Title :
An integral atrial wave identification based on spatiotemporal source separation: clinical validation
Author :
Castells, F. ; Ruiz, R. ; Rieta, J.J. ; Millet, J.
Author_Institution :
Bioeng., Electron. & Telemedicine Res. Group, Univ. Politecnica Valencia, Spain
Abstract :
This contribution presents a new methodology for the estimation of the atrial activity (AA) in supraventricular arrhythmias. The method is an adaptation of source separation techniques to this specific problem, including spatial and temporal information. An extensive clinical study over 45 patients demonstrates the suitability of the method. A spectral analysis of the estimated AA shows that the spectral concentration around the main peak is 13% in average higher than in the AA obtained with other source separation methods. Spectral differences between atrial fibrillation (AF) and atrial flutter (AFL) were also found: AF and AFL have main frequencies over 5 Hz and below 5 Hz respectively. In addition, AFL spectrum is characterized by the presence of harmonics.
Keywords :
blind source separation; electrocardiography; harmonics; medical signal processing; spatiotemporal phenomena; spectral analysis; atrial activity; atrial fibrillation; atrial flutter; harmonics; integral atrial wave identification; spatiotemporal source separation; spectral analysis; supraventricular arrhythmias; Additive white noise; Databases; Electrocardiography; Frequency; Gaussian noise; Independent component analysis; Principal component analysis; Source separation; Spatiotemporal phenomena; Statistical analysis;
Conference_Titel :
Computers in Cardiology, 2003
Print_ISBN :
0-7803-8170-X
DOI :
10.1109/CIC.2003.1291256