Using inertial signals to characterize main lower limb gait patterns in individuals post-stroke

Introduction/Background Different lower limb (LL) gait patterns (spastic, flexum, hyperextension, buckling) have been described post-stroke that require specific interventions (strengthening, botox, medication, orthosis). We explored whether the analysis of angular velocities of the femur and tibia, easily recorded using wearable inertial sensors, might be a good tool to assess LL gait deviations. The objective was to define how the sagittal angular velocity of the femur (ωF) and tibia (ωT) change with walking speed, hemiparesis and LL gait patterns. Material and method Twenty chronic hemiparetic individuals, with an inertial sensor (OPAL, APDM inc.) fixed at each tibia and thigh, walked over ground at self-selected walking speed in the laboratory. A physiotherapist noted the presence of LL gait pattern and recorded with a camera the LL movement. Participants also stayed 5 seconds in upright position and did 5 squats to calibrate the system. In addition, three healthy controls walked on a treadmill at comfortable, 0.8 and 0.4 m/s speeds. A preliminary analysis confirmed that treadmill data could be used as reference for data collected over ground. The normal range of ωT and ωF was determined using ± 2 SD around the mean values at similar walking speed. Results The speed changed the amplitude of ωT and ωF and the profiles of some individuals post-stroke differed from normal limits. In seven participants with specific LL gait pattern, their ωT and ωF revealed specific profiles, with typical deviations to each LL pattern. Clear deviations from the normal ωT and ωF profiles were also observed for two participants having deviations less observable with the naked eye. Conclusion These preliminary findings suggest that wearable inertial sensors have the potential to characterize specific LL movement deviations and add objective data to gait clinical assessment. It could be an interesting complement to assess the treatment effects.