Behavioral data. The average time taken to finish data collection after waking the participants was 18.1 ± 4.0 minutes (range 13 to 25 minutes). All participants self-reported that they fell asleep during the 50-min period provisioned for sleep. The average Karolinska Sleepiness Scale before the nap was 4.2 ± 1.1 (range 3 to 6), and after the nap, it increased to 5.7 ± 0.8 (range 4 to 7).
In terms of the range of postural sway acceleration in the sagittal plane (Fig. 1), ANOVA revealed a main effect of condition (F3,42 = 68.892, p < 0.001, partial eta squared = 0.831), and preplanned simple contrasts showed larger sway in the eyes-open foam-surface condition (F1,14 = 47.858, p < 0.001, partial eta squared = 0.774) and the eyes-closed foam-surface condition (F1,14 = 120.872, p < 0.001, partial eta squared = 0.896) than in the eyes-open firm-surface condition. ANOVA also revealed a main effect of time (F1,14 = 5.559, p = 0.033, partial eta squared = 0.284), with less sway after sleep.
In terms of the frequency of postural sway in the transverse plane (Fig. 2), no main effects were found, but a significant interaction effect of condition by sleep was observed (F3,42 = 3.517, p = 0.023, partial eta squared = 0.201). After sleep, the frequency increased in the eyes-open conditions but not in the eyes-closed conditions.
The postural sway in the middle-aged and elderly group seemed to be larger than that in the young group. However, in contrast to our primary hypothesis, there was neither main effect of group (F2,36 = 1.152, P = 0.327) nor interaction effect of group by time (F4,27 = 0.229, P = 0.922).
Furthermore, in contrast to our secondary hypothesis, a concurrent counting backwards task resulted in smaller sway dispersion; the main effect of dual-tasking condition was evident (F1,36 = 7.690, P = 0.009). There was also a significant interaction effect of dual-tasking condition by time by group (F4,72 = 2.902, P = 0.028). Preplanned simple contrasts, combing data of visual removal and visual addition, revealed significantly increased CoP sway from the pre to the post time interval only during single-tasking and only in the young group (t14 = -2.263, P = 0.040).
EEG data. Figure 3 illustrates the average scalp maps for the four component clusters in sensory association areas. Figure 4 shows the power changes in the alpha band as a function of condition and time. In the right occipital area, the Friedman test revealed a significant difference between sensory conditions (X2(3) = 43.380, p < 0.001). Post hoc analysis revealed significantly increased power in the eyes-closed firm-surface condition (Z = -3.845, p < 0.001) and the eyes-closed foam-surface condition (Z = -3.845, p < 0.001) than in the eyes-open firm-surface condition. In the left occipital area, the Friedman test revealed a significant difference between sensory conditions (X2(3) = 33.764, p < 0.001). Post hoc analysis revealed significantly increased power in the eyes-closed firm-surface condition (Z = -3.945, p < 0.001) and the eyes-closed foam-surface condition (Z = -3.717, p < 0.001) than in the eyes-open firm-surface condition. In the right parietal area, ANOVA revealed a main effect of sensory condition (F3,27 = 3.088, p = 0.044, partial eta squared = 0.255). Preplanned simple contrasts revealed significantly increased power in the eyes-closed firm-surface condition (F1,9 = 11.707, p = 0.008, partial eta squared = 0.565) than in the eyes-open firm-surface condition. Furthermore, ANOVA revealed a main effect of time in the right parietal area (F1,9 = 15.540, p = 0.003, partial eta squared = 0.633), showing decreased alpha power after sleep.
Figure 5 shows the power changes in the beta band. In the right occipital area, ANOVA revealed a main effect of sensory condition (F3,27 = 6.944, p = 0.026, partial eta squared = 0.436). Preplanned simple contrasts revealed nonsignificantly increased power compared to that in the eyes-open firm-surface condition, in the eyes-closed firm-surface condition (F1,9 = 7.325, p = 0.024, partial eta squared = 0.449) and in the eyes-closed foam-surface condition (F1,9 = 6.467, p = 0.032, partial eta squared = 0.418). In the left occipital area, the Friedman test revealed a significant difference between sensory conditions (X2(3) = 14.564, p = 0.002). Post hoc analysis revealed significantly increased power in the eyes-closed firm-surface condition (Z = -3.393, p = 0.001) and eyes-closed foam-surface condition (Z = -2.451, p = 0.014) than in the eyes-open firm-surface condition. Furthermore, the Wilcoxon signed-rank test revealed significantly increased beta power (Z = -2.241, p = 0.025) in the left occipital area after sleep.
Figure 6 shows the power changes in the fast beta band. In the right occipital area, ANOVA revealed a main effect of sensory condition (F3,27 = 4.681, p = 0.038, partial eta squared = 0.342). Preplanned simple contrasts revealed nonsignificantly increased power in the eyes-closed foam-surface condition (F1,9 = 5.267, p = 0.047, partial eta squared = 0.369) compared to that in the eyes-open firm-surface condition.
Figure 7 shows the power changes in the gamma band. In the right parietal area, ANOVA revealed a main effect of time (F1,9 = 6.626, p = 0.030, partial eta squared = 0.424), showing decreased gamma power after sleep.