A Convolutional Autoencoder-based Explainable Clustering Approach for Resting-State EEG Analysis

Abstract

Machine learning methods have frequently been applied to electroencephalography (EEG) data. However, while supervised EEG classification is well-developed, relatively few studies have clustered EEG, which is problematic given the potential for clustering EEG to identify novel subtypes or patterns of dynamics that could improve our understanding of neuropsychiatric disorders. There are established methods for clustering EEG using manually extracted features that reduce the richness of the feature space for clustering, but only a couple studies have sought to use deep learning-based approaches with automated feature learning to cluster EEG. Those studies involve separately training an autoencoder and then performing clustering on the extracted features, and the separation of those steps can lead to poor quality clustering. In this study, we propose an explainable convolutional autoencoder-based approach that combines model training with clustering to yield high quality clusters. We apply the approach within the context of schizophrenia (SZ), identifying 8 EEG states characterized by varying levels of δ activity. We also find that individuals who spend more time outside of the dominant state tend to have increased negative symptom severity. Our approach represents a significant step forward for clustering resting-state EEG data and has the potential to lead to novel findings across a variety of neurological and neuropsychological disorders in future years.

A Framework for Systematically Evaluating the Representations Learned by A Deep Learning Classifier from Raw Multi-Channel Electroencephalogram Data

“…(3) We assigned coefficients corresponding to frequency band 𝑓 in channel 𝑐 to values of zero. We could have randomly permuted coefficient values or reassigned them from a Gaussian distribution [10], [13], [15], [68]. However, doing so would have required repeatedly perturbing each channel and frequency band dozens of times, which would have been computationally prohibitive given subsequent steps.…”

“…Multiple modalities have been used to study neurological and neuropsychiatric disorders. A few of these modalities include EEG [1]- [3], [9], [10], [19]- [27], magnetoencephalography (MEG) [28]- [30], and functional magnetic resonance imaging (fMRI) [30]- [37]. Each modality offers both advantages and disadvantages.…”

“…Additionally, relative to MEG, EEG devices can be performed much more cheaply and are more widespread, making it better suited for deployment in a clinical setting. Within the domain of EEG analysis, both task [8], [11], [38] and resting-state [5], [6], [9], [10], [25], [39], [40] analyses are commonly performed. However, most brain activity spontaneously occurs (i.e., reflects brain networks that are unmodulated by any task), so resting-state activity better reflects the activity that is common to an individual with a disorder.…”

Explainable Fuzzy Clustering Framework Reveals Divergent Default Mode Network Connectivity Dynamics in Schizophrenia

“…These include electroencephalography (EEG), magnetoencephalography (MEG), and fMRI. All three modalities – EEG [16]–[18], MEG [19], and fMRI [13], [20]–[24] - have been used extensively for SZ analysis. EEG and MEG capture much higher resolution temporal information.…”