SQNN: A Spike-wave index Quantification Neural Network with a pre-labeling algorithm for epileptiform activity...

Area of Science:

• Neuroscience
• Medical Technology
• Artificial Intelligence

Background:

• Electrical status epilepticus during slow sleep (ESES) is characterized by significant epileptiform activity on EEG during sleep.
• Quantifying the spike-wave index (SWI) is crucial for diagnosing and predicting outcomes in children affected by ESES.
• Current SWI quantification methods can be time-consuming and subjective.

Purpose of the Study:

• To develop and validate a deep learning-based automated method for accurate SWI quantification.
• To improve the efficiency and reliability of SWI measurement in pediatric epilepsy.

Main Methods:

• A novel pre-labeling algorithm (PreLA) was developed using adaptive wavelet decomposition and slow-wave discrimination for efficient EEG labeling.
• A deep learning model, the SWI Quantification Neural Network (SQNN), was constructed to classify EEG data and identify abnormal events.
• SWI is automatically calculated based on the duration of abnormalities and signal length.

Main Results:

• The PreLA demonstrated effectiveness and robustness in labeling EEG data.
• The SQNN achieved accurate and reliable SWI quantification without requiring thresholds, with an average estimation error of 3.12%.
• The automated method is 100 times faster than manual expert analysis.

Conclusions:

• Deep learning offers a promising approach for the automatic quantification of SWI in ESES.
• The PreLA facilitates straightforward EEG data labeling for ESES syndromes.
• The proposed method shows significant potential for enhancing clinical diagnosis and prognosis in pediatric epilepsy.