Zero-Shot Heart Rate Variability Forecasting from Consumer Wearables Using Time Series Foundation Models

2026-07-22

AI Summary

This study addresses the challenge of accurately forecasting short-term heart rate variability (HRV) from fragmented, artifact-contaminated signals collected by consumer-grade wearable devices. It pioneers the application of time series foundation models (TSFMs) to zero-shot HRV prediction, demonstrating their ability to effectively model real-world data without task-specific fine-tuning. To preserve physiological dynamics during preprocessing, the authors propose a variability-preserving imputation method that augments linear interpolation with locally adaptive stochastic noise. Experimental results show that TSFMs—particularly Chronos and TimesFM—significantly outperform conventional baselines at context lengths of 32 and 64 time steps, achieving mean MASE scores between 0.81 and 0.87. These findings underscore the clinical potential of TSFMs for early detection of autonomic dysfunction and cardiac events.

Abstract

Short-term Heart Rate Variability (HRV) forecasting could provide clinicians with actionable lead time for detecting autonomic dysfunction and adverse cardiac events. Consumer wearable devices generate fragmented, artifact-rich HRV signals that challenge conventional forecasting approaches. In this study, we evaluated the forecasting ability of three Time Series Foundation Models (TSFMs), TimesFM, Chronos, and MOIRAI, against traditional baselines (Mean, Exponential Smoothing, and Exponentially Weighted Moving Average) on real-world wearable data collected from 49 healthy individuals. To address data fragmentation, we introduce a variability-preserving imputation method that augments linear interpolation with locally adaptive stochastic noise, retaining physiological dynamics essential for accurate forecasting. The results show that TSFMs outperformed all baselines without fine-tuning, achieving average Mean Absolute Scaled Error (MASE) between 0.81 and 0.87 across TSFMs and both context lengths (32 and 64 time steps), with Chronos and TimesFM as the top models, though MOIRAI showed limited gains over baselines. With up to a 2-hour forecast horizon, the results establish a baseline for TSFMs' performance on a real-world dataset, highlighting domain-specific fine-tuning as a promising direction for clinical deployment.

Problem

• Heart Rate Variability
• Zero-Shot Forecasting
• Consumer Wearables
• Time Series Foundation Models
• Autonomic Dysfunction

Innovation

• Time Series Foundation Models
• Zero-Shot Forecasting
• Heart Rate Variability
• Variability-Preserving Imputation
• Wearable Sensors