Improving everyday error detection, one picture at a time: a performance-based study of everyday task training.

OBJECTIVE Research suggests that dementia patients detect fewer action errors than age-matched controls; however, little is known about the derivation of their error-monitoring difficulties. The aims of the study are to evaluate a novel, task-training action intervention (TT-NAT) designed to increase error monitoring in dementia patients and to pinpoint the relation between error monitoring and neuropsychological processes. METHOD Participants (n = 45) with dementia were administered the Standard NAT, a performance-based test requiring completion of three everyday tasks. A second group (n = 42) was administered the TT-NAT, which includes a brief training session prior to the commencement of each task. All participants were compared on the following variables: total errors, proportion of errors detected, and proportion of errors corrected. Correlations between error-monitoring variables and neuropsychological tests of executive functioning and language were performed. RESULTS TT-NAT participants produced fewer total errors and detected significantly more errors than Standard NAT participants (z = 3.0; t = 3.36; p < .05). Error detection was strongly related to only the language composite index (r = .57, p = .00) in the TT-NAT, whereas it was moderately related to both the language (r = .31, p = .04) and executive composite (r = .36, p = .02) indices in the Standard NAT condition. CONCLUSION Review of task steps and objects before task performance may be a promising intervention for error-monitoring deficits in dementia patients; this finding has implications for neuropsychological rehabilitation of functional deficits in this population.