Automatic BASED scoring on scalp EEG in children with infantile spasms using convolutional neural network

Area of Science:

• Neurology
• Medical Imaging
• Artificial Intelligence

Background:

• The Burden of Amplitudes and Epileptiform Discharges (BASED) score is a key electroencephalogram (EEG) grading scale for infantile spasms.
• Manual EEG annotation for BASED scoring is time-consuming and labor-intensive.
• Convolutional neural networks (CNNs) show promise in EEG analysis but are underutilized for BASED scoring.

Purpose of the Study:

• To develop and evaluate an automated framework for BASED scoring using deep CNNs and EEG data.
• To investigate the feasibility of CNNs in accurately classifying EEG for infantile spasms based on the BASED score.
• To establish a more efficient and accurate method for infantile spasms diagnosis and treatment monitoring.

Main Methods:

• A deep convolutional neural network (CNN) model was designed for automatic EEG classification.
• Long-term EEG data from 36 patients with infantile spasms were annotated using four levels of the BASED score (5, 4, 3, and ≤2).
• The CNN model was trained and validated on the annotated EEG data to perform 4-label classification.

Main Results:

• The proposed CNN framework achieved a high accuracy of 96.9% on the validation set.
• The model demonstrated strong performance in classifying EEG data according to the BASED score levels.
• This represents a significant advancement in automating a previously manual and time-consuming scoring process.

Conclusions:

• Deep CNNs offer a feasible and highly accurate approach for automated BASED scoring of infantile spasms.
• The developed framework is a crucial step towards an efficient, automated diagnostic tool for infantile spasms.
• This study is the first to utilize CNNs for constructing a BASED score-based model.