Research report: Sex differences in the effects of lifetime intellectual cognitive reserve on language trajectory and risk of non-amnestic mild cognitive impairment: a longitudinal study

A lifetime of intellectual engagement—higher education, a mentally demanding job, and mentally stimulating hobbies after retirement—appears to help older adults hold on to their language abilities and may sharply reduce their risk of a common precursor to dementia, according to a new longitudinal study of more than 1,300 older adults in China.

The study, published in the European Journal of Ageing, followed 1,308 cognitively intact older adults from the Beijing Aging Brain Rejuvenation Initiative, a community-based prospective cohort. Participants had an average age of 65.54 years at baseline, and 69.5 percent were women. Each person was examined at baseline and at two follow-up visits, allowing the researchers to track both cognitive trajectories and new diagnoses of mild cognitive impairment, or MCI, an intermediate state between normal aging and dementia that carries an elevated risk of progression.

To capture intellectual engagement across the entire life course rather than at a single point, the researchers constructed a composite measure they call lifetime intellectual cognitive reserve, or LICR. It combined three components collected at baseline: early-life educational attainment, measured in years of schooling; midlife occupational complexity, scored across interactions with people, data, and things in a person’s work; and late-life mental leisure activities, rated through a 23-item questionnaire covering how frequently participants engaged in various intellectual pursuits after retirement. Using structural equation modeling, the team weighted these three factors to produce a single LICR score for each participant, ranging from −3.41 to 3.47, and also split the sample into lowest, middle, and highest groups by tertile.

Cognitive performance was assessed with a battery of nine neuropsychological tests, including measures of verbal learning, visuospatial construction, working memory, processing speed, and naming. Confirmatory factor analysis distilled these into five cognitive domains—working memory, episodic memory, processing speed, visuospatial ability, and language ability—which together predicted a general cognitive ability score. The researchers then used linear mixed-effects models to estimate how LICR related to baseline performance and to the annual rate of change in each domain, adjusting for a broad set of covariates including age, sex, depression, loneliness, smoking, drinking, physical activity, body mass index, and chronic diseases such as hypertension, diabetes, and cerebral small-vessel disease.

The headline finding concerned language. When the sample was split into groups, those in the highest LICR tertile declined 0.071 standard score units more slowly per year than those in the lowest tertile, a difference that reached statistical significance, while the middle group showed a trend in the same direction. Notably, this protective association was specific to language; episodic memory, processing speed, and general cognitive ability declined over the follow-up period, but only language showed a statistically significant slowing tied to LICR. The authors interpret this as evidence that lifetime intellectual reserve helps maintain crystallized intelligence—the accumulated knowledge and verbal skills built over decades—rather than fluid abilities such as processing speed.

Among the individual components, only educational attainment independently predicted slower language decline, at 0.014 standard score units per year of schooling, and higher education was also linked to a reduced risk of non-amnestic MCI. Interestingly, higher occupational complexity was associated with a slightly faster decline in processing speed, a counterintuitive result the authors suggest might reflect the stress and risk exposures of demanding professions or compensatory depletion of frontoparietal networks that such occupations rely on heavily across adulthood—though they caution these explanations remain speculative and require further testing. Mental leisure activity alone showed no significant relationship with the rate of decline in any domain, underscoring the paper’s central argument that no single proxy captures the cumulative protection of a lifetime of intellectual engagement.

The sex-stratified analyses revealed a striking asymmetry. At baseline, higher LICR was associated with better performance in both sexes, though the domains differed: men showed a larger association with processing speed, while women showed larger associations with episodic memory and visuospatial ability. During follow-up, however, only women with higher LICR showed slower language decline, at 0.04 standard score units per year, and among women in the highest LICR tertile the difference reached 0.11 standard score units per year compared with the lowest group. For the components, only higher educational attainment in women predicted slower language decline. The authors note that women generally outperform men on language tasks in older age, and that women face greater susceptibility to cognitive impairment and dementia, making this sex-specific protection potentially meaningful for targeted prevention.

The second major finding involved MCI risk. Over the two follow-up waves, 82 participants developed MCI. In multivariate Cox regression models, each unit of higher LICR was associated with an 18 percent lower risk of MCI overall and a 34 percent lower risk of the non-amnestic subtype, which involves impairment in domains other than memory, such as language, visuospatial skill, or executive function. When the highest LICR tertile was compared with the lowest, the hazard ratio for MCI was 0.44—corresponding to a 56 percent reduction in risk—and for non-amnestic MCI it was 0.09, a 91 percent reduction.

The authors are careful about the limits of these findings. Participants were volunteers living in urban Beijing rather than a random community sample, so generalization to rural populations and other cultures requires caution. Cognitive reserve remains a theoretical construct without universally accepted measurement standards, making direct comparison with other statistical approaches difficult. The number of incident MCI cases was too small to test whether the association with MCI risk differs between men and women, and the study included no neuroimaging, so the neural mechanisms linking intellectual reserve to preserved language and reduced impairment risk remain unexplored. The team also notes that clinically defined MCI is not a stable diagnosis—up to 24 to 50 percent of affected individuals revert to normal cognition—and 21 participants in this cohort did so at the second follow-up.

Even so, the study offers a data-driven argument for investing in cognitive health across the entire lifespan rather than only in old age. The authors call for larger population-based longitudinal studies, especially in women, and for multimodal neuroimaging to identify the neural plasticity that lifetime intellectual engagement may induce.

Subject of Research: Sex differences in how lifetime intellectual cognitive reserve affects language decline and mild cognitive impairment risk in older adults

Article Title: Sex differences in the effects of lifetime intellectual cognitive reserve on language trajectory and risk of non-amnestic mild cognitive impairment: a longitudinal study

Article References: Wang, D., Dang, M., Long, S., Wang, J., Hu, X., Li, X., & Zhang, Z. (2026). Sex differences in the effects of lifetime intellectual cognitive reserve on language trajectory and risk of non-amnestic mild cognitive impairment: a longitudinal study. European Journal of Ageing, 23(1), Article 43. https://doi.org/10.1007/s10433-026-00927-8

Image Credits: AI Generated

DOI: 10.1007/s10433-026-00927-8

Keywords: cognitive reserve, mild cognitive impairment, language decline, cognitive aging, sex differences, education, occupational complexity, mental leisure activities, dementia prevention, longitudinal study, crystallized intelligence, older women