As the population ages, cognitive impairment, including Alzheimer’s disease and other forms of dementia as well as lesser declines associated with aging, has become one of the greatest health threats. The total number of new cases of dementia each year worldwide is nearly 7.7 million, and the total number worldwide is projected to nearly double every 20 yr, to 65.7 million in 2030 and 115.4 million in 2050 (36).
Engagement in physical activity (PA) can positively influence the health of older adults and reduce the risk of many chronic diseases (28). Previous studies support the assertion that PA delays the onset of cognitive decline and the incidence of dementia or Alzheimer’s disease associated with aging (1,2,31). Most existing studies, however, have relied on self-reported PA assessments, which may be influenced by recall biases, health status, depression, cognitive ability, and other factors, especially for older adults. Objective PA measures have been increasingly used to overcome limitations of self-report measures, but the sample sizes of those present studies were relatively small (10,32). More objective assessments of PA are needed to better define the dose–response association of various PA outcomes with cognitive function.
In addition, important differences exist in PA levels across racial/ethnic groups (29). Racial/ethnic minorities are more likely to be misclassified by many self-report questions (4). To date, few investigations are based on national studies, including a large proportion of Black participants. A better understanding of the relationship between objectively measured PA and cognition in different racial/ethnic groups would significantly add to our knowledge of how to best promote healthy cognitive aging. Thus, the purpose of this study was to investigate the association of objectively measured PA with longitudinal cognitive function (including global cognitive function, memory and executive function) in White and Black older adults.
METHODS
Study cohort
The Reasons for Geographic and Racial Differences in Stroke (REGARDS) study is a national longitudinal study of Black and White older adults, enrolled January 2003–October 2007 in United States (19). REGARDS is designed to investigate causes of regional and racial disparities in stroke mortality. Using a computer-assisted telephone interview, trained interviewers obtained demographic information and medical history. In-home brief physical examination was conducted 3–4 wk after the telephone interview. Participants are followed every 6 months for possible stroke outcomes, annually for global cognitive status, and every 2 yr for assessment of memory and executive function. An ancillary accelerometer study was approved in October 2008 for implementation into the ongoing study (20). A question was added during telephone follow-up, asking participants if they were willing to wear an accelerometer and to complete a daily activity log for 1 wk. Upon acceptance, an accelerometer and a log sheet were mailed to participants, with instructions. Detailed design and methods have been described elsewhere (20). All participants gave informed consent, and the study was approved by the institutional review boards of Arizona State University, University of South Carolina, and University of Alabama at Birmingham.
Participant characteristics were collected at time of enrollment into the parent REGARDS study. Age, race, sex, and region of current residence (southeastern U.S. stroke belt vs the rest of the United States), highest education level, smoking (current, never, and past), physician diagnosis of stroke, history of hypertension, and health status were defined by self-report. Body mass index (BMI) was determined from measured height and weight. Blood pressure was defined as the average of two measurements taken using a standard protocol. Diabetes was defined as fasting glucose level >126 mL·dL−1 (or >200 mL·dL−1 if participant was nonfasting) or self-reported medication use for glucose control. Of the 12,146 participants who consented to wear an accelerometer, 8096 provided data after exclusions for failure to wear or return the device, device error, missing log sheet, and noncompliance with the criterion of >4 d with >10 h·d−1 of wear time. Details of inclusion and exclusion are described elsewhere (18,20).
Cognitive impairment was defined by each participant’s score on the Six-Item Screener (SIS) (9). SIS scores ranged from 0 to 6, with a score of <4 correct responses indicating cognitive impairment (9). No participant in this cohort developed incident cognitive impairment before PA variables were measured. After sensitivity analysis, cognitive tests conducted ±12 months of PA measurement were selected to achieve the largest sample size without affecting the results significantly. Participants were excluded if they 1) had no record of SIS within ±12 months of PA assessment (n = 696), 2) were identified as cognitively impaired within ±12 months of PA assessment (n = 323), 3) had no record of one or more items of participant characteristics (age, sex, race, etc.) mentioned (n = 85); 4) had no follow-up SIS assessment after the date of wearing the accelerometer (n = 338), and 5) had a stroke before the most recent follow-up SIS assessment (n = 202). A final sample of 6452 participants with usable accelerometer data and at least one complete follow-up SIS measure were included in the analyses.
Measures
An ActicalTM (Mini Mitter Respironics, Inc., Bend, OR) worn over the right hip while attached to a neoprene waistband provided estimates of the frequency, intensity, and duration of PA. Participants were instructed to put on the device after waking up each morning and take it off before going to bed each evening. Participants were instructed to wear the device for seven consecutive days and return the device immediately after. For all participants with usable data, absolute time spent in sedentary behavior (SED), light-intensity PA (LPA), and moderate-to-vigorous-intensity PA (MVPA) and proportions of total accelerometer wear time spent in SED (SED%), LPA (LPA%), and MVPA (MVPA%) were determined. Nonwear periods were defined as a string of 0 counts per minute for >120 consecutive minutes (21). Activity count cut points were applied to differentiate SED (0–49 counts per minute), LPA (50–1064 counts per minute), and MVPA (>1065 counts per minute), respectively (17,21,37). A laboratory-based validation study for ActicalTM was completed among older adults with similar demographic characteristics to our study (17).
Cognitive function was assessed during REGARDS follow-up telephone interviews using standardized scripts and scoring methods described previously (33). Cognitive impairment was defined by each participant’s score on the SIS. Four expanded cognitive battery tests with demonstrated usefulness in early identification of cognitive decline, Alzheimer’s disease, or dementia were also used (26). These tests were Word List Learning (WLL) and semantic fluency (animal fluency [AF]) from the Consortium to Establish a Registry for Alzheimer’s Disease battery (26), and letter fluency (LF), recall, and orientation items from the Montreal Cognitive Assessment (MoCA recall and orientation) (27). These tools assess domains of memory (WLL and MoCA recall and orientation) and executive function (AF and LF) (12,26). The z-scores based on the mean and SD for each specific test in domains of memory and executive function were created, respectively. The average z-scores of both tests under each domain were used as the z-score of memory or executive function. Participants were included if they had at least one baseline measure in memory or executive function domain within ±12 months of PA assessment and at least one follow-up measure of the same test after the PA assessment (37). All cognitive assessments have been proven to be valid and reliable for both White and Black participants (12,26,27,33), with no significant racial/ethnic differences having been reported.
Statistical analysis
Continuous data expressed as MVPA (min·wk−1) and MVPA% were skewed, thereby introducing concerns regarding the highly leveraged points. Thus, the primary independent variable of MVPA% was divided into quartiles, allowing for group comparisons and investigation of the dose–response relationship. Because of the inherent variability in daily accelerometer wear time, MVPA% rather than the absolute time spent in MVPA was used as the standard for quartiles to render results comparable among individuals. The main dependent variable was incident cognitive impairment defined as a shift from intact cognitive screening status (SIS score of 5 or 6) at the closest assessment to the baseline PA measurement (within ±12 months) to impaired cognitive screening status (SIS score of ≤4) at the latest follow-up assessment. Secondary dependent variables were the changes in z-scores of memory and executive function. Analyses for this report were based on data acquired through April 2015.
Differences in demographic variables, accelerometer variables, and cognitive tests across quartiles of MVPA% were tested by ANOVA or chi-square tests. Logistic regression analysis was used to estimate the odds ratio (OR) of cognitive impairment and the associated 95% confidence interval (CI), with adjustments made for age, sex, race, region of residence, education, BMI, hypertension, smoking, and diabetes. General linear regression models with adjustment for the previously mentioned confounders, baseline scores, and follow-up time intervals were used to assess the association between PA and changes in z-scores generated from WLL, AF, LF, and MoCA recall and orientation. Results were also stratified by race in this analysis. Regression analysis, including the interaction terms (i.e., sex × PA and race × PA), was also conducted, and there were no significant interactions between MVPA% and these confounders, respectively (P > 0.05). Thus, interaction terms were not included in the final models. All probability values were based on two-tailed tests; P < 0.05 indicated statistical significance. Analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC).
RESULTS
Demographics, accelerometer data, and cognitive function status are displayed in Tables 1 and 2 according to quartiles of MVPA%. Among the 6452 participants, 55.3% were women and 30.5% were Black. The mean ± SD age was 69.7 ± 8.5 yr. Participants were compliant wearing the accelerometer for 6.6 ± 0.8 d. Participants spent most of their wear time in SED (77.1% ± 9.3%) and LPA (21.4% ± 8.4%). MVPA% was extremely limited (1.5% ± 1.9%). Participants with higher MVPA% were significantly more likely to be men, White, younger, well educated, nonsmoker, without diabetes or hypertension, and lower in BMI and blood pressure. Participants with higher MVPA% had significantly better raw scores in WLL, AF, LF, and MoCA recall and orientation and higher baseline z-scores in executive function and memory (P < 0.001).
;)
TABLE 1: Characteristics of participants by quartile of MVPA%a (N = 6452).
TABLE 2: Baseline PA and cognitive scores by quartile of MVPA%.a
During a mean ± SD follow-up period of 2.9 ± 1.1 yr, 346 incident cases of cognitive impairment occurred according to SIS. White and Black adults experienced 209 and 137 cases of incident cognitive impairment, respectively. Higher MVPA% was significantly associated with less decreases in the cognitive scores of WLL, AF, LF, MoCA recall and orientation, executive function, and memory z-scores during follow-up (Table 3). The association between MVPA% and risk of cognitive impairment was significant in White and Black adults combined (Table 4). These relationships remained significant after adjustment for age, sex, race, region of residence, and education in model 2 (quartile 2: OR = 0.63, 95% CI = 0.48–0.84; quartile 3: OR = 0.53, 95% CI = 0.38–0.74; quartile 4: OR = 0.57, 95% CI = 0.40–0.81). In model 3, results were consistent with model 2, with participants in the higher MVPA% quartiles less likely to develop cognitive impairment. Stratified analyses indicated higher MVPA% quartiles were independently associated with lower risk of cognitive impairment among White adults in all models. However, there was no significant association between MVPA% and risk of cognitive impairment in Black adults in the adjusted models (i.e., quartile 4: OR = 0.69, 95% CI = 0.36–1.32 in model 3). Sensitivity analyses were conducted with a minimum of 2 yr between baseline assessment of PA and the next available SIS to exclude any underlying diseases at baseline, but most of the results were identical. Similar analyses were conducted using LPA% and SED% quartiles, but no significant results were found for any of the cohort groupings (see Table 1, Supplemental Digital Content 1, association of LPA% and SED% with incident cognitive impairment, https://links.lww.com/MSS/A742).
TABLE 3: Change of cognitive scores by quartiles of MVPA%a in follow-up.
TABLE 4: Incident cognitive impairment by quartiles of MVPA%a (N = 6452) in follow-up.
Adjusted linear regression models revealed some significant associations between MVPA% and changes in z-scores for executive function and memory when controlling for age, sex, race, region of residence, education, BMI, hypertension, smoking, diabetes, baseline scores, and follow-up time intervals (Table 5). Compared with the lowest MVPA% quartiles, all other quartiles maintained better memory performance (P < 0.05), whereas only the highest quartile significantly differed in maintenance of executive function (P = 0.026). In race stratified analyses, higher MVPA% quartiles were independently associated with better maintenance of executive function (quartile 4 of MVPA%) and memory (quartiles 2, 3, and 4 of MVPA%) (P < 0.05) in White adults. In Black adults, higher MVPA% quartiles were independently associated with better maintenance of memory (quartiles 3 and 4 of MVPA%, P < 0.05), but not executive function (P > 0.05). The preservation of memory over time was of similar magnitude among Black and White adults. Similar analyses were conducted using LPA% and SED% quartiles, but no significant results were observed (see Table 2, Supplemental Digital Content 2, association of LPA% with change in cognitive z-scores, https://links.lww.com/MSS/A743, and Table 3, Supplemental Digital Content 3, association of SED% with change in cognitive z-scores, https://links.lww.com/MSS/A744).
TABLE 5: Adjusteda linear regression model indicating the association of MVPA%b with the change in cognitive z-scoresc in follow-up.
DISCUSSION
This study is one of the first to examine the dose–response association between objectively measured PA and cognitive function in a large and racially/ethnically diverse older population. Our study is unique in demonstrating a significant relationship between objectively measured PA and cognitive function among 6452 White and Black older adults during follow-up. Even a minor difference in MVPA% (i.e., only a 3–5 min mean difference between quartiles 1 and 2) was associated with a significantly 36% lower risk of incident cognitive impairment and greater maintenance of executive function and memory over time. The study also indicated that higher MVPA% was independently associated with a 39%–47% less risk of cognitive impairment and maintenance of memory and executive function over time in White adults, as well as maintenance of memory performance in Black adults. These findings confirm our previous cross-sectional analysis in the same cohort (37) and support a dose–response relationship exists between engagement in MVPA% and cognitive performance over time. Interventions should be provided to prevent cognitive decline before symptoms are manifest.
The association between PA and cognitive function has been studied previously (11,34). A meta-analysis (13) of 16 prospective studies using self-reported PA indicated that the relative risks of incident dementia and Alzheimer’s disease in the higher PA categories were significantly reduced by 28%–45%. However, these previous studies were limited by the potential bias of self-report measures of PA and also have not been able to explore the effects of LPA and SED on cognitive function. A few studies with objectively measured data have reported that active older adults may have lower risks of cognitive impairment and dementia. One cross-sectional study (15) showed that the total accelerometer-measured PA (step counts per week) was positively associated with visual episodic memory and face–name memory accuracy in older adults. Another study (8) reported that greater MVPA (h·d−1) was related to lower volume of White matter lesions. In a recent longitudinal study, Barnes et al. (3) demonstrated higher levels of aerobic fitness (V˙O2max) predicted better performance on measures of attention and executive function during 6 yr of follow-up. Middleton et al. (25) reported a significant dose–response relationship between activity energy expenditure measured by doubly labeled water and incidence of cognitive impairment over 2–5 yr of follow-up. In the Rush Memory and Aging Project, accelerometer-measured total daily PA (counts per day) was substantially negatively associated with cognitive decline and incident Alzheimer’s disease in older individuals without dementia during a follow-up period of 4 yr (6). The sample sizes of these longitudinal studies were relatively small (N = 197–716) and primarily focused on White Americans. In addition, the methodological differences in objective measures preclude comparisons of studies of the optimal pattern of PA to prevent cognitive decline. With this current analysis, we can more confidently suggest even a small dose of daily free-living MVPA, rather than self-reported PA or structured exercise in clinical settings, was significantly associated with maintenance of cognitive performance.
In the current study, analyses revealed the association between PA and cognitive function remained robust in older adults after multiple adjustments, which confirms previous findings (24,35). The Chicago Health and Aging Project reported that more self-report PA was modestly (not independently) associated with less cognitive decline in a large community of White and Black older adults for 6 yr (30). In our REGARDS cohort, the stratified analyses indicated a similar association of MVPA% with memory performance over time among White and Black adults, but with executive function among only White adults. To our knowledge, this is one of the first studies to investigate how race/ethnicity may moderate the relationship between objectively measured PA and cognitive function in older adults. Further observation of this cohort over time is needed to better determine whether disparities exist between White and Black adults with respect to the PA and cognitive function dose–response relationship across specific cognitive domains.
Interestingly, the results regarding LPA% and SED% revealed no significant relationships with cognitive function change over time. These findings suggest PA intensity may also be an important independent factor influencing the relationship between PA and cognitive function in older adults. Other studies applying accelerometer measurements have reported intensity of PA is influential in the association between PA and memory and executive function in older adults (5,22). Previous studies have reported a significant association between LPA and various health measures (7), indicating additional follow-up may be necessary to investigate the association between LPA and cognitive function. For SED, computer using time has been positively associated with memory and executive function, whereas television viewing time was negatively associated with cognitive function in older adults (14,16,23). Unfortunately, we were not able to determine the nature of SED in our study. Because of conflicting results in this emerging area, additional investigations are necessary to fully explore the effect of SED on cognitive change.
This study has several strengths. First, objectively measured PA was used to examine the association of PA with cognitive function in a large population of White and Black older adults. This allowed for analyses with proportion of time spent being sedentary and in PA of varying intensity. Second, the sample was recruited from a well-characterized cohort of midlife and older Black and White adults living in the United States, and a substantial percentage of participants were quite compliant. Third, we explored the relationships of accelerometer derived PA with specific domains of memory and executive function, in addition to odds of impairment in global cognition. Fourth, this is one of the first studies to stratify results on the relationship between objectively measured PA and cognitive function by race.
The findings of this study were subject to limitations. Disadvantages of using hip-worn accelerometers include not being able to identify types of PA or capture upper-body movement. The application of standard cut points to differentiate time spent being sedentary and physically active at varying intensities does not account for potential differences in fitness levels, especially among older adults. Thus, the measured time in MVPA in our study may be underestimated. Also, the cohort was quite sedentary (>11 h·d−1, 77% of accelerometer wearing time), and there was a significant skewness of MVPA%. The limited variability of the dependent variable and trend of the association makes it difficult to determine a precise threshold of MVPA for reducing risk of cognitive impairment. There were no substantial differences by race–sex groups in those agreeing to participate in the accelerometer protocol compared with those who did not, with the exception that those who participated had higher education level (20). In addition, there were differences by race in the final proportions of usable data with Black adults providing a smaller yield (20). Thus, the results may not be generalizable to other REGARDS participants.
CONCLUSION
Higher levels of objectively measured MVPA% were independently associated with lower incidence of cognitive impairment in White older adults, with the greatest reduction in risk between the lowest and the next highest MVPA% quartiles (i.e., as little as 4 min·d−1 or 28 min·wk−1 of MVPA may be helpful in maintaining cognitive function). Higher levels of MVPA% were also independently and incrementally associated with better maintenance of memory over time for both Black and White adults, as well as maintenance of executive function for White adults. Race may moderate the relationship between PA and certain aspects of cognitive function in older adults. Neither LPA% nor SED% were independently associated with any measures of cognitive function in this cohort.
This work is supported by a cooperative agreement (no. U01 NS041588) and an investigator-initiated grant (no. R01NS061846) from the Department of Health and Human Service, National Institute of Neurological Disorders and Stroke, National Institutes of Health. Additional funding was supported by the Fundamental Research Funds for the Central Universities (grant no. GK201603128) and an unrestricted research grant from The Coca-Cola Company. The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. The authors thank the other investigators, staff, and participants of the REGARDS study for their valuable contributions. A full list of participating REGARDS investigators and institutions can be found at http://www.regardsstudy.org. We also acknowledge the contributions of Dr. Barbara Ainsworth, Dr. Matthew Buman, and Dr. Cheryl der Ananian who served on the first author’s doctoral dissertation committee at Arizona State University.
The results of the present study do not constitute endorsement by the American College of Sports Medicine. The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.
REFERENCES
1. Angevaren M, Vanhees L, Nooyens AC, Wendel-Vos CG, Verschuren WM. Physical activity and 5-year cognitive decline in the Doetinchem cohort study.
Ann Epidemiol. 2010;20(6):473–9.
2. Baker LD, Frank LL, Foster-Schubert K, et al. Aerobic exercise improves cognition for older adults with glucose intolerance, a risk factor for Alzheimer’s disease.
J Alzheimers Dis. 2010;22(2):569–79.
3. Barnes DE, Yaffe K, Satariano WA, Tager IB. A longitudinal study of cardiorespiratory fitness and cognitive function in healthy older adults.
J Am Geriatr Soc. 2003;51(4):459–65.
4. Bopp M, Wilcox S, Laken M, et al. Factors associated with physical activity among African-American men and women.
Am J Prev Med. 2006;30(4):340–6.
5. Brown BM, Peiffer JJ, Sohrabi HR, et al. Intense physical activity is associated with cognitive performance in the elderly.
Transl Psychiatry. 2012;2:e191.
6. Buchman AS, Boyle PA, Yu L, Shah RC, Wilson RS, Bennett DA. Total daily physical activity and the risk of AD and cognitive decline in older adults.
Neurology. 2012;78(17):1323–9.
7. Buman MP, Hekler EB, Haskell WL, et al. Objective light-intensity physical activity associations with rated health in older adults.
Am J Epidemiol. 2010;172(10):1155–65.
8. Burzynska AZ, Chaddock-Heyman L, Voss MW, et al. Physical activity and cardiorespiratory fitness are beneficial for white matter in low-fit older adults.
PLoS One. 2014;9(9):e107413.
9. Callahan CM, Unverzagt FW, Hui SL, Perkins AJ, Hendrie HC. Six-Item Screener to identify cognitive impairment among potential subjects for clinical research.
Med Care. 2002;40(9):771–81.
10. Davis MG, Fox KR. Physical activity patterns assessed by accelerometry in older people.
Eur J Appl Physiol. 2007;100(5):581–9.
11. Geda YE, Roberts RO, Knopman DS, et al. Physical exercise and mild cognitive impairment: a population-based study.
Arch Neurol. 2010;67(1):80–6.
12. Hachinski V, Iadecola C, Petersen RC, et al. National Institute of Neurological Disorders and Stroke—Canadian Stroke Network vascular cognitive impairment harmonization standards.
Stroke. 2006;37(9):2220–41.
13. Hamer M, Chida Y. Physical activity and risk of neurodegenerative disease: a systematic review of prospective evidence.
Psychol Med. 2009;39(1):3–11.
14. Hamer M, Stamatakis E. Screen-based sedentary behavior, physical activity, and muscle strength in the English longitudinal study of ageing.
PLoS One. 2013;8(6):e66222.
15. Hayes SM, Alosco ML, Hayes JP, et al. Physical activity is positively associated with episodic
memory in aging.
J Int Neuropsychol Soc. 2015;21(10):780–90.
16. Hoang TD, Reis J, Zhu N, et al. Effect of early adult patterns of physical activity and television viewing on midlife cognitive function.
JAMA Psychiatry. 2016;73(1):73–9.
17. Hooker SP, Feeney A, Hutto B, et al. Validation of the Actical activity monitor in middle-aged and older adults.
J Phys Act Health. 2011;8(3):372–81.
18. Hooker SP, Hutto B, Zhu W, et al.
Accelerometer measured sedentary behavior and physical activity in White and Black adults: the REGARDS study.
J Sci Med Sport. 2016;19(4):336–41.
19. Howard VJ, Cushman M, Pulley L, et al. The reasons for geographic and racial differences in stroke study: objectives and design.
Neuroepidemiology. 2005;25(3):135–43.
20. Howard VJ, Rhodes JD, Mosher A, et al. Obtaining
accelerometer data in a national cohort of Black and White adults.
Med Sci Sports Exerc. 2015;47(7):1531–7.
21. Hutto B, Howard VJ, Blair SN, et al. Identifying
accelerometer nonwear and wear time in older adults.
Int J Behav Nutr Phys Act. 2013;10:120.
22. Kerr J, Marshall SJ, Patterson RE, et al. Objectively measured physical activity is related to cognitive function in older adults.
J Am Geriatr Soc. 2013;61(11):1927–31.
23. Kesse-Guyot E, Charreire H, Andreeva VA, et al. Cross-sectional and longitudinal associations of different sedentary behaviors with cognitive performance in older adults.
PLoS One. 2012;7(10):e47831.
24. Landi F, Russo A, Barillaro C, et al. Physical activity and risk of cognitive impairment among older persons living in the community.
Aging Clin Exp Res. 2007;19(5):410–6.
25. Middleton LE, Manini TM, Simonsick EM, et al. Activity energy expenditure and incident cognitive impairment in older adults.
Arch Intern Med. 2011;171(14):1251–7.
26. Morris JC, Heyman A, Mohs RC, et al. The Consortium to Establish a Registry for Alzheimer’s disease (CERAD). Part I. Clinical and neuropsychological assessment of Alzheimer’s disease.
Neurology. 1989;39(9):1159–65.
27. Nasreddine ZS, Phillips NA, Bedirian V, et al. The Montreal cognitive assessment, MoCA: a brief screening tool for mild cognitive impairment.
J Am Geriatr Soc. 2005;53(4):695–9.
28. Nelson ME, Rejeski WJ, Blair SN, et al. Physical activity and public health in older adults: recommendation from the American College of Sports Medicine and the American Heart Association.
Med Sci Sports Exerc. 2007;39(8):1435–45.
29. Sofi F, Valecchi D, Bacci D, et al. Physical activity and risk of cognitive decline: a meta-analysis of prospective studies.
J Intern Med. 2011;269(1):107–17.
30. Sturman MT, Morris MC, Mendes de Leon CF. Physical activity, cognitive activity, and cognitive decline in a biracial community population.
Arch Neurol. 2005;62(11):1750–4.
31. Tierney MC, Moineddin R, Morra A, Manson J, Blake J. Intensity of recreational physical activity throughout life and later life cognitive functioning in women.
J Alzheimers Dis. 2010;22(4):1331–8.
32. Troiano RP. Physical activity in the United States measured by
accelerometer.
Med Sci Sports Exerc. 2008;40(1):181–8.
33. Wadley VG, Unverzagt FW, McGuire LC, et al. Incident cognitive impairment is elevated in the stroke belt: the REGARDS study.
Ann Neurol. 2011;70(2):229–36.
34. Willey JZ, Gardener H, Caunca MR, et al. Leisure-time physical activity associates with cognitive decline: the Northern Manhattan Study.
Neurology. 2016;86(20):1897–903.
35. Winker R, Lukas I, Perkmann T, et al. Cognitive function in elderly marathon runners: cross-sectional data from the marathon trial (APSOEM).
Wien Klin Wochenschr. 2010;122(23–24):704–16.
36. Wortmann M. Dementia: a global health priority—highlights from an ADI and World Health Organization Report.
Alzheimers Res Ther. 2012;4(5):40.
37. Zhu W, Howard VJ, Wadley VG, et al. Association between objectively measured physical activity and cognitive function in older adults—the Reasons for Geographic and Racial Differences in Stroke study.
J Am Geriatr Soc. 2015;63(12):2447–54.
