Causality-Driven Convolutional Manifold Attention Network for Electroencephalogram Signal Decoding

Deep learning-based methods have achieved remarkable success in brain-computer interfaces (BCIs). However, its inherent assumption of independent and identically distributed (i.i.d.) data renders it vulnerable to out-of distribution (OOD) scenarios. To address this limitation, the present study prop...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:IEEE transactions on pattern analysis and machine intelligence Ročník PP; s. 1 - 12
Hlavní autoři: Lu, Bin, Chen, Junxiang, Wang, Fuwang, Wen, Guilin, Fu, Rongrong, Hua, Changchun
Médium: Journal Article
Jazyk:angličtina
Vydáno: United States IEEE 27.10.2025
Témata:
ISSN:0162-8828, 1939-3539, 2160-9292, 1939-3539
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract Deep learning-based methods have achieved remarkable success in brain-computer interfaces (BCIs). However, its inherent assumption of independent and identically distributed (i.i.d.) data renders it vulnerable to out-of distribution (OOD) scenarios. To address this limitation, the present study proposed a causality-driven convolutional manifold attention network (CD-CMAN) that learned invariant representations from electroencephalogram (EEG) signals to enhance OOD generalization. The framework began with a spatiotemporal convolution module to extract rich temporal and spatial features. Guided by the defined structural causal model and leveraging the strengths of Riemannian geometry and deep learning, dual latent encoders with manifold attention units were crafted to explicitly separate spatiotemporal feature maps into semantic and variation latent factors. A reconstruction module with a dedicated loss was implemented to ensure these factors retaining informative, while the Hilbert-Schmidt independence criterion (HSIC) was introduced to enforce their statistical independence. Further, a variational information bottleneck and gradient reversal layer were incorporated to compress and disentangle the semantic and variation factors. Evaluations on two public datasets under both subject-dependent and subject independent settings demonstrated that CD-CMAN consistently outperforms comparative baselines. These findings suggest that the proposed model could provide a new solution for the practical application of BCI technology.
AbstractList Deep learning-based methods have achieved remarkable success in brain-computer interfaces (BCIs). However, its inherent assumption of independent and identically distributed (i.i.d.) data renders it vulnerable to out-of distribution (OOD) scenarios. To address this limitation, the present study proposed a causality-driven convolutional manifold attention network (CD-CMAN) that learned invariant representations from electroencephalogram (EEG) signals to enhance OOD generalization. The framework began with a spatiotemporal convolution module to extract rich temporal and spatial features. Guided by the defined structural causal model and leveraging the strengths of Riemannian geometry and deep learning, dual latent encoders with manifold attention units were crafted to explicitly separate spatiotemporal feature maps into semantic and variation latent factors. A reconstruction module with a dedicated loss was implemented to ensure these factors retaining informative, while the Hilbert-Schmidt independence criterion (HSIC) was introduced to enforce their statistical independence. Further, a variational information bottleneck and gradient reversal layer were incorporated to compress and disentangle the semantic and variation factors. Evaluations on two public datasets under both subject-dependent and subject independent settings demonstrated that CD-CMAN consistently outperforms comparative baselines. These findings suggest that the proposed model could provide a new solution for the practical application of BCI technology.
Deep learning-based methods have achieved remarkable success in brain-computer interfaces (BCIs). However, its inherent assumption of independent and identically distributed (i.i.d.) data renders it vulnerable to out-of distribution (OOD) scenarios. To address this limitation, the present study proposed a causality-driven convolutional manifold attention network (CD-CMAN) that learned invariant representations from electroencephalogram (EEG) signals to enhance OOD generalization. The framework began with a spatiotemporal convolution module to extract rich temporal and spatial features. Guided by the defined structural causal model and leveraging the strengths of Riemannian geometry and deep learning, dual latent encoders with manifold attention units were crafted to explicitly separate spatiotemporal feature maps into semantic and variation latent factors. A reconstruction module with a dedicated loss was implemented to ensure these factors retaining informative, while the Hilbert-Schmidt independence criterion (HSIC) was introduced to enforce their statistical independence. Further, a variational information bottleneck and gradient reversal layer were incorporated to compress and disentangle the semantic and variation factors. Evaluations on two public datasets under both subject-dependent and subject independent settings demonstrated that CD-CMAN consistently outperforms comparative baselines. These findings suggest that the proposed model could provide a new solution for the practical application of BCI technology.Deep learning-based methods have achieved remarkable success in brain-computer interfaces (BCIs). However, its inherent assumption of independent and identically distributed (i.i.d.) data renders it vulnerable to out-of distribution (OOD) scenarios. To address this limitation, the present study proposed a causality-driven convolutional manifold attention network (CD-CMAN) that learned invariant representations from electroencephalogram (EEG) signals to enhance OOD generalization. The framework began with a spatiotemporal convolution module to extract rich temporal and spatial features. Guided by the defined structural causal model and leveraging the strengths of Riemannian geometry and deep learning, dual latent encoders with manifold attention units were crafted to explicitly separate spatiotemporal feature maps into semantic and variation latent factors. A reconstruction module with a dedicated loss was implemented to ensure these factors retaining informative, while the Hilbert-Schmidt independence criterion (HSIC) was introduced to enforce their statistical independence. Further, a variational information bottleneck and gradient reversal layer were incorporated to compress and disentangle the semantic and variation factors. Evaluations on two public datasets under both subject-dependent and subject independent settings demonstrated that CD-CMAN consistently outperforms comparative baselines. These findings suggest that the proposed model could provide a new solution for the practical application of BCI technology.
Author Chen, Junxiang
Wen, Guilin
Hua, Changchun
Wang, Fuwang
Lu, Bin
Fu, Rongrong
Author_xml – sequence: 1
  givenname: Bin
  orcidid: 0009-0000-3678-1214
  surname: Lu
  fullname: Lu, Bin
  email: lubin@stumail.ysu.edu.cn
  organization: School of Electrical Engineering, Yanshan University, Qinhuangdao, China
– sequence: 2
  givenname: Junxiang
  orcidid: 0000-0002-8897-754X
  surname: Chen
  fullname: Chen, Junxiang
  email: juc91@pitt.edu
  organization: Department of Biomedical Informatics, University of Pittsburgh, Pittsburgh, PA, USA
– sequence: 3
  givenname: Fuwang
  surname: Wang
  fullname: Wang, Fuwang
  email: 20152622@neepu.edu.cn
  organization: School of Mechanical Engineering, Northeast Electric Power University, Jilin, China
– sequence: 4
  givenname: Guilin
  orcidid: 0000-0001-5940-2520
  surname: Wen
  fullname: Wen, Guilin
  email: glwen@ysu.edu.cn
  organization: School of Mechanical Engineering, Yanshan University, Qinhuangdao, China
– sequence: 5
  givenname: Rongrong
  orcidid: 0000-0003-0350-7619
  surname: Fu
  fullname: Fu, Rongrong
  email: frr1102@aliyun.com
  organization: School of Electrical Engineering, Yanshan University, Qinhuangdao, China
– sequence: 6
  givenname: Changchun
  orcidid: 0000-0001-6311-2112
  surname: Hua
  fullname: Hua, Changchun
  email: cch@ysu.edu.cn
  organization: School of Electrical Engineering, Yanshan University, Qinhuangdao, China
BackLink https://www.ncbi.nlm.nih.gov/pubmed/41144414$$D View this record in MEDLINE/PubMed
BookMark eNpFkMlOwzAQQC0EgrL8AEIoRy4pHsdx7GNVVolNAs6Rk0yKwbWLnYD4expa4DTS6L2R5u2STecdEnIIdAxA1enTw-T2eswoy8eZYLnIYIOMGAiaKqbYJhlRECyVkskdshvjK6XAc5ptkx0OwDkHPiLVVPdRW9N9pWfBfKBLpt59eNt3xjttk1vtTOttk0y6Dt2wTO6w-_ThLWl9SM4t1l3w6GpcvGjrZ0HPk0czG9QzrH1j3GyfbLXaRjxYzz3yfHH-NL1Kb-4vr6eTm7QGxbq0kbpohZBFA1poJVVV1bRoCo0MWJO3Sg4LaGVLBS-gAVVnmayozilf_kuzPXKyursI_r3H2JVzE2u0Vjv0fSwzJnJFcwCxRI_XaF_NsSkXwcx1-Cp_uywBtgLq4GMM2P4hQMshfvkTvxzil-v4S-loJRlE_BeAQZFzkX0Dfq6Axg
CODEN ITPIDJ
ContentType Journal Article
DBID 97E
RIA
RIE
AAYXX
CITATION
NPM
7X8
DOI 10.1109/TPAMI.2025.3625631
DatabaseName IEEE All-Society Periodicals Package (ASPP) 2005–Present
IEEE All-Society Periodicals Package (ASPP) 1998–Present
IEEE Electronic Library (IEL)
CrossRef
PubMed
MEDLINE - Academic
DatabaseTitle CrossRef
PubMed
MEDLINE - Academic
DatabaseTitleList
MEDLINE - Academic
PubMed
Database_xml – sequence: 1
  dbid: NPM
  name: PubMed
  url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed
  sourceTypes: Index Database
– sequence: 2
  dbid: RIE
  name: IEEE Electronic Library (IEL)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
– sequence: 3
  dbid: 7X8
  name: MEDLINE - Academic
  url: https://search.proquest.com/medline
  sourceTypes: Aggregation Database
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
Computer Science
EISSN 2160-9292
1939-3539
EndPage 12
ExternalDocumentID 41144414
10_1109_TPAMI_2025_3625631
11217546
Genre orig-research
Journal Article
GroupedDBID ---
-DZ
-~X
.DC
0R~
29I
4.4
53G
5GY
6IK
97E
AAJGR
AASAJ
AAWTH
ABAZT
ABQJQ
ABVLG
ACGFO
ACGFS
ACIWK
ACNCT
AENEX
AGQYO
AHBIQ
AKJIK
AKQYR
ALMA_UNASSIGNED_HOLDINGS
ASUFR
ATWAV
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CS3
DU5
E.L
EBS
EJD
F5P
HZ~
IEDLZ
IFIPE
IPLJI
JAVBF
LAI
M43
MS~
O9-
OCL
P2P
PQQKQ
RIA
RIE
RNS
RXW
TAE
TN5
UHB
~02
5VS
9M8
AAYXX
ABFSI
ADRHT
AETEA
AETIX
AGSQL
AI.
AIBXA
ALLEH
CITATION
FA8
H~9
IBMZZ
ICLAB
IFJZH
RNI
RZB
VH1
NPM
7X8
ID FETCH-LOGICAL-c192t-d8a7f6687d1a6a989bbc07d7ae212d5f98bbc01f8f06471d19c338b0a50492903
IEDL.DBID RIE
ISSN 0162-8828
1939-3539
IngestDate Wed Oct 29 11:34:17 EDT 2025
Wed Oct 29 02:22:14 EDT 2025
Sat Nov 29 06:59:19 EST 2025
Wed Nov 05 07:10:40 EST 2025
IsPeerReviewed true
IsScholarly true
Language English
License https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html
https://doi.org/10.15223/policy-029
https://doi.org/10.15223/policy-037
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c192t-d8a7f6687d1a6a989bbc07d7ae212d5f98bbc01f8f06471d19c338b0a50492903
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ORCID 0000-0002-8897-754X
0000-0003-0350-7619
0009-0000-3678-1214
0000-0001-5940-2520
0000-0001-6311-2112
PMID 41144414
PQID 3265905116
PQPubID 23479
PageCount 12
ParticipantIDs proquest_miscellaneous_3265905116
pubmed_primary_41144414
ieee_primary_11217546
crossref_primary_10_1109_TPAMI_2025_3625631
PublicationCentury 2000
PublicationDate 2025-Oct-27
PublicationDateYYYYMMDD 2025-10-27
PublicationDate_xml – month: 10
  year: 2025
  text: 2025-Oct-27
  day: 27
PublicationDecade 2020
PublicationPlace United States
PublicationPlace_xml – name: United States
PublicationTitle IEEE transactions on pattern analysis and machine intelligence
PublicationTitleAbbrev TPAMI
PublicationTitleAlternate IEEE Trans Pattern Anal Mach Intell
PublicationYear 2025
Publisher IEEE
Publisher_xml – name: IEEE
SSID ssj0014503
Score 2.4926803
Snippet Deep learning-based methods have achieved remarkable success in brain-computer interfaces (BCIs). However, its inherent assumption of independent and...
SourceID proquest
pubmed
crossref
ieee
SourceType Aggregation Database
Index Database
Publisher
StartPage 1
SubjectTerms Brain modeling
Causality
Decoding
Deep learning
EEG
Electroencephalography
Feature extraction
Invariant Representation
Manifolds
Motors
Reliability
Riemannian Manifold
Semantics
Symmetric matrices
Title Causality-Driven Convolutional Manifold Attention Network for Electroencephalogram Signal Decoding
URI https://ieeexplore.ieee.org/document/11217546
https://www.ncbi.nlm.nih.gov/pubmed/41144414
https://www.proquest.com/docview/3265905116
Volume PP
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVIEE
  databaseName: IEEE Electronic Library (IEL)
  customDbUrl:
  eissn: 2160-9292
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0014503
  issn: 0162-8828
  databaseCode: RIE
  dateStart: 19790101
  isFulltext: true
  titleUrlDefault: https://ieeexplore.ieee.org/
  providerName: IEEE
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LT8MwDLZg4gAH3o_xUpC4oULfSY7TAMFh0yRA2q1K0xQmQTtt3X4_dtryOHDgVlVJWsVO_H1xbANcCqlkkAaxwzOunVB7gYNcJXRMqJDHuVK4QtliE3w4FOOxHDXB6jYWxhhjL5-Za3q0vvys1As6KrtBbIDWLoxXYZVzXgdrfbkMwsiWQUYIg0sceUQbIePKm-dRb_CIXNCPrnG_juKA6sOEyAQQC4S_DJKtsPI32LRG537rn7-7DZsNumS9Wh12YMUUu7DVVm5gzULehY0faQj3IO2rxdziced2Rrsf65fFstFJHG6giklevmesV1X17Ug2rG-PM4S87K6upEMjT9-UzYH9wZ4mr9T1FtktWcd9eLm_e-4_OE3tBUcj5qucTCiex7HgmadiJYVMU-2iPJVBW5dFuRT0wstFTtGqXuZJjWQ3dVWElMOXbnAAnaIszBEw10-xW-5xndGBUyh0lPppYChml3LJdOGqFUAyrVNsJJaauDKxkktIckkjuS7s00x_t2wmuQsXrdASXCHk9lCFKRfzBAFqRFnIPGxzWEvzq3erBMd_jHoC6_RxMlY-P4VONVuYM1jTy2oyn52jGo7FuVXDT8aW1zk
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LT-MwEB4tXSSWA2-W7vIwEjeUkoeT2MeqgEDQCokicYscx4FKkKA25ffvjJMU9sCBWxTZluUZe77PngfAiZBKBmkQOXEWa4drL3CQq3DHcIU8zpXCFcoWm4hHI_H4KO-aYHUbC2OMsc5npkef9i0_K_WcrsrOEBugtePREvwMOfe9Olxr8WjAQ1sIGUEMbnJkEm2MjCvPxnf94TWyQT_s4YkdRgFViOHIBRAN8P9Mkq2x8jXctGbncv2bE96AtQZfsn6tEJvwwxRbsN7WbmDNVt6C1U-JCLchHaj5zCJy53xK5x8blMV7o5U43FAVk7x8yVi_qmr_SDaq_ccZgl52UdfSoZHfnpXNgv3K7idP1PUc-S3Zxx14uLwYD66cpvqCoxH1VU4mVJxHkYgzT0VKCpmm2kWJKoPWLgtzKeiHl4uc4lW9zJMa6W7qqhBJhy_dYBc6RVmYPWCun2K33It1RldOXOgw9dPAUNQuZZPpwmkrgOStTrKRWHLiysRKLiHJJY3kurBDK_3RslnkLhy3Qktwj9DDhypMOZ8lCFFDykPmYZvftTQXvVsl-PPFqEewcjUe3ia316Obv_CLJkKmy4_3oVNN5-YAlvV7NZlND60y_gMNvdmY
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Causality-Driven+Convolutional+Manifold+Attention+Network+for+Electroencephalogram+Signal+Decoding&rft.jtitle=IEEE+transactions+on+pattern+analysis+and+machine+intelligence&rft.au=Lu%2C+Bin&rft.au=Chen%2C+Junxiang&rft.au=Wang%2C+Fuwang&rft.au=Wen%2C+Guilin&rft.date=2025-10-27&rft.pub=IEEE&rft.issn=0162-8828&rft.spage=1&rft.epage=12&rft_id=info:doi/10.1109%2FTPAMI.2025.3625631&rft_id=info%3Apmid%2F41144414&rft.externalDocID=11217546
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0162-8828&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0162-8828&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0162-8828&client=summon