Segmentation of depth-EEG seizure signals: method based on a physiological parameter and comparative study

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Stereoelectroencephalography (SEEG) provides crucial data on brain electrical activity during epileptic seizures.
• Accurate segmentation of SEEG signals is essential for understanding seizure dynamics.
• Existing methods for SEEG signal analysis have limitations in robustness and ease of use.

Purpose of the Study:

• To introduce a novel, simple, nonparametric adaptive segmentation method for SEEG signals.
• To objectively compare the proposed method against existing literature methods.
• To evaluate the robustness and accuracy of the new method for seizure analysis.

Main Methods:

• Development of a nonparametric adaptive segmentation algorithm based on a physiologically relevant parameter.
• Comparative analysis of the proposed method with three established literature methods using identical SEEG datasets.
• Objective performance evaluation on the same basis for all tested methods.

Main Results:

• The proposed method demonstrates robustness across different types of signal changes during seizures.
• The new method is simpler to implement and conduct compared to other tested approaches.
• While robust and user-friendly, the method shows slightly lower accuracy in estimating precise change instants than one comparative method.

Conclusions:

• The presented adaptive segmentation method offers a robust and practical tool for analyzing SEEG signals in epilepsy.
• The method effectively segments seizure activity without requiring parameter readjustment.
• It provides clinically relevant seizure onset and offset timings, aligning with expert annotations.