학술논문
Single-trial classification of EEG in a visual object task using ICA and machine learning.
Document Type
Article
Source
Subject
*ELECTROENCEPHALOGRAPHY
*VISUAL perception
*MACHINE learning
*COGNITION
*TASK performance
*NEUROPSYCHOLOGY
*
*
*
*
*
Language
ISSN
0165-0270
Abstract
Highlights: [•] We consider machine learning in assessing information in different EEG data. [•] We train SVM classifiers using EEG data from a visual object stimuli task. [•] New data can be correctly labelled with ‘object present’ state well above chance. [•] Using one channel of ICA data as input increases classification accuracy to 87%. [•] We discuss how this method and IC sources might help studies of visual cognition. [Copyright &y& Elsevier]