relationship of cognitive reserve and measures of reserve with longitudinal cognitive change and the duration of preclinical, prodromal, and mild Alzheimer disease (AD) dementia remains to be fully character-ized. In our study, 660 b-amyloid-positive participants staged with preclinical AD, prodromal AD, and dementia due to AD from the Alzheimer's Disease Neuroimaging Initiative were selected. Cognitive reserve and brain reserve were defined by conventional proxies or the residual method at baseline. We evaluated the utility of these reserves in predicting longitudinal cognitive change by mixed effects models and used a multi-state model to esti-mate stage duration stratified by reserve groups. Corrected for age, sex, and APOE-e4 status, reserve was asso-ciated with cognitive decline, and the effects changed depending on the specific measures of reserve and the stage of AD. Reserves defined by the residual method were stronger predictors of cognitive change than those defined by conventional proxies. The estimated time from preclinical to mild AD dementia varied from 15-24 years based on the different reserve groups, and we observed a linear trend for the longest duration in individuals with high cognitive reserve/high brain reserve, followed by those with high cognitive reserve/low brain reserve, low cognitive reserve/high brain reserve, and low cognitive reserve/low brain reserve. This study showed a reduced risk of cognitive decline for individuals with higher level of reserve regardless of methods for measuring reserve. Interindividual differences in reserve may be important for clinical practice and trial design.(c) 2022 IBRO. Published by Elsevier Ltd. All rights reserved.