Interactive Effects of Physical Activity and Cognitive Activity on Cognition in Older Adults Without Mild Cognitive Impairment or Dementia

Abstract

Objective: The purpose of this secondary analysis was to test effects of interactions between accelerometer-measured physical activity and self-reported cognitive activity on cognition in older adults without cognitive impairment. Method: Participants were 742 older adults from the Rush Memory and Aging Project who completed annual clinical evaluations. A series of parallel growth models tested effects of interactions between physical activity and cognitive activity on cognition (global index, five domains) at Year 5, controlling for demographics, health factors, and corresponding cognition measures at Year 1. Results: Results were mixed, with significant physical and cognitive activity interactive effects for working and semantic memory. In models without interactions, higher physical and cognitive activities at Year 1 and less decline in cognitive activity over time were independently associated with better cognition at Year 5. Discussion: These findings may inform interventions that enhance physical and cognitive activities to prevent cognitive impairment in older adults.

Keywords

exercise physical activity cognition health behavior accelerometer

A growing body of research suggests that certain health behaviors can alter the trajectory of cognitive decline, perhaps by increasing cognitive reserve or compensatory resources that improve the brain’s resilience to damage or disease (Stern, 2012). Two key behaviors that have shown promise in promoting cognitive health include physical activity and cognitive activity (Brasure et al., 2018; Halloway, Wilbur, Schoeny, & Arfanakis, 2017; Lampit, Hallock, & Valenzuela, 2014). Physical activity includes bodily movements and can involve structured exercise, which is planned, repetitive, and purposive (e.g., aerobics class), or lifestyle physical activity, which includes both planned activities and unstructured activities (e.g., household chores, walking; Centers for Disease Control and Prevention, 2015). Cognitive activity involves participation in cognitively stimulating activities with little physical demand (e.g., reading; Bennett, Arnold, Valenzuela, Brayne, & Schneider, 2014).

Prior studies have often examined physical activity and cognitive activity with cognition separately (Bielak, Gerstorf, Anstey, & Luszcz, 2014; de Frias & Dixon, 2014; Hughes, Becker, Lee, Chang, & Ganguli, 2015; Lee, 2014; Rajan et al., 2015; Sturman et al., 2005; Verghese et al., 2003; Wang et al., 2013; Yu, Ryan, Schaie, Willis, & Kolanowski, 2009; Zhu, Qiu, Zeng, & Li, 2017) and have generally reported that more activity is associated with better cognitive function (Bielak et al., 2014; de Frias & Dixon, 2014; Lee, 2014; Rajan et al., 2015; Verghese et al., 2003; Wang et al., 2013; Zhu et al., 2017) or a lower risk of mild cognitive impairment (MCI) or dementia (Hughes et al., 2015; Verghese et al., 2003). It is possible, however, that physical activity and cognitive activity do not just affect cognitive function independently, but have complementary effects. The American Heart Association recently stated that maintaining lifestyle health behaviors over time is vital for preventing cognitive impairment, and that participating in multiple behaviors, such as physical activity and cognitive activity, may have interactive effects (Gorelick et al., 2017). In other words, the full effects of each behavior may depend on the presence of other behaviors. But, relatively few studies have tested this hypothesis, and findings have been mixed. To our knowledge, only five longitudinal studies examined the combined effects of physical activity and cognitive activity, and time spent in both types of activities was summed (Bielak et al., 2014; de Frias & Dixon, 2014; Hughes et al., 2015; Wang et al., 2013; Zhu et al., 2017). We are aware of only two longitudinal studies that examined the interactive effects of physical activity and cognitive activity (Lee, 2014; Yu et al., 2009) and neither found significant interactive effects on cognitive function. However, these studies examined physical activity and cognitive activity only at baseline. Given the well-established finding that participation in physical and cognitive activities decreases with age, further contributing to cognitive decline (Gorelick et al., 2017), studies are needed that examine how change in these activities over time affect cognition. Other limitations of prior research in this area include the use of self-report measures of physical activity, which are prone to overestimation (Troiano et al., 2008), and focus on structured exercise only. Lifestyle physical activity, such as household or occupational, captures the full spectrum of activities and provides information beyond participation in structured exercise. Furthermore, a device measuring physical activity (e.g., accelerometer) yields more accurate estimates of physical activity (Aadland & Ylvisåker, 2015), compared with self-report measures, and measures all lifestyle physical activity during daily life.

The purpose of this secondary analysis was to examine the effects of the interactions (both changes over time and baseline levels) between accelerometer-measured physical activity and self-reported cognitive activity on cognitive function in older adults without cognitive impairment. Cognitive function was measured using a validated battery of performance-based tests from which a global composite of cognition and five separate domains were derived. A secondary research question sought to explore the unique contributions of physical activity and cognitive activity on the cognitive function outcomes.

Method

This is a secondary analysis of data from the Rush Memory and Aging Project, an ongoing epidemiological study following participants on a yearly basis (Bennett et al., 2018). The project began in 1997, and accelerometer data were added in 2005. In this study, we used longitudinal data over a 4-year period. The first time point was the first assessment during which both cognitive function and accelerometer data were obtained.

Participants

Participants from this data set were recruited from retirement communities or subsidized housing facilities in the Chicago metropolitan area (Bennett et al., 2018). All participants provided written informed consent for data collection, including agreeing to annual clinical evaluations and organ donation. The study was approved by the Rush University Medical Center Institutional Review Board, and a waiver of consent was obtained for the current data analysis.

Participants underwent standardized clinical evaluations, including medical history and neurological examinations (Bennett et al., 2018). The only Rush Memory and Aging Project inclusion criteria are age 65 years or above and cognitive ability sufficient to complete the informed consent process at study enrollment. For inclusion in the current analysis, participants must have completed assessments during at least three consecutive time points in a 4-year period. Exclusion criteria for this study were dementia, MCI, or inadequate accelerometer data (Buchman et al., 2012).

Dementia was diagnosed using a three-step process. First, the battery of cognitive tests were scored by computer according to algorithms adjusted for years of education, as previously reported. Results from the cognitive testing were then reviewed by a board-certified or board-eligible neuropsychologist for signs of cognitive impairment. Finally, a clinician with experience diagnosing dementia in older adults (geriatrician, neurologist, or geriatric nurse practitioner) reviewed all available data and rendered a final diagnosis. Clinical judgment of Alzheimer’s dementia follows the recommendation of the joint working group of the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer’s Disease and Related Disorders Association (NINDS-ADRDA; McKhann et al., 1984). Participants who were cognitively impaired but did not meet the criteria for Alzheimer’s dementia were given the diagnosis of MCI, as previously described (Boyle, Wilson, Aggarwal, Tang, & Bennett, 2006).

A total of 1,392 participants completed a clinical evaluation and had valid Actical accelerometer data for the analytic baseline (Figure 1). Of these, 320 were excluded because of MCI and 39 because of Alzheimer’s dementia or other dementias. Another 291 were excluded because of missing data on key covariates. A total of 742 participants with complete questionnaire, accelerometer, and neurocognitive test data over a 4-year period were thus included in this study.

Figure 1.

Flowchart of subject participation.

Measures

Physical activity

Lifestyle physical activity was measured using an Actical (Mini Mitter, Bend, OR) piezoelectric triaxial accelerometer. The Actical is a battery-operated activity monitor worn on the wrist similar to a wrist watch, is waterproof, and can be worn while bathing or swimming. The Actical provides valid estimates of free-living physical activity (Crouter et al., 2011). Participants were instructed to wear the Actical continuously for 10 days on the nondominant wrist, starting at the time of the annual data collection appointment. After research staff received the Actical, raw data were downloaded and viewed using Respironics, Inc. (Bend, OR) software.

Data then were partitioned into 24-hr periods starting from 7 p.m. on the day of device placement to the time of retrieval. Only data from complete 24-hr periods were used to determine the average total daily lifestyle physical activity. Activity counts represented the area calculated by integrating the activity curve for each epoch, where nonzero values reflected activity. Activity counts were summed for each epoch (15 s; Buchman et al., 2012). Total daily lifestyle physical activity was the sum of all activity counts during a 24-hr period, averaged over the 7 days used for analysis. To correct for positively skewed data, total daily lifestyle physical activity was log-transformed. To facilitate presentation and interpretation of results, total daily physical activity was converted to a z score (Halloway, Arfanakis, Wilbur, Schoeny, & Pressler, 2018).

Cognitive activity

Cognitive activity was assessed utilizing a structured, self-report questionnaire (Appendix A). The questionnaire included seven activities involving information processing with minimal physical or social demand: reading magazines, reading books, reading newspapers, writing letters, visiting a library, attending a play, and playing games (e.g., chess, checkers, cards). A composite measure of cognitive activity was calculated by averaging the item scores (range = 1-5). Higher cognitive activity scores using this questionnaire were previously associated with slower cognitive decline and lower risk of MCI and Alzheimer’s dementia (Wilson, Segawa, Boyle, & Bennett, 2012).

Cognitive function

Cognitive function was assessed using a battery of 19 neurocognitive tests, which were conducted annually (Bennett et al., 2018). The tests evaluated five cognitive domains: episodic memory (Logical Memory Immediate Recall, Logical Memory Delayed Recall, Word List Memory, Word List Recall, Word List Recognition, East Boston Memory Test, East Boston Delayed Recall), semantic memory (Verbal Fluency, 15-item Boston Naming Test, 15-item Reading Test), working memory (Digit Span Forward, Digit Span Backward, Digit Ordering), perceptual speed (Number Comparison, Symbol Digit Test Modality, Stroop Word Reading, Stroop Color Naming), and visuospatial ability (Line Orientation, Progressive Matrices). Raw scores of all tests were converted to z scores. Composite scores for each domain were calculated by averaging z scores of the individual tests (Bennett et al., 2018). Global cognition was calculated by averaging z scores for the five domains.

Covariates

Covariates included demographics, depressive symptoms, vascular risk factors, and vascular disease burden. Demographics, including age, race, sex, and education, were documented at study entry. Depressive symptoms were assessed with a modified, 10-item version of the Center of Epidemiologic Studies Depression Scale (Bennett et al., 2018). The score (range = 0-10) was the total number of symptoms, with higher scores indicating more depressive symptoms. Vascular risk factors (hypertension, diabetes mellitus, and smoking history) and vascular disease burden (stroke, claudication, heart conditions, and congestive heart failure) were determined by self-reported yes/no questions and medication assessment. Each item was rated as absent (0) or present (1). The cumulative score for vascular risk factors ranged from 0 to 3, with higher scores indicating greater risk (Aggarwal et al., 2006), whereas the cumulative score for vascular disease burden ranged from 0 to 4, with higher scores indicating greater disease burden (Boyle, Buchman, Wilson, Leurgans, & Bennett, 2009).

Data Analysis

We employed latent curve growth modeling (LGCM) to examine longitudinal trajectories of physical activity and cognitive activity over 4 years. LGCM is a multivariate statistical method of structural equation modeling that allows modeling of repeated measures data to estimate changes in individual characteristics reflecting intraindividual growth patterns over time (Curran, Obeidat, & Losardo, 2010). For all models, standardized estimates are presented to aid interpretation of the relative strength of associations.

Parallel LGCM was used to examine the simultaneous longitudinal trajectories of physical activity and cognitive activity. The models included intercepts (level at Year 1) and linear time slopes (change per year) for physical activity and cognitive activity. We tested a series of two models: one series of models included interaction terms between physical activity and cognitive activity intercepts, and one series of models included interaction terms between physical activity and cognitive activity slopes. All models controlled for demographics, depressive symptoms, vascular risk factors, vascular disease burden, and a measure of the corresponding cognitive function at Year 1. Then, we tested models that included physical activity and cognitive activity intercepts and slopes, but excluded interactions, and controlled for covariates and the corresponding cognitive function measure at Year 1.

Descriptive statistics for baseline characteristics of the study sample were derived using SPSS Version 23 (IBM Corp., 2015). Mplus Version 8.0 (Muthén & Muthén, 2017) was used for LGCM. The p values <.05 were considered statistically significant.

Results

Of the 742 participants, 75.7% were female (Table 1). Their mean age was 79.3 years (SD = 7.1) and mean education level was 15 years (SD = 7.1). Participants reported low levels of depressive symptoms (M = 0.8 [SD = 1.4]). Of the 742 participants, 49 (6.6%) had scores indicating a high risk for major depression (Kohout, Berkman, Evans, & Cornoni-Huntley, 1993). Participants wore the Actical for a mean of 9.3 days (SD = 1.1). Compared with females, males had significantly more education (16.0 vs. 14.8 years), t(740) = −4.70, p < .001; fewer depressive symptoms (0.6 vs. 0.9), t(740) = 2.5, p = .013; less weekly cognitive activity (3.1 vs. 3.2), t(740) = 2.0, p = .042; and worse global cognition (0.1 vs. 0.2), t(740) = 3.4, p = .001; particularly worse episodic memory (0.1 vs. 0.3), t(740) = 5.9, p < .001; worse semantic memory (0.1 vs. 0.3), t(738) = 2.5, p = .012; and worse perceptual speed (0.03 vs. 0.2), t(737) = 2.6, p = .010. However, males had better visuospatial ability (0.5 vs. 0.1), t(737) = −5.6, p < .001.

Table 1.

Characteristics of Older Adults Without Mild Cognitive Impairment or Dementia.

Characteristic	Total(N = 742)		Female(n = 562)		Male(n = 180)
Characteristic	M (N)	SD (%)	M (n)	SD (%)	M (n)	SD (%)
Age (years)	79.3	7.3	79.2	7.1	79.6	7.0
Race (% non-Latino White)	707	95.3	532	94.7	175	97.2
Education (number of years completed)	15.1	7.1	14.8	2.8	16.0	3.3
Depressive symptoms (sum score)^a	0.8	1.9	0.9	1.5	0.6	1.1
Vascular risk factors^b	1.1	0.6	1.1	0.8	1.2	0.9
Vascular disease burden^c	0.4	0.7	0.4	0.7	0.4	0.7
Total daily physical activity (Year 1)^d	2.8	1.5	2.9	1.5	2.7	1.6
Cognitive activity (Year 1)^e	3.2	<0.1	3.2	0.6	3.1	0.6
Episodic memory (Year 5)	0.1	0.1	0.3	0.7	<0.1	0.7
Semantic memory (Year 5)	0.1	0.1	0.3	0.9	0.3	0.6
Working memory (Year 5)	0.1	0.6	0.1	0.8	0.1	0.9
Perceptual speed (Year 5)	<−0.1	0.8	<0.1	0.9	−0.2	0.9
Visuospatial ability (Year 5)	0.1	0.1	0.1	0.8	0.4	0.8
Global cognition (Year 5)	0.1	0.1	0.2	0.7	<0.1	0.7

Note. Year 1: n = 742, Year 2: n = 527, Year 3: n = 646, Year 4: n = 551, Year 5: n = 456.

Higher score indicates more depressive symptoms.

Range from 0 to 3.

Range from 0 to 4.

Total physical activity counts.

Mean score, ranging from 1 to 5.

Two separate series of models testing effects on the six cognitive function outcomes at Year 5 were conducted (for a total of 12 models). The first series of models tested the effects of the interaction between intercepts, and the second series of models tested the effects of the interaction between slopes. Both series of models included terms for physical activity intercept and slope, cognitive activity intercept and slope, covariates, and the corresponding cognitive function measure at Year 1. Across all six models that tested the effect of the interaction between intercepts, the interaction was a significant predictor of higher levels of only working memory at Year 5 (γ = 0.08, SE = 0.04, p = .040; Table 2). Across all six models that tested the effect of the interaction between slopes, the interaction was a significant predictor of higher levels of only semantic memory at Year 5 (γ = 0.13, SE = 0.14, p < .001; Table 3).

Table 2.

Results of Parallel Growth Models Testing the Effects of the Interaction Between Physical Activity and Cognitive Activity Intercepts on Cognition in 742 Older Adults Without Mild Cognitive Impairment or Dementia.

Variable	Cognitive function outcomes
	Episodic memory			Semantic memory			Working memory			Perceptual speed			Visuospatial ability			Global cognition
	Estimate^a	SE	p	Estimate^a	SE	p	Estimate^a	SE	p	Estimate^a	SE	p	Estimate^a	SE	p	Estimate^a	SE	p
PA intercept^b	0.48	0.26	.065	0.08	0.23	.747	−0.40	0.23	.093	−0.15	0.25	.561	0.01	0.01	.372	−0.04	0.22	.842
PA slope^b	−0.05	0.07	.428	0.01	0.06	.816	0.05	0.07	.512	0.04	0.07	.591	−0.02	0.02	.197	<0.01	0.06	.997
CA intercept^b	0.06	0.05	.173	0.06	0.04	.123	0.02	0.05	.642	<−0.01	0.05	.980	0.01	<0.01	.019	0.03	0.04	.399
CA slope^b	0.14	0.06	.014	0.16	0.06	.006	0.02	0.06	.683	0.19	0.06	.001	0.04	0.02	.125	0.13	0.05	.011
PA Intercept × CA Intercept	−0.06	0.04	.176	<−0.01	0.04	.925	0.08	0.04	.040	0.03	0.04	.357	<0.01	<0.01	.532	0.02	0.03	.508

Note. PA = physical activity; CA = cognitive activity.

For all models, standardized estimates are presented to aid interpretation of the relative strength of associations.

Main effects.

Table 3.

Results of Parallel Growth Models Testing the Effects of the Interaction Between Physical Activity and Cognitive Activity Slopes on Cognition in 742 Older Adults Without Mild Cognitive Impairment or Dementia.

Variable	Cognitive function outcomes
	Episodic memory			Semantic memory			Working memory			Perceptual speed			Visuospatial ability			Global cognition
	Estimate^a	SE	p	Estimate^a	SE	p	Estimate^a	SE	p	Estimate^a	SE	p	Estimate^a	SE	p	Estimate^a	SE	p
PA intercept^b	0.15	0.05	.003	0.03	0.05	.554	0.07	0.06	.200	0.08	0.05	.075	0.08	0.05	.117	0.01	179	.955
PA slope^b	−.05	0.06	.438	0.26	0.19	.173	0.13	0.13	.330	0.01	0.08	.857	−0.09	0.08	.263	−0.01	1.74	.997
CA intercept^b	0.06	0.05	.230	0.04	0.04	.223	0.03	0.05	.622	0.01	0.05	.869	0.11	0.05	.023	0.04	0.30	.907
CA slope^b	0.14	0.25	.576	1.08	.31	.001	0.23	0.26	.378	0.06	0.18	.075	0.27	0.28	.349	0.15	10.38	.988
PA Slope × CA Slope	<−0.01	0.18	.991	0.38	0.07	<.001	0.13	0.14	.371	−.08	0.12	.498	0.11	0.20	.572	0.02	7.38	.998

Note. PA = physical activity; CA = cognitive activity.

For all models, standardized estimates are presented to aid interpretation of the relative strength of associations.

Main effects.

Due to the small number of significant interaction effects, we tested parallel growth models that included physical activity and cognitive activity intercepts and slopes but excluded interaction terms. Again, these models controlled for covariates and the corresponding cognitive function measure at Year 1 (Table 4). These showed significant positive effects of physical activity intercept on episodic memory and global cognition at Year 5, as well as significant positive effects of cognitive activity intercept on visuospatial ability at Year 5. On average, participants showed declines in both physical activity and cognitive activity. However, slower rates of decline in cognitive activity were significantly associated with higher episodic memory, semantic memory, perceptual speed, and global cognition at Year 5, controlling for levels at Year 1. Thus, the independent effects of physical activity and cognitive activity remained significant when controlling for the other behavior. There were no significant associations between physical activity and cognitive activity intercepts or between physical activity and cognitive activity slopes.

Table 4.

Results of Parallel Growth Models Testing the Effects of Physical Activity and Cognitive Activity on Cognition in 742 Older Adults Without Mild Cognitive Impairment or Dementia.

Variable	Cognitive function outcomes
	Episodic memory			Semantic memory			Working memory			Perceptual speed			Visuospatial ability			Global cognition
	Estimate^a	SE	p	Estimate^a	SE	p	Estimate^a	SE	p	Estimate^a	SE	p	Estimate^a	SE	p	Estimate^a	SE	p
PA intercept^b	0.14	0.06	.017	0.03	0.06	.609	0.08	0.06	.483	0.05	0.08	.514	0.01	0.01	.372	0.11	0.05	.026
PA slope^b	−0.06	0.08	.461	0.17	0.08	.845	0.05	0.08	.569	0.05	0.10	.602	−.0.02	0.02	.197	−0.01	0.07	.835
CA intercept^b	0.06	0.05	.222	0.07	0.04	.086	0.03	0.05	.483	0.14	0.05	.748	0.01	<0.01	.019	0.06	0.04	.087
CA slope^b	0.14	0.07	.028	0.16	0.06	.010	0.02	0.06	.703	0.19	0.06	.002	0.04	0.02	.125	0.17	0.06	.003

Note. PA = physical activity; CA = cognitive activity.

For all models, standardized estimates are presented to aid interpretation of the relative strength of associations.

Main effects.

Discussion

In a series of growth models, we tested the effects of two interactions on the six cognitive function outcomes at Year 5: interaction between physical activity and cognitive activity at the first time point and interaction between changes in physical activity and changes in cognitive activity. Our results were mixed. Of the 12 models, only two contained significant interaction effects. We found significant effects of the interaction between physical activity and cognitive activity intercepts on higher levels of working memory and significant effects of the interaction between physical activity and cognitive activity slopes on higher levels of semantic memory.

Unlike the two earlier longitudinal studies that failed to find interaction effects (Lee, 2014; Yu et al., 2009), our results were mixed, with only two significant interaction effects. Despite the few interactive effects observed in our study, the observations of significant independent effects for both physical activity and cognitive activity imply that these activities should be addressed as distinct behaviors when considering cognitive function. Findings from intervention research also suggest potential interactive effects. For example, physical activity and cognitive activity intervention studies of healthy older adults have reported greater effects of cognitive activity on cognitive function among participants who exhibited greater changes in physical activity (Lauenroth, Ioannidis, & Teichmann, 2016). Although we had few interaction effects, the results from intervention research suggest that changes in physical activity and cognitive activity may potentially affect cognitive function synergistically; that is, the full effects of each behavior may depend on the presence of other behaviors.

To our knowledge, this is the first study to test interactive effects of changes over time in both physical activity and cognitive activity, instead of focusing on baseline levels alone. Furthermore, no previous study utilized an accelerometer to accurately measure total lifestyle physical activity. It is crucial to assess total lifestyle physical activity, rather than focusing on structured exercise, as lifestyle activities (e.g., household, occupational, transportation-related activities) have important health benefits, such as improving cardiovascular health and overall survival. Moreover, interventions emphasizing lifestyle physical activity are more appealing for older adults (Darden, Richardson, & Jackson, 2013). Nevertheless, existing physical and cognitive activity interventions mostly neglect lifestyle physical activity and focus on only structured exercise sessions (Lauenroth et al., 2016).

This secondary analysis had limitations. First, the lack of significant findings regarding the interactions should be interpreted cautiously. Although 742 participants afford adequate power to detect interactive effects, a single study is inadequate to conclude null effects. It is possible that there was inadequate variation in physical activity and cognitive activity to create significant interaction. Future research should utilize a randomized controlled intervention trial design to generate greater changes in physical activity and cognitive activity and explore these interactive effects. Second, although physical activity was measured objectively using an accelerometer, intensity of physical activity (e.g., light, moderate, vigorous) was not captured. Intensity is important to consider to properly inform and implement lifestyle physical activity recommendations (Aadland & Ylvisåker, 2015). We were also unable to determine an accurate measure of accelerometer wear time because the Actical accelerometer lacks a heat sensor component, and so removal of the device cannot always be distinguished from periods of no activity. Furthermore, the measure of cognitive activity was limited to self-report of seven common activities, but did not include other cognitive activities that involve technology that older adults may engage such as computer or tablet use. Innovative methods using mobile technology may be leveraged to capture all cognitive activities. Individuals may be able to report current activities in real time on a smart phone, similar to methods developed for other behaviors (Cain, Depp, & Jeste, 2009). Another limitation was that our sample was disproportionately female and White and had an education level higher than the United States’ average. This limitation, which was also present in earlier longitudinal studies (Bielak et al., 2014; de Frias & Dixon, 2014; Hughes et al., 2015; Rajan et al., 2015; Sturman et al., 2005; Verghese et al., 2003), reduces the ability to generalize the findings to other populations.

Taken together, our findings emphasize the need to further investigate the potentially nuanced effects of the interaction between physical activity and cognitive activity. In addition, the independent contributions of both physical activity and cognitive activity for the prevention of MCI and dementia must be considered. Future research is required to develop and test interventions that enhance both types of activity over time to prevent loss of cognition in older adults. Intervention research should target lifestyle physical activity, which older adults generally prefer over structured exercise (Darden et al., 2013).

Footnotes

Appendix

Late-life cognitive activity is a composite measure of frequency of participation in seven cognitively stimulating activities during the past year. Activities include reading, writing letters, visiting a library, and playing games such as chess or checkers. These items involve information processing or retention and have relatively few barriers to participation.

Participants are asked to rate each item on a 5-point scale. Values for Items 2 to 7 are flipped so that higher values indicate more frequent participation (see table below). The variable ranges from 1 to 5 and is calculated by averaging the individual item scores.

Participants are asked the following questions:

Response choices for each item:

The variable is calculated if at least half of the items are nonmissing.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Rush Alzheimer’s Disease Center (Grant NIH/NIA R01AG017917).

ORCID iDs

Shannon Halloway

Michael E. Schoeny

References

Aadland

Ylvisåker

(2015). Reliability of objectively measured sedentary time and physical activity in adults. PLoS ONE, 10(7), e0133296. doi:10.1371/journal.pone.0133296

Aggarwal

N. T.

Bienias

J. L.

Bennett

D. A.

Wilson

R. S.

Morris

M. C.

Schneider

J. A.

. . . Evans

D. A.

(2006). The relation of cigarette smoking to incident Alzheimer’s disease in a biracial urban community population. Neuroepidemiology, 26, 140-146. doi:10.1159/000091654

Bennett

D. A.

Arnold

S. E.

Valenzuela

M. J.

Brayne

Schneider

J. A.

(2014). Cognitive and social lifestyle: Links with neuropathology and cognition in late life. Acta Neuropathologica, 127, 137-150. doi:10.1007/s00401-013-1226-2

Bennett

D. A.

Buchman

A. S.

Boyle

P. A.

Barnes

L. L.

Wilson

R. S.

Schneider

J. A.

(2018). Religious orders study and rush memory and aging project. Journal of Alzheimer’s Disease, 64(Suppl 1), S161-S189. doi:10.3233/JAD-179939

Bielak

A. A. M.

Gerstorf

Anstey

K. J.

Luszcz

M. A.

(2014). Longitudinal associations between activity and cognition vary by age, activity type, and cognitive domain. Psychology and Aging, 29, 863-872. doi:10.1037/a0036960

Boyle

P. A.

Buchman

A. S.

Wilson

R. S.

Leurgans

S. E.

Bennett

D. A.

(2009). Association of muscle strength with the risk of Alzheimer disease and the rate of cognitive decline in community-dwelling older persons. Archives of Neurology, 66, 1339-1344. doi:10.1001/archneurol.2009.240

Boyle

P. A.

Wilson

R. S.

Aggarwal

N. T.

Tang

Bennett

D. A.

(2006). Mild cognitive impairment: Risk of Alzheimer disease and rate of cognitive decline. Neurology, 67, 441-445. doi:10.1212/01.wnl.0000228244.10416.20

Brasure

Desai

Davila

Nelson

V. A.

Calvert

Jutkowitz

. . . Kane

R. L.

(2018). Physical activity interventions in preventing cognitive decline and Alzheimer-type dementia: A systematic review. Annals of Internal Medicine, 168, 30-38. doi:10.7326/M17-1528

Buchman

A. S.

Boyle

P. A.

Shah

R. C.

Wilson

R. S.

Bennett

D. A.

(2012). Total daily physical activity and the risk of AD and cognitive decline in older adults. Neurology, 78, 1323-1329. doi:10.1212/WNL.0b013e3182535d35

10.

Cain

A. E.

Depp

C. A.

Jeste

D. V.

(2009). Ecological momentary assessment in aging research: A critical review. Journal of Psychiatric Research, 43, 987-996. doi:10.1016/j.jpsychires.2009.01.014

11.

Centers for Disease Control and Prevention. (2015). Glossary of terms: Physical activity. Retrieved from https://www.cdc.gov/physicalactivity/basics/glossary/index.htm

12.

Crouter

S. E.

DellaValle

D. M.

Horton

Haas

J. D.

Frongillo

E. A.

Bassett

D. R.

(2011). Validity of the Actical for estimating free-living physical activity. European Journal of Applied Physiology, 111, 1381-1389. doi:10.1007/s00421-010-1758-2

13.

Curran

P. J.

Obeidat

Losardo

(2010). Twelve frequently asked questions about growth curve modeling. Journal of Cognition and Development: Official Journal of the Cognitive Development Society, 11, 121-136. doi:10.1080/15248371003699969

14.

Darden

Richardson

Jackson

E. A.

(2013). Physical activity and exercise for secondary prevention among patients with cardiovascular disease. Current Cardiovascular Risk Reports, 7(6). doi:10.1007/s12170-013-0354-5

15.

de Frias

C. M.

Dixon

R. A

. (2014). Lifestyle engagement affects cognitive status differences and trajectories on executive functions in older adults. Archives of Clinical Neuropsychology, 29, 16-25. doi:10.1093/arclin/act089

16.

Gorelick

P. B.

Furie

K. L.

Iadecola

Smith

E. E.

Waddy

S. P.

Lloyd-Jones

D. M.

. . . Zerna

(2017). Defining optimal brain health in adults: A presidential advisory from the American Heart Association/American Stroke Association. Stroke, 48, e284-e303. doi:10.1161/STR.0000000000000148

17.

Halloway

Arfanakis

Wilbur

Schoeny

M. E.

Pressler

S. J.

(2018). Accelerometer physical activity is associated with greater gray matter volumes in older adults without dementia or mild cognitive impairment. The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences, gby010. doi:10.1093/geronb/gby010

18.

Halloway

Wilbur

Schoeny

M. E.

Arfanakis

(2017). Effects of endurance-focused physical activity interventions on brain health: A systematic review. Biological Research for Nursing, 19, 53-64. doi:10.1177/1099800416660758

19.

Hughes

T. F.

Becker

J. T.

Lee

C.-W.

Chang

C.-C. H.

Ganguli

(2015). Independent and combined effects of cognitive and physical activity on incident MCI. Alzheimer’s & Dementia: The Journal of the Alzheimer’s Association, 11, 1377-1384. doi:10.1016/j.jalz.2014.11.007

20.

IBM Corp. Released 2015. IBM SPSS Statistics for Windows, Version 23.0. Armonk, NY: IBM Corp.

21.

Kohout

F. J.

Berkman

L. F.

Evans

D. A.

Cornoni-Huntley

(1993). Two shorter forms of the CES-D (Center for Epidemiological Studies Depression) depression symptoms index. Journal of Aging and Health, 5, 179-193. doi:10.1177/089826439300500202

22.

Lampit

Hallock

Valenzuela

(2014). Computerized cognitive training in cognitively healthy older adults: A systematic review and meta-analysis of effect modifiers. PLoS Medicine, 11(11), e1001756. doi:10.1371/journal.pmed.1001756

23.

Lauenroth

Ioannidis

A. E.

Teichmann

(2016). Influence of combined physical and cognitive training on cognition: A systematic review. BMC Geriatrics, 16, Article 141. doi:10.1186/s12877-016-0315-1

24.

Lee

P.-L.

(2014). Cognitive function in midlife and beyond: Physical and cognitive activity related to episodic memory and executive functions. The International Journal of Aging and Human Development, 79, 263-278. doi:10.1177/0091415015574190

25.

McKhann

Drachman

Folstein

Katzman

Price

Stadlan

E. M.

(1984). Clinical diagnosis of Alzheimer’s disease: Report of the NINCDS-ADRDA work group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s disease. Neurology, 34, 939-944.

26.

Muthén

L. K.

Muthén

B. O.

(2017). Mplus user’s guide (8th ed.). Los Angeles, CA: Author.

27.

Rajan

K. B.

Barnes

L. L.

Skarupski

K. A.

Mendes

Leon

C. F.

Wilson

R. S.

Evans

D. A.

(2015). Physical and cognitive activities as deterrents of cognitive decline in a biracial population sample. The American Journal of Geriatric Psychiatry: Official Journal of the American Association for Geriatric Psychiatry, 23, 1225-1233. doi:10.1016/j.jagp.2015.07.008

28.

Stern

(2012). Cognitive reserve in ageing and Alzheimer’s disease. Lancet Neurology, 11, 1006-1012. doi:10.1016/S1474-4422(12)70191-6

29.

Sturman

M. T.

Morris

M. C.

Mendes

Leon

C. F.

Bienias

J. L.

Wilson

R. S.

Evans

D. A.

(2005). Physical activity, cognitive activity, and cognitive decline in a biracial community population. Archives of Neurology, 62, 1750-1754. doi:10.1001/archneur.62.11.1750

30.

Troiano

R. P.

Berrigan

Dodd

K. W.

Mâsse

L. C.

Tilert

McDowell

(2008). Physical activity in the United States measured by accelerometer. Medicine and Science in Sports and Exercise, 40, 181-188. doi:10.1249/mss.0b013e31815a51b3

31.

Verghese

Lipton

R. B.

Katz

M. J.

Hall

C. B.

Derby

C. A.

Kuslansky

. . . Buschke

(2003). Leisure activities and the risk of dementia in the elderly. The New England Journal of Medicine, 348, 2508-2516. doi:10.1056/NEJMoa022252

32.

Wang

H.-X.

Jin

Hendrie

H. C.

Liang

Yang

Cheng

. . . Gao

(2013). Late life leisure activities and risk of cognitive decline. The Journals of Gerontology, Series A: Biological Sciences & Medical Sciences, 68, 205-213. doi:10.1093/gerona/gls153

33.

Wilson

Barnes

Bennett

(2003). Assessment of lifetime participation in cognitively stimulating activities. Journal of Clinical and Experimental Neuropsychology, 25(5), 634-642. doi:10.1076/jcen.25.5.634.14572

34.

Wilson

R. S.

Segawa

Boyle

P. A.

Bennett

D. A.

(2012). Influence of late-life cognitive activity on cognitive health. Neurology, 78, 1123-1129. doi:10.1212/WNL.0b013e31824f8c03

35.

Ryan

L. H.

Schaie

K. W.

Willis

S. L.

Kolanowski

(2009). Factors associated with cognition in adults: The Seattle longitudinal study. Research in Nursing & Health, 32, 540-550. doi:10.1002/nur.20340

36.

Zhu

Qiu

Zeng

(2017). Leisure activities, education, and cognitive impairment in Chinese older adults: A population-based longitudinal study. International Psychogeriatrics, 29, 727-739. doi:10.1017/S1041610216001769