An improved BECT spike detection method with functional brain network features based on PLV.

Area of Science:

• Neuroscience
• Medical Imaging
• Computational Biology

Background:

• Benign childhood epilepsy with centro-temporal spikes (BECT) is characterized by specific EEG patterns.
• Accurate spike detection is crucial for clinical diagnosis of BECT.
• Traditional template matching methods face challenges due to individual variability.

Purpose of the Study:

• To propose an advanced spike detection method for BECT using functional brain networks and deep learning.
• To improve the accuracy and reliability of spike identification in pediatric epilepsy EEGs.

Main Methods:

• A novel method combining functional brain networks based on phase locking value (FBN-PLV) and deep learning is presented.
• Candidate spikes are identified using template matching and montage peak-to-peak analysis.
• Artificial neural networks (ANN) integrate time-domain spike features and FBN-PLV structural features for classification.

Main Results:

• The proposed FBN-PLV and ANN method demonstrated high performance on BECT EEG data.
• Achieved an accuracy (AC) of 97.6%, sensitivity (SE) of 98.3%, and specificity (SP) of 96.8% in detecting spikes.
• Validation was performed on EEG datasets from four BECT cases.

Conclusions:

• The FBN-PLV and deep learning approach offers a promising, accurate method for BECT spike detection.
• This technique can aid in the clinical diagnosis and management of pediatric epilepsy.
• The study highlights the potential of integrating network analysis with machine learning for EEG interpretation.