2014 News

Combined measurement of ECG, Breathing, and Seismocardiograms DataBase, CEBSDB. A dataset of 60 records from 20 volunteers has been contributed to PhysioBank by Miguel Angel Garcia Gonzalez and Ariadna Argelagos Palou from the Universitat Politecnica de Catalunya. Each record contains two ECGs, a respiration, and a seismocardiogram signals. The database was designed to check if errors in RR series are influenced by breathing, and if there is a significant variability in RR series estimated from ECGs versus seismocardiograms.

Two new software packages have been added to PhysioNet by Joachim Behar, Gari Clifford, and colleagues at the University of Oxford. FECGSyn is a new Foetal ECG Waveform Generator. The second package, the Random Search Toolbox, is a MATLAB toolbox for selecting model hyper-parameters via random search.

The top official results in Phase III were achieved by Alistair Johnson (87.9), Teo Soo-Kng (86.7), Thomas De Cooman (86.6), Jan Gierałtowski (86.4), Marcus Vollmer (86.2).

Information on how to prepare your talk for Computing in Cardiology is described here.

Two upcoming events this September at MIT will be of interest to PhysioNet users. The first is the International Critical Care Data Mining Marathon, a weekend event (5-7 September 2014). The International Critical Care Data Mining Marathon will bring together teams of engineers and clinicians to develop tools and crowd-source knowledge in clinical data, with a hands-on data mining workshop using the MIMIC-II Database.

The second event is the 2014 Computing in Cardiology conference held from September 7-10 2014 at Cambridge, Massachusetts. Computing in Cardiology is an international scientific conference that has been held annually since 1974 and hosting PhysioNet Challenges since 2000. This conference provides a forum for scientists and professionals from the fields of medicine, physics, engineering and computer science to discuss their current research in topics pertaining to computing in clinical cardiology and cardiovascular physiology.

The top official results in Phase II were achieved by Thomas De Cooman (86.2), Marcus Vollmer (86.1), Urska Pangerc (85.2), Filip Plesinger (85.0), and Alistair Johnson (84.6).

The top official results in Phase I were achieved by Marcus Vollmer (93.2), Urska Pangerc (89.2), Lars Johannesen (88.9), Quan Ding (88.9), and Teo Soo-Kng (88.7). Several other high-scoring entries were disqualified for reasons described in the FAQs.

The ECG-ID Database is a set of 310 ECGs from 90 volunteers, created and contributed to PhysioBank by Tatiana Lugovaya, who used the ECGs in her master's thesis. An excellent summary of this thesis, with a discussion of the challenges in using ECGs as biometrics, and a comparison of the author's methods and results with those of three previous studies, is also available.

This dataset, created and contributed by Luca Citi, Riccardo Poli, and Caterina Cinel, was generated as part of a study aimed at identifying the factors limiting the performance of brain-computer interfaces (BCIs) based on event-related potentials (ERPs), in order to improve the transfer rate and the usability of these interfaces. Twenty recordings of each of 10 participants include annotated 64-channel EEGs and 4-channel EOGs, generated while the participants focused on specified target characters displayed by a traditional matrix speller.

This data collection, contributed to PhysioNet by Kevin Sweeney and colleagues at the National University of Ireland in Maynooth, contains examples of functional near-infrared spectroscopy (fNIRS) and electroencephalogram (EEG) recordings that have been created for evaluating artifact removal methods. In each such recording, one or two pairs of similar physiological signals have been acquired from transducers in close proximity. One signal of each pair is contaminated by motion artifact, documented in each case by simultaneously recorded outputs of 3-axis accelerometers affixed to each transducer.

This collection of 552 CTGs from the Czech Technical University (CTU) in Prague and the University Hospital in Brno (UHB) was carefully selected from 9164 recordings recently collected at UHB. Each recording is up to 90 minutes long, and includes a fetal heart rate time series, a uterine contraction signal, and maternal, delivery, and fetal clinical details.

A contribution of simulated fetal PCGs is now available in PhysioBank. The synthetic PCGs exhibit a realistic range of signal-to-noise ratios, with simulated maternal heart and organ sounds, fetal movements, and environmental sounds.