Interpreting deep learning models for epileptic seizure detection on EEG signals

While Deep Learning (DL) is often considered the state-of-the art for Artificial Intel-ligence-based medical decision support, it remains sparsely implemented in clinical practice and poorly trusted by clinicians due to insufficient interpretability of neural network models. We have approached this issue in the context of online detection of epileptic seizures by developing a DL model from EEG signals, and associating certain properties of the model behavior with the expert medical knowledge. This has conditioned the preparation of the input signals, the network architecture, and the post-processing of the output in line with the domain knowledge. Specifically, we focused the discussion on three main aspects: (1) how to aggregate the classification results on signal segments provided by the DL model into a larger time scale, at the seizure-level; (2) what are the relevant frequency patterns learned in the first convolutional layer of different models, and their relation with the delta, theta, alpha, beta and gamma frequency bands on which the visual interpretation of EEG is based; and (3) the identification of the signal waveforms with larger contribution towards the ictal class, according to the activation differences highlighted using the DeepLIFT method. Results show that the kernel size in the first layer determines the interpretability of the extracted features and the sensitivity of the trained models, even though the final performance is very similar after post-processing. Also, we found that amplitude is the main feature leading to an ictal prediction, suggesting that a larger patient population would be required to learn more complex frequency patterns. Still, our methodology was successfully able to generalize patient inter-variability for the majority of the studied population with a classification F1-score of 0.873 and detecting 90% of the seizures.

Automated summary

This study addressed the lack of interpretability of neural network models in medical decision support by developing a deep learning model from EEG signals for online detection of epileptic seizures and associating model behavior with expert medical knowledge. The focus was on aggregating classification results, identifying frequency patterns, and recognizing signal waveforms. Results showed that the kernel size in the first layer significantly impacts feature interpretability and model sensitivity.