Limiting postural sway within specific boundaries while standing is essential for maintaining stability and preventing falls. This has driven the development of biofeedback-based balance training protocols designed to minimize postural sway. Although visual biofeedback protocols have been shown to effectively reduce sway, this reduction is frequently associated with increased muscular effort. Consequently, it remains uncertain whether biofeedback-induced sway reduction results in optimal postural control.

This study aims to investigate whole-body stability during upright posture using biofeedback protocols by applying a set of variables characterizing the mathematical optimization process that minimizes the postural sway in terms of CoP coordinates.

Seventeen participants performed three 60-second postural tasks: (1) eyes open (EO) as a control; (2) center of pressure (CoP) biofeedback; and (3) wrist-controlled laser biofeedback. Posturographic variables were extracted to characterize the optimization process, focusing on downhill, stability, and convergence properties. The downhill property included: (1) percentage of signal duration with a strictly decreasing cost function (t, %), and (2) average maximum absolute convergence rate (|?|, mm/s). Stability was quantified by: (1) number of local minima (min_N), (2) their average values (min_L, mm), and (3) variability (min_SD, mm). Convergence was assessed by: (1) global minimum value (min_G, mm), and (2) absolute difference between global and expected minima (|?_GL|, mm). Statistical analysis used repeated measures ANOVA with Holm correction (p < 0.05) to assess condition effects.

Significant main effects of condition were revealed on the following outcomes: Expected Local Minimum (min.L) (F(2, 32) = 3.320, p = 0.049; BF COP: M = 3.364, SD = 7.064; BF Laser: M = 10.761, SD = 19.650; OA: M = 0.691, SD = 1.062), Dispersion of Local Minima (min.SD) (F(2, 32) = 3.622, p = 0.038; BF COP: M = 3.185, SD = 6.777; BF Laser: M = 11.126, SD = 20.083; OA: M = 0.553, SD = 0.972), and Error (delta.GL) (F(2, 32) = 3.351, p = 0.048; BF COP: M = -3.263, SD = 6.862; BF Laser: M = -10.579, SD = 19.308; OA: M = -0.651, SD = 1.023). Post hoc comparisons showed significant differences between BF Laser and OA for dispersion (p = 0.044) and trends for Expected Local Minimum (p = 0.055) and Error (p = 0.054). No significant effects were found for other outcomes (p > 0.05).

This study shows that different biofeedback protocols affect postural control optimization differently. Wrist-controlled laser biofeedback (BF Laser) led to higher local minima, greater dispersion, and larger errors compared to center of pressure (CoP) biofeedback and the eyes-open (EO) condition, suggesting BF Laser imposes greater stability demands. In contrast, CoP biofeedback and EO demonstrated more stable and convergent control, indicating a more efficient optimization process.

These results highlight the importance of selecting biofeedback protocols that align with the goals of postural training, as not all protocols equally promote optimal postural control. Future research should explore the underlying mechanisms and long-term effects of these protocols to refine biofeedback-based balance training strategies.