Characterising grey-white matter relationships in recent-onset psychosis and its association with cognitive function

Abstract

ABSTRACTIndividuals with recent-onset psychosis (ROP) present widespread grey matter (GM) reductions and white matter (WM) abnormalities. However, relationships between GM and WM changes and their association with cognitive impairment, a key symptom of ROP, are unclear. Using a multiblock partial least squares correlation (MB-PLS-C) analysis, we examined multivariate GM-WM relationships and their association with cognitive abilities in ROP. We used T1 and diffusion-weighted images from 71 non-affective ROP individuals (age 22.1±3.2) and 71 matched controls. We performed MB-PLS-C between GM thickness and WM fractional anisotropy (FA) and between GM surface area and WM FA to identify multivariate GM-WM patterns and analysed correlations between these patterns and cognitive abilities. MB-PLS-C identified a ‘GM thickness’-‘WM FA’ pattern representing group differences, explaining 12.38% of the variance and associated with frontal and temporal GM regions and seven WM tracts, including the corticospinal tract. MB-PLS-C also identified a ‘GM surface area’-‘WM FA’ pattern showing group differences, explaining 18.92% and related with cingulate, frontal, temporal, and parietal GM regions and 15 WM tracts, including the inferior cerebellar peduncle. The ‘GM thickness’-‘WM FA’ pattern describing group differences was significantly correlated with processing speed in ROP. There was no association between cognition and the ‘GM surface area’-‘WM FA’ pattern. MB-PLS-C identified differential whole-brain GM-WM relationships, indicating a potential signature of brain alterations in ROP. Our findings of a relationship between cognitive function and GM-WM patterns for GM thickness rather than for surface area have implications for our understanding of brain-behaviour relationships neurodevelopmentally in psychosis.