Effects of delay and amplification of auditory feedback for walking: anticipation, variability, and frequency adaptation

When walking, we generate rich acoustic information through our footsteps. This sound stream contains information about footfall timing, dragging, loading rate, etc. The role of this endogenous auditory feedback in gait control remains underexplored. Building on work in delayed auditory feedback (DAF) for speech and on theories of sensorimotor coupling, we investigated whether manipulating the delay and amplification of self-produced footstep sounds modulates gait dynamics. Thirty healthy young adults walked overground while receiving real-time lateralized playback of the sound close to their feet using shoe-mounted microphones, belt-worn micro-computer, and headphones, all connected with cables for minimal latency. Across conditions, auditory feedback was delivered with no delay, low delay (12.5% step duration), or high delay (25% step duration), and at either full or half amplification. There was also a masked hearing condition with pink noise. Spatiotemporal gait parameters, namely cadence, speed, stride length, and coefficient of variation were analyzed as percent change relative to baseline. We found that amplification without delay reduced variability by almost 10% on average, consistent with strengthened sensorimotor coupling via enhanced perceptual access to foot-ground interaction dynamics. A second interesting finding was that delay increased the cadence of walking instead of reducing it, contrary to our expectations. We discussed how this effect can be explained by both Bayesian predictive coding and anticipatory dynamic systems with delayed feedback. We developed a theoretical model with anticipatory dynamics, with implications for closed-loop gait rehabilitation tools.