RAPIDLY GRASPING LEARNERS’ MATHEMATICAL PERFORMANCE FROM BRAIN ACTIVITY Kodai Suzuki, Satoshi Yazawa, Reiko Matsunaka, Toru Iwasawa and Kazuo Hiraki Graduate School of Art and Science, University of Tokyo 3-8-1, Komaba, Meguro-ku, Tokyo, Japan ABSTRACT For most people, it takes a lot of time to acquire mathematical skills. Mathematical problems are complex because they consist of many steps. Therefore, it is not easy to grasp the learners’ performance. It is ideal to know the level of learners in a reasonably short time. In this experiment, we aimed to grasp the learners’ performance from short time solving time estimation for the mathematical problems. Participants performed the task of estimating the solving time of mathematical problems. Time frequency analysis was conducted from Electroencephalogram (EEG) recordings. As a result, people with well mathematical performance had higher theta activity than those with poor performance. In conclusion, the present research suggests that brain activity measurement during solving time estimation can be effective in grasping the learners’ level in a short time. KEYWORDS Mathematical Education, Passive BCI, Problem Solving, Educational Technology, EEG, Time Frequency Analysis 1. INTRODUCTION Many people spend so long time to improve their mathematical skills. Because the cognitive processes involved in mathematical problem solving are very complex, it is very difficult to find an efficient learning method. It is important to be able to diagnose learners’ performance rapidly for the efficient improvement of mathematical skills. Passive brain computer interface (passive BCI), which captures the mental state of healthy users (Zander & Kothe, 2011), is a technology of interest and could be effective in achieving this end. In this experiment, in order to measure brain activity, we used Electroencephalogram (EEG). EEG has superb temporal resolution (Martinez et al., 2004). Using EEG, we can examine brain activity by isolating the epochs during solving time estimation for the mathematical problem from other brain activities. Theta activity has been linked to long-term memory, leading to faster problem solving (Lin et al., 2012). Resolving the problem from the fact retrieval showed higher theta activity than based on insights (Grabner & De Smedt, 2011). However, previous studies have dealt with simple problems such as arithmetic operations, so there was a challenge that multiple processes such as making plans and implementation could not be separated. In the present research, our goal is to grasp the learners’ performance in a short time when the learners are performing meta-cognition, focusing on understanding and planning. In the present research, participants performed a sort of meta-cognition task to estimate their solving time. We examined whether brain activity, especially theta power, was related to learning performance. 2. METHOD 2.1 Participants Fourteen high school students participated in this experiment (range: 15-16, 12 males and 2 females). Every participant is paid 2000 yen equally. We have obtained the consent of the participants and their parents. We International Conferences Interfaces and Human Computer Interaction 2020; and Game and Entertainment Technologies 2020 231