Abstract
Purpose: :
To test the hypothesis that psychophysically derived estimates of retinal ganglion cell density (RGCd) correlate with perimetric increment contrast sensitivity (CS) in a linear manner.
Methods: :
Previous research suggests that peripheral, high–contrast resolution acuity (PRA) is limited by a sampling process with density almost identical to RGCd. Consequently, RGCd may be estimated through the use of a PRA task. PRA was assessed at seven eccentricities from 9.5 – 70 degrees in the temporal visual field of six young, (24 – 45 years of age) normal subjects. Stimuli were vertical sinusoids drifting randomly to the right or left windowed through a circular aperture with Gaussian edge blur. The subjects’ task was to make a forced choice of the direction of drift while the spatial frequency was altered from trial to trial under the control of a ZEST procedure, which terminated after 40 trials. Subjects’ also performed a perimetric detection task at the same seven locations using a custom field, the full threshold algorithm and stimulus size five on a Humphrey perimeter. The decibel perimetric thresholds were converted to Weberian CS, (i.e. from logarithmic to linear units). All measures were repeated two to four times and testing for each subject took about two hours. Estimates of RGCd established from PRA were calculated for each location using an assumption of square packing. The RGCd was then multiplied by the area of the size five stimulus to determine how many RGCs fell within the area of the perimetric stimulus at each location. CS was plotted in linear units against the number of RGCs within the stimulus area.
Results: :
As expected, CS increased as the number of RGCs within the stimulus area increased (i.e. as eccentricity decreased). However, this relationship was not linear and was in the form of an asymptotic exponential. When CS was plotted against the log of the number of RGCs within the stimulus area, the relationship was linearized with r2 greater than 0.9 for all subjects. In addition, the slope of this function was quite variable between subjects and varied by a factor of three from shallowest to steepest.
Conclusions: :
The relationship between the number of RGCs within the stimulus and the perimetric increment contrast sensitivity is non–linear and quite individualized in human observers. This finding suggests that the link between the number of RGCs stimulated and the functional outcome may be only loosely defined within the population and the concept of a normal structure–function relationship may be difficult to apply when expressed in terms of population limits.
Keywords: contrast sensitivity • perimetry • visual fields