The Relationship Between Social Interaction and Executive Function in Older Adults: A Cross-Cultural Comparison

Molly Split, Drexel University, Philadelphia, United States
Kathryn Devlin, Drexel University, Philadelphia, United States
Maria Schultheis, Drexel University, Philadelphia, United States


Mild cognitive impairment (MCI) represents the intermediate stage between normal brain aging and early dementia. Early detection of MCI has become increasingly important because it allows clinicians to implement strategies earlier in the disease process that protect the brain against further decline. Increasing socialization and maintaining strong social relationships are behavioral strategies clinicians often recommend. Yet, little evidence clarifies what cognitive functions are necessary for fostering these experiences in older adults, especially older adults of color who are at elevated risk of dementia. Expanding on prior work, the current study examined the relationship between social interaction abilities and executive function (EF) in a diverse sample of community-dwelling older adults and explored the cross-cultural reliability of a previously developed self-report measure for assessing the cognitive demand of social interactions (Cognitive Assessment of Social Interaction; CASI).

Participants and Methods:

171 older adults without a prior neurocognitive diagnosis (54 Black non-Latino adults, 52 Latino adults of any race, and 65 White non-Latino adults) completed the CASI–Revised Self-Report (CASI-R - Self) which was previously determined to be reliable and valid for assessing cognitive difficulties with social interactions in older adults (Split et al., in preparation) and two online tests of EF using TestMyBrain.org (Digit Span and the Gradual Onset Continuous Performance Test; CPT). Hierarchical linear regressions were conducted to examine the relationship between CASI-R and EF performance and included gender, personality (Ten Item Personality Inventory), depression (Beck Depression Inventory – 2), and anxiety (State-Trait Anxiety Inventory – Trait) as covariates. Three separate exploratory factor analyses (EFA) were performed to examine measurement invariance of the CASI-R between the three ethnoracial groups.


After controlling for covariates, higher levels of cognitive complaints on the CASI-R related to fewer digits forward (β = -.23, p < .001), fewer digits backward (β = -.12, p = .03), and less accurate detectability between targets/non-targets on the CPT (β = -.36, p < .001). EF performance and CASI-R ratings did not significantly differ across ethnoracial groups, and reliability results revealed comparable Cronbach’s alpha (.983 for Black participants, .963 for Latino participants, and .980 for White participants). However, EFA results revealed different factor structures (3 factors for Black participants, 4 factors for Latino participants, and 2 factors for White participants).


Self-rated difficulties with social interactions related to poorer performances in basic attention, working memory, and disinhibition. Given that the current sample included older adults with no known cognitive impairment, findings suggest the CASI-R may detect subtle EF difficulties related to social functioning before more overt impairments emerge. Results also revealed that CASI-R item reliability and overall total scores do not significantly vary across ethnoracial groups; however, there is variability in the underlying factor structures. Although additional research is needed to clarify the cross-cultural nature of the CASI-R factors, findings indicate the CASI-R provides clinicians with a reliable measure for assessing the cognitive demands of social interactions in older adults and may assist with detecting early signs of cognitive difficulties across diverse ethnoracial communities.

Category: Aging

Keyword 1: mild cognitive impairment
Keyword 2: cognitive screening
Keyword 3: cross-cultural issues