The Neural Substrates for Cross-Domain Multi-Level Analogical Reasoning

This project aims to investigate the neural substrates for the cross-domain (such as syntax and Chinese characters) multi-level analogical reasoning by using the functional magnetic resonance imaging technique, and therefore provides a neuroscientific account for knowledge acquisition in the modes of both “up-down” and “left-right”. Here the “up-down” mode denotes to the hierarchical structure of knowledge, in which higher knowledge can generate and be abstracted from the lower ones, while the “left-right” mode refers to the cross-domain association between the knowledge from different domains. We are confident that through scrutinizing the neurobiology of the analogical reasoning, we may provide substantial evidence for deep learning, knowledge transfer, and creativity of humans.