Automated real-time epileptic seizure detection in scalp EEG recordings using an algorithm based on wavelet packet...

Area of Science:

• Biomedical Engineering
• Neuroscience
• Signal Processing

Background:

• Epileptic seizures are neurological disorders characterized by abnormal brain activity.
• Accurate and timely detection of seizures is crucial for patient management and treatment.
• Scalp electroencephalography (EEG) is a primary tool for monitoring brain activity and diagnosing epilepsy.

Purpose of the Study:

• To develop and validate a novel wavelet-based algorithm for real-time detection of epileptic seizures from scalp EEG.
• To create a patient-specific measure for quantifying seizure and non-seizure states.
• To evaluate the algorithm's performance in terms of sensitivity, false detection rate, and detection delay.

Main Methods:

• A moving-window analysis decomposes scalp EEG signals using wavelet packet transform.
• A patient-specific measure quantifies the separation between seizure and non-seizure states using wavelet coefficients.
• A combined seizure index (CSI) is calculated based on rhythmicity, energy, and inter-channel consistency.
• A cumulative sum test and analysis of channel alarms generate the final seizure alarm.

Main Results:

• The algorithm achieved a high sensitivity of 90.5% in detecting epileptic seizures.
• A low false detection rate of 0.51 h(-1) was observed.
• The median detection delay for seizures was 7 seconds.
• The algorithm demonstrated the ability to lateralize the seizure focus in temporal lobe epilepsy patients.

Conclusions:

• The proposed wavelet-based algorithm offers a promising solution for real-time epileptic seizure detection using scalp EEG.
• The algorithm exhibits high accuracy, low false alarms, and rapid detection, making it suitable for clinical applications.
• The capability to lateralize seizure onset is a significant advantage for surgical planning in epilepsy treatment.