That breakdown of the heart's exquisitely coordinated electrical activity is the essence of ventricular fibrillation, which causes more than 350,000 deaths a year in the United States alone. Factions of cardiac cells quietly riot against the normal rhythmic motions of the heart, disrupting the pumping process and impeding the flow of blood to vital organs.

Since the early 1980s, the implantable defibrillator has helped manage these unpredictable events in patients with heart disease. Surgically tucked beneath the skin and connected to the heart by electrodes, the battery-powered device continually monitors cardiac performance. If it detects the helter-skelter -- and lethal -- activity of ventricular fibrillation, it restores calm by delivering an electric shock.

Life-saving benefits, however, come at a price, and research has focused on building more efficient, less energy-hogging implants. "We don't have any detailed knowledge of what cardiac cells are doing in a physiological sense during defibrillation," says Nitish V. Thakor, professor of biomedical engineering at The Johns Hopkins University, who is modeling the heart's electrical processes at Pittsburgh Supercomputing Center. "Computer modeling provides the scientific rigor necessary for such an understanding, without having to do a zillion experiments."

From the patient's perspective, adds Matthew G. Fishler, "There are psychological, physiological and practical reasons to reduce the energy levels as much as possible." Repeated defibrillations at increasingly higher voltages are painful, can damage heart and surrounding skeletal tissue and hasten battery replacement, an expensive, invasive surgical procedure that takes a further toll on patients.

Wave of the Future

Under Thakor's direction, Fishler designed a computational model of what happens when either mono- or biphasic waves interact with fibrillating heart tissue. This model is the first of its kind to incorporate the gap junctions by which electrical charges travel from cell to cell, thereby simulating cardiac activity both inside and outside cells. "I can't simulate the whole heart," says Fishler, "so I've isolated a fundamental unit of fibrillation, a two-dimensional spiral rotor, sort of like a pinwheel." Rapidly moving and short lived, these roaming, marauding packs of excitation are believed the source of fibrillation.

Modeling these rotors, says Fishler, is computationally demanding. Generating the rotor itself, which represents 350 milliseconds in real time, required more than 100 hours of C90 processing. With the C90, Fishler overcame several problems he encountered with a Connection Machine CM-2. The C90's global memory allowed a significant performance gain in look-up table routines and the algorithms required to simulate defibrillation shocks are better suited to its vector architecture.

The Right Shock

The Shocking Aftermath Tissue activity two milliseconds after otherwise identical monophasic and biphasic shocks shows contrasting outcomes. For the monophasic shock (top), the leading edge of the rotor (the wavefront) remains steep. For the biphasic shock (bottom), however, the wavefront is graded, like the difference between a cliff and a gentle slope. The gradedness prevents continued propagation across the tissue by reducing the ability of excited and resting areas to interact, terminating the rotor. In contrast, the monophasic wavefront remains steep enough to maintain its integrity and continue its march through the tissue, foiling the defibrillation attempt.

Normal cardiac cells exist in three states: excited, at rest, or refractory -- the interval between excitation and rest. The farther along a cell is in refractory state, the easier it is to excite and, hence, defibrillate. Fishler's simulations showed little difference between mono- and biphasic shocks for cells in early and late refractory: no excitation early, full excitation late. In between early and late refractory, however, biphasics do what monophasics cannot -- induce some level of excitation. And the further along in refractory, the greater the excitation, creating a "graded" response. "By switching polarity, we've gained this graded response," says Fishler, emphasizing that "graded amounts of excitation throughout the heart are much better than random pockets of resting and excited regions."

This understanding of the interaction between low-voltage shocks and cardiac cells will aid future research. Just as pacemakers and defibrillators manage heart disease, says Thakor, implants incorporating novel waveforms could help the body's most important muscle do its job more efficiently: "The electrical and mechanical aspects of the heart are linked -- good electrical function leads to better mechanical output. If the heart gets damaged from failure, it may be possible to use electrical stimulation to improve mechanical function."

References, Acknowledgements & Credits