Discoverable, reproducible, and reusable cardiac models
presentationposted on 07.12.2016, 07:01 authored by David NickersonDavid Nickerson
Talk I presented at the 16th International Conference on Biomedical Engineering. Submitted abstract:
Computational models of the heart have evolved substantially since Denis Noble published his original 1962 modification of the Hodgkin & Huxley equations applicable to Purkinje fibre action and pacemaker potentials. As available computational power has grown the range of cardiac physiology and anatomy able to be represented by the mathematical models and numerical simulations has rapidly increased, to the point where we are now able to make use of multiscale and multiphysics cardiac models in scientific investigations. Models of cardiac electrophysiology, for example, are often used in the pharmacological industry as a screening tool for drug efficacy. While predictive cardiac models are not quite yet routine in clinical practice they are beginning to be utilised in specific situations. For example, ventricular mechanics models are used to help interpret cardiac MRI images and ventricular fluid flow observed using ultrasound.
While recent cardiac computational models can be very detailed biophysically and anatomically, scientists face many challenges when attempting to reproduce the results of a given modelling study - often as the first step in a desire to reuse an existing model. For instance, cardiac cellular models consisting of large systems of ordinary differential equations are frequently impossible to implement from the published paper alone. To address this problem, the Physiome and Virtual Physiological Human communities have developed standards for encoding model and simulation descriptions which enable scientists to archive and exchange their models in a reproducible manner.
The encoded model or simulation alone, however, is often not sufficient to enable scientists to discover and reuse the computational model. Here I will present recent developments aimed at improving the discoverability and reusability of cardiac models. Future work to encourage scientists to make use of these developments will also be presented.