Table 1.

Meta-analyses of RCTs evaluating the impact of exercise interventions on cognition.

First author, year		# of participants	Population (s)	Other selection criteria	Experimental exercise (# of studies)	Exercise dose	Control treatment	Meta-analysis results
Barha et al. (2017) [16]		39 RCTs (5,256 participants)	Healthy older adults 45+ years	(i) Mind-body exercise trials excluded (ii) Studies evaluating global cognition only excluded	AE (n = 19) RE (n = 9) Combined AE + RE (n = 13)	Frequency: 1–5 sessions/week Intensity: variable Duration:8–52 weeks	Active and inactive controls	(i) AE improved executive functioning relative to controls (g=2.064) (ii) AE interventions benefitted global cognitive function and executive functions more than RE (iii) Combined training benefitted global cognitive function and episodic memory more than AE and RE
Colcombe and Kramer (2003) [17]		18 RCTs 197 participants	Older adults, 55–80 years	(i) Date range 1966–2001 (ii) Measure of aerobic fitness	AE (49% of participants) or combined AE + RE (51% of participants)	Frequency: NR Intensity: NR Duration: 1–3 months (38%), 4–6 months (36%) 6+ months (26.7%)	NR	(i) Exercise had the greatest effect on executive function (g=0.68) (ii) Combined training produced larger improvements in cognition than AE alone (g=0.59 vs. 0.41) (iii) Long-term training associated with largest improvements in cognition (g=0.674)
Etnier et al. (1997) [18]		134 studies (17 clinical trials) 420 participants	All ages (exact ages NR)	(i) NR	AE and RE (numbers NR)	NR	NR	(i) Moderate effect of chronic exercise on cognition (g=0.33) (ii) Effect of clinical trials on cognition (g=0.18) (iii) Exercise session duration, frequency, and program duration were not significant moderators of cognition
Gothe and Mcauley (2015) [19]		22 (15 RCTs) 2,012 participants	All ages (mean age 62 years)	(i) Objective measure of cognitive function	Hatha yoga (n = 8) Iyengar yoga (n = 3) Integrated yoga therapy (n = 1), Sahaj yoga (n = 1)	Frequency: 1–5 sessions/wk Session duration: 45–120 mins Duration:1–6 months	Active and inactive	(i) Yoga had a moderate effect (g=0.33) on cognition (ii) Largest effect on attention and processing speed (g=0.29) followed by executive function (g=0.27) and memory (g=0.18)
Kelly et al. (2014) [20]		25 RCTs 2,217 participants	Healthy older adults with no cognitive impairment 50+ years	(i) Cardiovascular disease, other medical, psychiatric, or neurological conditions were excluded (ii) Date range 2002–2012	AE (n = 19), RE (n = 7), Tai chi (n = 3)	Frequency: 1–3 sessions/wk Intensity: variable Total duration: single bout to 1 year	No exercise, nonaerobic exercise, education, social or mental activities	(i) No improvements in AE vs stretching/toning on cognition (ii) Significant improvements for RE vs. stretching/toning in reasoning (p < 0.005) and Tai chi vs no exercise in attention (p < 0.001) and processing speed (p < 0.00001)
Northey et al. (2018) [21]		36 RCTs 2,748 participants	Older adults with and without cognitive impairment 50+ years	(i) Neurological and mental health populations excluded (ii) Unsupervised exercise interventions excluded (iii) Interventions <4 weeks excluded	AE (n = 18), RE (n = 13, multicomponent (n = 10), Tai chi (n = 4), yoga (n = 2)	Frequency: 1–5 sessions/wk Session duration: 20–90 min Intensity: variable, many NR Total duration: 6–52 weeks	Active (stretching, balance and tone, sham cognitive training, health education) and inactive (social interaction), no contact	(i) Improved cognitive function (SMD = 0.29; p < 0.01) with exercise of all types (ii) AE, RE, combined, and Tai chi interventions had significant effect estimates (p < 0.01) (iii) Session duration >45 min ≤ 60 min associated with improved cognition (p < 0.01) (iv) Moderate (p < 0.02) and vigorous (p < 0.01) intensity exercise (not low-intensity) associated with improved cognition
Sanders et al. (2019) [22]		36 RCTs 2,007 participants	Older adults with and without cognitive impairment 50+ years	(i) Treatment duration >4 weeks (ii) Excluded if did not specify exercise intensity (iii) Excluded studies if dose parameters were gradually increased	AE (n = 21) RE (n = 18) Multicomponent (n = 10) Balance (n = 2)	Frequency: 1–5 sessions/wk (mean = 2.62) Program duration: 4–52 weeks (mean = 22.3 weeks) Mean total dose: 2720 mins	Active and passive controls	(i) Small positive effect of exercise on executive function (d = 0.27) and memory (d = 0.24) in healthy older adults (ii) Exercise dose (type, session, duration, program duration, frequency, intensity) did not predict changes in cognition (iii) Shorter exercise sessions and higher frequency sessions predicted larger effects in those with cognitive impairment
Scherder et al. (2014) [23]		8 RCTs 642 participants	Sedentary healthy older adults 55+ years and older adults with cognitive impairment	(i) Published in English	AE (n = 8)	Frequency:1–7 sessions/wk Session duration: 30–60 mins Program duration: 4 weeks–1 year	Active and passive controls	(i) Walking improved set-shifting and inhibition in sedentary older adults without cognitive impairment (d = 0.36) (ii) No effect of walking on executive function among older adults with cognitive impairment (d = 0.14)
Smith et al. (2010) [24]		29 RCTs 2,049 participants	Healthy adults and MCI Mean age 18+ years 23 RCTs with older adults	(i) Treatment duration: >1 month	Supervised AE (n = 29)	Frequency: 1–5 sessions/wk Intensity: variable, many NR Duration: 8–72 weeks	Nonaerobic exercise, waitlist, education, stretching, social activities	(i) Modest improvements in attention and processing speed (g=0.158), executive function (g=0.123) (ii) Exercise intensity and duration did not moderate effects on memory (iii) Older adults demonstrated larger improvements in working memory compared to younger participants
Wayne et al. (2014) [25]		11 RCTs 2,553 participants	Older adults with and without cognitive impairment Mean age 60 years, except 1 study with adults	(i) Measure of cognitive function	Tai chi (n = 11)	Frequency: 1–4 sessions/week Intensity: variable, many NR Duration: 10 weeks–1 year	Active and nonactive controls	(i) Large effect of Tai chi vs no exercise on executive function (g=0.90) and moderate effect vs exercise controls (g=0.51) in healthy older adults
Wu et al. (2018) [26]		32 RCTs 3,624 participants	Older adults with and without cognitive impairment aged 55–80 years	(i) Measure of cognitive function	Tai chi (n = 18) Yoga (n = 8) Dance (n = 6)	Frequency: 1–4 sessions/week Intensity: NR Session duration: 20–60 mins/session, Duration: 2–48 wks	Active and nonactive controls	(i) Improvements in global cognition compared with controls (mean difference = 0.92), particularly cognitive flexibility, working memory, verbal fluency, and learning (ii) Moderate dose (60–120 mins per week) significantly improved global cognition compared to controls
Young et al. (2015) [27]		12 RCTs 754 participants	Healthy older adults >55 years	(i) Measure of CV fitness	AE (n = 12)	Frequency: 1–5 sessions per week Intensity: variable Duration: 8–26 weeks	No treatment, nonaerobic exercise, social or mental activities	(i) No benefit of AE vs active or inactive controls on any of the 11 cognitive domains (ii) Incresed fitness did not coincide with improvements in cognition.
Zhang et al. (2018) [28]		19 RCTs 2,539 participants	60+ years with and without cognitive impairment	(i) Measure of cognitive function (ii) Articles in English or Chinese	Tai chi (n = 12) Yoga (n = 4) Qigong (n = 2) Pilates (n = 1)	Frequency: 1–7 sessions/wk Session duration: 20–120 min Intensity: NR, Duration: 8–52 weeks.	Active or passive controls	(i) Benefits global cognition (g=0.23), executive functions (g=0.25–0.65), learning and memory (g=0.37–0.49), and language (g=0.35) (ii) Mind-body exercise more effective for older adults without cognitive impairment (iii) Total training time predictor of global cognition, executive function, and language

RCT, randomized controlled trial; MCI, mild cognitive impairment; NR, not reported; min, minutes; AE, aerobic exercise; RE, resistance exercise.