Functional Dissociation of Confident and Not-Confident Errors in the Spatial Delayed Response Task Demonstrates Impairments in Working Memory Encoding and Maintenance in Schizophrenia

Abstract

Even though extensively investigated, the nature of working memory (WM) deficits in patients with schizophrenia (PSZ) is not yet fully understood. In particular, the contribution of different WM sub-processes to the severe WM deficit observed in PSZ is a matter of debate. So far, most research has focused on impaired WM maintenance. By analyzing different types of errors in a spatial delayed response task (DRT), we have recently demonstrated that incorrect yet confident responses (which we labeled as false memory errors) rather than incorrect/not-confident responses reflect failures of WM encoding, which was also impaired in PSZ. In the present study, we provide further evidence for a functional dissociation between confident and not-confident errors by manipulating the demands on WM maintenance, i.e., the length over which information has to be maintained in WM. Furthermore, we investigate whether these functionally distinguishable WM processes are impaired in PSZ. Twenty-four PSZ and 24 demographically matched healthy controls (HC) performed a spatial DRT in which the length of the delay period was varied between 1, 2, 4, and 6 s. In each trial, participants also rated their level of response confidence. Across both groups, longer delays led to increased rates of incorrect/not-confident responses, while incorrect/confident responses were not affected by delay length. This functional dissociation provides additional support for our proposal that false memory errors (i.e., confident errors) reflect problems at the level of WM encoding, while not-confident errors reflect failures of WM maintenance. Schizophrenic patients showed increased numbers of both confident and not-confident errors, suggesting that both sub-processes of WM—encoding and maintenance—are impaired in schizophrenia. Combined with the delay length-dependent functional dissociation, we propose that these impairments in schizophrenic patients are functionally distinguishable.

An Electrophysiological Dissociation of Encoding vs. Maintenance Failures in Visual-Spatial Working Memory

“…Together, these ERP findings suggest a functional dissociation between confident and not-confident errors in the delayed orientation-discrimination task, with IC responses reflecting inefficient WM encoding (supporting Hypothesis 1) and not-confident errors most likely reflecting inefficient WM maintenance (supporting Hypothesis 2). This is in line with the previous behavioral evidence derived from studies that manipulated either the delay length (Mayer et al, 2018) or encoding demands (Mayer et al, 2011) in the spatial delayed response task. In addition, these findings are consistent with previous neuroimaging results (Lee et al, 2008) indicating that confident errors are not due to failures during the active maintenance of WM contents.…”

“…In contrast, incorrect/not-confident (IN) responses are more likely caused by the degradation of representations during the active maintenance of WM contents, resulting in judgments of low confidence. Consistent with these assumptions, we have demonstrated a functional dissociation between confident and not-confident errors in the spatial delayed response task with different delay lengths (Mayer et al, 2018). In line with the encoding hypothesis, we have also shown that the percentage of incorrect/confident (IC) responses in a visuo-spatial delayed response task decreased when the processes that support WM encoding were facilitated (Mayer et al, 2011).…”

“…In line with the encoding hypothesis, we have also shown that the percentage of incorrect/confident (IC) responses in a visuo-spatial delayed response task decreased when the processes that support WM encoding were facilitated (Mayer et al, 2011). Furthermore, this behavioral approach has been useful to dissociate encoding and maintenance deficits in patients with schizophrenia Mayer et al, 2014Mayer et al, , 2018.…”

“…One issue that complicates the investigation of failures in processes related specifically to WM encoding is the difficulty of isolating the encoding process in behavioral paradigms because performance measures (i.e., accuracy and reaction time) are compound measurements that potentially reflect processes associated with all different task phases (i.e., encoding, maintenance, and retrieval). In order to isolate WM encoding processes, we previously introduced a novel behavioral approach based on the analysis of different types of correct and erroneous responses depending on the trial-to-trial level of self-reported subjective response confidence (Lee et al, 2008;Mayer et al, 2011Mayer et al, , 2014Mayer et al, , 2018Rademaker et al, 2012;Peters et al, 2019). Specifically, we reasoned that incorrect responses that were, however, given with confidence most likely reflect a problem at the encoding stage.…”

“…For comparison with our previous clinical studiesMayer et al, 2014Mayer et al, , 2018, IQ was measured with the Mehrfachwahl-Wortschatz-Intelligenztest(Lehrl, 1999; the German version of the National Adult Reading Test), which is an established measure for estimating premorbid IQ in clinical samples.Frontiers in Psychology | www.frontiersin.org…”