Temporal Variations In Gait Phases, Not Cycle Definition, Influence Statistical Parametric Mapping Results

Statistical parametric mapping analysis (SPM) provides a statistical approach for comparing time-continuous curves, such as kinematics of gait cycles (Pataky, 2010). A recent study indicated that SPM results were sensitive to the definition of the start of each stride (Honert & Pataky, 2021). However, temporal alignment of phases within a stride may be necessary for meaningful comparisons between gait cycles (Helwig et al., 2011). PURPOSE: To compare SPM results of different walking speeds, using different gait cycle definitions, in temporally aligned and unaligned gait data. METHODS: 10 young healthy adults (f = 5, age = 25.2 ± 3.2 years, height = 1.72 ± 0.8 m, weight = 74.0 ± 12.2 kg) participated in this study. An instrumented mat was used to determine preferred speed overground. An instrumented treadmill, ten-camera Vicon motion capture system and Conventional Gait Model 2.4 were used to obtain biomechanical data of the legs (Nexus 2.13 software), while participants walked at 80, 90, 100, 110 and 120% of preferred speed. One hundred gait cycles at each speed for each participant were selected and temporally aligned, based on kinematic points of interest, to five gait phases (loading response, mid and terminal stance, pre-swing, initial and mid swing, and terminal swing), using custom MATLAB code. SPM with one-way repeated measures ANOVA was performed with five different stride definitions both with and without temporal alignment of gait phases. RESULTS: SPM results were inconsistent between different stride definitions for unaligned but were consistent for temporally aligned gait data (Figure 1). CONCLUSIONS: Consistent results of SPM for aligned data show that SPM is not sensitive to stride definition, but time normalization. Temporal alignment of phases is necessary for point-by-point comparison of gait kinematics using SPM .