Abstract
Abstract: :
Purpose: The pattern–pulse multifocal visual evoked potential (James, 2003 IOVS) is a method for the rapid concurrent testing of multiple visual field locations by presentation of temporally sparse pulses of pattern contrast. Compared to the more commonly used rapid contrast reversal, it appears to maintain the visual system in a high contrast–gain state, producing significantly larger response amplitudes. However because sparse stimuli involve fewer stimulus in a given time, variance of the estimated response can also be greater. We here study how this trade–off affects the signal to noise ratio (SNR) of the two methods. Methods: Stimuli were presented monocularly in 56 regions of a cortically–scaled dartboard of total diameter 48o at 75Hz frame rate. A 4x4 checkerboard pattern was either pulsed in contrast independently in each region at mean rate 1, 2, 4 or 9 pulses/s, or was contrast reversed at mean rate 38 or 19 reversals/s. Evoked potentials were recorded from 25 subjects, 3cm above vs 4cm below inion, for 4 runs of 55s for each stimulus condition. Elementary response waveforms were estimated for each region and quantified by mean square signal from 50–120ms post–stimulus, with noise quantified by mean square from 400–800ms, beyond the response window. The multiplicative effects of stimulus parameters were estimated by multiple linear regressions of decibel (dB) signal, noise and SNR power on stimulus and subject factors. Results: As mean stimulus rate varied from 1 to 38 events/s, signal power decreased by 17.1dB (+/–0.4SE), a factor of 7 times in RMS amplitude, while noise power decreased by 16.1dB (+/–0.3SE). The signal to noise ratio however rose between mean rates 1 and 4 pulses/s, then fell monotonically, dropping by 3.0dB (+/–0.4SE) between 4 pulses/s and the 38 reversals/s condition (highly significant, p<.001). Correcting for noise power included within the apparent signal power, this corresponds to a factor of at least 2 and up to 4 times more recording duration to obtain a criterion SNR. Conclusions: Response amplitude drops as stimulus rate increases, which more than cancels out the advantage of a greater number of presentations beyond an optimal rate, around 4Hz. That rate will require a recording duration ½ to ¼ as long as contrast reversing stimulation to achieve clinical utility.
Keywords: electrophysiology: clinical • visual cortex • visual fields