◆日 時 2018年1月19日（金）14:00~15:30
◆会 場 電気通信大学 東3号館306会議室
◆講 師 瀧山 健（東京農工大学大学院 工学研究院 先端電気電子部門・准教授）
◆司 会 庄野 逸 教授
◆題 目 運動学習・運動意思決定における予測表現（Prospective coding in human motor learning and decision making）（講演は日本語で行われます）
◆概 要 In our daily life, we make predictions in various situations, e.g., we predict tomorrow’s weather, outcomes of soccer matches, or stock price. In those predictions, our neural system receives some inputs (e.g., sky scene in predicting tomorrow’s weather) and represent future states (e.g., tomorrow’s weather). This representation of future states is referred to as prospective coding (ref. Komura et al., 2001). Here, I demonstrate that the prospective coding plays an essential role in human motor learning and motor decision making.
First, I explain about our computational model of motor learning. Diverse features of motor learning have been reported in numerous studies, but no single theoretical framework concurrently accounts for these features. We propose models for motor learning to explain these features in a unified way by extending a motor primitive framework (ref. Thoroughman & Shadmehr, 2000, Nature). Our model assumes that the recruitment pattern of motor primitives is determined by the predicted movement error of an upcoming movement (prospective error). I demonstrate that this model has a strong explanatory power to reproduce a wide variety of motor-learning-related phenomena that have been separately explained by different computational models.
Second, I explain about motor decision making in a competitive game. Although risk-seeking behavior in human motor decision making has been reported in several studies (e.g., Wu et al., 2009), those studies focused on an experiment with a single subject. In our daily life (especially in music or sports), our decision making (action selection) can be influenced by opponents in competitive games and partners in collaborative games; however, how decision making is affected by others remains unclear. Our experimental results demonstrate that subjects show risk-averse behavior at the onset of a competitive game, in contrast to risk-seeking behavior when they performed the same movement without any opponent. To understand the risk-averse behavior in a competitive game, we propose a computational model. Our computational model suggests that the risk-averse behavior is a result of optimization when our decision making is influenced by the predicted actions and results of ourselves and opponents (prospective outcome).
References:  K. Takiyama, M. Hirashima, D. Nozaki, Prospective errors determine motor learning, Nature Communications, 6, 5925: 1-12 (2015),  K. Ota, K. Takiyama, Competitive game influences risk-sensitivity in motor decision-making, Program No. 316.2. 2017 Washington, DC: Society for Neuroscience, 2017.
山田 幸生（脳科学ライフサポート研究センター 特任教授）