脳血流量の時間変動と局在特性を考慮した特徴量とノイズ除去手法の提案

－色彩環境下での知的作業課題判別－

抄録

This research proposed a model of cerebral blood flow responses due to work in a work environment. Based on the model, features were proposed to capture the temporal changes and localization of cerebral blood flows, focusing on the periodicity of the task. A method to remove the effect of environmental noise in the features was also proposed. To verify the usefulness of the proposed method, we conducted an intellectual task discrimination test in an office environment under a color environment. In the verification experiment, we measured the cerebral blood flow of participants using NIRS when they performed three types of intellectual tasks in four different color environments. We used SVM to discriminate the data of cerebral blood flow using the features calculated for each arbitrary segment. The maximum correct response rate was 0.717 after comparing the segmented interval, type of hemoglobin, and type of kernel function.