Change in neuropsychological functioning over one year in youth at clinical high risk for psychosis

Abstract

Schizophrenia and related psychotic disorders are associated with significant neuropsychological (NP) impairments. Yet the onset and developmental evolution of these impairments remains incompletely characterized. This study examined NP functioning over one year in a sample of youth at clinical high risk (CHR) for psychosis participating in a treatment study. We assessed functioning across six cognitive domains at two time points in a sample of 53 CHR and 32 healthy comparison (HC) subjects. Linear regression of HC one-year scores was used to predict one-year performance for CHR from baseline scores and relevant demographic variables. We used raw scores and MANOVAs of the standardized residuals to test for progressive impairment over time. NP functioning of CHR at one year fell significantly below predicted levels. Effects were largest and most consistent for a failure of normative improvement on tests of executive function. CHR who reached the highest positive symptom rating (6, severe and psychotic) on the Structured Interview of Prodromal Syndromes after the baseline assessment (n = 10/53) demonstrated a particularly large (d = -1.89), although non-significant, discrepancy between observed and predicted one-year verbal memory test performance. Findings suggest that, although much of the cognitive impairment associated with psychosis is present prior to the full expression of the psychotic syndrome, some progressive NP impairments may accompany risk for psychosis and be greatest for those who develop psychotic level symptoms.

Copyright © 2013 Elsevier B.V. All rights reserved.

Figures

Figure 1

Mean Standardized Residuals of One-Year Scores by CHR Symptom Group. CHR: Clinical High Risk; See Tables 2 or 3 for group descriptions. Residuals of one-year scores were calculated as observed minus predicted scores based on HC distribution; Residuals were standardized so that the M = 0, and SD = 1 for HC. Error bars represent SD. MANOVA overall effect of group: p < 0.001, significant for Verbal Memory and Executive Function domains. Significant (p < 0.01) pairwise comparisons were found between HC and CHR groups for these domains. All CHR performed significantly below predicted levels relative to HC on the executive function domain (p < 0.05). There were no significant differences between CHR groups.

Figure 2

CHR relative to HC Effect Sizes for Raw Test Scores at each Time Point. Effect sizes (ES) were calculated from CHR relative to HC raw test scores at each time point. The mean for tests with child-adult versions was obtained by weighting the separate ES for each child or adult test by the CHR N for that ES. ES were calculated from scaled scores for individuals who transferred from child to adult tests over time. Trail-Making time and WCST Perseverative Errors were log-transformed and made negative to maintain the same direction of effect across tests. See Tables 2 or 3 for group descriptions. See Supplement Tables 1–3 for raw score ES by test, domain, and time point.