iSGTW Feature - Be still, my heart

Feature - Be still, my heart

A whole ventricular model showing reentrant activity in the ventricles. The arrows show the direction of wave propagation. Image courtesy Thushka Maharaj

It's a familiar scene on television shows such as "ER": A patient's heart is beating improperly, a doctor puts a pair of electrical paddles against her chest, yells "clear," gives an electric shock - and whew, the heart starts beating again normally. The whole process, known as "defibrillation," is relatively common in real-life emergency rooms.

But what is really happening, and why does it work?

Thushka Maharaj, a doctoral student within the Computational Biology Group at the University of Oxford, hopes to answer these questions. She is part of an international collaboration studying the effects of applying electrical shocks to both healthy and diseased hearts to understand exactly how defibrillation works. In order to do so, she simulates on computers the application of electric shocks to heart tissue.

The broad picture is well-known: Within a normal, healthy heart, the muscle cells (myocardial cells) produce regular, powerful contractions that allow the heart to pump blood around the body. During "fibrillation," these contractions are no longer regular and powerful, but irregular - so the heart is unable to pump blood around the body. Shocking the heart into stopping completely - "defibrillation" - gives the myocardial cells a chance to get back into a regular rhythm again.

Simulation of the outer (epicardium) and inner (endocardium) layers of the heart. Image courtesy National Grid Service.

Enter the grid

But the exact mechanism is still a puzzle. To work it out, Maharaj and colleagues at the Integrative Biology Project run many sequential simulations, varying the parameters for the intensity of the electrical shock, the timing of its application and the heart's tissue properties. To obtain a mere 250 milliseconds of animated data, the researchers require 28 hours of processing time per variable.

"We use parallel code with around a million nodes, so it is pretty computationally intensive," explains Maharau.

But with the use of the UK's National Grid Service (NGS), Maharaj and her supervisor, Blanca Rodriguez, can run hundreds of sequential simulations on many CPUs.

"We can get 20ms of animation in 20 minutes using 32 CPUs on the NGS, which is a huge improvement," says Mahraj. In addition, the NGS' Storage Resource Broker, a data storage system developed at the San Diego Supercomputing Center, improves reliabilty and efficiency dramatically. This system allows users to store data in geographically distributed locations without having to keep track of where each file resides; the use of logical filenames (as opposed to physical filenames that indicate the path to a file) allows users to easily search for data across the locations.

Maharaj wonders if she would have been able to attempt her doctoral research without the grid.

"I didn't even know the NGS existed before starting my doctorate, but I don't think we could have run these simulations without using the NGS. And the benefits of services such as the Storage Resource Broker are immense - it's fantastic to be able to share data with colleagues all over the world so easily."

-Dan Drollette, iSGTW