New Delhi: A routine sleep study can contain hidden insights about a patient’s health that are usually overlooked at clinics. But artificial intelligence has made it easy. In a new study, an AI model helped capture such information and detected sleep patterns linked to cognitive decline, heart diseases, and death.
For the study, published in the journal Nature Communications on 3 August, AI researchers, sleep physicians, data scientists, and neuroscientists got together to build an AI model. The team worked under a 10-year research partnership between Cleveland Clinic and IBM, focused on using AI and quantum computing to accelerate discoveries in life sciences.
The AI model successfully identified “clinically meaningful” subtypes, categorising patients based on their long-term health risks. For this, it used standard sleep data and captured signals that are often missed in conventional summary measures.
“For decades we have distilled an overnight sleep study into a handful of summary measures,” said Reena Mehra, professor of medicine at the University of Washington and the study’s senior clinical author. “AI gives us the opportunity to move beyond those summaries and learn from the full richness of sleep physiology.”
Also read: Sleep disorders leave you tired? They may also alter your brain, US study finds
Extracting hidden information
Analysing sleep patterns — using data from the Cleveland Clinic Sleep Signals, Testing, and Reports Linked to Patient Traits (STARLIT) registry — the AI model divided patients into five distinct risk categories. It turned out that the patients identified by the model as high-risk had twice the risk of dying over the next five years compared to those classified as low-risk. This was not captured by the apnea-hypopnea index (AHI), which serves as a measure of the severity of obstructive sleep apnea (OSA).
The index records the average number of breathing pauses and shallow breathing events during a sleep per hour, which is then used by doctors to assess how bad the sleep apnea is in the patient. The study revealed that AHI could also contain vital information about how the heart, brain, lungs, and muscles respond during sleep.
In addition, the model also worked well in both men and women, since AHI has historically identified risk better in men.
“Modern AI lets us recover much more of the information contained in a night’s worth of sleep physiology, revealing clinically meaningful patient groups with very different long-term health risks,” said Jeffrey Rogers, the corresponding author of the study and an adjunct neurosurgery professor at the Yale School of Medicine. “These findings demonstrate that routine medical tests can contain substantially more physiologic information than current clinical practice extracts from them.”
