Beth Israel Deaconess Medical Center MARGRET & H. A. REY
for Nonlinear Dynamics in Medicine
Harvard Medical School

Google Search Tips 

Search WWW ReyLab


Mission Statement

Research Projects



Workshop / Talks


Related Links

Time Series Competition



ECG Wave-Maven

Center for Dynamical Biomarkers
George Moody is a leading expert on ambulatory ECG monitoring, has expertise in signal processing, data acquisition, and software and hardware development.

Since 1977, Mr. Moody has worked on topics related to automated processing and interpretation of physiologic signals, concentrating for much of that time on the long-term ECG and related signals. His research interests include robust methods in pattern recognition and power spectral density estimation, automated arrhythmia and ischemia detection, artificial intelligence-based medical decision support, and heart rate variability. He has invented a wide variety of innovative algorithms for physiologic signal processing, including the EDR (ECG-Derived Respiration) technique; a method for atrial fibrillation detection based on quasi-continuous Markov process-like models; ECG-based indices of physical activity; the TRIM (Turning-point/Recursive Improvement Method) algorithm, a widely-used ECG compression algorithm; and novel methods for robust estimation of principal components of continuous waveforms, such as QRS and ST waveforms in the ECG. He was also the first to apply the Lomb transformation of unevenly sampled data for power spectral density estimation of heart rate variability. In addition, he has designed and implemented many of the physiologic signal processing and analysis algorithms in regular use in our laboratory, including ARISTOTLE (a state-of-the-art arrhythmia detector that has been the nucleus of several successful commercial products) and WAVE (an extensible interactive graphical environment for exploratory data analysis of digital signals, in use by researchers worldwide).

Mr. Moody has had major roles in the development of most of the standard databases of digitized ECGs and other physiologic signals, including (among others) the MIT-BIH Arrhythmia Database (1980), the European Society of Cardiology ST-T Database (1990), the Massachusetts General Hospital/Marquette Foundation Waveform Database (1992), and, most recently, the MIMIC Database (see Resources). He has had a leading role in the standards development work of the Association for the Advancement of Medical Instrumentation (AAMI) with respect to automated ECG analysis. He contributed the methods for quantitative evaluation of ECG analysis algorithms, the reference implementations of those methods in software, and the descriptions of those methods contained in the recently adopted American National Standard for Ambulatory Electrocardiographs (ANSI/AAMI EC38-1994), while serving as a member of the AAMI ECG Committee and of its Ambulatory Monitoring Subcommittee, and as the chair of its Arrhythmia Monitoring Subcommittee. In earlier work for AAMI, he wrote much of the AAMI Recommended Practice for Testing and Reporting Performance Results of Ventricular Arrhythmia Detection Algorithms (AAMI ECAR-1987).