Individuals with spinal cord injury (SCI) who regain functional walking capacity present a high risk of falling. Since most falls happen during walking, a need exists to characterize gait stability more precisely. Wearable sensors can provide precise gait stability parameters that can augment clinical measurements while being more affordable and unobtrusive than lab assessments. The validity of gait stability parameters extracted from wearable sensors has not been evaluated in the SCI population. To evaluate the validity of gait stability parameters derived from wearable sensors in functional ambulators with SCI and able-bodied individuals (AB) through five walking conditions that progressively challenge stability.

Twenty-one individuals with a motor incomplete SCI who were able to walk without assistance for ≥20 meters and 19 AB individuals were recruited. Participants walked 20 meters under five conditions that challenged stability:

– normal (i.e. hard surface with vision);

– reduced vision;

– soft surface;

– reduced vision and soft surface;

– a dual cognitive task.

Participants wore five sensors (feet, shanks and pelvis) while walking. Inter-stride means and coefficients of variation (CV) of five gait parameters (cadence, double support percentage of gait cycle, gait cycle time, stride length, and stride speed) were used to characterize gait. Gait parameters were compared across conditions and groups using a mixed-effects model and Tukey test for post-hoc comparisons.

A significant difference existed between individuals with SCI and AB individuals for the mean and CV of all gait parameters (P≤0.03). Individuals with SCI had more variability than AB individuals and demonstrated lower gait capacity (e.g. lower cadence and greater double support percentage). In individuals with SCI, significant differences existed for most gait stability parameters between:

– normal walking and walking with reduced vision and soft surface (P≤0.01);

– walking with reduced vision and walking with reduced vision and soft surface (P≤0.05).

A significant difference also existed between normal walking and walking on soft surface for the mean cadence and stride speed (P≤0.02). No between-conditions differences existed in AB individuals.

Wearable sensors can augment measurements of walking under various challenging situations. Parameters extracted from these devices can reveal differences between populations. Walking on soft surface and reduced vision was the most challenging task in individuals with SCI.