Journal Logo

EPIDEMIOLOGY

Physical Activity from Childhood to Adulthood and Cognitive Performance in Midlife

HAKALA, JUUSO O.1,2; ROVIO, SUVI P.1; PAHKALA, KATJA1,2; NEVALAINEN, JAAKKO3; JUONALA, MARKUS4; HUTRI-KÄHÖNEN, NINA5; HEINONEN, OLLI J.2; HIRVENSALO, MIRJA6; TELAMA, RISTO6; VIIKARI, JORMA S. A.4; TAMMELIN, TUIJA H.7; RAITAKARI, OLLI T.1,8

Author Information
Medicine & Science in Sports & Exercise: May 2019 - Volume 51 - Issue 5 - p 882-890
doi: 10.1249/MSS.0000000000001862

Abstract

The prevalence of diagnosed dementia and milder cognitive deficits is increasing worldwide (1), making primary prevention a crucial target on the global public health agenda (2). The origin of cognitive deficits is multifactorial; for example, genetic-, cardiovascular-, and lifestyle-related risk factors may exert their influence already years or even decades before any clinical symptoms of cognitive deficits are detectable (2). Simultaneously, increased prevalence of unfavorable lifestyle (e.g., physical inactivity, smoking, and poor diet) (2) results in negative health consequences. Our previous study pinpointed the role of childhood cardiovascular risk factors as determinants for adulthood cognitive function (3). In addition, the role of adulthood physical activity (PA) as a determinant for later life cognitive performance has become clearer (4). However, the role of childhood PA as an independent contributing factor for adulthood cognitive performance has remained obscure.

Results from previous observational studies have mainly focused on revealing positive associations between midlife (5–9) or old age (10–14) PA and cognitive performance at old age. However, somewhat conflicting results have also been reported. One previous study found no association between adulthood PA and cognitive performance in old age (9), whereas another study suggested a negative association between lifelong strenuous PA and cognitive performance in late midlife (15). Evidence on the effects of childhood or adolescence PA on adulthood cognitive function are scarce. There are only two previous studies focusing on PA in childhood (16) or early adulthood (17) and cognitive performance in midlife. Similarly, only few previous studies have focused on the associations between adolescence PA and cognitive function in old age (15,18–20) or increased risk of mild cognitive impairment and early-onset dementia later in life (21). All these previous studies focusing on early life (i.e., childhood, adolescence, early adulthood) PA have suggested positive associations between PA and cognitive function. Furthermore, a previous study in adolescent population has reported a positive association between PA and several cognitive domains during growth and maturation (22), whereas another study suggested that PA may enhance brain development in adolescence which might reflect as better cognitive performance in old age (18). Importantly, it has also been suggested that subclinical cognitive decline may cause decline in PA from midlife which might confound the results on the effects of midlife PA on later cognitive function (23). However, some of previous observational studies have had short follow-up times (10,12) or queried PA retrospectively (5,9,15,18–20), and/or the possible confounders and mediators (e.g., education, systolic blood pressure, serum lipids, body mass index [BMI]) on the studied associations may not have been taken into account extensively. Therefore, there is still paucity of knowledge on the longitudinal and independent associations between childhood, adolescence and young adulthood PA and cognitive performance in adulthood.

Results from previous animal studies point toward similar associations than the observational studies indicating that PA may be associated with better spatial learning in adult (24), middle-age (25) and old rodents (26). Interestingly, one experimental study in rats supported the beneficial role of childhood PA on neurodevelopment as it suggested that early-life exercise may induce development of more complex neural circuitry in adulthood which may also result in a greater tolerance of later brain damage (27). Therefore, it is plausible to hypothesize that early-life PA might be associated with better cognitive function in midlife also in humans. The aim of the present study was to close the remaining gap of knowledge by elucidating the associations between longitudinal PA from childhood through adolescence to adulthood and midlife cognitive performance leveraging the data from the ongoing longitudinal Cardiovascular Risk in Young Finns Study (YFS).

MATERIALS AND METHODS

Participants

The YFS is a national ongoing longitudinal population-based study focusing on cardiovascular risk factors from childhood to adulthood. The first cross-sectional study was conducted in five Finnish university cities and their rural surroundings in 1980, when 3596 randomly selected individuals (boys and girls) age 3, 6, 9, 12, 15, and 18 yr participated in clinical examinations. Follow-up studies were conducted in 1983, 1986, 2001, 2007, and 2011. The study design of the YFS and more details on the YFS population and protocol has been reported elsewhere (28).

Cognitive performance

In 2011, cognitive performance was assessed in 2026 participants 34 to 49 yr old with the Cambridge Neuropsychological Test Automated Battery (CANTAB), including four tests that reflect different cognitive domains and neurodevelopmental entities: 1) the Paired Associates Learning (PAL) test assessing visual and episodic memory as well as visuospatial associative learning, 2) the Spatial Working Memory (SWM) test measuring working memory, executive function, problem solving, and the ability to conduct a self-organized search strategy, 3) the Reaction Time (RTI) test measuring motor and mental response speeds as well as response accuracy and impulsivity, and 4) the Rapid Visual Information Processing (RVP) test was used to assess visual processing, recognition, and sustained attention. Each of the four tests produced several variables. Principal component analyses were conducted to identify components accounting for the majority of the variation within each test. Test specific components were created to indicate performance in all studied cognitive domains. The principal components for cognitive performance were normalized using rank order normalization procedure resulting in four normally distributed components each with mean 0 and standard deviation (SD) 1. After that, the principal components were transformed so that greater value in the principal component indicates better cognitive performance (for example, higher value in the component for reaction time indicates better performance, not a longer reaction time). All available data for each cognitive test was used in the analyses, and therefore, the number of participants varies between the components (n = 177 were excluded due to technical reasons; n = 51 refused to participate in all or some of the tests). More detailed description and the validation of the cognitive data have been reported previously (29). Previous studies on CANTAB tests have shown adequate discriminate abilities for the CANTAB test battery among cognitively healthy adults (30). Furthermore, previous test–retest reliability analyses have shown adequate to high correlations (r = 0.71–0.89) among elderly population (31). Accordingly, the cognitive testing method used in the YFS may be considered adequate in discriminating the study subjects on a population level as done in the present study.

Physical activity

Physical activity was measured with a standardized self-administered questionnaire in all study phases from the age of nine (see Supplemental Tables 1 and 2, Supplemental Digital Content 1, https://links.lww.com/MSS/B474) and with a questionnaire administered by the parents for participants age 3 to 6 yr (see Supplemental Table 3, Supplemental Digital Content 1, https://links.lww.com/MSS/B474). The self-administered questionnaire included questions concerning the frequency and intensity of leisure-time PA, participation in sports club training, participation in competitive sport events, and the habitual way of spending leisure time. The questionnaire for the small children included questions concerning the child’s habitual way of playing indoors/outdoors, PA compared to other children at the same age and interests toward PA and sports. Based on these data, a PA index (PAI) was calculated in all study phases (32). Validation of the YFS PA data has been done in previous studies (32–34). The results from the validation analyses indicate that the YFS PA questionnaire is an acceptably valid subjective measure of PA as there was a significant moderate correlation between PAI index and the average number of daily pedometer steps (correlation coefficients 0.25–0.31) (34) even though the pedometer does not measure all possible aspects of PA (e.g., swimming, cycling). The reliability analyses conducted on the YFS PA questionnaire data showed significant correlations that varied between 0.44 and 0.69 among females, and between 0.49 and 0.76 among males in 1980 (32). Similarly, in 2001 the significant correlations varied between 0.59 and 0.85 among females, and between 0.74 and 0.85 among males (32).

To utilize all available repeatedly measured exposure data, the area under the curve (AUC) for continuous PA indices was evaluated to indicate a long-term exposure of PA (35). Subject-specific curves for PAI were estimated by mixed model regression splines (36). The covariance structure for the longitudinal setting was modeled by allowing for subject specific regression spline coefficients, which were incorporated as random effects to the model. To avoid overfitting, the number of knots was reduced (two knots on the calendar time from 1980 to 2011) for the subject-specific part from that of the fixed effects part (four knots on age from 3 to 34 yr). The mean profile was allowed to vary across birth cohorts and sex in terms of possibly different fixed effects parts. Similar to the approach of Lai et al. (35), the area AUC was evaluated as a measure of a long-term accumulation of the PAIs. For this study, the AUC variable for PAI was defined separately for childhood (age, 6–12 yr), adolescence (age, 12–18 yr), young adulthood (age, 18–24 yr), and from childhood to young adulthood (age, 6–24 yr).

Due to longer intervals between the adulthood follow-up studies, the AUC approach for the adulthood PA exposure would have relied on estimation from sparse data, which could have compromised their reliability. Therefore, we considered the AUC approach not to be applicable for adulthood PA exposure in the present study. To evaluate PA exposure in adulthood (between ages 24 and 37 yr), an average value of the PAI was calculated over the adulthood follow-up period (follow-up years, 2001–2011) during which each subject had one to three PAI assessments. Subjects with one adulthood PAI assessment (n = 695) were not excluded from the analyses as PA has previously been reported to remain stable in adulthood (37). For interpretability, the AUC variables and adulthood PA variable were standardized using rank normalization procedure resulting in normally distributed variables with mean 0 and SD 1.

Covariates

Age was defined in full years at the end of 2011. Socioeconomic status (SES) in childhood was determined as an annual income of the family in 1980 (38). Four annual family income strata at the time of baseline were determined: 1) <17,000 euros; 2) 17,000–27,000 euros; 3) 27,001–34,000 euros; 4) >34,000 euros. Childhood academic performance expressed as grade point average (i.e., mean of grades in all individual school subjects at baseline or either of the two subsequent follow-ups for those participants who were not of school age at baseline) was queried and used as a proxy for childhood cognitive ability. Adulthood education was queried in follow-up studies in 2001, 2007, and 2011. Maximum years of education was determined as a continuous variable from self-reported data concerning total years of education attained until the year 2011. Current smoking was queried throughout the follow-up time among participants 12 yr and older. Subjects who reported current smoking at any of the follow-up phases at the ages between 12 and 24 yr were classified as early-life smokers. Weight (kg) and height (m) were measured, and BMI was calculated as weight (kg) / height (m2). Standard methods were used for measuring systolic blood pressure and serum total cholesterol at baseline and all follow-up studies. Detailed description of the assessment of cognitive performance, PA and the covariates is presented in the Appendix (see document, Supplemental Digital Content 1, https://links.lww.com/MSS/B474).

Statistical analysis

Associations between categorical variables were studied with the χ2 test. Student’s t-test or the Wilcoxon rank sum test was applied for analyses for continuous variables. Linear regression analyses were conducted to investigate the associations for childhood/adolescence/young adulthood/adulthood PA and midlife cognitive performance. All regression analyses were conducted as multivariate models, adjusting first for sex, age, SES and PA exposure in adulthood for time frames between the ages 6 and 24 yr, as well as for PA exposure in childhood for adulthood PA exposure (model 1). After that, all analyses were further adjusted for childhood cognitive performance, adulthood years of education, systolic blood pressure, serum total cholesterol, and BMI at the time of cognitive testing (model 2). Possible effect modification of age and sex for the studied associations were analyzed by adding interaction terms (sex*PA, age*PA) into the fully adjusted models (model 2). All statistical analyses were performed using SAS 9.4, and the level of statistical significance was set at 0.05.

RESULTS

Representativeness of the study population

The representativeness of the study population participating in the cognitive testing was examined by comparing the baseline differences between the participants and nonparticipants (see Supplemental Table 4, Supplemental Digital Content 1, https://links.lww.com/MSS/B474). The participants were more often women (60.26%, P < 0.0001) and older (41.84 yr vs 40.92 yr, P < 0.0001) compared with the nonparticipants. Additionally, they originated from families with higher income (20.71% vs 7.85%, P = 0.003) and had better academic performance in childhood compared to the nonparticipants (7.77 vs 7.65, P < 0.0001). There were no significant differences between the participants and nonparticipants in PA from childhood to young adulthood or any of the covariates.

Characteristics of the study population

To compare participants with high and low PA exposure from childhood to young adulthood, the participants were divided into two groups according to their PA between ages 6 to 24 yr using the median as the cutoff value. The numbers of participants in each separate cognitive test and the differences in the background characteristics between the high and low PA groups are presented in Table 1. The participants in the high PA group were younger (P < 0.0001), more often men (P < 0.0001) and early-life nonsmokers (P < 0.0001) than the participants in the low PA group. The participants in the high PA group originated more often from families with higher income (P < 0.0001), and they also had more years of education in adulthood (P < 0.0001) than those in the low PA group. The participants in the high PA group had significantly better performance in all four cognitive domains compared to the participants in the low PA group (PAL test: −0.07 SD; 95% confidence interval [CI], −0.135 to−0.005 vs 0.07 SD; 95% CI, 0.001–0.133; P = 0.003; SWM test: −0.08 SD; 95% CI, −0.139 to −0.021 vs 0.08 SD; 95% CI, 0.016–0.144; P = 0.0002; RTI test: −0.15 SD; 95% CI, −0.216 to −0.084 vs 0.15 SD; 95% CI, 0.085–0.215; P < 0.0001; RVP test: −0.11 SD; 95% CI, −0.171 to −0.049 vs 0.11 SD; 95% CI, 0.047–0.173; P < 0.0001; Table 1). Additionally, as the cognitive performance may vary between women and men, the participants were divided into sex-specific high and low PA groups according to their PA between ages 6 to 24 yr using the sex-specific median as the cutoff value (see Supplemental Table 4, Supplemental Digital Content 1, https://links.lww.com/MSS/B474). The participants in the high PA group had significantly better performance in both sexes in memory and learning (PAL test: women: −0.02 SD; 95% CI, −0.107 to 0.067 vs 0.12 SD; 95% CI, 0.034–0.206; P = 0.033; men: −0.14 SD; 95% CI, −0.237 to −0.043 vs 0.03 SD; 95% CI, −0.065 to 0.125; P = 0.013) and in reaction time (RTI test: women: −0.31 SD; 95% CI, −0.396 to −0.224 vs −0.06 SD; 95% CI, −0.138 to 0.018; P < 0.0001; men: 0.07 SD; 95% CI, −0.032 to 0.172 vs 0.36 SD; 95% CI, 0.268–0.452; P < 0.0001). Men had significantly better performance in spatial working memory (SWM test: 0.08 SD; 95% CI, −0.011 to 0.171 vs 0.31; 95% CI, 0.216–0.404; P = 0.0006) and in visual processing and sustained attention (RVP test: −0.07 SD; 95% CI, −0.163 to 0.023 vs 0.21 SD; 95% CI, 0.115–0.305; P < 0.0001). No associations were found in women for spatial working memory or visual processing and sustained attention (see Supplemental Table 5, Supplemental Digital Content 1, https://links.lww.com/MSS/B474).

T1
TABLE 1:
Background characteristics and cognitive performance among low and high PA groups formed based on PA from childhood to young adulthood (age, 6–24 yr).

Cumulative PA from childhood to young adulthood and midlife cognitive performance

In the multivariate analyses adjusted for sex, age, family SES at baseline, and adulthood PA (model 1), the cumulative exposures to childhood, adolescence and young adulthood PA were found to be positively associated with reaction time in midlife (RTI test; childhood: β = 0.119 SD; 95% CI, 0.055–0.182; P = 0.0002; adolescence: β = 0.125 SD; 95% CI, 0.063–0.188; P < 0.0001; young adulthood: β = 0.135 SD; 95% CI, 0.063–0.207; P = 0.0002) (Table 2). Similarly, the level of adulthood PA was associated positively with reaction time (model 1: β = 0.045 SD; 95% CI, 0.013–0.077; P = 0.006) and visual processing and sustained attention in midlife (model 1: β = 0.041 SD; 95% CI, 0.010–0.072; P = 0.010) after adjusting for sex, age, family SES at baseline, and PA in childhood (age 6–12 yr). Subsequently, the analyses for all PA age frames were further adjusted for childhood academic performance, adulthood years of education, systolic blood pressure, serum total cholesterol, and BMI (model 2). The results from these further adjusted analyses remained essentially similar for all age frames for reaction time (childhood: β = 0.116 SD; 95% CI, 0.053–0.179; P = 0.0003; adolescence: β = 0.120 SD; 95% CI, 0.057–0.182; P = 0.0002; young adulthood: β = 0.127 SD, 95% CI 0.055–0.199, P = 0.0006; adulthood β = 0.036 SD; 95% CI, 0.004–0.069; P = 0.028). Based on our previous study that showed a −0.02 SD decline per year for reaction time in the YFS population (29), the association between childhood/adolescence/young adulthood PA correspond to ~6 yr effect of aging, whereas the association for adulthood PA corresponds to ~1.5 yr age effect. For RVP, the results attenuated when the analyses were further adjusted according to the model 2. No significant associations were found for other cognitive domains.

T2
TABLE 2:
Associations between cumulative exposure to PA and cognitive performance.

Effect modification of age and sex

The possible effect modification of age and sex for the association between PA from childhood to young adulthood (age 6–24 yr) or in adulthood (age 24–37 yr) and cognitive performance was studied by introducing interaction terms for each possible modifier (i.e., PA*sex, PA*age at the time of cognitive testing) separately into the fully adjusted linear regression models. No significant interactions were found for age in any of the studied cognitive domains.

For spatial working memory, a significant interaction was found for sex and adulthood PA (SWM test; β = 0.069 SD; 95% CI, 0.013–0.125; P = 0.015), whereas there tended to be an interaction between sex and PA from childhood to young adulthood (SWM test; β = 0.086 SD; 95% CI, −0.014 to 0.186; P = 0.091). For visual processing and sustained attention, a significant interaction was found for sex and PA from childhood to young adulthood (RVP test; β = 0.101 SD; 95% CI, 0.003–0.199; P = 0.043), but the interaction for adulthood PA was nonsignificant (RVP test; β = 0.044 SD; 95% CI, −0.011 to 0.099; P = 0.118). Due to the modifying effect of sex on the association between PA and spatial working memory as well as visual processing and sustained attention, analyses for these cognitive domains were conducted separately for men and women in all studied PA age windows.

Among men, the analyses adjusted for age, family SES at baseline, and PA in childhood (model 1), showed a significant association between adulthood PA and spatial working memory (SWM test; β = 0.045 SD; 95% CI, 0.001–0.089; P = 0.045) (Table 3). The association remained essentially similar after further adjustments for childhood academic performance, adulthood years of education, systolic blood pressure, serum total cholesterol, and BMI (model 2), but the statistical significance diluted (β = 0.035 SD; 95% CI, −0.010 to 0.079; P = 0.126). No significant associations for PA from childhood to young adulthood were found among men or for PA in any of the studied age frames among women for spatial working memory.

T3
TABLE 3:
Associations between cumulative exposure to PA, spatial working memory (SWM test), and visual processing and sustained attention (RVP test) separately among women and men.

For visual processing and sustained attention, the sex-stratified analyses adjusted for age, family SES at baseline, and adulthood PA (model 1) showed a significant association among men for young adulthood (RVP test; β = 0.101 SD; 95% CI, 0.001–0.200; P = 0.048) and adulthood PA (β = 0.064 SD; 95% CI, 0.018–0.110; P = 0.006). Additionally, there tended to be an association for adolescence PA (0.077 SD; 95% CI, −0.009 to 0.163; P = 0.081). All associations from the sex-stratified analyses for visual processing and sustained attention attenuated after further adjustments for childhood academic performance, adulthood years of education, systolic blood pressure, serum total cholesterol and BMI (model 2). The covariate that was mainly responsible for the dilution of the effect of PA on visual processing and sustained attention was childhood academic performance. The sex-stratified analyses for visual processing and sustained attention showed no significant associations for women.

DISCUSSION

This study showed that higher exposures to childhood, adolescence, and young adulthood PA were associated with better reaction time in midlife independent of midlife PA. Interestingly, higher exposure to adulthood PA was associated with better reaction time in midlife independent of childhood PA. Additionally, our results indicate that higher levels of PA in adolescence, young adulthood, and adulthood may associate positively with midlife visual processing and sustained attention among men. In addition, among men, higher level of adulthood PA may associate with better midlife spatial working memory.

Findings from the present study

To the best of our knowledge, this is the first longitudinal population-based study examining the association between cumulative exposure to PA from childhood to young adulthood and midlife cognitive performance independent of midlife PA. There are only a few previous studies which have measured the level of PA in early life (e.g., at the age of 11 or 15–25 yr) and examined associations with cognitive performance in midlife (16) or later (9,18–20), or with the risk for diagnosed late-life cognitive disorders (21). All these studies have reported positive associations between PA and cognitive function in some age frames (9,16,18–21). Together with our present study, the findings from the previous studies highlight the plausible significance of early-life PA for brain development during growth and maturation. However, these studies have had shorter follow-up times (e.g., 8 yr), applied different study designs (e.g., retrospective data) or not taken into account the accumulation of PA from childhood to early adulthood or the level of adulthood PA like our study. Additionally, in some of the previous studies, the outlook on cognitive function has been taken at relatively old age when the neuropathological processes causing cognitive deficits are most probably already ongoing (9,18,19). However, our results are supported by a previous cross-sectional study where PA found to associated with faster reaction time in individuals age 15 to 71 yr (39). Furthermore, in another study, PA in childhood and adolescence was observed to be positively associated with cognitive performance at the age of 50 yr (16). Even if that study did not examine the association of youth PA independently from adulthood PA, their results supported our conclusion that the benefits of the PA on cognitive performance have been gained by being physically active throughout the life course.

The novel results from the present study point out a different association for men and women for one of the studied cognitive domains, as higher cumulative exposure to PA from childhood to young adulthood was found to be associated with better visual processing and sustained attention (RVP test) among men but not among women. Our results are supported by a prior study showing similar association between PA and visual information processing among older men (19). In our previous analyses, men and women were found to differ in terms of midlife cognitive performance; men had faster reaction time as well as better performance in spatial working memory and visual processing and sustained attention compared to women, whereas women outperformed men in the test measuring memory and learning (29). Furthermore, during the intrauterine period, the brain develops differently in the male and female fetuses due to direct actions of different testosterone levels on the developing nerve cells and the interactions between the developing neurons and their environment (40). The process of sexual differentiation causes permanent structural and functional changes in the brain (40). These changes are believed to have a lasting effect on the sexual differentiation of the brain, which could also explain not only the sex differences in cognitive performance but also the differences in the determinants of midlife cognitive performance found in our present study. On the other hand, the differences in the associations between PA among men and women might also reflect the different participation activity in PA as well as different quality and preferences within PA among men and women.

Potential mechanisms

Increased neurogenesis among physically active subjects has been suggested to be the plausible biological mechanism explaining the positive association between PA and cognitive performance (41). Furthermore, PA may increase neuronal plasticity and to upregulate secretion of neurotrophins (42). Specifically, higher level of PA has been associated with the secretion of brain derived neurotrophic factor, a possible key mediator in maintaining or improving cognitive performance and a biological link between PA and better cognitive performance (25,43). Additionally, the vascular hypothesis suggests that the positive effects of PA on cognitive decline might be due to positive alterations on cardiovascular risk factors (e.g., high blood pressure and serum cholesterol levels) (43,44). The possible confounding/mediating role of cardiometabolic risk factors on the associations between, PA and cognitive function was taken into account in our analyses by adjusting for cardiometabolic risk factors. Importantly, in our study these adjustments did not alter the results, which suggest that also other possible pathways between PA and cognitive function may exist. Additionally, compromised vascular structure and function, such as endothelial dysfunction, might lead in a reduced capability to maintain the blood flow demands of the brain (45), which could subsequently affect also cognitive performance.

Strengths and limitations

Our study has several strengths. First, it is based on a large, randomly selected, population-based cohort making our study population to be representative of the general Finnish population. Second, our population has been followed-up from childhood to midlife for over 30 yr enabling us to study the lifelong associations between PA and cognitive function. Thirdly, our population is young and cognitively healthy which provides us a novel outlook to the associations between PA and cognitive function and highlights the possibilities for primordial prevention of cognitive deficits. This outlook is important considering the long subclinical phase behind the clinical symptoms of cognitive deficits. Furthermore, PA as well as cardiometabolic risk factors have been the key focuses in the YFS from the baseline. Therefore, the data on PA and cardiometabolic risk factors have been systematically collected since baseline using similar and standard methods in every follow-up study. Moreover, our computerized cognitive test battery may be considered to reflect accurately different cognitive domains and, at the end, different neurodevelopmental entities. Finally, computerized cognitive tests have many advantages including better precision, standardization and reliability compared to traditional noncomputerized tests.

There are some limitations that need to be considered. First, cognitive testing was conducted only in a single time point in midlife. Therefore, we were not able to elucidate the possible effects of childhood/adolescence PA on the changes in cognitive performance from childhood to adulthood in this study. Second, it has been previously reported that childhood intelligence quotient is a strong predictor of cognitive abilities in later life (44) and could therefore bias also our results. Even if we do not have an exact measurement of childhood cognitive ability/intelligence quotient, we have taken this possible bias into consideration by adjusting our analyses for childhood academic performance as a proxy for childhood cognitive performance.

Furthermore, self-reported measures of PA have been criticized for limited reliability and validity, particularly in samples of children and adolescents (46). To avoid this bias, we have validated our PA data in three previous studies (32–34) suggesting a good validity to the PA data in the YFS. We were not able to show associations for other cognitive domains except for reaction time which might be due to the age range of our study population. As our population is young and cognitively healthy, the variation in cognitive function might not be large enough to bring the associations visible. Therefore, the future follow-up studies of our population will complement the findings from the present study. With the respect to the establishment of causality, observational studies are prone to bias caused by reverse causation. A previous study presents that decline in PA may be due to a preclinical phase of dementia and suggests that the association between PA and cognitive function might not indicate a neuroprotective effect of PA (23). This possibility has to be taken into consideration also in relation to our findings. Therefore, we are not able to draw firm conclusions on the causal relations between PA and cognitive performance. Nevertheless, the use of existing population cohorts from childhood to adulthood provides an opportunity to test the hypothesis that early-life PA exposure is causally linked with adult cognitive performance.

CONCLUSIONS

This study showed that the cumulative exposure to PA from childhood to young adulthood is associated with reaction time in midlife, independently of midlife PA, whereas the associations for other cognitive domains were not observed. Hence, our results suggest that a physically active lifestyle should be adopted already during childhood, adolescence, and young adulthood and continued into midlife to ensure the plausible benefits of PA on midlife cognitive performance.

The Young Finns Study has been financially supported by the Academy of Finland: 134309 (EYE), 126925, 121584, 124282, 129378 (SALVE), 117797 (GENDI), 273971 (TULOS) and 41071 (SKIDI), the Social Insurance Institution of Finland, Kuopio, Tampere and Turku University Hospital Medical Funds, Juho Vainio Foundation, Sigrid Juselius Foundation, Yrjö Jahnsson Foundation, Paavo Nurmi Foundation, Finnish Foundation for Cardiovascular Research, Finnish Cultural Foundation, Tampere Tuberculosis Foundation and Emil Aaltonen Foundation.

Expert technical assistance in data management and statistical analyses by Irina Lisinen and Johanna Ikonen is gratefully acknowledged.

The authors report no relationships with industry. The funders of the study had no role in study design, data collection, data analysis, data interpretation, writing of the manuscript, or the decision to submit the manuscript for publication. The results of the present study do not constitute endorsement by ACSM. The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.

REFERENCES

1. Prince M, Bryce R, Albanese E, Wimo A, Ribeiro W, Ferri CP. The global prevalence of dementia: a systematic review and metaanalysis. Alzheimers Dement. 2013;9(1):63–75.
2. Scheltens P, Blennow K, Breteler MM, et al. Alzheimer’s disease. Lancet. 2016;388(10043):505–17.
3. Rovio SP, Pahkala K, Nevalainen J, et al. Cardiovascular risk factors from childhood and midlife cognitive performance: the Young Finns study. J Am Coll Cardiol. 2017;69(18):2279–89.
4. Blondell SJ, Hammersley-Mather R, Veerman JL. Does physical activity prevent cognitive decline and dementia? A systematic review and meta-analysis of longitudinal studies. BMC Public Health. 2014;14(1):510.
5. Chang M, Jonsson PV, Snaedal J, et al. The effect of midlife physical activity on cognitive function among older adults: AGES—Reykjavik study. Journals Gerontol - Ser A Biol Sci Med Sci. 2010;65 A(12):1369–74.
6. Sabia S, Nabi H, Kivimaki M, Shipley MJ, Marmot MG, Singh-Manoux A. Health behaviors from early to late midlife as predictors of cognitive function: the Whitehall II study. Am J Epidemiol. 2009;170(4):428–37.
7. Rovio S, Kåreholt I, Helkala E-L, et al. Leisure-time physical activity at midlife and the risk of dementia and Alzheimer’s disease. Lancet Neurol. 2005;4(11):705–11.
8. Iso-Markku P, Waller K, Kujala UM, Kaprio J. Physical activity and dementia: long-term follow-up study of adult twins. Ann Med. 2015;47(2):81–7.
9. Gow AJ, Pattie A, Deary IJ. Lifecourse activity participation from early, mid, and later adulthood as determinants of cognitive aging: the Lothian birth cohort 1921. J Gerontol B Psychol Sci Soc Sci. 2016;72(1):gbw124.
10. Podewils LJ, Guallar E, Kuller LH, et al. Physical activity, APOE genotype, and dementia risk: findings from the cardiovascular health cognition study. Am J Epidemiol. 2005;161(7):639–51.
11. Beydoun MA, Beydoun HA, Gamaldo AA, Teel A, Zonderman AB, Wang Y. Epidemiologic studies of modifiable factors associated with cognition and dementia: systematic review and meta-analysis. BMC Public Health. 2014;14(1):643.
12. Yaffe K, Barnes D, Nevitt M, Lui LY, Covinsky K. A prospective study of physical activity and cognitive decline in elderly women: women who walk. Arch Intern Med. 2001;161(14):1703–8.
13. Weuve J, Kang JH, Manson JE, Breteler MM, Ware JH, Grodstein F. Physical activity, including walking, and cognitive function in older women. JAMA. 2004;292(12):1454–61.
14. 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.
15. 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.
16. Dregan A, Gulliford MC. Leisure-time physical activity over the life course and cognitive functioning in late mid-adult years: a cohort-based investigation. Psychol Med. 2013;43(11):2447–58.
17. Hoang TD, Reis J, Zhu N, et al. Effect of early adult patterns of physical activity and television viewing on midlife cognitive function. JAMA Psychiat. 2016;73(1):73–9.
18. Middleton LE, Barnes DE, Lui LY, Yaffe K. Physical activity over the life course and its association with cognitive performance and impairment in old age. J Am Geriatr Soc. 2010;58(7):1322–6.
19. Dik M, Deeg DJ, Visser M, Jonker C. Early life physical activity and cognition at old age. J Clin Exp Neuropsychol. 2003;25(5):643–53.
20. Gill SJ, Friedenreich CM, Sajobi TT, et al. Association between lifetime physical activity and cognitive functioning in middle-aged and older community dwelling adults: results from the brain in motion study. J Int Neuropsychol Soc. 2015;21(10):816–30.
21. Nyberg J, Aberg MA, Schiöler L, et al. Cardiovascular and cognitive fitness at age 18 and risk of early-onset dementia. Brain. 2014;137(5):1514–23.
22. Álvarez-Bueno C, Pesce C, Cavero-Redondo I, Sánchez-López M, Martínez-Hortelano JA, Martínez-Vizcaíno V. The effect of physical activity interventions on children’s cognition and metacognition: a systematic review and meta-analysis. J Am Acad Child Adolesc Psychiatry. 2017;56(9):729–38.
23. Sabia S, Dugravot A, Dartigues J-F, et al. Physical activity, cognitive decline, and risk of dementia: 28 year follow-up of Whitehall II cohort study. BMJ. 2017;j2709.
24. van Praag H, Christie BR, Sejnowski TJ, Gage FH. Running enhances neurogenesis, learning, and long-term potentiation in mice. Proc Natl Acad Sci U S A. 1999;96(23):13427–31.
25. Marlatt MW. Running throughout middle-age improves memory function, hippocampal neurogenesis, and BDNF levels in female C57BL/6J mice. Dev Neurobiol. 2012;72(6):943–52.
26. van Praag H, Shubert T, Zhao C, Gage FH. Exercise enhances learning and hippocampal neurogenesis in aged mice. J Neurosci. 2005;25(38):8680–5.
27. Gomes da Silva S, De Almeida AA, Silva Araújo BH, Scorza FA, Cavalheiro EA, Arida RM. Early physical exercise and seizure susceptibility later in life. Int J Dev Neurosci. 2011;29(8):861–5.
28. Raitakari OT, Juonala M, Rönnemaa T, et al. Cohort profile: the cardiovascular risk in young Finns study. Int J Epidemiol. 2008;37(6):1220–6.
29. Rovio SP, Pahkala K, Nevalainen J, et al. Cognitive performance in young adulthood and midlife: relations with age, sex, and education—the cardiovascular risk in Young Finns study. Neuropsychology. 2016;30(5):532–42.
30. Robbins TW, James M, Owen AM, Sahakian BJ, McInnes L, Rabbitt P. Cambridge neuropsychological test automated battery (CANTAB): a factor analytic study of a large sample of normal elderly volunteers. Dementia. 1994;5(5):266–81.
31. Gonçalves MM, Pinho MS, Simões MR. Test–retest reliability analysis of the Cambridge neuropsychological automated tests for the assessment of dementia in older people living in retirement homes. Appl Neuropsychol Adult. 2016;23(4):251–63.
32. Telama R, Yang X, Viikari J, Välimäki I, Wanne O, Raitakari O. Physical activity from childhood to adulthood: a 21-year tracking study. Am J Prev Med. 2005;28(3):267–73.
33. Mansikkaniemi K, Juonala M, Taimela S, et al. Cross-sectional associations between physical activity and selected coronary heart disease risk factors in young adults. The cardiovascular risk in Young Finns study. Ann Med. 2012;44(7):733–44.
34. Hirvensalo M, Magnussen CG, Yang X, et al. Convergent validity of a physical activity questionnaire against objectively measured physical activity in adults: the cardiovascular risk in Young Finns study. Adv Physiol Educ. 2017;7(4):457–72.
35. Lai CC, Sun D, Cen R, et al. Impact of long-term burden of excessive adiposity and elevated blood pressure from childhood on adulthood left ventricular remodeling patterns: the Bogalusa heart study. J Am Coll Cardiol. 2014;64(15):1580–7.
36. Welham SJ. Smoothing spline models for longitudinal data. Longitudinal Data Analysis. Chapman and Hall/CRC; 2008. pp. 253–89.
37. Telama R. Tracking of physical activity from childhood to adulthood: a review. Obes Facts. 2009;2(3):187–95.
38. Liu RS, Burgner DP, Sabin MA, et al. Childhood infections, socioeconomic status, and adult cardiometabolic risk. Pediatrics. 2016;137(6):934–42.
39. Hillman CH, Motl RW, Pontifex MB, et al. Physical activity and cognitive function in a cross-section of younger and older community-dwelling individuals. Health Psychol. 2006;25(6):678–87.
40. Bao A-M, Swaab DF. Sexual differentiation of the human brain: relation to gender identity, sexual orientation and neuropsychiatric disorders. Front Neuroendocrinol. 2011;32(2):214–26.
41. Nokia MS, Lensu S, Ahtiainen JP, et al. Physical exercise increases adult hippocampal neurogenesis in male rats provided it is aerobic and sustained. J Physiol. 2016;594(7):1855–73.
42. Farmer J, Zhao X, Van Praag H, Wodtke K, Gage FH, Christie BR. Effects of voluntary exercise on synaptic plasticity and gene expression in the dentate gyrus of adult male Sprague-Dawley rats in vivo. Neuroscience. 2004;124(1):71–9.
43. Kennedy G, Hardman RJ, MacPherson H, Scholey AB, Pipingas A. How does exercise reduce the rate of age-associated cognitive decline? A review of potential mechanisms. J Alzheimers Dis. 2016;55(1):1–18.
44. Deary IJ, Corley J, Gow AJ, et al. Age-associated cognitive decline. Br Med Bull. 2009;92(1):135–52.
45. Toda N. Age-related changes in endothelial function and blood flow regulation. Pharmacol Ther. 2012;133(2):159–76.
46. Shephard RJ. Limits to the measurement of habitual physical activity by questionnaires. Br J Sports Med. 2003;37:197–206.
Keywords:

COGNITIVE PERFORMANCE; PHYSICAL ACTIVITY; CHILDHOOD; ADOLESCENCE; MIDLIFE; LONGITUDINAL; POPULATION-BASED

Supplemental Digital Content

Copyright © 2018 by the American College of Sports Medicine