Mediation of Developmental Risk Factors for Psychosis by White Matter Microstructure in Young Adults With Psychotic Experiences

IMPORTANCE White matter (WM) abnormalities have been identified in schizophrenia at the earliest stages of the disorder. Individuals in the general population with psychotic experiences (PEs) may show similar changes, suggesting dysfunction due to aberrant neurodevelopment. Studying such people is a powerful means of understanding the nature of neurodevelopmental problems without the confound of clinical management and allows other potential risk factors associated with the schizophrenia spectrum to be taken into account. OBJECTIVES To compare WM microstructure and myelination in young adults with and without PEs identified from a population-based cohort using diffusion and relaxometry magnetic resonance imaging and to quantify potential mediating effects of WM on several known risk factors for psychosis. DESIGN, SETTING, AND PARTICIPANTS In this case-control study, participants were drawn from the UK Avon Longitudinal Study of Parents and Children. Psychotic experiences were assessed using a semistructured interview. Magnetic resonance imaging was carried out at age 20 years in 123 participants who had PEs and 124 individuals serving as controls. Participants with PEs were subdivided into those with operationally defined suspected PEs, definite PEs, and psychotic disorder. MAIN OUTCOMES AND MEASURES Diffusion tensor magnetic resonance imaging and relaxometry-derived myelin water fractions were used to measure WM microstructure and myelination, respectively. Differences in quantitative WM indices were assessed using tract-based spatial statistics. A binary model and a continuum-like ordinal model of PEs were tested. RESULTS Among the 123 participants who had PEs (mean [SE] age, 20.01 [0.004] years), 37 were male and 86 were female. Among the 124 controls (mean [SE] age, 20.11 [0.004] years), 49 were male and 76 were female. Fractional anisotropy in left frontomedial WM was significantly reduced in individuals with PEs (Montreal Neurological Institute [MNI] coordinates, -18, 37, -2; P = .0046). The ordinal model identified a similar but more widespread effect, with a corresponding increase in radial diffusivity (MNI coordinates, -15, 29, 21; P = .0042). Low birth weight (rho = -0.155; P = .015) and childhood IQ (rho = -0.188; P = .003) were associated with the presence of PEs. Results of mediation analysis were consistent with the association between birth weight (21.1% mediation effect; P = 6.20 x 10(-3)) and childhood IQ (7.9% mediation effect; P = .041) and by PEs being mediated by fractional anisotropy changes in these regions. CONCLUSIONS AND RELEVANCE The results of the study imply the presence of abnormal WM microstructure in young adults with PEs. The results are consistent with the hypothesis that neurodevelopmental factors cause alterations in the cellular composition of WM circuits critical to higher cognitive function. Such alterations may first manifest in childhood as reduced IQ and later contribute to PEs in early adulthood.