• Top page
  • Timetable
  • Per session
  • Per presentation
  • How to
  • Meeting Planner

演題詳細

Oral

ブレイン・マシン/コンピュータ・インターフェース
BMI/BCI

開催日 2014/9/13
時間 17:10 - 18:10
会場 Room G(303)
Chairperson(s) 宮脇 陽一 / Yoichi Miyawaki (電気通信大学 先端領域教育研究センター / Center for Frontier Science and Engineering The University of Electro-Communications, Japan)
美馬 達哉 / Tatsuya Mima (京都大学大学院医学研究科附属脳機能総合研究センター / Human Brain Research Center, Kyoto University, Graduate School of Medicine, Japan)


Exploring the role of eye blinking in terms of attention/arousal: a method to quantify a person’s mental state via EEG, EOG, and classification analysis

  • O3-G-1-4
  • Jamilah Abdur-Rahim:1 Takeshi Ogawa:1 Motoaki Kawanabe:1 Shin Ishii:1,2 
  • 1:Dept. of Dynamic Brain Imaging, CMC, ATR, Kyoto, Japan 2:Dept. of System Science, Graduate School of Informatics, Kyoto Univ, Kyoto, Japan 

The role in which eye movements (blinking, saccades, fixation) play in regards to the type of task, task performance, attention, and arousal have been investigated in numerous modalities. Eye blinking is directly measured using an eye tracking device or electrooculography (EOG) and indirectly measured via electroencephalography (EEG). The goal of our work is to identify behavioral related eye blink attributes, confirm the existence of a connection between eye blink attributes and attention, and utilize eye blink attributes to predict a participant's engagement during a task.

Frequency ("how often the participant blinked") and speed ("how quickly a participant blinked") of eye blinks significantly decreased over the course of the experiment, thus displaying enough sensitivity to capture changes in behavior. Based on behavioral psychology literature, participants may have increased in attention (noted by the decrease in eye blink frequency) and decreased in arousal (noted by the increase in eye blink width).

In addition, changes in alpha/theta band at frontal areas correlated with changes in eye blinking; trials with few eye blinks corresponded to high theta power whereas trials with more frequent eye blinks corresponded to high alpha power. These results are in alignment with the existing EEG alpha desynchronization/theta synchronization framework to characterize attention.

Finally, a binary label for classifying attention and arousal was constructed from EOG frequency and speed attributes per trial, respectively. Artifacts were removed using ADJUST; independent component analysis based EEGLAB toolbox. The Sparse Linear Regression (SLR) classifier and Orbital Frontal Cortex (OFC)-Dorsal Lateral Pre Frontal Cortex (DLPFC) brain areas best predicted blink frequency with an accuracy of 65-82%. This may suggest that there is brain related information associated with blinking in EEG data at OFC-DLPFC.

Copyright © Neuroscience2014. All Right Reserved.