Detection of Mild Cognitive Impairment Using In‐clinic and Remote Unsupervised Digital Cognitive Assessments

Abstract

AbstractBackgroundThe Cogstate Brief Battery (CBB) is a computerized cognitive assessment validated for Alzheimer’s disease (AD) and unsupervised use. The CBB assesses processing speed, attention, visual learning, and working memory and is offered to cognitively normal (CN) and mild cognitive impairment (MCI) participants in ADNI‐3. In‐clinic visits are completed annually for MCI and every other year for CN, with both groups also able to complete unsupervised assessments at‐home within 14 days of the first in‐clinic visit and at up to 3 monthly intervals.MethodParticipants were 146 CN older adults (Mage = 72.1, SD = 6.43, age range 57–90 years, 58.2% females) and 37 older adults with MCI (Mage = 74.4, SD = 7.50, age range 61–89 years, 51.4% females). All participants underwent PET scans and confirmation of diagnosis at Baseline. Only participants with confirmed amyloid status were included in these analyses (i.e., all MCI participants were Aβ+ and all CU participants were Aβ‐). Participants completed the CBB in a supervised in‐clinic setting at Baseline, and again in an unsupervised remote setting within 90 days. Receiver Operating Characteristic analyses were conducted to ascertain whether the classification performance of the CBB in detecting MCI was similar in both settings.ResultAll CBB measures showed a significant ability to discriminate between CU Aβ‐ and MCI Aβ+ participants at both the supervised in clinic baseline (AUCs 0.63‐0.75) and initial remote visit (AUCs 0.63‐0.78). There was no significant difference for any CBB measure in classification performance (measured by AUC) between remote and supervised assessment (ps > 0.146).ConclusionThe CBB showed good ability to classify MCI‐related cognitive impairment, both supervised in clinic, and/or remotely without supervision.