New machine learning techniques could help measure timing of body clocks to make better health decisions

New machine learning methods can help time our body clocks, helping us all make better health decisions, such as when and how long to sleep.

This study, conducted by the University of Surrey and the University of Groningen, used a machine learning program to analyze metabolites in the blood to predict times in the body’s circadian timing system.

To date, the standard method of determining the timing of the circadian system is to measure the timing of our natural melatonin rhythms, specifically the timing to initiate melatonin production, known as the dim melatonin onset (DLMO).

Co-author of the study, Professor Debra Skene of the University of Surrey, said:

“After taking two blood samples from participants, our method was able to predict an individual’s DLMO with accuracy equal to or better than previous, more intrusive estimation methods.”

The research team collected time-series blood samples from 24 people (12 men and 12 women). All participants were healthy, did not smoke, and slept regularly for 7 days prior to his visit to the university’s clinical research facility. The researchers then used a targeted metabolomics approach to measure more than 130 metabolic rhythms. These metabolic data were then used in a machine learning program to predict circadian timing.

Professor Skane continued:

“We are excited about the new approach to predicting DLMO because it is more convenient than currently available tools and requires less sampling, but we are cautious. Our approach needs validation in different populations.” may pave the way for optimizing circadian therapy for rhythm sleep disorders and injury recovery.

“While smart devices and wearables provide valuable guidance on sleep patterns, our research has the potential to help individuals optimize their health and reduce their risk of serious illnesses associated with poor sleep. It paves the way for a truly personalized sleep and meal plan that’s aligned with the biology of your child: Did you eat at the wrong time?”

Our results may help develop affordable methods for estimating our own circadian rhythms that optimize behavior, diagnostic sampling, and timing of treatments.”

Professor Roelof Hut, co-author of the study at the University of Groningen

This research PNAS.