ISSN Print 2500–1094
ISSN Online 2542–1204
Bulletin of RSMU
BIOMEDICAL JOURNAL OF PIROGOV UNIVERSITY (MOSCOW, RUSSIA)
METHOD
Improving eye-brain-computer interface performance by using EEG frequency components
About authors
1 Department of Neurocognitive Technologies, Kurchatov Complex of NBICS Technologies,National Research Centre Kurchatov Institute, Moscow, Russia
2 Kurchatov Complex of NBICS Technologies,National Research Centre Kurchatov Institute, Moscow, Russia
3 Faculty of Cybernetics and Information Security,National Research Nuclear University MEPhI, Moscow, Russia
Correspondence should be addressed: Sergey Shishkin
pl. Akademika Kurchatova, d. 1, Moscow, Russia, 123182; ur.liam@nikghsihsgres
About paper
Funding: this work was partially supported by the Russian Science Foundation, grant no. 14-28-00234 (acquisition and preprocessing of experimental data), and the Russian Foundation for Basic Research, grant no. 15-29-01344 (evaluation of wavelet features significance for classification).
Received: 2016-04-08
Accepted: 2016-04-15
Published online: 2017-01-05
|
Fig. 1. Dependence of classification accuracy (AUC) for gaze fixations (control and non-control) on the method used for feature extraction from EEG recorded during gaze fixation
Legend: А — amplitude features only, В — wavelet features only, АВ — combined (amplitude-wavelet) set of features; Z1 — normalization type before PCA; Z2 — normalization type after PCA; features: normalization of separate features; trials — normalization of features within a single trial. Vertical lines represent 95 % confidence intervals.
Fig. 2. ROC curves (Receiver Operating Characteristic curves) for all subjects when using the amplitude-wavelet feature set, feature normalization before PCA, trial normalization after PCA (feature extraction method that allowed for the highest group averaged AUC value). Red line shows random classification, grey vertical line provides an example of strict requirements to the specificity of classifier (false positive rate = 0.1)
Effect of feature extraction methods on classification accuracy (AUC)
Note. Using multivariate analysis of variance (MANOVA), AUC dependence on normalization before PCA (Z1), normalization after PCA (Z2), feature set type (amplitude, wavelet, amplitude-wavelet) and their interaction (represented by ×) were analyzed. Statistically significant effect is shown in bold (p <0.05).
