Related Experiment Videos
Statistical mapping of scalp-recorded ictal EEG records using wavelet analysis
John J Battiston1, Terrance M Darcey, Adrian M Siegel
1Thayer School of Engineering, Dartmouth College, Hanover, New Hampshire, USA.
Epilepsia
|May 20, 2003
Summary
A new computerized method using wavelet transform (WT) objectively scores scalp-recorded seizures, improving seizure lateralization and localization compared to expert readers. Further validation is needed for full diagnostic accuracy.
Area of Science:
- Signal processing in neuroscience
- Computational analysis of electroencephalography (EEG)
Background:
- The wavelet transform (WT) is effective for analyzing time-varying signals.
- Objective scoring of scalp-recorded seizures is crucial for diagnosis and treatment.
Purpose of the Study:
- To develop and validate a computerized method for objective seizure scoring using WT.
- To improve the accuracy of seizure lateralization and localization.
Main Methods:
- A statistical scoring method using wavelet coefficients was developed.
- Spatiotemporal significance maps were created for each seizure.
- Hierarchical decision rules were established and tested on a dataset of 57 seizures from 18 patients with successful surgical outcomes.
Main Results:
- The computerized method correctly lateralized 48 seizures, outperforming expert readers (34 seizures).
- The computer correctly localized 41 seizures, compared to 31 by expert readers.
- The method demonstrated improved accuracy in estimating seizure onset lateralization and locus.
Conclusions:
- The WT-based method offers an objective approach for scoring ictal EEGs, minimizing artifact interference.
- The computerized method shows substantial improvement over expert scoring for seizure lateralization and localization.
- Further research is necessary to fully assess the diagnostic accuracy of this novel approach.