December 2002
Volume 43, Issue 13
ARVO Annual Meeting Abstract  |   December 2002
Relating Ganglion Cell Loss To Perimetric Defects: A Neural Model
Author Affiliations & Notes
  • WH Swanson
    Glaucoma Institute SUNY State College of Optometry New York NY
  • MW Dul
    Glaucoma Institute SUNY State College of Optometry New York NY
  • F Pan
    Glaucoma Institute SUNY State College of Optometry New York NY
  • Footnotes
    Commercial Relationships   W.H. Swanson, None; M.W. Dul, None; F. Pan, None. Grant Identification: NIH Grant EY07716, Novartis Ophthalmics
Investigative Ophthalmology & Visual Science December 2002, Vol.43, 2123. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      WH Swanson, MW Dul, F Pan; Relating Ganglion Cell Loss To Perimetric Defects: A Neural Model . Invest. Ophthalmol. Vis. Sci. 2002;43(13):2123.

      Download citation file:

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

  • Supplements

Abstract: : Purpose: To develop a quantitative neural model relating perimetric sensitivity to ganglion cell damage in patients with glaucoma. A related effort for normal subjects correlated perimetric sensitivity and effective ganglion cell density across eccentricity-dependent variations1, but normal effects of eccentricity can be characterized by variation in magnification factor2 and may not apply to decreased cell density in glaucomatous defects. Our neural model distinguishes between cell density and magnification factor. Methods: The model was implemented two stages: 1) sensitivities of ganglion cells in a mosaic were computed for standard perimetric stimuli, and 2) sensitivities of cortical populations were computed for inputs from the ganglion cell mosaics. The effects of ganglion cell damage were simulated by degrading mosaics in the first stage, and the effects of variation in magnification factor were simulated by varying peak spatial frequency of cortical populations in the second stage. Both localized and diffuse patterns of ganglion cell damage were simulated for a wide range of spatial scales. Results: With just two parameters (sensitivity and magnification) the neural model describes a wide range of data in the perimetric literature. For normal ganglion cell densities, the neural model fits spatial summation functions on log-log axes with a region of linear summation (slope of 1.0) below the magnification parameter and a region of probability summation (slope of 0.25) above it. For both diffuse and localized ganglion cell loss, the neural model predicts two key aspects of glaucomatous defects: abnormal spatial summation and increased test-retest variability for smaller stimuli. For large stimuli , if ganglion cell loss is less than 50% then defects of -6 dB or more reflect widespread cellular dysfunction in addition to cell death. Conclusions: A two-stage neural model for diffuse and localized damage of primate ganglion cells captures key features of the perimetric literature on effects of stimulus size in normal and glaucomatous eyes, and provides guidelines for distinguishing cell death from cell dysfunction. 1Garway-Heath et al. (2000) IOVS 41: 1774-1782; 2Latham, et al. (1993), IOVS 34:1691-1701

Keywords: 511 perimetry • 415 ganglion cells • 586 spatial vision 

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.