Abstract
Purpose. The electroretinogram (ERG) is the summed response from all levels of the retinal processing of light, and exhibits several profound nonlinearities in the underlying processing pathways. Accurate computational models of the ERG are important, both for understanding the multifold processes of light transduction to ecologically useful signals by the retina, and for their diagnostic capabilities for the identification and characterization of retinal disease mechanisms. There are, however, very few computational models of the ERG waveform, and none that account for the full extent of its features over time.
Methods. This study takes the neuroanalytic approach to modeling the ERG waveform, defined as a computational model based on the main features of the transmitter kinetics of the retinal neurons.
Results. The present neuroanalytic model of the human rod ERG is elaborated from the same general principles as that of Hood & Birch (1992), but incorporates the more recent understanding of the early nonlinear stages of ERG generation by Robson & Frishman (2014). As a result, it provides a substantially better match than previous models of rod responses in six different waveform features of the ERG flash intensity series on which the Hood & Birch model was based.
Conclusion. The neuroanalytic approach extends previous models of the component waves of the ERG, and can be structured to provide an accurate characterization of the full timecourse of the ERG waveform. The approach thus holds promise for advancing the theoretical understanding of the retinal kinetics of the light response.
