The test-retest variability of the photopic negative response of the multifocal electroretinogram

Purpose. The slow sequence multifocal flash electroretinogram (mfERG) contains the photopic negative response (mfPhNR) which occurs in time after the b-wave and originates from ganglion cell activity. We assessed the test-retest variability of the mfPhNR of human control subjects and compared it to that of behavioral perimetric sensitivities and nerve fiber layer thickness (NFLT) measurements from the same test locations on the same subjects.;Methods. mfPhNRs were recorded from 61 control subjects (22–79yrs) using a VERIS system on two separate days. Stimuli consisted of seven 12 degree equal sized hexagons. Monitor frame rate was 70Hz with 30 frames per m-step and an m-sequence exponent of 9. The flash strength was 9.9 cd-s/m2. Perimetric testing was performed on two different days using the SITA Standard 24-2 test. NFLT measurements for a circular scan around the optic nerve head were made using spectral domain optical coherence tomography (OCT) from Heidelberg. We specifically examined the temporal section of the circular scan associated with the papillomacular bundle area because this had good spatial correspondence with our electroretinogram stimulus. The major variable used for assessing reliability was the normalized standard deviation (NSD) of the mfPhNR amplitude, which was computed as the standard deviation of the test retest differences divided by the mean of both trials across all subjects.;Results. The NSD of mfPhNR amplitude quantified from the preceding b-wave peak were smaller (by 50–66%) at each test location than the NSD of mfPhNR amplitudes quantified from baseline. The NSD for mfPhNR amplitudes from the preceding b-wave peak (5%–10%) were lower than the NSD of perimetric sensitivities(11%–20%) from the same test locations with p-values of 0.01 to 0.012 for six test locations and p-value of 0.24 for one test location. The NSD of the NFLT measurement from the temporal section of the circular scan was 7% while the NSD of the mfPhNR and perimetric sensitivity for the same location were 12% and 14% respectively.;Conclusions. The test-retest variability of the mfPhNR was comparable to if not better than estimates of perimetric sensitivity from the same test locations. The test-retest variability of spectral domain OCT was lower than both mfPhNR (F=2.94 p<0.0001) and perimetric sensitivity (F=4 p<0.0001) in the papillomacular bundle area. The mfPhNR provides an objective strategy for quickly and simultaneously measuring retinal ganglion cell function from multiple locations.