Abstract
Purpose: :
This work is related to the efforts of the Boston Retinal Implant Project to develop a sub-retinal prosthesis to restore vision to the blind. As part of that effort, we want to develop more effective stimulation methods so as to improve the quality of elicited percepts and are studying how retinal neurons respond to electric stimulation. Here, we are studying how the properties of a dense band of sodium channels, located in the proximal axon of ganglion cells, influence the response to electric stimulation.
Methods: :
We developed a morphologically realistic multi-compartment retinal ganglion cell model in NEURON. The properties of individual compartments were based on previous work (Fohlmeister-Coleman-Miller) but modified to incorporate new anatomical findings. Stimulation was from a point source fixed within a plane 25 µm above the neuron.
Results: :
Simulated maps of threshold vs. stimulating electrode position were comparable to maps generated during physiological experiments. Threshold variations across different elements of the neuron were comparable to previous simulations, but the model revealed that thresholds were lowest at the distal edge of the band. Predictably, the site of spike initiation was generally at the distal edge of the band; however, the site of spike initiation moved closer to the stimulating electrode with increasing stimulation amplitudes. Similar to physiological experiments, minimum thresholds varied with changes in band properties, however, the model revealed that long bands far from the soma had the lowest thresholds. The range of thresholds arising from wide variations in band properties was limited to a factor of 2.
Conclusions: :
The model accurately reproduces many existing physiological results and confirms that the properties of the sodium channel band modulate the response to stimulation. The difference in threshold across types was smaller than those measured physiologically suggesting that factors other than band also influence threshold.
Keywords: ganglion cells • computational modeling • electrophysiology: non-clinical