The Reliability Of Movement Complexity During Steady And Non-steady State Treadmill Running

The complexity of human movement could play an important role in our health. Control entropy (CE)—a measure of complexity—accounts for the non-stationarity of many physiological signals, potentially making it a suitable clinical monitoring tool. However, few investigations exist into CE’s reliability over clinically relevant time frames. PURPOSE: To investigate the short-term test-retest reliability of CE during steady and non-steady state treadmill running. METHODS: Sixteen (9 males, 7 females) recreationally active rearfoot-striking adults (age: 27.4 ± 3.9 years; BMI: 22.9 ± 3.2 kg/m2) attended 2 sessions, 7 days apart. Participants ran for 3-minutes at a self-selected speed before completing three 40-second treadmill running trials in random order at 2.5, 3.0, and 3.5 m/s in standardized footwear with insole-embedded inertial measurement units (500 Hz). We defined steady-state (SS) as the 30-second period for which the participants were running at the constant speeds. The non-steady state (NS) condition included the periods in which the treadmill belt accelerated up to these speeds and then decelerated to 0.0 m/s. Control entropy, calculated as the mean sample entropy of overlapping sliding windows, was calculated for the left and right resultant foot accelerations using a custom MATLAB script and the PhysioNet Toolbox. The resultant foot accelerations were calculated using the Euclidean norm. Inter-session reliability was measured using intraclass correlation coefficient (ICC) model (2,1) and standard error of measurement (SEM). The SEM was used to calculate the minimal detectable change (MDC). RESULTS: Moderate to excellent short-term reliability was observed across all speeds for left (ICC: 0.73 - 0.90, p < 0.05) and right (0.84 - 0.88, p < 0.05) CE during SS, paired with low SEM (3.2% - 5.0%) and MDC (8.8% - 11.8%). For the NS, good to excellent short-term reliability was observed across all speeds for left (0.78 - 0.92, p < 0.05) and right (0.85 - 0.90, p < 0.05) CE. Low SEM (3.2% - 4.6%) and MDC (8.9% - 12.9%) were demonstrated across all speeds for NS. CONCLUSION: The reliability of CE calculated during SS and NS is similar over clinically relevant time frames. Clinicians wishing to use movement complexity as a rehabilitation monitoring tool can reliably monitor CE from a 40-second run over a 7-day period. Mitacs Accelerate IT24201