INS NYC 2024 Program

Poster	Poster Session 08 Program Schedule 02/16/2024 01:45 pm - 03:00 pm Room: Shubert Complex (Posters 1-60) Poster Session 08: Cognition \| Cognitive Reserve Variables Final Abstract #12 Understanding the Impact of Physical Activity on the Unity and Diversity of Executive Function in Older Adults Gabriell Champion, Georgia State University, Department of Psychology, Atlanta, United States Elizabeth Tighe, Georgia State University, Department of Psychology, Atlanta, United States Vonetta Dotson, Georgia State University, Department of Psychology, Gerontology Institute, Atlanta, United States Keith McGregor, Birmingham VA GRECC; University of Alabama at Birmingham, Department of Clinical and Diagnostic Services, Birmingham, United States Category: Aging Keyword 1: executive functions Objective: Aging is associated with declines in executive function (EF) leading to difficulties performing everyday tasks. Compared to non-active peers, healthy older adults who are physically active tend to have better EF. However, we know of no cognitive models that incorporate the effects of specific lifestyle behaviors that may mitigate these differences. Miyake and Friedman employed confirmatory factor analysis to describe a Unity and Diversity (U & D) framework where EF can be divided into what is common (Common EF), and what is unique to specific processes (Updating, Shifting). Research examining the U & D framework in older adults is limited. The current study attempted to replicate the U & D model in older adults including physical activity as a predictor. We hypothesized that differences in the degree to which individuals’ EFs load onto the latent factors would be associated with differences in physical activity levels. Participants and Methods: We report on a sample of 223 healthy older adults aged 60 to 85 years (M = 70.98, SD = 7.03) taken from the Human Connectome Project in Aging. Two models were computed in Mplus 8.0 using six cognitive assessments to estimate the three latent variables: Common EF, Updating, and Shifting. Model 1 replicated Miyake and Friedman’s bifactor model where all cognitive assessments loaded on the Common EF and updating and shifting assessments loaded on the Updating and Shifting factors. Model 2, similar to Model 1, was a structural model that included the IPAQ (physical activity questionnaire) as a separate independent variable predicting the three latent factors. Factor variance was set to 1.0 to set the scale of measurement and factor loadings were equal for the Updating and Shifting factors for model identification. Results: For Model 1, both factor loadings for the Shifting factor were zero and nonsignificant. We respecified Model 1 (Model 1a) so Shifting was removed, and its indicators only loaded on the Common EF factor. Model 1a had excellent fit indices x² (15) = 7.271, p = .508, CFI = 1.000, RMSEA = .001, where five tasks loaded significantly on the Common EF and updating tasks loaded significantly on the Updating factor. Model 2 was adjusted to account for the changes in Model 1a and had good overall fit, x² (12) = 13.266, p = .350, CFI = .989, RMSEA = .022. Model 2 had similar significant loadings as in Model 1a, but the IPAQ did not significantly predict the latent factors. Conclusions: A modified version of the U & D model that did not include Shifting or the IPAQ was found to fit the data. This could be due to the assessments selected for this analysis or the U & D framework could differ in older adults. Future analyses will continue to examine the U & D model and modified versions in samples of older adults. Combining the U & D model with other lifestyle measures and neuroimaging data could give additional insights to the mechanisms underlying EF in older adults.

Poster

02/16/2024
01:45 pm - 03:00 pm
Room: Shubert Complex (Posters 1-60)

Poster Session 08: Cognition | Cognitive Reserve Variables

Final Abstract #12

Understanding the Impact of Physical Activity on the Unity and Diversity of Executive Function in Older Adults

Gabriell Champion, Georgia State University, Department of Psychology, Atlanta, United States
Elizabeth Tighe, Georgia State University, Department of Psychology, Atlanta, United States
Vonetta Dotson, Georgia State University, Department of Psychology, Gerontology Institute, Atlanta, United States
Keith McGregor, Birmingham VA GRECC; University of Alabama at Birmingham, Department of Clinical and Diagnostic Services, Birmingham, United States

Category: Aging

Keyword 1: executive functions

Objective:

Aging is associated with declines in executive function (EF) leading to difficulties performing everyday tasks. Compared to non-active peers, healthy older adults who are physically active tend to have better EF. However, we know of no cognitive models that incorporate the effects of specific lifestyle behaviors that may mitigate these differences. Miyake and Friedman employed confirmatory factor analysis to describe a Unity and Diversity (U & D) framework where EF can be divided into what is common (Common EF), and what is unique to specific processes (Updating, Shifting). Research examining the U & D framework in older adults is limited. The current study attempted to replicate the U & D model in older adults including physical activity as a predictor. We hypothesized that differences in the degree to which individuals’ EFs load onto the latent factors would be associated with differences in physical activity levels.

Participants and Methods:

We report on a sample of 223 healthy older adults aged 60 to 85 years (M = 70.98, SD = 7.03) taken from the Human Connectome Project in Aging. Two models were computed in Mplus 8.0 using six cognitive assessments to estimate the three latent variables: Common EF, Updating, and Shifting. Model 1 replicated Miyake and Friedman’s bifactor model where all cognitive assessments loaded on the Common EF and updating and shifting assessments loaded on the Updating and Shifting factors. Model 2, similar to Model 1, was a structural model that included the IPAQ (physical activity questionnaire) as a separate independent variable predicting the three latent factors. Factor variance was set to 1.0 to set the scale of measurement and factor loadings were equal for the Updating and Shifting factors for model identification.

Results:

For Model 1, both factor loadings for the Shifting factor were zero and nonsignificant. We respecified Model 1 (Model 1a) so Shifting was removed, and its indicators only loaded on the Common EF factor. Model 1a had excellent fit indices x² (15) = 7.271, p = .508, CFI = 1.000, RMSEA = .001, where five tasks loaded significantly on the Common EF and updating tasks loaded significantly on the Updating factor. Model 2 was adjusted to account for the changes in Model 1a and had good overall fit, x² (12) = 13.266, p = .350, CFI = .989, RMSEA = .022. Model 2 had similar significant loadings as in Model 1a, but the IPAQ did not significantly predict the latent factors.

Conclusions:

A modified version of the U & D model that did not include Shifting or the IPAQ was found to fit the data. This could be due to the assessments selected for this analysis or the U & D framework could differ in older adults. Future analyses will continue to examine the U & D model and modified versions in samples of older adults. Combining the U & D model with other lifestyle measures and neuroimaging data could give additional insights to the mechanisms underlying EF in older adults.