Novel Vibration-Exercise Instrument With Dedicated Adaptive Filtering for Electromyographic Investigation of Neuromuscular Activation

Vibration exercise (VE) has been suggested as an effective methodology to improve muscle strength and power performance. Several studies link the effects of vibration training to enhanced neuromuscular demand, typically ascribed to involuntary reflex mechanisms. However, the underlying mechanisms are still unclear, limiting the identification of the most appropriate vibration training protocols. This study concerns the realization of a new vibration exercise system for the upper limbs. Amplitude, frequency, and baseline of the vibrating force, which is generated by an electromechanical actuator, can be adjusted independently. A second order model is employed to identify the relation between the generated force and the input voltage driving the actuator. Our results show a high correlation (0.99) between the second order model fit and the measured data, ensuring accurate control on the supplied force. The level of neuromuscular demand imposed by the system on the targeted muscles can be estimated by electromyography (EMG). However, EMG measurements during VE can be severely affected by motion artifacts. An adaptive least mean square algorithm is proposed to remove motion artifacts from the measured EMG data. Preliminary validation with seven volunteers showed excellent motion artifact removal, enabling reliable evaluation of the neuromuscular activation.