Electrical stimulation of retinal neurons is known to elicit visual sensations. When applied to the retina in a spatial pattern, electrical stimulation may be capable of providing rudimentary patterned vision that may be of benefit to sufferers of degenerative retinal disorders. No such device has yet been devised to provide for chronic study of the psychophysical perceptions elicited from a prosthesis for retinal stimulation. In this study, steps towards achieving this goal have been successfully carried out. Foregoing research was reviewed such that appropriate stimulation parameters were incorporated in the design of a 100 stimulation channel, complimentary metal oxide semiconductor (CMOS) integrated circuit, small enough in size so as to be capable of being implanted within the ocular anatomy or surrounding orbit. The device, and its associated external hardware and software were designed, modeled, fabricated, and interfaced with stimulating electrodes in acute testing in a highorder mammal (Ovis aries) so as to assess the capabilities of the device to elicit cortical potentials as a direct result of stimulation of the neural retina. Testing was performed under conditions similar to those anticipated in chronic in-situ configurations wherein radio-frequency telemetry was used to deliver power and configuration parameters to the device thus avoiding the passage of wires through tissue in order to communicate to the implant circuit. The results of the testing indicate that the circuit is indeed capable of eliciting physiological responses in the animal and evidence is present that these responses could be elicited in patterned form. Further work undertaken includes the development of surgical methods for implantation, and application of the prosthesis circuit in functional electronic stimulation.

Electrical stimulation of retinal neurons is known to elicit visual sensations. When applied to the retina in a spatial pattern, electrical stimulation may be capable of providing rudimentary patterned vision that may be of benefit to sufferers of degenerative retinal disorders. No such device has yet been devised to provide for chronic study of the psychophysical perceptions elicited from a prosthesis for retinal stimulation. In this study, steps towards achieving this goal have been successfully carried out. Foregoing research was reviewed such that appropriate stimulation parameters were incorporated in the design of a 100 stimulation channel, complimentary metal oxide semiconductor (CMOS) integrated circuit, small enough in size so as to be capable of being implanted within the ocular anatomy or surrounding orbit. The device, and its associated external hardware and software were designed, modeled, fabricated, and interfaced with stimulating electrodes in acute testing in a highorder mammal (Ovis aries) so as to assess the capabilities of the device to elicit cortical potentials as a direct result of stimulation of the neural retina. Testing was performed under conditions similar to those anticipated in chronic in-situ configurations wherein radio-frequency telemetry was used to deliver power and configuration parameters to the device thus avoiding the passage of wires through tissue in order to communicate to the implant circuit. The results of the testing indicate that the circuit is indeed capable of eliciting physiological responses in the animal and evidence is present that these responses could be elicited in patterned form. Further work undertaken includes the development of surgical methods for implantation, and application of the prosthesis circuit in functional electronic stimulation.

Relatively new technology called optical coherence tomography allows direct and non invasive in vivo imaging of retinal anatomy in human subjects. There are several interesting applications of this technique, including testing models relating retinal anatomy (structural measures) to behavioral thresholds of light sensitivity (functional measures). In addition to potentially improving our understanding of this relationship and how it changes during the course of neurodegenerative diseases of the eye such as glaucoma, analyses of these data may allow for early identification of glaucomatous neural damage in the retina, which has considerable clinical relevance. Here, the underlying assumptions and generalization of a previously developed model of the structure function relationship in glaucoma was tested by applying this model to a novel dataset. This model has been influential in the literature because it purports to accurately estimate the number of retinal ganglion cells; however, it was found to have several questionable assumptions and did not generalize well. Next, a new method of estimating the number of retinal ganglion cells from optical coherence tomography was developed. This method uses fewer and more defensible assumptions and demonstrated good agreement with independent histological estimates. Finally, a new method, using computer simulations, was developed for analyzing data from optical coherence tomography in order to distinguish early signs of glaucomatous changes in retinal anatomy from variability in structure among healthy retinas, and this method performed better than previously published techniques.

[Truncated abstract] Developmentally, the rodent retinocollicular projection is often thought of as a homogenous projection of retinal ganglion cell (RGC) axons, however the extensive period of RGC neurogenesis and sequential arrival of their axons into central targets such as the superior colliulus (SC) suggests otherwise. RGC axons are already present in the developing SC at embryonic (E) day 16.5-17. RGCs born on E15 have innervated the SC by birth, whereas axons derived from RGCs that are born last (E19) do not grow into the SC until postnatal (P) days 4-6 (Dallimore et al., 2002). These observations may go someway to explaining why, after SC lesions in rats at P2, there is greater growth distal to the lesion site compared to lesions made at P6 (Tan and Harvey, 1997b). It may be that the post lesion growth is simply de novo growth of axons from late-born RGCs rather than regeneration of pre-existing, injured axons. Early and late cohorts of growing RGC axons presumably encounter different developmental terrains as they grow from retina to central targets, possibly resulting in differences in developmental milestones and growth potentials. There may also be differences in guidance cues, further suggesting that gene expression in early vs late born RGCs may differ. To examine differences between early (E15) and late (E19) born RGCs during development, the time-course and extent of programmed RGC death in normal rat pups, and RGC death following the removal of target-derived trophic factors, was assessed. ... On the other hand, LCM captured GCL analysed for gene expression at P0 and P7 revealed decreases in AKT, Math5, Notch1, c-jun, DCC, Arginase-1 mRNA levels and a considerable decrease in GAP-43 expression. It is not surprising to see differences in gene expression between whole eye and the more specific GCL samples, as the cells in all layers of the retina have very different functions and different developmental profiles. It is important to note decreases in mRNA expression in the GCL for a number of the genes analysed at P0 and P7, reflecting cessation of RGC death and completion of axonal growth into central visual targets. I also examined at the protein level expression of DCC, Arginase1, c-Jun and Bcl-2 at birth (P0) in BrdU labeled RGCs born on E15 or E19. When comparing the percentage of double labelled cells compared to the total number of cells expressing each protein, Bcl-2, c-Jun and Arg1 were expressed more in E15 RGCs (22.90%, 72.71%, and 16.44% respectively in E15 RGCs, compared with 0.52%, 13.17% and 3.59% in E19 RGCs). In contrast, DCC was expressed more at birth in E19 RGCs (18.05% in E19 RGCs compared with 9.23% in E15 RGCs). This shows there is clearly a difference in the expression of proteins in the two cohorts of RGCs, which is consistent with PCR data and with their growth state as their axons encounter the changes in the newborn brain. The overall findings of this research suggest that seemingly homogenous populations of neurons are quite different in their developmental profile and in their response to injury. This work may provide new ways of determining better strategies for CNS repair and the most effective way of targeting cells for regeneration and survival.