Relationship between stimulus size and different components of the electroretinogram (ERG) elicited by flashed stimuli

摘要

Purpose To investigate how light stimulus conditions of varying spatial sizes affect components of the flash and long-flash electroretinogram (ERG) in normal subjects. Method Three stimulus conditions were generated by a Ganzfeld stimulator: a white flash on white background (WoW), a red flash on a blue background (RoB) and an L+M-cone isolating on-off (long flash) stimulus (Cone Iso). ERGs were recorded from six subjects (5 M, 1 F) with DTL electrodes to full-field (FF), 70 degrees, 60 degrees, 50 degrees, 40 degrees, 30 degrees and 20 degrees diameter circular stimuli. Amplitudes and peak times for a-, b-, d- and i-wave, and PhNR were examined. PhNR amplitudes were estimated in two different ways: from baseline (fB) and from preceding b-wave peak (fP). Results With decreasing stimulus size, amplitudes for all ERG waveform components attenuated and peak times increased, although the effect varied across different components. An exponential fit described the relationship between amplitudes and size of stimulated retinal area well for most components and conditions (R-2= 0.75-0.99), except for PhNR(fB) (R-2= - 0.16-0.88). For peak times, an exponential decay function also fitted the data well (R-2= 0.81-0.97), except in a few cases where the exponential constant was too small and a linear regression function was applied instead (a-wave Cone Iso, b- and i-wave WoW). The exponential constants for RoB amplitudes (b-wave, PhNR(fB), PhNR(fP)) were larger compared to their counterparts under WoW (p < 0.05), while there was no difference between the constants for a-wave amplitudes and peak times and for PhNR peak times. The exponential constants of amplitudes vs. area under WoW and Cone Iso were remarkably similar, while under RoB PhNR(fB) showed larger constants compared to either a- or b-wave (p < 0.05). Conclusion ERG components change in a predictable way with stimulus size and spectral characteristics of the stimulus under these conditions. This predictability could allow a modified version of these sets of stimuli to be tested for clinical applicability.