Exploring Convolutional Neural Network Architectures for EEG Feature Extraction

Abstract

The main purpose of this paper is to provide information on how to create a convolutional neural network (CNN) for extracting features from EEG signals. Our task was to understand the primary aspects of creating and fine-tuning CNNs for various application scenarios. We considered the characteristics of EEG signals, coupled with an exploration of various signal processing and data preparation techniques. These techniques include noise reduction, filtering, encoding, decoding, and dimension reduction, among others. In addition, we conduct an in-depth analysis of well-known CNN architectures, categorizing them into four distinct groups: standard implementation, recurrent convolutional, decoder architecture, and combined architecture. This paper further offers a comprehensive evaluation of these architectures, covering accuracy metrics, hyperparameters, and an appendix that contains a table outlining the parameters of commonly used CNN architectures for feature extraction from EEG signals.

Identifying EEG Biomarkers of Depression with Novel Explainable Deep Learning Architectures

“…Nevertheless, the use of manually engineered EEG features inherently limits the amount of information available for use by machine learning models. Moreover, deep learning architectures perform automated feature extraction that has the potential to uncover the most salient information in raw data [34]. The combination of the potentially richer feature space in raw EEG data and the potential for deep learning methods to uncover the most salient features has occasioned the application of deep learning models to raw EEG data with the goal of maximizing the utility of learned features [34].…”

“…Moreover, deep learning architectures perform automated feature extraction that has the potential to uncover the most salient information in raw data [34]. The combination of the potentially richer feature space in raw EEG data and the potential for deep learning methods to uncover the most salient features has occasioned the application of deep learning models to raw EEG data with the goal of maximizing the utility of learned features [34]. Nevertheless, deep learning methods applied to raw EEG data tend to be less explainable than methods applied to extracted features, which is highly problematic both for extracting scientific insights (i.e., biomarkers) from deep learning models [35] and for healthcare applications [36].…”