Automatic assessment of somatic depressive symptoms from physiological signals

点击翻译按钮获取中文摘要

To develop strategies for early prevention of MDD, it is crucial to identify biomarkers for depressive state and uncover the development model of MDD. Somatic depressive symptoms and the alterations of structure and functions in insula, the key region of brain interoceptive network (BIN), were reported as predictors of later development of MDD. However, it remained unclear whether the abnormalities of autonomic regulation and interoception are linked to the alterations of BIN in MDD. This two-year project aims to develop biomarkers for assessment of somatic symptoms and investigate the underlying neural mechanisms.In the first year, the neuromarkers for somatic symptoms were identified from magnetic resonance imaging data and linked to the abnormal cardiac autonomic regulation. The volume of 12 BIN regions including the cingulate, orbital frontal cortex, insula, and hippocampus were negatively correlated with somatic depressive symptoms. Among the 66 functional connectivity (FCs), a combination of 28 FCs achieved the highest prediction accuracy (correlation=0.805, p<0.01). Among them, 4 FCs between insula and anterior cingulate cortex (ACC) were significantly lower in MDD patients than in HC. Moreover, the reduced connectivity was correlated with decreased variability in inter-beat interval and decreased cardiac sympathetic and parasympathetic activity. The results suggest that alterations in the structure and function of the insula and ACC may underlie the somatic and autonomic abnormalities that play crucial roles in the development and remission of MDD.