Muscle synergy adaptation with fatigue development in constant-power cycling

Neuromuscular fatigue is known to alter muscle activation, but its effects on the modular control of cycling are not fully understood. In this study, we examined how fatigue during constant-power cycling affects the structure and coordination of lower-limb muscle synergies. Twenty recreational cyclists performed a constant-power test at 70% of their individual peak power until task failure. Bilateral surface electromyography (EMG) was recorded from seven lower-limb muscles, processed, and decomposed into muscle synergies using non-negative matrix factorization at three normalized time points: Initial (0-4% of total task duration), Midpoint (48-52%), and Final (96-100%). Two indices were computed from synergy weights: the Synergy Index (SI), reflecting flexor-extensor balance at the hip, knee, and ankle, and the Synergy Coordination Index (SCI), reflecting the size and compactness of the synergy space. Across all time points, four synergies reconstructed the EMG signals with high accuracy, as confirmed by the variance accounted for (global VAF >0.95, local VAF >0.85), and SI values were consistent across time points with balanced flexor-extensor contributions at the hip and extensor dominance at the knee and ankle. In contrast, SCI showed a significant effect of time (p < 0.01, Kendall's W = 0.32), with larger values at Midpoint and Final than at Initial, but no significant difference between Midpoint and Final. Thus, synergy overlap increased during the earlier phase of fatigue development and then remained elevated near task failure. These findings suggest that during constant-power cycling, the central nervous system preserves a fixed set of motor modules and joint-level balance. Initially, at 70% of peak power (P_max), muscle coordination remained flexible and adaptable to meet the task demands. However, as fatigue developed, muscle coordination became more constrained, reflected by an early increase in SCI that was maintained near task failure.