Global, mesoscale and local investigation into resting state fMRI brain network attributes of sleep quality

Poor sleep quality has been found to be associated with functional abnormalities in a few regions of the human brain. However, the brain is a dynamic network cooperation system, and it is necessary to study the relationship between sleep quality and the whole-brain network activity at multiple levels. A total of 115 college students underwent resting-state fMRI (rs-fMRI) and completed the Pittsburgh Sleep Quality Index (PSQI). Students were divided into good-quality sleepers (N = 65, PSQI<5) and poor-quality sleepers (N = 50, PSQI≥5). The fMRI data were analyzed using graph theory and machine learning methods to compare between-group differences in functional network topology at different levels. Global analysis shows the poor sleep quality group had lower small-worldness and higher characteristic path length. The mesoscale analysis demonstrates the subnetwork functions of the right middle frontal gyrus, bilateral superior parietal gyrus, bilateral caudate nucleus, and right superior temporal gyrus are important indicators that distinguish between the two groups. Local analysis shows the nodal degree of the left inferior occipital gyrus and left postcentral gyrus significantly differed between the two groups. Taken together, these findings deepen the macroscopic understanding of the relationship between sleep quality and resting functional network topology patterns.