Working memory in schizophrenia: a meta-analysis.

BACKGROUND Memory impairment is being recognized increasingly as an important feature of the neuropsychology of schizophrenia. Dysfunction of working memory, a system for the short-term storage and manipulation of information, may relate to a number of core symptoms of schizophrenia. Many studies have examined working memory function in schizophrenia but a clear understanding of the nature and extent of any deficit has been elusive. METHOD A systematic review and meta-analysis of studies comparing working memory function in subjects with schizophrenia and healthy controls was performed. Following a comprehensive literature search, meta-analyses were conducted on 36 measures of phonological, visuospatial and central executive working memory functioning, encompassing 441 separate results from 187 different studies. RESULTS Statistically significant effect sizes were found for all working memory measures, indicating deficits in schizophrenia groups. Some of these were robust findings in the absence of evidence of significant heterogeneity or publication bias. Meta-regression analyses showed that the working memory deficit was not simply explained by discrepancies in current IQ between schizophrenia and control groups. CONCLUSIONS Large deficits in working memory were demonstrated in schizophrenia groups across all three working memory domains. There were, however, no clear differences across subdomains or between particular working memory tasks. There was substantial heterogeneity across results that could only be partly explained.