基于脑电信号和瞬时情感强度的情感角色划分

doi:10.16180/j.cnki.issn1007-7820.2025.05.007

摘要/Abstract

摘要：

现有情感计算研究使用固定情感标签训练一个情感分类器,但经历情感刺激个体的脑电信号(Electroencephalogram,EEG)和真实情感变化均为动态而非恒定的,不同个体由情感刺激触发的生理和主观动态反应不同。针对该问题,文中依据自采集的情感点击强度和公开数据集SEED-IV中的脑电数据,将被试者分为高低情感代入角色并生成了两类情感强度标签。使用脑电信号特征和情感强度标签在4种机器学习进行瞬时情感强度回归预测,根据回归结果分析了两类情感角色群体在两类情感强度标签上的适用性。回归结果表明,不同情感角色在不同情绪下的瞬时情感强度变化不同,情感角色在相对应的情感强度标签上能够展现更好的回归结果,划分情感角色能更好地辅助分析个体瞬时情感变化。

关键词: 机器学习, 瞬时情感强度, 情感角色划分, 高低情感代入, 情感强度标签, 情感强度回归, 脑电信号, 情感计算, 情感强度变化

Abstract:

Existing affective computing studies use fixed affective labels to train an affective classifier. However, the EEG(Electroencephalogram) and true affective changes of individuals experiencing emotional stimuli are dynamic rather than constant, and different individuals have different physiological and subjective dynamic responses triggered by emotional stimuli. To address these issues, this study uses self-collected data on emotional click intensity and EEG data from the public dataset SEED-IV to categorize subjects into high and low emotional roles,which generates two types of affective intensity labels. Instantaneous affective intensity regression prediction is performed using electroencephalogram signal features and affective intensity labels across four machine learning models. According to the regression results, the applicability of two kinds of affective role groups to two kinds of affective intensity labels is analyzed. The regression results show that different emotional roles have different changes in the instantaneous emotional intensity under different emotions, and emotional roles can show better regression results on the corresponding emotional intensity label, and the division of emotional roles can better assist the analysis of individual instantaneous emotional changes.

Key words: machine learning, instantaneous affective intensities, emotional roles classification, high and low affective into, affective intensity labels, affective intensity regression, EEG, affective computing, affective intensity change

中图分类号:

TP391

李汭钉, 甘开宇, 尹钟. 基于脑电信号和瞬时情感强度的情感角色划分[J]. 电子科技, 2025, 38(5): 46-52.

LI Ruiding, GAN Kaiyu, YIN Zhong. Emotional Role Classification Based on the EEG and Instantaneous Affective Intensities[J]. Electronic Science and Technology, 2025, 38(5): 46-52.

图/表 12

图1

表1

表2

表3

图2

表4

表5

表6

表7

图3

图4

图5

参考文献 22

[1]	Schmidt M, Wilhelmy S, Gross D. Retrospective diagnosis of mental illness: Past and present[J]. The Lancet Psychiatry, 2020, 7(1):14-16.
[2]	Yang L, Qin S F. A review of emotion recognition methods from keystroke,mouse and touchscreen dynamics[J]. IEEE Access, 2021, 28(9):162197-162213.
[3]	García-Martínez B, Martínez-Rodrigo A, Alcaraz R, et al. A review on nonlinear methods using electroencephalographic recordings for emotion recognition[J]. IEEE Transactions on Affective Computing, 2021, 12(3):801-820.
[4]	Zontone P, Affanni A, Bernardini R, et al. Cardriver's sympathetic reaction detection through electrodermal activity and electrocardiogram measurements[J]. IEEE Transactions on Biomedical Engineering, 2020, 67(12): 3413-3424.
[5]	Shukla J, Barreda-Ángeles M, Oliver J, et al. Feature extraction and selection for emotion recognition from electrodermal activity[J]. IEEE Transactions on Affective Computing, 2021, 12(4):857-869.
[6]	Wang Z, Wang Y, Hu C, et al. Transformers for EEG-based emotion recognition:A hierarchical spatial information learning model[J]. IEEE Sensors Journal, 2022, 22(5): 4359-4368.
[7]	Wang Z, Wang Y, Zhang J, et al. Spatial-temporal feature fusion neural network for EEG-based emotion recognition[J]. IEEE Transactions on Instrumentation and Measurement, 2022, 71(2):1-12.
[8]	Sun Z Q, Wang C L, Tian E G, et al. ECG signal classification via combining hand-engineered features with deep neural network features[J]. Multimedia Tools and Applications, 2021, 81(10):13467-13488.
[9]	Foong R, Ang K K, Quek C, et al. Assessment of the efficacy of EEG-based MI-BCI with visual feedback and EEG correlates of mental fatigue for upper-limb stroke rehabilitation[J]. IEEE Transactions on Biomedical Engineering, 2020, 67(3):786-795.
[10]	Bagchi S, Bathula D R. EEG-Conv transformer for single-trial EEG-based visual stimulus classification[J]. Pattern Recognition, 2022, 12(9):108-117.
[11]	Varsehi H, Firoozabadi S M P. An EEG channel selection method for motor imagery based brain-computer interface and neurofeedback using Granger causality[J]. Neural Networks, 2021, 13(3):193-206.
[12]	Xu S H, Fang J, Hu X P, et al. Emotion recognition from gait analyses:Current research and future directions[J]. IEEE Transactions on Computational Social Systems, 2022, 37(4):1-15.
[13]	Ezzameli K, Mahersia H. Emotion recognition from unimodal to multimodal analysis:A review[J]. Information Fusion, 2023, 99(2):101-117.
[14]	Alonso-Fernandez F, Raja K B, Raghavendra R, et al. Cross-sensor periocular biometrics in a global pandemic:Comparative benchmark and novel multialgorithmic approach[J]. Information Fusion, 2022, 83(4):110-130.
[15]	Zheng W L, Liu W, Lu Y, et al. EmotionMeter:A multimodal framework for recognizing human emotions[J]. IEEE Transactions on Cybernetics, 2019, 49(3):1110-1122.
[16]	Kwak N, Chong-Ho C. Input feature selection by mutual information based on Parzen window[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002, 24(12):1667-1671.
[17]	Yin Z, Zhao M, Wang Y, et al. Recognition of emotions using multimodal physiological signals and an ensemble deep learning model[J]. Computer Methods and Programs in Biomedicine, 2017, 140(1):93-110.
[18]	郑展鹏, 尹钟. 基于集成SDAE和EEG的跨被试认知工作负荷识别[J]. 电子科技, 2021, 34(3):48-53.
	Zheng Zhanpeng, Yin Zhong. Inter-subject recognition of cognitive workload based on ensemble SDAE and EEG[J]. Electronic Science and Technology, 2021, 34(3):48-53.
[19]	Li M, Lu B L. Emotion classification based on gamma-band EEG[C]. Minnesota: Annual International Conference of the IEEE Engineering in Medicine and Biology Society, 2009:1223-1226.
[20]	Bai Z, Liu J, Hou F, et al. Emotion recognition with residual network driven by spatial-frequency characteristics of EEG recorded from hearing-impaired adults in response to video clips[J]. Computers in Biology and Medicine, 2023, 152(2):106344-106350.
[21]	Zolezzi D M, Alonso-Valerdi L M, Ibarra-Zarate D I. EEG frequency band analysis in chronic neuropathic pain:A linear and nonlinear approach to classify pain severity[J]. Computer Methods and Programs in Biomedicine, 2023, 23(10):107349-107351.
[22]	Kılıç B, Aydın S. Classification of contrasting discrete emotional states indicated by EEG based graph theoretical network measures[J]. Neuroinformatics, 2022, 20(4):863-877. doi: 10.1007/s12021-022-09579-2 pmid: 35286574

视频编号	情绪类别标签	电影片段来源
1	1	《半音》
2	2	《见鬼》
3	3	《宝贝计划》
4	0	《舌尖上的中国S2EP4》
5	2	《童眼》
6	0	《舌尖上的中国S2EP6》1
7	0	《舌尖上的中国S2EP6》2
8	1	《我家有喜EP32》
9	0	《舌尖上的中国S2EP2》
10	1	《分手合约》
11	2	《笔仙2》
12	1	《亲爱的》
13	1	《唐山大地震》
14	1	《养父EP3》
15	2	《笔仙3》
16	3	《北京2022冬奥会申奥宣传片》
17	2	《盂兰神功》1
18	2	《盂兰神功》2
19	3	《我家有喜EP64》
20	3	《你是我的生命EP10》
21	0	《舌尖上的中国S1EP1》
22	3	《听说》
23	0	《舌尖上的中国S1EP2》
24	3	《我家有喜EP5》

情感角色群体	被试者序号
H	4、6、9、12、15
L	1、2、3、5、7、8、10、11、13、14

情感角色群体	被试者序号
H	1、2、7、13
L	3、4、5、6、8、9、10、11、12,14、15、16、17、18、19、20

群体	标签	中性	开心	伤心	恐惧	平均值
高情感代入角色群体	H	0.175 5	0.120 8	0.135 0	0.213 5	0.161 2
高情感代入角色群体	L	0.159 4	0.075 9	0.115 8	0.218 4	0.142 4
低情感代入角色群体	H	0.220 1	0.221 8	0.206 1	0.168 8	0.204 2
低情感代入角色群体	L	0.219 6	0.203 6	0.201 8	0.144 5	0.192 4

群体	标签	中性	开心	伤心	恐惧	平均值
高情感代入角色群体	H	0.186 8	0.135 3	0.120 3	0.134 9	0.144 3
高情感代入角色群体	L	0.092 5	0.094 4	0.147 8	0.197 5	0.133 1
低情感代入角色群体	H	0.151 2	0.175 1	0.156 7	0.170 8	0.163 5
低情感代入角色群体	L	0.158 5	0.174 8	0.165 3	0.153 5	0.163 0