Implicit memory in schizophrenia: A meta-analysis

Objective Previous studies examining implicit memory in schizophrenia yielded inconsistent results. The present meta-analysis aimed at determining whether, compared to healthy controls, schizophrenic patients: (a) exhibited reduced priming in the whole set of studies; (b) were differentially impaired in conceptual/perceptual and production/identification tests; and (c) were less efficient in the use of semantic encoding processes. Method A systematic search in PsycINFO and PubMed led to the selection of 22 critical studies (31 effect sizes), comparing repetition priming in 836 schizophrenic patients and 760 healthy controls. Moderators were assessed by classifying implicit tasks into the perceptual/conceptual and identification/production categories, and by distinguishing between perceptual and conceptual encoding instructions. Results Overall, implicit memory was slightly, but significantly, impaired in schizophrenia (d = 0.179). Patients exhibited reduced priming in conceptually-driven tasks (d = 0.447), but intact priming in perceptually-driven tasks (d = 0.080). No significant difference was observed between identification and production priming (d = 0.064 vs. d = 0.243). Finally, priming in schizophrenic patients was significantly lower than that of controls when the encoding task required the analysis of the conceptual properties of the stimuli (d = 0.261). Conclusion Results suggest that schizophrenia is associated with a specific deficit in the use of conceptual processes, both at encoding and at retrieval. In contrast with theoretical expectations, high levels of response competition did not disproportionately impair the patients' performance.