Abstract
Purpose :
We are studying the effect of asymmetric anodic-first stimulation on threshold reduction in Argus II retinal prosthesis patients, and in a retinal ganglion cell (RGC) model in silico. Threshold reduction is important because prior work has shown that about half of the electrodes have thresholds higher than the safety limit and need to be grouped with neighboring electrodes to allow safe stimulation, consequently reducing the stimulation precision.
Methods :
Perception probability in response to asymmetric-anodic first and symmetric cathodic-first stimuli was measured in 2 human subjects with the Argus II implant. 20 pulses of each type with 0.2 or 0.1 ms pulse width (duration of the cathodic phase) were delivered at 20 Hz and the subjects were asked to report if they perceived a phosphene. Our in silico experiment included a ball and stick model of a single RGC in the NEURON software environment and a point current source for electrical stimulation. Pulse width was varied between 0.05 – 1 ms, and duration ratio (anodic to cathodic phase) between 2 – 30 for asymmetric anodic-first pulses. Stimulation was delivered at 20 hz for a total duration of 1.5 seconds.
Results :
In the human subject testing we observed an average of 78% and 94% increase in perception probabilities with asymmetric anodic-first stimulation in the first and second subjects respectively, when compared to a standard pulse (symmetric cathodic-first). Our in silico results show that the RGC activation threshold generally decreases with asymmetric anodic-first stimulation, and this effect is stronger with higher duration ratios and shorter pulse widths. The model predicts that the highest reduction of threshold compared to the standard pulse was -53.21 % (Pulse width = 0.05 ms, duration ratio = 30). When pulse widths of 0.5 ms or greater were used, the model predicts less than 10% reduction in threshold.
Conclusions :
Our clinical and in silico results confirm that the asymmetric anodic-first stimulation reduces retinal activation thresholds, confirming our previous in vitro findings. Inclusion of this novel pulse shape in clinic may provide more flexibility in retinal stimulation and improve the overall outcome.
This is a 2020 ARVO Annual Meeting abstract.