Abstract
Purpose :
The real-time stimulation of electrodes implanted in the visual pathway using a head-mounted camera creates visual perception through the activation of neurons at a close distance of the electrodes. Transferring camera frames to electrical pulses that create a meaningful and uniform percept requires understanding the relationship between the stimulation parameters and different percept properties such as brightness, location, shape, and size. Variable brightness across electrodes would cause images to look "mottled;" variable size could affect blurriness. Here, we studied the relative brightness and size of the percepts created by different Argus II epiretinal electrodes.
Methods :
Three people with Argus II implant participated in this study. We tested 10 and 5 electrodes for S1 and S2, respectively, and three groups of 4, 5, and 10 electrodes for S3. The relative brightness and size of the percepts were studied in a 2-AFC task with all pair combinations in the group repeated three times in random order. In each trial, a pair of electrodes was stimulated, one by one, at an amplitude level corresponding to a 0.75 detection rate specific to the given electrode (previously estimated using a Bayesian adaptive threshold test). The participant was asked to choose which electrode in the pair was brighter regardless of the size and which was bigger regardless of brightness. Electrodes were ranked from 0 to 1 based on the percentage of trials in which they were judged brighter or bigger, and ranks were plotted against threshold; the average rank is expected to be 0.5, and the standard deviation (SD) represents uncertainty.
Results :
Brightness ranks of all electrodes had a mean of 0.50 and a SD of 0.20 (range: 0.13 - 0.88). The average size rank was 0.50 with a SD of 0.17 (range: 0.07 - 0.81). Across electrodes, brightness ranks show a linear correlation with threshold (r = 0.45; p = 0.007), but the correlation between the size ranks and thresholds was not significant (r = 0.27; p = 0.1).
Conclusions :
We showed phosphenes size is not affected by threshold differences. However, electrodes with higher thresholds do create brighter percepts. Therefore, uneven brightness may be corrected by scaling stimulus amplitudes according to the threshold, but it is unlikely that the image blur can be reduced. This should be considered when normalizing brightness and size of the percept elicited by implanted electrodes.
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.