Increased cost of cognitive effort in preclinical Alzheimer’s disease

Abstract

AbstractBackgroundScores on neuropsychological tests depend on willingness to expend the necessary cognitive effort to complete each task. Group differences on cognitive test scores may partially reflect differences in perceived cognitive cost, or the willingness to exert cognitive effort, which will obscure estimates of “true” cognitive ability. It is critical, therefore, to obtain estimates of cognitive effort during neuropsychological testing. Subjective, questionnaire‐based measures of effort are prone to self‐report bias, so we administered an objective measure of cognitive effort (Cognitive Effort Discounting: COGED) to participants at the Knight ADRC. In COGED, participants complete increasingly difficult levels of a working memory task (the N‐Back) and then are asked to rate which “level” they would complete again for a given amount of monetary reward. We hypothesized that objective (COGED) but not subjective measures of effort would decrease in preclinical AD and this would be related to cognitive performance.MethodSixty‐five healthy older adults (Clinical Dementia Rating = 0) completed both the COGED task and a subjective effort rating scale (the NASA task load index). Participants were classified as amyloid‐positive (N=24) or negative (N=41) based either on amyloid PET scan or CSF Ab42/Ab40 ratio. Group differences on the effort task were analyzed using linear regression.ResultEffort ratings decreased across levels of N‐back (p<0.001), indicating the task accurately tracks perceived effort. As expected, amyloid‐positive participants scored lower on the N‐back task (p=.03) but more importantly, they also scored lower on COGED than the amyloid‐negative group (Figure 1, p=0.05), after controlling for age, education and working memory. There were no group differences on the NASA (Figure 2). Surprisingly, neither COGED scores nor the NASA were related to performance on the N‐back, suggesting that effort ratings are independent of task performance.ConclusionWe successfully adapted the COGED paradigm for remote assessment and administered the task to a group of healthy older adults from the Knight ADRC. This task accurately reflects effort / reward processing. Objective (but not subjective) measures of the ability to make these computations decline in preclinical AD but does not appear to affect concurrent cognitive assessments.