A Preliminary Study of the Cognitive Correlates and Efficacy of Working Memory Training for Comorbid Problem Gambling and Schizophrenia Spectrum Disorders

Background: Recent research has demonstrated elevated rates of problem gambling (PG) among persons with schizophrenia. This preliminary study examined the cognitive correlates associated with PG and schizophrenia spectrum disorders (schizophrenia and schizoaffective) and the efficacy of computerized working memory (WM) training on reducing PG severity. Methods: Participants with schizophrenia or schizoaffective disorders (N = 83) were administered WM tasks. Thereafter, participants were randomly assigned to: (1) WM training, (2) processing speed training (active control), or (3) no-training (passive control). Participants also completed the Problem Gambling Severity Index (PGSI) at baseline and at 10-week follow-up. Participants were categorized as endorsing PG if they scored 3+ on the PGSI. Results: Overall, 21 (25.3%) participants were classified as meeting threshold for PG. There were no significant differences between participants endorsing PG and those who did not on any cognitive task with small effect sizes. Of the 21 participants endorsing PG, 16 completed the follow-up measure of PGSI. Participants assigned to the WM group reported significant reductions in PGSI from baseline to follow-up with a large effect size. No significant reductions were found for processing speed or control conditions with large and small effects respectively. Conclusions: The current results do not support a baseline difference in WM between those who endorsed PG and those who did not. However, WM training may be effective for reducing problematic gambling among persons with PG and schizophrenia spectrum disorder.