Harnessing Machine Learning and Physiological Knowledge for a Novel EMG-Based Neural-Machine Interface