Virginia Cognitive Aging Project Research Articles

Testing the structural disconnection hypothesis: Myelin content correlates with memory in healthy aging

IntroductionThe "structural disconnection" hypothesis of cognitive aging suggests that deterioration of white matter (WM), especially myelin, results in cognitive decline, yet in vivo evidence is inconclusive. MethodsWe examined age differences in WM microstructure using Myelin Water Imaging and Diffusion Tensor Imaging in 141 healthy participants (age 20–79). We used the Virginia Cognitive Aging Project and the NIH Toolbox® to generate composites for memory, processing speed, and executive function. ResultsVoxel-wise analyses showed that lower myelin water fraction (MWF), predominantly in prefrontal WM, genu of the corpus callosum, and posterior limb of the internal capsule was associated with reduced memory performance after controlling for age, sex, and education. In structural equation modeling, MWF in the prefrontal white matter and genu of the corpus callosum significantly mediated the effect of age on memory, whereas fractional anisotropy (FA) did not. DiscussionOur findings support the disconnection hypothesis, showing that myelin decline contributes to age-related memory loss and opens avenues for interventions targeting myelin health.

Which variables moderate the relationship between depressive symptoms and global neurocognition across adulthood?

ABSTRACT The current study examined moderators of the relationship between depressive symptoms and global neurocognition in a large non-clinical community-dwelling sample spanning adulthood. Participants comprised 5,430 individuals between the ages of 18–99 years drawn from the Virginia Cognitive Aging Project. Depressive symptoms were measured via the Center for Epidemiologic Studies-Depression scale and neurocognition was operationalized as a composite variable comprising episodic memory, spatial visualization, processing speed, and reasoning tasks. Moderator variables included physical activity, cognitive activity, education, emotional stability, and openness. Hierarchical regressions were used to examine the influence of depressive symptoms and the moderators on neurocognition. Depressive symptoms significantly predicted neurocognition. Cognitive activity, years of education, and emotional stability moderated the depression-neurocognition relationship by buffering the impact of depressive symptoms on neurocognition. Cognitive activity engagement and level of education may function as a protective influence on those with higher levels of depressive symptoms, while emotional stability may be protective for individuals with lower levels of depressive symptoms. No differences in moderation were found across three age groups representing younger, middle, and older adults. Post-hoc analyses showed years of education and openness as moderators in a subsample excluding individuals with potentially clinically meaningful levels of depressive symptoms.

Does Neurocognition Predict Subjective Well-Being?

Greater subjective well-being (SWB) is associated with a myriad of positive outcomes across adulthood. While several studies have demonstrated a relationship between cognition and SWB, the current study extends previous work by examining the relationship between neurocognition and SWB across age and time. Data were drawn from 3,856 individuals between the ages of 18-99 years who participated in the Virginia Cognitive Aging Project, a prospective study of cognition in community dwelling adults. Participants completed a battery of neurocognitive tasks (assessing spatial visualization, episodic memory, reasoning, processing speed, and vocabulary) and measures assessing SWB (life satisfaction, positive affect, and negative affect). Results indicated that spatial visualization, episodic memory, and processing speed predicted life satisfaction only in specific age groups, but the magnitude of the coefficients were not significantly different between the groups, providing limited evidence of age moderation. Vocabulary was negatively associated with positive affect for all age groups. The temporal relationships between neurocognition and SWB were generally non-significant, and age did not moderate this relationship. Within the broader context of neurocognition, this study provides evidence that the relationship between neurocognition and SWB cross-sectionally may be partially age dependent for one facet of SWB, and the temporal relationship may be minimal.

Testing the Reserve Capacity Model: Does Race Moderate the Relationship Between Negative Emotions and Neurocognition?

To test whether race moderates the relationship between negative emotions and neurocognition by applying the reserve capacity model within a large sample that spans adulthood. The study sample (N = 1,020) consisted of community-dwelling adults between 18 and 84years of age who were drawn from the Virginia Cognitive Aging Project. Demographic variables were used to match a sample of Black participants to a sample of White participants. Race was examined as a moderator of the relationship between negative emotions (i.e., depressive symptoms, trait anxiety, and the negative affect subscale from the Positive and Negative Affect Schedule) and neurocognitive variables (episodic memory, reasoning, spatial visualization, and processing speed) with multiple-group structural equation modeling. After accounting for sociodemographic variables, depressive symptoms were negatively associated with processing speed in both groups, and with worse reasoning in the White subsample. Negative affect was associated with lower reasoning performance in both groups and with lower spatial visualization in the White subsample. Trait anxiety was not significantly associated with the neurocognitive constructs in either group. Multigroup structural equation models showed that the magnitudes of the associations were not different between the Black and White subsamples. Thus, race did not moderate the relationships between depressive symptoms, trait anxiety, and negative affect with neurocognition. Negative emotions are associated with lower performance on different neurocognitive tasks, but race does not moderate these relationships. Future research should examine perceived discrimination or other psychosocial variables when examining the relationships among negative emotions and neurocognition.

The temporal relationship between depressive symptoms and self-rated health across adulthood

Objective: Depressive symptoms have been found to relate to diminished self-rated health (SRH), which is a reliable index of general health. Despite such associations, there is limited research examining the bidirectional temporal relationship between these variables. The current study is the first to investigate the longitudinal relationship between depressive symptoms and SRH utilizing a cross-lagged panel analysis in a sample that spans adulthood (ages 18–93). Method: Data from the Virginia Cognitive Aging Project were used to examine the temporal relationship between depressive symptoms and SRH in a cross-lagged panel analysis using structural equation modeling. Results: A bidirectional temporal relationship, which was not moderated by age, was established between depressive symptoms and SRH. Conclusion: This article is the first to demonstrate that depressive symptoms and SRH influence each other reciprocally over time across adulthood, even after controlling for relevant variables. Considering the ubiquity and ramifications of depressive symptoms among American adults, these results highlight the importance of investigating mechanisms that could elucidate the link between the variables in question.

Disentangling the Influence of Data Contamination in Growth Curve Modeling: A Median Based Bayesian Approach

Growth curve models (GCMs), with their ability to directly investigate within-subject change over time and between-subject differences in change for longitudinal data, are widely used in social and behavioral sciences. While GCMs are typically studied with the normal distribution assumption, empirical data often violate the normality assumption in applications. Failure to account for the deviation from normality in data distribution may lead to unreliable model estimation and misleading statistical inferences. A robust GCM based on conditional medians was recently proposed and outperformed traditional growth curve modeling when outliers are present resulting in nonnormality. However, this robust approach was shown to perform less satisfactorily when leverage observations existed. In this work, we propose a robust double medians growth curve modeling approach (DOME GCM) to thoroughly disentangle the influence of data contamination on model estimation and inferences, where two conditional medians are employed for the distributions of the within-subject measurement errors and of random effects, respectively. Model estimation and inferences are conducted in the Bayesian framework, and Laplace distributions are used to convert the optimization problem of median estimation into a problem of obtaining the maximum likelihood estimator for a transformed model. A Monte Carlo simulation study has been conducted to evaluate the numerical performance of the proposed approach, and showed that the proposed approach yields more accurate and efficient parameter estimates when data contain outliers or leverage observations. The application of the developed robust approach is illustrated using a real dataset from the Virginia Cognitive Aging Project to study the change of memory ability.

The Relationship Between Social Support and Subjective Cognitive Functioning Across Adulthood.

Subjective cognitive functioning (SCF) is the self-appraisal of one's objective cognition, including both perceived cognitive ability and perceived cognitive impairment. Research has demonstrated perceived subjective cognitive impairment to be an important precursor to objective cognitive declines associated with increased age. Although previous work has shown a consistent positive relationship between social support and objective cognition throughout adulthood, few studies have examined the relationship between social support and SCF. Participants (N = 1,877; age range: 18-99) from the Virginia Cognitive Aging Project completed assessments of multiple domains of SCF and social support. Results demonstrated a significant association between negative support interactions and poorer SCF, beyond the influence of sociodemographic, well-being, and health factors. Exploratory analyses indicated that the magnitude of the relationships between social support and SCF were generally the same across age. These findings suggest that negative interactions may adversely impact an individual's self-appraisals of cognition.

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

ABSTRACT Using the bivariate dual change score approach, the present study investigated the directionality of the SMC-OMP association in a sample of healthy older adults (N = 2,057) from the Virginia Cognitive Aging Project. The sample was assessed throughout 10 years, five time points, and the impact of education, depressive symptoms, and low-memory functioning was tested. The Memory Functioning Questionnaire was used to assess SMC. There was a lack of longitudinal association with no significant coupling effects found between subjective and objective memory. After including depressive symptoms as a covariate, Frequency of Forgetting significantly predicted subsequent negative changes in OMP . A similar result was found for the low-memory functioning group after the inclusion of depression, with the frequency of memory complaints predicting subsequent memory decline . Our results do not support a predictive value of SMC without accounting for the influence of depressive symptoms and low-memory functioning.

Depressive symptoms as a predictor of memory decline in older adults: A longitudinal study using the dual change score model

The directionality of the longitudinal association between depressive symptoms and memory remains a topic of intense debate. A unidirectional association where depression impacts the change in memory (or vice-versa) and a bidirectional association where the trajectories of both dimensions affect each other lead to different clinical implications. Method: This study investigated the depression-memory longitudinal association in a sample of 2057 older adults aged between 60 and 99 years old from the Virginia Cognitive Aging Project. The bivariate dual change score model was used to investigate the directionality of the association between episodic memory and three dimensions of depression (somatic symptoms, depressed affect, and positive affect) throughout ten years (five measurement points), controlling for education and sex. Results: the bidirectional model showed the best fit between somatic symptoms and memory, with a significant coupling effect observed from initial somatic symptoms to subsequent changes in memory. For depressed and positive affect, the unidirectional model with initial levels of depression predicting following changes in memory showed the best fit to the data, with significant coupling effects observed. Higher initial levels of somatic symptoms and depressed affect predicted a subsequent decline in memory, and higher initial levels of positive affect predicted subsequent better memory performance. Statistical adjustments for covariates (education and sex) had no significant influence on these associations. Conclusions: Our findings support a unidirectional association with higher depressive symptoms preceding a steeper decline in memory in older adults. We discuss the clinical implications for depressive symptoms as a predictor of subsequent memory decline.

Mediators of the Relationship Between Cognition and Subjective Well-Being

Cognitive functioning has consistently found to predict subjective well-being (SWB), but it remains unclear why this relationship exists. Several potential explanations for this relationship have been offered in the literature, one of which is that a third variable accounts for this relationship. The current study examines this hypothesis by testing need for cognition (NfC), self-rated health, physical activity, cognitive activity, emotional stability, and conscientiousness as mediators of the relationship between cognitive ability (g) and indices of well-being. Data were drawn from the Virginia Cognitive Aging Project, a study of community-dwelling adults aged 18–99 (N = 4354). A composite variable representing cognition was created using measures of episodic memory, perceptual speed, reasoning ability, and spatial visualization. SWB was conceptualized as life satisfaction, positive affect, and negative affect; the three facets of SWB were treated as separate outcomes. Results indicate that NfC, self-rated health, emotional stability, and conscientiousness partially mediate the relationship between cognition and all facets of SWB. Neither physical activity nor cognitive activity mediated the relationship between cognition and SWB. The results of this study support the hypothesis that the cognition-SWB relationship can partially be explained by third variables and provide insight into the personality and individual difference characteristics that underlie this relationship.

Findings From the Virginia Cognitive Aging Project: Individual Differences, Well-Being, and Cognition

The Virginia Cognitive Aging Project (VCAP) is a cross-sectional and longitudinal study of cognitive functioning in a large sample of healthy community-dwelling adults between the ages of18-99 years (Salthouse,2009). Data are collected on several domains of cognitive functioning and subjective ratings of cognition, as well as a myriad of individual difference characteristics including self-reports of physical activity, cognitive activity, social support, personality, well-being, and affective measures. This symposium focuses on findings from VCAP that examine cross-sectional and longitudinal links between individual difference characteristics, indicators of well-being, and objective and subjective cognition. These topics include the cross-sectional assessment of >5,000 participants on the mediating role of Need for Cognition on the relationship between cognition and well-being (Yazdani & Siedlecki) and the relationship between social support and ratings of subjective cognition (Mueller & Minahan). Jung uses cross-lagged analyses to assess temporal relationships between physical and cognitive activity and cognition. Falzarano et al. present findings regarding the longitudinal relationship between subjective and objective measures of cognition. Finally, Minahan and Siedlecki present findings examining the temporal relationship between ratings of loneliness and depression over time. The symposium provides insights into the complex role of individual differences characteristics and cognitive functioning across the adult lifespan.

Investigating the Longitudinal Relationship Between Subjective Cognitive Complaints and Objective Cognition

Research examining the relationship between subjective cognitive complaints and objective cognitive performance has been mixed. Despite the lack of clear evidence demonstrating an association, subjective cognitive complaints are used as a criterion for the diagnosis of mild cognitive impairment and is considered a risk factor for Alzheimer’s disease. Cross-lagged panel analyses were used in the current study to examine the longitudinal relationships between subjective cognitive complaints (using the Memory Functioning Questionnaire) and objective cognition (e.g., reasoning, memory, spatial visualization, processing speed, and vocabulary) in healthy adults over 60 from the Virginia Cognitive Aging Project (N=441). Results indicated that objective and subjective cognition were only weakly related but that objective cognition is a stronger predictor of subjective cognitive complaints then vice versa. Although subjective cognitive complaints may be an early indicator of pathological aging, results indicate that subjective cognitive complaints may not be a valid predictor of objective cognitive decline.

A Cross-Lagged Panel Analysis of the Relationship Between Loneliness and Depression

Loneliness and depression have similar psychological features but are theoretically and statistically distinct (Hawkley et al., 2008). Most research has examined a unidirectional relationship between loneliness and depression (Cacioppo et al., 2006). Cacioppo et al. (2010), however, found that loneliness predicted subsequent depression, but depression did not predict subsequent loneliness. We extended this work by examining the reciprocal relationship between loneliness and depression in 1,560 healthy adults (ages ranging from 18-95 years at baseline) from the Virginia Cognitive Aging Project using cross-lagged panel analysis across three time points. Depression more strongly predicted subsequent loneliness (βs= 0.16 and 0.23, p < .05) than vice versa. These findings are consistent with the cognitive discrepancy theory, suggesting that affective processing, like depression, can negatively impact one’s perception of their social environment (Burholt & Scharf, 2014). Our findings highlight the need for increased longitudinal research examining the relationships among indices of well-being across adulthood.

Temporal Relationship Between Activity Engagement and Cognition

Previous studies have shown that activity engagement is related to cognitive function. However, few studies have examined the temporal order between activity engagement and various domains of cognition. Using data from the Virginia Cognitive Aging Project (baseline N =5430, Mage =51.28, SD =18.12), we examined the temporal relationships between engagement in physical and cognitive activity and different cognitive domains (reasoning, spatial visualization, episodic memory, processing speed, vocabulary) after controlling for age, education, self-rated health and depression. Cross-lagged panel analyses indicate that very few of the temporal relationships between activity level and cognition were significant except higher levels of cognitive activity significantly predicted better future processing speed, but not the reverse. Findings suggest the importance of engaging in cognitively stimulating activities, which help adults preserve processing speed over time. This study also highlights the importance of longitudinal design on various domains of cognition to help develop domain-specific interventions.

The Impact of Social Support on Subjective Cognition Across Adulthood

Increased age is associated with declines in objective cognition (OC). A related but distinct construct is subjective cognition (SC), which is an individual’s self-appraisal of their OC. Research shows that SC impairment is an important precursor to declines in OC (Sánchez-Benavidez et al., 2018). Research has also demonstrated a positive relationship between OC and social support (SS) across adulthood (La Fleur & Salthouse, 2017), but there is limited research on the relationship between SC and SS. Participants (N = 1,873; age range 18-99) from the Virginia Cognitive Aging Project completed assessments of multiple domains of SC, OC, and SS. Results from the current study showed a consistent, significant association between negative interactions with others and poorer SC (Betas ranged from -.077 to .103, p < .05), beyond the influence of sociodemographic, well-being, and health factors. Our findings suggest that negative interactions may adversely impact one’s self-appraisal of cognitive functioning.

Robust Bayesian growth curve modelling using conditional medians

Growth curve models have been widely used to analyse longitudinal data in social and behavioural sciences. Although growth curve models with normality assumptions are relatively easy to estimate, practical data are rarely normal. Failing to account for non-normal data may lead to unreliable model estimation and misleading statistical inference. In this work, we propose a robust approach for growth curve modelling using conditional medians that are less sensitive to outlying observations. Bayesian methods are applied for model estimation and inference. Based on the existing work on Bayesian quantile regression using asymmetric Laplace distributions, we use asymmetric Laplace distributions to convert the problem of estimating a median growth curve model into a problem of obtaining the maximum likelihood estimator for a transformed model. Monte Carlo simulation studies have been conducted to evaluate the numerical performance of the proposed approach with data containing outliers or leverage observations. The results show that the proposed approach yields more accurate and efficient parameter estimates than traditional growth curve modelling. We illustrate the application of our robust approach using conditional medians based on a real data set from the Virginia Cognitive Aging Project.

EXAMINING THE RELATIONSHIP BETWEEN COGNITIVE FUNCTIONING AND SUBJECTIVE WELL-BEING ACROSS AGE

Higher levels of subjective well-being (SWB) are associated with myriad of positive outcomes, including better physical health. Several variables have been shown to predict SWB, including cognitive functioning. The relationship between aspects of SWB (positive affect, negative affect, and life satisfaction) and cognition were examined in participants (N = 5, 125) between the ages of 18- 99 years from the Virginia Cognitive Aging Project (VCAP). Participants completed a battery of cognitive tasks, including tests of verbal episodic memory, processing speed, reasoning, spatial visualization, and vocabulary. Cross-sectional analyses were conducted using structural equation modeling, using full information maximum likelihood estimation. In the models, the five latent cognitive constructs simultaneously predicted each of the SWB outcome variables separately. Age, education, gender, and self-rated health were included as covariates. Results show that reasoning was a significant unique predictor of negative affect (-.30), vocabulary was a significant unique predictor of positive affect (-.21), and spatial visualization was a significant unique predictor of life satisfaction (.21). Age moderation was examined by dividing the sample into three age groups (younger, middle-aged, and older). There was some evidence of age moderation. Namely, spatial visualization was a significant unique predictor of life satisfaction in the younger sample only. Reasoning and processing speed predicted negative affect in the younger group, whereas only reasoning predicted negative affect in the older group. In conclusion, in a large community-based sample spanning adulthood, there is evidence that cognition predicts aspects of SWB but there is variation across SWB outcome variables, and across age.

INVESTIGATING THE RELATIONSHIP BETWEEN SOCIAL SUPPORT AND COGNITION AS MEDIATED BY HEALTH AND POSITIVE AFFECT

Decreased social networks are common in old age after major life events such as retirement, loss of loved ones, and declining health (Shankar et al., 2013). Diminished social ties are associated with increased feelings of loneliness and perceived isolation, which can have negative effects on cognition and physical health. The current study examines the relationship between social support (assessed via the Social Network Questionnaire) and overall cognitive performance (assessed as a latent construct comprising indicators that represent mean verbal episodic memory, processing speed, reasoning, and spatial visualization), and investigates positive affect and self-rated health as mediators of this relationship. The current study included 5,125 participants between the ages of 18-99 years from the Virginia Cognitive Aging Project (VCAP). Cross-sectional analyses were conducted using structural equation modeling. After controlling for age and education, results showed that a social support construct (comprising indicators representing each social network subscale) significantly and positively predicted cognitive performance (.59, p< .001). This relationship was reduced to .22 (p <.001) when positive affect was included as a mediator, and to .14 (p< .001) when self-rated health was included as a mediator. When the variables were included in a joint mediation model the relationship between social support and cognition was .20 (p < .001). Thus, health and positive affect are partial mediators of the relationship between social support and cognition and may help explain the relationship between social support and cognition. Furthermore, these findings provide additional evidence that social networks may play an important role in successful aging.

INVESTIGATING MODERATORS OF THE RELATIONSHIP BETWEEN SUBJECTIVE AND OBJECTIVE MEMORY

Subjective memory complaints (SMC) among older adults have been explored as an indicator of decline in objective memory functioning. While some research has found that SMC may be predictive of future cognitive impairment and dementia (Glodzik-Sobanska et al., 2007; Wang et al., 2004), others have suggested that SMC are common among healthy older adults (Cooper et al., 2011) and are not strongly related to objective memory performance. Researchers suggest that SMC may be more strongly related to affective factors (e.g., depression and anxiety; Rowell, Green, Teachman, & Salthouse, 2015). The current study examined the relationship between SMC, objective episodic memory performance (OEMP), along with depression and anxiety in a sample of 18-99 year olds (N = 5,430) from the Virginia Cognitive Aging Project (VCAP). Structural equation modeling with full information maximum likelihood estimation was used to investigate whether clinically-relevant depression and anxiety levels moderated the relationship between SMC and OEMP, controlling for age, education, gender, and health. OEMP was represented as a latent construct while the remaining variables were observed. Although depression and anxiety are significantly related to SMC (r’s = .29, .17, respectively), they are not correlated with OEMP. Furthermore, depression, but not anxiety, moderated the relationship between SMC and OEMP, such that those at risk for depression had a stronger relationship between SMC and OEMP (-.07, p<.05) compared to those not at risk (-.02, p=.31). This suggests that SMC may not be a valid indicator of OEMP as it may reflect variance from other sources, such as depression.

Examining Gender Differences in Neurocognitive Functioning Across Adulthood.

Previous research has shown that women have an advantage on verbal episodic memory and processing speed tasks, while men show an advantage on spatial ability measures. Previous work has also found differences in cognition across age. The current study examines gender differences in neurocognitive functioning across adulthood, whether age moderates this effect, and whether these differences remain consistent with practice across multiple testing sessions. Data from the Virginia Cognitive Aging Project were used, which included participants between the ages of 18 and 99 years (N = 5125). Participants completed measures assessing five cognitive domains: episodic memory, processing speed, reasoning, spatial visualization, and vocabulary. Results showed that gender was significantly related to memory, speed, and spatial visualization, but not to vocabulary or reasoning. Results of invariance analyses across men and women provided evidence of configural and metric invariance, along with partial scalar invariance. Additionally, there was little evidence that age or practice influenced the gender effect on neurocognition. Consistent with the previous research, these results suggest that there is a female advantage in episodic memory and processing speed, and a male advantage in spatial visualization. Gender was shown to influence cognition similarly across adulthood. Furthermore, the influence of gender remained the same across three sessions, which is consistent with the previous work that has shown that training does not differentially impact performance on spatial ability measures for females compared to males.