Visual Cortical Plasticity and the Implications for Sight Restoration Technologies

Abstract

Degenerative eye diseases are a leading cause of blindness in adults. While only limited treatment options for diseases such as age related macular degeneration and retinitis pigmentosa exist, there have been a variety of technologies developed that attempt to restore some visual functioning. These include electronic prostheses (either cortical or retinal), optogenetics, and gene therapy. However, to date, only retinal prosthetic devices have been implanted in patients, with cortical implant and optogenetic clinical trials still under way. These technologies are not likely to completely recreate normal vision, but rather provide essential visual input that can improve everyday functioning in patients. In the current state of sight restoration technology, these devices cause early on- and off-center retinal cells to fire simultaneously, rather than in a biologically complementary fashion (i.e. when on-cells fire, off-cells in the same location are suppressed). This creates deeply unnatural population responses that propagate from the retina to cortex. The central question of this thesis is whether patients have the potential to access cortical plasticity in adulthood that allows them to decode unnatural cell population responses in early visual processing caused by electronic prosthesis technologies. In Chapter 1, sighted participants were dichoptically presented with a combination of original and contrast-reversed images that resulted in a similar decoding challenge caused by simultaneous stimulation of on- and off-cells. Each image (I) and its contrast-reverse (Iʹ) was filtered using a radial checkerboard (F) in Fourier space and its inverse (Fʹ). [I * F′] + [Iʹ * F] was presented to one eye, and [I * F] + [Iʹ * F′] was presented to the other, such that regions of the image that produced on-center responses in one eye produced off-center responses in the other eye, and vice versa. We show that participants continuously improved in a naturalistic object discrimination task over 20 one-hour sessions using the distorted input. Pre- and post-tests suggest that performance improvements were due to two learning processes: learning to recognize objects with reduced visual information and learning to suppress contrast-reversed image information in a non–eye-selective manner In Chapter 2, we ask whether video game training provides any additional learning benefit than that witnessed previously, and examine whether learning generalizes outside of the trained conditions. While we did not find a large degree of transfer to novel stimuli as a result of video game training, we provide additional insights into mechanisms that explain learning on the task. Collectively, these studies examine the potential of human participants to learn to decode unnatural neural population responses in adulthood. Learning to decode these distorted stimuli will provide developers of retinal prosthesis technologies with a realistic, empirical expectation of the extent to which patients can learn to make use of their implant.