Unsupervised Performance of the CogState Brief Battery in the Brain Health Registry: Implications for Detecting Cognitive Decline

Introduction: The feasibility and validity of unsupervised, longitudinal brief computerized cognitive batteries is unknown. Methods: Participants aged 56–90 (N = 19476) from the Brain Health Registry (BHR) completed the CogState Brief Battery (CBB) at 6-month intervals over a period of 5 years. We used linear mixed-effects models to assess whether cross-sectional and longitudinal performance on CBB within BHR was associated with demographic and cognitive characteristics. We also defined a group of CBB decliners based on subject-specific slopes and estimated associations between decliner status and participant characteristics. Results: We found weak associations between longitudinal change in CBB and participant characteristics. Cross-sectional CBB scores were significantly associated with participant characteristics such as age, gender, ethnicity, self-reported disease status, and memory concern. CBB decliners were more likely to self-report mild cognitive impairment (MCI) and memory concerns. Discussion: Cross-sectional, remote CBB shows evidence of construct validity, but our results suggest that longitudinal assessment may not provide additional value for identifying those at risk for and with cognitive impairment.