Correlating digit span performance and event-related potentials to assess working memory

Event-related brain potentials (ERPs) were recorded during a computerized and modified version of the Digit Span Backwards (DB) task from the Wechsler Adult Intelligence Scale-Third Edition (WAIS-III). The modified DB version (ERP-DB task) was divided into two sections of 2, 4, 6 and 8 digits in length (Group 1) and 3, 5 and 7 digits in length (Group 2). Each trial had a study phase and a test phase. For the study phase, a series of digits was presented sequentially and aurally to 20 participants (10 for each group). For the test phase, a second series of digits was also presented sequentially and aurally that either corresponded to the reverse order of the digits in the study phase (correct condition) or had one digit in the sequence replaced by an incorrect digit (incorrect condition). The traditional DB task of the WAIS-III was also administered for comparison purposes. A prolonged positive slow wave (PSW) peaking between 450 and 750 ms was elicited to incorrect condition trials. For each participant, a derived measure was calculated from the ERP differentiation between correct and incorrect conditions. The derived measure was defined as the mean of the t-values obtained from the correct and incorrect waveform comparison, within the temporal interval that encompassed this component. The strongest statistical correlations between the derived measure and the traditional DB test scores were found at the Pz site (Group 1: r=0.79; Group 2: r=0.59). This statistical approach shows that it is possible to adequately relate an individual's performance on a traditional measure of working memory and ERP patterns. Overall, we believe that this kind of ERP approach holds promise as a technique for assessing quantitatively non-communicative patients.