Abstract

Working memory (WM) deficits are a persistent, disabling and relatively treatment-resistant feature of schizophrenia that may underlie many cognitive deficits and symptoms. They are associated with prefrontal cortex dysfunction. While most neuroimaging studies of WM demonstrate "task-related hypofrontality" in schizophrenic relative to healthy subjects, several recent studies have reported equal or increased prefrontal activity. These findings challenge central assumptions regarding cognitive deficits and prefrontal cortex dysfunction in schizophrenia. The goal of this review is to reconcile these seemingly discrepant findings. Methodological factors addressed include the use of intersubject averaging, WM task parameters and the reliability of the measures. Factors intrinsic to schizophrenia and their relevance to the selection of experimental methods and the interpretation of group data are also discussed. Both hypo- and hyperfrontality are hypothesized to be valid and informative reflections of prefrontal cortex dysfunction in schizophrenia. Due to the heterogeneity and variability of both performance and regional recruitment in schizophrenia, whether individual data is considered, the level and type of WM demands and the composition of the sample with regard to performance deficits all influence study outcome and contribute to discrepancies. Although the prefrontal cortex is consistently implicated in WM deficits, the basis of its dysfunction and its exact contribution remain unclear. Future work might focus on delineating the exact WM processes, domains and components that are deficient. In addition, variability in behavior and activation might best be regarded as intrinsic to schizophrenia and having a neural basis that requires explanation. In combination with other techniques, neuroimaging can identify the neural circuitry responsible for WM deficits and elucidate the contribution of each anatomical component.