Related Experiment Video
Updated: Jul 17, 2026
10:23
Equipment Setup and Artifact Removal for Simultaneous Electroencephalogram and Functional Magnetic Resonance Imaging for Clinical Review in Epilepsy
Published on: June 23, 2023
Evaluation of a BSS algorithm for artifacts rejection in epileptic seizure detection
Hui Liu1, Kenneth E Hild, J B Gao
1Dept. of Electr. & Comput. Eng., Florida Univ., Gainesville, FL, USA.
Summary
Blind Source Separation (BSS) preprocessing significantly reduced false alarms in electrocorticography (ECoG) seizure detection. This data-efficient method improved artifact rejection, enhancing the reliability of seizure detection algorithms.
Area of Science:
- Neuroscience
- Signal Processing
- Biomedical Engineering
Background:
- Intracranial electroencephalography (ECoG) is crucial for epilepsy research and clinical monitoring.
- Artifacts in ECoG data can compromise the accuracy of seizure detection algorithms.
- Developing data-efficient preprocessing methods is essential for reliable ECoG analysis.
Purpose of the Study:
- To evaluate the effectiveness of a data-efficient blind source separation (BSS) algorithm for artifact rejection in ECoG.
- To assess the impact of BSS preprocessing on seizure detection performance using recurrence time statistics (T1).
Main Methods:
- Application of a data-efficient blind source separation (BSS) algorithm to preprocess ECoG data.
- Artifact correction using BSS.
- Evaluation of recurrence time statistics (T1) feature from cleaned ECoG data.
- Comparison of seizure detection performance with and without BSS preprocessing.
Main Results:
- BSS preprocessing led to a significant reduction in false alarm rates for seizure detection.
- In one dataset, the false alarm rate decreased from 0.13 to 0.08 per hour at a 96% detection rate.
- In another dataset, the false alarm rate dropped from 0.34 to 0.21 per hour at a 100% detection rate.
Conclusions:
- Data-efficient BSS preprocessing effectively reduces artifacts in ECoG data.
- BSS-enhanced ECoG analysis improves the accuracy and reliability of seizure detection algorithms.
- This approach offers a promising strategy for improving epilepsy monitoring and diagnosis.
