Undergraduate Internship Programme: Summer 2024

Reverse engineering a model of neural control of breathing using machine learning

William Joy, University of Warwick

Supervisor: Liam Weaver

Liam Weaver was successful in securing £3,500 in BIOREME funding to supervise an 8 week undergraduate summer project during 2024.

The project aimed to reverse-engineer how the respiratory drive centre (RDC) in the brain responds to arterial blood gas (ABG) changes using machine learning. Breathing regulation, driven by the RDC, adjusts based on oxygen, carbon dioxide, and other ABG levels, but the exact mechanisms remain unknown.

To tackle this, Joy generated over 100,000 datapoints using a cardiopulmonary simulator, creating an artificial dataset based on six healthy patient parameters. Machine learning models were developed to predict breathing patterns—such as respiratory rate (RR), musculatory pressure (Pmin), and inspiration-to-expiration ratio (I:E Ratio)—from ABG inputs. The Narrow Neural Network outperformed for Pmin, while decision trees excelled for RR and I:E Ratio, achieving average errors under 19%.

The project highlighted the feasibility of machine learning in modelling RDC control but noted the need for further work on unhealthy patient simulations. Beyond technical achievements, the project enriched Joy’s machine learning expertise, introduced him to digital twin simulations, and prepared him for his next academic step: a PhD in Biomedical Systems Engineering.

“By undertaking this BIOREME project I feel much more comfortable entering the world of academia and believe it gave me a very solid foundation, which will allow me to hit the ground running when I start my course” William Joy