Continuous decoding of intended movements with a hybrid kinetic and kinematic brain machine interface

Abstract

Although most brain-machine interface (BMI) studies have focused on decoding kinematic parameters of motion, it is known that motor cortical activity also correlates with kinetic signals, including hand force and joint torque. In this experiment, a monkey used a cortically-controlled BMI to move a visual cursor and hit a sequence of randomly placed targets. By varying the contributions of separate kinetic and kinematic decoders to the movement of a virtual arm, we evaluated the hypothesis that a BMI incorporating both signals (Hybrid BMI) would outperform a BMI decoding kinematic information alone (Position BMI). We show that the trajectories generated by the Hybrid BMI during real-time decoding were straighter and smoother than those of the Position BMI. These results may have important implications for BMI applications that require controlling devices with inherent, physical dynamics or applying forces to the environment.