Attention State Related EEG Spectrum and Pupil Size in Vigilance Task

Abstract

Objectively recognizing Attention State is a particularly important task to ensure vigilance decrement in Vigilance Task. The aim of this study was to develop a vigilance task system using Electroencephalogram (EEG) signals to identify four kinds of Attention State including Thinking Otherthing (TO), Sleepy(S), Attention Outside (AO), and Attention to Screen (ATS). Simultaneously, Pupil dimension was recorded as reference. We approached this objective by firstly investigating the relevant EEG frequency band followed by deciding the appropriate feature extraction method. Two features were considered namely: 1. Wavelet Energy Percent, and 2. Wavelet Cosine Similarity. The results presented in this study indicated that wavelet energy percent extracted from alpha, beta, delta and theta bands seem to provide the necessary information for describing the aforementioned Attention State. Using the Wavelet Cosine Similarity (Dataset for attention state Analysis using electroencephalogram, subjective valuing and Vigilance Signals), our proposed method achieved a significant sensitivity and specificity. Deviation of pupil size also reflected the four kinds of attention state.