The bidirectional relationship between physical health and memory.

Abstract

Individuals typically experience changes in physical health and cognitive ability across the life span. Although these constructs dynamically relate to one another, the temporal ordering of dynamic changes in physical health and cognitive ability is not well-established. Therefore, we examined the temporal ordering of the dynamic, bidirectional relationship between physical health and memory across ages 50 -87 with Bivariate Dual Change Score Models (BDCSM). Employing a model-comparison approach, we tested whether inclusion of specific directional coupling parameters resulted in a meaningful improvement in model fit, controlling for education, gender, and race. The current sample included 9,103 individuals who participated in Waves 4 -11 (1998Waves 4 -11 ( -2012 of the Health and Retirement Study. Results indicated that both memory and physical health declined across ages 50 -87. Furthermore, level of memory at a given time point was positively associated with subsequent change in physical health, meaning higher memory was linked to less decline in physical health by the subsequent time point. The opposite effect, namely physical health predicting memory, was much weaker. Age differences were also evident in the bidirectional coupling model, indicating that old-old individuals (i.e., ages 75-87) exhibited a much stronger coupling effect from memory to change in physical health than younger individuals (i.e., ages 50 -74). In conclusion, memory buffers decline in physical health across mid-tolater life, and this effect is especially strong at older ages.

COVID-19 lockdowns’ effects on the quality of life, perceived health and well-being of healthy elderly individuals: A longitudinal comparison of pre-lockdown and lockdown states of well-being

“…A previous study reported an association between memory difficulties and worsened self-perceived health due to more limitations on activities of daily living and due to the fear of developing dementia ( Montejo et al, 2012 ). Lower memory capacity has also been linked to a decline in physical and self-perceived health, especially for those people between 75 and 87 years old ( Nelson et al, 2020 ). Accordingly, we hypothesize that those with troubles with their memory before the stay-at-home orders may have had worsened memory in the last year and, being more cognitively vulnerable, had therefore worsened perceived health.…”

Longitudinal association between subjective and objective memory in older adults: a study with the Virginia Cognitive Aging Project sample

“…Model fit was evaluated based on the guidelines proposed by Hu and Bentler (1999) for good fit. To compare the four models and choose the final, we adopted the same holistic approach proposed by Nelson, Jacobucci, Grimm, and Zelinski (2020), which includes the analysis of the Akaike Information Criterion (AIC) (Akaike, 1987) and the Bayesian Information Criterion (BIC) (Raftery, 1995) with lower values indicating better model fit. Following Nelson et al (2020) approach, we defined meaningful improvements in model fit as a difference in the model's information criteria higher than 10 points.…”

Longitudinal association between subjective and objective memory in older adults: a study with the Virginia Cognitive Aging Project sample

“…Model fit was evaluated based on the guidelines proposed by Hu and Bentler (1999) for good fit. To compare the four models and choose the final, we adopted the same holistic approach proposed by Nelson, Jacobucci, Grimm, and Zelinski (2020), which includes the analysis of the Akaike Information Criterion (AIC) (Akaike, 1987) and the Bayesian Information Criterion (BIC) (Raftery, 1995) with lower values indicating better model fit. Following Nelson et al (2020) approach, we defined meaningful improvements in model fit as a difference in the model's information criteria higher than 10 points.…”

“…In this section, the directionality of the association between subjective and objective memory is reported. The holistic approach protocol employed by Nelson et al (2020) was adopted to choose the final model. The estimates for each parameter of the four models (no coupling, unidirectional models, and full coupling) are shown in Table 4.…”