Pulse wave signals (e.g., blood pressure, PPG) are widely used in clinical and consumer settings (e.g. smartwatches) to assess cardiovascular function. Influenced by the heart, blood vessels, and autonomic and respiratory systems, they provide rich physiological information. Many indices and algorithms (including machine learning models) analyse pulse wave morphology to infer cardiovascular health.

However, evaluating these tools is challenging due to difficulties in measuring key reference variables (e.g., cardiac output), isolating individual factors in vivo, and capturing pulse waves from all relevant sites (especially central arteries). Clinical studies are also costly, time-consuming, and prone to measurement errors.
To facilitate this process, we use computational blood flow models to generate simulated pulse wave datasets that reflect real populations. This 'population-specific’ modelling approach enables rapid, cost-effective development and pre-clinical testing of pulse wave analysis algorithms across a wide range of cardiovascular conditions. These in silico datasets also help us explore the mechanisms behind observed populations trends and train machine learning algorithms across diverse physiological scenarios (Physiol Meas, 2024).

All the datasets created up to now - together with Matlab functions for processing and analysing the data - can be downloaded by clicking the above buttons.