Biopsychosocial contexts influence adult cognitive function concurrently and longitudinally

Abstract

Background: Cognitive aging is a complex process that impacts human behavior. Identifying the factors that preserve cognitive functioning is a public health priority, given that 20% of the US population will be at least 65 years old in the next decade. Biopsychosocial determinants of cognitive decline across the lifespan are often examined as ecological factors that independently moderate cognitive aging, despite the known complexity surrounding these relationships.

Objective: We aimed to address this gap by exploring the synergistic and simultaneous relationship between risk and protective factors on cognitive functioning.

Method: Using the MIDUS study datasets, we examined the relationships among physiological markers, friendship quality, and global cognition functioning, concurrently and longitudinally over ten years. Our participants included 929 healthy (417 men, 512 women) adults (average age at Time 1: 54.6 ± 11.6 years). Exploratory analyses examining the effects of racial minority status were also conducted.

Results: Cross-sectionally, age, and friendship quality moderated the relationship between vagally-mediated heart rate variability (vm-HRV) and cognition such that younger adults with greater friendship quality had a negative relationship between vm-HRV and cognitive performance; our unexpected finding suggests the heart-brain relationship is sensitive to the biopsychosocial environment. Longitudinally, higher IL-6 levels at Time 1 predicted poorer cognitive performance a decade later, but only among those with greater levels of friendship quality, especially for white-identifying individuals.

Conclusions: The relationships among physiological risk factors, social protective factors and cognitive functioning appear to be temporally different during mid-adulthood. Given many of the whole sample findings were not replicated within the racial minority subgroup, we suggest that these relationships should be examined in a larger and more diverse racial minority sample to determine whether this study lacked the power necessary to detect a relationship or if the relationships are in fact different by racial minority sub-group. In addition, future research should overcome the study's reliance on healthy adults and self-report measures of friendship quality by including adults with pre-existing cognitive impairments, and employing more real-time measures of friendship quality, such as daily diary or ecological momentary assessment.

Keywords: Aging; Biomarkers; CRP; Cognitive functioning; Friendship quality; HRV; IL-6.

Fig. 1

Proposed model of friendship quality moderating the relationship between physiological markers and cognitive functioning. Note. Given age is highly predictive of cognitive functioning, age was entered as a moderator in concurrent analyses as age may contextually shift the relationship among physiological markers, friendship quality, and cognitive functioning simultaneously. However, in the longitudinal analyses, age was entered as a covariate because cognitive functioning at time 1 was also included as covariate. Age was not expected to moderate the relationship among physiological markers, friendship quality, and cognitive functioning decline. The gray color highlights how age was incorporated into the model depending on the type of analysis, concurrent or longitudinal.

Fig. 2

Simple Slopes Showing a 3-Way Interaction among vm-HRV, Friendship Quality (FQ), and Age for Overall Sample on Concurrent Cognitive Functioning as well as a 2-Way Interaction between vm-HRV and Age on Concurrent Cognitive Functioning for the Racial Minority Sub-group Only. Note. *p < .05, **p < .01. Using PROCESS macro model 3, the 3-way interaction for the overall sample significantly predicted cognitive functioning when controlling for racial minority status, gender, education, and cardiovascular medication use. For the line representing younger adults with higher friendship quality, vm-HRV was negatively associated with cognitive functioning (β = −.164, p = .03). The other 3 lines' slopes indicating varying levels of friendship quality by age were not different than 0. The analyses were repeated by racial minority status. The White sub-sample analysis mirrored the overall sample findings; however, the racial minority sub-sample revealed a significant vm-HRV by age interaction, instead of a 3-way interaction with friendship quality. For the line indicating the younger racial minority sub-sample, vm-HRV was negatively associated with cognitive functioning (β = −.383, p = .008), while the slope of the line representing their older counterparts was not different than 0. vm-HRV = vagally mediated-heart rate variability, M2 = MIDUS wave 2 data collection, FQ = Friendship quality, SD = standard deviation.

Fig. 3

Simple slopes Showing a 2-way interaction between CRP and age on concurrent cognitive functioning by racial minority status. Note. *p < .05. Using PROCESS macro model 1, the 2-way interaction for the overall sample significantly predicted cognitive functioning when controlling for racial minority status, gender, education, and inflammation-related medication use. When further investigating the 2-way interaction by racial minority status, the interaction was only significant for the White sub-sample, not the racial minority sub-sample. For the line representing younger white adults, CRP was negatively associated with cognitive functioning (β = −.066, p = .049). The other 3 lines' slopes were not different than 0. CRPC-reactive protein, M2 = MIDUS wave 2 data collection, SD = standard deviation.

Fig. 4

Simple slopes Showing a 2-way interaction between IL-6 and friendship quality (FQ) on cognitive functioning a decade later by racial minority status. Note. *p < .05. Using PROCESS macro model 1, the 2-way interaction for the overall sample predicted M3 cognitive functioning when controlling for M2 cognitive functioning, M3 age, minority status, gender, education, and inflammation-related medication use. When further investigating the 2-way interaction by racial sub-group, the interaction was only significant for the White sub-sample, not the racial minority sub-sample. For the line representing white individuals with higher friendship quality, higher M2 IL-6 predicted poorer M3 cognitive functioning (β = −0.069, p = .047). The other 3 lines' slopes were not different than 0. IL-6 = Interleukin-6, M3 = MIDUS wave 3 data collection, M2 = MIDUS wave 2 data collection, SD = standard deviation.