融合共空间模式与脑网络特征的EEG抑郁识别

doi:10.3778/j.issn.1002-8331.2104-0329

摘要/Abstract

摘要： 提出共空间模式算法和脑网络拓扑属性融合的脑电信号（electroencephalography，EEG）特征，结合深度学习模型时序卷积网络（temporal convolution network，TCN）对抑郁组和对照组进行分类。根据相位锁值构建电极通道间相位同步性功能网络，分析不同频段下两种类别的功能连接模式。采用多特征融合方法将共空间模式特征和脑网络拓扑特征结合起来，最后结合Fisher score特征选择方法和分类器依赖结构，得到低维高效的特征子集并应用TCN进行分类。在抑郁数据集上的实验结果验证了所提策略的有效性。

关键词: 抑郁识别, 脑电信号（EEG）, 共空间模式, 时序卷积网络（TCN）, 特征选择

Abstract: This paper extracts the features of electroencephalography（EEG） signals based on common spatial pattern and brain connectivity. The depression group and the control group are recognized by using a deep learning model temporal convolution network（TCN） based on these EEG features. The phase synchronization functional network between channels is constructed according to the phase locking values, and the functional connection modes of two classes under different frequency bands are also analyzed. To contain more comprehensive information, the features between common spatial pattern and brain connectivity are combined. Finally, the Fisher score and classifier dependent structure are used for feature selection. As a result, these low-dimensional and efficient features are fed into a TCN classifier to detect depression. Experimental results on a depression dataset validate the effectiveness of the proposed strategy.

Key words: depression recognition, electroencephalography（EEG）, common spatial pattern, temporal convolution network（TCN）, feature selection

王怡忻, 朱湘茹, 杨利军. 融合共空间模式与脑网络特征的EEG抑郁识别[J]. 计算机工程与应用, 2022, 58(22): 150-158.

WANG Yixin, ZHU Xiangru, YANG Lijun. EEG Depression Recognition Based on Feature Fusion of Common Spatial Pattern and Brain Connectivity[J]. Computer Engineering and Applications, 2022, 58(22): 150-158.

参考文献

[1] SMITH K.Mental health：a world of depression[J].Nature News，2014，515（7526）：180-181.
[2] CUIJPERS P，STRANTENA V，ANDERSSON G，et al.Internet-administered cognitive behavior therapy for health problems：a systematic review[J].Journal of Behavioral Medicine，2008，31（2）：169-177.
[3] DELDIN P J，CHIU P.Cognitive restructuring and EEG in major depression[J].Biological Psychology，2005，70（3）：141-151.
[4] LU Q，WANG Y，LUO G，et al.Dynamic connectivity laterality of the amygdala under negative stimulus in depression：a MEG study[J].Neuroscience Letters，2013，547：42-47.
[5] HARVEY P O，FOSSATI P，POCHON J B，et al.Cognitive control，and brain resources in major depression：an fMRI study using the n-back task[J].Neuroimage，2005，26（3）：860-869.
[6] VIDEBECH P，RAVNKILDE B，KRISTENSEN S，et al.The Danish PET/depression project：poor verbal fluency performance despite normal prefrontal activation in patients with major depression[J].Psychiatry Research：Neuroimaging，2016，123（1）：49-63.
[7] PARVINNIA E，SABETI M，ZOLGHADRI J M，et al.Classification of EEG signals using adaptive weighted distance nearest neighbor algorithm[J].Journal of King Saud University Computer and Information Sciences，2014，26（1）：1-6.
[8] ZHU Y，WANG X，MATHIAK K，et al.Altered EEG oscillatory brain networks during music-listening in major depression[J].International Journal of Neural Systems，2021，31（3）：2150001.
[9] ZHENG W，YAO Z，XIE Y，et al.Identification of Alzheimer’s disease and mild cognitive impairment using networks constructed based on multiple morphological brain features[J].Biological Psychiatry：Cognitive Neuroscience and Neuroimaging，2018，3（10）：887-897.
[10] KOLES Z J.The quantitative extraction and topographic mapping of the abnormal components in the clinical EEG[J].Electroencephalography and Clinical Neurophy-
siology，1991，79（6）：440-447.
[11] RAMOSER H，GERKING J，PFURTSCHELLER G.Optimal spatial filtering of single trial EEG during imagined hand movement[J].IEEE Transactions on Rehabilitation Engineering，2000，8（4）：441-446.
[12] GUPTA R，FALK T H.Relevance vector classifier decision fusion and EEG graph-theoretic features for automatic affective state characterization[J].Neurocomputing，2016，174：875-884.
[13] GONUGUNTLA V，WANG Y，VELUVOLU K C.Phase synchrony in subject-specific reactive band of EEG for classification of motor imagery tasks[C]//Annual International Conference of the IEEE Engineering in Medicine and Biology Society，2013：2784-2787.
[14] LEE J，NICEWANDERW A.Thirteen ways to look at the correlation coefficient[J].The American Statistician，1988，42：59-66.
[15] SIKONJAM R，KONONENKO I.Theoretical and empirical analysis of Relief and ReliefF[J].Machine Learning，2003，53（1/2）：23-69.
[16] GAO J，WANG Z，YANG Y，et al.A novel approach for lie detection based on F-score and extreme learning machine[J].PLoS One，2013，8：e64704.
[17] ACHARYAUR，OHS L，HAGIWARA Y，et al.Automated EEG-based screening of depression using deep convolutional neural network[J].Computer Methods and Programs in Biomedicine，2018，161：103-113.
[18] LI X，ZHANG X，ZHU J，et al.Depression recognition using machine learning methods with different feature generation strategies[J].Artificial Intelligence in Medicine，2019，99：101696.
[19] ZHANG X，LI J，HOU K，et al.EEG-based depression detection using convolutional neural network with demographic attention mechanism[C]//2020 42nd Annual International Conference of the IEEE Engineering in Medicine & Biology Society，2020：128-133.
[20] LANG H，CHAN C W J，WANG Z.Automatic depression recognition using CNN with attention mechanism from videos[J].Neurocomputing，2021，422：165-175.
[21] HOCHREITER S，SCHMIDHUBER J.Long short-term memory[J].Neural Computation，1997，9（8）：1735-1780.
[22] CHE Z，PURUSHOTHAM S，CHO K，et al.Recurrent neural networks for multivariate time series with missing values[J].Scientific Reports，2018：6085.
[23] BAI S J，KOLTER Z，KOLTUN V.An empirical evaluation of generic convolutional and recurrent networks for sequence modeling[J].arXiv：1803.01271，2018.
[24] 李宇驰，李海芳，介丹，等.基于复杂网络的情感脑电相位同步性分析[J].计算机工程与应用，2017，53（18）：230-235.
LI Y C，LI H F，JIE D，et al.Phase synchronization analysis of emotional EEG based on complex network[J].Computer Engineering and Applications，2017，53（18）：230-235.
[25] RUBINOV M，SPORNS O.Complex network measures of brain connectivity：uses and interpretations[J].Neuroimage，2010，52（3）：1059-1069.
[26] OORD A V D，DIELEMAN S，ZEN H，et al.WaveNet：a generative model for raw audio[J].arXiv：1609.03499，2016.
[27] CHENY W，LIN C J.Combining SVMs with various feature selection strategies[M]//Feature extraction.Berlin，Heidelberg：Springer，2008：315-324.
[28] RADLOFF L S.The CES-D scale：a self-report depression scale for research in the general population[J].Applied Psychological Measurement，1977，1（3）：385-401.
[29] CELLA D，RILEY W，STONE A，et al.The patient-reported outcomes measurement information system（PROMIS） developed and tested its first wave of adult self-reported health outcome item banks：2005-2008[J].Journal of Clinical Epidemiology，2010，63（11）：79-94.
[30] FILIPPI M，VAN D，FORNITO A，et al.Assessment of system dysfunction in the brain through MRI-based connectomics[J].Lancet Neurology，2013，12（12）：1189-1199.
[31] MISEON S，CHANG H I，YONG W K，et al.Altered cortical functional network in major depressive disorder：a resting-state electroencephalogram study[J].Neuro Image Clinical，2018，19：1000-1007.
[32] FINGELOURT A A，RYTSALA H.Composition of brain oscillations in ongoing EEG during major depression disorder[J].Neuroscience Research，2006，56（2）：133-144.
[33] JAWORSLA N，BLIER P，FUSEEU W，et al.Alpha power，alpha asymmetry and anterior cingulate cortex activity in depressed males and females[J].Journal of Psychiatric Research，2012，46（11）：1483-1491.
[34] MOLLE M，MARSHALL L，FEHM H L，et al.EEG theta synchronization conjoined with alpha desynchronization indicate intentional encoding[J].European Journal of Neuroscience，2002，15（5）：923-928.
[35] PENG H，LONG F，DING C.Feature selection based on mutual information criteria of max-dependency，max-relevance，and min-redundancy[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2005，27（8）：1226-1238.