White Matter Microstructure and Sleep-wake Disturbances in Individuals at Ultra-high Risk of Psychosis

Aim: White matter changes in individuals at ultra-high risk for psychosis (UHR) may be involved in the transition to psychosis. Sleep-wake disturbances commonly precede the first psychotic episode and predict development of psychosis. We examined associations between white matter microstructure and sleep-wake disturbances in UHR individuals compared to healthy controls (HC), as well as explored the confounding effect of medication, substance use, and level of psychopathology.

Methods: Sixty-four UHR individuals and 35 HC underwent clinical interviews and diffusion weighted imaging. Group differences on global and callosal mean fractional anisotropy (FA) was tested using general linear modeling. Sleep-wake disturbances were evaluated using the subjective measures disturbed sleep index (DSI) and disturbed awakening index (AWI) from the Karolinska Sleep Questionnaire, supported by objective sleep measures from one-night actigraphy. The primary analyses comprised partial correlation analyses between global FA/callosal FA and sleep-wake measures. Secondary analyses investigated multivariate patterns of covariance between measures of sleep-wake disturbances and FA in 48 white matter regions of interest using partial least square correlations.

Results: Ultra-high risk for psychosis individuals displayed lower global FA ( = 14.56, < 0.001) and lower callosal FA ( = 11.34, = 0.001) compared to HC. Subjective sleep-wake disturbances were significantly higher among the UHR individuals (DSI: = 27.59, < 0.001, AWI: = 36.42, < 0.001). Lower callosal FA was correlated with increased wake after sleep onset ( = -0.34, = 0.011) and increased sleep fragmentation index ( = -0.31, = 0.019) in UHR individuals. Multivariate analyses identified a pattern of covariance in regional FA which were associated with DSI and AWI in UHR individuals ( = 0.028), but not in HC. Substance use, sleep medication and antipsychotic medication did not significantly confound these associations. The association with objective sleep-wake measures was sustained when controlling for level of depressive and UHR symptoms, but symptom level confounded the covariation between FA and subjective sleep-wake measures in the multivariate analyses.

Conclusion: Compromised callosal microstructure in UHR individuals was related to objectively observed disruptions in sleep-wake functioning. Lower FA in ventrally located regions was associated with subjectively measured sleep-wake disturbances and was partly explained by psychopathology. These findings call for further investigation of sleep disturbances as a potential treatment target.