Comparative Analysis of Entropy in Accelerations and Muscular Near-Infrared Spectroscopy (NIRS) Signals during Running: Dominant versus Non-Dominant Legs

Abstract

Sample entropy (SampEn) is a valuable tool for assessing variability in kinematic patterns and physiological signals during exercise (Estep et al., 2017). This variability may vary between dominant (D) and non-dominant (ND) limbs, particularly under different task-specific conditions. This study aimed to compare the variability in muscle oxygen saturation (SmO₂) and playerload (PL) between the D and ND legs during various running variations. After a familiarisation session, fifteen healthy individuals (age = 31.4 ± 9.3 years) performed six running variations randomly: high-knees (90° knee angle), tethered, shoulder-wide, shoulder-wide coreboard, treadmill (10.9 m/s at 0% slope), and slope treadmill (10.9 m/s at 6% slope). The running was carried out for 90 seconds, interspersed with a 5-minute rest (cadence: 130-150 bpm). To the PL data, two inertial measurement units (Blue Trident, Vicon, NZ) were placed on both ankles, while SmO2 was collected using two NIRS devices fixed on the vastus lateralis of both legs. SampEn was calculated from acceleration and physiological data as a measure of variability. A one-way ANOVA was conducted (JASP software version 0.16.4.0) to assess differences between the D and ND leg, considering a significance level of p≤0.05. No significant differences were found in PL between the D and ND legs across different running variations (0.3015 ≤ p ≤ 0.839). However, the results showed that the D leg exhibited higher SampEn values in three specific variations: high knees (D = 0.090 vs. ND = 0.087), tethered (D = 0.101 vs. ND = 0.096), and shoulder-wide (D = 0.114 vs. ND = 0.109), indicating higher variability. Conversely, the ND leg demonstrated increased complexity and irregularity in movement patterns during running performed on horizontal (D = 0.112 vs. ND = 0.122), sloped (D = 0.119 vs. ND = 0.123), and unstable (D = 0.162 vs. ND = 0.181) surfaces compared to the D leg. Likewise, no significant differences were observed in SmO₂ between D and ND legs across different running variations (0.101 ≤ p ≤ 0.882). However, interestingly, the ND leg generally exhibited higher SampEn values compared to the D leg in most conditions, except for high knees and slope treadmill, where the values were the same (N and ND = 0.017 and 0.014, respectively). This study revealed that the D leg exhibited more complex and unpredictable movement patterns compared to the ND leg, as evidenced by higher entropy values for accelerations. Surprisingly, the ND leg showed higher muscle SmO₂ variability. These findings emphasise the need to consider limb dominance and task-specific conditions when assessing movement pattern variability, offering valuable insights for sports training and performance optimisation.