July 2018
Volume 59, Issue 9
Open Access
ARVO Annual Meeting Abstract  |   July 2018
Modelling the contrast transducer for intense perimetric stimuli
Author Affiliations & Notes
  • Andrew J Anderson
    University of Melbourne, Parkville, Victoria, Australia
  • Shima Rashidi
    University of Melbourne, Parkville, Victoria, Australia
  • Andrew Turpin
    University of Melbourne, Parkville, Victoria, Australia
  • Footnotes
    Commercial Relationships   Andrew Anderson, None; Shima Rashidi, None; Andrew Turpin, Heidelberg Engineering GmBH (F)
  • Footnotes
    Support  Australian Research Council Future Fellowship FT120100407 (AJA)
Investigative Ophthalmology & Visual Science July 2018, Vol.59, 5117. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Andrew J Anderson, Shima Rashidi, Andrew Turpin; Modelling the contrast transducer for intense perimetric stimuli. Invest. Ophthalmol. Vis. Sci. 2018;59(9):5117.

      Download citation file:

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

  • Supplements

Purpose : To examine whether a physiologically plausible model for normal perimetric thresholds can also predict the normal visual system’s ability to discriminate between intense perimetric targets of the sort typically presented in areas of glaucomatous visual field loss.

Methods : The responses of the most active cortical neuron from a two-stage neuronal spiking model (Vision Research 2008;48:1859-1869) were determined for Humphrey Field Analyzer perimetric stimuli ranging from 40dB to 0dB, in 1dB steps. Neurometric functions relating the difference in spiking rate for two stimuli of differing intensities to the probability of detecting this difference were then generated for the lowest intensity stimulus pairs (a 27dB stimulus, and a slightly more intense comparator) reported in the two-alternative force choice suprathreshold contrast discrimination data of Anderson et al. (Invest Ophthalmol Vis Sci 2016;57:6397-6404) for four healthy observers. Only data from 9 degrees eccentricity was used, to match the eccentricity used in the model. Based on these neurometric functions, the probability of discriminating between the stimuli in the highest intensity pairs tested by Anderson et al. was then predicted. For comparison, we also examined the performance of a simple linear transducer, where neural firing increased linearly from a baseline rate of 9 impulses/s for a 35 dB, to maximum of 64 impulses/s for a 0 dB stimulus.

Results : The output of the two-stage model showed a saturating rate of firing as stimulus intensity increased. The two-stage model predicted near chance discrimination performance for intense stimulus pairs (ave 56%: range 50 to 61%) compared to the average performance of 94% (range: 83 to 100%) seen empirically. The simple linear model predicted average performance of 98% (94 to 100%). A likelihood analysis showed the linear model to be more likely (>12 log units) for all observers.

Conclusions : An unmodified two-stage model for normal perimetric detection cannot satisfactorily predict discrimination responses for intense perimetric targets. Models showing more linear contrast transducer properties are likely to be more satisfactory.

This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.


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.