Additional reading: assessment of consistency in sleep stage using machine learning algorithm

An interpretable machine learning algorithm was used to assess the inter-reliability (IRR) of sleep-center sleep stage annotations. The AI system was trained from 1 hospital learning rater. The results are compared with expert annotations to determine the IRR. Internal and central assessments were performed on 679 patients without sleep apnea in 6 sleep centers in Taiwan,

Inter-rater reliability (IRR) of sleep stage annotations between sleep centers are assessed using interpretable machine learning algorithms while analysis sensors are applied to various parts of the body to monitor brain activity (electroencephalogram [EEG]), eye movement (electrooculogram [EOG]) and muscle activity (electromyogram [EMG]) data to effectively assess the reliability and consistency of sleep scores and evaluate the quality of the sleep center.

The proposed artificial intelligence system proved effective in assessing IRR and hence the sleep center quality. Thus, the proposed algorithm can be used as a tool for scoring at least one sleep stage. The artificial intelligence system can be applied to a single institution to ensure consistency between assessors or to facilitate the evaluation of results from multiple institutions. While increasing the size of the patient body database, will also lead to a more accurate and reliable scoring system and effectively assist physicians in assessing patients.

In conclusion, the ECG belt provided signals comparable to patched ECG and could be used for the assessment of sleep apnea severity, especially during follow-up.

Share and learn

• Compared with Rooti Rx, why the ECG belt can sense the subject’s body data more accurately? Please explain in detail.
• What technology does the ECG belt use to capture a patient’s body parameters?