Classification of EEG-based attention for brain computer interface

Abstract

This paper reports on an EEG-based brain computer interface (BCI) development, which recognizes four levels of attention. In order to measure the levels of subject's attention, many types of biological signals can be recorded such as electroencephalogram (EEG), electrocardiogram(ECG), electrooculo-gram(EOG), and electromyogram (EMG). Among these methods EEG generally is used as the most effective one for assessing subject's cognitive functions. Recognizing attention levels can be used in a wide variety of applications such as students' attention level, clinical application in detecting Attention Deficit Hyperactivity Disorder (ADHD), and driver fatigue detecting system. Highlighting the four levels of attention is proposed here, in which the acquired signals from subjects are modeled in a designed task so that attention levels vary from non-attention conditions (closed eyes and reading task) to full attention conditions (mathematics task and vigilance). While the previous studies only worked on two levels of attention (low and high levels), the novelty of proposed method is in using four levels of attention. After proving the effectiveness of proposed system, the results reveal appropriate signal processing and classification methods for discriminating the levels of attention which can be used for boosting the BCI performance.