The Relationship Between White Matter Microstructure and General Cognitive Ability in Patients With Schizophrenia and Healthy Participants in the ENIGMA Consortium

Abstract

Objective: Schizophrenia has recently been associated with widespread white matter microstructural abnormalities, but the functional effects of these abnormalities remain unclear. Widespread heterogeneity of results from studies published to date preclude any definitive characterization of the relationship between white matter and cognitive performance in schizophrenia. Given the relevance of deficits in cognitive function to predicting social and functional outcomes in schizophrenia, the authors carried out a meta-analysis of available data through the ENIGMA Consortium, using a common analysis pipeline, to elucidate the relationship between white matter microstructure and a measure of general cognitive performance, IQ, in patients with schizophrenia and healthy participants.

Methods: The meta-analysis included 760 patients with schizophrenia and 957 healthy participants from 11 participating ENIGMA Consortium sites. For each site, principal component analysis was used to calculate both a global fractional anisotropy component (gFA) and a fractional anisotropy component for six long association tracts (LA-gFA) previously associated with cognition.

Results: Meta-analyses of regression results indicated that gFA accounted for a significant amount of variation in cognition in the full sample (effect size [Hedges' g]=0.27, CI=0.17-0.36), with similar effects sizes observed for both the patient (effect size=0.20, CI=0.05-0.35) and healthy participant groups (effect size=0.32, CI=0.18-0.45). Comparable patterns of association were also observed between LA-gFA and cognition for the full sample (effect size=0.28, CI=0.18-0.37), the patient group (effect size=0.23, CI=0.09-0.38), and the healthy participant group (effect size=0.31, CI=0.18-0.44).

Conclusions: This study provides robust evidence that cognitive ability is associated with global structural connectivity, with higher fractional anisotropy associated with higher IQ. This association was independent of diagnosis; while schizophrenia patients tended to have lower fractional anisotropy and lower IQ than healthy participants, the comparable size of effect in each group suggested a more general, rather than disease-specific, pattern of association.

Keywords: Cognition; ENIGMA; Intelligence; Meta-Analysis; Schizophrenia; White Matter.

FIGURE 1.. Scree plots for gFA and LA-gFA principal component analyses in a study of white matter microstructure and cognitive ability in schizophrenia^a

^a gFA=global fractional anisotropy component; LA-gFA=fractional anisotropy component for six long association tracts; ASRB=Australian Schizophrenia Research Bank; COBRE=Center for Biomedical Research Excellence; HUBIN=Human Brain Informatics; MCIC=MIND Clinical Imaging Consortium; MPRC=Maryland Psychiatric Research Center; TOP=Thematically Organized Psychosis (TOP) NORMENT Research Study.

FIGURE 2.. Forest plot for gFA meta-analysis in a study of white matter microstructure and cognitive ability in schizophrenia^a

^aThere was no significant difference between the observed effect size in patients compared with control participants (χ²=1.3, p=0.26). Effect values indicated with a vertical line are Hedges’ g group summaries for patients and control groups separately; the diamond represents summary statistics for the full sample. gFA=global fractional anisotropy component; ASRB=Australian Schizophrenia Research Bank; COBRE=Center for Biomedical Research Excellence; EDIN=Edinburgh; HUBIN=Human Brain Informatics; MCIC=MIND Clinical Imaging Consortium; MPRC=Maryland Psychiatric Research Center; TOP=Thematically Organized Psychosis (TOP) NORMENT Research Study. In the lower half of the figure, the “p” appended to site names indicates the values relating to the patient cohort for each sample.

FIGURE 3.. Forest plot for LA-gFA meta-analysis in a study of white matter microstructure and cognitive ability in schizophrenia^a

^aThere was no significant difference between the observed effect size in patients compared with control participants (χ²=0.55, p=0.46). Effect values indicated with a vertical line are Hedges’ g group summaries for patients and control groups separately; the diamond represents summary statistics for the full sample. LA-gFA=fractional anisotropy component for six long association tracts; ASRB=Australian Schizophrenia Research Bank; COBRE=Center for Biomedical Research Excellence; EDIN=Edinburgh; HUBIN=Human Brain Informatics; MCIC=MIND Clinical Imaging Consortium; MPRC=Maryland Psychiatric Research Center; TOP=Thematically Organized Psychosis (TOP) NORMENT Research Study. In the lower half of the figure, the “p” appended to site names indicates the values relating to the patient cohort for each sample.

FIGURE 4.. Leave-one-out meta-analysis for gFA and LA-gFA in a study of white matter microstructure and cognitive ability in schizophrenia^a

^a For both gFA and LA-gFA, 11 separate meta-analyses were carried out with n–1 site. The mean Hedges’ g effect size was taken for each meta-analysis with one site omitted for each iteration. The study name corresponds to the results when this site was omitted from the analysis. The association between both gFA and LA-gFA with IQ remained significant for each iteration, indicating that the results are not driven by a specific site. gFA=global fractional anisotropy component; LA-gFA=fractional anisotropy component for six long association tracts; ASRB=Australian Schizophrenia Research Bank; COBRE=Center for Biomedical Research Excellence; EDIN=Edinburgh; HUBIN=Human Brain Informatics; MCIC=MIND Clinical Imaging Consortium; MPRC=Maryland Psychiatric Research Center; TOP=Thematically Organized Psychosis (TOP) NORMENT Research Study. In the lower half of each panel, the “p” appended to site names indicates the values relating to the patient cohort for each sample.