Practical brain-machine interface system

Abstract

Over the last several decades, there have been lots of BMI studies. However, it is still difficult to use BMI system in real life. Here, we introduce our three BMI studies to overcome these problems. First, we predicted continuous movement trajectory from non-invasive MEG signals. Second, we proposed a new BMI prediction model to increase the prediction accuracy using external stereo camera. Finally, we showed that modes of the BMI system can be changed according to the user's brain state. Based on our results, we expect that practical and high accuracy BMI system will be possible by combining brain states and feedback information.