Neurotechnology startup Rythm is building Morpheo, an open-source and secure initiative to help develop machine learning models for automatic and predictive diagnosis of sleep disorders.Morpheo will explore and provide sleep clinicians with a free online software that will automatically analyze sleep patterns and help improve and speed diagnosis related to sleep. Morpheo will also enable software developers and machine learning experts as they create models for the diagnosis of other illnesses. Rythm jointly announced Morpheo with École Polytechnique and the Université Paris Descartes with €5 million in funding from the French government and BPI, a French investment bank.

A collaborative project, Morpheo brings patients, medical doctors, sleep experts, academic researchers, data scientists, web developers, and cryptologists together to aggregate and organize sleep data on an unprecedented scale. Through sophisticated artificial intelligence and blockchain technology, Morpheo helps visualize complex sleep data for clinicians and allows development of machine learning models on large medical datasets. All of the above is achieved while guaranteeing the privacy and security for all data owners.

“Artificial intelligence, ever-increasing computational power, and unprecedented amounts of data are transforming the healthcare industry,” says Hugo Mercier, CEO and co-founder at Rythm, in a release. “We are leveraging these advancements to understand one of the most important areas of health: sleep. We created Morpheo to provide a research tool that will pave the way for preventive and predictive medicine, enabling clinicians to treat pathologies earlier, faster and cheaper, and even to detect them before they happen.”

Once built and launched, Morpheo will be powered by a certified, open-source blockchain infrastructure that guarantees the privacy of data. Users cannot access each other’s data, and only limited, agreed upon machine learning algorithms can access the data for training. Using AI, Morpheo will integrate various sources of information, ranging from traditional medical polysomnography to sleep wearables (like the Dreem headband). With these datasets, Morpheo will identify patterns that characterize sleep pathologies and other related illnesses that are attributed to sleep data.

As Morpheo’s infrastructure is data and algorithm agnostic—in the longer term—it’s approach could be generalized to other healthcare possibilities and beyond. What makes Morpheo unique is its versatility in handling sensitive data to solve critical challenges across various fields that apply and benefit from artificial intelligence.