Changes in resting-state functional connectivity in patients with psychophysiological insomnia after cognitive–behavioral therapy

Abstract

Psychophysiological insomnia (PI) is characterized by cognitive, emotional, and physiological hyperarousals. The effectiveness of cognitive–behavioral therapy for insomnia (CBTi) has supported the concept of PI as a psychological disorder. However, the recent expansion in neuroimaging techniques has identified cortical, emotional, and sensory arousal in insomnia patients by demonstrating functional changes as well as structural changes in the brains these patients. Taken together with the psychological (cognitive-behavioral) aspects of insomnia, the neurobiological hyperarousal motivates the hyperarousal theory of insomnia. Functional connectivity (FC) changes in the default mode network (DMN) and in emotional, limbic, and somatosensory areas are being revealed and provide substantial evidence for hyperarousal in insomnia patients. However, despite the clinical effectiveness of CBTi in insomnia patients, resting-state FC changes after CBTi have not been identified. The purpose of the present study was to compare the resting-state FC between PI patients and good sleepers (GS) and to investigate resting-state FC changes after CBTi in PI patients. The present study included 13 patients in the PI group (mean age: 51.0 ± 10.2 years) and 18 controls in the GS group (mean age: 42.7 ± 12.3 years). The FC between subcortical seed regions (caudate, putamen, pallidum, amygdala, thalamus, and hippocampus) and DMN regions (anterior cingulate cortex, posterior cingulate cortex, paracingulate cortex, inferior parietal cortex) were analyzed in relation to the voxels of the whole brain. The FC between DMN regions were also examined by region-to-region analysis. Next, FC was compared in the PI group before and after 5 weeks of CBTi. The results were thresholded at a false discovery rate (FDR)-corrected cluster level of q < 0.05

ke indicates the cluster size in voxels. In the subcortical seed region analyses, compared to the GS group, the PI group exhibited stronger FC between the thalamus and prefrontal cortex (q = 0.031, ke = 101 for right frontal gyrus and q = 0.041, ke = 78 for bilateral frontal poles) and between the pallidum and precuneus (q = 0.009, ke = 118), but weaker FC between the pallidum and angular gyrus (q = 0.006, ke = 113), the caudate and orbitofrontal cortex (q = 0.040, ke = 86), and the hippocampus and fusiform gyrus (q = 0.029, ke = 100). In the DMN seed region analyses, compared to the GS group, the PI group exhibited stronger FC between the paracingulate gyrus and fusiform gyrus (q = 0.006, ke = 130), lingual gyrus (q = 0.006, ke = 121), and the left amygdala (q = 0.028, ke = 79), and stronger FC between the anterior cingulate cortex and the lingual gyrus (q = 0.045, ke = 88). However, ROI-to-ROI analysis between DMN structures showed no significant difference in FC in the PI group compared to the GS group. After CBTi, in the subcortical seed region analyses, the PI group exhibited decreased FC between the thalamus and parietal cortex (q = 0.022, ke = 61), the putamen and motor cortices (q = 0.034, ke = 56 for the right superior frontal gyrus and q = 0.034, ke = 54 for the left supplementary motor area), and the amygdala and lingual gyrus (q = 0.028, ke = 61), but increased FC between the caudate and supramarginal gyrus (q = 0.038, ke = 58), the pallidum and orbitofrontal cortex (q < 0.001, ke = 146), and the hippocampus and frontal/parietal gyri (q =0.002, ke = 113 for the frontal pole, q =0.047, ke = 51 for the left supramarginal gyrus, and q =0.047, ke = 50 for the right paracingulate gyrus). In the DMN seed region analyses, the PI group exhibited decreased FC between the posterior cingulate cortex and occipital cortex (q = 0.033, ke = 70), between the paracingulate gyrus and occipital cortex (q = 0.022, ke = 74), between the angular gyrus and fusiform gyrus (q = 0.013, ke = 76), and between the anterior cingulate cortex and pericentral cortex level (q = 0.034, ke = 55 for the left precentral gyrus and q = 0.034, ke = 54 for the left postcentral gyrus), but increased FC between the supramarginal gyrus and precuneus (q = 0.001, ke = 129). ROI-to-ROI analysis between DMN structures showed no significant difference in FC after CBTi in the PI group. The present study found significant FC changes in the subcortical and DMN regions in relation to the whole brain. After CBTi, significant FC changes were observed in the PI group. These findings demonstrate the existence of hyperarousal in PI and the therapeutic effect of CBTi on this, and may provide insight into the neurobiological mechanisms of CBTi.