An epilepsy classification based on FFT and fully convolutional neural network nested LSTM.

Area of Science:

• Neurology
• Artificial Intelligence
• Biomedical Signal Processing

Background:

• Epilepsy diagnosis is crucial for patient management and quality of life.
• Electroencephalogram (EEG)-based methods are effective and non-invasive for epilepsy detection.
• Current diagnostic frameworks can be improved with advanced computational approaches.

Purpose of the Study:

• To develop and evaluate a novel epilepsy classification method.
• To leverage Fast Fourier Transform (FFT) for feature extraction from EEG data.
• To integrate Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) for enhanced classification accuracy.

Main Methods:

• EEG data was preprocessed into training and validation sets.
• Fast Fourier Transform (FFT) was employed for feature extraction.
• A hybrid model combining CNN and LSTM was utilized for epilepsy classification.

Main Results:

• The proposed method achieved high accuracy, sensitivity, and specificity, exceeding 96% in most experiments.
• FFT features were identified as the most effective for classification within the CNN-LSTM framework.
• The model demonstrated consistent performance, with accuracy rates ranging from 97.95% to 99.83%.

Conclusions:

• The developed method accurately distinguishes between epileptic and non-epileptic individuals.
• The approach effectively categorizes epilepsy types and states (ictal/interictal).
• This automated EEG analysis technique shows significant potential for real-world clinical applications in epilepsy diagnosis.