Muscle synergy extraction during gait and cycling: effect of EMG detection at different locations along biarticular muscles

Muscle synergies represent a model about how the Central Nervous System (CNS) recruit different muscles, coordinating their activation during a functional task. Muscle strategies are usually estimated from surface EMG signals, recorded with bipolar technique, from the group of muscles involved in the movement under study; usually one bipolar detection system per muscle is used. However, studies in literature have suggested that some lower limb muscle (Rectus Femoris, Gastrocnemius Medialis, Vastus Medialis, Soleus) are not uniformly activated during a motor task, but separated regions are selectively recruited in different phases of the movement. This study aims to evaluate if muscle synergies extracted during gait and cycling depend on the electrode positioning on the muscle. High-density EMG technique has been used to simulate different electrode positioning. Eight healthy male subjects were recruited for this study. EMG signals were detected from Rectus Femoris, Gastrocnemius Medialis, Vastus Medialis, Vastus Lateralis, Soleus, Biceps Femoris, Gluteus Medius and Tibialis Anterior during i) over ground gait at normal and fast speed and ii) cycling with and without a pedal-straps to stick the foot. HD-sEMG was recorded from Rectus Femoris and Gastrocnemius Medialis muscles to distinguish proximal and distal muscle regions. Movement kinematics was recorded using a 12-camera optoelectronic system. The HD-sEMG signals recorded from Rectus Femoris and Gastrocnemius Medialis were used to simulate three bipolar detection systems for each muscle: one matching the SENIAM recommendation, one proximal and one distal. Muscle synergies were extracted considering all possible combinations of real and simulated signals. The extraction was performed with Matlab 2018b NNMF routine (Non-Negative Matrix Factorization). In general, the extracted synergies were repeatable and consistent with literature. During the gait task Rectus Femoris was mainly represented in the second synergy, that shows decreasing coefficient values moving from the proximal to the distal electrode positioning between the 60 and the 80% of the gait cycle, corresponding to the middle of the swing phase; this result seems to confirm the localized regional activation of Rectus Femoris for hip flexion and knee extension. Gastrocnemius Medialis is mainly represented in the third synergy whose coefficients show an anticipated and more gradual recruitment of the distal portion of the muscle. During the cycling task, significant differences in the coefficients of the third Synergy, mainly representing the Gastrocnemius Medialis muscle were observed. Even if the number, the coefficients and the weigths of the extracted synergies do not change significantly with different electrode positioning, some differences were observed in the coefficients for the synergies were the Gastrocnemius Medialis and Rectus Femoris are mainly represented.