A review of automatic sleep stage classification using machine learning algorithms based on heart rate variability.

Area of Science:

• * Cardiovascular Physiology and Computational Neuroscience.

Background:

• * Heart rate variability (HRV) is increasingly utilized due to its ease of collection, affordability, and relevance to psychophysiological and psychopathological conditions.
• * HRV analysis offers a non-invasive method for assessing autonomic nervous system function.
• * Sleep stage classification is crucial for diagnosing sleep disorders and monitoring overall health.

Purpose of the Study:

• * To review and evaluate machine learning algorithms for automatic sleep stage classification using HRV signals.
• * To compare HRV features, classification methods, and evaluation metrics employed in recent studies.
• * To demonstrate the effectiveness of AI-driven HRV analysis for sleep staging.

Main Methods:

• * Systematic review of machine learning algorithms applied to HRV-based sleep stage classification over the last 15 years.
• * Extraction and comparison of relevant HRV features.
• * Analysis of various classification algorithms (e.g., SVM, deep learning) and their performance metrics.

Main Results:

• * Machine learning algorithms leveraging HRV features achieve high accuracy, sensitivity, and specificity in sleep staging.
• * Feature selection and algorithm choice significantly impact classification performance.
• * Advancements in technology have led to improved performance of HRV-based sleep analysis.

Conclusions:

• * AI approaches using HRV signals show significant promise for accurate automatic sleep stage classification.
• * HRV-based sleep analysis with machine learning is a rapidly evolving field with substantial clinical potential.
• * Future developments are expected to yield more personalized and precise solutions in sleep medicine.