- Email: [email protected]
Feasibility of extraocular stimulation for a retinal prosthesis
Feasibility of extraocular stimulation for a retinal prosthesis Vivek Chowdhury,* MBBS; John W. Morley,† PhD; Minas T. Coroneo,* MD, FRANZCO ABSTRACT • RÉSUMÉ Background: We present a new approach to developing a retinal prosthesis for blind patients based on extraocular stimulation of the eye with disc electrodes. Methods: Experiments to assess the feasibility of using extraocular stimulation in a retinal prosthesis were carried out in anaesthetised adult cats (n = 6).A craniotomy and lateral orbital dissection were performed. Ball or disc electrodes were placed on the posterior scleral surface of the eye after incision of the periorbita. Cortical potentials evoked by electrical stimulation with these electrodes were recorded at the primary visual cortex. The viability of adapting the Nucleus 24 auditory brainstem implant (ABI) as an extraocular retinal prosthesis was also investigated. Results: Electrodes placed on the exterior of the eye could reliably evoke visual cortex responses for a variety of configurations.Threshold currents for eliciting an evoked response were lower than 100 μA with single pulses. Strength–duration curves and cortical activation maps were obtained for different stimulus paradigms. It was possible to excite the retina to evoke a cortical response using the electrodes and stimulus capabilities in a standard Nucleus 24 ABI. Interpretation: It is possible to electrically stimulate the retina with electrodes placed in an extraocular location.Threshold currents required to elicit a response were low, and comparable to epiretinal implants. Prototype electrodes, and a potential implant, were found to be effective at retinal stimulation. Contexte : Nous présentons une nouvelle approche pour développer une prothèse rétinienne pour patients aveugles, fondée sur la stimulation extraoculaire avec électrodes à disque. Méthodes : Des expériences visant à évaluer la possibilité d’appliquer une stimulation extraoculaire à une prothèse rétinienne ont été menées chez des chats adultes sous anesthésie (n = 6). L’on a pratiqué une craniotomie et une dissection orbitale latérale. Des électrodes en boule ou à disque ont été placées à la surface sclérale postérieure de l’œil après incision de la périorbite. Les potentiels corticaux suscités par la stimulation électrique avec ces électrodes ont été enregistrés dans le cortex visuel primaire. On a aussi examiné les chances de réussite de l’implant du tronc cérébral Nucleus 24 auditory brainstem implant (ABI) comme prothèse rétinienne extraoculaire.
From *the Department of Ophthalmology, Prince of Wales Clinical School, and †the School of Medical Sciences, University of New South Wales, Sydney, Australia Originally received Jul. 8, 2004 Accepted for publication Jan. 10, 2005
Retinal prosthesis—Chowdhury et al
Correspondence to: Dr. Vivek Chowdhury, PO Box 140, Randwick, NSW, 2031, Australia; fax 61-2-93851059; [email protected] This paper has been peer-reviewed. Can J Ophthalmol 2005;40:563–72
563
Retinal prosthesis—Chowdhury et al
Résultats : Les électrodes placées à l’extérieur de l’œil pouvaient susciter dans le cortex visuel des réactions à diverses configurations. Le seuil des courants pour susciter une réaction était inférieur à 100 μA avec pulsations simples. Les courbes intensité–durée et les cartes d’activation ont été obtenues pour divers paradigmes de stimulation. Il a été possible d’exciter la rétine pour susciter une réaction corticale avec des électrodes et les capacités de stimulation d’un Nucleus 24 ABI standard. Interprétation : Il est possible de stimuler électriquement la rétine avec des électrodes placées à l’extérieur de l’œil. Le seuil minimum des courants pour susciter une réaction était bas, et comparable aux implants épirétiniens. Des prototypes d’électrode et un implant potentiel se sont avérés efficaces pour stimuler la rétine.
R
etinitis pigmentosa is the most common cause of blindness in people under 70 years of age in the developed world. One and a half million people are affected with this disease worldwide.1 There is currently no clinical means of restoring visual perception to patients who have developed blindness (no light perception) as a result of retinitis pigmentosa or other inherited or acquired retinal dystrophies. There is also no clinical means of preventing the deterioration towards blindness that occurs in the retinal dystrophies.1,2 Experimental treatments such as gene therapy, neural cell transplantation, growth factors, vitamin supplementation, antioxidants, and regulators of apoptosis have shown no success in restoring visual sensations to blind patients in clinical trials.1 It is unlikely that any of these methods will be able to restore visual perceptions to blind patients in the medium term (5–10 years), if at all.2 This is due to the poor capacity of neural tissue, such as that in the retina, for repair and regeneration.3 Therefore, a way to restore vision based on bypassing the damaged elements of the visual pathway is required.4,5 The only experimental method that has successfully restored visual perceptions to irreversibly blind patients is through the electrical stimulation of eye, optic nerve, or brain with implanted electrodes.2,6–9 Of these 3 sites, electrical stimulation of the visual cortex is the only approach, at this time, to developing a bionic eye that has restored visual perceptions sufficient to increase the patient’s mobility and independence.10 Despite an intensive research effort by a number of well-funded groups over the past 10 years aimed at developing a visual prosthesis with intraocular electrodes at epiretinal9,11–13 or subretinal14,15 locations, electrical stimulation of the retina has not met with the success achieved by stimulation of the visual cortex.10,16 This is because there are significant prob-
564
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
lems to be overcome at the electrode–tissue interface before an intraocular retina-based bionic implant can be a viable treatment for blindness.16,17 There are 2 chronic human trials of intraocular retinal implants occurring. In a study by Chow et al, the authors note improvements in the visual acuity of implanted patients following subretinal stimulation that cannot be explained by a neuroprosthetic effect of the implanted device, leading them to suggest that low-level electrical stimulation of the retina is having an as-yet-undefined “neurotrophic effect.”18 In the second study by Humayun and colleagues, a 4 × 4 array of 16 platinum disc electrodes, each 520 μm diameter, with an interelectrode centre-to-centre spacing of 720 μm, was implanted with a tack at the epiretinal surface of a blind human patient.19 A cable from the electrode array passed through the sclera and tracked subcutaneously to a stimulator outside the orbit. Electrical stimulation of this array elicited the subjective perception of small spots of light in the patient’s visual field. Interfacing the device with a camera allowed the patient to detect the presence of light and large objects.19 We are investigating the feasibility of extraocular stimulation of the retina with an electrode array placed on the posterior scleral surface of the eye. Benefits of such an approach include no intraocular surgery, no implantation of an intraocular foreign body, and no cable passing through the sclera, which would leave a permanent defect in the integrity of the globe. We plan to adapt the neuroprosthetic technology that is available in the Nucleus 24 auditory brainstem implant (ABI) (Cochlear Ltd., Lane Cove, Australia). This implant has an electrode array of 21 platinum disc electrodes, each 700 μm diameter, with a 950-μm centre-to-centre interelectrode spacing, arranged in 3 rows in a silicone carrier.
Retinal prosthesis—Chowdhury et al
We present the results of a preliminary study of an extraocular retinal prosthesis, an approach to visual prosthesis development for which a device has hitherto not yet been developed.7,17,19 Our study investigated a number of electrode types, configurations, and stimulus methods to optimize the development of an extraocular retinal implant for the restoration of visual perception to blind patients. METHODS
Acute experiments were carried out in anaesthetised cats in accordance with the guidelines of the animal ethics committee of the University of New South Wales. In adult cats (n = 6) weighing between 2.5 kg and 5.5 kg, anaesthesia was induced with an intramuscular injection of ketamine (20 mg/kg) and xylazine (1 mg/kg). The animals were given a preoperative dose of subcutaneous atropine (0.2 mg/kg) and dexamethasone (1.5 mg/kg). After intubation, the animals were ventilated, and anaesthesia was maintained with a 70:30 mixture of nitrous oxide and oxygen, with 0.5% to 1% halothane. The electrocardiogram, end-tidal carbon dioxide, and core body temperature were monitored, and animals were checked regularly for the absence of reflexes to ensure adequate anaesthesia. At the end of the experiment, the cats were euthanased with an intravenous injection of pentobarbital. The cats underwent a bilateral craniotomy of the parietal bones, and the dura was removed to expose the primary visual cortex (cytoarchitectonic area 17) of both hemispheres. The right orbit was approached by removing the bone of the lateral orbital wall and retracting or removing the right temporalis muscle. The cerebral cortex was regularly irrigated with warmed (35°C) paraffin oil. The cats were placed in a stereotaxic frame, which provided a system of 3-dimentional coordinates used during neurosurgery to give a precise atlas of the brain. Platinum or silver ball electrodes were placed on the pial surface of the primary visual cortex. For the stimulating electrodes on the posterior scleral surface of the eye, we investigated a variety of types and configurations. This included silver ball electrodes (1 mm diameter), flat platinum disc electrodes of 2 and 4 mm diameter (Cochlear Ltd.), and a multielectrode array consisting of 21 platinum disc electrodes, each 700 μm diameter (Cochlear Ltd.). In some experiments, a contact-lens electrode was used as the stimulus return path (ERG-jet, Universo, La
Chaux-de-Fonds, Switzerland). We also experimented with different surgical techniques for implanting the electrode array. Electrical stimulation of the right retina by the extraocular electrodes evoked cortical potentials that we recorded using Scope 3.6.11 software and a PowerLab/4SP data acquisition system connected to a ML135 biopotential amplifier (ADInstruments Ltd., Castle Hill, Australia). Evoked potentials were averaged over 100 to 200 trials, after being filtered to within a frequency range of 10 Hz to 5 kHz, using a 50-Hz notch filter. A differential recording system was used, with an indifferent scalp clip and a ground wire connected to a right thigh subcutaneous pin. The integrity of the neural visual pathway was confirmed with recordings of electroretinograms and visual evoked potentials. Electrical stimuli were rectangular, constantcurrent, symmetrical, biphasic pulses delivered from a 2100 Isolated Pulse Stimulator (A-M Systems, Inc., Carlsborg, Wash.) or monophasic pulses from an ML180 Stimulus Isolator (ADInstruments Ltd.). Biphasic stimuli were also delivered using a Nucleus 24 ABI. RESULTS Ball electrodes
We investigated monopolar and bipolar extraocular stimulation of the retina with silver ball electrodes. In cat A, the stimulation electrode was placed on the posterior scleral surface of the eye, 5 mm lateral to the optic nerve attachment, in the horizontal meridian. In this animal, a 4-mm platinum disc electrode was implanted on the inferonasal aspect of the globe to serve as the current return path. In cat B, 2 silver ball electrodes with a 5-mm interelectrode spacing were placed at the central posterior surface of the globe, superolateral to the optic nerve head. The electrical evoked response to cathodal-first biphasic stimulation in a monopolar configuration from cat A was compared with the response to monophasic stimulation of bipolar electrodes in cat B (Fig. 1). Bipolar stimulation in cat B showed a typical negative–positive electrically evoked response.20 However, because of a prolonged artifact in cat A, which is the case with monopolar stimulation,21 the initial negative component of the cortical response in this cat was obscured (a comparison with control waves was performed). The first visible true response
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
565
Retinal prosthesis—Chowdhury et al
By determining the threshold current necessary to elicit a response at the visual cortex, we obtained a strength–duration curve for biphasic monopolar extraocular electrical stimulation of the retina (Fig. 2A). For phase durations greater than 200 μs, threshold currents for eliciting a visual cortex response were below 1.5 mA. We also determined the peak amplitude of the positive component of the evoked response in cat A for a range of stimulus current levels at a pulse width of 250 μs (Fig. 2B). Higher response amplitudes were obtained with increasing current up to 5 mA, after which there was a response plateau. Multipulse studies
Fig. 1—Evoked response at the primary visual cortex to retinal stimulation with 1-mA stimuli of 400-μs phase duration using ball electrodes. Before 20 ms, a stimulus artifact is punctuated by sharp afferent volley spikes.After 20 ms, bipolar stimulation in cat B shows a typical negative–positive evoked response with a negative peak at an average of 27 ms and a positive peak averaging 40 ms. With monopolar stimulation in cat A, a prolonged stimulus artifact (common with monopolar stimuli) obscures the first negative component of this wave. The first true response is a positive peak at 38 ms followed by another positive peak at 54 ms.
Table 1—Latency of evoked response of visual cortex to extraocular stimulation of the retina in cats. Mean time of response, ms (SD) (0) N1
(+) P1
(+) P2
Cat A*
—
36.53 (1.83)
51.70 (2.37)
Cat B
27.32 (1.42)
40.85 (3.48)
—
Peak
Note: ms = millisecond; SD = standard deviation; N = negative; P = positive. *Initial negative peak in cat A is obscured because of prolonged stimulus artifact from monopolar stimulation. Cat A exhibited 2 defined positive peaks, whereas cat B (bipolar stimulation) exhibited a single early positive peak.
in cat A is a positive peak at 38 ms. A second positive peak follows this wave at 54 ms. The average latency of the negative and positive peaks for cat B, and the 2 positive peaks for cat A, are shown in Table 1.
566
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
We investigated the effect of pulse trains on cortical response amplitude. The amplitude of the positive wave of the electrical evoked response was recorded for 2-mA, 250-μs cathodal-first biphasic stimuli to single pulses and pulse trains of 2, 3, or 4 pulses presented at 200 Hz. We also compared these responses to the amplitude of the cortical potential evoked by a double pulse at 100 Hz (Fig. 3). The highest cortical responses were obtained with 3 pulse trains. For double pulses, a longer interval between pulses evoked a larger cortical response. Cortical activation map
Extraocular stimulation of a localised region of the sclera leads to excitation of a localised region of the retina, which evokes localised cortical responses. We undertook monopolar extraocular stimulation of the retina with a ball electrode placed in the horizontal meridian 5 mm lateral to the attachment of the optic nerve on the globe. We used biphasic pulses of 250 μs pulse width and 2 mA current and recorded evoked responses at localised points on the primary visual cortex over the posteromedial regions of the lateral gyri of both hemispheres. Using a 1-mm silver-ball recording electrode, we moved anteroposteriorly in 1-mm steps to record evoked cortical potentials from HorsleyClarke stereotaxic coordinates A1 to P8 (Fig. 4). We recorded from the most medial position of the superior surface of the lateral gyrus that was accessible after a craniotomy that had preserved a narrow strip of bone overlying the sagittal sinus. Cortical responses in cat A were localised in the anteroposterior plane to Horsley-Clarke coordinates P6 and P7 in the left and right hemispheres, respectively. There was a second smaller peak of cortical activity anteriorly in
Retinal prosthesis—Chowdhury et al
Fig. 3—The effect of stimulus pulse trains on the electrically evoked response for frequencies of 100 Hz and 200 Hz.
Fig. 2—A: Strength–duration curve for extraocular retinal stimulation with biphasic cathodal–anodal symmetrical pulses using ball electrodes in a monopolar configuration.The threshold for detecting a reliable cortical response over background noise was defined as a positive peak amplitude of greater than 3 μV. B: Amplitude of the positive peak of the cortical evoked response to extraocular retinal stimulation with 250-μs cathodalfirst biphasic pulses.
the left hemisphere at P2, which was not present in the right hemisphere. Disc electrodes
We investigated extraocular stimulation of the retina with flat platinum disc electrodes in 3 different
configurations (Fig. 5A). In cat D, a 2-mm active electrode was sutured under the superior rectus muscle and a 4-mm return electrode was placed anterior to the attachment of the lateral rectus muscle (Fig. 5B). In cat E, a 2-mm active electrode was placed in the inferolateral quadrant of the right eye and a 4-mm return electrode was placed anterior to the attachment of the medial rectus muscle. In cat F, a 2-mm platinum disc was applied to the sclera, 5 mm lateral to the optic nerve attachment in the horizontal meridian, and a contact-lens electrode applied to the cornea was used for the current return path. The strength–duration curves for cortical activation with the electrode configurations in cats D and F is shown in Fig. 6. Electrical stimulation of the eye in cat E produced a right-cortex cortical activation map, which is shown along with that of cat A in Fig. 4. Anodal stimulation was more effective than cathodal stimulation at exciting the retina (cat F), demonstrated by the upward shift of the strength–duration curve. The curves flattened towards rheobase (the minimal strength of the electrical stimulus able to cause excitation) at pulse widths greater than 500 μs. Increasing the pulse width of the stimulus up to 1 ms led only to small decreases in the threshold current required for retinal activation. Cathodal stimulation was more effective with the electrode configuration of cat F than with that of cat D. In cat F, it was possible to obtain cortical responses from retinal stimulation
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
567
Retinal prosthesis—Chowdhury et al
Fig. 4—Bilateral cortical activation map for monopolar stimulation of the retina with an extraocular electrode placed 5 mm lateral to the attachment of the optic nerve in the horizontal meridian (cat A).The x-axis denotes Horsley-Clarke coordinates from 5 (A5) to –8 (P8). Also shown is the right-cortex cortical activation map for electrical stimulation with disc electrodes in cat E (see Fig. 5).
with threshold currents of less than 100 μA at pulse widths greater than 400 μs. Multielectrode array
We tested the components of a prototype extraocular retinal prosthesis in 2 cats. In cat C, we implanted a 21-electrode array along the horizontal meridian of the posterior sclera of the right globe, and bipolar biphasic pulses were delivered with a bench stimulator. In cat F, extraocular stimulation of the retina was carried out with a Nucleus 24 ABI, using neural response telemetry software and a portable programming system (Cochlear Ltd.). Strength–duration curves for both bench and implant stimulation are shown in Fig. 7. Electrode impedance measurements obtained with the reverse telemetry feature of the implant (cat F) gave a reading of 15.6 kΩ in the monopolar configuration (single grid electrode active, contact-lens electrode as the corneal return) and 8 kΩ in the “common ground” configuration (single grid electrode active, all other 20 electrodes on the grid shorted to the return path). Using bipolar stimulation (bench) of the electrode array, threshold currents were lower than 1 mA for pulse durations greater than 300 μs. Stimulation
568
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
levels with the implant are programmed in “device units” up to 255, which is the strongest stimulation level using the unit, equal to approximately 1.75 mA. Pulse widths up to the device maximum of 400 μs were investigated. Monopolar stimulation was more effective than common-ground stimulation. No responses could be elicited with a pulse width below 400 μs in common-ground mode. In monopolar configuration, using pulse widths of 100 μs and 200 μs, stimulation intensities at the upper limits of the device were required. INTERPRETATION
Electrical stimulation of the retina can elicit visual perceptions in blind patients,9 commonly described as small spots of light termed “phosphenes.” Major efforts are underway to develop a retinal prosthesis for blind patients that will restore a level of visual perceptions that is functionally useful.16,17,19,22,23 A medical device for this purpose must be surgically feasible24 and physically biocompatible to the body,25 and it must utilize safe stimulation parameters26,27 that give the highest possible number of stable, reproducible, and resolvable phosphene sensations.13 This study assessed the feasibility of using an extraocular
Retinal prosthesis—Chowdhury et al
A
B
SR
LR
IO Fig. 5—A:Three configurations of electrode placement on the eye. In cat D (blue), a 2-mm disc was placed under the superior rectus (SR) and a 4-mm return electrode was placed anterior to the lateral rectus (LR). In cat E (red), a 2-mm disc was placed inferior to the LR and posterior to the inferior oblique (IO) in the inferolateral quadrant of the eye. A 4-mm return electrode was placed anterior to the medial rectus (dotted). In cat F (green), a 2-mm disc was placed in the horizontal plane, 5 mm lateral to the attachment of the optic nerve, and stimulated with respect to a contact-lens electrode placed on the cornea. B: Picture of the right eye in cat D after implantating the electrodes and suturing the electrode tails to the periorbita.
approach to electrical stimulation of the retina, and we have proposed a device for this purpose. In 6 cats, it was possible to record evoked responses at the visual cortex to electrical stimuli applied to the retina with electrodes placed on the posterior surface of the globe. We investigated electrical stimulation with different electrode types in a variety of configurations (Fig. 5A), all of which were able to evoke visual cortex responses. It has been postulated that activity at the primary visual cortex is correlated to conscious perception,28 and we expect that the activity evoked in the visual pathway through this method of stimulation will likely evoke phosphene perception in human patients. We investigated the feasibility of implanting both single electrodes and electrode arrays on the surface of the sclera. In the cat shown in Fig. 5B, the disc electrode was slipped under the superior rectus muscle, and the electrode tail was then sutured to this muscle and then again to the periorbita with a purse-string suture. Local paralysis of the extraocular muscles to prevent movement will be necessary in future chronic implantations. The placement of extraocular electrodes requires neither intraocular surgery nor intraocular placement of foreign bodies against the delicate retinal tissues. It also does not require complicated methods, such as tacks, to attach the electrode arrays to the retinal surface.19 It would be possible in the future to use biological adhesives or
sutures to fix the silicone base, which carries the electrode, directly to the sclera. The firm scleral base for extraocular electrodes avoids mechanical trauma to the retina and decreases the likelihood of pathological effects due to charge transfer and heat generation.30 We examined the thresholds needed to elicit cortical evoked responses with extraocular electrodes. Current thresholds were variable amongst cats with different electrode configurations. The lowest thresholds were found with a 2-mm electrode placed on the posterior of the globe and stimulated with respect to a contact-lens electrode placed on the cornea. The threshold for eliciting an evoked response was as low as 100 μA (Fig. 6) at a pulse duration of 400 μs (0.04 μC per phase). This translates to a charge density at the electrode surface of 1.27 μC/cm2, which is well within safe charge injection levels in human studies for chronic neural stimulation.29 Higher thresholds occurred with monopolar stimulation using the Nucleus 24 ABI (Fig. 7). A threshold of 225 device units (255 units = 1.75 mA) was obtained when using 400-μs cathodal-first biphasic pulses. We assessed the ability of extraocular stimulation to produce localised phosphenes by investigating if extraocular stimulation produced localised activation at the primary visual cortex. The medial edge of the lateral gyrus (cytoarchitectonic area 17) contains a retinotopically mapped representation of the visual field.31 In cat A, the cortical response was localised
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
569
Retinal prosthesis—Chowdhury et al
Fig. 6—Strength–duration curves for current (mA) and pulse width (ms) for cat D and cat F using retinal stimulation with disc electrodes. In cat D, monopolar stimulation was performed with cathodal-first biphasic pulses. In cat F, monopolar stimulation was performed with cathodal or anodal monophasic pulses.
Fig. 7—Strength–duration curves for scleral stimulation with the 21-electrode array. Stimuli from implanted monopolar and common-ground electrode configurations (cat F) are measured in mA (right y-axis), and bipolar stimulation with a bench stimulator (cat C) is measured in device units (left y-axis, where 255 device units are ~1.75 mA). Monopolar stimulation involved activating one 700-μm platinum disc on the array and using a contact-lens electrode as the return path. Common-ground stimulation involved stimulating one active electrode from the array with respect to the remaining 20 electrodes, which were shorted to the return path. Bipolar stimulation with the bench stimulator utilized electrical stimulation between 2 electrodes on the array with a 2.85-mm centre-to-centre spacing. For the common-ground stimulus configuration, responses could only be obtained with a pulse width of 400 μs, which is the maximal pulse width programmable with the implant.
570
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
Retinal prosthesis—Chowdhury et al
posteriorly at a similar anteroposterior coordinate in both hemispheres (Fig. 4). Studies of pulse summation will aid in the identification of optimal stimulus parameters for an extraocular retinal prosthesis. Ideal frequencies for pulse trains were investigated with multipulse studies. Higher amplitudes of the evoked cortical responses were obtained with paired pulses at 100 Hz compared with those at 200 Hz. For 200-Hz stimuli, peak responses were obtained with 3 pulse trains (Fig. 3). As expected, anodal stimuli were less effective than cathodal stimuli in eliciting extracellular stimulation of retinal tissue (Fig. 6, cat F).21,30 We undertook a basic feasibility study of adapting technology in the Nucleus 24 ABI for extraocular retinal stimulation. This device is composed of an array of 21 platinum disc electrodes in a silicone carrier with a stimulator implant that can be powered and controlled transcutaneously. This array could elicit stimulation of the sclera under a variety of electrode configurations. The stimulator implant itself was able to evoke visual cortex responses with single biphasic pulses (Fig. 7). An extraocular retinal implant may possibly overcome some of the difficulties of using epiretinal stimulation for a visual prosthesis.16 We have shown the feasibility of such an approach by demonstrating an ease of surgical implantation, safe current thresholds for evoking cortical responses, and localised activation of the primary visual regions in the brain. Adaptation of the technology in the Nucleus 24 ABI will provide a basis for developing an extraocular prosthesis for human trial. Financial support for this work was received from the Ophthalmic Research Institute of Australia, the University of New South Wales, Retina Australia, the Brain Foundation, and the Australian National Health and Medical Research Council. Some equipment that was used in this project was provided at no expense by Cochlear Ltd. The 3 authors are directors and shareholders in Sydney Biotech Pty Ltd., which is collaborating with Cochlear Ltd. and the University of New South Wales to develop a visual prosthesis.
4. 5. 6.
7. 8. 9.
10. 11.
12. 13.
14. 15. 16.
17. 18.
19.
20.
21. 22.
REFERENCES 23. 1.
2. 3.
Sharma RK, Ehinger B. Management of hereditary retinal degenerations: present status and future directions. Surv Ophthalmol 1999;43:427–44. Chong NH, Bird AC. Management of inherited outer retinal dystrophies: present and future. Br J Ophthalmol 1999;83: 120–2. Mitchell RN, Cotran RS. Repair: Cell Regeneration, Fibrosis,
24.
25.
and Wound Healing. In: Kumar V, Cotran RS, Robbins SL, eds. Basic Pathology. 6th ed. Philadelphia, Pa: WB Saunders, 1997:chap 3. Scarlatis G. Optical prosthesis: visions of the future. JAMA 2000;283:2297. Larkin M. Artificial-vision research comes into focus. Lancet 2000;355:1080. Veraart C, Raftopoulos C, Mortimer JT, et al. Visual sensations produced by optic nerve stimulation using an implanted self-sizing spiral cuff electrode. Brain Res 1998;813:181–6. Maynard EM. Visual prostheses. Annu Rev of Biomed Eng 2001;3:145–68. Brindley GS, Lewin WS. The sensations produced by electrical stimulation of the visual cortex. J Physiol 1968;196: 479–93. Humayun MS, de Juan E Jr., Dagnelie G, et al. Visual perception elicited by electrical stimulation of retina in blind humans. Arch Ophthalmol 1996;114:40–6. Dobelle WH. Artificial vision for the blind by connecting a television camera to the visual cortex. ASAIO 2000;46:3–9. Grumet AE, Wyatt JL, Rizzo JF. Multi-electrode stimulation and recording in the isolated retina. J Neurosci Methods 2000; 101:31–42. Humayun MS. Intraocular retinal prosthesis. Trans Am Ophthalmol Soc 2001;99:271–300. Humayun MS, de Juan E Jr., Weiland JD, et al. Pattern electrical stimulation of the human retina. Vision Res 1999;39: 2569–76. Chow AY, Chow VY. Subretinal electrical stimulation of the rabbit retina. Neurosci Lett 1997;225:13–6. Stett A, Barth W, Weiss S, et al. Electrical multisite stimulation of the isolated chicken retina. Vision Res 2000;40: 1785–95. Rizzo JF, Wyatt J, Humayun M, et al. Retinal prosthesis: an encouraging first decade with major challenges ahead. Ophthalmology 2001;108:13–4. Margalit E, Maia M, Weiland JD, et al. Retinal prosthesis for the blind. Surv Ophthalmol 2002;47:335–56. Chow AY, Packo KH, Pollack JS, Schuchard, RA. Subretinal artificial silicon retina microchip implantation in retinitis pigmentosa patients: long term follow-up [abstract]. Invest Ophthalmol Vis Sci 2003;44:4205. Humayun MS, Weiland JD, Fujii GY, et al. Visual perception in a blind subject with a chronic microelectronic retinal prosthesis. Vision Res 2003;43:2573–81. Dawson WW, Radtke ND. The electrical stimulation of the retina by indwelling electrodes. Invest Ophthalmol Vis Sci 1977;16:249–52. Yeomans JS. Principles of Brain Stimulation. New York, NY: Oxford University Press; 1990. Eckmiller R. Learning retina implants with epiretinal contacts. Ophthalmic Res 1997;29:281–9. Jensen RJ, Rizzo JF 3rd, Ziv OR, et al. Thresholds for activation of rabbit retinal ganglion cells with an ultrafine, extracellular microelectrode. Invest Ophthalmol Vis Sci 2003;44: 3533–43. Walter P, Szurman P, Vobig M, et al. Successful long-term implantation of electrically inactive epiretinal microelectrode arrays in rabbits. Retina 1999;19:546–52. Majji AB, Humayun MS, Weiland JD, et al. Long-term his-
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
571
Retinal prosthesis—Chowdhury et al
tological and electrophysiological results of an inactive epiretinal electrode array implantation in dogs. Invest Ophthalmol Vis Sci 1999;40:2073–81. 26. McCreery DB, Agnew WF, Yuen TG, et al. Charge density and charge per phase as cofactors in neural injury induced by electrical stimulation. IEEE Trans Biomed Eng 1990;37: 996–1001. 27. Agnew WF, McCreery DB. Considerations for safety with chronically implanted nerve electrodes. Epilepsia 1990;31 (suppl 2):S27–32. 28. Lamme VA, Super H, Landman R, et al. The role of primary visual cortex (V1) in visual awareness. Vision Res 2000;40: 1507–21.
572
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
29. Gordon B, Lesser RP, Rance NE, et al. Parameters for direct cortical electrical stimulation in the human: histopathologic confirmation. Electroencephalogr Clin Neurophysiol 1990;75: 371–7. 30. Jayakar P. Physiological principles of electrical stimulation. Adv Neurol 1993;63:17–27. 31. Tusa RJ, Palmer LA, Rosenquist AC. The retinotopic organization of area 17 (striate cortex) in the cat. J Comp Neurol 1978;177:213–35. Key words: electric stimulation, implanted electrodes, phosphenes, retina, cats
From *the Department of Ophthalmology, Prince of Wales Clinical School, and †the School of Medical Sciences, University of New South Wales, Sydney, Australia Originally received Jul. 8, 2004 Accepted for publication Jan. 10, 2005
Retinal prosthesis—Chowdhury et al
Correspondence to: Dr. Vivek Chowdhury, PO Box 140, Randwick, NSW, 2031, Australia; fax 61-2-93851059; [email protected] This paper has been peer-reviewed. Can J Ophthalmol 2005;40:563–72
563
Retinal prosthesis—Chowdhury et al
Résultats : Les électrodes placées à l’extérieur de l’œil pouvaient susciter dans le cortex visuel des réactions à diverses configurations. Le seuil des courants pour susciter une réaction était inférieur à 100 μA avec pulsations simples. Les courbes intensité–durée et les cartes d’activation ont été obtenues pour divers paradigmes de stimulation. Il a été possible d’exciter la rétine pour susciter une réaction corticale avec des électrodes et les capacités de stimulation d’un Nucleus 24 ABI standard. Interprétation : Il est possible de stimuler électriquement la rétine avec des électrodes placées à l’extérieur de l’œil. Le seuil minimum des courants pour susciter une réaction était bas, et comparable aux implants épirétiniens. Des prototypes d’électrode et un implant potentiel se sont avérés efficaces pour stimuler la rétine.
R
etinitis pigmentosa is the most common cause of blindness in people under 70 years of age in the developed world. One and a half million people are affected with this disease worldwide.1 There is currently no clinical means of restoring visual perception to patients who have developed blindness (no light perception) as a result of retinitis pigmentosa or other inherited or acquired retinal dystrophies. There is also no clinical means of preventing the deterioration towards blindness that occurs in the retinal dystrophies.1,2 Experimental treatments such as gene therapy, neural cell transplantation, growth factors, vitamin supplementation, antioxidants, and regulators of apoptosis have shown no success in restoring visual sensations to blind patients in clinical trials.1 It is unlikely that any of these methods will be able to restore visual perceptions to blind patients in the medium term (5–10 years), if at all.2 This is due to the poor capacity of neural tissue, such as that in the retina, for repair and regeneration.3 Therefore, a way to restore vision based on bypassing the damaged elements of the visual pathway is required.4,5 The only experimental method that has successfully restored visual perceptions to irreversibly blind patients is through the electrical stimulation of eye, optic nerve, or brain with implanted electrodes.2,6–9 Of these 3 sites, electrical stimulation of the visual cortex is the only approach, at this time, to developing a bionic eye that has restored visual perceptions sufficient to increase the patient’s mobility and independence.10 Despite an intensive research effort by a number of well-funded groups over the past 10 years aimed at developing a visual prosthesis with intraocular electrodes at epiretinal9,11–13 or subretinal14,15 locations, electrical stimulation of the retina has not met with the success achieved by stimulation of the visual cortex.10,16 This is because there are significant prob-
564
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
lems to be overcome at the electrode–tissue interface before an intraocular retina-based bionic implant can be a viable treatment for blindness.16,17 There are 2 chronic human trials of intraocular retinal implants occurring. In a study by Chow et al, the authors note improvements in the visual acuity of implanted patients following subretinal stimulation that cannot be explained by a neuroprosthetic effect of the implanted device, leading them to suggest that low-level electrical stimulation of the retina is having an as-yet-undefined “neurotrophic effect.”18 In the second study by Humayun and colleagues, a 4 × 4 array of 16 platinum disc electrodes, each 520 μm diameter, with an interelectrode centre-to-centre spacing of 720 μm, was implanted with a tack at the epiretinal surface of a blind human patient.19 A cable from the electrode array passed through the sclera and tracked subcutaneously to a stimulator outside the orbit. Electrical stimulation of this array elicited the subjective perception of small spots of light in the patient’s visual field. Interfacing the device with a camera allowed the patient to detect the presence of light and large objects.19 We are investigating the feasibility of extraocular stimulation of the retina with an electrode array placed on the posterior scleral surface of the eye. Benefits of such an approach include no intraocular surgery, no implantation of an intraocular foreign body, and no cable passing through the sclera, which would leave a permanent defect in the integrity of the globe. We plan to adapt the neuroprosthetic technology that is available in the Nucleus 24 auditory brainstem implant (ABI) (Cochlear Ltd., Lane Cove, Australia). This implant has an electrode array of 21 platinum disc electrodes, each 700 μm diameter, with a 950-μm centre-to-centre interelectrode spacing, arranged in 3 rows in a silicone carrier.
Retinal prosthesis—Chowdhury et al
We present the results of a preliminary study of an extraocular retinal prosthesis, an approach to visual prosthesis development for which a device has hitherto not yet been developed.7,17,19 Our study investigated a number of electrode types, configurations, and stimulus methods to optimize the development of an extraocular retinal implant for the restoration of visual perception to blind patients. METHODS
Acute experiments were carried out in anaesthetised cats in accordance with the guidelines of the animal ethics committee of the University of New South Wales. In adult cats (n = 6) weighing between 2.5 kg and 5.5 kg, anaesthesia was induced with an intramuscular injection of ketamine (20 mg/kg) and xylazine (1 mg/kg). The animals were given a preoperative dose of subcutaneous atropine (0.2 mg/kg) and dexamethasone (1.5 mg/kg). After intubation, the animals were ventilated, and anaesthesia was maintained with a 70:30 mixture of nitrous oxide and oxygen, with 0.5% to 1% halothane. The electrocardiogram, end-tidal carbon dioxide, and core body temperature were monitored, and animals were checked regularly for the absence of reflexes to ensure adequate anaesthesia. At the end of the experiment, the cats were euthanased with an intravenous injection of pentobarbital. The cats underwent a bilateral craniotomy of the parietal bones, and the dura was removed to expose the primary visual cortex (cytoarchitectonic area 17) of both hemispheres. The right orbit was approached by removing the bone of the lateral orbital wall and retracting or removing the right temporalis muscle. The cerebral cortex was regularly irrigated with warmed (35°C) paraffin oil. The cats were placed in a stereotaxic frame, which provided a system of 3-dimentional coordinates used during neurosurgery to give a precise atlas of the brain. Platinum or silver ball electrodes were placed on the pial surface of the primary visual cortex. For the stimulating electrodes on the posterior scleral surface of the eye, we investigated a variety of types and configurations. This included silver ball electrodes (1 mm diameter), flat platinum disc electrodes of 2 and 4 mm diameter (Cochlear Ltd.), and a multielectrode array consisting of 21 platinum disc electrodes, each 700 μm diameter (Cochlear Ltd.). In some experiments, a contact-lens electrode was used as the stimulus return path (ERG-jet, Universo, La
Chaux-de-Fonds, Switzerland). We also experimented with different surgical techniques for implanting the electrode array. Electrical stimulation of the right retina by the extraocular electrodes evoked cortical potentials that we recorded using Scope 3.6.11 software and a PowerLab/4SP data acquisition system connected to a ML135 biopotential amplifier (ADInstruments Ltd., Castle Hill, Australia). Evoked potentials were averaged over 100 to 200 trials, after being filtered to within a frequency range of 10 Hz to 5 kHz, using a 50-Hz notch filter. A differential recording system was used, with an indifferent scalp clip and a ground wire connected to a right thigh subcutaneous pin. The integrity of the neural visual pathway was confirmed with recordings of electroretinograms and visual evoked potentials. Electrical stimuli were rectangular, constantcurrent, symmetrical, biphasic pulses delivered from a 2100 Isolated Pulse Stimulator (A-M Systems, Inc., Carlsborg, Wash.) or monophasic pulses from an ML180 Stimulus Isolator (ADInstruments Ltd.). Biphasic stimuli were also delivered using a Nucleus 24 ABI. RESULTS Ball electrodes
We investigated monopolar and bipolar extraocular stimulation of the retina with silver ball electrodes. In cat A, the stimulation electrode was placed on the posterior scleral surface of the eye, 5 mm lateral to the optic nerve attachment, in the horizontal meridian. In this animal, a 4-mm platinum disc electrode was implanted on the inferonasal aspect of the globe to serve as the current return path. In cat B, 2 silver ball electrodes with a 5-mm interelectrode spacing were placed at the central posterior surface of the globe, superolateral to the optic nerve head. The electrical evoked response to cathodal-first biphasic stimulation in a monopolar configuration from cat A was compared with the response to monophasic stimulation of bipolar electrodes in cat B (Fig. 1). Bipolar stimulation in cat B showed a typical negative–positive electrically evoked response.20 However, because of a prolonged artifact in cat A, which is the case with monopolar stimulation,21 the initial negative component of the cortical response in this cat was obscured (a comparison with control waves was performed). The first visible true response
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
565
Retinal prosthesis—Chowdhury et al
By determining the threshold current necessary to elicit a response at the visual cortex, we obtained a strength–duration curve for biphasic monopolar extraocular electrical stimulation of the retina (Fig. 2A). For phase durations greater than 200 μs, threshold currents for eliciting a visual cortex response were below 1.5 mA. We also determined the peak amplitude of the positive component of the evoked response in cat A for a range of stimulus current levels at a pulse width of 250 μs (Fig. 2B). Higher response amplitudes were obtained with increasing current up to 5 mA, after which there was a response plateau. Multipulse studies
Fig. 1—Evoked response at the primary visual cortex to retinal stimulation with 1-mA stimuli of 400-μs phase duration using ball electrodes. Before 20 ms, a stimulus artifact is punctuated by sharp afferent volley spikes.After 20 ms, bipolar stimulation in cat B shows a typical negative–positive evoked response with a negative peak at an average of 27 ms and a positive peak averaging 40 ms. With monopolar stimulation in cat A, a prolonged stimulus artifact (common with monopolar stimuli) obscures the first negative component of this wave. The first true response is a positive peak at 38 ms followed by another positive peak at 54 ms.
Table 1—Latency of evoked response of visual cortex to extraocular stimulation of the retina in cats. Mean time of response, ms (SD) (0) N1
(+) P1
(+) P2
Cat A*
—
36.53 (1.83)
51.70 (2.37)
Cat B
27.32 (1.42)
40.85 (3.48)
—
Peak
Note: ms = millisecond; SD = standard deviation; N = negative; P = positive. *Initial negative peak in cat A is obscured because of prolonged stimulus artifact from monopolar stimulation. Cat A exhibited 2 defined positive peaks, whereas cat B (bipolar stimulation) exhibited a single early positive peak.
in cat A is a positive peak at 38 ms. A second positive peak follows this wave at 54 ms. The average latency of the negative and positive peaks for cat B, and the 2 positive peaks for cat A, are shown in Table 1.
566
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
We investigated the effect of pulse trains on cortical response amplitude. The amplitude of the positive wave of the electrical evoked response was recorded for 2-mA, 250-μs cathodal-first biphasic stimuli to single pulses and pulse trains of 2, 3, or 4 pulses presented at 200 Hz. We also compared these responses to the amplitude of the cortical potential evoked by a double pulse at 100 Hz (Fig. 3). The highest cortical responses were obtained with 3 pulse trains. For double pulses, a longer interval between pulses evoked a larger cortical response. Cortical activation map
Extraocular stimulation of a localised region of the sclera leads to excitation of a localised region of the retina, which evokes localised cortical responses. We undertook monopolar extraocular stimulation of the retina with a ball electrode placed in the horizontal meridian 5 mm lateral to the attachment of the optic nerve on the globe. We used biphasic pulses of 250 μs pulse width and 2 mA current and recorded evoked responses at localised points on the primary visual cortex over the posteromedial regions of the lateral gyri of both hemispheres. Using a 1-mm silver-ball recording electrode, we moved anteroposteriorly in 1-mm steps to record evoked cortical potentials from HorsleyClarke stereotaxic coordinates A1 to P8 (Fig. 4). We recorded from the most medial position of the superior surface of the lateral gyrus that was accessible after a craniotomy that had preserved a narrow strip of bone overlying the sagittal sinus. Cortical responses in cat A were localised in the anteroposterior plane to Horsley-Clarke coordinates P6 and P7 in the left and right hemispheres, respectively. There was a second smaller peak of cortical activity anteriorly in
Retinal prosthesis—Chowdhury et al
Fig. 3—The effect of stimulus pulse trains on the electrically evoked response for frequencies of 100 Hz and 200 Hz.
Fig. 2—A: Strength–duration curve for extraocular retinal stimulation with biphasic cathodal–anodal symmetrical pulses using ball electrodes in a monopolar configuration.The threshold for detecting a reliable cortical response over background noise was defined as a positive peak amplitude of greater than 3 μV. B: Amplitude of the positive peak of the cortical evoked response to extraocular retinal stimulation with 250-μs cathodalfirst biphasic pulses.
the left hemisphere at P2, which was not present in the right hemisphere. Disc electrodes
We investigated extraocular stimulation of the retina with flat platinum disc electrodes in 3 different
configurations (Fig. 5A). In cat D, a 2-mm active electrode was sutured under the superior rectus muscle and a 4-mm return electrode was placed anterior to the attachment of the lateral rectus muscle (Fig. 5B). In cat E, a 2-mm active electrode was placed in the inferolateral quadrant of the right eye and a 4-mm return electrode was placed anterior to the attachment of the medial rectus muscle. In cat F, a 2-mm platinum disc was applied to the sclera, 5 mm lateral to the optic nerve attachment in the horizontal meridian, and a contact-lens electrode applied to the cornea was used for the current return path. The strength–duration curves for cortical activation with the electrode configurations in cats D and F is shown in Fig. 6. Electrical stimulation of the eye in cat E produced a right-cortex cortical activation map, which is shown along with that of cat A in Fig. 4. Anodal stimulation was more effective than cathodal stimulation at exciting the retina (cat F), demonstrated by the upward shift of the strength–duration curve. The curves flattened towards rheobase (the minimal strength of the electrical stimulus able to cause excitation) at pulse widths greater than 500 μs. Increasing the pulse width of the stimulus up to 1 ms led only to small decreases in the threshold current required for retinal activation. Cathodal stimulation was more effective with the electrode configuration of cat F than with that of cat D. In cat F, it was possible to obtain cortical responses from retinal stimulation
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
567
Retinal prosthesis—Chowdhury et al
Fig. 4—Bilateral cortical activation map for monopolar stimulation of the retina with an extraocular electrode placed 5 mm lateral to the attachment of the optic nerve in the horizontal meridian (cat A).The x-axis denotes Horsley-Clarke coordinates from 5 (A5) to –8 (P8). Also shown is the right-cortex cortical activation map for electrical stimulation with disc electrodes in cat E (see Fig. 5).
with threshold currents of less than 100 μA at pulse widths greater than 400 μs. Multielectrode array
We tested the components of a prototype extraocular retinal prosthesis in 2 cats. In cat C, we implanted a 21-electrode array along the horizontal meridian of the posterior sclera of the right globe, and bipolar biphasic pulses were delivered with a bench stimulator. In cat F, extraocular stimulation of the retina was carried out with a Nucleus 24 ABI, using neural response telemetry software and a portable programming system (Cochlear Ltd.). Strength–duration curves for both bench and implant stimulation are shown in Fig. 7. Electrode impedance measurements obtained with the reverse telemetry feature of the implant (cat F) gave a reading of 15.6 kΩ in the monopolar configuration (single grid electrode active, contact-lens electrode as the corneal return) and 8 kΩ in the “common ground” configuration (single grid electrode active, all other 20 electrodes on the grid shorted to the return path). Using bipolar stimulation (bench) of the electrode array, threshold currents were lower than 1 mA for pulse durations greater than 300 μs. Stimulation
568
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
levels with the implant are programmed in “device units” up to 255, which is the strongest stimulation level using the unit, equal to approximately 1.75 mA. Pulse widths up to the device maximum of 400 μs were investigated. Monopolar stimulation was more effective than common-ground stimulation. No responses could be elicited with a pulse width below 400 μs in common-ground mode. In monopolar configuration, using pulse widths of 100 μs and 200 μs, stimulation intensities at the upper limits of the device were required. INTERPRETATION
Electrical stimulation of the retina can elicit visual perceptions in blind patients,9 commonly described as small spots of light termed “phosphenes.” Major efforts are underway to develop a retinal prosthesis for blind patients that will restore a level of visual perceptions that is functionally useful.16,17,19,22,23 A medical device for this purpose must be surgically feasible24 and physically biocompatible to the body,25 and it must utilize safe stimulation parameters26,27 that give the highest possible number of stable, reproducible, and resolvable phosphene sensations.13 This study assessed the feasibility of using an extraocular
Retinal prosthesis—Chowdhury et al
A
B
SR
LR
IO Fig. 5—A:Three configurations of electrode placement on the eye. In cat D (blue), a 2-mm disc was placed under the superior rectus (SR) and a 4-mm return electrode was placed anterior to the lateral rectus (LR). In cat E (red), a 2-mm disc was placed inferior to the LR and posterior to the inferior oblique (IO) in the inferolateral quadrant of the eye. A 4-mm return electrode was placed anterior to the medial rectus (dotted). In cat F (green), a 2-mm disc was placed in the horizontal plane, 5 mm lateral to the attachment of the optic nerve, and stimulated with respect to a contact-lens electrode placed on the cornea. B: Picture of the right eye in cat D after implantating the electrodes and suturing the electrode tails to the periorbita.
approach to electrical stimulation of the retina, and we have proposed a device for this purpose. In 6 cats, it was possible to record evoked responses at the visual cortex to electrical stimuli applied to the retina with electrodes placed on the posterior surface of the globe. We investigated electrical stimulation with different electrode types in a variety of configurations (Fig. 5A), all of which were able to evoke visual cortex responses. It has been postulated that activity at the primary visual cortex is correlated to conscious perception,28 and we expect that the activity evoked in the visual pathway through this method of stimulation will likely evoke phosphene perception in human patients. We investigated the feasibility of implanting both single electrodes and electrode arrays on the surface of the sclera. In the cat shown in Fig. 5B, the disc electrode was slipped under the superior rectus muscle, and the electrode tail was then sutured to this muscle and then again to the periorbita with a purse-string suture. Local paralysis of the extraocular muscles to prevent movement will be necessary in future chronic implantations. The placement of extraocular electrodes requires neither intraocular surgery nor intraocular placement of foreign bodies against the delicate retinal tissues. It also does not require complicated methods, such as tacks, to attach the electrode arrays to the retinal surface.19 It would be possible in the future to use biological adhesives or
sutures to fix the silicone base, which carries the electrode, directly to the sclera. The firm scleral base for extraocular electrodes avoids mechanical trauma to the retina and decreases the likelihood of pathological effects due to charge transfer and heat generation.30 We examined the thresholds needed to elicit cortical evoked responses with extraocular electrodes. Current thresholds were variable amongst cats with different electrode configurations. The lowest thresholds were found with a 2-mm electrode placed on the posterior of the globe and stimulated with respect to a contact-lens electrode placed on the cornea. The threshold for eliciting an evoked response was as low as 100 μA (Fig. 6) at a pulse duration of 400 μs (0.04 μC per phase). This translates to a charge density at the electrode surface of 1.27 μC/cm2, which is well within safe charge injection levels in human studies for chronic neural stimulation.29 Higher thresholds occurred with monopolar stimulation using the Nucleus 24 ABI (Fig. 7). A threshold of 225 device units (255 units = 1.75 mA) was obtained when using 400-μs cathodal-first biphasic pulses. We assessed the ability of extraocular stimulation to produce localised phosphenes by investigating if extraocular stimulation produced localised activation at the primary visual cortex. The medial edge of the lateral gyrus (cytoarchitectonic area 17) contains a retinotopically mapped representation of the visual field.31 In cat A, the cortical response was localised
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
569
Retinal prosthesis—Chowdhury et al
Fig. 6—Strength–duration curves for current (mA) and pulse width (ms) for cat D and cat F using retinal stimulation with disc electrodes. In cat D, monopolar stimulation was performed with cathodal-first biphasic pulses. In cat F, monopolar stimulation was performed with cathodal or anodal monophasic pulses.
Fig. 7—Strength–duration curves for scleral stimulation with the 21-electrode array. Stimuli from implanted monopolar and common-ground electrode configurations (cat F) are measured in mA (right y-axis), and bipolar stimulation with a bench stimulator (cat C) is measured in device units (left y-axis, where 255 device units are ~1.75 mA). Monopolar stimulation involved activating one 700-μm platinum disc on the array and using a contact-lens electrode as the return path. Common-ground stimulation involved stimulating one active electrode from the array with respect to the remaining 20 electrodes, which were shorted to the return path. Bipolar stimulation with the bench stimulator utilized electrical stimulation between 2 electrodes on the array with a 2.85-mm centre-to-centre spacing. For the common-ground stimulus configuration, responses could only be obtained with a pulse width of 400 μs, which is the maximal pulse width programmable with the implant.
570
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
Retinal prosthesis—Chowdhury et al
posteriorly at a similar anteroposterior coordinate in both hemispheres (Fig. 4). Studies of pulse summation will aid in the identification of optimal stimulus parameters for an extraocular retinal prosthesis. Ideal frequencies for pulse trains were investigated with multipulse studies. Higher amplitudes of the evoked cortical responses were obtained with paired pulses at 100 Hz compared with those at 200 Hz. For 200-Hz stimuli, peak responses were obtained with 3 pulse trains (Fig. 3). As expected, anodal stimuli were less effective than cathodal stimuli in eliciting extracellular stimulation of retinal tissue (Fig. 6, cat F).21,30 We undertook a basic feasibility study of adapting technology in the Nucleus 24 ABI for extraocular retinal stimulation. This device is composed of an array of 21 platinum disc electrodes in a silicone carrier with a stimulator implant that can be powered and controlled transcutaneously. This array could elicit stimulation of the sclera under a variety of electrode configurations. The stimulator implant itself was able to evoke visual cortex responses with single biphasic pulses (Fig. 7). An extraocular retinal implant may possibly overcome some of the difficulties of using epiretinal stimulation for a visual prosthesis.16 We have shown the feasibility of such an approach by demonstrating an ease of surgical implantation, safe current thresholds for evoking cortical responses, and localised activation of the primary visual regions in the brain. Adaptation of the technology in the Nucleus 24 ABI will provide a basis for developing an extraocular prosthesis for human trial. Financial support for this work was received from the Ophthalmic Research Institute of Australia, the University of New South Wales, Retina Australia, the Brain Foundation, and the Australian National Health and Medical Research Council. Some equipment that was used in this project was provided at no expense by Cochlear Ltd. The 3 authors are directors and shareholders in Sydney Biotech Pty Ltd., which is collaborating with Cochlear Ltd. and the University of New South Wales to develop a visual prosthesis.
4. 5. 6.
7. 8. 9.
10. 11.
12. 13.
14. 15. 16.
17. 18.
19.
20.
21. 22.
REFERENCES 23. 1.
2. 3.
Sharma RK, Ehinger B. Management of hereditary retinal degenerations: present status and future directions. Surv Ophthalmol 1999;43:427–44. Chong NH, Bird AC. Management of inherited outer retinal dystrophies: present and future. Br J Ophthalmol 1999;83: 120–2. Mitchell RN, Cotran RS. Repair: Cell Regeneration, Fibrosis,
24.
25.
and Wound Healing. In: Kumar V, Cotran RS, Robbins SL, eds. Basic Pathology. 6th ed. Philadelphia, Pa: WB Saunders, 1997:chap 3. Scarlatis G. Optical prosthesis: visions of the future. JAMA 2000;283:2297. Larkin M. Artificial-vision research comes into focus. Lancet 2000;355:1080. Veraart C, Raftopoulos C, Mortimer JT, et al. Visual sensations produced by optic nerve stimulation using an implanted self-sizing spiral cuff electrode. Brain Res 1998;813:181–6. Maynard EM. Visual prostheses. Annu Rev of Biomed Eng 2001;3:145–68. Brindley GS, Lewin WS. The sensations produced by electrical stimulation of the visual cortex. J Physiol 1968;196: 479–93. Humayun MS, de Juan E Jr., Dagnelie G, et al. Visual perception elicited by electrical stimulation of retina in blind humans. Arch Ophthalmol 1996;114:40–6. Dobelle WH. Artificial vision for the blind by connecting a television camera to the visual cortex. ASAIO 2000;46:3–9. Grumet AE, Wyatt JL, Rizzo JF. Multi-electrode stimulation and recording in the isolated retina. J Neurosci Methods 2000; 101:31–42. Humayun MS. Intraocular retinal prosthesis. Trans Am Ophthalmol Soc 2001;99:271–300. Humayun MS, de Juan E Jr., Weiland JD, et al. Pattern electrical stimulation of the human retina. Vision Res 1999;39: 2569–76. Chow AY, Chow VY. Subretinal electrical stimulation of the rabbit retina. Neurosci Lett 1997;225:13–6. Stett A, Barth W, Weiss S, et al. Electrical multisite stimulation of the isolated chicken retina. Vision Res 2000;40: 1785–95. Rizzo JF, Wyatt J, Humayun M, et al. Retinal prosthesis: an encouraging first decade with major challenges ahead. Ophthalmology 2001;108:13–4. Margalit E, Maia M, Weiland JD, et al. Retinal prosthesis for the blind. Surv Ophthalmol 2002;47:335–56. Chow AY, Packo KH, Pollack JS, Schuchard, RA. Subretinal artificial silicon retina microchip implantation in retinitis pigmentosa patients: long term follow-up [abstract]. Invest Ophthalmol Vis Sci 2003;44:4205. Humayun MS, Weiland JD, Fujii GY, et al. Visual perception in a blind subject with a chronic microelectronic retinal prosthesis. Vision Res 2003;43:2573–81. Dawson WW, Radtke ND. The electrical stimulation of the retina by indwelling electrodes. Invest Ophthalmol Vis Sci 1977;16:249–52. Yeomans JS. Principles of Brain Stimulation. New York, NY: Oxford University Press; 1990. Eckmiller R. Learning retina implants with epiretinal contacts. Ophthalmic Res 1997;29:281–9. Jensen RJ, Rizzo JF 3rd, Ziv OR, et al. Thresholds for activation of rabbit retinal ganglion cells with an ultrafine, extracellular microelectrode. Invest Ophthalmol Vis Sci 2003;44: 3533–43. Walter P, Szurman P, Vobig M, et al. Successful long-term implantation of electrically inactive epiretinal microelectrode arrays in rabbits. Retina 1999;19:546–52. Majji AB, Humayun MS, Weiland JD, et al. Long-term his-
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
571
Retinal prosthesis—Chowdhury et al
tological and electrophysiological results of an inactive epiretinal electrode array implantation in dogs. Invest Ophthalmol Vis Sci 1999;40:2073–81. 26. McCreery DB, Agnew WF, Yuen TG, et al. Charge density and charge per phase as cofactors in neural injury induced by electrical stimulation. IEEE Trans Biomed Eng 1990;37: 996–1001. 27. Agnew WF, McCreery DB. Considerations for safety with chronically implanted nerve electrodes. Epilepsia 1990;31 (suppl 2):S27–32. 28. Lamme VA, Super H, Landman R, et al. The role of primary visual cortex (V1) in visual awareness. Vision Res 2000;40: 1507–21.
572
CAN J OPHTHALMOL—VOL. 40, NO. 5, 2005
29. Gordon B, Lesser RP, Rance NE, et al. Parameters for direct cortical electrical stimulation in the human: histopathologic confirmation. Electroencephalogr Clin Neurophysiol 1990;75: 371–7. 30. Jayakar P. Physiological principles of electrical stimulation. Adv Neurol 1993;63:17–27. 31. Tusa RJ, Palmer LA, Rosenquist AC. The retinotopic organization of area 17 (striate cortex) in the cat. J Comp Neurol 1978;177:213–35. Key words: electric stimulation, implanted electrodes, phosphenes, retina, cats