Control of a Vehicle-formed BMI system for Rats by Neural Signals Recorded in the Motor Cortex

We have developed a brain-machine interface (BMI) that we call the RatCar. The system consists of a vehicle body to support a rat, a locomotion estimator based on neural signals recorded in the rat´s motor cortex, and an optical behavior recorder. We chronically implanted wire electrodes in the motor cortex of a rat´s brain to record neural signals while the position and the body direction of the rat were simultaneously calculated. Then we applied a linear model to estimate the locomotion speed and directional changes of the rat using a weighted linear model for the neural firing rates. Finally, the vehicle was controlled according to the estimated results. This paper describes the system design, an algorithm for neural signal processing, experimental setups, and results of the vehicle control with estimated locomotion states. Although the recorded neural signals contained a large number of errors due to a lack of information and motion artifacts, rough control of the vehicle on a plain floor was achieved.