Dynamic Functional Network Connectivity Features Overlap Across Psychotic Disorders

Background
Dynamic functional network connectivity (dFNC) is a powerful approach that has shown the ability to differentiate patients and healthy control (HC). There is a need to better understand the relationship between schizophrenia (SZ), bipolar (BP), and schizoaffective disorder (SAD). Here we focus on evaluating dFNC features overlap across multiple psychotic disorders.
Methods
We used rest fMRI (rs-fMRI), demographics, and clinical information from the Bipolar–Schizophrenia Network on Intermediate Phenotypes (B-SNIP) cohort, including SZ (N=102), BP (N=102), SAD, (N=102) disorders, and HC (N=118) groups. We did not observe significant age and sex differences across groups. Using neuromark independent component analysis, followed by a sliding window approach, we estimated the dFNC for each participant. Next, we estimated three distinct dFNC states using the k-means clustering method. Then, we estimated the occupancy rate (OCR), the proportional amount of time each participant spent in each state, and compared them across HC and patient groups.
Results
Results showed that unliked BP, the OCR of the SZ and SAD group are significantly higher than that of HC in a state with relatively lower whole-brain functional connectivity (corrected p<0.01). Additionally, we found that SZ and SAD have lower OCR than HC in a state with relatively higher functional connectivity among sensory networks (corrected p<0.05).
Conclusions
We identified an overlap between SZ and SAD diseases on OCR estimated from dFNC, while we did not observe any overlap on this feature between BP and the other two conditions. Results provide useful insights into the relationship between brain function and mental disorders.