Abstract

Several years ago, Stork and Falk¹ introduced a technique to convert flicker sensitivity data into temporal impulse-response functions (IRF). A slightly modified procedure was later suggested by Swanson.² When this technique is applied to data collected foveally and at photopic brightness levels from normal observers, a plausible, biphasic waveform for the IRF is obtained. Lower flicker sensitivities, however, such as those existing in ophthalmology patients at extrafoveal retinal locations or at low luminance levels, may yield IRF waveforms with a very steep initial flank and late oscillations. The steep initial flank is caused by the first-order extrapolation beyond the measured critical flicker frequency. Late oscillations can be attributed to the assumed flat sensitivity characteristic below the lowest frequency tested. Using the data and models of Kelly, Watson, Roufs, and others, as well as data collected from 20 normal observers and 100 retinitis pigmentosa (RP) patients, I have modified the Stork-Falk procedure to eliminate the influence of sensitivity scaling in normal observers and to yield results that are more plausible in view of our understanding of changes in RP photoreceptors. The choices of high- and low-frequency extrapolations and the applicability of the minimum-phase assumption implicit in the procedure will be discussed.

© 1990 Optical Society of America