Deep Learning for Metabolic Rate Estimation from Biosignals: A Comparative Study of Architectures and Signal Selection

2025-11-12

AI Summary

This study addresses the problem of accurate metabolic equivalent of task (MET) estimation from wearable physiological signals—namely heart rate, respiration, and acceleration. We systematically decouple and evaluate the independent effects of neural architecture design and input signal selection. We propose the first adaptive modeling framework that jointly accounts for activity intensity and inter-individual variability. Our analysis identifies minute ventilation as the optimal single-signal feature. Among architectures—including Transformer, CNN, ResNet, and attention-augmented variants—the Transformer achieves the lowest overall RMSE (0.87 W/kg) and exceptional accuracy during light activities (0.29 W/kg; NRMSE = 0.04). Furthermore, multi-signal fusion substantially enhances performance of lightweight models. All code and pre-trained models are publicly released.

Technology Category

• Machine Learning: Hardware-aware ML
• Cognitive Modeling & Cognitive Systems: Adaptive Behavior
• Search and Optimization: Metareasoning and Metaheuristics

Application Category

• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Energy expenditure estimation aims to infer human metabolic rate from physiological signals such as heart rate, respiration, or accelerometer data, and has been studied primarily with classical regression methods. The few existing deep learning approaches rarely disentangle the role of neural architecture from that of signal choice. In this work, we systematically evaluate both aspects. We compare classical baselines with newer neural architectures across single signals, signal pairs, and grouped sensor inputs for diverse physical activities. Our results show that minute ventilation is the most predictive individual signal, with a transformer model achieving the lowest root mean square error (RMSE) of 0.87 W/kg across all activities. Paired and grouped signals, such as those from the Hexoskin smart shirt (five signals), offer good alternatives for faster models like CNN and ResNet with attention. Per-activity evaluation revealed mixed outcomes: notably better results in low-intensity activities (RMSE down to 0.29 W/kg; NRMSE = 0.04), while higher-intensity tasks showed larger RMSE but more comparable normalized errors. Finally, subject-level analysis highlights strong inter-individual variability, motivating the need for adaptive modeling strategies. Our code and models will be publicly available at https://github.com/Sarvibabakhani/deeplearning-biosignals-ee .

Problem

• Systematically evaluates neural architectures and signal selection for metabolic rate estimation
• Compares deep learning models across single/paired/grouped biosignals during physical activities
• Analyzes performance variations across activity intensities and individual subject differences

Innovation

• Transformer model achieves best accuracy with minute ventilation
• CNN and ResNet with attention work well with grouped signals
• Study reveals need for adaptive modeling across individuals