Transient PERG extracted from Steady-State PERG in Glaucoma suspects

Abstract

Purpose : To use the Continuous Loop Averaging Deconvolution (CLAD) technique to extract a per-stimulus PERG (Transient) from steady state (SS) PERG at jittered high-frequency stimulation rates, and compare CLAD PERG data from a pool of glaucoma suspects (N=23) with those of age matched controls (N=16).

Methods : PERG were recorded in response to contrast-reversal of horizontal gratings (LED display JÖRVEC Corp, mean luminance 800 cd/m², contrast ratio 99%) presented under 3 conditions: Conventional transient PERG at 3.91 rev/sec, conventional SS-PERG at 15.63 rev/sec, and jittered ‘quasi’ steady state (Q-SS) PERG at 15.63 mean rev/sec. Q-SS responses were deconvolved using CLAD (PMID: 25477321; 24618216), extracting the equivalent per-stimulus transient response at 15.63 rev/sec, obtaining a response with peaks and troughs similar to a conventional transient PERG. For conventional and deconvolved transient responses, waveform peaks (N35, P50, N95) were identified and latencies and amplitudes were measured. For conventional SS-PERG responses, analysis was done in the frequency domain, by looking at the phase and magnitude of the response at the frequency bin corresponding to the rate of stimulation (15.63 rev/sec).

Results : In all three conditions, there was a significant difference between the control and glaucoma suspect groups (p<0.05), with the CLAD deconvolved response having the most significant difference between the two groups. Receiver operator curves (ROC) generated for the three cases resulted in area under the curve (AUC) values of 0.789 for conventional transient responses (P50 to N95 amplitude), 0.683 for conventional SS PERG (SS magnitude) and 0.897 for deconvolved Q-SS responses (P50 to N95 amplitude).

Conclusions : Results show that while all the methods were effective differentiating between control and glaucoma suspect groups, the CLAD deconvolved responses had the highest specificity, suggesting that the CLAD technique unveils some clinically relevant information that is obscured in the SS response when using conventional analysis techniques. This has the potential to greatly expand the diagnostic utility of the PERG response.

This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.

 

A) Average waveforms of the ‘quasi’ steady state (Q-SS) PERG responses and the respective average of deconvolved waveforms in glaucoma suspects and normal controls showing waveform peaks (N35, P50, N95). B) ROC curves of the transient PERG, steady-state (SS) PERG and the deconvolved responses.