• DocumentCode
    2317637
  • Title

    Separating the atrial and ventricular components in atrial fibrillation. Are 64 leads better than 12?

  • Author

    Haigh, AJ ; Murray, A. ; Langley, P.

  • Author_Institution
    Newcastle Univ., Newcastle upon Tyne
  • fYear
    2007
  • fDate
    Sept. 30 2007-Oct. 3 2007
  • Firstpage
    281
  • Lastpage
    284
  • Abstract
    It is important to be able to separate the atrial activity from the dominant ventricular activity of the body surface ECG signals in atrial fibrillation so that the effect of therapies such as drug cardioversion and ablation can be adequately assessed. Principal Component Analysis (PCA) and Independent Component Analysis (ICA) are algorithms which can be used but their application has been limited to standard 12-lead ECG recordings. Our aim was to investigate whether better separation of the atrial and ventricular activities could be achieved using additional leads. We applied PCA and ICA to ECG recordings of 12 and 64 leads in 5 patients in atrial fibrillation. We measured and compared between the 12 lead and 64 lead analyses the amplitude of residual ventricular QRS which remained in the separated atrial activity and the dominant fibrillatory frequency. We found that there was little difference between 12 and 64 leads in separating the atrial and ventricular activity.
  • Keywords
    electrocardiography; independent component analysis; medical signal processing; principal component analysis; 12 leads; 64 leads; atrial components; atrial fibrillation; body surface ECG signals; dominant ventricular activity; independent component analysis; principal component analysis; residual ventricular QRS; ventricular components; Atrial fibrillation; Cardiology; Drugs; Electrocardiography; Electrodes; Frequency; Independent component analysis; Medical treatment; Principal component analysis; Signal analysis;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computers in Cardiology, 2007
  • Conference_Location
    Durham, NC
  • ISSN
    0276-6547
  • Print_ISBN
    978-1-4244-2533-4
  • Electronic_ISBN
    0276-6547
  • Type

    conf

  • DOI
    10.1109/CIC.2007.4745476
  • Filename
    4745476