Alterations in white matter microstructure in individuals at persistent risk for psychosis

Abstract

Abnormalities in brain white matter (WM) are reported in youth at-risk for psychosis. Yet, the neurodevelopmental time course of these abnormalities remains unclear. Thus, longitudinal diffusion-weighted imaging (DWI) was used to investigate WM abnormalities in youth at-risk for psychosis. A subset of individuals from the Philadelphia Neurodevelopmental Cohort (PNC) completed two DWI scans approximately 20 months apart. Youths were identified through structured interview as having subthreshold persistent psychosis risk symptoms (n = 46), and were compared to healthy typically developing participants (TD; n = 98). Analyses were conducted at voxelwise and regional levels. Nonlinear developmental patterns were examined using penalized splines within a generalized additive model. Compared to TD, youth with persistent psychosis risk symptoms had lower whole-brain WM fractional anisotropy (FA) and higher radial diffusivity (RD). Voxelwise analyses revealed clusters of significant WM abnormalities within the temporal and parietal lobes. Lower FA within the cingulum bundle of hippocampus and cerebrospinal tracts were the most robust deficits in individuals with persistent psychosis symptoms. These findings were consistent over two visits. Thus, it appears that WM abnormalities are present early in youth with persistent psychosis risk symptoms, however, there is little evidence to suggest that these features emerge in late adolescence or early adulthood. Future studies should seek to characterize WM abnormalities in younger individuals and follow individuals as subthreshold psychotic symptoms emerge.

Figure 1.

Flow diagram of Persistent Psychosis and Typically Developing youth as enrolled and followed as part of the Philadelphia Neurodevelopmental Cohort. Of the 1601 individuals imaged as part of the PNC, 452 individuals screened positive for psychosis spectrum (PS) symptoms and 498 were considered healthy. At follow-up, 208 PS and 210 healthy were re-enrolled. Of these, 106 PS has persistent or worsening symptoms, while 153 healthy were still considered healthy. Of those imaged at follow-up, 46 youth with persistent psychosis risk symptoms and 98 TD had high quality diffusion weighted imaging acquired. Note: Not all individuals at baseline or follow-up completed the DWI portion of the neuroimaging protocol^, or data quality was considered too poor for analysis.

Figure 2.

Mean (+/−95% CI) whole brain diffusion metrics in typically developing (TD) and Persistent Psychosis Risk. On average, those with persistent psychosis risk had lower average fractional anisotropy and higher radial diffusivity as compared to TD. Fitted values across the two time points are shown for each individual in the left panel for each metric. There were no group differences in mean or axial diffusivity. Analyses were corrected for linear and non-linear effects of age, and sex, race, and temporal signal-to-noise ratio. Age was centered for all analyses. *p<0.05, FDR corrected.

Figure 3.

Regional differences in fractional anisotropy (FA) in typically developing (TD) and Persistent Psychosis Risk. On average, those with Persistent Psychosis Risk had lower average fractional anisotropy in the Cerebrospinal Tracts (CST) and the Cingulum Bundle of the Hippocampus (CGH) as compared to TD. There were no group differences in regional MD, RD or AD. Analyses were corrected for linear and non-linear effects of age, and sex, race, and temporal signal-to-noise ratio. Age was centered for all analyses. *p<0.05, false discovery rate (FDR) corrected. Mean (+/− 95% CI).