The effects of random element loss on letter identification: implications for visual acuity loss in patients with retinitis pigmentosa.

The hypothesis that reductions in Snellen acuities in patients with retinitis pigmentosa are due solely to losses of photoreceptors was tested by measuring the effects of random losses of sampling elements on letter identification. Sampling element losses were mimicked by setting the luminance of randomly selected pixels equal to the luminance of the surround. The amount of pixel blanking ranged from 0 to 90%. Letters varying in retinal subtense from 5 to 17 min arc were presented for 500 msec. Although letter identification accuracy decreased with increasing pixel blanking for all letter sizes, performance remained relatively high even when a majority of the pixels was blanked. The data suggest that unless the loss of cone photoreceptors in greater than 80%, loss of sampling elements alone can not account for letter acuities poorer than 20/40. In addition to loss of cone photoreceptors in patients with RP, there are histological reports of photoreceptor abnormalities and psychophysical studies of visual sensory deficits. It is conceivable that these alone, or in combination with losses of photoreceptors, could account for decreased visual acuity. In a series of experiments, stimulus parameters were manipulated in order to mimic the effects of some of these abnormalities and deficits and the effects on letter identification were examined. The results of these experiments demonstrated that sampling element loss interacts with sensory factors (e.g. luminance and contrast sensitivity) and perceptual factors (e.g. set size and letter orientation) to reduce letter identification accuracy. The implication of these results is that decreases in letter acuity observed in patients with retinitis pigmentosa cannot be attributed solely to a random loss of sampling elements in the underlying retina, but may be due to the combination of photoreceptor degeneration and other sensory and perceptual factors.