Abstract
Instability of neural recordings from the brain is a big problem to develop invasive brain-machine interfaces (BMIs) including our vehicular system for a rat "RatCar". Its ability to estimate actual locomotions from motor cortical activities has strictly depended on local functionality around the electrodes. It has been almost impossible to select specific neurons to record, and to keep the selection after the implant surgery. Meanwhile, neurons are known to be functionally plastic to change their own behaviour according to synaptic inputs. We have recently started a challenge to induce plastic changes of the motor cortex by applying micro-current stimuli to forcibly correlate their activities with voluntary locomotion. In this work, our methodology and preliminary results are described towards a more persistent controller for BMI systems.
