Project description
Atrial fibrillation (AF) is the most common heart rhythm disorder in the world, affecting 1 in 4 adults aged over 55. Some episodes of AF can terminate spontaneously, while others persist and require intervention. However, distinguishing these in advance remains a major clinical challenge. This project aims to identify predictive markers of AF termination using machine learning applied to ECG recordings in the emergency department (ED). Students will extract features such as dominant frequency, spectral organisation, wavelet energy, entropy, and phase singularity dynamics, then train predictive models (e.g., XGBoost, random forests, transformer-based architectures). Outcome: A prototype early-warning model for AF termination, with potential utility for AF monitoring systems and integration into personalised AF management pipelines.
Assumed knowledge
Programming
Supervisors research focus
Computational cardiology, data science, signal processing, machine learning, biomedical engineering, biophysics
Note: You need to register interest in projects from different supervisors (not a number of projects with the one supervisor).
You must also contact each supervisor directly to discuss both the project details and your suitability to undertake the project.