Study on the feature extraction of combination of mVEP and SSVEP evoked by visual stimulus
10.3760/cma.j.issn.1673-4181.2015.05.003
- VernacularTitle:视觉诱发的联合mVEP和SSVEP特征成分提取研究
- Author:
Wei YUE
;
Jinxiu SUN
;
Suogang WANG
- Publication Type:Journal Article
- Keywords:
Motion-onset visual evoked potential;
Steady-state visual evoked potentials;
Combination;
Feature
- From:
International Journal of Biomedical Engineering
2015;38(5):266-270,后插4
- CountryChina
- Language:Chinese
-
Abstract:
Objective To propose A brain computer interface paradigm based on the combination of the motion-onset visual evoked potential(mVEP) and the steady state visual evoked potential(SSVEP).Methods By designing a 3 ×3 character spelling matrix,a vertical white bar in the column of matrix which flicks at a pre-set frequency induced the corresponding SSVEP.The vertical white bar also randomly moved horizontally so as to induce mVEP.Then the two types of features were extracted by time frequency analysis.Finally the support vector machine was applied to compare the target character identification rate between the proposed paradigm and the single mVEP paradigm.Results The target character identification accuracy of subject S1 and S6 was improved by about 2％ comparing the proposed paradigm to the single paradigm.Other subjects achieved the improvement of 6％ for the same performance comparison.The averaged identification accuracy of the proposed paradigm could reach 91.4％ if the same stimulus repeated for more than 3 times,while the accuracy of single paradigm achieved 85.6％.Conclusions The proposed brain computer interface paradigm could effectively induce many kinds of brain feature signals.The identification accuracy by the proposed paradigm is higher than that by the single paradigm for various numbers of repeated trials.The proposed paradigm of combined visual stimulus merges the motion induction and the flash frequency modulation together and hence reduces the stimulation time and increases the efficiency of the feature extraction.
