Hiroyuki Nakahara^1
1RIKEN Brain Science Institute, Japan

Under the theme of symposium, context-dependent preferences, I will discuss reward-guided learning and decision-making with two particular cases: learning environmental structures under reinforcement learning (RL) and others' mental processes under social decision making. I demonstrate that dopamine activity and related basal ganglia circuit incorporate context dependence for better learning and decision-making under normative understanding of RL. I show you how we also extend RL theory into the realm of social cognition. Combining human fMRI with modeling, I present how one learns to predict the minds of others and also takes account of rewards to others, so as to adjust one’s own decisions accordingly. I also plan to present results of some ongoing studies.