New experimental research into sight restoration for the blind has revealed that accounting for eye positions may be needed to create a more natural approach to visual perception using prosthetic devices for neurostimulation. Visual sensations, or phosphenes, in both the sighted and the blind are able to be induced and manipulated through electrostimulation of the visual cortex. These generated phosphenes are considered retina-centered. However, in the blind, even after years without sight, eye movements still affect the spatial localization of visual sensations generated in this manner.

In order to function well, cortical visual prostheses must be able to compensate for any localization errors caused by voluntary and involuntary eye movements. “There’s an inherent dissociation between [cortical] stimulation and eye movements,” noted the research team. “It’s unknown whether and how robustly the brain maintains the oculomotor circuitry in patients with bare or no light perception.”

They hypothesized that if electric stimulation of the medial occipital lobe results in retina-centered phosphenes, then the head-centered locations of the perceived visual sensations should correlate with eye position.

The study included three experiments for examining the perceived locations of phosphenes generated by visual cortical prostheses. Study subjects’ visual cortices were stimulated with a cortical visual prosthesis and their eye positions were recorded with timestamps. Each subject was asked to indicate the location of the perceived percepts with a handheld marker. Time between stimulation onset and response was measured. The researchers mapped the phosphenes between spatiotopic and retinotopic coordinates.

The researchers found that the locations of cortical-stimulated percepts changed based on the subject’s eye position at stimulation onset. They were able to remap measured responses based on measured eye positions to pinpoint retinal locations associated with the implanted electrodes.

“The ability to remap the locations perceived by stimulating various electrodes to corrected locations based on eye position, with variability reduced to pointing error, validates our hypothesis that stimulation of the selected cortical areas is indeed retinotopic,” the researchers noted.

“Mapping based on the eye position at the time of the stimulation and not on the time of action demonstrates that spatial localization of the phosphene occurs at the time of stimulation,” they continued. “This implies that in a future cortical visual prosthesis, corrections for eye movements should be done in real-time, based on the instantaneous eye position at the time of stimulation.” The team believes such a correlation will be critical so that the brain can integrate the visual information at the proper location in space.

“Our present results, that perceived locations in coordinates of the head-mounted camera can be mapped to defined retinal locations, demonstrate that eye scanning is likewise feasible with a cortical prosthesis that completely bypasses the eyes,” they concluded.