Transformer-Based Decomposition of Electrodermal Activity for Real-World Mental Health Applications

TL;DR

This paper introduces the Feel Transformer, a novel deep learning model for decomposing Electrodermal Activity signals into meaningful components, demonstrating improved robustness and potential for real-world mental health monitoring.

Contribution

The study presents the first Transformer-based model for EDA decomposition, specifically designed for in-the-wild data, with mechanisms to ensure physiologically meaningful separation without supervision.

Findings

Feel Transformer outperforms traditional methods in noisy real-world data

The model maintains a balance between feature fidelity and robustness

Potential applications include stress prediction and mental health interventions

Abstract

Decomposing Electrodermal Activity (EDA) into phasic (short-term, stimulus-linked responses) and tonic (longer-term baseline) components is essential for extracting meaningful emotional and physiological biomarkers. This study presents a comparative analysis of knowledge-driven, statistical, and deep learning-based methods for EDA signal decomposition, with a focus on in-the-wild data collected from wearable devices. In particular, the authors introduce the Feel Transformer, a novel Transformer-based model adapted from the Autoformer architecture, designed to separate phasic and tonic components without explicit supervision. The model leverages pooling and trend-removal mechanisms to enforce physiologically meaningful decompositions. Comparative experiments against methods such as Ledalab, cvxEDA, and conventional detrending show that the Feel Transformer achieves a balance between feature fidelity (SCR frequency, amplitude, and tonic slope) and robustness to noisy, real-world data. The model demonstrates potential for real-time biosignal analysis and future applications in stress prediction, digital mental health interventions, and physiological forecasting.