R.G. Mark

The impact of the MIT-BIH Arrhythmia Database

The MIT-BIH Arrhythmia Database was the first generally available set of standard test material for evaluation of arrhythmia detectors, and it has been used for that purpose as well as for basic research into cardiac dynamics at about 500 sites worldwide since 1980. It has lived a far longer life than any of its creators ever expected. Together with the American Heart Association Database, it played an interesting role in stimulating manufacturers of arrhythmia analyzers to compete on the basis of objectively measurable performance, and much of the current appreciation of the value of common databases, both for basic research and for medical device development and evaluation, can be attributed to this experience. In this article, we briefly review the history of the database, describe its contents, discuss what we have learned about database design and construction, and take a look at some of the later projects that have been stimulated by both the successes and the limitations of the MIT-BIH Arrhythmia Database.

A database for evaluation of algorithms for measurement of QT and other waveform intervals in the ECG

Presents a QT database designed for evaluation of algorithms that detect waveform boundaries in the ECG. The database consists of 105 fifteen-minute excerpts of two-channel ECG Holter recordings, chosen to include a broad variety of QRS and ST-T morphologies. Waveform boundaries for a subset of beats in these recordings have been manually determined by expert annotators using an interactive graphic display to view both signals simultaneously and to insert the annotations. Examples of each morphology were included in this subset of annotated beats; at least 30 beats in each record, 3622 beats in all, were manually annotated in the database. In 11 records, two independent sets of annotations have been included, to allow inter-observer variability studies. The QT Database is available on a CD-ROM in the format previously used for the MIT-BIH Arrhythmia Database and the European ST-T Database, from which some of the recordings in the QT Database have been obtained.

The apnea-ECG database

Sleep apnea is a sleep disorder with a high prevalence in the adult male population. Sleep apnea is regarded as an independent risk factor for cardiovascular sequelae such as ischemic heart attacks and stroke. The diagnosis of sleep apnea requires polysomnographic studies in sleep laboratories with expensive equipment and attending personnel. Sleep apnea can be treated effectively using nasal ventilation therapy (nCPAP). Early recognition and selection of patients with sleep related breathing disorders is an important task. Although it has been suggested that this can be done on the basis of the ECG, careful quantitative studies of the accuracy of such techniques are needed. An annotated database with 70 nighttime ECG recordings has been created to support such studies. The annotations were based on visual scoring of disordered breathing during sleep.

A database to support development and evaluation of intelligent intensive care monitoring

Development and evaluation of automated decision support systems requires large amounts of well-characterized, reproducible test data. The MIMIC (Multi-parameter Intelligent Monitoring for Intensive Care) Database is intended to meet these needs. The database, currently nearing completion, will include 100 patient records, each typically containing between 24 and 48 hours of continuous data recorded from patient monitors in the medical, surgical, and cardiac intensive care units of Bostons Beth Israel Hospital. Each record will be accompanied by detailed clinical data derived from the patients medical record and from the hospitals on-line medical information systems. We select patients to record from those likely to be hemodynamically unstable during the planned recording period. We expect to complete the selection of the recordings to be included in the database by the end of 1996, and to make the database available to other researchers shortly thereafter.

Integration of real-time and off-line clinical data in the MIMIC database

The MIMIC (Multi-parameter Intelligent Monitoring for Intensive Care) database is a collection of recorded physiological signals and clinical data that is intended to support the development and evaluation of automated decision support systems for intensive care settings. Each patient record typically includes 24 to 48 hours of continuous real-time signals (e.g. ECGs, blood pressure waveforms or respiration) and derived measurements (e.g. heart rate or systolic blood pressure), together with clinical data derived from the patients medical record and from the hospitals clinical computing systems. The clinical data are encoded using structured HTML (readable using a standard Web browser, or by a decision support system or other automated application). An important element of this encoding is a reconstruction of the medical record in the form of a time-line of events, which permits a detailed and repeatable simulation of the inputs that are potentially available to a real-time intelligent ICU monitor.

The European ST-T database: development, distribution and use

The European project for the development of an ST-T annotated database originated from the Concerted Action on ambulatory monitoring, set up by the European Community in 1985. The goal was to define an electrocardiographic (ECG) database for assessing the quality of ambulatory ECG monitoring (AECG) systems. Thirteen research groups from 8 countries continued providing AEGG tapes and annotating beat-by-beat the selected 2-channel records, each 2 hours in duration. ST segment and T-wave changes were identified and their onset, offset and peak beats annotated in addition to QRSs, beat types, rhythm and signal quality changes. The first set of 50 records was completed and stored on compact-disk ROM. It includes more than 200 ST segment and almost 300 T-wave changes. In cooperation with the developers of the Massachusetts Institute of Technology (MIT)-Bostons Beth Israel Hospital (BIH) arrhythmia database, the annotation scheme was devised to be consistent with both MIT-BIH and American Heart Association (AHA) formats.<<ETX>>

Performance measures for algorithms to detect transient ischemic ST segment changes

The availability of the European ST-T database makes it possible to perform quantitative, reproducible performance tests of methods for detecting and measuring transient ischemic ST changes in the electrocardiogram (ECG). The authors propose an evaluation protocol and performance measurements for use with the database. They describe methods for evaluating the accuracy of ST episode detection, measurement of ischemia duration, and measurements of ST deviation. Bootstrap estimation is used to derive expected lower bounds on performance measures and to assess their utility as predictors of performance. These methods are illustrated by a case study in which an evaluation is presented of a two-channel algorithm for automated detection of ischemic ST episodes.<<ETX>>

Stimulating rapid research advances via focused competition: the Computers in Cardiology Challenge 2000

Obstructive sleep apnea is associated with a variety of serious health risks. Standard methods for detecting and quantifying sleep apnea are based on respiration monitoring, but previous studies have suggested that apnea detection based on the ECG alone might be possible. The authors therefore offered a challenge to the research community, to demonstrate the efficacy of ECG-based methods for apnea detection using a large, well-characterized, and representative set of data made freely available via the Internet. The goal of the contest was to stimulate effort and advance the state of the art in this clinically significant problem, and to foster both friendly competition and wide-ranging collaborations, The event was an outstanding success, with most entrants achieving 90% to 100% accuracy in identifying subjects with significant apnea, and minute-by-minute apnea detection accuracy between 85% and 93%, comparable to the concurrence of human experts scoring full polysomnograms.

PhysioNet: a research resource for studies of complex physiologic and biomedical signals

PhysioNet (http://www.physionet.org/) is a web-based resource supplying well-characterized physiologic signals and related open-source software to the biomedical research community. Inaugurated in September 1999 under the auspices of the NIHs National Center for Research Resources (NCRR), PhysioNet provides an on-line forum for free dissemination and exchange of research data and software, with facilities for cooperative analysis of data and evaluation of new analytic methods. As of September 2000, PhysioBank, the data archive made available via PhysioNet, contained roughly 35 gigabytes of recorded signals and annotations. PhysioNet is a public service of the Research Resource for Complex Physiologic Signals, a cooperative project initiated by researchers at Bostons Beth Israel Deaconess Medical Center/Harvard Medical School, Boston University, McGill University, and MIT.

Repolarization alternans detection using the KL transform and the beatquency spectrum

Electrical alternans at the repolarization phase of the cardiac cycle have been reported to be related with increased risk for ventricular arrhythmias (VA). Here, the authors present and study a method for detecting ST-T complex alternans based on the Karhunen-Loeve (KL) transform and a beatquency transform (Fourier transform of KL series with beat order as temporal reference). The ratio of the power in the band around 0.5 cycles/beat with respect to the total power is then used as indicator of the period 2 alternans presence. This is similarly done for period 4 alternans. To study the performance of this detection system the authors have generated a simulated alternating ECG with added noise (moving artifacts, EMG and baseline wander noise). The results for SNR around 15 dB for the ST-T complex (normal value for ECG signals) give a sensitivity and specificity for period 2 alternans detection of over 95%. This method applied to the entire European ST-T database gives 25 alternating episodes, 56% corresponding to ischemic episodes, showing that 5.3% of the ischemic episodes include alternans episodes.