Do Late-Life Cognitive Activities Moderate the Relationship Between Education and Cognition?: Findings from the Aging, Demographics, and Memory Study (ADAMS)

Background: Based on the cognitive reserve theory, education and cognitive leisure activities are regarded as major constructs of earlier and later-life cognitive reserve. Indeed, an established body of study found the protective effects of both constructs on cognition among older adults. However, few studies focused on seeking the mechanism of how those two predictors interact to influence cognition. Moreover, the definitions of cognitive leisure activities have often been diverse that which specific activities dedicate cognition in greater degree remain to be answered. Until recently, dementia risk has been the most popular outcome variable to measure one’s cognition. Cognitively impaired not demented (CIND) status holds comparable value of interest with an increasing demand for earlier prevention of dementia. Our study investigated the moderating role of each subcategory of cognitive leisure activities on the relationship between education and cognition, categorized into normal cognition, CIND, and dementia.  

Methods: Data were drawn from a national sample of the Aging, Demographics, and Memory Study, a supplement sample from the larger Health and Retirement Study. The final sample size included 704 older Americans age 70 and above. Education was measured in years of formal education (0-17). Factor analysis was conducted to draw subcategories from seven cognitive leisure activities resulting in two factors: verbal (reading, writing, and discussing) and visuospatial activities (doing arts, doing puzzles, and playing games). Each category was measured in average of frequency score (1-4). Multinomial logistic regression was followed to test main and moderating effects of each activities on cognition. CIND sample was a reference group to be compared with normal cognition or dementia. The clinical diagnosis of cognition was given by the expert consensus panel. Other covariates such as sociodemographic variables, depression, chronic conditions, self-rated health, and presence of apolipoprotein E ε4 genotype were included in all analyses using STATA software.  

Results: The study revealed that individuals with greater engagement in verbal and visuospatial activities were more likely to have normal cognition but less likely to have dementia when compared to CIND. Education was related with normal cognition but not with dementia (vs. CIND). Only verbal activities had significant interaction effect. Individuals with higher education had more benefit of engaging in verbal activities for decreased odds of having dementia (vs. CIND).

Implications: Findings highlight previous challenges faced by researchers considering the complex constructs for cognitive leisure activities. Moreover, our study further suggests that different approach should be made for normal and CIND population. The former seems to be affected by both education and cognitive leisure activities to remain normal while the latter seem to be majorly affected by cognitive leisure activities to endure CIND. In specific, if an individual is yet normal in cognition, education still matters to prevent the manifestation of cognitive impairment. If he/she is already impaired in cognition, education does not come into play but engagement in cognitive activities matters upfront. This suggests that it is still worthwhile to encourage cognitively stimulating activities for individuals with cognitive impairment but more focus of verbal activities for those with higher education.