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. In order to develop more effective stimulation methods so as to improve the quality of elicited percepts, we are studying how retinal neurons respond to electric stimulation. In a previous study, we determined that bipolar cells were more sensitive to lower frequencies (5-25 Hz) of sinusoidal stimulation than to higher frequencies (i.e. 100 Hz). Here, we studied the factors that influence this sensitivity difference.
Methods: :
The model bipolar cell was traced from a rat rod bipolar cell and constructed in NEURON. A simplified two compartment model was also developed in Matlab. Both models were first implemented with a passive membrane to explore the contribution of morphology to the frequency response. Then, T- and L-type calcium channels were incorporated to determine their influence on the frequency response.
Results: :
We measured how the voltage in the bipolar cell axon terminal changed as a function of stimulus frequency. In both passive models the frequency response was relatively flat until 794 Hz. Changes to all but one parameter had only a modest effect on the frequency response. Increasing the intra-axonal resistance by a factor of 8 decreased the cutoff frequency to 117Hz. Conversely, decreasing the intra-axonal resistance caused the cutoff frequency to increase drastically. It was not possible in either passive model to duplicate the 25 Hz frequency sensitivity that was observed physiologically. When Ca channels were introduced to the models, the peak frequency response approached those observed physiologically: T-type channels exhibited a bandpass response with peak frequencies ranging from 4-24 Hz. For L-type channels, the response profile was low-pass or bandpass (depending on stimulus amplitude) with cutoff frequencies ranging from 65-350Hz.
Conclusions: :
The sensitivity of bipolar cells to low frequency stimulation observed physiologically is likely mediated by the kinetics of Ca channels. This suggests that the ion channels endogenous to a given neuron contribute significantly to the cell’s response to stimulation. Knowledge of the distribution and kinetics of ion channels within each neuron may therefore lead to improved methods of preferential activation (targeting of specific neuronal classes).
Keywords: bipolar cells • computational modeling • electrophysiology: non-clinical