Transformer-Based Decomposition of Electrodermal Activity for Real-World Mental Health Applications
Charalampos Tsirmpas, Stasinos Konstantopoulos, Dimitris Andrikopoulos, Konstantina Kyriakouli, Panagiotis Fatouros · 2025
Read Transformer-Based Decomposition of Electrodermal Activity for Real-World Mental Health Applications by Charalampos Tsirmpas, Stasinos Konstantopoulos, Dimitris Andrikopoulos, Konstantina Kyriakouli, Panagiotis Fatouros (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-06378.txt.
Introduction
Charalampos Tsirmpas and co-authors split a skin signal into its fast and slow parts, in a 2025 paper. The signal, electrodermal activity, tracks sweat gland response and hints at stress and feeling. Wearable gear now records it in daily life, where noise runs high. Splitting it cleanly is key to reading what the body is doing.
The split parts are a fast phasic stream, tied to brief events, and a slow tonic drift beneath it. The authors compare older rule-based and math tools with learning-based ones. They add a new model, the Feel Transformer, which splits the signal with no labeled examples. It pools data and strips trends so the parts stay true to body behavior. Tests show it holds up well on messy real-world data, a step toward live stress tracking.
Catalog record
| CREATOR | Charalampos Tsirmpas, Stasinos Konstantopoulos, Dimitris Andrikopoulos, Konstantina Kyriakouli, Panagiotis Fatouros |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 41 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-06378
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Transformer-Based Decomposition of Electrodermal Activity for Real-World Mental Health Applications Charalampos Tsirmpas, Stasinos Konstantopoulos, Dimitris Andrikopoulos, Konstantina Kyriakouli, Panagiotis Fatouros 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. [orcid=0009-0000-9110-7716,[email protected]] *1 [orcid=0000-0002-2586-1726,[email protected]] [orcid=0000-0001-7141-7661] [orcid=0000-0002-6081-4423] [orcid=0000-0003-1685-0600,[email protected]] keywords biosignal processing ,phasic and tonic decomposition ,time series detrending ,electrodermal activity ,wearable sensors ,deep learning ,transformer models ,digital mental health ,real-world signal analysis † † copyrightyear: 2025 † † copyright: Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 † † address: Feel Therapeutics Inc., San Francisco, California, US † † address: Siglyx, Athens, Greece † † address: Johannes Kepler University, Linz, Austria † † corresp: Corresponding author. 1 Introduction Real-time physiological monitoring using wearable sensors is increasingly recognized as a valuable tool in psychiatric practice, offering continuous, objective data to support mental health assessment and intervention ( Abdullah and Choudhury 2018 ; Sano et al. 2018 ; Singh and Sharma 2024 ) Electrodermal Activity (EDA), in particular, is a measure of changes in the electrical properties of the skin, influenced by the activity of sweat glands, which are regulated by the autonomic nervous system (ANS). Specifically, EDA reflects the sympathetic branch of the ANS, associated with physiological arousal and responses to emotional or stressful stimuli ( Boucsein 2012 ) . When an individual experiences heightened emotional arousal, the sympathetic nervous system stimulates eccrine sweat glands, particularly on the palms and soles, leading to measurable fluctuations in skin conductance ( Dawson et al. 2012 ; Shields et al. 1987 ) . These fluctuations are captured as changes in electrical conductance on the skin’s surface, typically measured in microsiemens (uS). EDA is widely used in psychophysiological research due to its sensitivity to emotional, cognitive, and stress-related processes ( Rahma et al. 2022 ; Buchwald et al. 2019 ) . It has been applied extensively in fields such as psychology, neuroscience, and human-computer interaction for applications ranging from stress and anxiety detection to emotion recognition and mental health monitoring ( Greco et al. 2016 ) . More recently, measuring EDA has become increasingly accessible with the advent of wearable sensors, allowing continuous and unobtrusive monitoring of physiological arousal in everyday environments. Research-grade devices such as the Feel Monitor ( Therapeutics 2024 ) , Empatica E4 ( Empatica 2024 ) , and Shimmer3 GSR+ ( Sensing 2024 ) have been widely adopted in clinical and research settings ( Milstein and Gordon 2020 ; Udovičić et al. 2017 ; Fatouros et al. 2025 ; Andrikopoulos et al. 2024 ; Tsirmpas et al. 2023 ; Mourtakos et al. 2021 ) , while consumer-grade devices like the Fitbit and Garmin wearables are increasingly being used in mental health monitoring ( de Vries et al. 2023 ; Damme et al. 2024 ) . These platforms facilitate the detection of stress patterns and physiological correlates of emotional states, facilitating early identification of psychiatric symptoms and supporting timely, personalized interventions ( Poh et al. 2010 ) . EDA signals are typically divided into two primary components: phasic and tonic, each representing different aspects of ANS arousal. The phasic component reflects short-term fluctuations that are closely tied to discrete external or internal stimuli, such as sudden loud noises, visual stimuli, or cognitive tasks that elicit an arousal response. These rapid changes, also known as skin conductance responses (SCR), occur in reaction to specific events and are characterized by sharp, transient increases in skin conductance followed by a gradual return to baseline Boucsein 2012 . Phasic activity provides insights into an individual’s immediate, momentary responses to stimuli, making it useful for studying arousal dynamics, event-related responses, and stimulus-triggered physiological changes Boucsein 2012 . The tonic component, on the other hand, reflects overall arousal levels over longer periods. It is often referred to as the skin conductance level (SCL) and is indicative of an individual’s general physiological state or emotional baseline Boucsein 2012 . The tonic component changes gradually and is influenced by factors such as stress, attention, or sustained emotional states rather than immediate stimuli. SCL provides essential context for interpreting the individual’s broader arousal state Boucsein 2012 . Given the distinct physiological information each component provides, it becomes evident that decomposing EDA signals into their phasic and tonic components is critical for a comprehensive understanding of ANS functioning. Given the distinct physiological information each component provides, it becomes evident that decomposing EDA signals into their phasic and tonic components is critical for a comprehensive understanding of ANS functioning. This methodological distinction has direct clinical implications: accurately isolating phasic activity can help identify acute stress responses in real time, while monitoring tonic levels may reveal chronic stress or emotional dysregulation patterns. Such insights can inform timely, personalized interventions in patients with mood or anxiety disorders, making EDA decomposition not just a technical requirement but a clinically actionable step in advancing precision mental health care. By isolating these elements, researchers can better distinguish between momentary, stimulus-induced reactions and the broader, baseline arousal states that evolve over time. However, a significant challenge in evaluating EDA decomposition techniques is the lack of a universally accepted ground truth for distinguishing phasic and tonic EDA components, which directly impacts the clinical validation and broader adoption of EDA-based monitoring tools in psychiatric care. Without clear ground truth, it becomes difficult to objectively assess the performance of different decomposition algorithms, which in turn complicates their integration into clinical workflows that require reliability and interpretability. This methodological gap limits the utility of EDA signals in informing treatment personalization, and delays the implementation of real-time stress interventions based on these physiological markers. In lab conditions, researchers trigger ANS responses, which allows them to directly localize SCRs in the signal. Naturally, the duration of such studies is limited. This means that such experiments do not offer the opportunity to observe meaningful changes in the slow-changing SCL. Continuous monitoring, on the other hand, allows the study of SCL but lacks explicit information about where SCRs have occurred. Bach and Friston 2013 also discuss this lack of a definitive reference, and how it hinders the objective assessment of the accuracy and reliability of different methods. They assert, however, that phasic/tonic methods are more accurate than directly analysing the EDA signal as was standard practice previously. In this article, we investigate whether using deep learning—specifically, a Transformer architecture in a non-autoregressive setting—can be used to address the challenge of EDA decomposition without requiring detailed supervision. This