PhysioNet


PhysioNet is a unique resource for research and education, offering free access to large collections of physiological data and related open-source software. In cooperation with the annual Computing in Cardiology conference, PhysioNet also hosts an annual series of challenges, focusing research on unsolved problems in clinical and basic science. PhysioNet is managed by members of the MIT Laboratory for Computational Physiology. For an overview of PhysioNet resources, see our content overview or search our content.



A Community Resource

The PhysioNet resource, established in 1999, is intended to stimulate current research and new investigations in the study of complex biomedical and physiologic signals. It has three closely interdependent components:

  • An extensive archive of well-characterized digital recordings of physiologic signals, time series, and related data for use by the biomedical research community. PhysioNet includes collections of cardiopulmonary, neural, and other biomedical signals from healthy subjects and patients with a variety of conditions with major public health implications, including sudden cardiac death, congestive heart failure, epilepsy, gait disorders, sleep apnea, and aging. These collections include data from a wide range of studies, as developed and contributed by members of the research community.

  • A large and growing library of software for physiologic signal processing and analysis, detection of physiologically significant events using both classical techniques and novel methods based on statistical physics and nonlinear dynamics, interactive display and characterization of signals, creation of new databases, simulation of physiologic and other signals, quantitative evaluation and comparison of analysis methods, and analysis of nonequilibrium and nonstationary processes.

  • A collection of popular tutorials and educational materials, offering expert guidance in approaches for exploring and analysing health data and physiologic signals. A unifying theme for these resources is a focus on the extraction of “hidden” information from biomedical data, providing information that may have diagnostic or prognostic value in medicine, or explanatory or predictive power in basic research.

PhysioNet is not only the name of the Resource, but also of its web site, physionet.org. The website was established by the Resource as its mechanism for free and open dissemination and exchange of recorded biomedical signals and open-source software for analyzing them, by providing facilities for cooperative analysis of data and evaluation of proposed new algorithms. In addition to providing free electronic access to data and software, the PhysioNet web site offers service and training via on-line tutorials to assist users at entry and more advanced levels. In cooperation with the annual Computing in Cardiology conference, PhysioNet hosts an annual series of challenges, in which researchers and students address unsolved problems of clinical or basic scientific interest using data and software provided by PhysioNet.

All data and software included in PhysioNet are carefully reviewed. We invite you to participate in the ongoing review process. By sharing common data sets, and software in source form, the research community benefits from access to materials that have been rigorously scrutinized by many investigators. We further invite researchers to contribute data and software for review and possible inclusion in our resource collection. Please review our guidelines for contributors before submitting material.



Brief History


Beginning in the mid-1970s, members of the PhysioNet team who were then working on some of the first microcomputer-based instruments for cardiac arrhythmia monitoring foresaw the usefulness of establishing shared databases of well-characterized ECG recordings, as a basis for evaluation, iterative improvement, and objective comparison of algorithms for automated arrhythmia analysis. A five-year effort culminated in the publication of the MIT-BIH Arrhythmia Database in 1980, which soon became the standard reference collection of its type, used by over 500 academic, hospital, and industry researchers and developers worldwide during the 1980s and 1990s. Other databases of ECGs and eventually other physiologic signals followed. By 1999, the MIT group distributed CD-ROMs containing 11 such collections, and had participated in the development of several others.

PhysioNet was established in 1999 as the outreach component of the Research Resource for Complex Physiologic Signals, a cooperative project initiated by a diverse group of computer scientists, physicists, mathematicians, biomedical researchers, clinicians, and educators at Boston's Beth Israel Deaconess Medical Center/Harvard Medical School, Boston University, and McGill University, all working together with the MIT group. Many of us have worked together for 20 years or even longer on problems relating to characterizing and understanding the dynamics of human physiology, the implications of dynamical change in diagnosis and treatment of pathophysiology, novel and robust techniques for physiologic monitoring in ambulatory subjects and critical care patients, and applications of model-based reasoning to medical decision support in intensive care. The MIT group contributed its 11 databases, and the software it had developed for exploring and analyzing them, to establish PhysioBank and PhysioToolkit. Free availability of these resources via the Internet catalyzed an even greater explosion of interest in them, as researchers and students worldwide who had no previous access to such data or software began new programs of research, and specialists began comparing their methods. These initial contributions were quickly supplemented by additional collections of data and software from their collaborators, and soon after, from many researchers worldwide. PhysioBank and PhysioToolkit have grown to many times their original sizes, and most of the growth has been thanks to the hard work and generosity of an international community of researchers.

At the time PhysioNet was established, members of the PhysioNet team at MIT were preparing to host Computers in Cardiology 2000. We hoped to introduce PhysioNet to our international colleagues who would be attending CinC, by encouraging participation in an activity that made effective use of the facilities provided by PhysioNet to stimulate rapid progress on an unsolved problem of practical clinical significance. A timely contribution of data made it possible to create the first PhysioNet/CinC Challenge, which attracted the attention of more than a dozen teams to the subject of detecting sleep apnea from the ECG. Their efforts were broadly successful, they discussed their findings at CinC 2000, and an annual tradition was born. For a summary, see our timeline.



Funding


PhysioNet was originally established under the auspices of the NIH's National Center for Research Resources in 1999 (which was abolished in 2011). We gratefully acknowledge support from the following organizations:

  • NIH National Institute of Biomedical Imaging and Bioengineering and the National Institute of General Medical Sciences, under grant 2R01GM104987-09.


Current Research


Methods for assessment of signal quality and detection of events in weakly correlated multiparameter data; false alarm reduction in the ICU; methods for multivariate trend analysis and forecasting, with applications in intensive care; cardiovascular system modeling (including adaptation to microgravity and orthostatic intolerance); novel signal processing techniques for automated or semi-automated patient diagnosis; web-enabled signal processing, with applications in research and telemedicine; data mining algorithms for efficient searching in very long time series; networked instrumentation for acquisition and remote viewing of real-time physiologic data (Roger Mark, George Moody, Li-wei Lehman, Benjamin Moody, Chen Xie, Felipe Torres Fabregas, Tom Pollard).

Algorithms that quantify the transient and local properties of nonstationary physiologic signals and the cross-interactions among multiparameter signals; application of these techniques to detect changes that may precede the onset of catastrophic physiologic events, including epileptic seizures and sudden cardiac death; techniques for quantifying the dynamics of physiologic control; mathematical/physiological modeling of these control mechanisms; identification of new measures related to nonlinear dynamics and fractal scaling that have diagnostic/prognostic use in life-threatening cardiopulmonary pathologies (Madalena Costa, Ary Goldberger).




Contact us