Steady-state PERG adaptation: a conspicuous component of response variability with clinical significance

Purpose: To investigate within-test variability of the steady-state PERG (SS-PERG).

Methods: SS-PERGs were recorded in response to black-white horizontal gratings (1.6 cycles/deg, 98% contrast, 15.63 reversals/s, LED display, 25 deg square field, 800 cd/sqm mean luminance) using skin electrodes. PERG and noise (± reference) signals were averaged over 1024 epochs (~ 2.2 min) and Fourier analyzed to retrieve SS-PERG amplitude and phase. SS-PERGs were split into 16 partial averages (samples) of 64 epochs each, and corresponding amplitudes and phases combined in polar coordinates to assess their dispersion (within-test variability). To assess time-dependent variability, samples were clustered in four successive time segments of ~ 33 s each. Amplitude adaptation was defined as amplitude difference between initial and final clusters, and PERG phase adaptation as the corresponding phase difference. To determine the dynamic range of SS-PERG adaptation, recording was performed in normal controls of different age (n = 32) and patients with different severity of optic nerve dysfunction (early manifest glaucoma, EMG, n = 7; non-arteritic ischemic optic neuropathy, NAION, n = 5).

Results: Amplitude adaptation was largest in younger controls (amplitude adaptation ÷ noise, SNR = 9.5, 95% CI 13.1, 5.9) and progressively decreased with increasing age (older subjects, SNR = 5.5, 95% CI 9.2, 1.8) and presence of disease (EMG: SNR = 2.4, 95% CI 3.5, 1.4; NAION: SNR = 1.9, 95% CI 6.5,-2.2). In 11 young subjects, amplitude adaptation was repeatable (test-retest in two sessions a week apart; intraclass correlation coefficient = 0.59). Phase adaptation was not significantly different from zero in all groups.

Conclusions: SS-PERG adaptation accounts for a sizeable portion of the within-test variability. As it has robust SNR, sufficient test-retest variability, and is altered in disease, it may have physiological and clinical significance. This study suggests that SS-PERG protocols should include adaptation in addition to SS-PERG amplitude and phase/latency.