Machine Learning Approaches on Examination of Visual Fatigue Effect on Prefrontal Cortex Activity for LED Luminescent Sentence

Abstract

This study aims to examine the relationship between visual fatigue and prefrontal cortex (PFC) activity during a backlit text reading task. We conducted a 10 min luminescent sentence reading task in a dark room. Time-course of oxy-Hb in the PFC was measured using a 22-channel wearable near-infrared spectroscope (NIRS). Subjective fatigue was assessed using the “Jikaku-sho shirabe” questionnaire. The critical flicker-fusion frequency (CFF) and LF/HF were assessed as objective index of fatigue and autonomic nervous system. 192 trials consisted of twelve healthy young volunteers and twelve LED treatments were analyzed. In the analysis, a linear non-gaussian acyclic (LiNGAM) model for causal discovery was introduced in addition to the correlation analysis. Our results suggested that the development of visual fatigue and central fatigue is related to the deactivation conditions in the medial PFC and Left PFC.