Altered brain functional network segregation and integration in patients with insomnia disorder

Insomnia disorder (ID) is frequently accompanied by depressive symptoms, complicating clinical management. However, alterations in interregional brain interactions associated with depressive symptoms in patients with ID remain unclear. We enrolled 40 healthy controls (HCs), 43 patients with insomnia disorder without clinically significant depressive symptoms (ID-ND), and 20 patients with insomnia disorder with clinically significant depressive symptoms (ID-D), and collected clinical assessments and MRI data. The nested-spectral partition (NSP) method was applied to comprehensively characterize differences in functional brain networks among the three groups. Linear regression models and correlation analyses were further conducted to examine associations between clinical measures and neuroimaging metrics. The ID-D group showed more severe insomnia, depressive, and anxiety symptoms than the ID-ND group. ID-ND exhibited significantly greater whole-brain and network-level functional integration than HCs, and these findings were stable in stratified bootstrap analyses. At the regional level, ID-ND showed greater integration than HCs, mainly involving the prefrontal cortex, sensorimotor regions, insula, and parietal lobe. The functional integration within the prefrontal cortex, temporal gyrus, and fusiform gyrus was also higher in the ID-ND group than in the ID-D group, and these regional differences spatially corresponded with molecular maps of relatively high 5-HT1A, 5-HT2A and GABAA receptor distribution. In the ID-D group, depression severity was positively correlated with integration in the subcortical network, visual network, dorsal attention network, and somatosensory-motor network, but none of these survived FDR correction. Notably, the linear regression model revealed opposing association patterns of insomnia and depressive symptom severity with whole-brain functional segregation and integration in patients with ID. Our study investigated the neural network differences among ID subtypes, providing a foundation for understanding underlying mechanisms and informing intervention strategies.