**6. Summary**

In this study we proposed a new detection method that uses oxy-Hb and its differential as indexes for application to the NIRS-BCI rehabilitation system, detecting brain activity from the data measured using NIRS. First, we developed a BCI system to control robot arm using NIRS, and confirmed that NIRS-BCI system can control machine and device. When a detection method with a simple threshold is employed, no stable operation was made during the tasks. This study proposes a method by which oxy-Hb is plotted on the horizontal axis and the differential of oxy-Hb on the vertical axis, and brain activity is judged from the area of the plotted trajectory. As a result, we confirmed that the proposed detection method enables highly accurate detection with little time delay compared with the conventional detection method, during both grasping tasks and imagined grasping tasks.

Next, we applied NIRS-BCI system developed for rehabilitation. We also confirmed that the proposed detection method enables highly accurate detection compared with the detection method with a simple threshold during both grasping tasks and imagined grasping tasks. These results confirm the validity of the proposed detection method for the NIRS-BC rehabilitation system. We gathered experiment data from healthy men in the present study, but in future studies, we will develop portable NIRS-BCI rehabilitation system, and conduct experiments on patients with hemiplegia to put this system into practical use.

Furthermore, the detection method proposed in this study uses the oxy-Hb level and its differential alone as indexes, so it is applicable to the brain activity of the prefrontal area while the brain is performing cognition tasks, where deoxy-Hb exhibits various fluctuation patterns (Toichi M. et. al, 2004). Therefore, BCI can be expected to be applied to measure other than the motor area (e.g., the field of entertainment).
