WO2018024842A1

WO2018024842A1 - Retinal photoreceptor mosaic simulator

Info

Publication number: WO2018024842A1
Application number: PCT/EP2017/069711
Authority: WO
Inventors: Brian Vohnsen; Denise VALENTE DOS SANTOS
Original assignee: University College Dublin, National University Of Ireland, Dublin
Priority date: 2016-08-03
Filing date: 2017-08-03
Publication date: 2018-02-08
Also published as: GB201613418D0

Abstract

A method of manufacturing a retinal photoreceptor mosaic simulator is described. The method comprising the steps of: depositing a photoresist onto a substrate; exposing the photoresist with radiation to define a pattern of structures on the photoresist, the pattern corresponding to a distribution of retina photoreceptors cells in a retina of an eye; and producing a film comprising a latent image defining a plurality of dielectric structures; wherein a refractive index contrast between the plurality of dielectric structures and their surrounds in the latent image corresponds to a refractive index contrast of the retina photoreceptor cells and their surrounds in the retina of the eye.

Description

RETINAL PHOTORECEPTOR MOSAIC SIMULATOR

Field of the invention The present invention relates to a retinal photoreceptor mosaic simulator and a method of manufacturing a retinal photoreceptor mosaic simulator.

Background of the invention (i) Current physical eye models used in laboratories worldwide model the refractive properties of the cornea and lens. They use a lens (or a combination of lenses) to mimic the refractive properties of the anterior eye and a CCD camera or screen to capture and monitor the images in the retinal plane. The camera pixels are sensitive to light being incident at any angle without discrimination and thus if scattering or diffractive components are analysed, such as in the design of intraocular lenses, contact lenses and spectacles, or when modelling vision with cataract, incorrect effective retinal images will be captured.

(ii) Retinal implants, or bionic eye implants, have for the past decade been increasingly used to partially restore vision in patients being blind due to absence of functioning photoreceptors but with other retinal layers being intact. The pixelated implants are surgically inserted into the retina and produce a small electric current when exposed to light whereby neural responses are stimulated and vision activated. This allows to partially recover vision in eyes affected by retinitis pigmentosa (RP) and possibly advanced age-related macular degeneration (AMD). Current implants are based on camera technology without directional selectivity whereby they capture also diffuse light that has undergone multiple scattering within the eye of the patients. This causes glare and prevents the full visual potential of implants to be used when recovering vision.

Another prior invention suggests a contact lens device, made of opaque material with circular channels in an area correspondent to the portion of the retina that still maintains optical perception. The light passing through these holes is faced to strike the posterior retina in a perpendicular manner, resulting in increased visual acuity. The portion of the retina not having sufficient light sensitivity is covered by an opaque portion of the contact lenses without holes. Thus it essentially blocks unwanted light at the pupil but cannot remove unwanted intraocular scattering of light which is especially problematic in diseased eyes with low visual acuity. Summary of the invention

It is an object of the present invention to provide a device for simulating the optical functioning of the photoreceptor mosaic in the retina and a method of manufacturing such a device.

To simulate the optical functioning of the photoreceptor mosaic in the retina, an array of refractive-index-enhanced dielectric structures has been invented. The refractive index enhancement is understood as an incremental increase of the refractive index in the dielectric structures (e.g. caused by the method of manufacture used).

In a first aspect of the invention, there is provided a method of manufacturing a retinal photoreceptor mosaic simulator, the method comprising the steps of: depositing a photoresist onto a substrate; exposing the photoresist with radiation to define a pattern of structures on the photoresist, the pattern corresponding to a distribution of retina photoreceptors cells in a retina of an eye; and producing a film comprising a latent image defining a plurality of dielectric structures; wherein a refractive index contrast between the plurality of dielectric structures and their surrounds in the latent image corresponds to a refractive index contrast of the retina photoreceptor cells and their surrounds in the retina of the eye.

In one embodiment of the invention, a distribution of the plurality of dielectric structures in the latent image corresponds to a distribution of retina photoreceptor cells in the retina of the eye. In an embodiment of the invention, the plurality of dielectric structures are spaced apart from each other in the latent image and a refractive index of a dielectric structure is higher than a refractive index of a space between two of the plurality of dielectric structures. In another embodiment of the invention, the developing step comprises: postexposure baking the photoresist to catalytically complete a photoreaction initiated during the exposing step. In an embodiment of the invention, the exposing step comprises: overlaying the photoresist with a photomask; the photomask having the pattern of structures corresponding to the distribution of retina photoreceptors cells in the retina of the eye; and illuminating the photoresist via the photomask with an ultra violet light source.

In one embodiment of the invention, the ultra violet light source is a collimated ultra violet light source.

In another embodiment of the invention, the exposing step comprises: scanning a focused beam of ultra violet light over the photoresist to form the pattern of structures on the photoresist.

In an embodiment of the invention, the exposing step comprises: scanning infra-red laser beams over the photoresist to form the pattern of structures on the photoresist. For example, two or three photon induced refractive index changes could be produced by pulsed infrared laser sources. Such scanning with an intense pulsed infrared beam would achieve substantially the same result by multiphoton absorption as the UV light achieves. In a further embodiment of the invention, the exposing step comprises: scanning an electronic beam over the photoresist to form the pattern of structures on the photoresist.

In an embodiment of the invention, the beam or beams (UV, IR, or e-beam) can be scanned with respect to the film. In another embodiment of the invention, the film can be scanned with respect to the beam or beams used.

In one embodiment of the invention, each of the plurality of dielectric structures is about 2 to about 8 microns in diameter and the plurality of dielectric structures are spaced apart from each other by a distance of about 2.5 microns to about 13 microns.

In another embodiment of the invention, each of the dielectric structures is about 5 microns in diameter and the plurality of dielectric structures are spaced apart from each other by a distance of about 10 microns.

In one embodiment of the invention, the plurality of dielectric structures are substantially cylindrical in shape.

In another embodiment of the invention, the plurality of dielectric structures are substantially non-cylindrical in shape. For example, cones or cuboids may be designed to better match the shape of the pixels in the light-sensitive electronic chips. In a further embodiment of the invention, the plurality of dielectric structures comprise of substantially cylindrical dielectric structures and/or substantially non- cylindrical dielectric structures.

In an embodiment of the invention, the photoresist is a positive photoresist. In a further embodiment of the invention, the photoresist is a positive photoresist of the type AZ40XT.

In another embodiment, the photoresist is a negative photoresist. In a further embodiment, the photoresist is a negative photoresist of the type SU-8.

In one embodiment of the invention, the substrate is one of a glass substrate, a silica substrate, a fused silica substrate or a polymer substrate. In a further embodiment of the invention, the depositing of the photoresist onto the substrate comprises one of spin coating, dip coating or spray coating the photoresist onto the substrate.

In a second aspect of the invention, there is provided a retinal photoreceptor mosaic simulator comprising: a film comprising a latent image defining a plurality of dielectric structures; wherein a refractive index contrast between the plurality of dielectric structures and their surrounds in the latent image corresponds to a refractive index contrast of retina photoreceptor cells and their surrounds in a retina of an eye. In one embodiment of the invention, a distribution of the plurality of dielectric structures in the latent image corresponds to a distribution of retina photoreceptor cells in the retina of the eye.

In another embodiment of the invention, the plurality of dielectric structures are spaced apart from each other in the latent image and a refractive index of a dielectric structure is higher than a refractive index of a space between two of the plurality of dielectric structures.

In a further embodiment, the plurality of dielectric structures are configured to act as a waveguide for guiding light from a retinal image plane to a detection plane.

In one embodiment of the invention, each one of the plurality of dielectric structures is configured to act as an independent waveguide for guiding light from a retinal image plane to a detection plane.

In an embodiment of the invention, the plurality of dielectric structures are configured to act as an angular low-pass filter for dampening the influence of oblique light rays. In a further embodiment of the invention, each one of the plurality of dielectric structures is configured to act as an angular low-pass filter for dampening the influence of oblique light rays.

In another embodiment of the invention, each one of the plurality of dielectric structures is about 2 to about 8 microns in diameter and the plurality of dielectric structures are spaced apart from each other by a distance of about 2.5 microns to about 13 microns.

In a further embodiment, each one of the plurality of dielectric structures is about 5 microns in diameter and the plurality of dielectric structures are spaced apart from each other by a distance of about 10 microns. In one embodiment of the invention, the plurality of dielectric structures are substantially cylindrical in shape.

In another embodiment of the invention, the plurality of dielectric structures are non- cylindrical in shape.

In a third aspect of the present invention, there is provided an image capturing system comprising the retinal photoreceptor mosaic simulator of the present invention.

In an embodiment of the invention, the image capturing system comprises one or more of: an eye model, a retinal model, a retinal implant, a bionic eye implant, a bionic lens implant, an intraocular lens, a corneal inlay, spectacle glasses, a contact lens or a camera.

Brief description of the drawings

Embodiments of the present invention will now be described, by way of example only, with reference to the accompanying figures in which: Fig. 1 shows a schematic view of 1 (a) lateral and 1 (b) top view of the UV-exposed photoresist.

Fig 2 shows a schematic of an experimental setup (not to scale) used during the photolithography process in accordance with an embodiment of the invention;

Fig 3 is an interferometric measurement showing the increased refractive index as a bending of the interference fringes at the location of the cylindrical structures.

Fig 4(a) shows a schematic of the incident beam at different positions at the pupil and oblique incidence at the retina;

Fig. 4(b) is a graph showing transmitted total power (normalized) through the array of elevated refractive-index cylinders as a function of the angle of incidence: experimentally verified (squares) and relative luminance efficiency of the Stiles- Crawford effect of the first kind with a characteristic directionality of p = 0.01/mm².

-p?

This curve is the fitting of a Gaussian distribution function ⁼ 10 where r is the distance from the pupil centre; Fig. 5(a) shows a refractive physical eye model and the insertion of a retinal simulator in the image plane, or a conjugated image plane, and 5(b) shows a retinal microelectronic implant with the addition of a retinal simulator for narrowed angular sensitivity; and

Fig. 6 shows example images captured with a CCD camera and using an optical diffuser to simulate intraocular scattering 6(a) without and 6(b) with the retinal model inserted in front of the camera.

Figure 7(a) shows a cross-section of light distribution for different values of length from 10μηι to Ι ΟΟμηη modelled in Comsol™ of an infinite array of high-index cylinders, 5μηι diameter, separated by 10μηι in an hexagonal transparent lattice, with n_core = 1 .595 and n_dad =1.591 and wavelength λ= 543nm.

Figure 7(b) shows a cross-section of light distribution in the presence of absorption in the surrounding cladding for different values of length from ΙΟμηι to ΙΟΟμηι modelled in Comsol™ of an infinite array of high-index cylinders, 5μηι diameter, separated by ΙΟμηι in an hexagonal transparent lattice, with n_core = 1.595 and n_clad = 1.591 and wavelength λ = 543nm.

Figure 8(a) shows transmitted power fraction through the cores as a function of angle of incidence: (I) without absorption in the surround cladding; (II) with absorption.

Figure 8(b) shows directional sensitivity as a function of waveguide length with and without absorption in the surround cladding. The presence of the absorber avoids the leakage of light to neighbouring cylinders, increasing the directionality of the waveguides.

Figure 9 shows an example setup (not to scale) used to analyse the influence of the photoresist operating as an angular low-pass filter on the image quality. The lens L_t is mounted in a translation stage to controllably induce defocus in the wavefront. L₁ = 19mm, L₂ = 150mm, Objective Olympus Uplan 20X, Iris Φ = 6mm

Figure 10(a) shows an analysis of the contribution of the retina phantom to the visual performance. An USAF 1951 paper target was used as object. In the images are observed group -2 and elements 4 and 5 of the target. On the left, resultant images with a photoresist without printed cylinders; on the right, the retinal images with printed cylinders. Different defocus conditions were analysed from 0μm displacement out of the focal plane, up to 195μm. The blurring progression is slower in the presence of the array. Figure 10(b) shows Matlab simulations of the projected image. The images present defocus aberration generated by a displacement of l_i in 195μηι out of the focal plane and were low-pass angular filtered by retinal simulators with different p values.

Description of Exemplary Embodiments

Cone and rod photoreceptors in the human retina are elongated cells (diameter in the range of 2 to 8 microns (μηι) and total length up to 100 μηι) that have directional light capture properties similar to those of optical fibres. This ensures efficient transmission from the inner (30 to 50 μηι) to the outer (20 to 50 μηι) photoreceptor segments where visual pigments are densely packed and the visual cycle is triggered by absorption. The refractive index of the inner segments is approximately 1 .35 and thus it is only slightly larger than that of the surrounding matrix of approximately 1 .34. The outer segment has higher refractive index due to the dense packing of pigments with a typical value of 1 .43. The directional light-capture efficiency and thus the acceptance angle of the photoreceptors is approximately matched to the size of the eye pupil. It dampens the impact of intraocular light scattering that could otherwise cause glare and blurry vision. Retinal disease can alter the alignment and composition of the photoreceptors and lead to irreversible vision loss or blindness.

In Fig. 1 , lateral and top view sketches show areas 1 exposed by UV light during the photolithography process and areas 2 where the UV light was blocked according to embodiment of the invention. The exposed areas 2 ("grey areas") show an increased refractive index compared with the unexposed areas 1 ("white areas") enabling each column in the ensemble to operate as an independent waveguide. With regard to Fig. 1 , the "white areas" 1 have the refractive index of the resist (n_resist), whereas the grey areas 2 have the increased (elevated) refractive index which is n_resist + Δη (where, in this example, Δη = 0.004).

In other words, Fig. 1 shows an array of refractive-index-enhanced dielectric cylinders 1 in a substantially uniform matrix which has been designed in order to simulate the optical functioning of the photoreceptor mosaic in the human retina. In this embodiment, the structure has been made by use of UV photolithography with the positive photoresist AZ40XT. This material facilitates large film thickness via spin-coating in a single step and with good adhesion onto a transparent glass substrate. The lithographic process results in a latent image with refractive index contrast similar to that of the photoreceptor cells and their surrounds. Thus, conventional post-exposure chemical etching of the photoresist in combination with liquid immersion is not required to achieve a suitable refractive index contrast resembling that of the human retina.

Fig 2 shows a schematic of an experimental setup (not to scale) used during the photolithography process in accordance with an embodiment of the invention. With reference to fig. 2, UV exposure was performed with a beam of UV light from an Olympus U-LH100HG 3, at a wavelength of approximately 365 nm, passing through an iris 4 and an UV-passing filter (Thorlabs FGUV1 1 S) 5. The light is then optionally collimated by an achromatic lens 6 and illuminates the chrome photomask 7 (other opaque masks, not based on chrome, are also feasible) having a hexagonal lattice of 5 μηι circles separated by 10 μηι. The pattern of the chrome photomask is printed in the photoresist 8 placed on top of the photomask (e.g. in hard contact in this example).

In this embodiment, the photoresist 8 was deposited onto a fused silica substrate using spin coating and posteriorly baked to remove the solvent. To obtain elongated cylindrical structures, as is shown in Fig. 2, contact printing was performed by illuminating through the chrome photomask 7 overlaying the photoresist 8 with a UV beam (e.g. the collimated UV beam). The photomask 8 has a hexagonal lattice of about 5μηι diameter circles with a spacing of about 10μηι between circles, similar to the distribution of retina cone photoreceptor cells near the fovea but can easily be altered with alternative mask configurations. In this embodiment, the photoresist 8 is post-exposure baked to catalytically complete the photoreaction initiated during exposure and thereby obtain the desired refractive index changes. During the UV exposure the AZ40xt photoresist experiences a photoreaction that is completed with the post-exposure baking. This photoreaction changes both chemical and optical properties of the photoresist. In other embodiments, post-exposure baking is not performed. In chemically amplified resists, the post-exposure bake catalytically performs and completes the photo reaction initiated during exposure. However for non-chemically amplified resists, such as the AZ® and Tl resists distributed by MicroChemicals®, post-exposure baking is not required for this purpose. In this embodiment, it is not necessary to develop the film. The UV exposure will produce the latent image as the refractive index changes. In other words, the photoresist is exposed to UV light which will create the "developed" film with latent images of elevated refractive indices in the exposed parts of the film. The photoresist 8 was successfully developed experimentally with printed dielectric cylinders with elevated refractive index in the hexagonal lattice of about 5 μηι circles separated by about 10 μηι, similar to the distribution of retinal cones in the vicinity of the foveal region.

Elevated refractive index is understood to mean that the refractive index is slightly higher (e.g. about 0.004 in this embodiment) in the exposed parts of the resist than the parts of the resist not exposed to UV light. The absolute value of the refractive index is not important for the film with latent images with elevated refractive index structures (e.g. cylinders) to act as a waveguide as only the refractive index difference between the cylinders and the surrounding matrix define the properties. Thus it is not necessary to measure the absolute value of the refractive indices, though this could be done with other techniques such as by total internal reflection to determine the critical angle. The refractive index of the resist used in this embodiment is about 1 .7 though the exact value will be wavelength dependent.

The resulting film has a thickness of approximately 50μηι and a latent image can be observed with optical microscopy. Fig. 3 is an interferometric measurement showing the increased refractive index as a bending of the interference fringes 10 at the location of the cylindrical structures. In this measurement coherent laser light transmitted by the structure was made to interfere with a plane reference wave of light. Changes of the refractive index in the photoresist 8 were examined with interferometry revealing an increment of about Δη=0.004 within the cylinders which is approximately 40% of the value found for the inner segments and their surrounds in the human retina. In other words, local distortion of the fringe pattern within each cylindrical region demonstrates that the photolithographic process has produced an increment of Δη=0.004 of the refractive index determined from the bending of the fringes 10. The purpose of the invention is not to absorb the light in visual pigments but only to guide the light towards the detector (e.g. CCD). Thus, retinal photoreceptor mosaic simulator 40 of the present invention does not need to reproduce the refractive index variation of the inner and outer segments of the photoreceptors cells exactly but only to have a relatively uniform guidance of light from the retinal image plane to the detection plane. A key parameter that is important in terms of guiding of the light is the refractive index difference (contrast) between the waveguides and the surrounding matrix material. Therefore, with respect to the interferometry measurements, the inventors of the present invention are mainly concerned with measuring the change of refractive index rather than the absolute value. The refractive index of the resist would be known to the manufacturer although it is wavelength dependent. It is in the range of 1 .67 - 1 .70 typically, depending on the chosen resist. Thus, in this embodiment, the refractive index changes for the wave-guiding are very small (about 0.2% increase though it is feasible to be able to control it up to at least about 1 %). Angular tuning of light coupling is used to probe the directionality of the cylindrical refractive-index enhanced structures. When oblique light is incident onto a waveguide less light is coupled and transmitted within the core. Thus it acts as an angular low-pass filter that dampens the influence of unwanted oblique light rays associated with either aberrations or intraocular light scattering. In the human eye this angular dependence is known as the Stiles-Crawford effect of the first kind (SCE-I) and it is manifested psychophysical^ by a reduction in visual sensitivity to light rays that enter the eye near to the pupil rim. Thus, the photoreceptor and visual response depends on the obliqueness of the incident light at the retina with highest sensitivity for light that is incident near the eye pupil centre. The invented retinal photoreceptor mosaic simulator 40 mimics the angular dependence of the SCE-I in the fraction of transmitted light for any given angle of incidence as verified by angular transmission measurements in Fig. 4. Fig 4(a) shows a schematic of the incident beam at different positions at the pupil and oblique incidence at the retina and Fig. 4(b) is a graph showing transmitted total power (normalized) through the array of elevated refractive-index cylinders as a function of the angle of incidence: experimentally verified (squares) and relative luminance efficiency of the Stiles- Crawford effect of the first kind with a characteristic directionality of p = 0.01/mm².

This curve is the fitting of a Gaussian distribution function 7 =10 ' where r is the distance from the pupil centre. There are at least two problems that may be solved by this invention operating as a phantom of the photoreceptor mosaic. As shown in Fig. 5, the retinal photoreceptor mosaic simulator 40 of the present invention can be added onto (i) physical eye models or (ii) retinal implants that restore vision.

Fig. 5(a) shows a refractive physical eye model 30 and the insertion of a retinal simulator 40 in the image plane, or a conjugated image plane, and 5(b) shows a retinal microelectronic implant 50 with the addition of a retinal simulator 40 for narrowed angular sensitivity.

(i) Current physical eye models 30 used in laboratories worldwide accurately model the refractive properties of the cornea and lens but do not take the photoreceptor structure and packing into account when estimating effective retinal images. They use a lens (or a combination of lenses) 60 to mimic the refractive properties of the anterior eye and a CCD camera or screen 65 to capture and monitor the images in the retinal plane. The camera pixels are sensitive to light being incident at any angle without discrimination and thus if scattering or diffractive components are analysed, such as in the design of intraocular lenses, contact lenses and spectacles, or when modelling vision with cataract, incorrect effective retinal images will be captured. Realization of improved eye models with a retina-simulating phantom will allow to objectively model the role of photoreceptor structure and packing on vision and thereby deduce its role on common analysis methods such as point-spread function (PSF), modulation transfer function (MTF), and wavefront analysis before posterior neural responses of the visual system. The retinal photoreceptor mosaic simulator 40 can be used to examine the SCE-I, the hue shift of the related Stiles-Crawford effect of the 2nd kind (SCE-II), the role of aberrations, accommodation and emmetropization of the eye in a physical setting. (ii) Retinal implants 50, or bionic eye implants, have for the past decade been increasingly used to partially restore vision in patients being blind due to absence of functioning photoreceptors but with other retinal layers being intact. The pixelated implants are surgically inserted into the retina and produce a small electric current when exposed to light whereby neural responses are stimulated and vision activated. This allows to partially recover vision in eyes affected by retinitis pigmentosa (RP) and possibly advanced age-related macular degeneration (AMD). Current implants are based on camera technology without directional selectivity whereby they capture also diffuse light that has undergone multiple scattering within the eye of the patients. This causes glare and prevents the full visual potential of implants to be used when recovering vision. The developed retinal photoreceptor mosaic simulator 40 can be overlaid onto existing implants acting as a low-pass angular filter that resembles the optical properties of the photoreceptors in the healthy eye fundus.

Although retinal implant technology to restore vision to the blind is still in its early phases major potential for growth is to be expected as 1/4000 people are globally affected by RP and AMD is a leading cause of blindness with 7000 new cases reported annually in Ireland alone.

No physical eye model currently incorporates photoreceptor-simulating structures. In turn, prior inventions have been made for retinal simulators to overcome degenerated retinal function in the human eye. These are mainly based on photosensitive surfaces that, with different mechanisms, can electrically simulate the inner retinal layer and ganglion cells thereby bypassing deteriorated photoreceptors but they do not reproduce the reduced angular sensitivity associated with the SCE-I.

Microelectronic devices have been constructed and implanted in clinical trials resulting in light perception and image performance although with very low visual acuity (VA) due to the limited number of pixels of such devices. The best reported performance is VA = 20/546 which is still only about 3% of normal vision in the healthy eye. Moreover, vision is only black and white with no proper colour response. Retinal implants have been surgically inserted into the retina of blind RP patients using technology developed by predominantly two companies: (i) Retina Implant AG and (ii) Second Sight Inc.. They fall into two categories: one that relies on the anterior optics of the human eye to focus images onto the device and another that makes use of an external camera to send radio-wave signals to an implanted electronic receiver.

(i) Retina Implant AG has implanted at least 29 sub-retinal (behind the retina, replacing the lost photoreceptor layer) Alpha IMS implants with more than 86% of treated patients being able to see light stimuli and 72% with increased mobility and capability for daily tasks. The light-sensitive silicon implant contains an array of gold- wired photodiodes and signal amplifiers and is approximately 3 mm ^χ 3 mm ^χ 70 μηι in size. It operates essentially as a digital camera within the eye. It is electrically powered by induction.

(ii) Second Sight has implanted the Argus II epi-retinal (attached onto the retina) prosthesis in 30 patients. It has conductive tips that extend into the ganglion cell layer where it stimulates neural responses. It makes use of an external digital camera attached to a pair of eye glasses along with a video processing unit that transmits by radio frequency real-time captured images to the implant by induction. In this manner it bypasses the anterior optics of the patient's eye that may be degraded by disease, but in turn it requires head motion rather than just eye motion to change the direction of gaze. The Alpha IMS technology has a total of up to 1500 electrodes whereas the Argus II has only 60 electrodes (each with a 200 μηι diameter). It is to be expected that the number of pixels will continue to rise as the technology improves although it is still orders of magnitude below the more than 6 million cones (and more than 150,000 cones/mm2 at the fovea) and 90 million rods. The cost of each Argus II implant is superior to 150,000 USD and it is the first technology to receive FDA approval in the USA.

Physical eye models used to analyse the performance of intraocular lenses, contact lenses, corneal inlays, spectacle glasses, simply use a camera in the retinal image plane. Such settings do not include the narrow angular selectivity of the photoreceptors and thus it is known that the simulated retinal images acquired differs from visual sensations. This is partially attributed to neural adaptation, but the exclusion of a physical retina model is questionable. This becomes increasingly important for large pupil diameters or when analysing the possible role of aberrations or scattering of light such as in patient eyes. The add-on of a physical retinal model in such systems will be beneficial and may potentially allow for the development of improved optical designs.

In contrast to the retinal photoreceptor mosaic simulator 40 of the present invention, current retinal implant technologies do not emulate the directional sensitivity of the photoreceptors and thereby fail to discriminate against obliquely incident light originating either from the rim of the eye pupil where aberrations are large or from intraocular scattering 70 causing glare and reduced visual acuity. Since current implants are made of highly reflective materials intraocular scattering 70 will be strong and stray light in eyes with implants is therefore problematic. This is most problematic for implants that rely on the optics of the patient's own eye (the technology of Retina Implant AG) but much less so when images are supplied by an external camera attached to opaque glasses worn by the patient (Second Sight Inc.). The latter technology injects electrical current locally into the retinal ganglion cells thereby avoiding the use of light in the patient eye. Its main drawback is that it is attached to the head rather than the eye and thus a change of gaze requires rotation of the head rather than the eye.

With respect to the aforementioned contact lens design that incorporates the Stiles- Crawford effect it does not restore the visual response of damaged photoreceptors and therefore cannot restore vision in RP and AMD nor can it alleviate intraocular scattering of light 70.

The invented retinal photoreceptor mosaic simulator 40 can be implemented as a physical optics retinal model for dampening the visual impact of obliquely incident light at the retina whether caused by aberrations or intraocular scattering 70. It accomplishes this by the angular coupling efficiency shown in Fig. 4 that limits transmission to a narrow angular range set by the characteristic directionality related to the refractive index difference between the cylindrical structures and the surrounding matrix material within the photoresist layer.

This can allow for:

(i) Improved physical eye models 30 used in the testing of refractive optics and intraocular lens design whether mounted directly onto the imaging CCD camera 65 or in a conjugate retinal image plane.

(ii) Improved retinal implant 50 microchips when using the optics of the human eye 80 (such as in the Retina Implant AG technology). Although resolution of current implant technology is low (due to the small number of pixels) it is to be expected that future implants will have smaller pixels with high density whereby the need for improved visual acuity and discrimination against unwanted light components becomes increasingly important. When overlaid to a photosensitive material, the retinal photoreceptor mosaic will reproduce the natural angular filtering of photoreceptors in the healthy human eye fundus.

As mentioned above, the photoresist sample was successfully developed experimentally with printed dielectric cylinders with elevated refractive index in a hexagonal lattice of 5 μηι circles separated by 10 μηι, similar to the distribution of retinal cones in the vicinity of the foveal region. The transmission efficiency of the structure as a function of angle of incidence was analysed and verified to decrease for increasing angles of incidence with a characteristic Stiles-Crawford directionality of 0.01 /mm² which is similar to the photoreceptor rods. Other photoresists or other chemical processing may further increase the directionality to a range commonly associated with cones, namely, 0.05 - 0.10 /mm². Likewise, non-cylindrical refractive index elevations are also possible using, for example, cones or cuboids that may be designed to better match the shape of the pixels in the light-sensitive electronic chips.

It is envisaged that the retinal photoreceptor mosaic simulator 40 of the present invention can be used for optical image transmission that together with a CCD camera (or screen) 65 acts as an artificial retina in an advanced eye model 30 where aberrations and light scattering can be controlled to analyse its predicted visual performance. Fig. 6 shows an example of images captured with a CCD camera and using an optical diffuser to simulate intraocular scattering (a) without and (b) with the retinal model inserted in front of the camera. The angular selectivity of the resist structure is still low when compared to the eye but with larger refractive index difference, possibly using another photoresist, or with rescaling optics before the imaging CCD camera, directionality similar to that of the human retina should be feasible.

In summary, the retinal photoreceptor mosaic of the present invention will reproduce/mimic the natural angular filtering of the retinal photoreceptors in the eye. Accordingly, the present invention adds a significant improvement to existing eye models and retinal implants which do not take into account the angular filtering of the retinal photoreceptors. Although based on the eye and retinal implants, the same technology may be integrated into image capturing devices such as cameras to reduce their angular response and thereby make them less sensitive to aberrations or scattering of light. This may well prove useful for cameras such as those in mobile phones or in web- cameras.

In an alternative embodiment, the method of manufacturing the retinal photoreceptor mosaic simulator 40 comprises manufacturing (for example 3D printing) the retinal photoreceptor mosaic simulator 40 with two different materials which will likewise allow for two different refractive indices. This could be an alternative to the use of photoresist.

Further experiments and simulations were performed with the photoresist sample discussed above. These experiments and simulations will now be discussed with reference to figures 7(a) to 10(b).

1. With respect to the directional sensitivity of the sample Figure 7(a) shows a cross-section of light distribution for different values of length from 10μηι to 100μηι modelled in Comsol™ of an infinite array of high-index cylinders, 5μηι diameter, separated by 10μηι in an hexagonal transparent lattice, with n_COre = 1 .595 and n_C|_aCi =1.591 and wavelength λ= 543nm. Figure 7(b) shows a cross-section of light distribution in the presence of absorption in the surrounding cladding for different values of length from ΙΟμηι to ΙΟΟμητ modelled in Comsol™ of an infinite array of high-index cylinders, 5μτη diameter, separated by ΙΟμηι in an hexagonal transparent lattice, with n_core = 1.595 and ⁿciad = 1-591 and wavelength λ = 543nm.

Figure 8(a) shows transmitted power fraction through the cores as a function of angle of incidence: (I) without absorption in the surround cladding; (II) with absorption. Figure 8(b) shows directional sensitivity as a function of waveguide length with and without absorption in the surround cladding. The presence of the absorber avoids the leakage of light to neighbouring cylinders, increasing the directionality of the waveguides.

The guided modes of an optical fiber are formed by total internal reflection (TIR) at the core, causing light rays to self-interfere. However, the short propagation distance would not allow the radiation modes to be properly dissipated. As a result, a combination of guided and non-guided light is measured at the exit of the waveguides, reducing its effective directionality.

The dependence on length of the refractive-index-enhanced cylinders (L) in a hexagonal infinite lattice was analysed using the software Comsol Multiphasics™. The simulations were performed in 3D, assuming the same refractive indices, geometry and distribution as in the photoresist. The results are presented in Figs. 7(a)-8(b).

For L < 50μηι the light distribution across the array shows a poor confinement of light at the core, above it, the guiding through the core becomes perceptible (Fig. 7(a)). The desired decoupling between guided and radiative modes is translated into an increasingly narrow distribution of transmitted power through the high-index cylinders (Fig. 8(a)(ll) (right)). The directional coupling at small angles (Θ < 6.5°) was calculated using the Gaussian distribution function η = 10^~pr2 showing a sigmoidal growth of the directional parameter, p, as a function of waveguide length (Fig. 8(b)).

At larger angles (Θ > 6.5°) the incident light is above the acceptance angle and is no longer guided. However, due to the short distances and the transparency of the film, the light radiates through neighbouring cylinders. This leakage could be avoided by the introduction of an absorber in the surround cladding, that would also assist the dissipation of the radiant modes, increasing the directionality even further (see Figs. 7(b), 8(a) and 8(b)). For a 50μηι long array in the presence of absorption (n_clad = 1.591 - 0.010, the directionality parameter was calculated as p = 0.053mm^-2 , reaching the range commonly associated with foveal cones in the human retina. 2. With respect to the improvement of the image quality

Figure 10(a) shows an analysis of the contribution of the retina phantom to the visual performance. An USAF 1951 paper target was used as object. In the images are observed group -2 and elements 4 and 5 of the target. On the left, resultant images with a photoresist without printed cylinders; on the right, the retinal images with printed cylinders. Different defocus conditions were analysed from 0μm displacement out of the focal plane, up to 195μm. The blurring progression is slower in the presence of the array.

Figure 10(b) shows Matlab™ simulations of the projected image. The images present defocus aberration generated by a displacement of l_i in 195μηι out of the focal plane and were low-pass angular filtered by retinal simulators with different p values.

The contribution of the retina phantom (retinal photoreceptor mosaic simulator) on the visual performance was analysed. As shown in Fig. 9, a USAF 1951 paper target 100 was used as an object, illuminated by white light and imaged by an artificial eye 120, composed of an iris (Φ = 6mm) 130 and achromatic lens (L1 , / = 19mm) simulating the anterior eye. The photoresist 140 is placed at the image plane which, in turn is conjugated with a CCD camera 150 by a microscope 160. To compare the effect of the low-pass angular filtering, the retinal images were analysed at two different regions of the photoresist 140: with printed cylinders and without printed cylinders (see Fig. 10(a)). Furthermore, a set of images shifted gradually in the xy plane is digitally combined to fill the gaps between the waveguides. That shifting would, in the human eye, be accomplished by tremor and saccades.

As shown in Fig. 10(a), no appreciable difference is observed between the images obtained with or without printed cylinders when the target is in focus (planar wavefront). However, in the presence of defocus, the images with printed cylinders show improvement in contrast and resolution, with a slower progression of the blurring. The image filtering would be even more pronounced with a retinal phantom (retinal photoreceptor mosaic simulator) with higher p. The potential enhancement of contrast and resolution was simulated using Matlab™ with L₁ displaced 195mm out of focal plane and p = 0.00mm^-2 up to p = 0.10mm^-2 (Fig.10(b)). The implementation described above and illustrated in the drawings is just one possible implementation (with variations as described). The examples described are purely illustrative and the skilled reader will appreciate that many further modifications and variations are possible within the scope of the invention described herein.

Claims

1 . A method of manufacturing a retinal photoreceptor mosaic simulator, the method comprising the steps of:

depositing a photoresist onto a substrate;

exposing the photoresist with radiation to define a pattern of structures on the photoresist, the pattern corresponding to a distribution of retina photoreceptors cells in a retina of an eye; and

producing a film comprising a latent image defining a plurality of dielectric structures;

wherein a refractive index contrast between the plurality of dielectric structures and their surrounds in the latent image corresponds to a refractive index contrast of the retina photoreceptor cells and their surrounds in the retina of the eye.

2. The method of claim 1 , wherein a distribution of the plurality of dielectric structures in the latent image corresponds to a distribution of retina photoreceptor cells in the retina of the eye.

3. The method of claim 1 or claim 2, wherein the plurality of dielectric structures are spaced apart from each other in the latent image and a refractive index of a dielectric structure is higher than a refractive index of a space between two of the plurality of dielectric structures.

4. The method of any preceding claim, wherein the producing step comprises: post-exposure baking the photoresist to catalytically complete a photoreaction initiated during the exposing step.

5. The method of any preceding claim, wherein the exposing step comprises: overlaying the photoresist with a photomask; the photomask having the pattern of structures corresponding to the distribution of retina photoreceptors cells in the retina of the eye; and

illuminating the photoresist via the photomask with an ultra violet light source.

6. The method of claim 5, wherein the ultra violet light source is a collimated ultra violet light source.

7. The method of any of claims 1 to 4, wherein the exposing step comprises: scanning a focused beam of ultra violet light over the photoresist to form the pattern of structures on the photoresist.

8. The method of any of claims 1 to 4, wherein the exposing step comprises: scanning infra-red laser beams over the photoresist to form the pattern of structures on the photoresist.

9. The method of any of claims 1 to 4, wherein the exposing step comprises: scanning an electronic beam over the photoresist to form the pattern of structures on the photoresist.

10. The method of any preceding claim, wherein each of the plurality of dielectric structures is about 2 to about 8 microns in diameter and wherein the plurality of dielectric structures are spaced apart from each other by a distance of about 2.5 microns to about 13 microns.

1 1 . The method of claim 10, wherein each of the dielectric structures is about 5 microns in diameter and wherein the plurality of dielectric structures are spaced apart from each other by a distance of about 10 microns.

12. The method of any preceding claim, wherein the plurality of dielectric structures are substantially cylindrical in shape.

13. The method of any of claims 1 to 1 1 , wherein the plurality of dielectric structures are substantially non-cylindrical in shape.

14. The method of any preceding claim, wherein the photoresist is a positive photoresist.

15. The method of claim 14, wherein the photoresist is a positive photoresist of the type AZ40XT.

16. The method of any of claims 1 to 13, wherein the photoresist is a negative photoresist.

17. The method of claim 16, wherein the photoresist is a negative photoresist of the type SU-8.

18. The method of any preceding claim, wherein the substrate is one of a glass substrate, a silica substrate, a fused silica substrate or a polymer substrate.

19. The method of any preceding claim, wherein the depositing of the photoresist onto the substrate comprises one of spin coating, dip coating or spray coating the photoresist onto the substrate.

20. A retinal photoreceptor mosaic simulator comprising:

a film comprising a latent image defining a plurality of dielectric structures; wherein a refractive index contrast between the plurality of dielectric structures and their surrounds in the latent image corresponds to a refractive index contrast of retina photoreceptor cells and their surrounds in a retina of an eye.

21 . The retinal photoreceptor mosaic simulator of claim 20, wherein a distribution of the plurality of dielectric structures in the latent image corresponds to a distribution of retina photoreceptor cells in the retina of the eye.

22. The retinal photoreceptor mosaic simulator of claim 20 or claim 21 , wherein the plurality of dielectric structures are spaced apart from each other in the latent image and a refractive index of a dielectric structure is higher than a refractive index of a space between two of the plurality of dielectric structures.

23. The retinal photoreceptor mosaic simulator of any of claims 20 to 22, wherein the plurality of dielectric structures are configured to act as a waveguide for guiding light from a retinal image plane to a detection plane.

24. The retinal photoreceptor mosaic simulator of claim 23, wherein each one of the plurality of dielectric structures is configured to act as an independent waveguide for guiding light from a retinal image plane to a detection plane.

25. The retinal photoreceptor mosaic simulator of any of claims 20 to 24, wherein the plurality of dielectric structures are configured to act as an angular low-pass filter for dampening the influence of oblique light rays.

26. The retinal photoreceptor mosaic simulator of any of claims 20 to 25, wherein each of the plurality of dielectric structures is about 2 to about 8 microns in diameter and wherein the dielectric structures are spaced apart from each other by a distance of about 2.5 microns to about 13 microns.

27. The retinal photoreceptor mosaic simulator of claim 26, wherein each of the dielectric structures is about 5 microns in diameter and wherein the dielectric structures are spaced apart from each other by a distance of about 10 microns.

28. The retinal photoreceptor mosaic simulator of any of claims 20 to 27, wherein the plurality of dielectric structures are substantially cylindrical in shape.

29. The retinal photoreceptor mosaic simulator of any of claims 20 to 27, wherein the plurality of dielectric structures are non-cylindrical in shape.

30. An image capturing system comprising the retinal photoreceptor mosaic simulator of any of claims 20 to 29.

31 . The image capturing system of claim 30, wherein the image capturing system comprises one or more of: an eye model, a retinal model, a retinal implant, a bionic eye implant, a bionic lens implant, an intraocular lens, a corneal inlay, spectacle glasses, a contact lens or a camera.

32. A method, a retinal photoreceptor mosaic simulator, or an image capturing system substantially as described herein with reference to the accompanying drawings.