Wednesday 26 March 2025
Researchers have developed an artificial intelligence-powered agent that can accurately estimate heart rates from photoplethysmogram (PPG) signals, a type of wearable data. This achievement could lead to more efficient and personalized healthcare, where patients receive timely and relevant information about their physical condition.
The PPG signal is generated by the changes in light absorption as blood flows through vessels, and it can be used to monitor various physiological parameters such as heart rate and rhythm. However, processing these signals requires expertise and specialized software, making it challenging for non-experts to analyze them.
To address this issue, a team of researchers has created an agent that uses large language models (LLMs) to extract accurate heart rates from PPG signals. The agent is designed to integrate user interaction, data sources, and analytical tools to generate reliable health insights. In a case study, the agent was evaluated on a dataset of PPG and electrocardiogram (ECG) recordings collected from individuals in a remote health monitoring study.
The results show that the agent significantly outperformed benchmark models, achieving lower error rates and more accurate heart rate estimations. The agent’s performance was benchmarked against OpenAI GPT-4o-mini and GPT-4o, with ECG serving as the gold standard for heart rate estimation. The agent’s accuracy was evaluated using various metrics, including mean absolute error (MAE), root mean square error (RMSE), mean absolute percentage error (MAPE), and median absolute difference (MAD).
The agent’s superiority can be attributed to its ability to integrate user interaction and data sources seamlessly. Unlike traditional methods that require manual uploading of PPG signals, the agent can automatically retrieve relevant data from wearable devices or sensors. This integration enables the agent to process large amounts of data in real-time, providing patients with timely and personalized health insights.
The potential applications of this technology are vast. Patients could receive instant feedback on their heart rate and rhythm, allowing them to make informed decisions about their physical activity and overall well-being. Healthcare professionals could also use this information to monitor patient health more closely, identifying potential issues before they become severe.
Furthermore, the agent’s ability to process PPG signals in real-time opens up possibilities for remote health monitoring. Patients can wear wearable devices or sensors that transmit data to the agent, which would then provide them with personalized health insights and recommendations. This could revolutionize healthcare by making it more accessible and convenient for patients worldwide.
Cite this article: “Artificial Intelligence-Powered Agent Accurately Estimates Heart Rates from Wearable Data”, The Science Archive, 2025.
Artificial Intelligence, Heart Rate Estimation, Photoplethysmogram, Ppg Signals, Wearable Data, Healthcare, Personalized Medicine, Remote Health Monitoring, Large Language Models, Machine Learning.
