In this thesis the laser ray tracing (LRT) technique for measurement of ocular aberrations has been implemented, validated and applied, in conjunction with complementary techniques, to the study of ocular aberrations in human eyes. In particular, we studied optical aberrations in myopic and hyperopic eyes and the optical changes induced by refractive surgery for myopia and hyperopia. We have studied the impact of the optimisation of some experimental parameters on the estimation of the wave aberration. We demonstrated that although the polarisation state and wavelength of the illumination light affected the intensity patterns of the images obtained using reflectometric aberrometry (LRT and Hartmann Shack sensor), these changes did not affect the estimation of aberrations. We also showed that the difference in the defocus term (focus shift) due to the use of different wavelengths is in agreement with the Longitudinal Chromatic Aberration of the Indiana Chromatic Eye Model for average normal eyes, although intersubject variability is not negligible. In addition, we studied experimentally the influence of the geometrical distribution and density of the pupil sampling on the estimation of aberrations using artificial and normal human eyes, and performed numerical simulations to extend our results to "abnormal"eyes. We found that the spatial distribution of the samples can be more important than the number of samples, for both our measured as well as our simulated "abnormal" eyes. Experimentally, we did not find large differences across patterns except, as expected, for undersampled patterns. We found that hyperopic eyes tended to have more positive asphericity and greater total and corneal spherical aberrati on than myopic eyes, as well as greater 3rd and higher order aberrations. Although we found no significant differences between groups in terms of internal aberrations, internal spherical aberration showed a significant age-related shift toward less negative values in the hyperopic group. We also assessed the impact of the LASIK corneal surgery, a popular surgical technique for correction of refractive errors, on the optical quality for both myopic and hyperopic standard techniques. Third and higher order ocular and an terior corneal aberrations increased with the surgery. Ocular and corneal spherical aberration changed towards more positive values with myopic LASIK, and towards more negative values with hyperopic LASIK. Changes in internal spherical aberration were of opposite sign than those induced in corneal spherical aberration. Changes induced by hyperopic LASIK were larger than those induced by myopic LASIK for a similar attempted correction.

The presence of a bright light source in the visual field, particularly when viewed against a dark background, can generate a form of discomfort, which is often described as ‘discomfort glare’. The mechanisms for discomfort glare remain poorly understood, even after 50 years of multidisciplinary research in this field. The aim of this investigation was to investigate a number of relevant parameters that can affect discomfort glare in order to gain insights into the corresponding mechanisms. We measured retinal illuminance levels for discomfort glare at threshold as a function of source size, eccentricity and surrounding background luminance. In addition, the pupil size was measured throughout and related to the measured thresholds for discomfort glare. A group of 50 subjects with normal visual acuity and no clinical signs of eye disease took part in the primary study that measured discomfort glare thresholds as a function of source size. A light ‘homogenizer’ was used to integrate the concentrated light output from a quad LED light source. Pulse frequency modulation was used to control the intensity of the source and continuous pupil size measurements made it possible to calculate retinal illuminance. Discomfort glare thresholds were estimated by measuring the retinal illuminance of the glare source at threshold using a staircase procedure. Discomfort glare thresholds were measured as a function of glare source area, eccentricity and background luminance. The amplitude of pupil constriction was also measured both below and above the discomfort glare threshold. A model of contrast vision with the filtering of a photoreceptor signal through centre-surround ganglion cells was developed to account for the small size dependence of discomfort glare thresholds that was observed experimentally. Another model for scattered light was applied to compute the corresponding pupil constriction amplitude caused by the integrated photoreceptor signals generated by the glare source both within and outside the stimulus area. The threshold for discomfort glare decreased gradually with glare source size and increased with background luminance and showed little dependence on glare source eccentricity. The effect of forward light scatter in the eye was also investigated and a model was developed to account for the continued increase in pupil response amplitude well above the discomfort glare threshold. The effect of glare source size on discomfort glare thresholds could be predicted by a model involving photoreceptor saturation and edge response. When the scattered light outside the stimulus area was also taken into account, the pupil constriction amplitude increased log-linearly with stimulus retinal illuminance both below and above discomfort glare thresholds. These findings suggest that discomfort glare depended largely on the localised retinal illuminance and could be accounted for by the saturation of photoreceptor signals in the retina. The results and the pupil modeling work also suggest that the pupil response to light flux increments continued well above the discomfort glare threshold, largely as a result of light scattered outside the area of the glare source.

Background: Sub-Saharan Africa faces an epidemic of diabetes. The prevalence and incidence of sight threatening diabetic retinopathy in developed countries and associations between systemic factors, including glycaemic control, blood pressure and blood lipid levels, are well documented. In contrast the epidemiological literature from sub-Saharan Africa is sparse. In this resource-poor setting, population-specific variables such as a high burden of infectious disease and anaemia are likely to affect the spectrum of pathology encountered. I aimed to investigate the prevalence, incidence and progression of diabetic retinopathy in Southern Malawi, to investigate the risk factors for diabetic retinopathy severity and progression in this population and to characterise endothelial function in Southern Malawian subjects with diabetes. Methods: I established the Malawi Diabetic Retinopathy Study, a 24 month prospective cohort study. Subjects were systematically sampled from two hospital-based, primary care diabetes clinics. Visual acuity, glycaemic control, systolic blood pressure, HIV status, urine albumin–creatinine ratio, and haemoglobin and serum lipid levels were assessed. Retinopathy was graded at an accredited reading centre using modified Wisconsin grading of four-field mydriatic photographs. Additionally, in order to investigate DR progression at five years, a cohort of subjects recruited to a cross-sectional study of diabetes complications in 2007 were traced and assessed. In a nested case-control study, serum markers of endothelial dysfunction and pulse amplitude tonometry were measured in a subset of subjects from the main cohort plus subjects without diabetes. Results: 357 subjects were recruited to the 24 month cohort study. At baseline 13.4% subjects were HIV-positive and 15.1% were anaemic. Baseline prevalence rates of any retinopathy, sight threatening diabetic retinopathy and proliferative retinopathy were 50.1% (95% CI 44.9–55.3), 29.4% (95% CI 24.7–34.1) and 7.3% (95% CI 4.6–10.0), respectively. Cumulative incidence at 2 years of sight threatening diabetic retinopathy for subjects with level 10 (no retinopathy), level 20 (background) and level 30 at baseline was 2.7% (95% CI 0.1-5.3), 27.3% (16.4-38.2) and 25.0% (0-67.4), respectively. In a multivariate logistic analysis, 2 step progression of diabetic retinopathy at 2 years was associated with HbA1c (odds ratio 1.27, 95%CI 1.12-1.45), baseline grade of DR (1.39, 1.02-1.91) and HIV infection (OR 0.16, 0.03-0.78). At 2 years, rates of progression to visual loss were: ≥15 letters lost in 17 subjects (5.8%), moderate visual impairment (<60 letters) in 3 subjects (1.0%), severe visual impairment ( < 50 letters) in 5 subjects (1.7%). The five year incidence of sight threatening diabetic retinopathy in subjects recruited to the 2007 cross sectional study, for those with level 10 and level 20 retinopathy at baseline, was 19.4% (11.3-27.4) and 81.3% (62.1-100), respectively. In the case control study of endothelial function higher serum VEGF and E-selectin were associated with having diabetes in multivariate regression. Serum VCAM-1 was associated with death in multivariate regression. Conclusions: I report the first cohort study of diabetic retinopathy from sub-Saharan Africa. I found a prevalence of sight threatening diabetic retinopathy, in subjects attending diabetes clinics, approximately 3 times that reported in recent European studies and a prevalence of proliferative retinopathy approximately 10 times higher. Progression to sight threatening diabetic retinopathy occurred approximately 3 times more frequently than reported in Europe. The negative association of HIV infection with retinopathy progression is a new finding. I report the first evidence from sub-Saharan Africa of endothelial dysfunction in subjects with diabetes and of an association between levels of endothelial biomarkers and mortality in these subjects. Results presented in this thesis highlight the urgent need for provision of services for retinopathy detection and management to avoid a large burden of vision loss.

Aim: I aimed to investigate methods of delivering neuroprotection to the retina: a) using magnetofection (a non-viral transfection technique) in ocular tissue, b) using stem cells as a vector of neuroprotection delivery and c) harnessing galvanotaxis to direct migration of stem cells into the retina. Methods: For the recipient test bed of the experiments in my thesis, I used retinal explants from mice. The viability of retinal explants was determined using retinal ganglion cell health as a read-out in the form of Sholl plots. I also explored magnetofection with oscillation as a non-viral technique for neuroprotection delivery in ocular tissue using cornea and retinal explants. Using the galvanotaxis technique, I explored its use in directing neural stem cells in-vitro with an electrotactic chamber and ex-vivo on a retinal explant in a modified Boyden chamber. Finally, I magnetofected neural stem cells to over-express BDNF and directed their migration into the retina using galvanotaxis. Results: The retinal explants had a viability of up to 3 days based on the Sholl plots of retinal ganglion cells. Magnetofection with oscillation transfected cornea endothelium and the retinal ganglion cell layer with GFP in the explant models. It also transfected neural stem cells with BDNF-myc. Directed migration with neural stem cells occurred in the electrotactic chamber as well as in the retinal explant model. In the absence of electric field, no migration into the retina occurred. Neurospheres transfected with BDNF-myc also migrated into the retina when exposed to an electric field. Conclusion: Within the period of up to 3 days, retinal explants can be used to investigate neuroprotective therapeutic agents using Sholl plots of retinal ganglion cells. Magnetofection with oscillation is a novel non-viral technique for potentially transfecting the eye in the anterior and posterior segments. Neurospheres can be directed to migrate into the retina using an electric field in the ex-vivo model.

The purpose of this thesis was to develop further the concept of Dynamic Noise Perimetry (DNP). The influence of 4 different strengths of Gaussian filter on the DNP stimulus edge, without and with a noise mask, was separately investigated in 15 normal individuals at three eccentricities. The DNP threshold was not affected by the filtering. The critical check size of the noise mask was investigated in 11 normal individuals at three eccentricities for 8 different checks per cycle. The critical check size at the fovea was 4 checks per cycle and in the periphery between 2 and 4 checks per cycle. The influence of optical defocus was investigated in 11 normal individuals at three eccentricities. For a defocus of +4.00DS, sensitivity without the noise mask declined by approximately 1dB; with the noise mask sensitivity increased by 1dB. The original ‘Proof of Concept’ threshold algorithm, which enabled the estimation of threshold at one location in approximately 3 minutes, underwent numerous modifications. The final iteration permitted threshold estimation at 45 locations in approximately 7 minutes. Five of the ten individuals with open angle glaucoma who had undergone DNP and standard automated perimetry (SAP) in 2007 were re-examined, using an identical protocol, after a follow-up of four years. The abnormality with DNP at baseline was present at the follow-up in all five individuals and was more severe in 3 individuals. Only 2 individuals exhibited abnormality by SAP. The influence of the learning effect on the outcome of DNP was evaluated, in one designated eye at each of the five weekly visits, for 10 ‘young’ and 8 ‘elderly’ normal individuals naïve to perimetry. Optimum performance was essentially achieved at the third visit without and with the noise mask.

Despite the association of neuronal cell loss with a wide range of neurodegenerative disorders, the mechanisms leading to this cell death remain poorly understood. In this thesis I have investigated these mechanisms and tested whether they represent viable targets for therapeutic intervention. The adult mouse retinal explant is a popular model of axotomy-induced neuronal degeneration but has been limited by the lack of morphometric data. Since dendritic pruning is well-evidenced to precede cell loss in neurodegenerative diseases, including glaucoma and Alzheimer’s disease, I investigated whether the quantification of dendritic morphology of retinal ganglion cells in the retinal explant could be used as a more sensitive measure of neuronal health after axotomy. I report here that retinal ganglion cell dendrite loss precedes cell loss by at least 7 days and that this retraction is substantially retarded following treatment with brain-derived neurotrophic factor applied at the time of explantation. Perhaps most importantly, I demonstrate for the first time in this model that delayed application of brain-derived neurotrophic factor significantly protects against dendritic retraction of retinal ganglion cells. The work outlined in this thesis thus supports the targeting of dendritic outgrowth and/or synaptic connectivity for the treatment of neurodegenerative disorders.

It is estimated that there are millions of corneal operations performed annually in the United Kingdom and United States following pathology and/or acute injury. Equally, millions elect to have corneal refractive surgery to remedy visual disorders such has myopia, prespreopia, astigamatism and associated keratoconus. Following all of these procedures, there is a risk of complications from scar formation, cell loss, and corneal oedema which may adversely affect the surgical outcome and resulting best vision. Through research into how the cornea heals, it is hoped that these factors could be better understood and ultimately used to increase patient satisfaction. The projects are outlined as followed: 1) Using X-ray scattering techniques to investigate the role of Pax 6 on the collagen ultra- structure of murine corneas 2) Oral mucosal fibroblasts as a novel wound healing treatment of LASIK injured ovine cornea 3) Using X-ray scattering techniques to investigate the ultra-structural changes of a 12 year post-operative penetrating keratoplasty for keratoconous 4) Using small angle X-ray scattering techniques to investigate limbal collagen ultra-structure of keratoconic and normal human tissue In this wound healing study, a range of different pathological and acutely injured mammalian corneas (human, sheep, and mouse) were evaluated using a range of quantitative techniques. These techniques encompass small and wide angle X-ray scattering, mechanical analysis, immunofluorescence, spectrophotometry, and cell/organ culture model. We feel that investigating various species and disorders has provided a more comprehensive and broad overview of the various maladies which can occur during the wound healing process.

Cataracts are the commonest cause of preventable blindness in the world. During surgery the natural lens is replaced with a polymeric intraocular lens (IOL), leaving the capsular bag in situ. The most common postoperative complication is scarring which is known as posterior capsule opacification (PCO). PCO occurs when residual lens epithelial cells (LECs) dedifferentiate and migrate onto the previously cell free posterior capsule. By modifying the IOL surface properties we can manipulate the cellular response. BioInteractions Ltd. is an innovative supplier of biomaterials, which aim to minimise, the host response, and provided the materials for this study. The aim of this study was to evaluate potential IOL coatings to reduce PCO. This can either be achieved by enabling a monolayer of LECs to attach to the posterior surface of the IOL, thus sandwiching the IOL to the capsular bag, or prohibiting cell attachment to the IOL entirely. Materials and Methods: Various coatings were investigated incorporating functional groups of poly ethylene glycol (PEG), sulphates, sulfonates, glycosaminoglycans (Heparin, (HEP) hyaluronic aid, (HA) and chondroitin sulphate (CS)) and zwitterionic monomers (10-30%). Ways to prevent dedifferentiation was also evaluated. LECs were seeded onto all coatings and monitored for a period of 7 – 14 days in cell culture. LECs were examined morphologically, cell nuclei were counted and growth curves were plotted. Water contact angle (CA) measurements were taken to measure the wettability of the coatings. Scanning electron microscope (SEM) analysis was performed to examine the topography of the coating. White light interferometry (WLI) analysis was conducted to analysis the surface roughness. Dedifferentiation of LECs and the use of TGFβ3 to neutralise or prevent dedifferentiation were also investigated. Results and Discussions: Coatings with a greater number of water-based layers were the most hydrophilic, and did not offer the appropriate cell binding sites required to promote cell attachment. In general, little cell attachment was observed on HEP and HA coatings provided by BioInteractions Ltd., cell attachment varied on CS coatings provided by BioInteractions Ltd. When HA and CS were covalently bound onto amine coated coverslips a reduction in cell attachment was observed. The LEC response varied across different ratios of zwitterionic monomer within the coatings. Zwitterionic coatings were not cytotoxic to LECs and surface analysis demonstrated no clear link between wettability and roughness compared to cell attachment. Addition of transforming growth factor beta 2 (TGFβ2) was chosen as a successful dedifferentiation model. Addition of TGFβ3 had little influence at reversing dedifferentiation however it may offer some protection against differentiation. PCR analysis showed a change in regulation of collagens, integrins, matrix metallopeptidase and fibronectin 1 genes, when LECs were incubated with TGFβ2, TGFβ3 or untreated (control LECs). These genes may play important roles in PCO. Conclusions Incorporation of functional groups influenced the cellular response, however the coatings with more water-based layers prohibit cell attachment. The cellular response varied depending on GAG type and the conformation of GAG on the surface coating. HA and CS bound to amine-coated coverslips prohibited cell attachment at higher concentrations, indicating their potential to prohibit LEC attachment. There was no clear link between wettability and cell attachment on the novel zwitterionic coatings. The ratio of zwitterionic-component:arylic-based monomer(s) influenced cell attachment. TGFβ2 successfully dedifferentiated LECs. Further work is required to understand the influence of TGFβ3 on dedifferentiation.

Pocket and portable electronic low vision aids (PELVAs) are relatively new devices designed to enhance vision of people with visual impairment. Therefore, the aim of this thesis was to evaluate their use and prescribing patterns for PELVAs for patients with visual impairment and to inform clinicians of the functions or attributes that are most important when considering their prescription. Firstly, the parameters (including magnification screen size, luminance contrast, and resolution) of PELVAs were assessed. Magnification and screen size were also compared with the manufacturers’ data. Secondly, data (age, gender, ocular condition, visual acuity, living situation, registration of visual impairment, type of low vision aid prescribed) from 6,668 patients who attended the Low Vision Service for Wales were analyzed to assess the clinicians’ prescribing patterns for PELVAs. Thirdly, reading frequency and duration, and self-reported effectiveness of PELVAs and optical low vision aids for patients with visual impairment were evaluated by using a telephone interview based upon the Manchester Low Vision Questionnaire. Finally, reading speed was measured for normally sighted subjects with simulated visual impairment who used a PELVA and an optical low vision aid. The factors that could predict reading speed were investigated. PELVAs enhanced the luminance contrast of high and low contrast letters which may be beneficial for patients with contrast reduction, for example due to cataracts. Variations were found between reported and measured magnification of the PELVA. Only 10% of adult patients were prescribed a PELVA, and younger males were more likely to use them. A larger proportion of children (36.5%) were prescribed a PELVA. Patients who used PELVAs rated them highly for near vision tasks and were more likely to use PELVAs for reading once a day but for a long duration, whereas optical low vision aids were used more frequently and for a shorter duration. It was found that the reading speed was significantly improved using PELVA and optical low vision for subjects with simulated visual impairment; Bailey-Lovie near visual acuity, high and low contrast Colenbarnder near visual acuity, and number of visible characters were significant predictors of reading speed.

Glaucoma is a progressive optic neuropathy that causes irreversible vision loss and is the second leading cause of blindness worldwide. Glaucoma is characterised by loss of retinal ganglion cells (RGC) and the proposed site of primary damage is the lamina cribrosa (LC), where RGC axonal transport is disrupted causing subsequent RGC damage and eventual cell death. Current detection for primary open angle glaucoma (POAG) is based upon clinical measures such as intraocular pressure (IOP), visual field loss and changes to the optic nerve head (ONH). However, for there to be an indication that there is a problem using these measures, often RGC damage has already occurred. Therefore it is crucial to determine ocular parameters that alter in the earliest stage of disease, prior to vision loss occurring. In this thesis optical coherence tomography (OCT) was used to assess the optic nerve heads and maculae of control eyes and eyes with preperimetric, early and advanced glaucoma in order to characterise changes that could potentially be used as biomarkers for the earliest stages of the disease. A custom built 1050 nm research OCT was used to acquire datasets from the macula and optic nerve heads of eyes glaucomatous and control eyes in vivo. Analysis of the inner retinal layers at the macula was performed to indirectly assess RGC integrity. At the ONH the prelamina and LC volume and regional depth and thicknesses were investigated. Additionally, nerve fibre layer and Bruch’s membrane parameters were assessed. Finally, LC beam coherence and orientation were probed in order to determine whether regional or glaucomatous changes ould be detected at the LC connective tissue microstructure. Prelamina depth and thickness was shown to be an indicator of early and preperimetric glaucoma (p < 0.01), although the volume of the prelamina did not change with increasing stage of glaucoma (p > 0.01). Border nerve fibre layer revealed significant thinning in early glaucoma compared to control, and the superior peripapillary nerve fibre layer was thinner in preperimetric glaucoma than control. The ratio of inner plexiform layer (IPL) : ganglion cell layer (GCL) showed significant differences between control eyes and preperimetric glaucoma, and as such has potential to be a useful biomarker for indicating the earliest stages of disease. Both the GCL and IPL were thinner in early glaucoma than control (p < 0.01), a hypothesis that cell body shrinkage and death occurs in preperimetric glaucoma and dendritic loss occurs in early glaucoma, when vision loss is first apparent, is suggested. Additionally, LC beams showed greater coherence in the superior and inferior poles than the temporal region, indicating that the shows regional variation but that further research is required to characterise changes. In conclusion, 1050 nm OCT was used to probe microstructural parameters of the optic nerve head in vivo to characterise changes that could be used as a potential biomarker for early glaucoma. ONH and retinal parameters have been identified that, with further research, may be used to differentiate between control eyes and those with preperimetric and early glaucoma. These have the potential to help identify those ONHs at risk of glaucoma damage.