Alterations of functional connectivity network related to working memory in ultra-high risk for psychosis and first-episode schizophrenia

Abstract

Higher cognitive functions require functional interactions between multiple specialized neuronal networks in the brain. Disturbances in functional connectivity have been proposed as a major pathophysiological mechanism for schizophrenia (SPR). In the present study, functional connectivity networks were explored using small-worldness and synchronization likelihood (SL) during a working memory task. Thirteen first-episode SPR patients, 11 UHR individuals, 13 healthy controls were recruited. EEG data were collected during a verbal 0- and 2-back task in all participants. For SL, there was significantly different pattern of change in UHR subjects and SPR patients as the working memory load varied (significant working memory load by group interaction, theta: p<0.001, F=21.768; alpha: p<0.001, F=22.779; beta: p<0.001, F=30.318, and gamma: p<0.001, F=17.452). Specifically, while SL significantly decreased as the working memory load increased in the controls (theta: p=0.017, alpha: p=0.003, beta: p=0.004, gamma: p=0.005), these working memory load effect was not observed in UHR subjects. In SPR patients, SL was significantly higher than controls or UHR during the 0-back task, but decreased as the working memory load increased (significant working memory load effect, theta: p<0.001, alpha: <0.001, beta: <0.001, gamma: <0.001). For small-worldness, a trend of the group effect was seen in the theta band (p=0.088, F=2.616), with decreased small-worldness in SPR patients compared to controls (p=0.031) as resulted from the pairwise comparisons. In correlation analyses, increased SL in UHR subjects was associated with lower working memory performance, lower neurocognitive function and more severe symptoms. These findings collectively support the disconnection hypothesis of schizophrenia and suggest an alteration of functional connectivity in the prodromal phase. These changes may be associated with impairments in functional trimming needed for modular specialization of brain networks for higher cognitive demands.