Computational Physiology Course Content
The first 3 days are driven by demonstrating the way computational physiology contributes to the “standard” clinical workflow: data collection, model building, simulation, visualisation, and tying it all together to lead to clinical predictions/outcomes/tools. Each session begins with a lecture from an active computational physiologist presenting current research topics in the exemplar organ system being used in that session.
- Introduction to computational physiology and the standard clinical workflow.
- Introduction to the two main software tools to be used during the DTP CP module: OpenCOR and MAP Client.
- Data collection – (heart exemplar)
- Cardiac lecture
- Understanding source data formats: DICOM
- Image processing and manual segmentation, understanding how to interpret clinical data (e.g., MRI images), how generated data will be used.
- Discussion of automated segmentation methods.
- Model construction – (breast exemplar)
- Breast lecture
- Basic introduction to geometric finite element models
- Creating meshes, fitting meshes
- Discussion on other aspects of model building (material fields, mathematical models)
- Simulation – (GI exemplar)
- GI lecture
- Introduction to numerical methods (focus on numerical integration)
- Exploration of different integration methods and their parameterisation
- Multiscale simulation (electrical propagation in a tissue sheet), and discussion on computational cost/benefit decisions.
- Creating and simulating ordinary differential equation models
- Reproducibility and exchange
- Visualisation – (lung exemplar)
- Lung lecture
- Viewing and interacting with computational models
- Types of graphics, appropriateness of graphics for visualising different types and resolutions of data.
- Time varying visualisations.
- Exporting visualisations for web distribution/viewing.
- Incorporating clinical data (images) into visualisations.
- Complete workflows – (MSK exemplar)
- MSK lecture
- Demonstrate workflows which make use of all the above to define workflows going from source data to clinically relevant predictions/outcomes.
- Discuss best practices for creating complete workflows, especially regarding being able to share and collaborate.
- Best practices in computational physiology
- Physiome lecture
- Reproducibility and reusable mathematical models
- Collaboration, versioning, and model discovery
- Bond graph-based approach to modularity and thermodynamically balanced model construction.
- Projects (more being added each year)
- Mechanical parameter estimation
- Geometric femur fitting
- Bond graph model of the Hill muscle model