16 December 2019
BIOS USES MACHINE LEARNING TO AUTOMATICALLY EXTRACT NEURAL SIGNALS; DISCOVERS RESPIRATORY BIOMARKER
Vancouver, Canada and Cambridge, UK – 16 December 2019 – BIOS, a leading neural engineering startup, has announced that it has successfully used AI to automatically identify a biomarker of respiration activity in neural signals of the vagus nerve. This marks a major advance in BIOS’s quest to decode the human nervous system and create a new type of healthcare which can surpass traditional drugs with neural algorithms that deliver continuous, responsive, personalised care for billions of patients suffering from chronic disease. The team revealed its findings at the Neural Information Processing Systems (NeurIPS) 2019 conference in Vancouver, Canada.
This latest development follows the announcement in May 2019 that BIOS had automatically extracted the neural signals regulating physiological biomarkers such as glucose and blood pressure using an AI-enabled neural interface, making it the first to automatically isolate biomarkers in neural data. The vagus nerve is the largest highway of neural data between the brain and organs. Within this vast amount of messy information of signals going between the brain and organs in the vagus nerve, BIOS was able to reliably and automatically extract neural biomarkers thanks to its unique method of analysing nerve activity that “decodes” the signals of the nervous system. The paper the team presented opens the box on this research method that BIOS is using to map neural signals and systems, which was most recently used to discover the respiratory biomarker.
BIOS’s AI technology picks the signal from the biological noise consistently across multiple subjects throughout various time points. Traditional methods require hand analysis repeated on any single time point to find a signal in a single subject. Biomarkers for neural and bodily function are normally discovered through other means such as blood tests or fMRI scans, whereas this neural biomarker was discovered through machine learning analysis of raw neural recordings. This is significant because being able to use machine learning to find biomarkers of organ function in neural data and be precise about which nerve activity relates to a specific condition will mean effective neural treatments can be developed to replace drugs.
By 2020, chronic diseases are expected to account for more than 70 per cent of all deaths and 60% of the global burden of disease.[i] Many common chronic conditions often occur as a result of a failure or change in our neural pathways. By using machine learning to isolate these damaged signals and for the first time rewrite their code, neural interface technology can revolutionise healthcare by delivering an alternative to drugs. BIOS’ platform approach means that its technology can help any number of partners develop ground-breaking AI treatments to improve the lives of chronic sufferers.
Additionally, BIOS’s results show that the models can identify nerve activity relevant to specific conditions in an unsupervised environment, which vastly accelerates treatment discovery, meaning the approach could be a huge boost for research and bring us closer to neural interface treatments becoming available.
Emil Hewage, Co-Founder and CEO at BIOS, said: “Being able to precisely link nerve activity to specific conditions is nothing short of a game changer. This is a critical breakthrough that will expedite the development of neural interface technology to improve the quality of life of people suffering from chronic respiratory-related illnesses, and we see similar potential for this approach in other chronic diseases including cardiovascular disease and inflammation. Understanding the ‘language’ of the nervous systems means we can develop treatments that communicate directly with organs and nerves, and is a huge leap forward in transforming healthcare through AI and machine learning.”
To access the method paper outlining the discovery which was presented at the NeurIPS conference (“Coordinate-VAE: Unsupervised clustering and de-noising of peripheral nervous system data”), click here.
BIOS combines neural engineering with machine learning to crack the code of the human nervous system. Our neural interface platform interprets the language from the brain to the body to enable more effective treatments for a range of chronic conditions, from heart disease to diabetes. Through our hardware interface and machine learning software, clinical experts can discover, translate, and scale new algorithmic treatments that are personalised, responsive, and targeted. BIOS is positioned to be the platform on which a new generation of AI treatments can be built, helping billions of people suffering from chronic disease to improve their quality of life.
Co-founded by Cambridge University graduates Emil Hewage, a computational neuroscientist, and Oliver Armitage, a biomechanical engineer, BIOS is made up of a wide range of experts from neuroscience, machine learning, software engineering, applied biomaterials, biotechnology, and medicine. Combined the BIOS team has 25 patents, 150+ peer-reviewed publications, and 100+ years of industry experience.
Head of Communications & Partnerships