Hand in bed holding reading glasses

AI for a good night’s sleep

Almost one billion people suffer from sleep apnoea. Unfortunately, our current diagnostic metrics are simplistic at best, simply measuring the frequency of breathing cessations.

An Horizon 2020 project, supported in Australia by a $500,000 collaborative National Health and Medical Research Council grant, hopes to address that by coming up with an artificial intelligence solution that takes into account symptoms and comorbidities of sleep apnoea. It will be used to develop individual treatments addressing each patient’s symptoms.

The project is being coordinated by Reykjavik University with 37 partners across Europe and the University of Queensland (UQ) in Australia.

“An increase in the number of patients will pose challenges to health care and its resources,” Professor Juha Töyräs, Head of Biomedical Engineering at the School of Information Technology and Electrical Engineering, UQ. He is also professor of Medical Physics and Engineering at University of Eastern Finland.

“It is therefore important that the project also takes into account therapeutic perspectives and explores the cost-effectiveness of the entire chain of examinations and care.”

The ultimate objective of the project, which runs to 2025, is to draw up new international guidelines for the diagnostics of sleep apnoea in collaboration with the European Sleep Research Society and national sleep research societies.

The key to the study is the creation of a data-rich patient database of more than 30,000 sleep recordings collected in hospitals and research centres, and augmented by sleep questionnaires and through consumer products, such as smartwatches.

Sleep apnoea has been associated with an increased risk for stroke, cardiovascular diseases, type II diabetes, and daytime symptoms that erode productivity and quality of life and even road safety.