Dependence of the pattern-ERG on variations of the stimulus pattern

1. In order to study the relative contributions of luminance and pattern driven mechanisms to the pattern electroretinogram (PERG) we varied the spatial frequency and spatial profile of the PERG stimulus waveform. 2. We hypothesized that high-pass spatially filtered stimulus waveforms ('Craik waves'), which preserve the edge contours, but suppress the luminance modulation of a square wave, would selectively stimulate contrast-specific generators of the PERG, while low-pass filtered stimulus waveforms ('smooth waves'), which suppress the edge contours of a square wave, would selectively stimulate luminance-dependant PERG generators. 3. PERG recordings at several spatial frequencies revealed little difference in the PERGs elicited by Craik wave or smooth wave stimuli matched for luminance modulation. The response to a square wave stimulus equal to the sum of the Craik wave and smoothwave stimuli was roughly predicted by the vector sum of the separate responses to the smooth wave and Craik wave stimuli. 4. A consistent effect of stimulus spatial frequency on the phase lag of the second-harmonic response component was seen for all stimulus waveforms, with a 90 deg increase in phase lag occurring as the stimulus spatial frequency increased from 0 to 1.0 c/deg. Over this range, there was little dependence of response amplitude on stimulus spatial frequency. 5. Control experiments using stimuli with artificially reduced contrast or simulated optical blur yielded marked reductions in PERG response amplitude, with small phase changes opposite to those seen with increasing spatial frequency, ruling out optical degradation as the cause of the phase effect. 6. We are thus able to demonstrate both pattern- and luminance-dependent non-linear mechanisms underlying the PERG by means of manipulations of the spatial stimulus waveform.