April 2011
Volume 52, Issue 14
Free
ARVO Annual Meeting Abstract  |   April 2011
Visualization of Fundus Vessel Pulsation Using Principal Component Analysis
Author Affiliations & Notes
  • Fabrice Moret
    Sect. Visual Function and Electrophysiology,
    Eye Hospital, University of Freiburg, Freiburg, Germany
  • Wolf A. Lagrèze
    Sect. Neuroophthalmology,
    Eye Hospital, University of Freiburg, Freiburg, Germany
  • Charlotte M. Poloschek
    Sect. Neuroophthalmology,
    Eye Hospital, University of Freiburg, Freiburg, Germany
  • Michael Bach
    Sect. Visual Function and Electrophysiology,
    Eye Hospital, University of Freiburg, Freiburg, Germany
  • Footnotes
    Commercial Relationships  Fabrice Moret, None; Wolf A. Lagrèze, None; Charlotte M. Poloschek, None; Michael Bach, None
  • Footnotes
    Support  Deutsche Forschungsgemeinschaft BA877/19-2
Investigative Ophthalmology & Visual Science April 2011, Vol.52, 1339. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Fabrice Moret, Wolf A. Lagrèze, Charlotte M. Poloschek, Michael Bach; Visualization of Fundus Vessel Pulsation Using Principal Component Analysis. Invest. Ophthalmol. Vis. Sci. 2011;52(14):1339.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose: : Vascular pulsations such as spontaneous venous pulsations or other fainter pulsatile movements are difficult to observe because of eye movements. Recording of fundus movies and aligning (registering) the images addresses slow drifts but the images still contain distracting microsaccadic distortions and noise. We propose a method to overcome these issues, allowing observation of minute retinal pulsations.

Methods: : We analyzed 5-second image sequences recorded with a near-infrared SLO (Spectralis HRA+OCT) in ten volunteers. The images were first registered, microsaccade-distorted images were then manually rejected, and the remaining image sequences decomposed using a Principal Component Analysis (PCA), a form of blind source analysis. A movie was then constructed based on the first five principal components (these had pulsatile features).

Results: : The detection of pulsatile features was improved through each of the processing steps (registration, cleaning, PCA filtering). Comparison of the raw input images and the PCA outcome indicated that the automatic steps did not introduce any artifacts of their own. After PCA processing, we obtained clear visualization of spontaneous venous pulsation in five of ten subjects - previous work reports that 80%-90% of the population displays spontaneous venous pulsation. Arterial pulsation could be observed in nine of ten subjects, together with additional features such as pulsation of arterioles down to 70 µm diameter, complete venous collapse, overall optic nerve head tissue pulsation, and moving mechanical links between veins and arteries.

Conclusions: : Disentangling pulsatile motion from other dynamic components of retinal images yields an unprecedented resolution in physiologic motion of the retinal vessel structure.

Keywords: imaging/image analysis: non-clinical • retina • optic nerve 
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×