Dissociable contributions of working memory capacity and lapses of attention to cognitive deficits in schizophrenia

Working Memory (WM) impairment is a central feature of schizophrenia, yet it remains unclear whether WM deficits primarily reflect reduced storage capacity or inconsistent attentional engagement of available capacity. Traditional trial-averaged measures obscure this distinction. In this study, we combined a multi-item visual WM task with computational modeling to distinguish maximum storage capacity from trial-to-trial consistency of attentional engagement in people with schizophrenia (PSZ, N=52) and healthy control subjects (HCS, N=41). Participants completed two task variants that differed in response demands: a Free Response Order condition that permitted self-selected report order and a Random Response Order condition that imposed an externally determined order. PSZ showed robust impairments in average WM performance across tasks. Model-based analyses revealed that these impairments reflected both a reduction in maximum WM capacity and reduced consistency of attentional engagement. Capacity estimates were stable across task contexts, whereas consistency was selectively reduced in PSZ when response order was externally imposed. Time-resolved analyses demonstrated that group differences were stable across the experimental session and were not attributable to learning or fatigue effects. These findings indicate that WM deficits in schizophrenia arise from distinct impairments in both storage capacity and attentional consistency, and that task context influences the expression of these deficits. Parsing these components advances mechanistic models of cognitive dysfunction in schizophrenia and may help guide the design of cognitive assessments that are more sensitive to underlying process-level impairments.