Medial prefrontal cortex (mPFC) is involved in several higher-order processes such as working memory, reward evaluation, and behavioral inhibition. Some previous studies have revealed that mPFC contributes to regulation of an impulse for behavioral inhibition. However, most of them were concentrated on roles of some neuro-transmitters in mPFC, and the neural mechanism of the function is still unknown. Applying brain-machine interface (BMI) methods may illuminate the function and its plasticity of that brain area. Actually, there are much evidence that BMI can be operated by neuronal activity of motor-related and perceptual areas and the BMI trainings can change functional connections and plasticity in those areas. On the other hand, few researches have done to see whether BMIs can be run with neuronal activity of mPFC. Here, we first recorded neuronal population activity of mPFC in rats when they performed a task which required them to wait 10 seconds until a tone was presented and then to poke their noses into a small hole and get reward. Then, using a BMI system based on real-time and automatic sorting with independent component analysis (Takahashi & Sakurai, 2005) and by transforming the sorted activities of neuronal population in mPFC into electrical signals to control a task-operating computer, we trained the rats to get reward by modulating firing rates of their mPFC neurons with auditory and visual stimuli to give them feedback about changes of firing rates. We are now analyzing the data to examine whether the BMI method can be operated by the neuronal activity of mPFC as other brain areas and to see if the training with the BMI enhances or weaken the function of mPFC in regulation of an impulse and learning of behavioral inhibition.