Contextualized cognition: Clarifying associations between remote unsupervised performance and clinically relevant contextual factors.

Objective: Ambulatory cognitive assessments show promise as scalable, low friction performance metrics that address limitations of clinician-administered testing. However, influences of contextual factors on unsupervised testing are not well understood. This study evaluated associations between remote, unsupervised, ambulatory cognitive assessments and self-ratings of mood, arousal, stress, pain, and distraction during task performance. Method: 190 participants (ages 18-89) engaged in a 14-day ecological momentary assessment (EMA) protocol using the Mobile Monitoring of Cognitive Change (M2C2) platform. Participants completed ambulatory assessments of processing speed (Symbol Search) and working memory (Rotation Span) and provided self-ratings of contextual factors 5 times daily. Age stratified multilevel modeling (<65, ≥65 years of age) was used to analyze within- and between-person associations between context and performance. Results: In multilevel models where all 5 contextual factors were examined, within-person variation in momentary arousal and distraction levels were significantly associated with variation in performance on each cognitive task, with higher arousal and lower distraction being associated with better performance across both age groups. Significant between-person associations were limited to adults ≥65 years of age, where characteristic levels of arousal and distraction were associated with Rotation Span performance. Conclusions: This study highlighted distraction and arousal at time of task completion as important for understanding variation in remote, unsupervised cognitive testing. Monitoring of these associations using EMA may complement weaknesses of traditional supervised testing approaches, such as testing at a single time or set intervals. It also may facilitate individualized treatment recommendations for cognitive difficulties associated with contextual factors.