Abstract
Purpose: :
Our goal was to determine if a combination of percepts can be induced in the non–human primate by electrical stimulation of intracortical electrodes in area V1. In earlier experiments we demonstrated that 2–dimensional spatial information could be transferred to the primate cortex by training an animal to saccade to phosphenes. In this study we tried to use a combination of orientation cues to induce an integrated percept using simultaneous stimulation of multiple electrodes.
Methods: :
The receptive field (RF) locations were mapped and the orientation tuning measured for each of 96 intracortical electrodes chronically implanted in the right operculum of a rhesus monkey. Orientation tuning was measured with a field of randomly positioned line segments moving in directions {0, 22.5, <font face="symbol">¼</font>, 337.5°}. To test the theory that a summation of many orientation cues could form a coherent percept, ten channels were stimulated simultaneously, all of them contacting neurons tuned to similar orientations. In each trial, two groups of electrodes were stimulated in sequence, each group corresponding to a set of similarly oriented neurons. The animal was trained to indicate in a 2AFC task whether the progression of orientations was clockwise or counterclockwise.
Results: :
Visual response data (firing rates) for the 96 channels were fit with a 2D Gaussian to estimate the midpoint and the tuning width of the receptive field, and fits converged for 89 of the 96 electrodes. RF centers ranged from –9° to –2° horizontally and 1° to 5° vertically. Sixty–eight channels showed significant orientation tuning. We systematically explored the effects of multichannel stimulation by activating neurons tuned for similar orientations in a two–alternative forced–choice (2AFC) task. Despite attempts with various combinations of frequency, current and temporal pulse distribution, we were unable to train a monkey to report the sensation of orientation after 201,452 trials.
Conclusions: :
Although other researchers have demonstrated that it is possible to bias direction, speed, depth and orientation percepts with single–channel microstimulation, the construction of a de novo percept is more problematic than anticipated. This suggests that visual prosthesis strategies based on the assemblage of point percepts (phosphenes) may be the only viable approach to a first–generation prosthetic visual reconstruction in humans.