Current challenges in the development of a cortical brain-machine interface

Recent advances in cortically controlled brain-machine interfaces (BMIs) have demonstrated that goal-directed movement of external devices is possible in real-time using multi-electrode recordings from cortex. A number of challenges are currently being confronted to further advance BMI research to the next level. These include choosing the optimal decoding algorithm for the type of control to be performed, localizing the optimal cortical site for reliable control, and focusing on the most suitable electrophysiological signal for practical use in a BMI. We present results that attempt to address these challenges based on multi-electrode recording from multiple motor cortical areas in behaving monkeys.