Tracking and classifying reasoning process of prefrontal cortex using portable fNIRS system

Abstract

It is known that the prefrontal cortex (PFC) plays a key role in relational reasoning. Because PFC and reasoning ability are proportional, the reasoning test tool can be used for diagnosing PFC related diseases. The limitation of RPM, the existing reasoning diagnosis method, is that difficulty adjustment is not quantitative, and the number of problems that can be created is limited. In addition, it can only be seen as a score, so it is ambiguous to distinguish whether it is training or high intelligence. Therefore, this study proposes a new cognitive reasoning test that combines a brain monitoring device which is a portable fNIRS system and CLEVR dataset which is difficulty controllable and capable of generating numerous unique questions. Based on the CLEVR dataset, a Korean version of the proposed reasoning dataset was created, classified according to type (R: Relational / NR: Non-Relational) and difficulty (Easy / Hard), and a cognitive experiment was built using proposed dataset as an input. Then, the subjects were recruited and the data were measured with fNIRS system. When brain activation of R and NR was compared, R was significantly higher than that of NR. Also, in terms of localization, the difference in R and NR was larger in the right RLPFC than in the left RLPFC, which is in agreement with previous studies that the right RLPFC is involved in the integration of visuo-spatial relations. However, the difference between R and NR was not significant in some subjects.