Online Artifact Removal for Brain-Computer Interfaces Using Support Vector Machines and Blind Source Separation
2007

Online Artifact Removal for Brain-Computer Interfaces

Sample size: 4 publication 10 minutes Evidence: moderate

Author Information

Author(s): Sebastian Halder, Michael Bensch, Jürgen Mellinger, Martin Bogdan, Andreas Kübler, Niels Birbaumer, Wolfgang Rosenstiel

Primary Institution: University of Tübingen

Hypothesis

Can a combination of blind source separation and support vector machines effectively remove artifacts from EEG signals in brain-computer interfaces?

Conclusion

The proposed artifact removal method effectively isolates and removes artifacts from EEG signals, improving the performance of brain-computer interfaces.

Supporting Evidence

  • The artifact filter was tested on three BCI datasets, showing improved performance in cases of uncorrelated muscle and ocular activity.
  • Classification rates for SVMs were above 90%, indicating high accuracy in identifying artifacts.
  • AMUSE algorithm showed significantly lower runtime compared to other methods, making it suitable for online applications.

Takeaway

This study shows a way to clean up brain signals so that people can use their thoughts to control computers better, even when there are distractions like eye blinks or muscle movements.

Methodology

The study used blind source separation and support vector machines to classify and remove artifacts from EEG data recorded from participants.

Limitations

The method may not completely remove muscle artifacts, which can still interfere with BCI performance.

Participant Demographics

Participants included healthy individuals and those with paralysis.

Digital Object Identifier (DOI)

10.1155/2007/82069

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication