A BMI-based flashing light recognition system on free-moving rats

Research on flashing light stimulus dependence of activity in the visual system is a meaningful supplement to the understanding of visual system of central nervous system (CNS). Meanwhile, the novel method brain machine interfaces (BMIs) makes it possible to read out visual perception and have been applied to separate different visual perception of CNS. The combination of BMIs based visual perception system and rat robot navigation system may be applied to detect the condition of flashing light of unknown environment in the future. In this paper, we introduced a BMI-based system to evaluate the effectiveness of flashing light recognition with neural signal recorded from primary visual cortex of free-moving rats. We first analyzed features of neural signal and found significant difference between Stimulus state and Control state as well as discernible difference between high and low flashing frequencies. The results of average areas under receiver operating characteristic (ROC) curve (AUC) of cross-validation with prediction algorithm reached 0.9238 between high flashing rate stimulation and Control state, 0.8942 between low flashing rate and Control state in dark circumstance.