Relationships Between NIH Toolbox Cognitive Battery Subtests and In Vivo Amyloid and Tau in People with Amnestic Mild Cognitive Impairment and Dementia of the Alzheimer’s Type

Abstract Background The computerized NIH Toolbox Cognition Battery (NIHTB‐CB) was designed to assess cognitive functioning across the lifespan. Previous studies demonstrated that NIHTB‐CB measures discriminate between healthy controls (HCs), individuals with amnestic mild cognitive impairment (aMCI), and individuals with dementia of the Alzheimer’s type (DAT). Scores on NIHTB‐CB tasks also correspond with performance on well‐validated neuropsychological measures of the same cognitive domains. We are among the first to evaluate how NIHTB‐CB subtest performances and in vivo Alzheimer’s disease (AD) pathology are related in people with aMCI and DAT. Method As part of a larger clinical trial in the Research Program on Cognition and Neuromodulation Based Interventions, 139 participants with aMCI or DAT (Table 1) underwent neuropsychological assessment, including five iPad NIHTB‐CB subtests: Flanker, List Sort, Card Sort, Pattern Separation, and Picture Sequence. Participants also completed positron emission tomography (PET) scans of amyloid and tau. Amyloid PET scans were normalized to the centiloid value, and tau PET standardized uptake value ratios (SUVr) were merged into a bilateral meta‐region‐of‐interest based on Braak staging. Stepwise linear regressions were used to evaluate the relationship between age and education with NIHTB‐CB subtests (Model 1), as well as in vivo AD pathophysiology with NIHTB‐CB subtests beyond age and education (Model 2). Result In vivo amyloid was significantly related to List Sort (β = ‐0.31, p<.001, ΔR2 = .09) and Card Sort performance (β = ‐0.25, p = .004, ΔR2 = .06), explaining significantly more subtest variance beyond age and education. In vivo tau was significantly related to performance on all NIHTB‐CB subtests and explained significantly more subtest variance than demographic factors (Flanker: β = ‐0.25, p = .004, ΔR2 = .06; List Sort: β = ‐0.40, p<.001, ΔR2 = .15; Card Sort: β = ‐0.32, p<.001, ΔR2 = .10; Pattern Separation: β = ‐0.18, p = .04, ΔR2 = .03; Picture Sequence: β = ‐0.23, p = .01, ΔR2 = .05). Conclusion In vivo AD pathophysiology is related to NIHTB‐CB subtest performances in individuals with aMCI or DAT and explains significantly more NIHTB‐CB subtest variance than demographic factors alone. This work adds growing support to the relationship between NIHTB‐CB subtests and in vivo AD biomarkers.