Review of uncertainty in heart rate variability and challenges with the control of its impact in artificial intelligence-driven medical applications

Abstract

As global adoption of artificial intelligence (AI) continues to rise, attention has been paid to the performance of large language models (LLMs) in healthcare with emphasis on a distinct gap between the controlled studies and the real clinical settings. Heart rate variability (HRV) has emerged as an essential non-invasive tool for understanding cardiac autonomic function, evaluating and identifying various health impairments. Thus, HRV is increasingly utilized within LLMs. This narrative review comprehensively describes a range of AI-driven applications incorporating HRV based on representative cases selected from searching results. We analyse the factors' influence on data quality and the effects of performance at model output with focus on the gap between expectations and reality. The notable advances in validation technique are described. The significance of latest research on data technology is also summarized with focus on bridging the gap between clinical trust and advanced models of the next phase.