Consistency of Feature Importance Algorithms for Interpretable EEG Abnormality Detection

Studies in health technology and informatics, vol. 296, pp. 33–40

Abstract

Recent advances in machine learning show great potential for automatic detection of abnormalities in electroencephalography (EEG). While simple and interpretable models combined with expert-comprehensible input features offer full control of the decision making process, these methods commonly lag behind complex deep learning and feature extraction methods in terms of performance. Here we study a feasibility of a bridging solution, where deep learning is combined with interpretable input and an algorithm computing the importance of particular EEG features in the decision process. We built a convolutional neural network with multi-channel EEG frequency bands as input and investigated four different methods for feature importance attribution: Layer-wise Relevance Propagation (LRP), DeepLIFT, Integrated Gradients (IG) and Guided GradCAM. Our analysis showed consistency between the first three methods, and deviating attributions of the fourth method, suggesting the importance of using a package of methods together to ensure the robustness of medical interpretation.

Authors 6

1. Felix Knispel Aachen Institute of Medical Informatics Medical Faculty

   RWTH Aachen University

   Affiliation as printed

   Institute of Medical Informatics, Medical Faculty, RWTH Aachen University, Aachen, Germany

2. Alexander Brenner

   University of Münster

   Affiliation as printed

   Institute of Medical Informatics, University of Münster, Münster, Germany

3. Rainer Röhrig Aachen Institute of Medical Informatics Medical Faculty

   RWTH Aachen University

   Affiliation as printed

   Institute of Medical Informatics, Medical Faculty, RWTH Aachen University, Aachen, Germany

4. Yvonne G. Weber Aachen Department of Epileptology, Neurology Medical Faculty

   RWTH Aachen University

   Affiliation as printed

   Department of Epileptology, Neurology, Medical Faculty, RWTH Aachen University, Aachen, Germany

5. Julian Varghese

   University of Münster

   Affiliation as printed

   Institute of Medical Informatics, University of Münster, Münster, Germany

6. Ekaterina Kutafina corresponding Aachen Institute of Medical Informatics Medical Faculty

   AGH University of Krakow · RWTH Aachen University

   Affiliation as printed

   Faculty of Applied Mathematics, AGH University of Science and Technology, Krakow, Poland

   Institute of Medical Informatics, Medical Faculty, RWTH Aachen University, Aachen, Germany