NET NEWS: Complex Signals

Science  31 Mar 2000:
Vol. 287, Issue 5462, pp. 2371
DOI: 10.1126/science.287.5462.2371b

Physiological processes can be a bear to measure and evaluate. Everything from heartbeats to brain activity during sleep can generate nonlinear, wildly fluctuating patterns of data. A new site called PhysioNet aims to standardize how physiologists make sense of the mayhem.

“The goal is to do for physiology what GenBank did for molecular biology,” says project director Ary Goldberger of Beth Israel Deaconess Medical Center in Boston. Researchers register DNA sequences with GenBank whenever they publish a paper; likewise, Goldberger hopes that physiologists will make their data recordings and analysis algorithms available for the rest of the field to evaluate.

The site, sponsored by the National Institutes of Health, includes a bank that stores recordings, such as brain activity in epileptics, gait patterns, and cardiovascular data from people at risk for sudden cardiac death. Complementing the bank is a tool site, where researchers share and compare their software and analysis models. The site is hosting a contest for the best algorithm to detect evidence of sleep apnea in cardiac signals. A few online tutorials coach newcomers on the uses and limitations of various techniques.

Many disagreements in the physiology community—such as whether certain types of cardiac activity are chaotic or periodic—arise because people work on relatively restricted samples, says Goldberger. But he hopes that “if we meet as a community over common data sets, lots of disagreements can be resolved.”

Navigate This Article