Sleep disordered breathing detection using heart rate variability and R-peak envelope spectrogram.

Area of Science:

• Biomedical Engineering
• Cardiology
• Sleep Medicine

Background:

• Sleep disordered breathing (SDB) is a prevalent condition affecting cardiovascular health.
• Accurate SDB detection often relies on polysomnography, which is resource-intensive.
• Non-invasive, accessible methods for SDB detection are needed.

Purpose of the Study:

• To develop and validate a novel method for detecting SDB using single-lead electrocardiogram (ECG) recordings.
• To investigate the efficacy of combining heart rate variability (HRV) and R-peak envelope (RPE) features for SDB detection.
• To assess the performance of a machine learning classifier utilizing these combined features.

Main Methods:

• Extraction of textural features from normalized gray-level cooccurrence matrices of time-frequency plots of HRV and RPE sequences.
• Selection of an optimal subset of textural features for classification.
• Utilizing a multi-layer perceptron (MLP) as the classifier for SDB detection.

Main Results:

• The proposed method achieved high average training performance: 100.0% sensitivity, 99.9% specificity, and 99.9% accuracy.
• Mean testing performance demonstrated strong results: 99.0% sensitivity, 96.7% specificity, and 97.8% accuracy.
• The approach showed significant improvement in SDB detection from single-lead ECG.

Conclusions:

• Combining HRV and RPE analysis from single-lead ECG offers a promising, accurate, and non-invasive approach for SDB detection.
• The textural feature extraction and MLP classification strategy effectively identifies SDB.
• This method has the potential to enhance SDB screening and diagnosis in clinical settings.