Walking modulates visual detection performance according to stride cycle phase

While the effects of exercise on cognition are well-studied, the impact of locomotion on perception within the stride cycle remains largely unexplored. We take experiments beyond static laboratory settings to examine the influence of active observation during walking on continuous visual performance. Previous research has hinted at a link between visual performance and gait phases, showing that responses to visual information depend on the phase of human locomotion. Eye movements and visuomotor coordination also show coupling to locomotion phases. This suggests dynamic demands within each step may modulate visual processing during walking. This study uses wireless, position-tracked virtual reality to continuously probe visual detection performance during natural walking, offering a fine-grained analysis of performance within each stride cycle. We hypothesize that visual performance will show rhythmic changes synchronized with the gait cycle.

Existing research has generally focused on the overall effects of walking and exercise on cognitive function over extended periods rather than within-stride cycle modulations. Studies have shown that moderate exercise can improve performance on cognitive tasks and enhance neuroplasticity. Other studies have investigated the effects of dual-tasking (walking while performing another task) on gait parameters. More recent research demonstrates that walking or light exercise enhances visual processing compared to stationary conditions, potentially through elevated response gain in early visual processing areas. However, the majority of these studies have not investigated temporal changes within the stride cycle. This study aims to fill this gap by examining performance oscillations within the gait cycle.

Thirty-six participants (22 female; mean age = 19.6, SD = 2.6) with normal or corrected-to-normal vision performed a visual detection task in a wireless virtual reality environment. They walked along a 9.5m track at a self-selected comfortable pace while responding to briefly presented visual targets within a drifting circular annulus. Target contrast was adaptively adjusted to maintain accuracy around 75% using a QUEST staircase procedure. Head position was tracked at 90 Hz to define stride cycles. Performance (accuracy, reaction time, response likelihood) was analyzed relative to the phase of the stride cycle using a Fourier analysis, with a non-parametric shuffling procedure to assess significance. Bayesian inference was used to determine the population prevalence of oscillations. Eye movements were also recorded and analyzed to rule out gaze-related artifacts. Data analysis was conducted using MATLAB and JASP, including psychometric fits using the psignifit toolbox.

Walking increased the threshold for visual target detection compared to stationary conditions, despite maintaining similar overall accuracy (around 75%). Reaction times were faster while walking. Analysis revealed clear oscillations in accuracy, reaction times, and response likelihood within the stride cycle, predominantly at approximately 2 cycles per stride (cps). These oscillations were highly prevalent in the sample. Bayesian inference showed a high prevalence of oscillations at 2 cps (accuracy MAP = 0.32, [0.18, 0.49]; reaction time MAP = 0.30, [0.16, 0.47]; response likelihood MAP = 0.41, [0.26, 0.58]). A substantial portion of participants also exhibited oscillations at 4 cps. While the exact frequency was idiosyncratic to each participant, the phase of the oscillations was remarkably consistent across individuals. Optimal performance for accuracy and reaction time was observed during the swing phase, while response likelihood peaked around heel strike. Analyses ruled out mechanical artifacts caused by footfalls as the explanation for the observed oscillations.

The findings demonstrate a rhythmic modulation of visual performance synchronized with the gait cycle, suggesting a close interplay between locomotion and visual processing. The oscillations, primarily at 2 cps, may reflect the resetting of neural oscillations with each step, a phenomenon observed in other action-perception paradigms. The observed oscillations might also arise from reverberation frequencies within neural circuits related to both motor control and sensory processing. The elevated sensory cortical response gain associated with locomotion might interact with stride cycle phases to produce these oscillations. The study's results extend prior research on brain-body coupling, particularly the influence of cardio-respiratory rhythms on perception, by showing how locomotion also rhythmically modulates sensory performance. The idiosyncratic frequency variations could be linked to cardiorespiratory exertion, a topic for future investigation. The timing of perceptual modulations aligns with previous research showing phasic periods of increased sensory demand during walking, particularly before heel strike when vestibular signals are weighted more heavily for balance.

This study provides compelling evidence for rhythmic modulation of visual perception during walking, closely linked to the phases of the stride cycle. This novel finding opens several avenues for future research, including investigating attention allocation over the stride cycle, visual field uniformity of modulations, and similar effects in auditory or tactile tasks. The established methodology provides a valuable tool for further exploration of the dynamic interplay between locomotion and perception.

The study's use of a virtual reality environment might limit the generalizability of the findings to real-world walking scenarios. The relatively simple visual detection task might not fully capture the complexities of natural visual processing during locomotion. Future research could investigate more ecologically valid tasks and real-world walking conditions. The relatively homogeneous sample of university students may limit the generalizability of the findings to other populations. The study did not directly measure neural activity to confirm the hypothesized mechanisms underlying the observed oscillations.