERB bracis-33661 175 24 that that SCONJ bracis-33661 175 25 the the DET bracis-33661 175 26 balancing balancing NOUN bracis-33661 175 27 of of ADP bracis-33661 175 28 the the DET bracis-33661 175 29 metrics metric NOUN bracis-33661 175 30 was be AUX bracis-33661 175 31 obeyed obey VERB bracis-33661 175 32 and and CCONJ bracis-33661 175 33 that that SCONJ bracis-33661 175 34 there there PRON bracis-33661 175 35 were be VERB bracis-33661 175 36 no no DET bracis-33661 175 37 problems problem NOUN bracis-33661 175 38 related relate VERB bracis-33661 175 39 to to ADP bracis-33661 175 40 the the DET bracis-33661 175 41 adaptation adaptation NOUN bracis-33661 175 42 of of ADP bracis-33661 175 43 the the DET bracis-33661 175 44 models model NOUN bracis-33661 175 45 to to ADP bracis-33661 175 46 the the DET bracis-33661 175 47 class class NOUN bracis-33661 175 48 imbalance imbalance NOUN bracis-33661 175 49 . . PUNCT bracis-33661 176 1 table table NOUN bracis-33661 176 2 3 3 NUM bracis-33661 176 3 . . PUNCT bracis-33661 177 1 statistical statistical ADJ bracis-33661 177 2 analysis analysis NOUN bracis-33661 177 3 using use VERB bracis-33661 177 4 the the DET bracis-33661 177 5 wilcoxon wilcoxon ADJ bracis-33661 177 6 test test NOUN bracis-33661 177 7 to to PART bracis-33661 177 8 evaluate evaluate VERB bracis-33661 177 9 the the DET bracis-33661 177 10 best good ADJ bracis-33661 177 11 configuration configuration NOUN bracis-33661 177 12 among among ADP bracis-33661 177 13 the the DET bracis-33661 177 14 combinations combination NOUN bracis-33661 177 15 of of ADP bracis-33661 177 16 gnns gnns ADJ bracis-33661 177 17 and and CCONJ bracis-33661 177 18 eegraph eegraph PROPN bracis-33661 177 19 neighborhood neighborhood NOUN bracis-33661 177 20 levels level NOUN bracis-33661 177 21 . . PUNCT bracis-33661 178 1 adopting adopt VERB bracis-33661 178 2 a a DET bracis-33661 178 3 confidence confidence NOUN bracis-33661 178 4 level level NOUN bracis-33661 178 5 of of ADP bracis-33661 178 6 95 95 NUM bracis-33661 178 7 % % NOUN bracis-33661 178 8 ( ( PUNCT bracis-33661 178 9 \(\alpha \(\alpha NOUN bracis-33661 178 10 \ \ NOUN bracis-33661 178 11 ) ) PUNCT bracis-33661 178 12 = = SYM bracis-33661 178 13 0.05 0.05 NUM bracis-33661 178 14 ) ) PUNCT bracis-33661 178 15 , , PUNCT bracis-33661 178 16 the the DET bracis-33661 178 17 symbols symbol NOUN bracis-33661 178 18 > > X bracis-33661 178 19 , , PUNCT bracis-33661 178 20 < < X bracis-33661 178 21 , , PUNCT bracis-33661 178 22 and and CCONJ bracis-33661 178 23 \(\approx \(\approx PRON bracis-33661 178 24 \ \ NOUN bracis-33661 178 25 ) ) PUNCT bracis-33661 178 26 indicate indicate VERB bracis-33661 178 27 that that SCONJ bracis-33661 178 28 the the DET bracis-33661 178 29 variable variable NOUN bracis-33661 178 30 in in ADP bracis-33661 178 31 the the DET bracis-33661 178 32 column column NOUN bracis-33661 178 33 is be AUX bracis-33661 178 34 respectively respectively ADV bracis-33661 178 35 better well ADJ bracis-33661 178 36 , , PUNCT bracis-33661 178 37 worse bad ADJ bracis-33661 178 38 , , PUNCT bracis-33661 178 39 or or CCONJ bracis-33661 178 40 equivalent equivalent ADJ bracis-33661 178 41 to to ADP bracis-33661 178 42 the the DET bracis-33661 178 43 variable variable NOUN bracis-33661 178 44 in in ADP bracis-33661 178 45 the the DET bracis-33661 178 46 row row NOUN bracis-33661 178 47 . . PUNCT bracis-33661 179 1 inner inner ADJ bracis-33661 179 2 h=4 h=4 PROPN bracis-33661 179 3 results result NOUN bracis-33661 179 4 omitted omit VERB bracis-33661 179 5 without without ADP bracis-33661 179 6 affecting affect VERB bracis-33661 179 7 the the DET bracis-33661 179 8 analysis.full analysis.full PROPN bracis-33661 179 9 size size NOUN bracis-33661 179 10 table table NOUN bracis-33661 179 11 the the DET bracis-33661 179 12 results result NOUN bracis-33661 179 13 obtained obtain VERB bracis-33661 179 14 in in ADP bracis-33661 179 15 the the DET bracis-33661 179 16 table table NOUN bracis-33661 179 17     SPACE bracis-33661 179 18 2 2 NUM bracis-33661 179 19 shows show VERB bracis-33661 179 20 a a DET bracis-33661 179 21 significant significant ADJ bracis-33661 179 22 improvement improvement NOUN bracis-33661 179 23 for for ADP bracis-33661 179 24 all all DET bracis-33661 179 25 gnns gnns ADJ bracis-33661 179 26 proposed propose VERB bracis-33661 179 27 architectures architecture NOUN bracis-33661 179 28 in in ADP bracis-33661 179 29 the the DET bracis-33661 179 30 target target NOUN bracis-33661 179 31 metric metric ADJ bracis-33661 179 32 f1 f1 NOUN bracis-33661 179 33 - - PUNCT bracis-33661 179 34 score score NOUN bracis-33661 179 35 when when SCONJ bracis-33661 179 36 compared compare VERB bracis-33661 179 37 to to ADP bracis-33661 179 38 the the DET bracis-33661 179 39 reference reference NOUN bracis-33661 179 40 lstm lstm NOUN bracis-33661 179 41 architecture architecture NOUN bracis-33661 179 42 . . PUNCT bracis-33661 180 1 in in ADP bracis-33661 180 2 addition addition NOUN bracis-33661 180 3 to to ADP bracis-33661 180 4 a a DET bracis-33661 180 5 significant significant ADJ bracis-33661 180 6 improvement improvement NOUN bracis-33661 180 7 of of ADP bracis-33661 180 8 almost almost ADV bracis-33661 180 9 18 18 NUM bracis-33661 180 10 % % NOUN bracis-33661 180 11 in in ADP bracis-33661 180 12 the the DET bracis-33661 180 13 target target NOUN bracis-33661 180 14 metric metric NOUN bracis-33661 180 15 of of ADP bracis-33661 180 16 the the DET bracis-33661 180 17 study study NOUN bracis-33661 180 18 and and CCONJ bracis-33661 180 19 a a DET bracis-33661 180 20 better well ADJ bracis-33661 180 21 balance balance NOUN bracis-33661 180 22 between between ADP bracis-33661 180 23 precision precision NOUN bracis-33661 180 24 and and CCONJ bracis-33661 180 25 recall recall NOUN bracis-33661 180 26 , , PUNCT bracis-33661 180 27 a a DET bracis-33661 180 28 better well ADJ bracis-33661 180 29 fit fit NOUN bracis-33661 180 30 of of ADP bracis-33661 180 31 the the DET bracis-33661 180 32 gnns gnns ADJ bracis-33661 180 33 architectures architecture NOUN bracis-33661 180 34 compared compare VERB bracis-33661 180 35 to to ADP bracis-33661 180 36 the the DET bracis-33661 180 37 lstm lstm NOUN bracis-33661 180 38 is be AUX bracis-33661 180 39 also also ADV bracis-33661 180 40 noticeable noticeable ADJ bracis-33661 180 41 since since SCONJ bracis-33661 180 42 the the DET bracis-33661 180 43 precision precision NOUN bracis-33661 180 44 of of ADP bracis-33661 180 45 the the DET bracis-33661 180 46 lstm lstm PROPN bracis-33661 180 47 network network NOUN bracis-33661 180 48 has have VERB bracis-33661 180 49 signs sign NOUN bracis-33661 180 50 of of ADP bracis-33661 180 51 bias bias NOUN bracis-33661 180 52 . . PUNCT bracis-33661 181 1 the the DET bracis-33661 181 2 eegraph eegraph PROPN bracis-33661 181 3 strategy strategy NOUN bracis-33661 181 4 for for ADP bracis-33661 181 5 the the DET bracis-33661 181 6 graph graph NOUN bracis-33661 181 7 structure structure NOUN bracis-33661 181 8 is be AUX bracis-33661 181 9 shown show VERB bracis-33661 181 10 in in ADP bracis-33661 181 11 the the DET bracis-33661 181 12 table table NOUN bracis-33661 181 13 according accord VERB bracis-33661 181 14 to to ADP bracis-33661 181 15 the the DET bracis-33661 181 16 number number NOUN bracis-33661 181 17 of of ADP bracis-33661 181 18 neighborhood neighborhood NOUN bracis-33661 181 19 levels level NOUN bracis-33661 181 20 ( ( PUNCT bracis-33661 181 21 h h NOUN bracis-33661 181 22 ) ) PUNCT bracis-33661 181 23 , , PUNCT bracis-33661 181 24 considered consider VERB bracis-33661 181 25 from from ADP bracis-33661 181 26 1 1 NUM bracis-33661 181 27 to to ADP bracis-33661 181 28 4 4 NUM bracis-33661 181 29 . . PUNCT bracis-33661 181 30 with with ADP bracis-33661 181 31 exception exception NOUN bracis-33661 181 32 of of ADP bracis-33661 181 33 gcnii gcnii NOUN bracis-33661 181 34 , , PUNCT bracis-33661 181 35 smaller small ADJ bracis-33661 181 36 neighborhood neighborhood NOUN bracis-33661 181 37 levels level NOUN bracis-33661 181 38 like like ADP bracis-33661 181 39 1 1 NUM bracis-33661 181 40 and and CCONJ bracis-33661 181 41 2 2 NUM bracis-33661 181 42 usually usually ADV bracis-33661 181 43 provide provide VERB bracis-33661 181 44 better well ADJ bracis-33661 181 45 results result NOUN bracis-33661 181 46 in in ADP bracis-33661 181 47 terms term NOUN bracis-33661 181 48 of of ADP bracis-33661 181 49 the the DET bracis-33661 181 50 target target NOUN bracis-33661 181 51 metric metric ADJ bracis-33661 181 52 f1 f1 NOUN bracis-33661 181 53 - - PUNCT bracis-33661 181 54 score score NOUN bracis-33661 181 55 , , PUNCT bracis-33661 181 56 as as ADV bracis-33661 181 57 well well ADV bracis-33661 181 58 as as ADP bracis-33661 181 59 in in ADP bracis-33661 181 60 terms term NOUN bracis-33661 181 61 of of ADP bracis-33661 181 62 accuracy accuracy NOUN bracis-33661 181 63 and and CCONJ bracis-33661 181 64 recall recall NOUN bracis-33661 181 65 . . PUNCT bracis-33661 182 1 this this PRON bracis-33661 182 2 means mean VERB bracis-33661 182 3 that that SCONJ bracis-33661 182 4 a a DET bracis-33661 182 5 strategy strategy NOUN bracis-33661 182 6 to to PART bracis-33661 182 7 increment increment VERB bracis-33661 182 8 the the DET bracis-33661 182 9 number number NOUN bracis-33661 182 10 of of ADP bracis-33661 182 11 edges edge NOUN bracis-33661 182 12 based base VERB bracis-33661 182 13 solely solely ADV bracis-33661 182 14 on on ADP bracis-33661 182 15 neighborhood neighborhood NOUN bracis-33661 182 16 levels level NOUN bracis-33661 182 17 , , PUNCT bracis-33661 182 18 without without ADP bracis-33661 182 19 considering consider VERB bracis-33661 182 20 weight weight NOUN bracis-33661 182 21 differences difference NOUN bracis-33661 182 22 during during ADP bracis-33661 182 23 the the DET bracis-33661 182 24 edge edge NOUN bracis-33661 182 25 construction construction NOUN bracis-33661 182 26 steps step NOUN bracis-33661 182 27 , , PUNCT bracis-33661 182 28 does do AUX bracis-33661 182 29 not not PART bracis-33661 182 30 contribute contribute VERB bracis-33661 182 31 to to ADP bracis-33661 182 32 an an DET bracis-33661 182 33 enhancement enhancement NOUN bracis-33661 182 34 of of ADP bracis-33661 182 35 network network NOUN bracis-33661 182 36 learning learn VERB bracis-33661 182 37 for for ADP bracis-33661 182 38 the the DET bracis-33661 182 39 given give VERB bracis-33661 182 40 problem problem NOUN bracis-33661 182 41 . . PUNCT bracis-33661 183 1 on on ADP bracis-33661 183 2 the the DET bracis-33661 183 3 other other ADJ bracis-33661 183 4 hand hand NOUN bracis-33661 183 5 , , PUNCT bracis-33661 183 6 it it PRON bracis-33661 183 7 also also ADV bracis-33661 183 8 indicates indicate VERB bracis-33661 183 9 that that SCONJ bracis-33661 183 10 the the DET bracis-33661 183 11 relationships relationship NOUN bracis-33661 183 12 among among ADP bracis-33661 183 13 adjacent adjacent ADJ bracis-33661 183 14 electrodes electrode NOUN bracis-33661 183 15 contribute contribute VERB bracis-33661 183 16 decisively decisively ADV bracis-33661 183 17 to to PART bracis-33661 183 18 achieve achieve VERB bracis-33661 183 19 better well ADJ bracis-33661 183 20 predictive predictive ADJ bracis-33661 183 21 performance performance NOUN bracis-33661 183 22 . . PUNCT bracis-33661 184 1 table table NOUN bracis-33661 184 2     SPACE bracis-33661 184 3 3 3 NUM bracis-33661 184 4 shows show VERB bracis-33661 184 5 the the DET bracis-33661 184 6 wilcoxon wilcoxon ADJ bracis-33661 184 7 statistical statistical ADJ bracis-33661 184 8 test test NOUN bracis-33661 184 9 between between ADP bracis-33661 184 10 all all DET bracis-33661 184 11 the the DET bracis-33661 184 12 architectures architecture NOUN bracis-33661 184 13 used use VERB bracis-33661 184 14 in in ADP bracis-33661 184 15 the the DET bracis-33661 184 16 experiment experiment NOUN bracis-33661 184 17 , , PUNCT bracis-33661 184 18 making make VERB bracis-33661 184 19 explicit explicit ADJ bracis-33661 184 20 the the DET bracis-33661 184 21 superiority superiority NOUN bracis-33661 184 22 of of ADP bracis-33661 184 23 all all DET bracis-33661 184 24 gnns gnns ADJ bracis-33661 184 25 techniques technique NOUN bracis-33661 184 26 in in ADP bracis-33661 184 27 comparison comparison NOUN bracis-33661 184 28 to to ADP bracis-33661 184 29 lstm lstm NOUN bracis-33661 184 30 . . PUNCT bracis-33661 185 1 the the DET bracis-33661 185 2 results result NOUN bracis-33661 185 3 of of ADP bracis-33661 185 4 the the DET bracis-33661 185 5 sgc sgc PROPN bracis-33661 185 6 algorithm algorithm NOUN bracis-33661 185 7 with with ADP bracis-33661 185 8 h=1 h=1 PROPN bracis-33661 185 9 are be AUX bracis-33661 185 10 statistically statistically ADV bracis-33661 185 11 superior superior ADJ bracis-33661 185 12 to to ADP bracis-33661 185 13 all all DET bracis-33661 185 14 other other ADJ bracis-33661 185 15 gnns gnns NOUN bracis-33661 185 16 and and CCONJ bracis-33661 185 17 eegraph eegraph PROPN bracis-33661 185 18 configurations configuration NOUN bracis-33661 185 19 . . PUNCT bracis-33661 186 1 for for ADP bracis-33661 186 2 h=2 h=2 PROPN bracis-33661 186 3 and and CCONJ bracis-33661 186 4 h=3 h=3 PROPN bracis-33661 186 5 , , PUNCT bracis-33661 186 6 sgc sgc PROPN bracis-33661 186 7 and and CCONJ bracis-33661 186 8 gcn gcn NOUN bracis-33661 186 9 are be AUX bracis-33661 186 10 statistically statistically ADV bracis-33661 186 11 equivalent equivalent ADJ bracis-33661 186 12 and and CCONJ bracis-33661 186 13 outperform outperform VERB bracis-33661 186 14 all all DET bracis-33661 186 15 other other ADJ bracis-33661 186 16 architectures architecture NOUN bracis-33661 186 17 , , PUNCT bracis-33661 186 18 demonstrating demonstrate VERB bracis-33661 186 19 the the DET bracis-33661 186 20 potential potential NOUN bracis-33661 186 21 of of ADP bracis-33661 186 22 both both DET bracis-33661 186 23 techniques technique NOUN bracis-33661 186 24 . . PUNCT bracis-33661 187 1 table table NOUN bracis-33661 187 2 4 4 NUM bracis-33661 187 3 . . PUNCT bracis-33661 188 1 statistical statistical ADJ bracis-33661 188 2 analysis analysis NOUN bracis-33661 188 3 using use VERB bracis-33661 188 4 the the DET bracis-33661 188 5 wilcoxon wilcoxon ADJ bracis-33661 188 6 test test NOUN bracis-33661 188 7 to to PART bracis-33661 188 8 evaluate evaluate VERB bracis-33661 188 9 the the DET bracis-33661 188 10 models model NOUN bracis-33661 188 11 regardless regardless ADV bracis-33661 188 12 of of ADP bracis-33661 188 13 the the DET bracis-33661 188 14 eegraph eegraph NOUN bracis-33661 188 15 modeling modeling NOUN bracis-33661 188 16 strategy strategy NOUN bracis-33661 188 17 . . PUNCT bracis-33661 189 1 adopting adopt VERB bracis-33661 189 2 a a DET bracis-33661 189 3 confidence confidence NOUN bracis-33661 189 4 level level NOUN bracis-33661 189 5 of of ADP bracis-33661 189 6 95 95 NUM bracis-33661 189 7 % % NOUN bracis-33661 189 8 ( ( PUNCT bracis-33661 189 9 \(\alpha \(\alpha NOUN bracis-33661 189 10 \ \ NOUN bracis-33661 189 11 ) ) PUNCT bracis-33661 189 12 = = SYM bracis-33661 189 13 0.05 0.05 NUM bracis-33661 189 14 ) ) PUNCT bracis-33661 189 15 , , PUNCT bracis-33661 189 16 the the DET bracis-33661 189 17 symbols symbol NOUN bracis-33661 189 18 > > X bracis-33661 189 19 , , PUNCT bracis-33661 189 20 < < X bracis-33661 189 21 , , PUNCT bracis-33661 189 22 and and CCONJ bracis-33661 189 23 \(\approx \(\approx PRON bracis-33661 189 24 \ \ NOUN bracis-33661 189 25 ) ) PUNCT bracis-33661 189 26 indicate indicate VERB bracis-33661 189 27 that that SCONJ bracis-33661 189 28 the the DET bracis-33661 189 29 variable variable NOUN bracis-33661 189 30 in in ADP bracis-33661 189 31 the the DET bracis-33661 189 32 column column NOUN bracis-33661 189 33 is be AUX bracis-33661 189 34 respectively respectively ADV bracis-33661 189 35 better well ADJ bracis-33661 189 36 , , PUNCT bracis-33661 189 37 worse bad ADJ bracis-33661 189 38 , , PUNCT bracis-33661 189 39 or or CCONJ bracis-33661 189 40 equivalent equivalent ADJ bracis-33661 189 41 to to ADP bracis-33661 189 42 the the DET bracis-33661 189 43 variable variable NOUN bracis-33661 189 44 in in ADP bracis-33661 189 45 the the DET bracis-33661 189 46 row.full row.full NOUN bracis-33661 189 47 size size NOUN bracis-33661 189 48 table table NOUN bracis-33661 189 49 table table NOUN bracis-33661 189 50 5 5 NUM bracis-33661 189 51 . . PUNCT bracis-33661 189 52 statistical statistical ADJ bracis-33661 189 53 analysis analysis NOUN bracis-33661 189 54 using use VERB bracis-33661 189 55 the the DET bracis-33661 189 56 wilcoxon wilcoxon ADJ bracis-33661 189 57 test test NOUN bracis-33661 189 58 to to PART bracis-33661 189 59 evaluate evaluate VERB bracis-33661 189 60 the the DET bracis-33661 189 61 eegraph eegraph NOUN bracis-33661 189 62 modeling modeling NOUN bracis-33661 189 63 strategy strategy NOUN bracis-33661 189 64 , , PUNCT bracis-33661 189 65 regardless regardless ADV bracis-33661 189 66 of of ADP bracis-33661 189 67 the the DET bracis-33661 189 68 model model NOUN bracis-33661 189 69 architecture architecture NOUN bracis-33661 189 70 . . PUNCT bracis-33661 190 1 adopting adopt VERB bracis-33661 190 2 a a DET bracis-33661 190 3 confidence confidence NOUN bracis-33661 190 4 level level NOUN bracis-33661 190 5 of of ADP bracis-33661 190 6 95 95 NUM bracis-33661 190 7 % % NOUN bracis-33661 190 8 ( ( PUNCT bracis-33661 190 9 \(\alpha \(\alpha NOUN bracis-33661 190 10 \ \ NOUN bracis-33661 190 11 ) ) PUNCT bracis-33661 190 12 = = SYM bracis-33661 190 13 0.05 0.05 NUM bracis-33661 190 14 ) ) PUNCT bracis-33661 190 15 , , PUNCT bracis-33661 190 16 the the DET bracis-33661 190 17 symbols symbol NOUN bracis-33661 190 18 > > X bracis-33661 190 19 , , PUNCT bracis-33661 190 20 < < X bracis-33661 190 21 , , PUNCT bracis-33661 190 22 and and CCONJ bracis-33661 190 23 \(\approx \(\approx PRON bracis-33661 190 24 \ \ NOUN bracis-33661 190 25 ) ) PUNCT bracis-33661 190 26 indicate indicate VERB bracis-33661 190 27 that that SCONJ bracis-33661 190 28 the the DET bracis-33661 190 29 variable variable NOUN bracis-33661 190 30 in in ADP bracis-33661 190 31 the the DET bracis-33661 190 32 column column NOUN bracis-33661 190 33 is be AUX bracis-33661 190 34 respectively respectively ADV bracis-33661 190 35 better well ADJ bracis-33661 190 36 , , PUNCT bracis-33661 190 37 worse bad ADJ bracis-33661 190 38 , , PUNCT bracis-33661 190 39 or or CCONJ bracis-33661 190 40 equivalent equivalent ADJ bracis-33661 190 41 to to ADP bracis-33661 190 42 the the DET bracis-33661 190 43 variable variable NOUN bracis-33661 190 44 in in ADP bracis-33661 190 45 the the DET bracis-33661 190 46 row.full row.full NOUN bracis-33661 190 47 size size NOUN bracis-33661 190 48 table table NOUN bracis-33661 190 49 when when SCONJ bracis-33661 190 50 comparing compare VERB bracis-33661 190 51 the the DET bracis-33661 190 52 models model NOUN bracis-33661 190 53 regardless regardless ADV bracis-33661 190 54 of of ADP bracis-33661 190 55 the the DET bracis-33661 190 56 h h NOUN bracis-33661 190 57 - - PUNCT bracis-33661 190 58 eegraph eegraph NOUN bracis-33661 190 59 architecture architecture NOUN bracis-33661 190 60 , , PUNCT bracis-33661 190 61 observed observe VERB bracis-33661 190 62 in in ADP bracis-33661 190 63 table table NOUN bracis-33661 190 64     SPACE bracis-33661 190 65 4 4 NUM bracis-33661 190 66 , , PUNCT bracis-33661 190 67 the the DET bracis-33661 190 68 sgc sgc PROPN bracis-33661 190 69 model model PROPN bracis-33661 190 70 consistently consistently ADV bracis-33661 190 71 outperforms outperform NOUN bracis-33661 190 72 gcn gcn ADJ bracis-33661 190 73 , , PUNCT bracis-33661 190 74 sage sage NOUN bracis-33661 190 75 , , PUNCT bracis-33661 190 76 and and CCONJ bracis-33661 190 77 gcnii gcnii NOUN bracis-33661 190 78 , , PUNCT bracis-33661 190 79 indicating indicate VERB bracis-33661 190 80 it it PRON bracis-33661 190 81 is be AUX bracis-33661 190 82 the the DET bracis-33661 190 83 most most ADV bracis-33661 190 84 effective effective ADJ bracis-33661 190 85 model model NOUN bracis-33661 190 86 in in ADP bracis-33661 190 87 this this DET bracis-33661 190 88 context context NOUN bracis-33661 190 89 . . PUNCT bracis-33661 191 1 gcn gcn NOUN bracis-33661 191 2 also also ADV bracis-33661 191 3 performs perform VERB bracis-33661 191 4 well well ADV bracis-33661 191 5 , , PUNCT bracis-33661 191 6 surpassing surpass VERB bracis-33661 191 7 sage sage NOUN bracis-33661 191 8 and and CCONJ bracis-33661 191 9 gcnii gcnii NOUN bracis-33661 191 10 . . PUNCT bracis-33661 192 1 sage sage NOUN bracis-33661 192 2 shows show VERB bracis-33661 192 3 better well ADJ bracis-33661 192 4 performance performance NOUN bracis-33661 192 5 than than ADP bracis-33661 192 6 gcnii gcnii NOUN bracis-33661 192 7 but but CCONJ bracis-33661 192 8 is be AUX bracis-33661 192 9 outperformed outperform VERB bracis-33661 192 10 by by ADP bracis-33661 192 11 both both DET bracis-33661 192 12 gcn gcn NOUN bracis-33661 192 13 and and CCONJ bracis-33661 192 14 sgc sgc PROPN bracis-33661 192 15 . . PUNCT bracis-33661 193 1 in in ADP bracis-33661 193 2 addition addition NOUN bracis-33661 193 3 to to ADP bracis-33661 193 4 the the DET bracis-33661 193 5 previous previous ADJ bracis-33661 193 6 statistical statistical ADJ bracis-33661 193 7 analysis analysis NOUN bracis-33661 193 8 , , PUNCT bracis-33661 193 9 one one PRON bracis-33661 193 10 can can AUX bracis-33661 193 11 see see VERB bracis-33661 193 12 in in ADP bracis-33661 193 13 table table NOUN bracis-33661 193 14     SPACE bracis-33661 193 15 2 2 NUM bracis-33661 193 16 that that PRON bracis-33661 193 17 gcnii gcnii NOUN bracis-33661 193 18 model model NOUN bracis-33661 193 19 has have VERB bracis-33661 193 20 very very ADV bracis-33661 193 21 lower low ADJ bracis-33661 193 22 predictive predictive ADJ bracis-33661 193 23 performance performance NOUN bracis-33661 193 24 in in ADP bracis-33661 193 25 comparison comparison NOUN bracis-33661 193 26 to to ADP bracis-33661 193 27 all all DET bracis-33661 193 28 other other ADJ bracis-33661 193 29 gnns gnn NOUN bracis-33661 193 30 . . PUNCT bracis-33661 194 1 it it PRON bracis-33661 194 2 is be AUX bracis-33661 194 3 important important ADJ bracis-33661 194 4 to to PART bracis-33661 194 5 note note VERB bracis-33661 194 6 that that SCONJ bracis-33661 194 7 we we PRON bracis-33661 194 8 used use VERB bracis-33661 194 9 only only ADV bracis-33661 194 10 a a DET bracis-33661 194 11 single single ADJ bracis-33661 194 12 hyperparameters hyperparameter NOUN bracis-33661 194 13 configuration configuration NOUN bracis-33661 194 14 for for ADP bracis-33661 194 15 the the DET bracis-33661 194 16 gcnii gcnii NOUN bracis-33661 194 17 architecture architecture NOUN bracis-33661 194 18 , , PUNCT bracis-33661 194 19 which which PRON bracis-33661 194 20 could could AUX bracis-33661 194 21 be be AUX bracis-33661 194 22 a a DET bracis-33661 194 23 potential potential ADJ bracis-33661 194 24 limitation limitation NOUN bracis-33661 194 25 in in ADP bracis-33661 194 26 fully fully ADV bracis-33661 194 27 evaluating evaluate VERB bracis-33661 194 28 its its PRON bracis-33661 194 29 performance performance NOUN bracis-33661 194 30 , , PUNCT bracis-33661 194 31 as as SCONJ bracis-33661 194 32 it it PRON bracis-33661 194 33 may may AUX bracis-33661 194 34 indicate indicate VERB bracis-33661 194 35 an an DET bracis-33661 194 36 stronger strong ADJ bracis-33661 194 37 dependence dependence NOUN bracis-33661 194 38 about about ADP bracis-33661 194 39 the the DET bracis-33661 194 40 choice choice NOUN bracis-33661 194 41 of of ADP bracis-33661 194 42 \(\alpha \(\alpha X bracis-33661 194 43 \ \ NOUN bracis-33661 194 44 ) ) PUNCT bracis-33661 194 45 and and CCONJ bracis-33661 194 46 \(\beta \(\beta PRON bracis-33661 194 47 \ \ NOUN bracis-33661 194 48 ) ) PUNCT bracis-33661 194 49 hyperparameters hyperparameter NOUN bracis-33661 194 50 , , PUNCT bracis-33661 194 51 related relate VERB bracis-33661 194 52 respectively respectively ADV bracis-33661 194 53 to to ADP bracis-33661 194 54 the the DET bracis-33661 194 55 initial initial ADJ bracis-33661 194 56 residual residual ADJ bracis-33661 194 57 and and CCONJ bracis-33661 194 58 identity identity NOUN bracis-33661 194 59 mapping mapping NOUN bracis-33661 194 60 contributions contribution NOUN bracis-33661 194 61 . . PUNCT bracis-33661 195 1 because because SCONJ bracis-33661 195 2 of of ADP bracis-33661 195 3 that that PRON bracis-33661 195 4 , , PUNCT bracis-33661 195 5 we we PRON bracis-33661 195 6 remove remove VERB bracis-33661 195 7 gcnii gcnii NOUN bracis-33661 195 8 results result NOUN bracis-33661 195 9 from from ADP bracis-33661 195 10 our our PRON bracis-33661 195 11 next next ADJ bracis-33661 195 12 statistical statistical ADJ bracis-33661 195 13 analysis analysis NOUN bracis-33661 195 14 , , PUNCT bracis-33661 195 15 as as SCONJ bracis-33661 195 16 we we PRON bracis-33661 195 17 are be AUX bracis-33661 195 18 interested interested ADJ bracis-33661 195 19 in in ADP bracis-33661 195 20 identifying identify VERB bracis-33661 195 21 the the DET bracis-33661 195 22 contribution contribution NOUN bracis-33661 195 23 of of ADP bracis-33661 195 24 different different ADJ bracis-33661 195 25 neighborhood neighborhood NOUN bracis-33661 195 26 levels level NOUN bracis-33661 195 27 in in ADP bracis-33661 195 28 our our PRON bracis-33661 195 29 eegraph eegraph NOUN bracis-33661 195 30 modeling modeling NOUN bracis-33661 195 31 strategy strategy NOUN bracis-33661 195 32 . . PUNCT bracis-33661 196 1 table table NOUN bracis-33661 196 2     SPACE bracis-33661 196 3 5 5 NUM bracis-33661 196 4 shows show VERB bracis-33661 196 5 that that SCONJ bracis-33661 196 6 one one NUM bracis-33661 196 7 level level NOUN bracis-33661 196 8 neighborhood neighborhood NOUN bracis-33661 196 9 ( ( PUNCT bracis-33661 196 10 h=1 h=1 NOUN bracis-33661 196 11 ) ) PUNCT bracis-33661 196 12 outperforms outperform VERB bracis-33661 196 13 the the DET bracis-33661 196 14 other other ADJ bracis-33661 196 15 neighborhood neighborhood NOUN bracis-33661 196 16 configurations configuration NOUN bracis-33661 196 17 across across ADP bracis-33661 196 18 gcn gcn PROPN bracis-33661 196 19 , , PUNCT bracis-33661 196 20 sgc sgc ADJ bracis-33661 196 21 and and CCONJ bracis-33661 196 22 sage sage NOUN bracis-33661 196 23 gnns gnn NOUN bracis-33661 196 24 . . PUNCT bracis-33661 197 1 conversely conversely ADV bracis-33661 197 2 , , PUNCT bracis-33661 197 3 the the DET bracis-33661 197 4 h=4 h=4 PROPN bracis-33661 197 5 configuration configuration NOUN bracis-33661 197 6 is be AUX bracis-33661 197 7 less less ADV bracis-33661 197 8 effective effective ADJ bracis-33661 197 9 compared compare VERB bracis-33661 197 10 to to ADP bracis-33661 197 11 the the DET bracis-33661 197 12 others other NOUN bracis-33661 197 13 , , PUNCT bracis-33661 197 14 maybe maybe ADV bracis-33661 197 15 due due ADP bracis-33661 197 16 to to ADP bracis-33661 197 17 the the DET bracis-33661 197 18 difficult difficult ADJ bracis-33661 197 19 to to PART bracis-33661 197 20 propagate propagate VERB bracis-33661 197 21 salient salient ADJ bracis-33661 197 22 information information NOUN bracis-33661 197 23 over over ADP bracis-33661 197 24 a a DET bracis-33661 197 25 denser dense ADJ bracis-33661 197 26 graph graph NOUN bracis-33661 197 27 . . PUNCT bracis-33661 198 1 in in ADP bracis-33661 198 2 summary summary NOUN bracis-33661 198 3 , , PUNCT bracis-33661 198 4 the the DET bracis-33661 198 5 results result NOUN bracis-33661 198 6 presented present VERB bracis-33661 198 7 in in ADP bracis-33661 198 8 this this DET bracis-33661 198 9 section section NOUN bracis-33661 198 10 highlight highlight VERB bracis-33661 198 11 the the DET bracis-33661 198 12 substantial substantial ADJ bracis-33661 198 13 potential potential NOUN bracis-33661 198 14 of of ADP bracis-33661 198 15 leveraging leverage VERB bracis-33661 198 16 the the DET bracis-33661 198 17 links link NOUN bracis-33661 198 18 between between ADP bracis-33661 198 19 electrodes electrode NOUN bracis-33661 198 20 in in ADP bracis-33661 198 21 eeg eeg NOUN bracis-33661 198 22 classification classification NOUN bracis-33661 198 23 tasks task NOUN bracis-33661 198 24 . . PUNCT bracis-33661 199 1 furthermore furthermore ADV bracis-33661 199 2 , , PUNCT bracis-33661 199 3 they they PRON bracis-33661 199 4 also also ADV bracis-33661 199 5 show show VERB bracis-33661 199 6 that that SCONJ bracis-33661 199 7 eegraph eegraph NOUN bracis-33661 199 8 modeling modeling NOUN bracis-33661 199 9 strategy strategy NOUN bracis-33661 199 10 allow allow VERB bracis-33661 199 11 easy easy ADJ bracis-33661 199 12 integration integration NOUN bracis-33661 199 13 of of ADP bracis-33661 199 14 eeg eeg PROPN bracis-33661 199 15 data datum NOUN bracis-33661 199 16 with with ADP bracis-33661 199 17 gnns gnn NOUN bracis-33661 199 18 , , PUNCT bracis-33661 199 19 a a DET bracis-33661 199 20 powerful powerful ADJ bracis-33661 199 21 machine machine NOUN bracis-33661 199 22 learning learning NOUN bracis-33661 199 23 framework framework NOUN bracis-33661 199 24 . . PUNCT bracis-33661 200 1 overall overall ADV bracis-33661 200 2 , , PUNCT bracis-33661 200 3 the the DET bracis-33661 200 4 results result NOUN bracis-33661 200 5 confirm confirm VERB bracis-33661 200 6 that that SCONJ bracis-33661 200 7 the the DET bracis-33661 200 8 sgc sgc PROPN bracis-33661 200 9 and and CCONJ bracis-33661 200 10 1 1 NUM bracis-33661 200 11 - - NOUN bracis-33661 200 12 eegraph eegraph NOUN bracis-33661 200 13 is be AUX bracis-33661 200 14 the the DET bracis-33661 200 15 best well ADV bracis-33661 200 16 - - PUNCT bracis-33661 200 17 performing perform VERB bracis-33661 200 18 model model NOUN bracis-33661 200 19 . . PUNCT bracis-33661 201 1 5 5 NUM bracis-33661 201 2 conclusion conclusion NOUN bracis-33661 201 3 in in ADP bracis-33661 201 4 this this DET bracis-33661 201 5 study study NOUN bracis-33661 201 6 , , PUNCT bracis-33661 201 7 we we PRON bracis-33661 201 8 successfully successfully ADV bracis-33661 201 9 developed develop VERB bracis-33661 201 10 four four NUM bracis-33661 201 11 gnns gnns ADJ bracis-33661 201 12 architectures architecture NOUN bracis-33661 201 13 based base VERB bracis-33661 201 14 on on ADP bracis-33661 201 15 convolutional convolutional ADJ bracis-33661 201 16 dynamics dynamic NOUN bracis-33661 201 17 to to PART bracis-33661 201 18 predict predict VERB bracis-33661 201 19 ppc ppc NOUN bracis-33661 201 20 results result NOUN bracis-33661 201 21 . . PUNCT bracis-33661 202 1 these these DET bracis-33661 202 2 architectures architecture NOUN bracis-33661 202 3 effectively effectively ADV bracis-33661 202 4 explored explore VERB bracis-33661 202 5 both both DET bracis-33661 202 6 nonlinear nonlinear ADJ bracis-33661 202 7 transformations transformation NOUN bracis-33661 202 8 and and CCONJ bracis-33661 202 9 local local ADJ bracis-33661 202 10 connections connection NOUN bracis-33661 202 11 between between ADP bracis-33661 202 12 the the DET bracis-33661 202 13 vertices vertex NOUN bracis-33661 202 14 . . PUNCT bracis-33661 203 1 furthermore furthermore ADV bracis-33661 203 2 , , PUNCT bracis-33661 203 3 we we PRON bracis-33661 203 4 proposed propose VERB bracis-33661 203 5 eegraph eegraph NOUN bracis-33661 203 6 , , PUNCT bracis-33661 203 7 a a DET bracis-33661 203 8 modeling modeling NOUN bracis-33661 203 9 strategy strategy NOUN bracis-33661 203 10 to to PART bracis-33661 203 11 transform transform VERB bracis-33661 203 12 raw raw ADJ bracis-33661 203 13 eeg eeg NOUN bracis-33661 203 14 data datum NOUN bracis-33661 203 15 into into ADP bracis-33661 203 16 graph graph NOUN bracis-33661 203 17 structures structure NOUN bracis-33661 203 18 , , PUNCT bracis-33661 203 19 which which PRON bracis-33661 203 20 leverages leverage VERB bracis-33661 203 21 the the DET bracis-33661 203 22 connections connection NOUN bracis-33661 203 23 between between ADP bracis-33661 203 24 the the DET bracis-33661 203 25 vertices vertex NOUN bracis-33661 203 26 ( ( PUNCT bracis-33661 203 27 electrodes electrode NOUN bracis-33661 203 28 ) ) PUNCT bracis-33661 203 29 of of ADP bracis-33661 203 30 the the DET bracis-33661 203 31 graph graph NOUN bracis-33661 203 32 structures structure NOUN bracis-33661 203 33 to to ADP bracis-33661 203 34 the the DET bracis-33661 203 35 fullest full ADJ bracis-33661 203 36 extent extent NOUN bracis-33661 203 37 . . PUNCT bracis-33661 204 1 our our PRON bracis-33661 204 2 experimental experimental ADJ bracis-33661 204 3 setup setup NOUN bracis-33661 204 4 provided provide VERB bracis-33661 204 5 a a DET bracis-33661 204 6 comprehensive comprehensive ADJ bracis-33661 204 7 comparison comparison NOUN bracis-33661 204 8 between between ADP bracis-33661 204 9 the the DET bracis-33661 204 10 proposed propose VERB bracis-33661 204 11 gnns gnns PROPN bracis-33661 204 12 architectures architecture NOUN bracis-33661 204 13 and and CCONJ bracis-33661 204 14 an an DET bracis-33661 204 15 lstm lstm NOUN bracis-33661 204 16 architecture architecture NOUN bracis-33661 204 17 , , PUNCT bracis-33661 204 18 demonstrating demonstrate VERB bracis-33661 204 19 the the DET bracis-33661 204 20 satisfactory satisfactory ADJ bracis-33661 204 21 performance performance NOUN bracis-33661 204 22 of of ADP bracis-33661 204 23 our our PRON bracis-33661 204 24 proposals proposal NOUN bracis-33661 204 25 . . PUNCT bracis-33661 205 1 in in ADP bracis-33661 205 2 conclusion conclusion NOUN bracis-33661 205 3 , , PUNCT bracis-33661 205 4 the the DET bracis-33661 205 5 results result NOUN bracis-33661 205 6 obtained obtain VERB bracis-33661 205 7 from from ADP bracis-33661 205 8 our our PRON bracis-33661 205 9 study study NOUN bracis-33661 205 10 demonstrate demonstrate VERB bracis-33661 205 11 that that SCONJ bracis-33661 205 12 the the DET bracis-33661 205 13 proposed propose VERB bracis-33661 205 14 gnn gnn PROPN bracis-33661 205 15 architectures architecture NOUN bracis-33661 205 16 outperformed outperform VERB bracis-33661 205 17 the the DET bracis-33661 205 18 lstm lstm NOUN bracis-33661 205 19 in in ADP bracis-33661 205 20 terms term NOUN bracis-33661 205 21 of of ADP bracis-33661 205 22 various various ADJ bracis-33661 205 23 performance performance NOUN bracis-33661 205 24 metrics metric NOUN bracis-33661 205 25 . . PUNCT bracis-33661 206 1 the the DET bracis-33661 206 2 comparison comparison NOUN bracis-33661 206 3 was be AUX bracis-33661 206 4 carried carry VERB bracis-33661 206 5 out out ADP bracis-33661 206 6 considering consider VERB bracis-33661 206 7 different different ADJ bracis-33661 206 8 modeling modeling NOUN bracis-33661 206 9 architectures architecture NOUN bracis-33661 206 10 of of ADP bracis-33661 206 11 eeg eeg NOUN bracis-33661 206 12 data datum NOUN bracis-33661 206 13 as as ADP bracis-33661 206 14 a a DET bracis-33661 206 15 graph graph NOUN bracis-33661 206 16 . . PUNCT bracis-33661 207 1 this this DET bracis-33661 207 2 accomplishment accomplishment NOUN bracis-33661 207 3 aligns align VERB bracis-33661 207 4 with with ADP bracis-33661 207 5 the the DET bracis-33661 207 6 primary primary ADJ bracis-33661 207 7 objective objective NOUN bracis-33661 207 8 of of ADP bracis-33661 207 9 our our PRON bracis-33661 207 10 article article NOUN bracis-33661 207 11 , , PUNCT bracis-33661 207 12 which which PRON bracis-33661 207 13 aimed aim VERB bracis-33661 207 14 to to PART bracis-33661 207 15 explore explore VERB bracis-33661 207 16 the the DET bracis-33661 207 17 spatial spatial ADJ bracis-33661 207 18 structure structure NOUN bracis-33661 207 19 of of ADP bracis-33661 207 20 eeg eeg PROPN bracis-33661 207 21 and and CCONJ bracis-33661 207 22 establish establish VERB bracis-33661 207 23 its its PRON bracis-33661 207 24 potential potential ADJ bracis-33661 207 25 contribution contribution NOUN bracis-33661 207 26 to to PART bracis-33661 207 27 performance performance NOUN bracis-33661 207 28 improvement.throughout improvement.throughout ADP bracis-33661 207 29 the the DET bracis-33661 207 30 experiments experiment NOUN bracis-33661 207 31 , , PUNCT bracis-33661 207 32 it it PRON bracis-33661 207 33 was be AUX bracis-33661 207 34 observed observe VERB bracis-33661 207 35 that that SCONJ bracis-33661 207 36 the the DET bracis-33661 207 37 gnns gnns ADJ bracis-33661 207 38 architectures architecture NOUN bracis-33661 207 39 consistently consistently ADV bracis-33661 207 40 achieved achieve VERB bracis-33661 207 41 higher high ADJ bracis-33661 207 42 accuracy accuracy NOUN bracis-33661 207 43 , , PUNCT bracis-33661 207 44 f1 f1 NOUN bracis-33661 207 45 - - PUNCT bracis-33661 207 46 score score NOUN bracis-33661 207 47 , , PUNCT bracis-33661 207 48 and and CCONJ bracis-33661 207 49 recall recall NOUN bracis-33661 207 50 values value NOUN bracis-33661 207 51 compared compare VERB bracis-33661 207 52 to to ADP bracis-33661 207 53 the the DET bracis-33661 207 54 lstm lstm NOUN bracis-33661 207 55 . . PUNCT bracis-33661 208 1 this this PRON bracis-33661 208 2 suggests suggest VERB bracis-33661 208 3 that that SCONJ bracis-33661 208 4 , , PUNCT bracis-33661 208 5 leveraging leverage VERB bracis-33661 208 6 the the DET bracis-33661 208 7 topological topological ADJ bracis-33661 208 8 characteristics characteristic NOUN bracis-33661 208 9 of of ADP bracis-33661 208 10 eeg eeg PROPN bracis-33661 208 11 data datum NOUN bracis-33661 208 12 through through ADP bracis-33661 208 13 graph graph NOUN bracis-33661 208 14 - - PUNCT bracis-33661 208 15 based base VERB bracis-33661 208 16 approaches approach NOUN bracis-33661 208 17 can can AUX bracis-33661 208 18 effectively effectively ADV bracis-33661 208 19 enhance enhance VERB bracis-33661 208 20 the the DET bracis-33661 208 21 performance performance NOUN bracis-33661 208 22 of of ADP bracis-33661 208 23 eeg eeg NOUN bracis-33661 208 24 analysis analysis NOUN bracis-33661 208 25 tasks task NOUN bracis-33661 208 26 . . PUNCT bracis-33661 209 1 looking look VERB bracis-33661 209 2 forward forward ADV bracis-33661 209 3 , , PUNCT bracis-33661 209 4 further further ADJ bracis-33661 209 5 studies study NOUN bracis-33661 209 6 will will AUX bracis-33661 209 7 focus focus VERB bracis-33661 209 8 on on ADP bracis-33661 209 9 developing develop VERB bracis-33661 209 10 novel novel ADJ bracis-33661 209 11 layers layer NOUN bracis-33661 209 12 tailored tailor VERB bracis-33661 209 13 specifically specifically ADV bracis-33661 209 14 for for ADP bracis-33661 209 15 the the DET bracis-33661 209 16 gnn gnn PROPN bracis-33661 209 17 architecture architecture NOUN bracis-33661 209 18 to to PART bracis-33661 209 19 achieve achieve VERB bracis-33661 209 20 even even ADV bracis-33661 209 21 better well ADJ bracis-33661 209 22 performance performance NOUN bracis-33661 209 23 on on ADP bracis-33661 209 24 the the DET bracis-33661 209 25 prognostic prognostic NOUN bracis-33661 209 26 of of ADP bracis-33661 209 27 patients patient NOUN bracis-33661 209 28 in in ADP bracis-33661 209 29 coma coma NOUN bracis-33661 209 30 . . PUNCT bracis-33661 210 1 by by ADP bracis-33661 210 2 customizing customize VERB bracis-33661 210 3 the the DET bracis-33661 210 4 gnn gnn NOUN bracis-33661 210 5 architecture architecture NOUN bracis-33661 210 6 to to PART bracis-33661 210 7 better well ADV bracis-33661 210 8 suit suit VERB bracis-33661 210 9 the the DET bracis-33661 210 10 characteristics characteristic NOUN bracis-33661 210 11 of of ADP bracis-33661 210 12 eeg eeg PROPN bracis-33661 210 13 data datum NOUN bracis-33661 210 14 , , PUNCT bracis-33661 210 15 we we PRON bracis-33661 210 16 can can AUX bracis-33661 210 17 potentially potentially ADV bracis-33661 210 18 uncover uncover VERB bracis-33661 210 19 additional additional ADJ bracis-33661 210 20 improvements improvement NOUN bracis-33661 210 21 in in ADP bracis-33661 210 22 the the DET bracis-33661 210 23 accuracy accuracy NOUN bracis-33661 210 24 and and CCONJ bracis-33661 210 25 effectiveness effectiveness NOUN bracis-33661 210 26 of of ADP bracis-33661 210 27 eeg eeg NOUN bracis-33661 210 28 analysis analysis NOUN bracis-33661 210 29 techniques technique NOUN bracis-33661 210 30 . . PUNCT bracis-33661 211 1 in in ADP bracis-33661 211 2 the the DET bracis-33661 211 3 future future NOUN bracis-33661 211 4 , , PUNCT bracis-33661 211 5 we we PRON bracis-33661 211 6 will will AUX bracis-33661 211 7 refine refine VERB bracis-33661 211 8 the the DET bracis-33661 211 9 modeling modeling NOUN bracis-33661 211 10 strategy strategy NOUN bracis-33661 211 11 to to PART bracis-33661 211 12 represent represent VERB bracis-33661 211 13 eeg eeg NOUN bracis-33661 211 14 data datum NOUN bracis-33661 211 15 into into ADP bracis-33661 211 16 graph graph NOUN bracis-33661 211 17 structures structure NOUN bracis-33661 211 18 and and CCONJ bracis-33661 211 19 extend extend VERB bracis-33661 211 20 its its PRON bracis-33661 211 21 application application NOUN bracis-33661 211 22 to to ADP bracis-33661 211 23 other other ADJ bracis-33661 211 24 public public ADJ bracis-33661 211 25 datasets dataset NOUN bracis-33661 211 26 . . PUNCT bracis-33661 212 1 this this DET bracis-33661 212 2 refinement refinement NOUN bracis-33661 212 3 will will AUX bracis-33661 212 4 focus focus VERB bracis-33661 212 5 on on ADP bracis-33661 212 6 employing employ VERB bracis-33661 212 7 network network NOUN bracis-33661 212 8 structural structural ADJ bracis-33661 212 9 optimization optimization NOUN bracis-33661 212 10 techniques technique NOUN bracis-33661 212 11 to to PART bracis-33661 212 12 enhance enhance VERB bracis-33661 212 13 the the DET bracis-33661 212 14 accuracy accuracy NOUN bracis-33661 212 15 and and CCONJ bracis-33661 212 16 efficiency efficiency NOUN bracis-33661 212 17 of of ADP bracis-33661 212 18 the the DET bracis-33661 212 19 graph graph NOUN bracis-33661 212 20 - - PUNCT bracis-33661 212 21 based base VERB bracis-33661 212 22 models model NOUN bracis-33661 212 23 , , PUNCT bracis-33661 212 24 thereby thereby ADV bracis-33661 212 25 improving improve VERB bracis-33661 212 26 the the DET bracis-33661 212 27 overall overall ADJ bracis-33661 212 28 predictive predictive ADJ bracis-33661 212 29 performance performance NOUN bracis-33661 212 30 and and CCONJ bracis-33661 212 31 applicability applicability NOUN bracis-33661 212 32 in in ADP bracis-33661 212 33 real real ADJ bracis-33661 212 34 - - PUNCT bracis-33661 212 35 world world NOUN bracis-33661 212 36 scenarios scenario NOUN bracis-33661 212 37 . . PUNCT bracis-33661 213 1 references reference NOUN bracis-33661 213 2 alhagry alhagry PROPN bracis-33661 213 3 , , PUNCT bracis-33661 213 4 s. s. PROPN bracis-33661 213 5 , , PUNCT bracis-33661 213 6 fahmy fahmy PROPN bracis-33661 213 7 , , PUNCT bracis-33661 213 8 a.a a.a PROPN bracis-33661 213 9 . . PROPN bracis-33661 213 10 , , PUNCT bracis-33661 213 11 el el PROPN bracis-33661 213 12 - - PUNCT bracis-33661 213 13 khoribi khoribi NOUN bracis-33661 213 14 , , PUNCT bracis-33661 213 15 r.a r.a PROPN bracis-33661 213 16 . . PROPN bracis-33661 213 17 : : PUNCT bracis-33661 214 1 emotion emotion NOUN bracis-33661 214 2 recognition recognition NOUN bracis-33661 214 3 based base VERB bracis-33661 214 4 on on ADP bracis-33661 214 5 eeg eeg NOUN bracis-33661 214 6 using use VERB bracis-33661 214 7 lstm lstm PROPN bracis-33661 214 8 recurrent recurrent ADJ bracis-33661 214 9 neural neural ADJ bracis-33661 214 10 network network NOUN bracis-33661 214 11 . . PUNCT bracis-33661 215 1 int int NOUN bracis-33661 215 2 . . PUNCT bracis-33661 216 1 j. j. PROPN bracis-33661 216 2 adv adv PROPN bracis-33661 216 3 . . PUNCT bracis-33661 217 1 comput comput PROPN bracis-33661 217 2 . . PUNCT bracis-33661 218 1 sci sci PROPN bracis-33661 218 2 . . PUNCT bracis-33661 218 3 appl appl PROPN bracis-33661 218 4 . . PROPN bracis-33661 219 1 8(10 8(10 NUM bracis-33661 219 2 ) ) PUNCT bracis-33661 219 3 ( ( PUNCT bracis-33661 219 4 2017 2017 NUM bracis-33661 219 5 ) ) PUNCT bracis-33661 219 6 google google NOUN bracis-33661 219 7 scholar scholar NOUN bracis-33661 219 8     SPACE bracis-33661 219 9 altıntop altıntop NOUN bracis-33661 219 10 , , PUNCT bracis-33661 219 11 ç.g ç.g PROPN bracis-33661 219 12 . . PROPN bracis-33661 219 13 , , PUNCT bracis-33661 219 14 latifoğlu latifoğlu PROPN bracis-33661 219 15 , , PUNCT bracis-33661 219 16 f. f. PROPN bracis-33661 219 17 , , PUNCT bracis-33661 219 18 akın akın PROPN bracis-33661 219 19 , , PUNCT bracis-33661 219 20 a.k a.k PROPN bracis-33661 219 21 . . PROPN bracis-33661 219 22 , , PUNCT bracis-33661 219 23 cetin cetin PROPN bracis-33661 219 24 , , PUNCT bracis-33661 219 25 b. b. PROPN bracis-33661 219 26 : : PUNCT bracis-33661 219 27 a a DET bracis-33661 219 28 novel novel ADJ bracis-33661 219 29 approach approach NOUN bracis-33661 219 30 for for ADP bracis-33661 219 31 detection detection NOUN bracis-33661 219 32 of of ADP bracis-33661 219 33 consciousness consciousness NOUN bracis-33661 219 34 level level NOUN bracis-33661 219 35 in in ADP bracis-33661 219 36 comatose comatose ADJ bracis-33661 219 37 patients patient NOUN bracis-33661 219 38 from from ADP bracis-33661 219 39 eeg eeg NOUN bracis-33661 219 40 signals signal NOUN bracis-33661 219 41 with with ADP bracis-33661 219 42 1 1 NUM bracis-33661 219 43 - - PUNCT bracis-33661 219 44 d d ADJ bracis-33661 219 45 convolutional convolutional ADJ bracis-33661 219 46 neural neural ADJ bracis-33661 219 47 network network NOUN bracis-33661 219 48 . . PUNCT bracis-33661 220 1 biocybern biocybern NOUN bracis-33661 220 2 . . PROPN bracis-33661 220 3 biomed biome VERB bracis-33661 220 4 . . PUNCT bracis-33661 221 1 eng eng PROPN bracis-33661 221 2 . . PUNCT bracis-33661 221 3 42(1 42(1 NOUN bracis-33661 221 4 ) ) PUNCT bracis-33661 221 5 , , PUNCT bracis-33661 221 6 16–26 16–26 NUM bracis-33661 221 7 ( ( PUNCT bracis-33661 221 8 2022 2022 NUM bracis-33661 221 9 ) ) PUNCT bracis-33661 221 10 article article NOUN bracis-33661 221 11     SPACE bracis-33661 221 12 google google PROPN bracis-33661 221 13 scholar scholar NOUN bracis-33661 221 14     SPACE bracis-33661 221 15 baldo baldo PROPN bracis-33661 221 16 júnior júnior PROPN bracis-33661 221 17 , , PUNCT bracis-33661 221 18 s. s. PROPN bracis-33661 221 19 , , PUNCT bracis-33661 221 20 carneiro carneiro PROPN bracis-33661 221 21 , , PUNCT bracis-33661 221 22 m.g m.g AUX bracis-33661 221 23 . . PROPN bracis-33661 221 24 , , PUNCT bracis-33661 221 25 destro destro NOUN bracis-33661 221 26 - - PUNCT bracis-33661 221 27 filho filho NOUN bracis-33661 221 28 , , PUNCT bracis-33661 221 29 j.b j.b PROPN bracis-33661 221 30 . . PROPN bracis-33661 221 31 , , PUNCT bracis-33661 221 32 zhao zhao PROPN bracis-33661 221 33 , , PUNCT bracis-33661 221 34 l. l. PROPN bracis-33661 221 35 , , PUNCT bracis-33661 221 36 tinós tinós PROPN bracis-33661 221 37 , , PUNCT bracis-33661 221 38 r. r. PROPN bracis-33661 221 39 : : PUNCT bracis-33661 221 40 classification classification NOUN bracis-33661 221 41 of of ADP bracis-33661 221 42 coma coma NOUN bracis-33661 221 43 etiology etiology NOUN bracis-33661 221 44 using use VERB bracis-33661 221 45 convolutional convolutional ADJ bracis-33661 221 46 neural neural ADJ bracis-33661 221 47 networks network NOUN bracis-33661 221 48 and and CCONJ bracis-33661 221 49 long long ADJ bracis-33661 221 50 - - PUNCT bracis-33661 221 51 short short ADJ bracis-33661 221 52 term term NOUN bracis-33661 221 53 memory memory NOUN bracis-33661 221 54 networks network NOUN bracis-33661 221 55 . . PUNCT bracis-33661 222 1 in in ADP bracis-33661 222 2 : : PUNCT bracis-33661 222 3 2023 2023 NUM bracis-33661 222 4 international international ADJ bracis-33661 222 5 joint joint ADJ bracis-33661 222 6 conference conference NOUN bracis-33661 222 7 on on ADP bracis-33661 222 8 neural neural ADJ bracis-33661 222 9 networks network NOUN bracis-33661 222 10 ( ( PUNCT bracis-33661 222 11 ijcnn ijcnn PROPN bracis-33661 222 12 ) ) PUNCT bracis-33661 222 13 , , PUNCT bracis-33661 222 14 pp pp ADP bracis-33661 222 15 . . PUNCT bracis-33661 222 16     SPACE bracis-33661 223 1 1–8 1–8 X bracis-33661 223 2 . . PUNCT bracis-33661 223 3 ieee ieee NOUN bracis-33661 223 4 ( ( PUNCT bracis-33661 223 5 2023 2023 NUM bracis-33661 223 6 ) ) PUNCT bracis-33661 223 7 google google NOUN bracis-33661 223 8 scholar scholar NOUN bracis-33661 223 9     SPACE bracis-33661 223 10 bissaro bissaro NOUN bracis-33661 223 11 , , PUNCT bracis-33661 223 12 l.z l.z PROPN bracis-33661 223 13 . . PROPN bracis-33661 223 14 , , PUNCT bracis-33661 223 15 junior junior ADJ bracis-33661 223 16 , , PUNCT bracis-33661 223 17 o.o.n o.o.n ADJ bracis-33661 223 18 . . PROPN bracis-33661 223 19 , , PUNCT bracis-33661 223 20 filho filho PROPN bracis-33661 223 21 , , PUNCT bracis-33661 223 22 j.b.d j.b.d PROPN bracis-33661 223 23 . . PROPN bracis-33661 223 24 , , PUNCT bracis-33661 223 25 jin jin PROPN bracis-33661 223 26 , , PUNCT bracis-33661 223 27 y. y. PROPN bracis-33661 223 28 , , PUNCT bracis-33661 223 29 carneiro carneiro PROPN bracis-33661 223 30 , , PUNCT bracis-33661 223 31 m.g m.g PROPN bracis-33661 223 32 . . PROPN bracis-33661 223 33 : : PUNCT bracis-33661 223 34 towards towards ADP bracis-33661 223 35 the the DET bracis-33661 223 36 prognosis prognosis NOUN bracis-33661 223 37 of of ADP bracis-33661 223 38 patients patient NOUN bracis-33661 223 39 in in ADP bracis-33661 223 40 coma coma NOUN bracis-33661 223 41 using use VERB bracis-33661 223 42 echo echo NOUN bracis-33661 223 43 state state NOUN bracis-33661 223 44 networks network NOUN bracis-33661 223 45 for for ADP bracis-33661 223 46 eeg eeg NOUN bracis-33661 223 47 analysis analysis NOUN bracis-33661 223 48 . . PUNCT bracis-33661 224 1 procedia procedia NOUN bracis-33661 224 2 comput comput NOUN bracis-33661 224 3 . . PUNCT bracis-33661 225 1 sci sci PROPN bracis-33661 225 2 . . PROPN bracis-33661 226 1 222 222 NUM bracis-33661 226 2 , , PUNCT bracis-33661 227 1 509–518 509–518 NUM bracis-33661 227 2 ( ( PUNCT bracis-33661 227 3 2023 2023 NUM bracis-33661 227 4 ) ) PUNCT bracis-33661 227 5 article article NOUN bracis-33661 227 6     SPACE bracis-33661 227 7 google google PROPN bracis-33661 227 8 scholar scholar NOUN bracis-33661 227 9     SPACE bracis-33661 227 10 carneiro carneiro PROPN bracis-33661 227 11 , , PUNCT bracis-33661 227 12 m.g m.g PROPN bracis-33661 227 13 . . PROPN bracis-33661 227 14 , , PUNCT bracis-33661 227 15 ramos ramos PROPN bracis-33661 227 16 , , PUNCT bracis-33661 227 17 c.d c.d PROPN bracis-33661 227 18 . . PROPN bracis-33661 227 19 , , PUNCT bracis-33661 227 20 destro destro NOUN bracis-33661 227 21 - - PUNCT bracis-33661 227 22 filho filho NOUN bracis-33661 227 23 , , PUNCT bracis-33661 227 24 j.b j.b PROPN bracis-33661 227 25 . . PROPN bracis-33661 227 26 , , PUNCT bracis-33661 227 27 zhu zhu PROPN bracis-33661 227 28 , , PUNCT bracis-33661 227 29 y.t y.t PROPN bracis-33661 227 30 . . PROPN bracis-33661 227 31 , , PUNCT bracis-33661 227 32 ji ji PROPN bracis-33661 227 33 , , PUNCT bracis-33661 227 34 d. d. PROPN bracis-33661 227 35 , , PUNCT bracis-33661 227 36 zhao zhao PROPN bracis-33661 227 37 , , PUNCT bracis-33661 227 38 l. l. PROPN bracis-33661 227 39 : : PUNCT bracis-33661 227 40 high high ADJ bracis-33661 227 41 - - PUNCT bracis-33661 227 42 level level NOUN bracis-33661 227 43 classification classification NOUN bracis-33661 227 44 for for ADP bracis-33661 227 45 eeg eeg NOUN bracis-33661 227 46 analysis analysis NOUN bracis-33661 227 47 . . PUNCT bracis-33661 228 1 in in ADP bracis-33661 228 2 : : PUNCT bracis-33661 228 3 2023 2023 NUM bracis-33661 228 4 international international ADJ bracis-33661 228 5 joint joint ADJ bracis-33661 228 6 conference conference NOUN bracis-33661 228 7 on on ADP bracis-33661 228 8 neural neural ADJ bracis-33661 228 9 networks network NOUN bracis-33661 228 10 ( ( PUNCT bracis-33661 228 11 ijcnn ijcnn PROPN bracis-33661 228 12 ) ) PUNCT bracis-33661 228 13 , , PUNCT bracis-33661 228 14 pp pp ADP bracis-33661 228 15 . . PUNCT bracis-33661 228 16     SPACE bracis-33661 229 1 1–8 1–8 X bracis-33661 229 2 . . PUNCT bracis-33661 229 3 ieee ieee NOUN bracis-33661 229 4 ( ( PUNCT bracis-33661 229 5 2023 2023 NUM bracis-33661 229 6 ) ) PUNCT bracis-33661 229 7 google google NOUN bracis-33661 229 8 scholar scholar NOUN bracis-33661 229 9     SPACE bracis-33661 229 10 chen chen PROPN bracis-33661 229 11 , , PUNCT bracis-33661 229 12 m. m. NOUN bracis-33661 229 13 , , PUNCT bracis-33661 229 14 wei wei PROPN bracis-33661 229 15 , , PUNCT bracis-33661 229 16 z. z. PROPN bracis-33661 229 17 , , PUNCT bracis-33661 229 18 huang huang PROPN bracis-33661 229 19 , , PUNCT bracis-33661 229 20 z. z. PROPN bracis-33661 229 21 , , PUNCT bracis-33661 229 22 ding ding PROPN bracis-33661 229 23 , , PUNCT bracis-33661 229 24 b. b. PROPN bracis-33661 229 25 , , PUNCT bracis-33661 229 26 li li PROPN bracis-33661 229 27 , , PUNCT bracis-33661 229 28 y. y. NOUN bracis-33661 229 29 : : PUNCT bracis-33661 229 30 simple simple ADJ bracis-33661 229 31 and and CCONJ bracis-33661 229 32 deep deep ADJ bracis-33661 229 33 graph graph NOUN bracis-33661 229 34 convolutional convolutional ADJ bracis-33661 229 35 networks network NOUN bracis-33661 229 36 . . PUNCT bracis-33661 230 1 in in ADP bracis-33661 230 2 : : PUNCT bracis-33661 230 3 international international ADJ bracis-33661 230 4 conference conference NOUN bracis-33661 230 5 on on ADP bracis-33661 230 6 machine machine NOUN bracis-33661 230 7 learning learning NOUN bracis-33661 230 8 , , PUNCT bracis-33661 230 9 pp pp ADJ bracis-33661 230 10 . . PUNCT bracis-33661 230 11 1725–1735 1725–1735 NUM bracis-33661 230 12 . . PUNCT bracis-33661 230 13 pmlr pmlr NOUN bracis-33661 230 14 ( ( PUNCT bracis-33661 230 15 2020 2020 NUM bracis-33661 230 16 ) ) PUNCT bracis-33661 230 17 google google PROPN bracis-33661 230 18 scholar scholar NOUN bracis-33661 230 19     SPACE bracis-33661 230 20 davis davis PROPN bracis-33661 230 21 , , PUNCT bracis-33661 230 22 a.e a.e PROPN bracis-33661 230 23 . . PROPN bracis-33661 230 24 , , PUNCT bracis-33661 230 25 gimenez gimenez PROPN bracis-33661 230 26 , , PUNCT bracis-33661 230 27 a. a. NOUN bracis-33661 230 28 : : PUNCT bracis-33661 230 29 cognitive cognitive ADJ bracis-33661 230 30 - - PUNCT bracis-33661 230 31 behavioral behavioral ADJ bracis-33661 230 32 recovery recovery NOUN bracis-33661 230 33 in in ADP bracis-33661 230 34 comatose comatose ADJ bracis-33661 230 35 patients patient NOUN bracis-33661 230 36 following follow VERB bracis-33661 230 37 auditory auditory ADJ bracis-33661 230 38 sensory sensory ADJ bracis-33661 230 39 stimulation stimulation NOUN bracis-33661 230 40 . . PUNCT bracis-33661 231 1 j. j. PROPN bracis-33661 231 2 neurosci neurosci PROPN bracis-33661 231 3 . . PUNCT bracis-33661 231 4 nurs nurs PROPN bracis-33661 231 5 . . PUNCT bracis-33661 232 1 35(4 35(4 NUM bracis-33661 232 2 ) ) PUNCT bracis-33661 232 3 , , PUNCT bracis-33661 232 4 202–209 202–209 NUM bracis-33661 232 5 ( ( PUNCT bracis-33661 232 6 2003 2003 NUM bracis-33661 232 7 ) ) PUNCT bracis-33661 232 8 article article NOUN bracis-33661 232 9     SPACE bracis-33661 232 10 google google PROPN bracis-33661 232 11 scholar scholar NOUN bracis-33661 232 12     SPACE bracis-33661 232 13 demir demir PROPN bracis-33661 232 14 , , PUNCT bracis-33661 232 15 a. a. NOUN bracis-33661 232 16 , , PUNCT bracis-33661 232 17 koike koike NOUN bracis-33661 232 18 - - PUNCT bracis-33661 232 19 akino akino NOUN bracis-33661 232 20 , , PUNCT bracis-33661 232 21 t. t. PROPN bracis-33661 232 22 , , PUNCT bracis-33661 232 23 wang wang PROPN bracis-33661 232 24 , , PUNCT bracis-33661 232 25 y. y. PROPN bracis-33661 232 26 , , PUNCT bracis-33661 232 27 haruna haruna PROPN bracis-33661 232 28 , , PUNCT bracis-33661 232 29 m. m. NOUN bracis-33661 232 30 , , PUNCT bracis-33661 232 31 erdogmus erdogmus NOUN bracis-33661 232 32 , , PUNCT bracis-33661 232 33 d. d. PROPN bracis-33661 232 34 : : PUNCT bracis-33661 232 35 eeg eeg PROPN bracis-33661 232 36 - - PUNCT bracis-33661 232 37 gnn gnn NOUN bracis-33661 232 38 : : PUNCT bracis-33661 232 39 graph graph NOUN bracis-33661 232 40 neural neural ADJ bracis-33661 232 41 networks network NOUN bracis-33661 232 42 for for ADP bracis-33661 232 43 classification classification NOUN bracis-33661 232 44 of of ADP bracis-33661 232 45 electroencephalogram electroencephalogram NOUN bracis-33661 232 46 ( ( PUNCT bracis-33661 232 47 eeg eeg NOUN bracis-33661 232 48 ) ) PUNCT bracis-33661 232 49 signals signal NOUN bracis-33661 232 50 . . PUNCT bracis-33661 233 1 in in ADP bracis-33661 233 2 : : PUNCT bracis-33661 233 3 2021 2021 NUM bracis-33661 233 4 43rd 43rd ADJ bracis-33661 233 5 annual annual ADJ bracis-33661 233 6 international international ADJ bracis-33661 233 7 conference conference NOUN bracis-33661 233 8 of of ADP bracis-33661 233 9 the the DET bracis-33661 233 10 ieee ieee NOUN bracis-33661 233 11 engineering engineering NOUN bracis-33661 233 12 in in ADP bracis-33661 233 13 medicine medicine NOUN bracis-33661 233 14 and and CCONJ bracis-33661 233 15 biology biology NOUN bracis-33661 233 16 society society NOUN bracis-33661 233 17 ( ( PUNCT bracis-33661 233 18 embc embc PROPN bracis-33661 233 19 ) ) PUNCT bracis-33661 233 20 , , PUNCT bracis-33661 233 21 pp pp ADJ bracis-33661 233 22 . . PUNCT bracis-33661 233 23 1061–1067 1061–1067 NUM bracis-33661 233 24 . . X bracis-33661 233 25 ieee ieee NOUN bracis-33661 233 26 ( ( PUNCT bracis-33661 233 27 2021 2021 NUM bracis-33661 233 28 ) ) PUNCT bracis-33661 233 29 google google NOUN bracis-33661 233 30 scholar scholar NOUN bracis-33661 233 31     SPACE bracis-33661 233 32 di di X bracis-33661 233 33 perri perri PROPN bracis-33661 233 34 , , PUNCT bracis-33661 233 35 c. c. PROPN bracis-33661 233 36 , , PUNCT bracis-33661 233 37 thibaut thibaut PROPN bracis-33661 233 38 , , PUNCT bracis-33661 233 39 a. a. PROPN bracis-33661 233 40 , , PUNCT bracis-33661 233 41 heine heine PROPN bracis-33661 233 42 , , PUNCT bracis-33661 233 43 l. l. PROPN bracis-33661 233 44 , , PUNCT bracis-33661 233 45 soddu soddu NOUN bracis-33661 233 46 , , PUNCT bracis-33661 233 47 a. a. NOUN bracis-33661 233 48 , , PUNCT bracis-33661 233 49 demertzi demertzi NOUN bracis-33661 233 50 , , PUNCT bracis-33661 233 51 a. a. NOUN bracis-33661 233 52 , , PUNCT bracis-33661 233 53 laureys laurey NOUN bracis-33661 233 54 , , PUNCT bracis-33661 233 55 s. s. PROPN bracis-33661 233 56 : : PUNCT bracis-33661 233 57 measuring measure VERB bracis-33661 233 58 consciousness consciousness NOUN bracis-33661 233 59 in in ADP bracis-33661 233 60 coma coma NOUN bracis-33661 233 61 and and CCONJ bracis-33661 233 62 related related ADJ bracis-33661 233 63 states state NOUN bracis-33661 233 64 . . PUNCT bracis-33661 234 1 world world PROPN bracis-33661 234 2 j. j. PROPN bracis-33661 234 3 radiol radiol PROPN bracis-33661 234 4 . . PUNCT bracis-33661 235 1 6(8 6(8 NUM bracis-33661 235 2 ) ) PUNCT bracis-33661 235 3 , , PUNCT bracis-33661 235 4 589 589 NUM bracis-33661 235 5 ( ( PUNCT bracis-33661 235 6 2014 2014 NUM bracis-33661 235 7 ) ) PUNCT bracis-33661 235 8 article article NOUN bracis-33661 235 9     SPACE bracis-33661 235 10 google google PROPN bracis-33661 235 11 scholar scholar NOUN bracis-33661 235 12     SPACE bracis-33661 235 13 gerrard gerrard NOUN bracis-33661 235 14 , , PUNCT bracis-33661 235 15 p. p. NOUN bracis-33661 235 16 , , PUNCT bracis-33661 235 17 zafonte zafonte PROPN bracis-33661 235 18 , , PUNCT bracis-33661 235 19 r. r. PROPN bracis-33661 235 20 , , PUNCT bracis-33661 235 21 giacino giacino PROPN bracis-33661 235 22 , , PUNCT bracis-33661 235 23 j.t j.t PROPN bracis-33661 235 24 . . PROPN bracis-33661 235 25 : : PUNCT bracis-33661 236 1 coma coma NOUN bracis-33661 236 2 recovery recovery NOUN bracis-33661 236 3 scale scale NOUN bracis-33661 236 4 - - PUNCT bracis-33661 236 5 revised revise VERB bracis-33661 236 6 : : PUNCT bracis-33661 236 7 evidentiary evidentiary ADJ bracis-33661 236 8 support support NOUN bracis-33661 236 9 for for ADP bracis-33661 236 10 hierarchical hierarchical ADJ bracis-33661 236 11 grading grading NOUN bracis-33661 236 12 of of ADP bracis-33661 236 13 level level NOUN bracis-33661 236 14 of of ADP bracis-33661 236 15 consciousness consciousness NOUN bracis-33661 236 16 . . PUNCT bracis-33661 237 1 arch arch NOUN bracis-33661 237 2 . . PUNCT bracis-33661 238 1 phys phy NOUN bracis-33661 238 2 . . PUNCT bracis-33661 239 1 med med PROPN bracis-33661 239 2 . . PUNCT bracis-33661 239 3 rehabil rehabil PROPN bracis-33661 239 4 . . PUNCT bracis-33661 240 1 95(12 95(12 NUM bracis-33661 240 2 ) ) PUNCT bracis-33661 240 3 , , PUNCT bracis-33661 240 4 2335–2341 2335–2341 NUM bracis-33661 240 5 ( ( PUNCT bracis-33661 240 6 2014 2014 NUM bracis-33661 240 7 ) ) PUNCT bracis-33661 240 8 article article NOUN bracis-33661 240 9     SPACE bracis-33661 240 10 google google PROPN bracis-33661 240 11 scholar scholar NOUN bracis-33661 240 12     SPACE bracis-33661 240 13 graves grave NOUN bracis-33661 240 14 , , PUNCT bracis-33661 240 15 a. a. NOUN bracis-33661 240 16 : : PUNCT bracis-33661 240 17 generating generate VERB bracis-33661 240 18 sequences sequence NOUN bracis-33661 240 19 with with ADP bracis-33661 240 20 recurrent recurrent ADJ bracis-33661 240 21 neural neural ADJ bracis-33661 240 22 networks network NOUN bracis-33661 240 23 . . PUNCT bracis-33661 241 1 arxiv arxiv PROPN bracis-33661 241 2 preprint preprint VERB bracis-33661 241 3 arxiv:1308.0850 arxiv:1308.0850 NOUN bracis-33661 241 4 ( ( PUNCT bracis-33661 241 5 2013 2013 NUM bracis-33661 241 6 ) ) PUNCT bracis-33661 241 7 hamilton hamilton PROPN bracis-33661 241 8 , , PUNCT bracis-33661 241 9 w. w. PROPN bracis-33661 241 10 , , PUNCT bracis-33661 241 11 ying ying PROPN bracis-33661 241 12 , , PUNCT bracis-33661 241 13 z. z. PROPN bracis-33661 241 14 , , PUNCT bracis-33661 241 15 leskovec leskovec PROPN bracis-33661 241 16 , , PUNCT bracis-33661 241 17 j. j. PROPN bracis-33661 241 18 : : PUNCT bracis-33661 241 19 inductive inductive ADJ bracis-33661 241 20 representation representation NOUN bracis-33661 241 21 learning learn VERB bracis-33661 241 22 on on ADP bracis-33661 241 23 large large ADJ bracis-33661 241 24 graphs graph NOUN bracis-33661 241 25 . . PUNCT bracis-33661 242 1 adv adv PROPN bracis-33661 242 2 . . PUNCT bracis-33661 242 3 neural neural ADJ bracis-33661 242 4 info info NOUN bracis-33661 242 5 . . PUNCT bracis-33661 242 6 process process NOUN bracis-33661 242 7 . . PUNCT bracis-33661 243 1 syst syst PROPN bracis-33661 243 2 . . PUNCT bracis-33661 244 1 30 30 NUM bracis-33661 244 2 ( ( PUNCT bracis-33661 244 3 2017 2017 NUM bracis-33661 244 4 ) ) PUNCT bracis-33661 244 5 google google NOUN bracis-33661 244 6 scholar scholar NOUN bracis-33661 244 7     SPACE bracis-33661 244 8 hochreiter hochreiter PROPN bracis-33661 244 9 , , PUNCT bracis-33661 244 10 s. s. PROPN bracis-33661 244 11 , , PUNCT bracis-33661 244 12 schmidhuber schmidhuber PROPN bracis-33661 244 13 , , PUNCT bracis-33661 244 14 j. j. PROPN bracis-33661 244 15 : : PUNCT bracis-33661 244 16 long long ADJ bracis-33661 244 17 short short ADJ bracis-33661 244 18 - - PUNCT bracis-33661 244 19 term term NOUN bracis-33661 244 20 memory memory NOUN bracis-33661 244 21 . . PUNCT bracis-33661 245 1 neural neural ADJ bracis-33661 245 2 comput comput NOUN bracis-33661 245 3 . . PUNCT bracis-33661 246 1 9(8 9(8 NUM bracis-33661 246 2 ) ) PUNCT bracis-33661 246 3 , , PUNCT bracis-33661 246 4 1735–1780 1735–1780 NUM bracis-33661 246 5 ( ( PUNCT bracis-33661 246 6 1997 1997 NUM bracis-33661 246 7 ) ) PUNCT bracis-33661 246 8 article article NOUN bracis-33661 246 9     SPACE bracis-33661 246 10 math math NOUN bracis-33661 246 11     SPACE bracis-33661 246 12 google google PROPN bracis-33661 246 13 scholar scholar NOUN bracis-33661 246 14     SPACE bracis-33661 246 15 huang huang PROPN bracis-33661 246 16 , , PUNCT bracis-33661 246 17 y. y. PROPN bracis-33661 246 18 , , PUNCT bracis-33661 246 19 yang yang PROPN bracis-33661 246 20 , , PUNCT bracis-33661 246 21 j. j. PROPN bracis-33661 246 22 , , PUNCT bracis-33661 246 23 liao liao PROPN bracis-33661 246 24 , , PUNCT bracis-33661 246 25 p. p. PROPN bracis-33661 246 26 , , PUNCT bracis-33661 246 27 pan pan PROPN bracis-33661 246 28 , , PUNCT bracis-33661 246 29 j. j. PROPN bracis-33661 246 30 : : PUNCT bracis-33661 246 31 fusion fusion NOUN bracis-33661 246 32 of of ADP bracis-33661 246 33 facial facial ADJ bracis-33661 246 34 expressions expression NOUN bracis-33661 246 35 and and CCONJ bracis-33661 246 36 eeg eeg NOUN bracis-33661 246 37 for for ADP bracis-33661 246 38 multimodal multimodal NOUN bracis-33661 246 39 emotion emotion NOUN bracis-33661 246 40 recognition recognition NOUN bracis-33661 246 41 . . PUNCT bracis-33661 247 1 comput comput NOUN bracis-33661 247 2 . . PUNCT bracis-33661 248 1 intell intell PROPN bracis-33661 248 2 . . PUNCT bracis-33661 249 1 neurosci neurosci PROPN bracis-33661 249 2 . . PUNCT bracis-33661 250 1 2017(1 2017(1 NUM bracis-33661 250 2 ) ) PUNCT bracis-33661 250 3 , , PUNCT bracis-33661 250 4 2107451 2107451 NUM bracis-33661 250 5 ( ( PUNCT bracis-33661 250 6 2017 2017 NUM bracis-33661 250 7 ) ) PUNCT bracis-33661 250 8 google google NOUN bracis-33661 250 9 scholar scholar NOUN bracis-33661 250 10     SPACE bracis-33661 250 11 iwaki iwaki NOUN bracis-33661 250 12 , , PUNCT bracis-33661 250 13 t. t. PROPN bracis-33661 250 14 , , PUNCT bracis-33661 250 15 hayashi hayashi PROPN bracis-33661 250 16 , , PUNCT bracis-33661 250 17 m. m. NOUN bracis-33661 250 18 , , PUNCT bracis-33661 250 19 hori hori PROPN bracis-33661 250 20 , , PUNCT bracis-33661 250 21 t. t. PROPN bracis-33661 250 22 : : PUNCT bracis-33661 250 23 changes change NOUN bracis-33661 250 24 in in ADP bracis-33661 250 25 alpha alpha ADJ bracis-33661 250 26 band band NOUN bracis-33661 250 27 eeg eeg NOUN bracis-33661 250 28 activity activity NOUN bracis-33661 250 29 in in ADP bracis-33661 250 30 the the DET bracis-33661 250 31 frontal frontal ADJ bracis-33661 250 32 area area NOUN bracis-33661 250 33 after after ADP bracis-33661 250 34 stimulation stimulation NOUN bracis-33661 250 35 with with ADP bracis-33661 250 36 music music NOUN bracis-33661 250 37 of of ADP bracis-33661 250 38 different different ADJ bracis-33661 250 39 affective affective ADJ bracis-33661 250 40 content content NOUN bracis-33661 250 41 . . PUNCT bracis-33661 251 1 percept percept PROPN bracis-33661 251 2 . . PUNCT bracis-33661 252 1 mot mot PROPN bracis-33661 252 2 . . PUNCT bracis-33661 253 1 skills skill NOUN bracis-33661 253 2 84(2 84(2 NUM bracis-33661 253 3 ) ) PUNCT bracis-33661 253 4 , , PUNCT bracis-33661 253 5 515–526 515–526 PROPN bracis-33661 253 6 ( ( PUNCT bracis-33661 253 7 1997 1997 NUM bracis-33661 253 8 ) ) PUNCT bracis-33661 253 9 article article NOUN bracis-33661 253 10     SPACE bracis-33661 253 11 math math NOUN bracis-33661 253 12     SPACE bracis-33661 253 13 google google PROPN bracis-33661 253 14 scholar scholar NOUN bracis-33661 253 15     SPACE bracis-33661 253 16 kipf kipf PROPN bracis-33661 253 17 , , PUNCT bracis-33661 253 18 t.n t.n PROPN bracis-33661 253 19 . . PROPN bracis-33661 253 20 , , PUNCT bracis-33661 253 21 welling well VERB bracis-33661 253 22 , , PUNCT bracis-33661 253 23 m. m. NOUN bracis-33661 253 24 : : PUNCT bracis-33661 253 25 semi semi ADJ bracis-33661 253 26 - - ADJ bracis-33661 253 27 supervised supervised ADJ bracis-33661 253 28 classification classification NOUN bracis-33661 253 29 with with ADP bracis-33661 253 30 graph graph NOUN bracis-33661 253 31 convolutional convolutional ADJ bracis-33661 253 32 networks network NOUN bracis-33661 253 33 . . PUNCT bracis-33661 254 1 arxiv arxiv PROPN bracis-33661 254 2 preprint preprint VERB bracis-33661 254 3 arxiv:1609.02907 arxiv:1609.02907 NOUN bracis-33661 254 4 ( ( PUNCT bracis-33661 254 5 2016 2016 NUM bracis-33661 254 6 ) ) PUNCT bracis-33661 254 7 kumari kumari PROPN bracis-33661 254 8 , , PUNCT bracis-33661 254 9 r.s.s r.s.s ADJ bracis-33661 254 10 . . PROPN bracis-33661 254 11 , , PUNCT bracis-33661 254 12 jose jose PROPN bracis-33661 254 13 , , PUNCT bracis-33661 254 14 j.p j.p PROPN bracis-33661 254 15 . . PROPN bracis-33661 254 16 : : PUNCT bracis-33661 255 1 seizure seizure VERB bracis-33661 255 2 detection detection NOUN bracis-33661 255 3 in in ADP bracis-33661 255 4 eeg eeg NOUN bracis-33661 255 5 using use VERB bracis-33661 255 6 time time NOUN bracis-33661 255 7 frequency frequency NOUN bracis-33661 255 8 analysis analysis NOUN bracis-33661 255 9 and and CCONJ bracis-33661 255 10 svm svm NOUN bracis-33661 255 11 . . PUNCT bracis-33661 256 1 in in ADP bracis-33661 256 2 : : PUNCT bracis-33661 256 3 2011 2011 NUM bracis-33661 256 4 international international ADJ bracis-33661 256 5 conference conference NOUN bracis-33661 256 6 on on ADP bracis-33661 256 7 emerging emerge VERB bracis-33661 256 8 trends trend NOUN bracis-33661 256 9 in in ADP bracis-33661 256 10 electrical electrical ADJ bracis-33661 256 11 and and CCONJ bracis-33661 256 12 computer computer NOUN bracis-33661 256 13 technology technology NOUN bracis-33661 256 14 , , PUNCT bracis-33661 256 15 pp pp ADJ bracis-33661 256 16 . . PUNCT bracis-33661 257 1 626–630 626–630 NOUN bracis-33661 257 2 . . PUNCT bracis-33661 258 1 ieee ieee PROPN bracis-33661 258 2 ( ( PUNCT bracis-33661 258 3 2011 2011 NUM bracis-33661 258 4 ) ) PUNCT bracis-33661 258 5 google google PROPN bracis-33661 258 6 scholar scholar NOUN bracis-33661 258 7     SPACE bracis-33661 258 8 kustermann kustermann PROPN bracis-33661 258 9 , , PUNCT bracis-33661 258 10 t. t. PROPN bracis-33661 258 11 , , PUNCT bracis-33661 258 12 et et PROPN bracis-33661 258 13 al al PROPN bracis-33661 258 14 . . PROPN bracis-33661 258 15 : : PUNCT bracis-33661 259 1 electroencephalography electroencephalography NOUN bracis-33661 259 2 - - PUNCT bracis-33661 259 3 based base VERB bracis-33661 259 4 power power NOUN bracis-33661 259 5 spectra spectra PROPN bracis-33661 259 6 allow allow VERB bracis-33661 259 7 coma coma NOUN bracis-33661 259 8 outcome outcome NOUN bracis-33661 259 9 prediction prediction NOUN bracis-33661 259 10 within within ADP bracis-33661 259 11 24 24 NUM bracis-33661 259 12 h h NOUN bracis-33661 259 13 of of ADP bracis-33661 259 14 cardiac cardiac ADJ bracis-33661 259 15 arrest arrest NOUN bracis-33661 259 16 . . PUNCT bracis-33661 260 1 resuscitation resuscitation NOUN bracis-33661 260 2 142 142 NUM bracis-33661 260 3 , , PUNCT bracis-33661 260 4 162–167 162–167 NUM bracis-33661 260 5 ( ( PUNCT bracis-33661 260 6 2019 2019 NUM bracis-33661 260 7 ) ) PUNCT bracis-33661 260 8 article article NOUN bracis-33661 260 9     SPACE bracis-33661 260 10 math math NOUN bracis-33661 260 11     SPACE bracis-33661 260 12 google google PROPN bracis-33661 260 13 scholar scholar NOUN bracis-33661 260 14     SPACE bracis-33661 260 15 lehembre lehembre NOUN bracis-33661 260 16 , , PUNCT bracis-33661 260 17 r. r. PROPN bracis-33661 260 18 , , PUNCT bracis-33661 260 19 et et PROPN bracis-33661 260 20 al al PROPN bracis-33661 260 21 . . PROPN bracis-33661 260 22 : : PUNCT bracis-33661 260 23 resting rest VERB bracis-33661 260 24 - - PUNCT bracis-33661 260 25 state state NOUN bracis-33661 260 26 eeg eeg NOUN bracis-33661 260 27 study study NOUN bracis-33661 260 28 of of ADP bracis-33661 260 29 comatose comatose ADJ bracis-33661 260 30 patients patient NOUN bracis-33661 260 31 : : PUNCT bracis-33661 260 32 a a DET bracis-33661 260 33 connectivity connectivity NOUN bracis-33661 260 34 and and CCONJ bracis-33661 260 35 frequency frequency NOUN bracis-33661 260 36 analysis analysis NOUN bracis-33661 260 37 to to PART bracis-33661 260 38 find find VERB bracis-33661 260 39 differences difference NOUN bracis-33661 260 40 between between ADP bracis-33661 260 41 vegetative vegetative ADJ bracis-33661 260 42 and and CCONJ bracis-33661 260 43 minimally minimally ADV bracis-33661 260 44 conscious conscious ADJ bracis-33661 260 45 states state NOUN bracis-33661 260 46 . . PUNCT bracis-33661 261 1 funct funct ADJ bracis-33661 261 2 . . PUNCT bracis-33661 262 1 neurol neurol NOUN bracis-33661 262 2 . . PUNCT bracis-33661 263 1 27(1 27(1 NUM bracis-33661 263 2 ) ) PUNCT bracis-33661 263 3 , , PUNCT bracis-33661 263 4 41 41 NUM bracis-33661 263 5 ( ( PUNCT bracis-33661 263 6 2012 2012 NUM bracis-33661 263 7 ) ) PUNCT bracis-33661 263 8 google google PROPN bracis-33661 263 9 scholar scholar NOUN bracis-33661 263 10     SPACE bracis-33661 263 11 liu liu PROPN bracis-33661 263 12 , , PUNCT bracis-33661 263 13 j. j. PROPN bracis-33661 263 14 , , PUNCT bracis-33661 263 15 meng meng PROPN bracis-33661 263 16 , , PUNCT bracis-33661 263 17 h. h. PROPN bracis-33661 263 18 , , PUNCT bracis-33661 263 19 nandi nandi PROPN bracis-33661 263 20 , , PUNCT bracis-33661 263 21 a. a. PROPN bracis-33661 263 22 , , PUNCT bracis-33661 263 23 li li PROPN bracis-33661 263 24 , , PUNCT bracis-33661 263 25 m. m. NOUN bracis-33661 263 26 : : PUNCT bracis-33661 263 27 emotion emotion NOUN bracis-33661 263 28 detection detection NOUN bracis-33661 263 29 from from ADP bracis-33661 263 30 eeg eeg NOUN bracis-33661 263 31 recordings recording NOUN bracis-33661 263 32 . . PUNCT bracis-33661 264 1 in in ADP bracis-33661 264 2 : : PUNCT bracis-33661 264 3 2016 2016 NUM bracis-33661 264 4 12th 12th ADJ bracis-33661 264 5 international international ADJ bracis-33661 264 6 conference conference NOUN bracis-33661 264 7 on on ADP bracis-33661 264 8 natural natural ADJ bracis-33661 264 9 computation computation NOUN bracis-33661 264 10 , , PUNCT bracis-33661 264 11 fuzzy fuzzy ADJ bracis-33661 264 12 systems system NOUN bracis-33661 264 13 and and CCONJ bracis-33661 264 14 knowledge knowledge NOUN bracis-33661 264 15 discovery discovery PROPN bracis-33661 264 16 ( ( PUNCT bracis-33661 264 17 icnc icnc ADJ bracis-33661 264 18 - - PUNCT bracis-33661 264 19 fskd fskd NOUN bracis-33661 264 20 ) ) PUNCT bracis-33661 264 21 , , PUNCT bracis-33661 264 22 pp pp ADP bracis-33661 264 23 . . PUNCT bracis-33661 265 1 1722–1727 1722–1727 NUM bracis-33661 265 2 . . PUNCT bracis-33661 266 1 ieee ieee NOUN bracis-33661 266 2 ( ( PUNCT bracis-33661 266 3 2016 2016 NUM bracis-33661 266 4 ) ) PUNCT bracis-33661 266 5 google google PROPN bracis-33661 266 6 scholar scholar NOUN bracis-33661 266 7     SPACE bracis-33661 266 8 ma ma PROPN bracis-33661 266 9 , , PUNCT bracis-33661 266 10 s. s. PROPN bracis-33661 266 11 , , PUNCT bracis-33661 266 12 cui cui PROPN bracis-33661 266 13 , , PUNCT bracis-33661 266 14 j. j. PROPN bracis-33661 266 15 , , PUNCT bracis-33661 266 16 chen chen PROPN bracis-33661 266 17 , , PUNCT bracis-33661 266 18 c.l c.l PROPN bracis-33661 266 19 . . PROPN bracis-33661 266 20 , , PUNCT bracis-33661 266 21 xiao xiao PROPN bracis-33661 266 22 , , PUNCT bracis-33661 266 23 w. w. PROPN bracis-33661 266 24 , , PUNCT bracis-33661 266 25 liu liu PROPN bracis-33661 266 26 , , PUNCT bracis-33661 266 27 l. l. PROPN bracis-33661 266 28 : : PUNCT bracis-33661 266 29 an an DET bracis-33661 266 30 improved improved ADJ bracis-33661 266 31 bi bi NOUN bracis-33661 266 32 - - ADJ bracis-33661 266 33 lstm lstm ADJ bracis-33661 266 34 eeg eeg PROPN bracis-33661 266 35 emotion emotion NOUN bracis-33661 266 36 recognition recognition NOUN bracis-33661 266 37 algorithm algorithm NOUN bracis-33661 266 38 . . PUNCT bracis-33661 267 1 j. j. PROPN bracis-33661 267 2 netw netw PROPN bracis-33661 267 3 . . PUNCT bracis-33661 268 1 intell intell PROPN bracis-33661 268 2 7(3 7(3 NUM bracis-33661 268 3 ) ) PUNCT bracis-33661 268 4 , , PUNCT bracis-33661 268 5 623–639 623–639 NUM bracis-33661 268 6 ( ( PUNCT bracis-33661 268 7 2022 2022 NUM bracis-33661 268 8 ) ) PUNCT bracis-33661 268 9 math math NOUN bracis-33661 268 10     SPACE bracis-33661 268 11 google google PROPN bracis-33661 268 12 scholar scholar NOUN bracis-33661 268 13     SPACE bracis-33661 268 14 panda panda NOUN bracis-33661 268 15 , , PUNCT bracis-33661 268 16 r. r. PROPN bracis-33661 268 17 , , PUNCT bracis-33661 268 18 khobragade khobragade PROPN bracis-33661 268 19 , , PUNCT bracis-33661 268 20 p. p. NOUN bracis-33661 268 21 , , PUNCT bracis-33661 268 22 jambhule jambhule NOUN bracis-33661 268 23 , , PUNCT bracis-33661 268 24 p. p. PROPN bracis-33661 268 25 , , PUNCT bracis-33661 268 26 jengthe jengthe PROPN bracis-33661 268 27 , , PUNCT bracis-33661 268 28 s. s. PROPN bracis-33661 268 29 , , PUNCT bracis-33661 268 30 pal pal NOUN bracis-33661 268 31 , , PUNCT bracis-33661 268 32 p. p. PROPN bracis-33661 268 33 , , PUNCT bracis-33661 268 34 gandhi gandhi PROPN bracis-33661 268 35 , , PUNCT bracis-33661 268 36 t. t. PROPN bracis-33661 268 37 : : PUNCT bracis-33661 268 38 classification classification NOUN bracis-33661 268 39 of of ADP bracis-33661 268 40 eeg eeg PROPN bracis-33661 268 41 signal signal NOUN bracis-33661 268 42 using use VERB bracis-33661 268 43 wavelet wavelet NOUN bracis-33661 268 44 transform transform NOUN bracis-33661 268 45 and and CCONJ bracis-33661 268 46 support support VERB bracis-33661 268 47 vector vector NOUN bracis-33661 268 48 machine machine NOUN bracis-33661 268 49 for for ADP bracis-33661 268 50 epileptic epileptic ADJ bracis-33661 268 51 seizure seizure NOUN bracis-33661 268 52 diction diction NOUN bracis-33661 268 53 . . PUNCT bracis-33661 269 1 in in ADP bracis-33661 269 2 : : PUNCT bracis-33661 269 3 2010 2010 NUM bracis-33661 269 4 international international ADJ bracis-33661 269 5 conference conference NOUN bracis-33661 269 6 on on ADP bracis-33661 269 7 systems system NOUN bracis-33661 269 8 in in ADP bracis-33661 269 9 medicine medicine NOUN bracis-33661 269 10 and and CCONJ bracis-33661 269 11 biology biology NOUN bracis-33661 269 12 , , PUNCT bracis-33661 269 13 pp pp ADJ bracis-33661 269 14 . . PUNCT bracis-33661 270 1 405–408 405–408 NUM bracis-33661 270 2 . . PUNCT bracis-33661 271 1 ieee ieee NOUN bracis-33661 271 2 ( ( PUNCT bracis-33661 271 3 2010 2010 NUM bracis-33661 271 4 ) ) PUNCT bracis-33661 271 5 google google PROPN bracis-33661 271 6 scholar scholar NOUN bracis-33661 271 7     SPACE bracis-33661 271 8 portnova portnova PROPN bracis-33661 271 9 , , PUNCT bracis-33661 271 10 g.v g.v PROPN bracis-33661 271 11 . . PROPN bracis-33661 271 12 , , PUNCT bracis-33661 271 13 atanov atanov PROPN bracis-33661 271 14 , , PUNCT bracis-33661 271 15 m.s m.s PROPN bracis-33661 271 16 . . PROPN bracis-33661 271 17 : : PUNCT bracis-33661 272 1 nonlinear nonlinear ADJ bracis-33661 272 2 eeg eeg PROPN bracis-33661 272 3 parameters parameter NOUN bracis-33661 272 4 of of ADP bracis-33661 272 5 emotional emotional ADJ bracis-33661 272 6 perception perception NOUN bracis-33661 272 7 in in ADP bracis-33661 272 8 patients patient NOUN bracis-33661 272 9 with with ADP bracis-33661 272 10 moderate moderate ADJ bracis-33661 272 11 traumatic traumatic ADJ bracis-33661 272 12 brain brain NOUN bracis-33661 272 13 injury injury NOUN bracis-33661 272 14 , , PUNCT bracis-33661 272 15 coma coma NOUN bracis-33661 272 16 , , PUNCT bracis-33661 272 17 stroke stroke NOUN bracis-33661 272 18 and and CCONJ bracis-33661 272 19 schizophrenia schizophrenia NOUN bracis-33661 272 20 . . PUNCT bracis-33661 273 1 aims aims PROPN bracis-33661 273 2 neurosci neurosci PROPN bracis-33661 273 3 . . PUNCT bracis-33661 274 1 5(4 5(4 NUM bracis-33661 274 2 ) ) PUNCT bracis-33661 274 3 , , PUNCT bracis-33661 274 4 221 221 NUM bracis-33661 274 5 ( ( PUNCT bracis-33661 274 6 2018 2018 NUM bracis-33661 274 7 ) ) PUNCT bracis-33661 274 8 article article NOUN bracis-33661 274 9     SPACE bracis-33661 274 10 google google PROPN bracis-33661 274 11 scholar scholar NOUN bracis-33661 274 12     SPACE bracis-33661 274 13 sabeti sabeti NOUN bracis-33661 274 14 , , PUNCT bracis-33661 274 15 m. m. NOUN bracis-33661 274 16 , , PUNCT bracis-33661 274 17 katebi katebi PROPN bracis-33661 274 18 , , PUNCT bracis-33661 274 19 s. s. PROPN bracis-33661 274 20 , , PUNCT bracis-33661 274 21 boostani boostani PROPN bracis-33661 274 22 , , PUNCT bracis-33661 274 23 r. r. PROPN bracis-33661 274 24 , , PUNCT bracis-33661 274 25 price price NOUN bracis-33661 274 26 , , PUNCT bracis-33661 274 27 g. g. NOUN bracis-33661 274 28 : : PUNCT bracis-33661 274 29 a a DET bracis-33661 274 30 new new ADJ bracis-33661 274 31 approach approach NOUN bracis-33661 274 32 for for ADP bracis-33661 274 33 eeg eeg PROPN bracis-33661 274 34 signal signal NOUN bracis-33661 274 35 classification classification NOUN bracis-33661 274 36 of of ADP bracis-33661 274 37 schizophrenic schizophrenic ADJ bracis-33661 274 38 and and CCONJ bracis-33661 274 39 control control VERB bracis-33661 274 40 participants participant NOUN bracis-33661 274 41 . . PUNCT bracis-33661 275 1 expert expert NOUN bracis-33661 275 2 syst syst PROPN bracis-33661 275 3 . . PUNCT bracis-33661 276 1 appl appl PROPN bracis-33661 276 2 . . PUNCT bracis-33661 277 1 38(3 38(3 NUM bracis-33661 277 2 ) ) PUNCT bracis-33661 277 3 , , PUNCT bracis-33661 277 4 2063–2071 2063–2071 NUM bracis-33661 277 5 ( ( PUNCT bracis-33661 277 6 2011 2011 NUM bracis-33661 277 7 ) ) PUNCT bracis-33661 277 8 article article NOUN bracis-33661 277 9     SPACE bracis-33661 277 10 math math NOUN bracis-33661 277 11     SPACE bracis-33661 277 12 google google PROPN bracis-33661 277 13 scholar scholar NOUN bracis-33661 277 14     SPACE bracis-33661 277 15 sheykhivand sheykhivand NOUN bracis-33661 277 16 , , PUNCT bracis-33661 277 17 s. s. PROPN bracis-33661 277 18 , , PUNCT bracis-33661 277 19 mousavi mousavi PROPN bracis-33661 277 20 , , PUNCT bracis-33661 277 21 z. z. PROPN bracis-33661 277 22 , , PUNCT bracis-33661 277 23 rezaii rezaii PROPN bracis-33661 277 24 , , PUNCT bracis-33661 277 25 t.y t.y PROPN bracis-33661 277 26 . . PROPN bracis-33661 277 27 , , PUNCT bracis-33661 277 28 farzamnia farzamnia PROPN bracis-33661 277 29 , , PUNCT bracis-33661 277 30 a. a. NOUN bracis-33661 277 31 : : PUNCT bracis-33661 277 32 recognizing recognize VERB bracis-33661 277 33 emotions emotion NOUN bracis-33661 277 34 evoked evoke VERB bracis-33661 277 35 by by ADP bracis-33661 277 36 music music NOUN bracis-33661 277 37 using use VERB bracis-33661 277 38 cnn cnn PROPN bracis-33661 277 39 - - PUNCT bracis-33661 277 40 lstm lstm PROPN bracis-33661 277 41 networks network NOUN bracis-33661 277 42 on on ADP bracis-33661 277 43 eeg eeg NOUN bracis-33661 277 44 signals signal NOUN bracis-33661 277 45 . . PUNCT bracis-33661 278 1 ieee ieee NOUN bracis-33661 278 2 access access NOUN bracis-33661 278 3 8 8 NUM bracis-33661 278 4 , , PUNCT bracis-33661 278 5 139332–139345 139332–139345 NUM bracis-33661 278 6 ( ( PUNCT bracis-33661 278 7 2020 2020 NUM bracis-33661 278 8 ) ) PUNCT bracis-33661 278 9 article article NOUN bracis-33661 278 10     SPACE bracis-33661 278 11 google google PROPN bracis-33661 278 12 scholar scholar NOUN bracis-33661 278 13     SPACE bracis-33661 278 14 shoeb shoeb NOUN bracis-33661 278 15 , , PUNCT bracis-33661 278 16 a.h a.h PROPN bracis-33661 278 17 . . PROPN bracis-33661 278 18 , , PUNCT bracis-33661 278 19 guttag guttag NOUN bracis-33661 278 20 , , PUNCT bracis-33661 278 21 j.v j.v PROPN bracis-33661 278 22 . . PUNCT bracis-33661 278 23 : : PUNCT bracis-33661 279 1 application application NOUN bracis-33661 279 2 of of ADP bracis-33661 279 3 machine machine NOUN bracis-33661 279 4 learning learn VERB bracis-33661 279 5 to to ADP bracis-33661 279 6 epileptic epileptic ADJ bracis-33661 279 7 seizure seizure NOUN bracis-33661 279 8 detection detection NOUN bracis-33661 279 9 . . PUNCT bracis-33661 280 1 in in ADP bracis-33661 280 2 : : PUNCT bracis-33661 280 3 proceedings proceeding NOUN bracis-33661 280 4 of of ADP bracis-33661 280 5 the the DET bracis-33661 280 6 27th 27th ADJ bracis-33661 280 7 international international ADJ bracis-33661 280 8 conference conference NOUN bracis-33661 280 9 on on ADP bracis-33661 280 10 machine machine NOUN bracis-33661 280 11 learning learning NOUN bracis-33661 280 12 ( ( PUNCT bracis-33661 280 13 icml-10 icml-10 PROPN bracis-33661 280 14 ) ) PUNCT bracis-33661 280 15 , , PUNCT bracis-33661 280 16 pp pp PROPN bracis-33661 280 17 . . PUNCT bracis-33661 281 1 975–982 975–982 NUM bracis-33661 281 2 ( ( PUNCT bracis-33661 281 3 2010 2010 NUM bracis-33661 281 4 ) ) PUNCT bracis-33661 281 5 google google PROPN bracis-33661 281 6 scholar scholar NOUN bracis-33661 281 7     SPACE bracis-33661 281 8 soleymani soleymani NOUN bracis-33661 281 9 , , PUNCT bracis-33661 281 10 m. m. NOUN bracis-33661 281 11 , , PUNCT bracis-33661 281 12 asghari asghari ADV bracis-33661 281 13 - - PUNCT bracis-33661 281 14 esfeden esfeden ADJ bracis-33661 281 15 , , PUNCT bracis-33661 281 16 s. s. PROPN bracis-33661 281 17 , , PUNCT bracis-33661 281 18 fu fu PROPN bracis-33661 281 19 , , PUNCT bracis-33661 281 20 y. y. PROPN bracis-33661 281 21 , , PUNCT bracis-33661 281 22 pantic pantic NOUN bracis-33661 281 23 , , PUNCT bracis-33661 281 24 m. m. NOUN bracis-33661 281 25 : : PUNCT bracis-33661 281 26 analysis analysis NOUN bracis-33661 281 27 of of ADP bracis-33661 281 28 eeg eeg NOUN bracis-33661 281 29 signals signal NOUN bracis-33661 281 30 and and CCONJ bracis-33661 281 31 facial facial ADJ bracis-33661 281 32 expressions expression NOUN bracis-33661 281 33 for for ADP bracis-33661 281 34 continuous continuous ADJ bracis-33661 281 35 emotion emotion NOUN bracis-33661 281 36 detection detection NOUN bracis-33661 281 37 . . PUNCT bracis-33661 282 1 ieee ieee PROPN bracis-33661 282 2 trans trans PROPN bracis-33661 282 3 . . PROPN bracis-33661 282 4 affect affect VERB bracis-33661 282 5 . . PUNCT bracis-33661 283 1 comput comput NOUN bracis-33661 283 2 . . PUNCT bracis-33661 284 1 7(1 7(1 NUM bracis-33661 284 2 ) ) PUNCT bracis-33661 284 3 , , PUNCT bracis-33661 284 4 17–28 17–28 NUM bracis-33661 284 5 ( ( PUNCT bracis-33661 284 6 2015 2015 NUM bracis-33661 284 7 ) ) PUNCT bracis-33661 284 8 article article NOUN bracis-33661 284 9     SPACE bracis-33661 284 10 math math NOUN bracis-33661 284 11     SPACE bracis-33661 284 12 google google PROPN bracis-33661 284 13 scholar scholar NOUN bracis-33661 284 14     SPACE bracis-33661 284 15 temko temko NOUN bracis-33661 284 16 , , PUNCT bracis-33661 284 17 a. a. NOUN bracis-33661 284 18 , , PUNCT bracis-33661 284 19 thomas thomas PROPN bracis-33661 284 20 , , PUNCT bracis-33661 284 21 e. e. PROPN bracis-33661 284 22 , , PUNCT bracis-33661 284 23 marnane marnane PROPN bracis-33661 284 24 , , PUNCT bracis-33661 284 25 w. w. PROPN bracis-33661 284 26 , , PUNCT bracis-33661 284 27 lightbody lightbody PROPN bracis-33661 284 28 , , PUNCT bracis-33661 284 29 g. g. PROPN bracis-33661 284 30 , , PUNCT bracis-33661 284 31 boylan boylan PROPN bracis-33661 284 32 , , PUNCT bracis-33661 284 33 g. g. PROPN bracis-33661 284 34 : : PUNCT bracis-33661 284 35 eeg eeg NOUN bracis-33661 284 36 - - PUNCT bracis-33661 284 37 based base VERB bracis-33661 284 38 neonatal neonatal ADJ bracis-33661 284 39 seizure seizure NOUN bracis-33661 284 40 detection detection NOUN bracis-33661 284 41 with with ADP bracis-33661 284 42 support support NOUN bracis-33661 284 43 vector vector NOUN bracis-33661 284 44 machines machine NOUN bracis-33661 284 45 . . PUNCT bracis-33661 285 1 clin clin PROPN bracis-33661 285 2 . . PUNCT bracis-33661 286 1 neurophysiol neurophysiol PROPN bracis-33661 286 2 . . PUNCT bracis-33661 287 1 122(3 122(3 NUM bracis-33661 287 2 ) ) PUNCT bracis-33661 287 3 , , PUNCT bracis-33661 287 4 464–473 464–473 NUM bracis-33661 287 5 ( ( PUNCT bracis-33661 287 6 2011 2011 NUM bracis-33661 287 7 ) ) PUNCT bracis-33661 287 8 article article NOUN bracis-33661 287 9     SPACE bracis-33661 287 10 google google PROPN bracis-33661 287 11 scholar scholar NOUN bracis-33661 287 12     SPACE bracis-33661 287 13 thakor thakor PROPN bracis-33661 287 14 , , PUNCT bracis-33661 287 15 n.v n.v PROPN bracis-33661 287 16 . . PROPN bracis-33661 287 17 , , PUNCT bracis-33661 287 18 tong tong PROPN bracis-33661 287 19 , , PUNCT bracis-33661 287 20 s. s. PROPN bracis-33661 287 21 : : PUNCT bracis-33661 287 22 advances advance NOUN bracis-33661 287 23 in in ADP bracis-33661 287 24 quantitative quantitative ADJ bracis-33661 287 25 electroencephalogram electroencephalogram NOUN bracis-33661 287 26 analysis analysis NOUN bracis-33661 287 27 methods method NOUN bracis-33661 287 28 . . PUNCT bracis-33661 288 1 annu annu NOUN bracis-33661 288 2 . . PUNCT bracis-33661 289 1 rev rev PROPN bracis-33661 289 2 . . PROPN bracis-33661 289 3 biomed biome VERB bracis-33661 289 4 . . PUNCT bracis-33661 290 1 eng eng PROPN bracis-33661 290 2 . . PROPN bracis-33661 291 1 6 6 NUM bracis-33661 291 2 , , PUNCT bracis-33661 291 3 453–495 453–495 NUM bracis-33661 291 4 ( ( PUNCT bracis-33661 291 5 2004 2004 NUM bracis-33661 291 6 ) ) PUNCT bracis-33661 291 7 article article NOUN bracis-33661 291 8     SPACE bracis-33661 291 9 math math NOUN bracis-33661 291 10     SPACE bracis-33661 291 11 google google PROPN bracis-33661 291 12 scholar scholar NOUN bracis-33661 291 13     SPACE bracis-33661 291 14 wang wang PROPN bracis-33661 291 15 , , PUNCT bracis-33661 291 16 s. s. PROPN bracis-33661 291 17 , , PUNCT bracis-33661 291 18 qu qu PROPN bracis-33661 291 19 , , PUNCT bracis-33661 291 20 j. j. PROPN bracis-33661 291 21 , , PUNCT bracis-33661 291 22 zhang zhang PROPN bracis-33661 291 23 , , PUNCT bracis-33661 291 24 y. y. PROPN bracis-33661 291 25 , , PUNCT bracis-33661 291 26 zhang zhang PROPN bracis-33661 291 27 , , PUNCT bracis-33661 291 28 y. y. NOUN bracis-33661 291 29 : : PUNCT bracis-33661 291 30 multimodal multimodal ADJ bracis-33661 291 31 emotion emotion NOUN bracis-33661 291 32 recognition recognition NOUN bracis-33661 291 33 from from ADP bracis-33661 291 34 eeg eeg NOUN bracis-33661 291 35 signals signal NOUN bracis-33661 291 36 and and CCONJ bracis-33661 291 37 facial facial ADJ bracis-33661 291 38 expressions expression NOUN bracis-33661 291 39 . . PUNCT bracis-33661 292 1 ieee ieee NOUN bracis-33661 292 2 access access NOUN bracis-33661 292 3 11 11 NUM bracis-33661 292 4 , , PUNCT bracis-33661 292 5 33061–33068 33061–33068 NUM bracis-33661 292 6 ( ( PUNCT bracis-33661 292 7 2023 2023 NUM bracis-33661 292 8 ) ) PUNCT bracis-33661 292 9 article article NOUN bracis-33661 292 10     SPACE bracis-33661 292 11 math math NOUN bracis-33661 292 12     SPACE bracis-33661 292 13 google google PROPN bracis-33661 292 14 scholar scholar NOUN bracis-33661 292 15     SPACE bracis-33661 292 16 wu wu PROPN bracis-33661 292 17 , , PUNCT bracis-33661 292 18 f. f. PROPN bracis-33661 292 19 , , PUNCT bracis-33661 292 20 zhang zhang PROPN bracis-33661 292 21 , , PUNCT bracis-33661 292 22 t. t. PROPN bracis-33661 292 23 , , PUNCT bracis-33661 292 24 jr jr PROPN bracis-33661 292 25 . . PROPN bracis-33661 292 26 , , PUNCT bracis-33661 292 27 a.h.s a.h.s PROPN bracis-33661 292 28 . . PROPN bracis-33661 292 29 , , PUNCT bracis-33661 292 30 fifty fifty NUM bracis-33661 292 31 , , PUNCT bracis-33661 292 32 c. c. NOUN bracis-33661 292 33 , , PUNCT bracis-33661 292 34 yu yu PROPN bracis-33661 292 35 , , PUNCT bracis-33661 292 36 t. t. PROPN bracis-33661 292 37 , , PUNCT bracis-33661 292 38 weinberger weinberger PROPN bracis-33661 292 39 , , PUNCT bracis-33661 292 40 k.q k.q PROPN bracis-33661 292 41 . . PROPN bracis-33661 292 42 : : PUNCT bracis-33661 292 43 simplifying simplify VERB bracis-33661 292 44 graph graph NOUN bracis-33661 292 45 convolutional convolutional ADJ bracis-33661 292 46 networks network NOUN bracis-33661 292 47 . . PUNCT bracis-33661 293 1 arxiv arxiv PROPN bracis-33661 293 2 preprint preprint VERB bracis-33661 293 3 arxiv:1902.07153 arxiv:1902.07153 NOUN bracis-33661 293 4 ( ( PUNCT bracis-33661 293 5 2019 2019 NUM bracis-33661 293 6 ) ) PUNCT bracis-33661 294 1 zheng zheng PROPN bracis-33661 294 2 , , PUNCT bracis-33661 294 3 w.l w.l PROPN bracis-33661 294 4 . . PROPN bracis-33661 294 5 , , PUNCT bracis-33661 294 6 et et PROPN bracis-33661 294 7 al al PROPN bracis-33661 294 8 . . PROPN bracis-33661 294 9 : : PUNCT bracis-33661 294 10 predicting predict VERB bracis-33661 294 11 neurological neurological ADJ bracis-33661 294 12 outcome outcome NOUN bracis-33661 294 13 in in ADP bracis-33661 294 14 comatose comatose ADJ bracis-33661 294 15 patients patient NOUN bracis-33661 294 16 after after ADP bracis-33661 294 17 cardiac cardiac ADJ bracis-33661 294 18 arrest arrest NOUN bracis-33661 294 19 with with ADP bracis-33661 294 20 multiscale multiscale ADJ bracis-33661 294 21 deep deep ADJ bracis-33661 294 22 neural neural ADJ bracis-33661 294 23 networks network NOUN bracis-33661 294 24 . . PUNCT bracis-33661 295 1 resuscitation resuscitation NOUN bracis-33661 295 2 169 169 NUM bracis-33661 295 3 , , PUNCT bracis-33661 295 4 86–94 86–94 NUM bracis-33661 295 5 ( ( PUNCT bracis-33661 295 6 2021 2021 NUM bracis-33661 295 7 ) ) PUNCT bracis-33661 295 8 article article NOUN bracis-33661 295 9     SPACE bracis-33661 295 10 math math NOUN bracis-33661 295 11     SPACE bracis-33661 295 12 google google PROPN bracis-33661 295 13 scholar scholar NOUN bracis-33661 295 14     SPACE bracis-33661 295 15 zheng zheng PROPN bracis-33661 295 16 , , PUNCT bracis-33661 295 17 w.l w.l PROPN bracis-33661 295 18 . . PROPN bracis-33661 295 19 , , PUNCT bracis-33661 295 20 et et PROPN bracis-33661 295 21 al al PROPN bracis-33661 295 22 . . PROPN bracis-33661 295 23 : : PUNCT bracis-33661 295 24 predicting predict VERB bracis-33661 295 25 neurological neurological ADJ bracis-33661 295 26 outcome outcome NOUN bracis-33661 295 27 from from ADP bracis-33661 295 28 electroencephalogram electroencephalogram NOUN bracis-33661 295 29 dynamics dynamic NOUN bracis-33661 295 30 in in ADP bracis-33661 295 31 comatose comatose ADJ bracis-33661 295 32 patients patient NOUN bracis-33661 295 33 after after ADP bracis-33661 295 34 cardiac cardiac ADJ bracis-33661 295 35 arrest arrest NOUN bracis-33661 295 36 with with ADP bracis-33661 295 37 deep deep ADJ bracis-33661 295 38 learning learning NOUN bracis-33661 295 39 . . PUNCT bracis-33661 296 1 ieee ieee PROPN bracis-33661 296 2 trans trans PROPN bracis-33661 296 3 . . PROPN bracis-33661 296 4 biomed biome VERB bracis-33661 296 5 . . PUNCT bracis-33661 297 1 eng eng PROPN bracis-33661 297 2 . . PUNCT bracis-33661 298 1 69(5 69(5 NUM bracis-33661 298 2 ) ) PUNCT bracis-33661 298 3 , , PUNCT bracis-33661 298 4 1813–1825 1813–1825 NUM bracis-33661 298 5 ( ( PUNCT bracis-33661 298 6 2021 2021 NUM bracis-33661 298 7 ) ) PUNCT bracis-33661 298 8 article article NOUN bracis-33661 298 9     SPACE bracis-33661 298 10 math math NOUN bracis-33661 298 11     SPACE bracis-33661 298 12 google google PROPN bracis-33661 298 13 scholar scholar NOUN bracis-33661 298 14     SPACE bracis-33661 298 15 zhong zhong PROPN bracis-33661 298 16 , , PUNCT bracis-33661 298 17 p. p. PROPN bracis-33661 298 18 , , PUNCT bracis-33661 298 19 wang wang PROPN bracis-33661 298 20 , , PUNCT bracis-33661 298 21 d. d. PROPN bracis-33661 298 22 , , PUNCT bracis-33661 298 23 miao miao PROPN bracis-33661 298 24 , , PUNCT bracis-33661 298 25 c. c. NOUN bracis-33661 298 26 : : PUNCT bracis-33661 298 27 eeg eeg NOUN bracis-33661 298 28 - - PUNCT bracis-33661 298 29 based base VERB bracis-33661 298 30 emotion emotion NOUN bracis-33661 298 31 recognition recognition NOUN bracis-33661 298 32 using use VERB bracis-33661 298 33 regularized regularize VERB bracis-33661 298 34 graph graph NOUN bracis-33661 298 35 neural neural ADJ bracis-33661 298 36 networks network NOUN bracis-33661 298 37 . . PUNCT bracis-33661 299 1 ieee ieee PROPN bracis-33661 299 2 trans trans PROPN bracis-33661 299 3 . . PROPN bracis-33661 299 4 affect affect VERB bracis-33661 299 5 . . PUNCT bracis-33661 300 1 comput comput NOUN bracis-33661 300 2 . . PUNCT bracis-33661 301 1 13(3 13(3 NUM bracis-33661 301 2 ) ) PUNCT bracis-33661 301 3 , , PUNCT bracis-33661 301 4 1290–1301 1290–1301 NUM bracis-33661 301 5 ( ( PUNCT bracis-33661 301 6 2022 2022 NUM bracis-33661 301 7 ) ) PUNCT bracis-33661 301 8 article article NOUN bracis-33661 301 9     SPACE bracis-33661 301 10 math math NOUN bracis-33661 301 11     SPACE bracis-33661 301 12 google google PROPN bracis-33661 301 13 scholar scholar NOUN bracis-33661 301 14     SPACE bracis-33661 301 15 download download NOUN bracis-33661 301 16 references reference NOUN bracis-33661 301 17 acknowledgment acknowledgment NOUN bracis-33661 301 18 authors author NOUN bracis-33661 301 19 thank thank VERB bracis-33661 301 20 the the DET bracis-33661 301 21 financial financial ADJ bracis-33661 301 22 support support NOUN bracis-33661 301 23 given give VERB bracis-33661 301 24 by by ADP bracis-33661 301 25 brazilian brazilian ADJ bracis-33661 301 26 national national PROPN bracis-33661 301 27 council council PROPN bracis-33661 301 28 for for ADP bracis-33661 301 29 scientific scientific ADJ bracis-33661 301 30 and and CCONJ bracis-33661 301 31 technological technological ADJ bracis-33661 301 32 development development NOUN bracis-33661 301 33 cnpq cnpq NOUN bracis-33661 301 34 ( ( PUNCT bracis-33661 301 35 grant grant VERB bracis-33661 301 36 number number NOUN bracis-33661 301 37 420212/2023 420212/2023 NUM bracis-33661 301 38 - - SYM bracis-33661 301 39 0 0 NUM bracis-33661 301 40 ) ) PUNCT bracis-33661 301 41 , , PUNCT bracis-33661 301 42 and and CCONJ bracis-33661 301 43 the the DET bracis-33661 301 44 center center NOUN bracis-33661 301 45 for for ADP bracis-33661 301 46 artificial artificial ADJ bracis-33661 301 47 intelligence intelligence NOUN bracis-33661 301 48 ( ( PUNCT bracis-33661 301 49 c4ai c4ai NOUN bracis-33661 301 50 - - PUNCT bracis-33661 301 51 usp usp NOUN bracis-33661 301 52 ) ) PUNCT bracis-33661 301 53 , , PUNCT bracis-33661 301 54 with with ADP bracis-33661 301 55 support support NOUN bracis-33661 301 56 by by ADP bracis-33661 301 57 the the DET bracis-33661 301 58 são são PROPN bracis-33661 301 59 paulo paulo PROPN bracis-33661 301 60 research research PROPN bracis-33661 301 61 foundation foundation PROPN bracis-33661 301 62 ( ( PUNCT bracis-33661 301 63 grants grant NOUN bracis-33661 301 64 number number NOUN bracis-33661 301 65 2019/07665 2019/07665 NUM bracis-33661 301 66 - - SYM bracis-33661 301 67 4 4 NUM bracis-33661 301 68 and and CCONJ bracis-33661 301 69 2024/02783 2024/02783 NUM bracis-33661 301 70 - - SYM bracis-33661 301 71 7 7 NUM bracis-33661 301 72 ) ) PUNCT bracis-33661 301 73 and and CCONJ bracis-33661 301 74 by by ADP bracis-33661 301 75 the the DET bracis-33661 301 76 ibm ibm PROPN bracis-33661 301 77 corporation corporation NOUN bracis-33661 301 78 . . PUNCT bracis-33661 302 1 we we PRON bracis-33661 302 2 also also ADV bracis-33661 302 3 thank thank VERB bracis-33661 302 4 nvidia nvidia PROPN bracis-33661 302 5 corporation corporation NOUN bracis-33661 302 6 by by ADP bracis-33661 302 7 the the DET bracis-33661 302 8 donation donation NOUN bracis-33661 302 9 of of ADP bracis-33661 302 10 a a DET bracis-33661 302 11 titan titan NOUN bracis-33661 302 12 v v ADP bracis-33661 302 13 gpu gpu NOUN bracis-33661 302 14 used use VERB bracis-33661 302 15 in in ADP bracis-33661 302 16 this this DET bracis-33661 302 17 research research NOUN bracis-33661 302 18 . . PUNCT bracis-33661 303 1 author author NOUN bracis-33661 303 2 information information NOUN bracis-33661 303 3 authors author NOUN bracis-33661 303 4 and and CCONJ bracis-33661 303 5 affiliations affiliation NOUN bracis-33661 303 6 faculty faculty NOUN bracis-33661 303 7 of of ADP bracis-33661 303 8 computing computing NOUN bracis-33661 303 9 , , PUNCT bracis-33661 303 10 federal federal ADJ bracis-33661 303 11 university university PROPN bracis-33661 303 12 of of ADP bracis-33661 303 13 uberlândia uberlândia PROPN bracis-33661 303 14 , , PUNCT bracis-33661 303 15 uberlândia uberlândia PROPN bracis-33661 303 16 , , PUNCT bracis-33661 303 17 brazil brazil PROPN bracis-33661 303 18 odelmo odelmo PROPN bracis-33661 303 19 o. o. PROPN bracis-33661 303 20 nascimento nascimento PROPN bracis-33661 303 21 junior junior PROPN bracis-33661 303 22 , , PUNCT bracis-33661 303 23     SPACE bracis-33661 303 24 dhara dhara PROPN bracis-33661 303 25 l. l. PROPN bracis-33661 303 26 c. c. PROPN bracis-33661 303 27 assis assis PROPN bracis-33661 303 28 , , PUNCT bracis-33661 303 29     SPACE bracis-33661 303 30 joão joão PROPN bracis-33661 303 31 b. b. PROPN bracis-33661 303 32 destro destro PROPN bracis-33661 303 33 filho filho PROPN bracis-33661 303 34     SPACE bracis-33661 303 35 & & CCONJ bracis-33661 303 36     SPACE bracis-33661 303 37 murillo murillo PROPN bracis-33661 303 38 g. g. PROPN bracis-33661 303 39 carneiro carneiro PROPN bracis-33661 303 40 department department PROPN bracis-33661 303 41 of of ADP bracis-33661 303 42 computing computing NOUN bracis-33661 303 43 and and CCONJ bracis-33661 303 44 mathematics mathematic NOUN bracis-33661 303 45 , , PUNCT bracis-33661 303 46 university university PROPN bracis-33661 303 47 of of ADP bracis-33661 303 48 são são PROPN bracis-33661 303 49 paulo paulo PROPN bracis-33661 303 50 , , PUNCT bracis-33661 303 51 ribeirão ribeirão PROPN bracis-33661 303 52 preto preto PROPN bracis-33661 303 53 , , PUNCT bracis-33661 303 54 brazil brazil PROPN bracis-33661 303 55 liang liang PROPN bracis-33661 303 56 zhao zhao PROPN bracis-33661 303 57     SPACE bracis-33661 303 58 & & CCONJ bracis-33661 303 59     SPACE bracis-33661 303 60 murillo murillo PROPN bracis-33661 303 61 g. g. PROPN bracis-33661 303 62 carneiro carneiro PROPN bracis-33661 303 63 authors author NOUN bracis-33661 303 64 odelmo odelmo PRON bracis-33661 303 65 o. o. PROPN bracis-33661 303 66 nascimento nascimento PROPN bracis-33661 303 67 juniorview juniorview PROPN bracis-33661 303 68 author author NOUN bracis-33661 303 69 publications publication NOUN bracis-33661 303 70 search search NOUN bracis-33661 303 71 author author NOUN bracis-33661 303 72 on on ADP bracis-33661 303 73 : : PUNCT bracis-33661 303 74 pubmed pubmed PROPN bracis-33661 303 75     SPACE bracis-33661 303 76 google google PROPN bracis-33661 303 77 scholar scholar PROPN bracis-33661 303 78 dhara dhara PROPN bracis-33661 303 79 l. l. PROPN bracis-33661 303 80 c. c. PROPN bracis-33661 303 81 assisview assisview PROPN bracis-33661 303 82 author author NOUN bracis-33661 303 83 publications publication NOUN bracis-33661 303 84 search search NOUN bracis-33661 303 85 author author NOUN bracis-33661 303 86 on on ADP bracis-33661 303 87 : : PUNCT bracis-33661 303 88 pubmed pubmed PROPN bracis-33661 303 89     SPACE bracis-33661 303 90 google google PROPN bracis-33661 303 91 scholar scholar PROPN bracis-33661 303 92 joão joão PROPN bracis-33661 303 93 b. b. PROPN bracis-33661 303 94 destro destro PROPN bracis-33661 303 95 filhoview filhoview PROPN bracis-33661 303 96 author author NOUN bracis-33661 303 97 publications publication NOUN bracis-33661 303 98 search search NOUN bracis-33661 303 99 author author NOUN bracis-33661 303 100 on on ADP bracis-33661 303 101 : : PUNCT bracis-33661 303 102 pubmed pubmed PROPN bracis-33661 303 103     SPACE bracis-33661 303 104 google google PROPN bracis-33661 303 105 scholar scholar PROPN bracis-33661 303 106 liang liang PROPN bracis-33661 303 107 zhaoview zhaoview PROPN bracis-33661 303 108 author author NOUN bracis-33661 303 109 publications publication NOUN bracis-33661 303 110 search search NOUN bracis-33661 303 111 author author NOUN bracis-33661 303 112 on on ADP bracis-33661 303 113 : : PUNCT bracis-33661 303 114 pubmed pubmed PROPN bracis-33661 303 115     SPACE bracis-33661 303 116 google google PROPN bracis-33661 303 117 scholar scholar NOUN bracis-33661 303 118 murillo murillo PROPN bracis-33661 303 119 g. g. PROPN bracis-33661 303 120 carneiroview carneiroview PROPN bracis-33661 303 121 author author NOUN bracis-33661 303 122 publications publication NOUN bracis-33661 303 123 search search NOUN bracis-33661 303 124 author author NOUN bracis-33661 303 125 on on ADP bracis-33661 303 126 : : PUNCT bracis-33661 303 127 pubmed pubmed PROPN bracis-33661 303 128     SPACE bracis-33661 303 129 google google PROPN bracis-33661 303 130 scholar scholar NOUN bracis-33661 303 131 corresponding correspond VERB bracis-33661 303 132 author author NOUN bracis-33661 303 133 correspondence correspondence NOUN bracis-33661 303 134 to to ADP bracis-33661 303 135 odelmo odelmo PROPN bracis-33661 303 136 o. o. PROPN bracis-33661 303 137 nascimento nascimento PROPN bracis-33661 303 138 junior junior PROPN bracis-33661 303 139 . . PUNCT bracis-33661 304 1 editor editor NOUN bracis-33661 304 2 information information NOUN bracis-33661 304 3 editors editor NOUN bracis-33661 304 4 and and CCONJ bracis-33661 304 5 affiliations affiliation NOUN bracis-33661 304 6 universidade universidade VERB bracis-33661 304 7 federal federal PROPN bracis-33661 304 8 fluminense fluminense PROPN bracis-33661 304 9 , , PUNCT bracis-33661 304 10 niterói niterói PROPN bracis-33661 304 11 , , PUNCT bracis-33661 304 12 brazil brazil PROPN bracis-33661 304 13 aline aline PROPN bracis-33661 304 14 paes paes PROPN bracis-33661 304 15 instituto instituto PROPN bracis-33661 304 16 tecnológico tecnológico PROPN bracis-33661 304 17 de de PROPN bracis-33661 304 18 aeronáutica aeronáutica PROPN bracis-33661 304 19 , , PUNCT bracis-33661 304 20 são são PROPN bracis-33661 304 21 josé josé PROPN bracis-33661 304 22 dos dos PROPN bracis-33661 304 23 campos campos PROPN bracis-33661 304 24 , , PUNCT bracis-33661 304 25 brazil brazil PROPN bracis-33661 304 26 filipe filipe PROPN bracis-33661 304 27 a. a. PROPN bracis-33661 304 28 n. n. PROPN bracis-33661 304 29 verri verri PROPN bracis-33661 304 30 rights right NOUN bracis-33661 304 31 and and CCONJ bracis-33661 304 32 permissions permission NOUN bracis-33661 304 33 reprints reprint NOUN bracis-33661 304 34 and and CCONJ bracis-33661 304 35 permissions permission VERB bracis-33661 304 36 copyright copyright NOUN bracis-33661 304 37 information information NOUN bracis-33661 304 38 © © ADP bracis-33661 304 39 2025 2025 NUM bracis-33661 304 40 the the DET bracis-33661 304 41 author(s author(s NOUN bracis-33661 304 42 ) ) PUNCT bracis-33661 304 43 , , PUNCT bracis-33661 304 44 under under ADP bracis-33661 304 45 exclusive exclusive ADJ bracis-33661 304 46 license license NOUN bracis-33661 304 47 to to ADP bracis-33661 304 48 springer springer NOUN bracis-33661 304 49 nature nature PROPN bracis-33661 304 50 switzerland switzerland PROPN bracis-33661 304 51 ag ag PROPN bracis-33661 304 52 about about ADP bracis-33661 304 53 this this DET bracis-33661 304 54 paper paper NOUN bracis-33661 304 55 cite cite VERB bracis-33661 304 56 this this DET bracis-33661 304 57 paper paper NOUN bracis-33661 304 58 nascimento nascimento PROPN bracis-33661 304 59 junior junior PROPN bracis-33661 304 60 , , PUNCT bracis-33661 304 61 o.o o.o PROPN bracis-33661 304 62 . . PROPN bracis-33661 304 63 , , PUNCT bracis-33661 304 64 assis assis NOUN bracis-33661 304 65 , , PUNCT bracis-33661 304 66 d.l.c d.l.c NOUN bracis-33661 304 67 . . PROPN bracis-33661 304 68 , , PUNCT bracis-33661 304 69 destro destro VERB bracis-33661 304 70 filho filho NOUN bracis-33661 304 71 , , PUNCT bracis-33661 304 72 j.b j.b PROPN bracis-33661 304 73 . . PROPN bracis-33661 304 74 , , PUNCT bracis-33661 304 75 zhao zhao PROPN bracis-33661 304 76 , , PUNCT bracis-33661 304 77 l. l. PROPN bracis-33661 304 78 , , PUNCT bracis-33661 304 79 carneiro carneiro PROPN bracis-33661 304 80 , , PUNCT bracis-33661 304 81 m.g m.g PROPN bracis-33661 304 82 . . PROPN bracis-33661 304 83 ( ( PUNCT bracis-33661 304 84 2025 2025 NUM bracis-33661 304 85 ) ) PUNCT bracis-33661 304 86 . . PUNCT bracis-33661 305 1 modeling model VERB bracis-33661 305 2 eeg eeg PROPN bracis-33661 305 3 data datum NOUN bracis-33661 305 4 into into ADP bracis-33661 305 5 graphs graph NOUN bracis-33661 305 6 for for ADP bracis-33661 305 7     SPACE bracis-33661 305 8 the the DET bracis-33661 305 9     SPACE bracis-33661 305 10 prognostic prognostic NOUN bracis-33661 305 11 of of ADP bracis-33661 305 12     SPACE bracis-33661 305 13 patients patient NOUN bracis-33661 305 14 in in ADP bracis-33661 305 15     SPACE bracis-33661 305 16 coma coma NOUN bracis-33661 305 17 using use VERB bracis-33661 305 18 graph graph NOUN bracis-33661 305 19 neural neural ADJ bracis-33661 305 20 networks network NOUN bracis-33661 305 21 . . PUNCT bracis-33661 306 1 in in ADP bracis-33661 306 2 : : PUNCT bracis-33661 306 3 paes paes PROPN bracis-33661 306 4 , , PUNCT bracis-33661 306 5 a. a. PROPN bracis-33661 306 6 , , PUNCT bracis-33661 306 7 verri verri NOUN bracis-33661 306 8 , , PUNCT bracis-33661 306 9 f.a.n f.a.n ADJ bracis-33661 306 10 . . PUNCT bracis-33661 306 11 ( ( PUNCT bracis-33661 306 12 eds ed NOUN bracis-33661 306 13 ) ) PUNCT bracis-33661 306 14 intelligent intelligent ADJ bracis-33661 306 15 systems system NOUN bracis-33661 306 16 . . PUNCT bracis-33661 306 17 bracis bracis PROPN bracis-33661 306 18 2024 2024 NUM bracis-33661 306 19 . . PUNCT bracis-33661 307 1 lecture lecture NOUN bracis-33661 307 2 notes note NOUN bracis-33661 307 3 in in ADP bracis-33661 307 4 computer computer NOUN bracis-33661 307 5 science science NOUN bracis-33661 307 6 ( ( PUNCT bracis-33661 307 7 ) ) PUNCT bracis-33661 307 8 , , PUNCT bracis-33661 307 9 vol vol NOUN bracis-33661 307 10 15415 15415 NUM bracis-33661 307 11 . . PUNCT bracis-33661 308 1 springer springer NOUN bracis-33661 308 2 , , PUNCT bracis-33661 308 3 cham cham PROPN bracis-33661 308 4 . . PUNCT bracis-33661 309 1 https://doi.org/10.1007/978-3-031-79038-6_16 https://doi.org/10.1007/978-3-031-79038-6_16 PROPN bracis-33661 309 2 download download NOUN bracis-33661 309 3 citation citation NOUN bracis-33661 309 4 .ris .ris PUNCT bracis-33661 309 5 .enw .enw PROPN bracis-33661 309 6 .bib .bib PUNCT bracis-33661 310 1 doi doi PROPN bracis-33661 310 2 : : PUNCT bracis-33661 310 3 https://doi.org/10.1007/978-3-031-79038-6_16 https://doi.org/10.1007/978-3-031-79038-6_16 PROPN bracis-33661 310 4 published publish VERB bracis-33661 310 5 : : PUNCT bracis-33661 310 6 31 31 NUM bracis-33661 310 7 january january PROPN bracis-33661 310 8 2025 2025 NUM bracis-33661 310 9 publisher publisher NOUN bracis-33661 310 10 name name NOUN bracis-33661 310 11 : : PUNCT bracis-33661 310 12 springer springer NOUN bracis-33661 310 13 , , PUNCT bracis-33661 310 14 cham cham PROPN bracis-33661 310 15 print print PROPN bracis-33661 310 16 isbn isbn PROPN bracis-33661 310 17 : : PUNCT bracis-33661 310 18 978 978 NUM bracis-33661 310 19 - - SYM bracis-33661 310 20 3 3 NUM bracis-33661 310 21 - - PUNCT bracis-33661 310 22 031 031 NUM bracis-33661 310 23 - - PUNCT bracis-33661 310 24 79037 79037 NUM bracis-33661 310 25 - - SYM bracis-33661 310 26 9 9 NUM bracis-33661 310 27 online online ADJ bracis-33661 310 28 isbn isbn NOUN bracis-33661 310 29 : : PUNCT bracis-33661 310 30 978 978 NUM bracis-33661 310 31 - - SYM bracis-33661 310 32 3 3 NUM bracis-33661 310 33 - - PUNCT bracis-33661 310 34 031 031 NUM bracis-33661 310 35 - - PUNCT bracis-33661 310 36 79038 79038 NUM bracis-33661 310 37 - - SYM bracis-33661 310 38 6 6 NUM bracis-33661 310 39 ebook ebook NOUN bracis-33661 310 40 packages package NOUN bracis-33661 310 41 : : PUNCT bracis-33661 310 42 computer computer NOUN bracis-33661 310 43 sciencecomputer sciencecomputer NOUN bracis-33661 310 44 science science NOUN bracis-33661 310 45 ( ( PUNCT bracis-33661 310 46 r0 r0 NOUN bracis-33661 310 47 ) ) PUNCT bracis-33661 310 48 share share VERB bracis-33661 310 49 this this DET bracis-33661 310 50 paper paper NOUN bracis-33661 310 51 anyone anyone PRON bracis-33661 310 52 you you PRON bracis-33661 310 53 share share VERB bracis-33661 310 54 the the DET bracis-33661 310 55 following follow VERB bracis-33661 310 56 link link NOUN bracis-33661 310 57 with with ADP bracis-33661 310 58 will will AUX bracis-33661 310 59 be be AUX bracis-33661 310 60 able able ADJ bracis-33661 310 61 to to PART bracis-33661 310 62 read read VERB bracis-33661 310 63 this this DET bracis-33661 310 64 content content NOUN bracis-33661 310 65 : : PUNCT bracis-33661 310 66 get get VERB bracis-33661 310 67 shareable shareable ADJ bracis-33661 310 68 linksorry linksorry NOUN bracis-33661 310 69 , , PUNCT bracis-33661 310 70 a a DET bracis-33661 310 71 shareable shareable ADJ bracis-33661 310 72 link link NOUN bracis-33661 310 73 is be AUX bracis-33661 310 74 not not PART bracis-33661 310 75 currently currently ADV bracis-33661 310 76 available available ADJ bracis-33661 310 77 for for ADP bracis-33661 310 78 this this DET bracis-33661 310 79 article article NOUN bracis-33661 310 80 . . PUNCT bracis-33661 311 1 copy copy VERB bracis-33661 311 2 shareable shareable ADJ bracis-33661 311 3 link link NOUN bracis-33661 311 4 to to PART bracis-33661 311 5 clipboard clipboard NOUN bracis-33661 311 6 provided provide VERB bracis-33661 311 7 by by ADP bracis-33661 311 8 the the DET bracis-33661 311 9 springer springer NOUN bracis-33661 311 10 nature nature PROPN bracis-33661 311 11 sharedit sharedit PROPN bracis-33661 311 12 content content NOUN bracis-33661 311 13 - - PUNCT bracis-33661 311 14 sharing share VERB bracis-33661 311 15 initiative initiative NOUN bracis-33661 311 16 keywords keyword NOUN bracis-33661 311 17 graph graph VERB bracis-33661 311 18 neural neural ADJ bracis-33661 311 19 networks network NOUN bracis-33661 311 20 electroencephalogram electroencephalogram VERB bracis-33661 311 21 coma coma NOUN bracis-33661 311 22 outcome outcome NOUN bracis-33661 311 23 sequential sequential ADJ bracis-33661 311 24 data datum NOUN bracis-33661 311 25 graph graph NOUN bracis-33661 311 26 - - PUNCT bracis-33661 311 27 based base VERB bracis-33661 311 28 learning learn VERB bracis-33661 311 29 eegraph eegraph NOUN bracis-33661 311 30 publish publish VERB bracis-33661 311 31 with with ADP bracis-33661 311 32 us us PROPN bracis-33661 311 33 policies policy NOUN bracis-33661 311 34 and and CCONJ bracis-33661 311 35 ethics ethic NOUN bracis-33661 311 36 profiles profile NOUN bracis-33661 311 37 murillo murillo PROPN bracis-33661 311 38 g. g. PROPN bracis-33661 311 39 carneiro carneiro PROPN bracis-33661 311 40 view view PROPN bracis-33661 311 41 author author NOUN bracis-33661 311 42 profile profile NOUN bracis-33661 311 43 search search NOUN bracis-33661 311 44 search search NOUN bracis-33661 311 45 by by ADP bracis-33661 311 46 keyword keyword NOUN bracis-33661 311 47 or or CCONJ bracis-33661 311 48 author author NOUN bracis-33661 311 49 search search NOUN bracis-33661 311 50 navigation navigation NOUN bracis-33661 311 51 find find VERB bracis-33661 311 52 a a DET bracis-33661 311 53 journal journal NOUN bracis-33661 311 54 publish publish VERB bracis-33661 311 55 with with ADP bracis-33661 311 56 us we PRON bracis-33661 311 57 track track VERB bracis-33661 311 58 your your PRON bracis-33661 311 59 research research NOUN bracis-33661 311 60 discover discover VERB bracis-33661 311 61 content content NOUN bracis-33661 311 62 journals journal NOUN bracis-33661 311 63 a a DET bracis-33661 311 64 - - PUNCT bracis-33661 311 65 z z NOUN bracis-33661 311 66 books book NOUN bracis-33661 311 67 a a DET bracis-33661 311 68 - - PUNCT bracis-33661 311 69 z z NOUN bracis-33661 311 70 publish publish NOUN bracis-33661 311 71 with with ADP bracis-33661 311 72 us us PROPN bracis-33661 311 73 journal journal PROPN bracis-33661 311 74 finder finder PROPN bracis-33661 311 75 publish publish VERB bracis-33661 311 76 your your PRON bracis-33661 311 77 research research NOUN bracis-33661 311 78 language language NOUN bracis-33661 311 79 editing edit VERB bracis-33661 311 80 open open ADJ bracis-33661 311 81 access access NOUN bracis-33661 311 82 publishing publishing NOUN bracis-33661 311 83 products product NOUN bracis-33661 311 84 and and CCONJ bracis-33661 311 85 services service NOUN bracis-33661 311 86 our our PRON bracis-33661 311 87 products product NOUN bracis-33661 311 88 librarians librarian VERB bracis-33661 311 89 societies society NOUN bracis-33661 311 90 partners partner NOUN bracis-33661 311 91 and and CCONJ bracis-33661 311 92 advertisers advertiser NOUN bracis-33661 311 93 our our PRON bracis-33661 311 94 brands brand NOUN bracis-33661 311 95 springer springer NOUN bracis-33661 311 96 nature nature PROPN bracis-33661 311 97 portfolio portfolio PROPN bracis-33661 311 98 bmc bmc PROPN bracis-33661 311 99 palgrave palgrave PROPN bracis-33661 311 100 macmillan macmillan PROPN bracis-33661 311 101 apress apress PROPN bracis-33661 311 102 discover discover VERB bracis-33661 311 103 your your PRON bracis-33661 311 104 privacy privacy NOUN bracis-33661 311 105 choices choice NOUN bracis-33661 311 106 / / SYM bracis-33661 311 107 manage manage NOUN bracis-33661 311 108 cookies cookie NOUN bracis-33661 311 109 your your PRON bracis-33661 311 110 us us PROPN bracis-33661 312 1 state state NOUN bracis-33661 312 2 privacy privacy NOUN bracis-33661 312 3 rights right NOUN bracis-33661 312 4 accessibility accessibility NOUN bracis-33661 312 5 statement statement NOUN bracis-33661 312 6 terms term NOUN bracis-33661 312 7 and and CCONJ bracis-33661 312 8 conditions condition NOUN bracis-33661 312 9 privacy privacy NOUN bracis-33661 312 10 policy policy NOUN bracis-33661 312 11 help help NOUN bracis-33661 312 12 and and CCONJ bracis-33661 312 13 support support VERB bracis-33661 312 14 legal legal ADJ bracis-33661 312 15 notice notice NOUN bracis-33661 312 16 cancel cancel VERB bracis-33661 312 17 contracts contract NOUN bracis-33661 312 18 here here ADV bracis-33661 312 19 129.74.145.123 129.74.145.123 NUM bracis-33661 312 20 hesburgh hesburgh PROPN bracis-33661 312 21 library library PROPN bracis-33661 312 22 er er INTJ bracis-33661 312 23 unit unit NOUN bracis-33661 312 24 ( ( PUNCT bracis-33661 312 25 3005732405 3005732405 NUM bracis-33661 312 26 ) ) PUNCT bracis-33661 312 27 northeast northeast ADJ bracis-33661 312 28 research research NOUN bracis-33661 312 29 libraries library NOUN bracis-33661 312 30 ( ( PUNCT bracis-33661 312 31 nerl nerl PROPN bracis-33661 312 32 ) ) PUNCT bracis-33661 312 33 ( ( PUNCT bracis-33661 312 34 8200828607 8200828607 NUM bracis-33661 312 35 ) ) PUNCT bracis-33661 312 36 nerl nerl VERB bracis-33661 312 37 ta ta X bracis-33661 312 38 account account NOUN bracis-33661 312 39 ( ( PUNCT bracis-33661 312 40 3006206169 3006206169 NUM bracis-33661 312 41 ) ) PUNCT bracis-33661 312 42 university university NOUN bracis-33661 312 43 of of ADP bracis-33661 312 44 notre notre PROPN bracis-33661 312 45 dame dame PROPN bracis-33661 312 46 hesburgh hesburgh PROPN bracis-33661 312 47 library library NOUN bracis-33661 312 48 ( ( PUNCT bracis-33661 312 49 3000184373 3000184373 NUM bracis-33661 312 50 ) ) PUNCT bracis-33661 313 1 © © ADP bracis-33661 313 2 2025 2025 NUM bracis-33661 313 3 springer springer NOUN bracis-33661 313 4 nature nature NOUN