Journal of Alzheimer's Disease · Nov 18, 2026
Sex-specific cognitive predictors of conversion from mild cognitive impairment to Alzheimer's disease
By Lajoie X, Boutin S, Clarke N et al.
Background: Mild cognitive impairment (MCI) involves cognitive decline without impaired daily functioning. Although it increases risk for Alzheimer's disease (AD), many remain stable or revert to normal cognition. Identifying those most likely to convert is essential for early intervention. Despite women comprising two-thirds of AD cases, sex differences in cognitive predictors of MCI conversion remain underexplored. Objective: To identify baseline neuropsychological and sociodemographic factors associated with progression from MCI to AD and determine whether these conversion predictors differ by sex. Methods: Using baseline data from the ADNI cohort and longitudinal follow-up to determine clinical outcomes, we compared baseline profiles of 435 MCI non converter individuals and 328 MCI converters, who later converted to AD, using linear regression and classification approaches, including sex effects and sex-stratified analyses. Results: Converters had greater baseline impairments not only in episodic memory and executive function, but also language. Significant, but small, sex-by-conversion interactions in memory indicated slightly more pronounced conversion-related decline among women. Machine learning classifier achieved a mean bootstrap accuracy of 74.0% in distinguishing converters from non-converters, slightly improving in sex-specific models for women (76.2%) but not for men (71.4%). Predictive features differed by sex: memory (recognition), global cognition and executive function were most informative in men, while functional ability, memory (encoding), demographics and language were stronger predictors in women. Conclusions: These findings highlight the value of neuropsychological assessments for predicting progression from MCI to AD and underscore the importance of incorporating sex-specific approaches to improve the clinical relevance of dementia risk prediction.