基于线性估计的抑郁症脑磁图研究

Research on depression MEG based on linear estimator

目的为了研究抑郁症脑磁图的病理特征,使用信息熵对不同情绪刺激的脑磁图复杂度进行量化分析。方法采用线性估计的静态熵、条件熵和互信息,对正性、中性和负性情绪图片刺激下抑郁症患者脑磁图的特征进行对比性分析。结果在中性情绪图片刺激下,抑郁症脑磁图41个通道的静态熵和51个通道的互信息显著高于健康对照组(P<0.05),且均主要集中于额区和颞区。相比之下,条件熵在不同情绪刺激和不同维度下对抑郁症脑磁图的识别效果较差。当维度值改变时,对线性估计条件熵和互信息的区分效果影响较小。结论抑郁症患者较易受到情绪刺激影响,并且静态熵和互信息可以有效地表征抑郁症脑磁图非线性差异,为抑郁症的研究提供帮助。

Objective To study the pathological characteristics of depression magnetoencephalography(MEG),the complexity of MEG for different emotional stimuli was quantified by using entropy.Methods Comparative analysis of the characteristics of the MEG of depression patients under positive,neutral,and negative emotional picture stimuli was made by using linear estimator of static entropy,conditional entropy,and mutual information.Results Under neutral mood picture stimuli,static entropy of 41 channels and mutual information of 51 channels were significantly higher in depression MEG than in healthy controls(P<0.05),and both were mainly located in frontal and temporal regions.In comparison,conditional entropy was less effective in identifying depression MEG across emotional stimuli and dimensions.When dimensional values changed,there was less effect on the discriminative effect of linear estimator of conditional entropy and mutual information.Conclusions Depression is more susceptible to emotional stimuli,and static entropy and mutual information can effectively characterize nonlinear differences in depression MEG for the study of depression.