Limits...
Enhanced detection of visual-evoked potentials in brain-computer interface using genetic algorithm and cyclostationary analysis.

Gupta CN, Palaniappan R - Comput Intell Neurosci (2007)

Bottom Line: Next, bandpass or lowpass filtering is performed to reduce the EEG artifacts using these identified frequency ranges.The CS and GA methods need to be applied only to the training data; for the test data, the knowledge of the cyclic frequency bands and unmixing matrix would be sufficient for enhanced VEP detection.Hence, the framework could be used for online VEP detection.

View Article: PubMed Central - PubMed

Affiliation: Department of Computing and Electronic Systems, University of Essex, Wivenhoe Park, Colchester CO4 3SQ, UK.

ABSTRACT
We propose a novel framework to reduce background electroencephalogram (EEG) artifacts from multitrial visual-evoked potentials (VEPs) signals for use in brain-computer interface (BCI) design. An algorithm based on cyclostationary (CS) analysis is introduced to locate the suitable frequency ranges that contain the stimulus-related VEP components. CS technique does not require VEP recordings to be phase locked and exploits the intertrial similarities of the VEP components in the frequency domain. The obtained cyclic frequency spectrum enables detection of VEP frequency band. Next, bandpass or lowpass filtering is performed to reduce the EEG artifacts using these identified frequency ranges. This is followed by overlapping band EEG artifact reduction using genetic algorithm and independent component analysis (G-ICA) which uses mutual information (MI) criterion to separate EEG artifacts from VEP. The CS and GA methods need to be applied only to the training data; for the test data, the knowledge of the cyclic frequency bands and unmixing matrix would be sufficient for enhanced VEP detection. Hence, the framework could be used for online VEP detection. This framework was tested with various datasets and it showed satisfactory results with very few trials. Since the framework is general, it could be applied to the enhancement of evoked potential signals for any application.

No MeSH data available.


Trials for target colour block and its cyclicspectrum.
© Copyright Policy - open-access
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2266790&req=5

fig14: Trials for target colour block and its cyclicspectrum.

Mentions: The presented framework was also tested offline from adataset for a biometric application. Similar to the Donchin paradigm, theapplication had seven blocks of colours which were flashed to evoke P300components. Sequences were block randomised, which means, after seven flasheseach colour was flashed once, after fourteen flashes each colour was flashedtwice. Forty trials were recorded (each trial had 7 flashes of the colourblock). The subject was asked to focus on a single-colour block (say red) andalso keep a count of the number of times it flashed, which enabled monitoringthe performance of the subject. The colour blocks were flashed for 100millisecond with an interstimulus interval of 300 millisecond. EEG recordingswere carried out on a Biosemi Active Two system using 34 channels (32 on ascalp and 2 on either mastoids); however, only channel Cz was used. Data wassampled at 256 Hz with no filtering. The subject was a male aged 27 who hadexperience of using the BCIs before, with no known neurological disorders. Theperformance of the framework for target and nontarget color blocks is discussedbelow diagrammatically. It can be seen from Figure 14 that the target-trialdata is cyclic in time domain and also that the magnitude of the cyclic spectrumis much higher than that of the nontarget data as in Figure 15. The lagparameter for cyclostationary analysis was set to length of data. After somepreliminary experimentation, five-trial cyclic spectrums were again found to beoptimum for analysis as it seemed to highlight the VEP signal band appreciably.A threshold for the magnitude of the cyclic spectra was used to obtain the VEPsignal frequency band of (0–10Hz) for lowpass filtering. Based on thisobtained band from cyclostationary analysis, a lowpass filter for (targetcolour block and nontarget colour block) was used as in Section 3.1 to remove nonoverlapping EEG artifacts and the output is shown in Figures 16-17.


Enhanced detection of visual-evoked potentials in brain-computer interface using genetic algorithm and cyclostationary analysis.

Gupta CN, Palaniappan R - Comput Intell Neurosci (2007)

Trials for target colour block and its cyclicspectrum.
© Copyright Policy - open-access
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2266790&req=5

fig14: Trials for target colour block and its cyclicspectrum.
Mentions: The presented framework was also tested offline from adataset for a biometric application. Similar to the Donchin paradigm, theapplication had seven blocks of colours which were flashed to evoke P300components. Sequences were block randomised, which means, after seven flasheseach colour was flashed once, after fourteen flashes each colour was flashedtwice. Forty trials were recorded (each trial had 7 flashes of the colourblock). The subject was asked to focus on a single-colour block (say red) andalso keep a count of the number of times it flashed, which enabled monitoringthe performance of the subject. The colour blocks were flashed for 100millisecond with an interstimulus interval of 300 millisecond. EEG recordingswere carried out on a Biosemi Active Two system using 34 channels (32 on ascalp and 2 on either mastoids); however, only channel Cz was used. Data wassampled at 256 Hz with no filtering. The subject was a male aged 27 who hadexperience of using the BCIs before, with no known neurological disorders. Theperformance of the framework for target and nontarget color blocks is discussedbelow diagrammatically. It can be seen from Figure 14 that the target-trialdata is cyclic in time domain and also that the magnitude of the cyclic spectrumis much higher than that of the nontarget data as in Figure 15. The lagparameter for cyclostationary analysis was set to length of data. After somepreliminary experimentation, five-trial cyclic spectrums were again found to beoptimum for analysis as it seemed to highlight the VEP signal band appreciably.A threshold for the magnitude of the cyclic spectra was used to obtain the VEPsignal frequency band of (0–10Hz) for lowpass filtering. Based on thisobtained band from cyclostationary analysis, a lowpass filter for (targetcolour block and nontarget colour block) was used as in Section 3.1 to remove nonoverlapping EEG artifacts and the output is shown in Figures 16-17.

Bottom Line: Next, bandpass or lowpass filtering is performed to reduce the EEG artifacts using these identified frequency ranges.The CS and GA methods need to be applied only to the training data; for the test data, the knowledge of the cyclic frequency bands and unmixing matrix would be sufficient for enhanced VEP detection.Hence, the framework could be used for online VEP detection.

View Article: PubMed Central - PubMed

Affiliation: Department of Computing and Electronic Systems, University of Essex, Wivenhoe Park, Colchester CO4 3SQ, UK.

ABSTRACT
We propose a novel framework to reduce background electroencephalogram (EEG) artifacts from multitrial visual-evoked potentials (VEPs) signals for use in brain-computer interface (BCI) design. An algorithm based on cyclostationary (CS) analysis is introduced to locate the suitable frequency ranges that contain the stimulus-related VEP components. CS technique does not require VEP recordings to be phase locked and exploits the intertrial similarities of the VEP components in the frequency domain. The obtained cyclic frequency spectrum enables detection of VEP frequency band. Next, bandpass or lowpass filtering is performed to reduce the EEG artifacts using these identified frequency ranges. This is followed by overlapping band EEG artifact reduction using genetic algorithm and independent component analysis (G-ICA) which uses mutual information (MI) criterion to separate EEG artifacts from VEP. The CS and GA methods need to be applied only to the training data; for the test data, the knowledge of the cyclic frequency bands and unmixing matrix would be sufficient for enhanced VEP detection. Hence, the framework could be used for online VEP detection. This framework was tested with various datasets and it showed satisfactory results with very few trials. Since the framework is general, it could be applied to the enhancement of evoked potential signals for any application.

No MeSH data available.