The present study relates a four-stage information-processing model of inductive reasoning to four brain regions. We assume that there is a fusiform gyrus region-of-interest (ROI) where a stimulus is visually recognized, a DLPFC ROI where an underlying rule is identified, a caudate ROI where a rule is applied, and a motor ROI where hand movements are programmed during inductive reasoning process. Then, an fMRI experiment was performed to articulate the roles of these four regions. The present study is a 2 (task: rule induction vs. rule application) × 2 (period length: simple vs. complex) × 2 (priming effect: prime vs. target) design. As predicted, both the fusiform gyrus ROI and the motor ROI showed no effects of task, period length, and priming effect, and respectively reflected encoding of stimuli and button-pressing response. The DLPFC ROI responded to task and period length, and was confirmed to play a crucial role in rule identification. The caudate showed no effect of task and responded to period length and priming effect, and was verified to be responsible for rule application. The exploratory analysis also demonstrated our assumptions. Thus, the main stream of information-processing in inductive reasoning process can be described by using the four ROIs. |