The Mobile Toolbox for assessing cognition in older adults: associations with standardized cognitive testing and amyloid and tau PET.

Background: Remote, smartphone‐based cognitive assessments such as the Mobile Toolbox (MTB) can improve our ability to detect subtle, yet meaningful cognitive changes that are affected early on by Alzheimer's disease (AD) pathology. The goal of our study was to investigate which MTB tests are sensitive to subtle cognitive deficits due to early amyloid and tau deposition. Method: The MTB will be administered to N = 100 clinically normal (CN) older adults who have retrospective and prospective clinical, cognitive and AD biomarker data from three well‐characterized prospective cohort‐studies of CN older adults: the Harvard Aging Brain Study and the affiliated Subjective Cognitive Decline and Instrumental Activities of Daily Living cohorts. The MTB includes six measures of fluid cognition, including episodic memory (Picture Sequence Memory and Face‐Name Associations), executive functions (Dimensional Change Card Sort, Flanker), working memory (Memory for Sequence), processing speed (Number‐Symbol Match), and two measures of crystallized cognition (Spelling and Vocabulary). Using confirmatory factor analysis, MTB baseline data will be used to determine which MTB tests are most strongly associated with established clinical assessments (i.e., measures included in the Preclinical Alzheimer's Cognitive Composite‐5 [PACC5]) as well as amyloid and tau biomarkers on positron emission tomography (PET) imaging. Result: Preliminary MTB data was available on 15 participants (age 74.3±8.2, 63% female, 16.6±2.9 years of education, MMSE 29±1.5). Table 1 shows exploratory correlation analyses, revealing that both MTB crystallized measures were associated with education but not age or sex. As expected, worse performance on MTB memory measures were associated with lower PACC5 scores, and greater amyloid and tau burden, albeit at trend‐level. Executive function and processing speed measures were not related to amyloid but did show associations with entorhinal tau. Conclusion: While preliminary, these findings suggest that the MTB may help elucidate the subtle cognitive deficits that are associated with early AD pathology. In addition to memory measures, the MTB fluid measures capture various aspects of cognitive function that are also relevant to preclinical AD. Data collection is ongoing, and next steps include examining which (combination of) MTB measures are most sensitive for detecting emerging AD pathology and to track cognitive change over time. [ABSTRACT FROM AUTHOR]