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. The brightness and contrast of normal vision is mediated in part by the frequency with which retinal ganglion cells generate action potentials. There is conflicting evidence however as to the ability of high-frequency pulse trains to generate high-frequency spike trains. Here, we report on the response of five different types of retinal ganglion cells to high-frequency pulse trains.
Methods: :
Cell-attached patch clamping was used to record spikes from rabbit retinal GCs in the isolated rabbit retina. Stimuli consisted of pulse trains with rates ranging from 100-300 pulses per second (PPS); the duration of each pulse was 0.2 ms. The stimulating electrode was a small (10kΩ) stimulating electrode placed 25µm above the ganglion cell.
Results: :
At 100 PPS, each of the five ganglion cell types that we measured reliably elicited a single action potential in response to each cathodal pulse in the train. At higher rates, only a percentage of pulses elicited standard action potentials; the response to the other pulses in the train was much smaller (amplitude) than the typical spike. Interestingly, all pulses elicited either a spike or the small response - never both. In addition, the number of total responses (big plus small) was equal to the total number of pulses. Latency analysis revealed that the timing of the small response was slightly faster than that of the typical spike.
Conclusions: :
The amplitude and kinetics of the small waveform are consistent with that of an action potential initiated within the initial segment of the axon and back-propagated to the soma. Since it is likely that all action potentials are initiated in the initial segment, large action potentials represent those cases in which the spike initiated in the initial segment is converted into a somatic action potential. Therefore, the small waveform represents the case where the spike initiated in the initial segment is not converted into an action potential. Thus, our results suggest that all ganglion cell types reliably elicit an action potential for each pulse but the efficacy with which band initiated spikes are converted into somatic spikes varies across cell types.
Keywords: ganglion cells • ion channels • electrophysiology: non-clinical