Table 1

Results of moderator analysis

Moderator	No. of effect sizes	Estimate Mean (95% CI)	Q statistic	Omnibus test of moderators
Exercise moderators
Mode			Q₃₂₈=781.68; p<0.01	Q₅=39.53; p<0.01
Aerobic	153	0.24 (0.10 to 0.37)
Resistance training	80	0.29 (0.13 to 0.44)
Multicomponent training	47	0.33 (0.14 to 0.53)
Tai chi	25	0.52 (0.32 to 0.71)
Yoga	28	0.13 (−0.10 to 0.36)
Duration			Q₃₁₈=789.68; p<0.01	Q₃=27.83; p<0.01
Short (≤45 min)	36	0.09 (−0.28 to 0.46)
Medium (>45 to ≤60 min)	263	0.31 (0.16 to 0.46)
Long (>60 min)	24	0.33 (−0.04 to 0.65)
Frequency			Q₃₂₉=804.58; p<0.01	Q₃=24.12; p<0.01
Low (≤2)	92	0.32 (0.13 to 0.52)
Medium (3–4)	229	0.24 (0.07 to 0.40)
High (5-–7)	13	0.69 (0.10 to 1.28)
Intensity			Q₂₀₇=264.61; p<0.01	Q₃=13.55; p<0.01
Low	71	0.10 (−0.02 to 0.23)
Moderate	57	0.17 (0.03 to 0.33)
High	83	0.16 (0.04 to 0.27)
Length			Q₃₃₀=807.48 P<0.01	Q₃=23.32; p<0.01
Short (4–12 weeks)	78	0.31 (0.09 to 0.54)
Medium (13–26 weeks)	170	0.28 (0.10 to 0.47)
Long (>26 weeks)	86	0.27 (0.03 to 0.52)
Cognitive moderators
Cognitive domain			Q₃₀₃=795.06; p<0.01	Q₅=28.08; p<0.01
Global cognition	6	0.16 (−0.14 to 0.47)
Attention	87	0.27 (0.14 to 0.41)
Executive function	94	0.34 (0.20 to 0.47)
Memory	81	0.36 (0.22 to 0.50)
Working memory	36	0.29 (0.12 to 0.45)
Study design moderators
Control group			Q₃₂₈=785.37; p<0.01	Q₄=25.52; p<0.01
Active	120	0.13 (−0.06 to 0.32)
Education	17	0.48 (0.14 to 0.82)
No contact	189	0.34 (0.17 to 0.51)
Social	7	0.20 (−0.58 to 0.98)
Cognitive status			Q₃₃₀=797.56; p<0.01	Q₃=21.62; p<0.01
MCI - Yes	197	0.28 (0.11 to 0.44)
MCI - No	41	0.36 (0.04 to 0.68)
Unclear	96	0.28 (0.05 to 0.51)

MCI, mild cognitive impairment.