Abstract
Over the past few decades, the use of heart rate variability (HRV) has expanded significantly due to its ease of collection, affordability, and its clinical relevance to psychophysiological processes and psychopathological disorders. This study aims to demonstrate the effectiveness of an artificial intelligence approach based on HRV signals for automatic sleep stage classification. This review examines machine learning algorithms for HRV-based sleep stage classification over the past 15 years. It also compares the HRV features extracted, the classification algorithms used, and the evaluation parameters employed. Existing studies indicate that with advances in technology, machine learning algorithms utilizing HRV features for sleep staging achieve high accuracy, sensitivity, and specificity. The use of HRV for sleep analysis via machine learning algorithms is an active area of research with broad application potential. As technology progresses and data accumulation increases, this approach is expected to offer more accurate and personalized solutions for sleep medicine and health management.