Awake brain surgery is performed for the purpose of preserving brain functions and removing brain tumors as possible when the tumors are in or near the eloquent area like the language area and the motor area. In this surgery, detecting functionally essential cortical areas is important process which is performed with direct electrical stimulation using an electrode by surgeons, and this process is called cortical mapping. Cortical mapping requires sophisticated techniques of surgeons because it needs a lot of experience and knowledge for selecting electrical stimulation position and current intensity. Hence it is needed to extract techniques of expert surgeons by analyzing recorded surgery videos performed by expert surgeons. Surgery videos recoded during operations, however, are not suitable for analysis due to these videos are recorded for postoperative evaluation. Then tagging specific scenes like electrical stimulation has been performed in manual with a lot of labor. Therefore, in this paper, we propose a method to automatically detect the position of electrical stimulation in surgery videos of cortical mapping. The proposed method employs the shape characteristics of an electrode, a classifier using Real AdaBoost and a tracking using mean-shift algorithm. We apply the proposed method to three different recorded surgery videos, and experimental results show that our method detects the tip of an electrode with high accuracy.