Saudi Researchers Develop AI for Rapid and Accurate Sleep Apnea Detection
In a significant scientific advancement, researchers in Saudi Arabia have developed an innovative AI model for detecting obstructive sleep apnea, a condition that affects over a billion people globally. This model utilizes unidirectional electrocardiography and advanced AI techniques.
The study, led by Malak Al-Murshid at the University Center for Sleep Medicine and Research at King Saud University, was published in the journal Frontiers in Artificial Intelligence. It demonstrates that the model employs deep learning based on attention transformers, providing faster and more accurate diagnoses compared to traditional tests.
Results indicated that the new model achieved an F1 score improvement of 13% over previous studies, with the capability to detect apneic events with an accuracy of up to one second. This offers physicians precise and timely diagnostic insights while reducing costs compared to full sleep studies, which are time-consuming and require manual analysis.
The model is notable for using a single biomarker, the electrocardiogram, with intelligent local encoding via an autoencoder. This allows for the processing of raw data without complex pre-analyses and ensures efficiency even with noisy data.
This study reflects Saudi Arabia's ambition to leverage artificial intelligence to enhance its healthcare sector and transition towards a knowledge-based economy post-oil, positioning the kingdom at the forefront of technological innovation both regionally and globally.
