Time Dissociative Characteristics of Numerical Inductive Reasoning: Behavioral and ERP Evidence

Although some preliminary spatial localization results have been reported, the temporal characteristics of human inductive reasoning process have not been investigated. In the present study, event-related potential (ERP) was used to explore the time course of inductive reasoning process. Based on pilot studies and some other related research, we hypothesized that the process of numerical inductive reasoning is composed of number recognition, strategy formation, hypothesis generation and validation, and the above three stages are (partially) dissociable over time. A typical task of inductive reasoning, function-finding, was adopted. Induction tasks and calculation tasks were performed in the experiments, respectively. The mean reaction time of induction tasks was much longer than that of calculation tasks as expected. Statistical analysis revealed that induction showed no significant separations from calculation for the early ERP components and the slow waveforms after about 600ms, while marked dissociations appeared for the late components in the time window of about 250-600ms. It can be preliminarily concluded that, the early components before about 250ms may reflect the process of attention and number recognition, the late components may relate to strategy formation, and hypothesis generation and validation may be performed in about 600-1200ms. On the whole, the results on the behavioral data and ERP data support our hypothesis.