Subjective Cognitive Concerns are Associated with Worse Performance on Mobile-App Based Cognitive Assessment: An Observational Study in Cognitively Normal Older Adults (Preprint)

BACKGROUND Subjective cognitive concerns (SCC) may be among the earliest clinical symptoms of dementia. There is growing interest in applying mobile app-based cognitive assessment to remotely screen for cognitive status in preclinical dementia, but the relationship between SCC and relevant mobile assessment metrics is uncertain. OBJECTIVE We characterized the relationship between SCC and adherence, satisfaction, and performance on mobile-app assessments in cognitively unimpaired older adults. METHODS Participants (N=122, Mage=69.85, Meducation=16.52, %female=66.7, %White=86.2) completed 8 assessment days using Mobile Monitoring of Cognitive Change (M2C2), an app-based testing platform, with brief daily sessions within morning, afternoon, and evening time windows (24 total testing sessions). M2C2 includes digital working memory, processing speed, and episodic memory tasks. Participants provided feedback about their satisfaction and motivation related to M2C2 upon study completion. SCC was assessed using the Cognitive Function Instrument (CFI). Regression analyses evaluated the association between SCC and adherence, satisfaction, and performance on M2C2, controlling for age, sex, and loneliness (UCLA Loneliness Scale). Linear-mixed effects models evaluated whether SCC predicted M2C2 subtest performance over the 8-day testing period, controlling for covariates. RESULTS SCC was not associated with app satisfaction or protocol motivation, but it was significantly associated with lower rates of protocol adherence (ß=-0.19, p=.04, 95% CI -0.642, -0.016). Higher SCC endorsement significantly predicted worse overall episodic memory performance (ß=-0.201, p = .021, 95% CI -0.02, -0.01), but not working memory or processing speed. There was a main effect of SCC on working memory performance at day 1 (Est=-1.017, SE=0.49, p=0.038) and a significant interaction between SCC and working memory over the 8-day period (Est=0.047, SE=0.02, p=0.039), such that SCC was associated with initially worse, then progressively better working memory performance. CONCLUSIONS SCCs are associated with worse overall memory performance on mobile-app assessments, patterns of cognitive inefficiency (variable working memory) and mildly diminished adherence across an 8-day assessment period. Findings suggest that mobile app assessments may be sensitive to subtle cognitive changes, with important implications for early detection and treatment for individuals at risk for dementia.