The statistical analysis indicated there was a significant main time (F1,12=13.5 p = 0.003, ηp2:0.53), group (F1,12=28.1 p = 0.001, ηp2:0.70) and interaction effect (F1,12=40.5 p = 0.001, ηp2: 0.77) for MoCA. In details, we observed a significant main time (F1,12=7.9 p = 0.016, ηp2:0.40), group (F1,12=5.0 p = 0.044, ηp2:0.30) and interaction effect (F1,12=13.6 p = 0.003, ηp2: 0.53) for attention and working memory. In addition, there were significant differences at short term memory (time: F1,12=3.2 p = 0.101, ηp2:0.21), group: F1,12=12.9 p = 0.004, ηp2:0.52 and interaction effect: F1,12=27.0 p = 0.001, ηp2: 0.69) and executive function and visuospatial power (time: F1,12=0.1 p = 0.991, ηp2:0.01), group: F1,12=22.8 p = 0.001, ηp2:0.66 and interaction effect: F1,12=38.8 p = 0.001, ηp2: 0.76) between groups. However, there were not significant differences between group at orientation (time: F1,12=0.6 p = 0.468, ηp2:0.05), group: F1,12=0.7 p = 0.436, ηp2:0.05 and interaction effect: F1,12=2.2 p = 0.165, ηp2: 0.15), language (time: F1,12=4.5 p = 0.055, ηp2:0.27), group: F1,12=0.23 p = 0.636, ηp2:0.02 and interaction effect: F1,12=3.8 p = 0.075, ηp2: 0.24) and abstraction (time: F1,12=0.02 p = 0.901, ηp2:0.01), group: F1,12=3.3 p = 0.096, ηp2:0.21 and interaction effect: F1,12=1.9 p = 0.190, ηp2: 0.14). For GDS, there was no significant main time (F1,12=0.2 p = 0.631, ηp2:0.02), but a significant in group (F1,12=23.7 p = 0.001, ηp2:0.66) and interaction effect existed (F1,12=21.2 p = 0.001, ηp2:0.64).
***Figure 2 near here***
Descriptive statistics of performance and perceptual parameters pre- and post-intervention are summarized in Table 1. In overall, TG compare to CG demonstrated substantial improvements in all performance indices following a 12-week intervention. We found a significant main time (F1,12=6.4 p = 0.026, ηp2:0.35), group (F1,12=40.0 p = 0.001, ηp2:0.77) and interaction effect (F1,12=53.7 p = 0.001, ηp2:0.82) for 6 min walking. For chair sit and reach, there was no significant main time (F1,12=0.9 p = 0.342, ηp2:0.07), though a significant in group (F1,12=87.6 p = 0.001, ηp2:0.88) and interaction effect existed (F1,12=135.9 p = 0.001, ηp2:0.92). Furthermore, Following the 12-week intervention, we found a significant main time (F1,12=11.2 p = 0.006, ηp2:0.48), group (F1,12=80.2 p = 0.001, ηp2:0.87) and interaction effect (F1,12=61.3 p = 0.001, ηp2:0.84) for strength of preacher biceps curl. For strength of knee extensions, there also was a significant main time (F1,12=6.1 p = 0.030, ηp2:0.34), group (F1,12=25.1 p = 0.001, ηp2:0.68) and interaction effect (F1,1238.8.2 p = 0.001, ηp2:0.76). For strength of handgrip, there was no significant main time (F1,12=2.3 p = 0.152, ηp2:0.16), but significant in group (F1,12=63.6 p = 0.001, ηp2:0.84) and interaction effect (F1,12=74.2 p = 0.001, ηp2:0.86).
Table 1
Performance and psychological characteristics of participants pre- and post-intervention
Variable | Group | Pre | Post | % change | Cohen’s d | p |
6 min walking (m) | TG | 177.0 ± 81.5 | 318.9 ± 85.5 | 96.9 | 1.9 | 0.001 |
CG | 180.0 ± 66.2 | 174.2 ± 62.9 | -2.6 | -0.5 |
Knee extension (kg) | TG | 10.8 ± 5.6 | 23.1 ± 10.6 | 134.6 | 1.6 | 0.001 |
CG | 10.7 ± 4.9 | 10.0 ± 4.5 | -4.5 | -0.6 |
Biceps curl (kg) | TG | 6.4 ± 1.7 | 10.6 ± 2.5 | 70.2 | 2.6 | 0.001 |
CG | 6.4 ± .1.6 | 6.1 ± 1.5 | -3.1 | -0.3 |
Hand grip (kg) | TG | 23.1 ± 9.1 | 31.6 ± 8.9 | 47.9 | 2.4 | 0.001 |
CG | 22.61 ± 8.3 | 21.3 ± 8.1 | -7.1 | -1.4 |
30 s stand-up (N) | TG | 10.3 ± 3.4 | 18.5 ± .37 | 94.8 | 1.9 | 0.001 |
CG | 9.7 ± 3.1 | 9.2 ± 2.3 | -2.7 | -0.5 |
Timed Up and Go test (s) | TG | 11.8 ± 2.5 | 6.4 ± 1.4 | -45.6 | -3.5 | 0.001 |
CG | 11.8 ± 2.7 | 12.4 ± 2.9 | 5.7 | 0.8 |
Chair sit and reach (cm) | TG | 18.5 ± 8.1 | 26.9 ± 7.6 | 54.5 | 3.9 | 0.001 |
CG | 19.8 ± 8.1 | 18.7 ± 7.1 | -3.7 | -0.5 |
MoCA | TG | 18.6 ± 3.5 | 23.9 ± 2.3 | 28.4 | 1.7 | 0.001 |
CG | 19.0 ± 2.1 | 17.9 ± 2.2 | -3.3 | -0.5 |
GDS | TG | 5.4 ± 2.9 | 2.6 ± 1.9 | -49.5 | -1.4 | 0.001 |
CG | 4.4 ± 2.5 | 4.5 ± 2.5 | 6.8 | 0.2 |
Alpha wave (%) | TG | 80.0 ± 5.5 | 83.2 ± 2.7 | 4.3 | 0.7 | 0.002 |
CG | 78.8 ± 7.7 | 78.0 ± 7.6 | -1.0 | -0.6 |
Beta wave (%) | TG | 3.7 ± 3.1 | 8.9 ± 3.3 | 218.3 | 2.6 | 0.001 |
CG | 3.1 ± 1.6 | 3.0 ± 1.1 | 5.0 | -0.2 |
Theta wave (%) | TG | 16.2 ± 6.5 | 7.8 ± 3.7 | -51.8 | -1.9 | 0.001 |
CG | 18.1 ± 6.9 | 19.0 ± 7.0 | 5.9 | 0.6 |
TG: Training Group, CG: Control Group; MoCA: The Montreal Cognitive Assessment; GDS: Geriatric Depression Scale |
For functional indices, we found a significant main time (F1,12=12.7 p = 0.004, ηp2:0.52), group (F1,12=90.9 p = 0.001, ηp2:0.88) and interaction effect (F1,12=172.1 p = 0.001, ηp2: 0.94) for timed up and go test. In addition, there was a significant main time (F1,12=29.0 p = 0.001, ηp2:0.71), group (F1,12=54.6 p = 0.001, ηp2:0.82) and interaction effect (F1,12=41.1 p = 0.001, ηp2: 0.77) for chair stand.
***Table 1 near here***
***Figure 3 near here***
Following 12 weeks combined training, the percentage of resting average frequency of brain waves in occipital region in the TG increased significantly by 14.5% from, change from alpha range to beta frequency (11.51 Hz to 13.15) Hz, but there was no significant change (-1.4%) in control group (11.13 Hz to 10.95 Hz). Descriptive statistics of the brain waves are presented in Table 1. The results of repeated measure ANOVA showed there were significant main time (within group) (F1,12 =11.4 p = 0.005, ηp2:0.48), group (between group)(F1,12 =63.7 p = 0.001, ηp2: 0.84) and interaction (group × time) effects (F1,12=39.7 p = 0.001, ηp2:0.77) for resting average frequency of brain waves (Fig. 3). We found no significant main time (F1,12=3.2 p = 0.098, ηp2:0.21), and group (F1,12=3.6 p = 0.080, ηp2:0.23) effect, though a significant interaction effect existed (F1,12 =6.7 p = 0.024, ηp2:0.36) for percentage of alpha waves. While for percentage of beta waves, significant main time (F1,12=19.2 p = 0.001, ηp2:0.62), group (F1,12=77.2 p = 0.001, ηp2:0.86) and interaction effect (F1,12=82.1 p = 0.001, ηp2:0.87) were observed. For percentage of theta waves, there were significant main time (F1,12=14.7 p = 0.002, ηp2:0.55), group (F1,12 =39.5 p = 0.001, ηp2:0.77) and interaction effect (F1,12=46.2 p = 0.001, ηp2:0.79). There were significant main time (F1,12=10.5 p = 0.007, ηp2:0.47), group (F1,12=29.1 p = 0.001, ηp2:0.71) and interaction effect (F1,12=33.7 p = 0.001, ηp2:0.74) for theta/alpha ratio (Fig. 3).
Table. 2 presents the bivariate Pearson’s correlation coefficient (r) between the percentage change of performance parameters and MoCA and GDS. In general, there were moderate to large, positive correlations between MoCA changes and performance induces. Moderate, negative correlations was found between changes in GDS, and performance induces. In addition, MoCA correlated negatively with theta/alpha ratio, while GDS correlated positively.
***Table 2 near here***
Table 2
The Pearson’s correlation coefficient between the variables.
| | Hand grip | Knee extension | Biceps curl | 30 s stand-up | Timed Up and Go test | 6 min walking | Chair sit & reach | theta/alpha ratio |
MoCA | r | 0.81 | 0.78 | 0.63 | 0.68 | -0.68 | 0.36 | 0.74 | -0.56 |
p | 0.001 | 0.001 | 0.001 | 0.001 | 0.001 | 0.067 | 0.001 | 0.003 |
GDS | r | -0.50 | -0.55 | -0.61 | -0.203 | 0.56 | -0.57 | -0.47 | 0.62 |
p | 0.010 | 0.004 | 0.001 | 0.319 | 0.003 | 0.003 | 0.016 | 0.001 |
The mean (SD) of HR during the intervention period was presented in the Fig. 4. The training begun at 50% of maximal HR and reached to 70% of maximal HR toward the end of intervention. The range of HR were 80 to 125 beat per minute.
***Figure 4 near here***