Deconstructing processing speed deficits in schizophrenia: application of a parametric digit symbol coding test

Abstract

Cognitive processing inefficiency, often measured using digit symbol coding tasks, is a putative vulnerability marker for schizophrenia and a reliable indicator of illness severity and functional outcome. Indeed, performance on the digit symbol coding task may be the most severe neuropsychological deficit patients with schizophrenia display at the group level. Yet, little is known about the contributions of simpler cognitive processes to coding performance in schizophrenia (e.g. decision making, visual scanning, relational memory, motor ability). We developed an experimental behavioral task, based on a computerized digit symbol coding task, which allows the manipulation of demands placed on visual scanning efficiency and relational memory while holding decisional and motor requirements constant. Although patients (n=85) were impaired on all aspects of the task when compared to demographically matched healthy comparison subjects (n=30), they showed a particularly striking failure to benefit from the presence of predictable target information. These findings are consistent with predicted impairments in cognitive processing speed due to schizophrenia patients' well-known memory impairment, suggesting that this mnemonic deficit may have consequences for critical aspects of information processing that are traditionally considered quite separate from the memory domain. Future investigation into the mechanisms underlying the wide-ranging consequences of mnemonic deficits in schizophrenia should provide additional insight.

Figure 1. Schematic Illustration of the Parametric Digit Symbol Coding Task

In this illustration of a trial of the computerized digit symbol coding task, the single digit-symbol pair in the middle of the figure represents the target pair, which the participant must check against the reference set, represented by the row of digit-symbol pairings along the top of the figure. For publication, bright blue features of symbols were transformed to gray scale. Reference set size was varied parametrically (3, 6, or 9 pairs) between blocks of trials. Additionally, blocks included either a fixed reference set, in which the digit-symbol pairings remained constant from trial to trial, or a random reference set, in which digits were randomly reassigned to the symbols at the beginning of each new trial. Performance during the fixed condition is enhanced by memory of the reference set. Conversely, performance during the random condition is particularly dependent upon visual scanning, because the target pair must be compared to the reconfigured reference set on each trial. In all trials, subjects indicate via button press, if the target pair was identical to one the digit-symbol pairs in the reference set. The arrows reminding the subject which button to press to indicate that the target is the same as or different than a pairing in the reference set remained visible throughout task performance.

Figure 2. Digit Symbol Coding Performance

Performance is plotted in number of correct responses provided within a 60 second block of trials, separately for each set size and condition. Patient (N=85) averages are depicted as filled circles and control (N=30) averages are filled squares. Main effects of Condition, Set Size, and Diagnostic Group are apparent, whereby performance was better in the fixed condition, and at lower set sizes. Controls also performed better than patients, a difference that was greater in the fixed condition than in the random condition.