Obstructive Sleep Apnoea Detection Gets AI Boost From Saudi Study Using ECG

A Saudi research team has built an intelligent model that detects obstructive sleep apnea using artificial intelligence and electrocardiography signals. The work, led by Dr. Malak Al-Marshad at King Saud University, aims to support faster and cheaper diagnosis for a condition affecting more than one billion people around the world.

The model is reported to be more efficient than conventional polysomnography, which is the current standard test for obstructive sleep apnea. Polysomnography typically needs overnight monitoring in a sleep laboratory, specialised equipment, and expert analysis, making diagnosis slow, labour-intensive, and expensive for hospitals and patients.

The study appears in the journal "Frontiers in Artificial Intelligence" and focuses on analysing a single unidirectional ECG signal. The system uses transformer-based AI, similar in principle to large language models, together with autoencoder-based positional encoding. This design allows direct processing of raw ECG data, avoiding heavy manual preprocessing or handcrafted feature extraction often used in earlier medical algorithms.

Results show that the attention transformer-based deep learning model improves performance compared with earlier research on obstructive sleep apnea detection. The study reports a 13% gain in F1 score over previous work. The model also reaches high temporal accuracy, identifying apnea events with precision down to one second, even when tested on noisy signals that reflect real clinical situations.

The research highlights growing global interest in artificial intelligence applications within sleep medicine and respiratory disorders. King Saud University has become an active centre in this field. According to the 2025 ScholarGPS rankings, King Saud University holds 18th place worldwide for sleep medicine research output over the previous five years.

The same ScholarGPS 2025 listing ranks Professor Ahmed BaHammam of the College of Medicine at King Saud University fifth globally among sleep medicine scientists during that period. The new AI model, developed at the University Sleep Medicine and Research Center and Medical City, offers physicians decision support that is faster, more affordable, and reliable, potentially improving obstructive sleep apnea screening and follow-up in Saudi Arabia and internationally.

With inputs from SPA