Abstract
Purpose: :
Visual function cannot be directly measured in preclinical animal models. The multifocal electroretinogram (mfERG) is an emerging tool to assess drug efficacy and toxicity at the retina. In animal models of age-related macular degeneration (AMD), mfERG trace analysis is challenging due to smaller amplitudes, peak delays, low signal-to-noise ratios and motion artifacts. In this study, 4 mfERG trace analysis methods were compared to optimize retinal status determination in a rat laser AMD model.
Methods: :
Choroidal neovascularization in the AMD model was induced with 8 radial laser spots at a constant distance from the optic disk in Brown Norway rats and compared to unlasered controls (n=6/group). mfERG analysis was performed 4 wk after induction, using contact lens electrodes and a 127 non-scaled hexagonal pattern stimuli. Four trace analysis methods were employed: 1) An image processing program determined the numbers and % total affected hexagon area in a 3D display. 2) Overall amplitude obtained by averaging each N1-P1 hexagonal response per trace. 3) Ring ratios obtained by averaging ring N1-P1 amplitude expressed as a ratio of central ring response. 4) A template-scaling method developed by Hood and Li where both amplitude and implicit time are scaled to baseline traces using a least-squares fitting procedure. The goodness-of-fit was expressed as Statfit, which relates the scaled response to the mean. A normal Statfit value was 0-0.4 while those ≥ 0.7 were considered abnormal.
Results: :
The % total affected hexagon area measurement suggested a difference between lasered and control animals (11.3+5.4 % vs 6.3+3.4 %, p=0.06).Overall amplitude in AMD eyes was 1.1±0.1 µV not significantly different from the control (1.3±0.3 µV, p=0.4). Similarly, ring ratio analysis of AMD eyes was 2.1±0.2, not significantly different from control (2.4±0.3, p=0.3). Template scaling analysis showed two distinct response clusters. Statfit response in the lasered eyes was significantly greater than control (0.6±0.2 vs 0.2±0.1, p<0.001).
Conclusions: :
The 3D image analysis is based on a relative value grading scheme with values not concretely related to underlying retinal pathology. In the overall amplitude and ring ratio analyses, direct amplitude comparison does not account for motion artifacts, low signal-to-noise ratios or delayed peak response. These analysis methods are, therefore, less desirable. The template-scaling method is preferred, because it incorporates amplitudes and implicit time and also accounts for inter-animal variability.
Keywords: age-related macular degeneration • electroretinography: non-clinical • retina