Analyzing Neural Correlations Between Numerical Induction and Letter Induction Based on Data-Brain Driven Integration Evidence

Abstract

Numerical induction and letter induction are two kinds of important subtypes of induction. Analyzing their neural correlations is very important for understanding the common mechanism of induction. Previous comparative studies on number cognition and letter comprehension were mainly based on a group of comparative experiment designs and their neuroimaging data. However, because of the many-to-many structure-function relationships, it is difficult to understand neural correlations between number cognition and letter comprehension, especially in complex cognitive functions, such as induction, only based on single-task or few-task neuroimaging data within an experimental lab. This paper proposes a systematic approach to analyze the similarity and disimilarity of neural pattern between numerical and letter induction by using Data-Brain driven integration evidence. Under the four dimensions of Data-Brain, a group of internal and external evidence is collected. A three stages multi-task analytical method is proposed to understand the similarity and disimilarity of neural pattern between numerical and letter induction, by combining meta-analysis and representational similarity. Results show that more activation specific for inductive reasoning is left MFG and IFG. And number inductive reasoning and letter inductive reasoning have high neural pattern similarity in the IFG and MFG, and a significant main effect of inductive reasoning is in the left MFG. Other hand, the method can supplementary proof some results, it has important implications for understand the brain mechanism of information processing.