Domain-Invariant Representation Learning from EEG with Private Encoders

TL;DR

This paper introduces a domain-invariant EEG representation learning method using private encoders and MMD-based alignment, improving emotion classification and preserving privacy across datasets.

Contribution

It presents a novel multi-source architecture with private encoders and MMD alignment for domain-invariant EEG features, enhancing generalization and privacy.

Findings

Outperforms state-of-the-art in EEG emotion classification

Preserves dataset privacy through private encoders

Improves test-time generalization across domains

Abstract

Deep learning based electroencephalography (EEG) signal processing methods are known to suffer from poor test-time generalization due to the changes in data distribution. This becomes a more challenging problem when privacy-preserving representation learning is of interest such as in clinical settings. To that end, we propose a multi-source learning architecture where we extract domain-invariant representations from dataset-specific private encoders. Our model utilizes a maximum-mean-discrepancy (MMD) based domain alignment approach to impose domain-invariance for encoded representations, which outperforms state-of-the-art approaches in EEG-based emotion classification. Furthermore, representations learned in our pipeline preserve domain privacy as dataset-specific private encoding alleviates the need for conventional, centralized EEG-based deep neural network training approaches with shared parameters.