A high-performance brain-machine interface (BMI) using image information

Sensory feedback is very important for movement control. However the feedback information has not been considered in the brain-machine interface (BMI) studies. Here, we propose a novel prediction method, the feedback-prediction algorithm (FPA), to generate feedback information from the positions of objects and using the feedback to predict movements. The FPA modifies the predicted direction toward the target and modulates the magnitude of the predicted vector to easily reach the target by combining feedback information. We demonstrated that combining feedback information for movement prediction considerably improves prediction accuracy. The proposed method, FPA, will promote the development of a practical BMI system.