Tactile Biofeedback That Targets Stance Time Acutely Modulates Propulsion Mechanics in Healthy Gait

Technical Summary: Tactile Biofeedback Targeting Stance Time Modulates Propulsion Mechanics

Problem Statement
Stroke survivors frequently exhibit temporal and propulsive asymmetries in their gait, which contribute to increased metabolic costs, reduced walking speed, and compromised balance control. While temporal improvements are critical, they often show resistance to change. Previous research suggests that effective weight bearing (stance time) on the paretic limb is necessary for sufficient push-off and propulsion. Although biofeedback has been shown to improve stance time and propulsion independently in chronic stroke populations, most existing techniques target symmetry ratios. This approach can be difficult for patients to follow and may induce unwanted compensatory movements in the non-paretic limb. Furthermore, prior studies have rarely analyzed whether altering one specific gait parameter (stance time) via biofeedback acutely induces changes in other interdependent features, specifically propulsive mechanics. There is a need to determine if instructing a specific increase in stance time for a single limb can simultaneously improve propulsive features, potentially offering a unified, low-cost rehabilitation strategy compared to external assistive devices like exoskeletons.

Methodology
The study utilized a within-subjects design with eight healthy adults (mean age 27.6 ± 3.8 years) walking on a split-belt instrumented treadmill at a fixed speed of 1.0 m/s.

• Biofeedback System: A custom real-time controller (MATLAB) processed motion capture data (100 Hz) and ground reaction forces (1000 Hz) to detect gait events (heel strike, toe-off) and peak propulsive timing.
• Stimulation: Tactile feedback was delivered via six vibrating tactors embedded in custom insoles (three per foot). Subjects received vibration on only one limb at a time.
• Instruction Protocol: Subjects were instructed to synchronize their push-off with the onset of vibration. A "success" vibration (constant 300ms pulse) indicated the previous step met the target stance time, while a "failure" vibration (pulsing 40ms on/20ms off) indicated the target was not met.
• Experimental Conditions: Targets were calculated using symmetry ratios (ranging from 0.775 to 1.225) applied to the most recent 10 baseline steps. This created "Small," "Medium," and "Large" stance time increase targets for either the left or right limb.
• Data Analysis: Stance times and propulsive variables (Trailing Limb Angle [TLA], peak propulsive force, and propulsive impulse) were analyzed using linear mixed-effects models to assess condition effects and correlations.

Key Contributions

1. Acute Modulation of Propulsion via Temporal Feedback: The study demonstrates that providing tactile biofeedback instructing only an increase in stance time (without direct instruction on force or angle) acutely increases key propulsive features.
2. Specificity of Stance Time Instruction: The system successfully instructed specific, graded increases in stance time for a single limb, though some carry-over effects were observed in the uninstructed limb.
3. Quantification of Relationships: The research establishes strong statistical correlations between instructed stance time increases and improvements in TLA, peak propulsive force, and propulsive impulse in healthy gait.

Results

• Stance Time: During the "Large" stance time increase target, subjects significantly increased their stance time by 0.23 ± 0.02 seconds ($p < 0.001$) compared to no feedback. While the instructed limb showed the greatest increase, the uninstructed limb also showed significant increases for small, medium, and large targets, likely due to a tendency toward symmetric walking.
• Propulsive Features: The "Large" feedback condition resulted in significant increases in:
  • Trailing Limb Angle (TLA): +8.1 ± 0.8° ($p = 0.001$)
  • Peak Propulsive Force: +8.1 ± 1.1% Body Weight ($p = 0.002$)
  • Propulsive Impulse: +2.9 ± 0.3% Body Weight·s ($p = 0.002$)
• Correlations: Stance time changes were significantly correlated with TLA ($R^2 = 0.80$) and propulsive impulse ($R^2 = 0.81$), and to a lesser degree with peak propulsive force ($R^2 = 0.72$), after accounting for subject variance.
• Specificity: While the system could distinguish between different levels of instruction (e.g., Small vs. Large), the distinction was most robust between the Baseline and all other conditions, with some overlap between intermediate levels.

Significance and Claims
The authors claim that this study provides evidence for a biofeedback tool capable of influencing propulsive features by instructing only stance times. The significance lies in the potential application of this technology for stroke rehabilitation. Specifically, the authors propose that a tactile device instructing stance time increases could serve as a "cheaper alternative" to external assistances (such as exoskeletons) or complex ground reaction force biofeedback systems. By leveraging the relationship between stance time and propulsion, such a device could simultaneously target temporal and propulsive deficits in the paretic limb of stroke survivors.

The authors remain modest regarding the immediate clinical application, noting that the results were obtained from healthy individuals on a treadmill at a fixed speed. They acknowledge that overground walking introduces different velocity constraints and strategies that may alter the stance-propulsion relationship. Consequently, the paper concludes that while the mechanism is promising, future work is required to validate these effects in individuals with stroke during overground walking before confirming its efficacy as a unified rehabilitation strategy.