Alterations in white matter evident before the onset of psychosis

Abstract

Background: Psychotic disorders are associated with widespread reductions in white matter (WM) integrity. However, the stage at which these abnormalities first appear and whether they are correlates of psychotic illness, as opposed to an increased vulnerability to psychosis, is unclear. We addressed these issues by using diffusion tensor imaging (DTI) to study subjects at ultra high risk (UHR) of psychosis before and after the onset of illness.

Methods: Thirty-two individuals at UHR for psychosis, 32 controls, and 15 patients with first-episode schizophrenia were studied using DTI. The UHR subjects and controls were re-scanned after 28 months. During this period, 8 UHR subjects had developed schizophrenia. Between-group differences in fractional anisotropy (FA) and diffusivity were evaluated cross sectionally and longitudinally using a nonparametric voxel-based analysis.

Results: At baseline, WM DTI properties were significantly different between the 3 groups (P < .001). Relative to controls, first-episode patients showed widespread reductions in FA and increases in diffusivity. DTI indices in the UHR group were intermediate relative to those in the other 2 groups. Longitudinal analysis revealed a significant group by time interaction in the left frontal WM (P < .001). In this region, there was a progressive reduction in FA in UHR subjects who developed psychosis that was not evident in UHR subjects who did not make a transition.

Conclusions: People at UHR for psychosis show alterations in WM qualitatively similar to, but less severe than, those in patients with schizophrenia. The onset of schizophrenia may be associated with a progressive reduction in the integrity of the frontal WM.

Fig. 1.

Detailed characteristics of the UHR group. *Age (mean ± SD): 22.75 ± 2.92; gender (male/females): 7/1. **Age (mean ± SD): 25.18 ± 3.70; gender (male/females): 4/1; 3 subjects were receiving antipsychotic medication (Quetiapine). †Age (mean ± SD): 23.67 ± 4.13; gender (male/females): 12/12. ‡Age (mean ± SD): 27.03 ± 4.33; gender (male/females): 7/10; 2 subjects were receiving antipsychotic medication (Risperidone, Quetiapine).

Fig. 2.

Linear changes in FA across three groups. Figure (A) displays areas of reduced fractional anisotropy (FEP < UHR < HC); images are radiologically oriented (participant's left is to the right). The figure (B and C) shows differences in FA for significant clusters (median ± range).

Fig. 3.

Linear changes in RD across 3 groups. Figure (A) displays areas of increased radial diffusivity (FEP > UHR > HC); images are radiologically oriented (participant's left is to the right). The figure (B) shows differences in RD for significant clusters (median ± range). Radial diffusivity is measured in 10⁻⁶ mm²/s.

Fig. 4.

Group × time interaction analysis results. Fractional anisotropy map (A) indicating the areas where there is an interaction between group membership and time between scans. Image is radiologically oriented (participant's left is to the right). Plot of FA values (median ± 95% CI) extracted from the significant cluster (B).