Abstract
Background
While much is known about the effects of physical exercise in adult humans, literature on the oldest-old (≥ 85 years old) is sparse. The present study explored the relationship between self-reported engagement in physical exercise and cognition in the oldest-old.
Methods
The sample included 184 cognitively healthy participants (98 females, MoCA mean score = 24.81) aged 85 to 99 years old (mean = 88.49 years). Participants completed the Community Healthy Activities Model Program for Seniors (CHAMPS) questionnaire and a cognitive battery including NIH-TB, Coding, Symbol Search, Letter Fluency, and Stroop task. Three groups of participants – sedentary (n = 58; MoCA mean score = 24; 36 females; mean age = 89.03), cardio (n = 60; MoCA mean score = 25.08; 29 females; mean age = 88.62), and cardio + strength training (n = 66; MoCA mean score = 25.28; 33 females; mean age = 87.91) – were derived from responses on CHAMPS.
Results
Analyses controlled for years of education, NIH-TB Crystallized Composite, and metabolic equivalent of tasks. The cardio + strength training group had the highest cognitive performances overall and scored significantly better on Coding (p < 0.001) and Symbol Search (p < 0.05) compared to the sedentary group. The cardio + strength training group scored significantly better on Symbol Search, Letter Fluency, and Stroop Color-Word compared to the cardio group (p < 0.05).
Conclusions
Our findings suggest self-reported exercise in the oldest-old is linked to better performance on cognitive measures of processing speed and executive functioning, and that there may be a synergistic effect of combining aerobic and resistance training on cognition.
Supplementary Information
The online version contains supplementary material available at 10.1007/s11357-023-00885-4.
Keywords: Cognition, Cross-sectional study, Healthy aging, Physical exercise
Introduction
Cognitive benefits are conferred to those who engage in physical exercise as they age; however, a fundamental question in geriatrics is whether exercise continues to provide these cognitive benefits to the oldest-old: individuals aged 85 or older. In this study, we sought to examine the relationship between exercise and cognition in the oldest-old.
A growing body of evidence suggests exercise induces structural and chemical changes in the brain. In a previous review, Basso and Suzuki [1] detailed how exercise can work to change the human brain. On electroencephalography (EEG), aerobic exercise has been found to increase P300 amplitude and decrease P300 latency [2], and to increase hippocampal activity measured by functional magnetic resonance imaging (fMRI) [3]. Older adults (65 to 79 years old) who underwent an aerobic exercise intervention showed greater volumetric increases in anterior cingulate cortex, middle frontal gyrus, and surrounding white matter tracts compared to controls on structural MRI [4]. In adults from the UK Biobank, better cardiorespiratory fitness predicted larger gray matter volumes while greater amounts of moderate-to-vigorous physical activity predicted larger bilateral hippocampal volumes [5]. Furthermore, brain-derived neurotrophic factor (BDNF) is upregulated after either aerobic or resistance exercise [6, 7]. In rodent models, increases in BDNF following physical exercise are found in the hippocampus [8], prefrontal cortex [9], and amygdala [10]. Similarly, better cardiovascular health in humans is associated with BDNF related neuroplastic potential [11]. BDNF is an essential precursor of neuroplasticity and long-term potentiation. Structural changes in the brain due to exercise may therefore be partially modulated via changes in BDNF levels. Other neurochemical changes due to exercise include changes in dopaminergic signaling in the striatum [12], changes in frontal cortex norepinephrine levels [13], changes in serotonin levels [14], and changes in glutamate and GABA levels in the anterior cingulate and visual cortex [15]. Furthermore, the changes brought on by physical exercise appear to extend to psychological functions.
Chan and colleagues [16] reviewed the therapeutic benefits of physical exercise for mood. In general, they found ample evidence to indicate physical exercise promotes positive affect and diminishes negative moods. Resistance training more consistently improves mood and decreases anxiety compared to aerobic exercise [17]. Long-term, habitual exercisers more frequently experience positive affect than non-exercisers [18]. Finally, only 15 to 30 min of exercise of moderate intensity is needed to improve stress resilience and mood [16].
Numerous studies have demonstrated the impact of physical exercise on cognition. Sanchez & McGough [19] reviewed observational studies that evaluated the relationship between self-reported physical exercise and cognitive aging. Overall, they found that individuals engaged in physical exercise better maintained their cognitive ability into old age and that rates of dementia were lower for exercisers. In particular, reported intensity of physical activity was positively correlated with improved processing speed and memory [20] and executive functioning [21]. Finally, randomized clinical trials of resistance training generally find improvements in measures of executive functioning and processing speed [22–25]. Collectively, these studies suggest that physical exercise may improve cognition, particularly within the executive domain and fluid cognitive skills. However, many of the studies to date have focused on a large range of ages or have specified narrow ranges (i.e., 20 to 59 years old, 65 years and older). Additionally, observational studies in the elderly have tended to conflate physical activity with physical exercise (i.e., backyard gardening is operationalized the same as bicycling 25 miles). What is not clear is whether these cognitive improvements are also found in the oldest-old and whether the form of physical exercise (i.e., aerobic and/or resistance) can modify the relationship.
Studying the oldest-old is important because of the insights that can be gleaned from this unique demographic. They have survived far past the average life expectancy and may provide information on longevity in aging. Additionally, they are at high risk for cognitive decline and factors limiting cognitive impairment in this age group are of relevance. The intention of the present study then was to explore the relationship between self-reported physical exercise and cognition in the oldest-old, individuals aged 85 years and older. We hypothesized that performance on cognitive tests of fluid ability, executive functioning, and processing speed would be better in oldest-old individuals that reported engagement with both aerobic and resistance exercise over those that reported only aerobic exercise. Further, both groups would perform better than sedentary individuals on the same cognitive tasks.
Methods
Study Design
The current study utilized a cross-sectional design in partnership between four sites: University of Florida, University of Miami, University of Arizona, and University of Alabama-Birmingham. Cognitively intact, physically able participants were recruited from the metro areas of the respective sites. The study was approved by the institutional review board of each study site, and every participant was provided informed consent and agreed to participate. Figure 1 delineates the intake process for potential recruits and Table 1 details inclusion and exclusion criteria. Study procedures consisted of three study visits: initial phone interview, an in-person screen, and baseline cognitive testing.
Fig. 1.
Intake process for potential recruits. Abbreviations: TICS-M Telephone Interview for Cognitive Status – Modified, MoCA Montreal Cognitive Assessment
Table 1.
Inclusion and exclusion criteria
| Inclusion Criteria | 1. Age 85 and older |
| 2. No major physical disability | |
| 3. Independent in basic activities of daily living | |
| 4. Normal cognition defined as not below 1.0 standard deviations for the age and education expected normative values for this group. Participants with and without subjective cognitive complaints will be included | |
| 5. Independent in instrumental activities of daily living | |
| Exclusion Criteria | 1. Any uncontrolled medical condition expected to limit life expectancy or interfere with participation in the study (i.e., unstable cancer, severe depression, or anxiety by DSM-5 criteria) |
| 2. Unable to follow study protocol and instructions due to existing cognitive deficits, dysphasia, or hemineglect | |
| 3. Active substance abuse or alcohol dependence by DSM-5 criteria | |
| 4. Less than 6th grade reading level | |
| 5. Vision or hearing deficits that would preclude administration of the cognitive measures | |
| 6. MRI contraindication(s) | |
| 7. Unwilling or unable to provide written informed consent |
DSM-5 Diagnostic and Statistical Manual of Mental Disorders – Fifth Edition, MRI magnetic resonance imaging
At phone-screen, participants were administered the Telephone Interview for Cognitive Status – Modified (TICS-M) [26]. Participants at or above the cut-off score of 29 [27] on the TICS-M were invited to an in-person screening visit for a thorough evaluation of cognitive status and physical health.
After successful phone-screen, participants were mailed the Community Healthy Activities Model Program for Seniors (CHAMPS) [28] questionnaire to be completed at-home and brought to the in-person screening visit for review by study staff. CHAMPS is a self-report questionnaire that assesses weekly frequency and duration of a variety of lifestyle physical activities pertinent for older adults. It includes activities of various intensities (from light to vigorous) such as walking, running, hiking, swimming, bicycling, dancing, tennis, aerobics, yoga/tai chi, gardening, and housework.
The in-person screening visit occurred approximately two weeks after the phone-screen. During this visit, participants underwent a comprehensive evaluation of sensory, motor, and cognitive abilities. Further information regarding health history, medical history, and activities of daily living were also collected. Participants who scored 22 or higher on the Montreal Cognitive Assessment (MoCA) [29] proceeded to the next part of the study; participants scoring below 22 on the MoCA were subjected to a consensus process between the investigators of each study site to determine cognitive status as informed by performance on other cognitive tasks. During consensus, if the pattern of performance on individual cognitive tests indicated normal cognition (not less than 1 standard deviation below group means) then the participant proceeded to the next part of the study; in the full sample, 14 participants underwent consensus.
The baseline visit occurred approximately 3 to 7 days after the in-person screening visit. This visit consisted of a comprehensive neuropsychological evaluation, including the National Institutes of Health Cognitive Toolbox (NIH-TB) and several standard neuropsychological tests. Because the literature primarily indicates benefits in fluid cognition, executive function, and processing speed as a result of physical exercise, the current study evaluated the NIH-TB Fluid Composite, letter/semantic fluency, Stroop Color-Word inhibition test, and WAIS-IV Coding and Symbol Search tests.
Participants
A total of 206 participants completed the study. Of these, 194 completed the CHAMPS. From participants’ yes/no responses on CHAMPS’ prompts, we derived three physical exercise groups totaling 184 participants: 1) sedentary (n = 58) 2) cardio (n = 60) and 3) cardio + strength training (n = 66). A fourth group of 10 participants (strength training) was excluded from analyses due to small sample size. Supplementary Table 1 details how the groups were formed. The cardio group was created from participants who endorsed either jogging, biking, swimming, aerobics, walking fast, or walking uphill in the past month. The cardio + strength training group was created from participants who endorsed the cardio group activities AND further endorsed strength training OR light strength training in the past month; strength training refers to resistance-based exercises which are intended to specifically improve muscular fitness (such as hand-held weights, elastic bands, weight machines, or calisthenics). The sedentary group consisted of participants who reported no engagement in the aforementioned activities in the past month. Although screened participants did not exhibit or report physical disabilities, arthritic status for each group was included to illustrate general musculoskeletal dysfunction. 11 participants in the final sample underwent the consensus process due to MoCA scores below 22. Finally, metabolic equivalent of tasks (METs) [28] were used to estimate caloric expenditure/week and provide data on duration and intensity of exercise. Weekly caloric expenditure was calculated following Stewart et al. [28]: for each exercise, weekly duration * METs * 210 * (body weight in kg/200), then sum the products to acquire caloric expenditure/week from all exercise activities. Table 2 contains sample characteristics.
Table 2.
Sample characteristics by physical exercise group
| Characteristics | Sedentary (n = 58) | Cardio (n = 60) | Cardio + Strength Training (n = 66) | Total (n = 184) |
|---|---|---|---|---|
| Age (mean, SD) | 89.03 (3.37) | 88.62 (3.03) | 87.91 (2.99) | 88.49 (3.14) |
| White (%) | 96.6 | 95 | 95.5 | 95.7 |
| Female (%) | 62.1 | 48.3 | 50 | 53.2 |
| Years of Education (mean, SD) | 15.6 (2.91) | 16.02 (2.76) | 16.91 (3.31) | 16.21 (3.05) |
| MoCA (mean, SD) | 24 (2.29) | 25.08 (2.22) | 25.28 (2.64) | 24.81 (2.45) |
| NIH-TB Crystallized Composite (mean, SD) | 109.54 (7.61) | 114.19 (9.59) | 114.95 (8.66) | 112.98 (8.94) |
| NIH-TB Fluid Composite (mean, SD) | 77.02 (9.36) | 78.81 (8.19) | 82.65 (9.15) | 79.61 (9.18) |
| WAIS-IV Coding (mean, SD) | 38.8 (9.56) | 45.19 (11.27) | 49.38 (12.26) | 44.72 (11.92) |
| WAIS-IV Symbol Search (mean, SD) | 18.16 (5.16) | 19.02 (4.86) | 21.65 (5.74) | 19.7 (5.47) |
| Letter Fluency (mean, SD) | 23.43 (7.76) | 23.23 (7.68) | 27.44 (8.32) | 24.8 (8.14) |
| Stroop Color-Word (mean, SD) | 24.55 (9.05) | 24.55 (9.05) | 29.48 (8.37) | 26.64 (9.13) |
| Arthritis (%) | 61 | 58 | 59 | 60 |
| METs Calories/Week (mean, SD) | 1358.24 (1131.07) | 3069.2 (2565.85) | 3403.58 (1662.69) | 2649.82 (2072.27) |
METs metabolic equivalent of tasks, MoCA Montreal Cognitive Assessment, NIH-TB National Institutes of Health Cognitive Toolbox, WAIS-IV Wechsler Adult Intelligence Scale – Fourth Edition
Cognitive Measures
The NIH-TB Fluid Composite [30] is a measure composed of working memory, executive functioning, and memory tasks consisting of five measures of fluid abilities: Dimensional Change Card Sort Test [31], the Flanker Inhibitory Control and Attention Test [31], the Picture Sequence Memory Test [32], the List Sorting Working Memory Test [33], and the Pattern Comparison Processing Speed Test [34]. Nolin et al. [35] validated the use of the NIH-TB in this cohort.
The NIH-TB Crystallized Composite [30] is a measure of premorbid intellectual function which consists of two measures of crystallized abilities: Picture Vocabulary Test [36] and the Oral Reading Recognition Test [36]. The NIH-TB is an iPad administered battery of tests whereas all other cognitive tasks were traditional paper and pencil with a stopwatch for time-tracking.
WAIS-IV Coding [37] is a measure of processing speed which requires participants to match symbols to numbers as fast as they can within 120 s.
WAIS-IV Symbol Search [37] is a measure of processing speed which requires participants to identify target symbols within a group of non-target symbols as fast as they can within 120 s.
Letter Fluency [29] is a measure of verbal ability and executive control which requires participants to name as many words as they can within 60 s beginning with a specified target letter.
Stroop Color-Word [38] is a measure of inhibitory ability and executive control which requires participants to name the ink color that a set of words were written in as fast as they can within 45 s while inhibiting the automatic reading response.
Data Management
Data were entered and managed using REDCap electronic data capture tools hosted at each study site [39, 40]. REDCap (Research Electronic Data Capture) is a secure, web-based software platform designed to support data capture for research studies, providing 1) an intuitive interface for validated data capture; 2) audit trails for tracking data manipulation and export procedures; 3) automated export procedures for seamless data downloads to common statistical packages; and 4) procedures for data integration and interoperability with external sources.
Statistical Analysis
ANCOVA with Tukey-adjusted post hoc tests were used to evaluate for differences on each cognitive measure between the three physical exercise groups. Physical exercise group was entered as the independent variable, and except for the NIH-TB Crystallized Composite, each cognitive measure was entered as a dependent variable. NIH-TB Crystallized Composite scores and participants’ years of education were entered as covariates in each model to adjust for baseline intelligence and educational attainment given that NIH-TB normative data do not yet exist in this age group. METs were included as a covariate to adjust for caloric expenditure from all physical activities reported on CHAMPS. Alpha set to 0.05. Finally, Mahalanobis D [41], a measure of multivariate effect size that controls for correlation between variables, was calculated for each group comparison to illustrate group differences accounting for unique contributions from the cognitive measures.
An a priori power analysis was conducted using G*Power version 3.1.9.6 [42] to determine the minimum sample size required to test the study hypothesis. Results indicated the required sample size to achieve 80% power for detecting a medium effect, at a significance criterion of α = 0.05, was n = 158 for ANCOVA. Thus, the final sample size of 184 is adequate to test the study hypothesis.
Results
NIH-TB Fluid Composite
Levene’s test and normality checks were carried out and all assumptions were met. The overall model was significant [F(4, 168) = 5.606, p < 0.001] and there was a significant difference in NIH-TB Fluid Composite [30] scores [F(2, 168) = 3.273, p = 0.040] between the physical exercise groups. After controlling for multiple comparisons, post hoc tests did not show a significant difference between the three physical exercise groups on the NIH-TB Fluid Composite score.
WAIS-IV Coding
Levene’s test and normality checks were carried out and the assumptions met. The overall model was significant [F(4, 171) = 9.984, p < 0.001] and there was a significant difference in Wechsler Adult Intelligence Scale (WAIS-IV) Coding [37] scores [F(2, 171) = 7.325, p < 0.001] between the physical exercise groups. Post hoc tests showed there was a significant difference between the sedentary group and the cardio + strength training group (p < 0.001) of large effect (d = 0.787). Comparing the estimated marginal means (Fig. 2) showed that the highest score was from the cardio + strength training group (mean = 48.682, se = 1.395) compared to the cardio group (mean = 44.164, se = 1.417) and sedentary group (mean = 40.351, se = 1.552).
Fig. 2.
A Group scores on NIH-TB Fluid Composite. B) Group scores on WAIS-IV Coding. C) Group scores on WAIS-IV Symbol Search. D) Group scores on Letter Fluency. E) Group scores on Stroop Color-Word. Bars indicate 95% CI of estimated marginal means. *, **, *** indicate Tukey’s HSD p < 0.05, 0.01, 0.001, respectively. Abbreviations: METs metabolic equivalent of tasks, NIH-TB National Institutes of Health Cognitive Toolbox, WAIS-IV Wechsler Adult Intelligence Scale – Fourth Edition
WAIS-IV Symbol Search
Levene’s test and normality checks were carried out and all assumptions were met. The overall model was significant [F(4, 171) = 5.789, p < 0.001] and there was a significant difference in WAIS-IV Symbol Search [37] scores [F(2, 171) = 4.496, p = 0.013] between the physical exercise groups. Post hoc tests showed there was a significant difference between the sedentary group and the cardio + strength training group (p = 0.030) of medium effect (d = 0.533) and a significant difference between the cardio group and the cardio + strength training group (p = 0.029) of medium effect (d = 0.476). Comparing the estimated marginal means (Fig. 2) showed that the highest score was from the cardio + strength training group (mean = 21.411, se = 0.671) compared to the cardio group (mean = 18.985, se = 0.682) and sedentary group (mean = 18.696, se = 0.747).
Letter Fluency
Levene’s test and normality checks were carried out and all assumptions were met. The overall model was significant [F(4, 171) = 8.546, p < 0.001] and there was a significant difference in Letter Fluency [29] scores [F(2, 171) = 5.098, p = 0.007] between the physical exercise groups. Post hoc tests showed there was a significant difference between the cardio group and the cardio + strength training group (p = 0.007) of medium effect (d = 0.568). Comparing the estimated marginal means (Fig. 2) showed that the highest score was from the cardio + strength training group (mean = 27.327, se = 0.985) compared to the cardio group (mean = 23.080, se = 1.001) and sedentary group (mean = 23.908, se = 1.096).
Stroop Color-Word
Levene’s test and normality checks were carried out and the assumptions met. The overall model was significant [F(4, 165) = 6.949, p < 0.001] and there was a significant difference in Stroop Color-Word [38] scores [F(2, 165) = 3.546, p = 0.031] between the physical exercise groups. Post hoc tests showed there was a significant difference between the cardio group and the cardio + strength training group (p = 0.034) of medium effect (d = 0.473). Comparing the estimated marginal means (Fig. 2) showed that the highest score was from the cardio + strength training group (mean = 29.000, se = 1.123) compared to the cardio group (mean = 24.993, se = 1.156) and sedentary group (mean = 25.573, se = 1.272).
Table 3 summarizes ANCOVA results. Supplementary Table 2 and Supplementary Table 3 summarizes ANCOVA results excluding METs and including sex, respectively.
Table 3.
ANCOVA results
| Variables | p | η2 | d |
|---|---|---|---|
| Fluid Reasoning (n = 172) | |||
| Overall model*** | < 0.001 | ||
| Physical Exercise Group* | 0.040 | 0.035 | |
| Crystallized Composite*** | < 0.001 | 0.079 | |
| Years of education | 0.522 | 0.002 | |
| METs | 0.860 | 0.000 | |
| Sed | Cardio | 0.972 | 0.046 | |
| Sed | Cardio + Strength | 0.078 | 0.458 | |
| Cardio | Cardio + Strength | 0.072 | 0.411 | |
| WAIS-IV Coding (n = 175) | |||
| Overall model*** | < 0.001 | ||
| Physical Exercise Group*** | < 0.001 | 0.072 | |
| Crystallized Composite*** | < 0.001 | 0.092 | |
| Years of education | 0.416 | 0.003 | |
| METs | 0.689 | 0.001 | |
| Sed | Cardio | 0.184 | 0.360 | |
| Sed | Cardio + Strength*** | < 0.001 | 0.787 | |
| Cardio | Cardio + Strength | 0.057 | 0.427 | |
| WAIS-IV Symbol Search (n = 175) | |||
| Overall model*** | < 0.001 | ||
| Physical Exercise Group* | 0.013 | 0.046 | |
| Crystallized Composite*** | < 0.001 | 0.072 | |
| Years of education | 0.219 | 0.008 | |
| METs | 0.896 | 0.000 | |
| Sed | Cardio | 0.958 | 0.057 | |
| Sed | Cardio + Strength* | 0.030 | 0.533 | |
| Cardio | Cardio + Strength* | 0.029 | 0.476 | |
| Letter Fluency (n = 175) | |||
| Overall model*** | < 0.001 | ||
| Physical Exercise Group** | 0.007 | 0.050 | |
| Crystallized Composite*** | < 0.001 | 0.105 | |
| Years of education | 0.586 | 0.001 | |
| METs | 0.146 | 0.010 | |
| Sed | Cardio | 0.850 | 0.111 | |
| Sed | Cardio + Strength | 0.075 | 0.457 | |
| Cardio | Cardio + Strength** | 0.007 | 0.568 | |
| Stroop Color-Word (n = 169) | |||
| Overall model*** | < 0.001 | ||
| Physical Exercise Group* | 0.031 | 0.037 | |
| Crystallized Composite*** | < 0.001 | 0.108 | |
| Years of education | 0.703 | 0.001 | |
| METs | 0.630 | 0.001 | |
| Sed | Cardio | 0.942 | 0.068 | |
| Sed | Cardio + Strength | 0.137 | 0.405 | |
| Cardio | Cardio + Strength* | 0.034 | 0.473 | |
*, **, *** indicate Tukey-adjusted p < 0.05, p < 0.01, p < 0.001, respectively
WAIS-IV Wechsler Adult Intelligence Scale – Fourth Edition, METs metabolic equivalent of tasks
Mahalanobis D
After aggregating the cognitive measures in this study, the magnitude of D between the sedentary group and the cardio group was 0.66. The magnitude of D between the sedentary group and the cardio + strength training group was 0.99. The magnitude of D between the cardio group and the cardio + strength training group was 0.69. Figure 3 visualizes the aggregate effect sizes.
Fig. 3.
Visualization of aggregate effect size comparisons between each physical exercise group. A) 74.5% of the Cardio group is above the mean of the Sedentary group. B) 83.9% of the Cardio and Strength Training group is above the mean of the Sedentary Group. C) 75.5% of the Cardio and Strength Training group is above the mean of the Cardio group
Discussion
In this sample of cognitively intact oldest-old adults, we found that 68.5% reported participating in some form of physical exercise. When categorized into sedentary, cardio, or cardio + strength training groups, significant group differences in performance on tests related to executive functioning (i.e., Stroop Color-Word and Letter Fluency) and processing speed (i.e., WAIS-IV Coding and WAIS-IV Symbol Search) were observed. After post hoc corrections, those who performed cardiovascular exercise in addition to strength-based exercise (the cardio + strength training group) tended to have the best performances and were significantly better on WAIS-IV Coding and WAIS-IV Symbol Search as compared to the sedentary group. The cardio + strength training group performed significantly better on WAIS-IV Symbol Search, Letter Fluency, and Stroop Color-Word compared to the cardio group. The cardio group did not significantly differ from the sedentary group in any of the measures.
Our results align with past findings: physical exercise, particularly resistance training, predicts stronger executive functioning. In a double-blind randomized controlled trial, Eckardt, Braun, and Kibele [22] evaluated the effects of resistance training on executive functioning in 68 healthy older adults aged 65 to 79 years old. Two groups – free-weights or machine-weights – completed two workouts a week for 10 weeks. The researchers found that the free-weight group had small to medium effect improvements on tasks of working memory, processing speed, and response inhibition compared to the machine-weight group. Although our study did not discriminate between free-weights and machine-weights, engagement in some form of resistance training was related to better cognitive performances on similar tasks.
In another study [23], 20 participants between 60 to 80 years old underwent a 12-week resistance training program. Cognitive abilities were assessed using the NIH-TB [43]. They found an improvement of large effect (d = 1.27) on the NIH-TB Fluid Composite, a measure composed of working memory, executive functioning, and memory tasks. There was no improvement on the NIH-TB Crystallized Composite, a measure composed of reading and vocabulary tasks. Of note, the current finding on the NIH-TB Fluid Composite was not significant after correction for multiple comparisons and likely reflects what was found by Nolin [35]: the executive functioning domain lacked construct validity in this cohort, which was driven by unexpected uncorrelation with the Dimensional Change Card Sort test and the Flanker Inhibitory Control and Attention Test. Since the composite depends on these two tests, it may be that fluid cognitive ability in the oldest-old is diminished in normal aging and that there is a reliance on other cognitive functions to complete the tasks composing the NIH-TB Fluid Composite.
In a third study [25], the researchers conducted a randomized controlled trial of 43 participants 65 years and older who had a Clinical Dementia Rating [44] of 0.5 (i.e., mild cognitive impairment with spared functional independence). The intervention group received 16 weeks of resistance training. Both groups had their cognitive abilities assessed via the Frontal Assessment Battery (FAB) [45], which measures several executive functioning abilities such as abstract reasoning, verbal fluency, planning and organization, interference control, and inhibitory control. In the resistance training group, they found a small effect improvement on a task of processing speed and a moderate effect improvement on the FAB.
We found that participants who reported engaging in both aerobic and anaerobic exercise had an advantage on executive functioning pointing to potentially superior gains from resistance training or suggesting a potential synergistic effect. However, it is possible better executive function leads to 1) a higher likelihood of following through with multiple types of exercise and 2) being healthy enough to engage in multiple types of exercise. The effect size differences on executive functioning from randomized controlled trials using resistance training interventions tend to be larger (approximately 0.3 to 1.3) than effect sizes (approximately 0.1 to 0.3) from randomized controlled trials using aerobic training [46]. However, our results are generally smaller in effect size compared to previous studies and this may be due to many factors. First, our study is cross-sectional and retrospective compared to previous studies which were randomized controlled trials that inherently have higher power. Second, previous findings were interventional in nature and researchers could verify participants’ actual involvement with the physical exercise. While CHAMPS can provide useful information about physical exercise, self-report physical exercise measures are highly variable by nature and lack precision; for example, self-report measures of physical activity only correlate between 0.1 to 0.4 with gold-standard actigraphy [47–50]. Additionally, participants tend to under-report participation in light to moderate activities while over-reporting participation in vigorous activities. Assuming our sample follows similar patterns, weekly caloric expenditure values may be understated in the sedentary group and overstated in the other two groups.
However, our study extends previous findings by focusing on an understudied but growing population of the oldest-old, and by including different forms of physical exercise in the analysis. We were interested in cognitive differences between oldest-old participants that primarily engaged in aerobic training or primarily engaged in strength or resistance training. Because of sample size issues with the strength training only group (n = 10) we included a combined cardio + strength training group. It was this combined group which showed findings consistent with previous studies. Considering the previous randomized controlled trials which used resistance training interventions, the fact that the combined group generally performed better than the other physical exercise groups fits with the literature. Surprisingly, the cardio group did not show better cognitive performances compared to the sedentary group on any of the individual cognitive tasks. However, the aggregate effect of the cognitive tasks demonstrated improvement of the cardio group over the sedentary group. Some randomized controlled trials have found aerobic training to improve cognitive abilities [51, 52].
Our analysis also includes a novel approach by calculating Mahalanobis D to evaluate the aggregate effect of the physical exercise group on tests of fluid ability, executive functioning, and processing speed. Although the effect sizes for each cognitive test were relatively moderate, when effect sizes were combined whilst removing the shared variance, we found a large difference between the sedentary and cardio + strength training group and a moderate difference between the cardio and cardio + strength training group. This aligns well with Yoon and colleagues’ trial finding an effect size of 0.79 on the FAB, which conceptually resembles the cognitive tests evaluated in our study.
Strengths of the current study include a novel sample group. Physical exercise, aerobic or anaerobic, is not well characterized in the oldest-old and this study helps to fill that gap in the literature. Another important strength of this study is the emphasis on normal aging rather than a pathological process (e.g., neurodegenerative disease) or a preternatural process (e.g., Super-agers). Other strengths include the aggregate effect size approach and controlling for physical exercise duration and intensity.
Limitations of the current study include lack of generalizability. By design, we were interested in cognitively intact and physically healthy oldest-old adults. Identifying the characteristics of this group is important to understand what variables may be related to successful aging. However, this group then is biased towards healthy adults 85 years and older and results of our study may not be applicable for other age groups or oldest-old adults who may have cognitive and/or physical impairments. Another limitation is the use of CHAMPS to categorize physical exercise groups. As mentioned before, self-reported physical exercise tends to be imprecise when compared against actigraphy data. Similarly, the current study did not include reported lifetime involvement in physical exercise. Participation in physical exercise earlier in life is associated with more positive health outcomes, may increase levels of cognitive reserve, and how much physical exercise each participant engaged in throughout life could have varied considerably as they aged [53]. The current study only captures physical exercise involvement in the past month and cannot clarify the impact of consistent, long-term physical exercise on cognition. Finally, our findings are associative and we cannot make claims of causality.
Future studies should include randomized controlled trials of oldest-old individuals with physical exercise as an intervention. Our findings of high involvement in physical exercise in this age group suggests it is feasible to find participants who could complete such a study. Future studies should also include actigraphy data to obtain more objective measures of physical activity. Additionally, structural and chemical differences in the brains of this age group should also be evaluated in the context of exercise and cognitive improvement because these differences may partially explain the relationships between cognitive ability and physical exercise. Furthermore, future studies could compare these results to a normative sample of people 85 years and older to show the positive benefit of intact cognition or how physical exercise may be an important factor in the preservation of cognition in old age. Finally, an important future study may use physical exercise as an intervention for improving executive subtype mild cognitive impairment.
Conclusion
The cardio + strength training group performed significantly better on the cognitive measures compared to the sedentary group, as well as the cardio group on some measures. For two measures, Letter Fluency and Stroop Color-Word, the cardio + strength training group also performed significantly better than the cardio group. The cardio group did not significantly differ from the sedentary group on any of the cognitive measures. However, when effect sizes of cognitive tasks were aggregated we found larger differences with increasing modalities of physical exercise over the sedentary group. Our findings suggest exercise in the oldest-old is linked with better performance on cognitive measures of processing speed and executive functioning and that there may be an additive or synergistic effect of combining aerobic exercise with resistance training.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
The bulk of the work was supported by the McKnight Brain Research Foundation. The authors would like to acknowledge support from the National Institutes of Health (grant numbers P30AG019610, P30AG072980, R56AG067200, R01AG0720445, R01AG064587, and T32AG061892), the State of Arizona and Arizona Department of Health Services, and the Arizona Research Institute for Biomedical Imaging.
Authors’ contributions
BDH participated in the design of the study, wrote the manuscript, and conducted data analysis; RC, GEA, KV, TR, BL conceived of and designed the study. All authors provided manuscript edits and suggestions. All authors have read and approved the final version of the manuscript and agree with the order of presentation of the authors.
Data Availability
Data is available upon request.
Declarations
Disclosures
The authors declare no competing interests.
Footnotes
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Basso JC, Suzuki WA. The effects of acute exercise on mood, cognition, neurophysiology, and neurochemical pathways: a review. Brain Plasticity. 2017;2(2):127–152. doi: 10.3233/BPL-160040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hillman CH, Pontifex MB, Raine LB, Castelli DM, Hall EE, Kramer AF. THE Effect of acute treadmill walking on cognitive control and academic achievement in preadolescent children. Neuroscience. 2009;159(3):1044. doi: 10.1016/J.NEUROSCIENCE.2009.01.057. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.van Dongen EV, Kersten IHP, Wagner IC, Morris RGM, Fernández G. Physical exercise performed four hours after learning improves memory retention and increases hippocampal pattern similarity during retrieval. Curr Biol. 2016;26(13):1722–1727. doi: 10.1016/J.CUB.2016.04.071. [DOI] [PubMed] [Google Scholar]
- 4.Colcombe SJ, Erickson KI, Scalf PE, et al. Aerobic exercise training increases brain volume in aging humans. J Gerontol-Series A Biol Sci Med Sci. 2006;61(11):1166–1170. doi: 10.1093/GERONA/61.11.1166. [DOI] [PubMed] [Google Scholar]
- 5.Raichlen DA, Klimentidis YC, Bharadwaj PK, Alexander GE. Differential associations of engagement in physical activity and estimated cardiorespiratory fitness with brain volume in middle-aged to older adults. Brain Imaging Behav. 2020;14(5):1994–2003. doi: 10.1007/S11682-019-00148-X/TABLES/2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Dinoff A, Herrmann N, Swardfager W, et al. The Effect of exercise training on resting concentrations of peripheral brain-derived neurotrophic factor (BDNF): A meta-analysis. PLoS One. 2016;11(9). 10.1371/JOURNAL.PONE.0163037 [DOI] [PMC free article] [PubMed]
- 7.Dinoff A, Herrmann N, Swardfager W, Lanctôt KL. The effect of acute exercise on blood concentrations of brain-derived neurotrophic factor in healthy adults: a meta-analysis. Eur J Neurosci. 2017;46(1):1635–1646. doi: 10.1111/EJN.13603. [DOI] [PubMed] [Google Scholar]
- 8.Lu J, Xu Y, Hu W, et al. Exercise ameliorates depression-like behavior and increases hippocampal BDNF level in ovariectomized rats. Neurosci Lett. 2014;573:13–18. doi: 10.1016/J.NEULET.2014.04.053. [DOI] [PubMed] [Google Scholar]
- 9.Cefis M, Prigent-Tessier A, Quirié A, Pernet N, Marie C, Garnier P. The effect of exercise on memory and BDNF signaling is dependent on intensity. Brain Struct Funct. 2019;224(6):1975–1985. doi: 10.1007/S00429-019-01889-7/FIGURES/6. [DOI] [PubMed] [Google Scholar]
- 10.Liu YF, Chen HI, Wu CL, et al. Differential effects of treadmill running and wheel running on spatial or aversive learning and memory: roles of amygdalar brain-derived neurotrophic factor and synaptotagmin I. J Physiol. 2009;587(Pt 13):3221. doi: 10.1113/JPHYSIOL.2009.173088. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Cabral DF, Bigliassi M, Cattaneo G, et al. Exploring the interplay between mechanisms of neuroplasticity and cardiovascular health in aging adults: A multiple linear regression analysis study. Auton Neurosci. 2022;242. 10.1016/J.AUTNEU.2022.103023 [DOI] [PMC free article] [PubMed]
- 12.Fisher BE, Li Q, Nacca A, et al. Treadmill exercise elevates striatal dopamine D2 receptor binding potential in patients with early Parkinson’s disease. NeuroReport. 2013;24(10):509–514. doi: 10.1097/WNR.0B013E328361DC13. [DOI] [PubMed] [Google Scholar]
- 13.Pagliari R, Peyrin L. Norepinephrine release in the rat frontal cortex under treadmill exercise: a study with microdialysis. 1995;78(6):2121-2130.10.1152/JAPPL.1995.78.6.2121 [DOI] [PubMed]
- 14.Young SN. How to increase serotonin in the human brain without drugs. J Psychiatry Neurosci. 2007;32(6):394. Accessed February 9, 2022. /pmc/articles/PMC2077351/ [PMC free article] [PubMed]
- 15.Maddock RJ, Casazza GA, Fernandez DH, Maddock MI. Acute Modulation of Cortical Glutamate and GABA Content by Physical Activity. J Neurosci. 2016;36(8):2449. doi: 10.1523/JNEUROSCI.3455-15.2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Chan JSY, Liu G, Liang D, Deng K, Wu J, Yan JH. Special issue – therapeutic benefits of physical activity for mood: a systematic review on the effects of exercise intensity, duration, and modality. 2018;153(1):102–125. 10.1080/00223980.2018.1470487 [DOI] [PubMed]
- 17.Battaglia C, Di Cagno A, Fiorilli G, et al. Participation in a 9-month selected physical exercise programme enhances psychological well-being in a prison population. Crim Behav Ment Health. 2015;25(5):343–354. doi: 10.1002/CBM.1922. [DOI] [PubMed] [Google Scholar]
- 18.Reed J, Buck S. The effect of regular aerobic exercise on positive-activated affect: A meta-analysis. Psychol Sport Exerc. 2009;10(6):581–594. doi: 10.1016/J.PSYCHSPORT.2009.05.009. [DOI] [Google Scholar]
- 19.Kirk-Sanchez NJ, McGough EL. Physical exercise and cognitive performance in the elderly: current perspectives. Clin Interv Aging. 2014;9:51–62. doi: 10.2147/CIA.S39506. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Angevaren M, Vanhees L, Wendel-Vos W, et al. Intensity, but not duration, of physical activities is related to cognitive function. Eur J Cardiovasc Prev Rehabil. 2007;14(6):825–830. doi: 10.1097/HJR.0B013E3282EF995B. [DOI] [PubMed] [Google Scholar]
- 21.Bixby WR, Spalding TW, Haufler AJ, et al. The unique relation of physical activity to executive function in older men and women. Med Sci Sports Exerc. 2007;39(8):1408–1416. doi: 10.1249/MSS.0B013E31806AD708. [DOI] [PubMed] [Google Scholar]
- 22.Eckardt N, Braun C, Kibele A. Instability Resistance Training improves Working Memory, Processing Speed and Response Inhibition in Healthy Older Adults: A Double-Blinded Randomised Controlled Trial. Sci Rep. 2020;10(1):1–12. doi: 10.1038/s41598-020-59105-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Macaulay TR, Pa J, Kutch JJ, et al. 12 weeks of strength training improves fluid cognition in older adults: A nonrandomized pilot trial. PLoS One. 2021;16(7):e0255018. doi: 10.1371/JOURNAL.PONE.0255018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Yoon DH, Kang D, Kim HJ, Kim JS, Song HS, Song W. Effect of elastic band-based high-speed power training on cognitive function, physical performance and muscle strength in older women with mild cognitive impairment. Geriatr Gerontol Int. 2017;17(5):765–772. doi: 10.1111/ggi.12784. [DOI] [PubMed] [Google Scholar]
- 25.Yoon DH, Lee JY, Song W. Effects of Resistance Exercise Training on Cognitive Function and Physical Performance in Cognitive Frailty: A Randomized Controlled Trial. J Nutr Health Aging. 2018;22(8):944–951. doi: 10.1007/S12603-018-1090-9. [DOI] [PubMed] [Google Scholar]
- 26.Detection of Dementia in the Elderly Using Telephone Screening of Cognitive Status. Accessed February 12, 2022. https://oce.ovid.com/article/00010291-199304000-00006/HTML
- 27.Graff-Radford NR, Ferman TJ, Lucas JA, et al. A cost effective method of identifying and recruiting persons over 80 free of dementia or mild cognitive impairment. Alzheimer Dis Assoc Disord. 2006;20(2):101–104. doi: 10.1097/01.WAD.0000213813.35424.D2. [DOI] [PubMed] [Google Scholar]
- 28.Stewart AL, Mills KM, King AC, Haskell WL, Gillis D, Ritter PL. CHAMPS physical activity questionnaire for older adults: outcomes for interventions. Med Sci Sports Exerc. 2001;33(7):1126–1141. doi: 10.1097/00005768-200107000-00010. [DOI] [PubMed] [Google Scholar]
- 29.Nasreddine ZS, Phillips NA, Bédirian V, et al. The montreal cognitive assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc. 2005;53(4):695–699. doi: 10.1111/J.1532-5415.2005.53221.X. [DOI] [PubMed] [Google Scholar]
- 30.Heaton RK, Akshoomoff N, Tulsky D, et al. Reliability and validity of composite scores from the nih toolbox cognition battery in adults. J Int Neuropsychol Soc. 2014;20(6):588. doi: 10.1017/S1355617714000241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Zelazo PD, Anderson JE, Richler J, et al. NIH toolbox cognition battery (CB): validation of executive function measures in adults. J Int Neuropsychol Soc. 2014;20(6):620. doi: 10.1017/S1355617714000472. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Dikmen SS, Bauer PJ, Weintraub S, et al. Measuring episodic memory across the lifespan: nih toolbox picture sequence memory test. J Int Neuropsychol Soc. 2014;20(6):611. doi: 10.1017/S1355617714000460. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Tulsky DS, Carlozzi N, Chiaravalloti ND, et al. NIH toolbox cognition battery (NIHTB-CB): the list sorting test to measure working memory. J Int Neuropsychol Soc. 2014;20(6):599. doi: 10.1017/S135561771400040X. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Carlozzi NE, Tulsky DS, Chiaravalloti ND, et al. NIH toolbox cognitive battery (NIHTB-CB): the nihtb pattern comparison processing speed test. J Int Neuropsychol Soc. 2014;20(6):630. doi: 10.1017/S1355617714000319. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Nolin SA, Cowart H, Merritt S, et al. Validity of the NIH toolbox cognitive battery in a healthy oldest-old 85+ sample. J Int Neuropsychol Soc. Published online October 14, 2022:1–10. 10.1017/S1355617722000443 [DOI] [PMC free article] [PubMed]
- 36.Gershon RC, Cook KF, Mungas D, et al. Language Measures of the NIH Toolbox Cognition Battery. J Int Neuropsychol Soc. 2014;20(6):642. doi: 10.1017/S1355617714000411. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Wechsler D. Wechsler Adult Intelligence Scale-Fourth Edition (WAIS-IV) San Antonio, TX: Pearson. Published online; 2008. [Google Scholar]
- 38.Golden C. Stroop Color and Word Test: A Manual for Clinical and Experimental Uses. Wood Dale, Illinois: Stoelting Company. Published online; 1978. [Google Scholar]
- 39.Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)—A metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42(2):377–381. doi: 10.1016/J.JBI.2008.08.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Harris PA, Taylor R, Minor BL, et al. The REDCap consortium: Building an international community of software platform partners. J Biomed Inform. 2019;95:103208. doi: 10.1016/J.JBI.2019.103208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Del Giudice M. Multivariate misgivings: is D a valid measure of group and sex differences? Evol Psychol. 2013;11(5):1067–1076. doi: 10.1177/147470491301100511. [DOI] [PubMed] [Google Scholar]
- 42.Faul F, Erdfelder E, Lang AG, Buchner A. G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods. 2007;39(2):175–191. doi: 10.3758/BF03193146. [DOI] [PubMed] [Google Scholar]
- 43.Weintraub S, Dikmen SS, Heaton RK, et al. The cognition battery of the nih toolbox for assessment of neurological and behavioral function: validation in an adult sample. J Int Neuropsychol Soc. 2014;20(6):567. doi: 10.1017/S1355617714000320. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Morris JC. Clinical dementia rating: a reliable and valid diagnostic and staging measure for dementia of the alzheimer type. Int Psychogeriatr. 1997;9(S1):173–176. doi: 10.1017/S1041610297004870. [DOI] [PubMed] [Google Scholar]
- 45.Dubois B, Slachevsky A, Litvan I, Pillon B. The FAB. Neurology. 2000;55(11):1621–1626. doi: 10.1212/WNL.55.11.1621. [DOI] [PubMed] [Google Scholar]
- 46.Kelly ME, Loughrey D, Lawlor BA, Robertson IH, Walsh C, Brennan S. The impact of exercise on the cognitive functioning of healthy older adults: a systematic review and meta-analysis. Ageing Res Rev. 2014;16(1):12–31. doi: 10.1016/J.ARR.2014.05.002. [DOI] [PubMed] [Google Scholar]
- 47.Orsini N, Bellocco R, Bottai M, et al. Validity of self-reported total physical activity questionnaire among older women. Eur J Epidemiol. 2008;23(10):661–667. doi: 10.1007/S10654-008-9273-Z. [DOI] [PubMed] [Google Scholar]
- 48.Quinlan C, Rattray B, Pryor D, et al. The accuracy of self-reported physical activity questionnaires varies with sex and body mass index. PLoS One. 2021;16(8):e0256008. doi: 10.1371/JOURNAL.PONE.0256008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Warner ET, Wolin KY, Duncan DT, Heil DP, Askew S, Bennett GG. Differential accuracy of physical activity self-report by weight status. Am J Health Behav. 2012;36(2):168. doi: 10.5993/AJHB.36.2.3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Winckers ANE, MacKenbach JD, Compernolle S, et al. Educational differences in the validity of self-reported physical activity Energy balance-related behaviours. BMC Public Health. 2015;15(1):1–5. doi: 10.1186/S12889-015-2656-7/TABLES/2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Roig-Coll F, Castells-Sánchez A, Lamonja-Vicente N, et al. Effects of aerobic exercise, cognitive and combined training on cognition in physically inactive healthy late-middle-aged adults: the projecte moviment randomized controlled trial. Front Aging Neurosci. 2020;12:376. doi: 10.3389/FNAGI.2020.590168/BIBTEX. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Stern Y, Mackay-Brandt A, Lee S, et al. Effect of aerobic exercise on cognition in younger adults: A randomized clinical trial. Neurology. 2019;92(9):e905. doi: 10.1212/WNL.0000000000007003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Roychowdhury D. using physical activity to enhance health outcomes across the life span. J Funct Morphol Kinesiol. 2020;5(1). 10.3390/JFMK5010002 [DOI] [PMC free article] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data is available upon request.