基于跨被试和跨会话的脑电信号情感识别方法
打开文本图片集
关键词：脑电信号；领域自适应方法；多头自注意力机制；长短期记忆网络；情感识别中图分类号：TP391 文献标志码：A 文章编号：1001-3695（2025）10-018-3034-07doi： 10.19734/j .issn. 1001-3695.2025.03.0062
Cross-subject and cross-session EEG-based approach to emotion recognition
Shi Pengcheng"，b，Wang Hailong a，b† ，Liu Lin a，b （20 （a.Colegefedoeceobalit ，China）
Abstract：Electroencephalographic（EEG）signalisacritical indicatorforrecognizing emotional states，andemploying EEGbasedemotionrecognitionconstitutesakeycomponentinthediagnosisandrehabilitationof psychiatricdisorders.Mostcuret researchinthefieldofemotionrecognitionfocusesonthespatial-temporalinformationofEEGsignalsanddoesnotadequately accountforthevariabilityofEEGsgnalsacrossdiferent individuals（crossubject）andacrosdierenttimeperiods（rosssession）.Domainadaptationmethods effectivelyaddressthese kinds of problems.This paper proposed amulti-source domainadaptation methodforEEG-basedemotionrecognition.Ittransferedeach sourcedomainseparatelytothe targetdomain， therebyavoiding interferencefrom multipleEEGsourcedomains.Firstly，itintroducedalinear-model-basedsharedfeatureextractortoobtainlow-leveldomain-invariantfeatures.Next，itincorporatedamulti-headself-attention-baseddomain-specific featureextractortocapturedomain-specificcharacteristics.Throughafeatureprocessingmechanism，itprogressvelyeliminateddomain-specificinterferencefactorssuchasphysiological diferencesandemotionalfluctuations.Finally，itemployeda long short-termmemory-baseddomai-specificclassfertocapturethetemporaldependenciesintheinputdatanhancing the model's ability to represent complex features.This paper conducted experiments on the SEED and SEED-IV datasets.For cross-subject evaluation，the accuracies are 90.02% and 91.57% ，respectively. For cross-session evaluation，the accuracies are 67.09% and 69.27% ，respectively. The experiments show that the proposed method outperforms several baseline models inbothcro-subjectandcrosssessionemotionrecogitiontass，fullyvalidatingteodel’sfectivenessandgeneraation capability.
Key words：EEGsignal；domain-adaptivemethod;multi-headdself-atentionmechanism；longshort-term memorynetwork; emotion recognition
0 引言
情感是一种综合、复杂的生理和心理状态，可以通过生理和行为信号来表征。（剩余17313字）
