基于睡眠障碍患者的脑结构网络预测抑郁发病风险OA

Objective:To evaluate the predictive value of brain structural network characteristics for the risk of depression in patients with sleep disorders.Methods:A two-year prospective follow-up was performed in 129 patients with sleep disorders enrolled from Outpatient Departments of Neurology and Otolaryngology,Affiliated Yixing Hospital of Jiangsu University from January 2020 to December 2021.A total of 50 age-and gender-matched healthy controls were recruited by local community advertisements at the same time period.All subjects underwent 3.0 T MRI and brain structural networks were constructed.They were then followed up for 2 years.During the period,all patients were assessed for depression occurrence by Hamilton Depression Scale(HAMD)and divided into group of sleep disorders with depression and group of sleep disorders without depression accordingly.Differences in baseline information and brain structural networks were compared among the 3 groups,and independent influencing factors for depression in patients with sleep disorders were analyzed by multivariate Logistic regression.The predictive value of brain structural network indexes in depression in patients with sleep disorders was analyzed by receiver operator characteristic(ROC)curve.Partial correlation analysis was conducted to explore the correlation between brain structural network indexes and HAMD scores in patients with sleep disorders.Results:Compared with the healthy control group,the group of sleep disorders without depression and group of sleep disorders with depression had significantly lower global efficiency,and node efficiency of the left amygdala,right fusiform gyrus,left superior frontal gyrus and left hippocampus(all P＜0.05).Compared with the group of sleep disorders without depression,the group of sleep disorders with depression had significantly lower global efficiency,statistically higher clustering coefficient,and significantly lower node efficiency of the left hippocampus,left amygdala and right superior occipital gyrus(all P＜0.05).Logistic regression analysis revealed that global efficiency,nodal efficiency of the left hippocampus,and nodal efficiency of the left amygdala were independent influencing factors for depression in patients with sleep disorders(P＜0.05).The ROC curve analysis results indicated that the area under the curve for predicting depression in patients with sleep disorders by the regression model constructed with the above three factors was 0.882(95％CI:0.815-0.953,P＜0.01).Partial correlation analysis indicated that global efficiency,nodal efficiency of the left hippocampus,and nodal efficiency of the left amygdala in the group of sleep disorders with depression were negatively correlated with HAMD scores at the end of follow-up(r=-0.672,-0.618,-0.649,all P＜0.01).Conclusion:Baseline global efficiency,nodal efficiency of the left hippocampus,and nodal efficiency of the left amygdala could be used to effectively predict the risk of depression in patients with sleep disorders and are closely associated with depression severity.