Evaluating the repeatability of the multifocal electroretinogram in the cone dominant 13-lined ground squirrel

To assess the repeatability of the multifocal electroretinogram (mfERG) in the 13-lined ground squirrel (13-LGS).

13-LGS (Ictidomys tridecemlineatus, 2 M/4F, n = 12 eyes) were equally divided between 61-segment or 103-segment protocols containing two consecutive mfERG scans per session, and two total sessions performed one week apart per eye. The trough-to-peak amplitudes were analyzed under three conditions: raw, normalized to the optic nerve head (ONH), and normalized to the visual streak (VS). For intrasession repeatability, the two consecutive scans within each session were analyzed. For intersession repeatability analyses, one scan was randomly chosen from each day.

Intrasession repeatability of the mfERG trough-to-peak amplitude data averaged at 63% (Raw), 55% (ONH), and 50% (VS). There was no significant difference in repeatability between each day's intrasession repeatability values for all normalization conditions (Raw: Wilcoxon t-test, p = 0.2334; ONH: Paired t-test, p = 0.7803; VS: Wilcoxon t-test, p = 0.3804). Intersession percent repeatability of mfERG trough-to-peak amplitude data averaged at 72% (Raw), 61% (ONH), and 53% (VS). There was a statistically significant difference between the groups (Friedman test, p = 0.0038). This was evident in the Benjamini-Hochberg method of controlling the false discovery rate (FDR) where there was a significant difference comparing Raw versus VS (p = 0.0130) and ONH versus VS (p = 0.0011). There was no difference comparing Raw versus ONH (p = 0.1076).

Overall intrasession and intersession repeatability of mfERG amplitude was relatively poor in our sample, though not markedly different than that reported in some other species and normalization methods did result in improved repeatability. As animal models are critical for vision research, these repeatability estimates will prove useful in interpreting future data collected following interventions or in longitudinal monitoring of disease models.