Motor variability as an index of fatigue in dynamic actions: a perspective from the optimal movement variability theory

Fatigue is a complex process that affects both force production and movement execution. Traditional measures, such as countermovement jump (CMJ) height, assess performance decrements but fail to capture compensatory movement adaptations. Nonlinear analysis of motor variability derived from acceleration signals provides a novel approach to monitoring fatigue by detecting subtle changes in movement execution. This study examined fatigue induced by three resistance training modalities-power, hypertrophy, and maximal strength-on motor variability during squats. Forty-four participants performed 10 squats at 70% of 1 repetition maximum (RM) before and after a training session, with follow-up assessments at 24, 48, and 72 h. Lower-back acceleration was recorded using inertial measurement units (IMUs). Acceleration variability was analyzed in terms of magnitude [standard deviation (SD)] and temporal structure [fuzzy entropy (FuEn); detrended fluctuation analysis (DFA)]. CMJ height served as a traditional marker of fatigue. Significant reductions in CMJ height were observed across the three training modalities (<i>P</i> < 0.05). No significant changes were found in SD for any modality (<i>P</i> > 0.05). FuEn increased after hypertrophy (<i>P</i> < 0.01; ES = 0.07) and maximal strength training (<i>P</i> = 0.01; ES = 0.03), but not after power training (<i>P</i> = 0.99). DFA decreased following hypertrophy (<i>P</i> = 0.02; ES = 0.03) and maximal strength sessions (<i>P</i> = 0.02; ES = 0.03), with no significant change after power training (<i>P</i> = 0.78). Nonlinear analysis of motor variability through acceleration signals provides valuable insight into fatigue-induced movement adaptations, complementing traditional metrics. This cost-effective approach offers practical applications for optimizing training and rehabilitation, particularly when high-intensity assessments are impractical.<b>NEW & NOTEWOIRTHY</b> Traditional fatigue assessments often overlook subtle movement adaptations. This study applies nonlinear motor variability analysis using inertial measurement units to detect fatigue-induced changes during resistance training. By using fuzzy entropy and detrended fluctuation analysis, we demonstrate how different training modalities influence movement patterns and recovery. This approach offers a cost-effective and ecologically valid tool for monitoring fatigue, optimizing training, and reducing injury risk in both athletic and clinical populations.