The Neural Mechanism Of Numeral Induction

As an important factor of human's intelligence, inductive reasoning is a major way for people not only to adapt to the environment, learn and acquire knowledge, but also to science research. By definition, inductive reasoning could be regard as a process which through a number of specific stimuli (the premise) observed to produce a general rule (the conclusion). People have researched this issue for amost 30 years. A typical cognitive induction model could be described as a process of hypothesis generation and selection. Specifically, people decide the relevant information to depict the current situation base on researching a lot of information. Early behavior researches focused on inductive power and the characteristics of development of children's inductive reasoning in order to establish a domain-general structural model of induction. The neural mechanism of inductive reasoning was reviewed based on the research of neuropsychology and brain imaging. The neuropsychological research have shown that left lateral frontal cortex was greater activity during induction, and the process of relational integration in induction was associated with bilateral rostrolateral PFC (RLPFC,BA10) and right dorsolateral prefrontal cortex (BA9/46).In order to study the dynamics process of induction, it is necessary to use the ERP method because of its advantage of time resolution. Although some researches on hypothesis have revealed the P300 may be a Sign of hypothesis generation, there was no real hypothesis generated in these tasks. In such task, people adjust their hypothesis according to feedback controlled by the experiment. The research on temopral characteristics of induction mainly uses these tasks including syllogism semantic induction task, category induction task and function founding.ms). The prior one more referred to the induction power but not to how the hypothesis generated and verified. In addition, the sentential task may be affected by different background knowledge. In the category induction task and function founding task, people need abstracted the common attributes among the stimuli and generated the rule. Different from these tasks above, the number series complete task more referred to relations abstraction to get a general rule. This task aim to create a solution space by evaluating the stimuli and inferring the rule base on the prior item. Therefore, this research purpose to explore the temporal characteristics of induction reasoning by number series complete task.There are two experiments in this study. Experiment I is aim to reveal the difference between the numeral induction task and the calculation task on ERPs. In this experiment, the rule period of these tasks is one and subjects could infer the rule by 3 numbers. The results showed that all the premises numbers induced N1, P2, N2, P3 the frontal LNC and the posterior LPC. It descripts a time course of numerical induction including number recognition, relation detection, hypothesis generation and validation. P2,P3 and LPC induced by Num 3 is Significantly larger than the ones induced by Numl and Num 2, while N2 is much smaller than the prior two numbers. It suggested that Num3 paid a key role in induction task and the difference may refer to the process of hypothesis validation. Compared to the calculation task, Numl in induction task induced smaller N1 and larger LPC; Num 2 induced larger P2 and LPC, while Num3 induced larger P3 and LPC. Such results revealed the difference on ERPs between the numerical induction task and calculation tasks, which refer to the hypothesis generation and validation. In experimentâ…¡, there are two kinds of rule period (period one vs. period two) task. It is used to explore the electrophysiological relevance of hypothesis modification and validation, and the premise quantity effect. The results showed that Num3 in the induction task with period two induced larger N2 and P3 than the ones induced by Num3 of induction task with period one. It suggested that subjects may deal with more conflict information and P3 may be a Sig.n of the hypothesis modification. Moreover, the results also showed a premise number effect reflected by the difference of N1 and N2 among Num3, Num4 and Num5 in the induction task with period one. The different N1 may reflect the changes of attention distribution after subjects validate the hypothesis, and the N2 probably reflect metacognitive monitoring in the process of hypothesis validation.The two experiment results have suggested that there are main three stages of induction as follow: the first one is the preparation of induction including attention, perception register and object recognition, which refer to mostly all cognitive tasks; then is the relevant information analysis according to task demand. In present study the relevant information is the algorithm between numbers. Finally, it is the process of hypothesis generation and validation in which there would associated with the modifying and testing hypothesis process by more premises presented. There will be a dynamic process involved the metacognitive monitoring until the final hypothesis is confirmed. Moreover, working memory and central control system may pay an important role in this stage.