FEEL: Quantifying Heterogeneity in Physiological Signals for Generalizable Emotion Recognition

TL;DR

FEEL is a comprehensive benchmarking study evaluating emotion recognition models across diverse physiological datasets, highlighting the importance of domain knowledge and heterogeneity for model generalization.

Contribution

This study introduces FEEL, the first large-scale benchmark for emotion recognition using physiological signals across multiple datasets and modeling approaches.

Findings

Contrastive signal-language pretraining models achieve highest F1 scores.

Handcrafted features outperform raw signal models in low-resource settings.

Models trained on real-life data generalize well to lab and wearable device datasets.

Abstract

Emotion recognition from physiological signals has substantial potential for applications in mental health and emotion-aware systems. However, the lack of standardized, large-scale evaluations across heterogeneous datasets limits progress and model generalization. We introduce FEEL, the first large-scale benchmarking study of emotion recognition using electrodermal activity (EDA) and photoplethysmography (PPG) signals across 19 publicly available datasets. We evaluate 16 architectures spanning traditional machine learning, deep learning, and self-supervised pretraining approaches, structured into four representative modeling paradigms. Our study includes both within-dataset and cross-dataset evaluations, analyzing generalization across variations in experimental settings, device types, and labeling strategies. Our results showed that fine-tuned contrastive signal-language pretraining (CLSP) models (71/114) achieve the highest F1 across arousal and valence classification tasks, while simpler models like Random Forests, LDA, and MLP remain competitive (36/114). Models leveraging handcrafted features (107/114) consistently outperform those trained on raw signal segments, underscoring the value of domain knowledge in low-resource, noisy settings. Further cross-dataset analyses reveal that models trained on real-life setting data generalize well to lab (F1 = 0.79) and constraint-based settings (F1 = 0.78). Similarly, models trained on expert-annotated data transfer effectively to stimulus-labeled (F1 = 0.72) and self-reported datasets (F1 = 0.76). Moreover, models trained on lab-based devices also demonstrated high transferability to both custom wearable devices (F1 = 0.81) and the Empatica E4 (F1 = 0.73), underscoring the influence of heterogeneity. More information about FEEL can be found on our website https://alchemy18.github.io/FEEL_Benchmark/.