|Home | About | Journals | Submit | Contact Us | Français|
Retinal prostheses provide the capability to blind patients to detect motion and locate large objects. To avoid activating axons of passage, which can create streak-like perceptions, long pulse stimulation can be used to bypass axons and achieve focal retinal activation. Safety is a concern because long pulses require more charge than short pulses to elicit a response from neural tissue. Future implants will require smaller electrodes to improve resolution, but increased charge density may result, which is another safety concern. We developed a method to study the effects of electrical stimulation in the retina in real time using OCT (Optical Coherence Tomography) imaging combined with micropositioning of a stimulating electrode over the retina in an animal model. When using a 250-micron diameter electrode and stimulating for 30 minutes (frequency: 333 Hz), charge density: 1.22 mC/cm2, we observed an increase in retinal thickness from 154.3 μm ± 7.04 μm to 179.67 μm ± 0.47μm, a 16.66 % ± 5.49% increase compared to baseline. The region of increased thickness extended laterally for 0.56 mm ± 0.009 mm. When stimulating with a charge density of 1.63 mC/cm2, we observed an increase in retinal thickness from 160.3 μm ± 2.05 μm to 190 μm ± 0.81μm, a 19.52 % ± 1.86% increase compared to baseline. The region of increased thickness expanded laterally for 1.27 mm ± 0.19 mm.
Retinitis pigmentosa  and aged related macular degeneration  are two diseases that affect primarily the photoreceptor layer, but leaves the inner retinal cells (bipolar cells and ganglion cells) relatively intact , allowing the retina to be electrically stimulated to restore a sense of vision. Retinal prostheses have demonstrated the capability to elicit the sensation of light and to give subjects the ability to detect motion and the location of large objects .
Studies of retinal stimulation safety are needed for higher resolution arrays and new stimulation paradigms. The resolution of a retinal prosthesis is correlated to the number and size of electrodes in its array . To be able to achieve a high-resolution retinal prosthesis, simulations of artificial vision have suggested that 600 to 1000 electrodes, between 50–100 microns diameter will be required in a 5mm diameter on the retina  . Small electrodes will focus the stimulus but also require high charge density. Retinal cells (ganglion cells and bipolar cells) can be activated selectively depending on the pulse parameters used for stimulation. Short pulses (1ms) produce elongated phosphenes by activating both ganglion cell axons and bipolar cells , while long pulses (25 ms) produce focal phosphenes by activating bipolar cells and avoiding ganglion cell axons . Safety is a concern when using long pulses to stimulate, because long pulses require more charge than short pulses to have the same effect on neural tissue.
Several studies have been reported about the safety of chronic and acute electrical stimulation in the retina. Most of these studies have relied on post-mortem analysis like hematoxylin staining , or propidium iodide (PI) fluorescent dye . The effects of electrical stimulation in the retina were evaluated in real time using OCT (Optical Coherence Tomography) imaging in an in-vitro model . These studies have helped to understand the anatomy of the retina at one time point after electrical stimulation , , , and showed the value of real-time imaging to analyze the retina during the stimulation period. However, since this method was in vitro, long-term effects of stimulation could not be investigated. We propose an in vivo experimental model more similar to current and future epiretinal prostheses.
We studied the effects of electrical stimulation in the retina in real time using OCT (Optical Coherence Tomography) images in a rabbit eye, using an in-vivo model and a platinum/Iridium disk electrode (75 and 250 um diameter). 75 um corresponds to the electrode size proposed for a high resolution epiretinal prosthesis and 250 um is similar to electrodes in a commercially available epiretinal prosthesis. Using an in-vivo model allows us to study the retina in its natural condition.
Adult pigmented rabbits (Irish Farms, Norco, CA); approximately 2 months old were used for all experiments. Experiments were performed in the left eye of each animal (n=13). Rabbits were anesthetized with ketamine-xylazine (100 mg/kg and 20 mg/kg) and euthanized at the end of the procedure.
All animals were maintained on a daily 12 h light/day cycle prior experiment. All procedures conformed to the Guide for Care and Use of Laboratory Animals (National Institute of Health). The University of Southern California Institutional Animal Care and Use committee reviewed and approved all procedures.
OCT images were taken as control before stimulation (HRA+OCT Spectralis, HEIDELBERG). After the eye was dilated, an incision was made in the sclera, 3 mm from the limbus and a 25 gauge valved trocar (Alcon) was used to help keep the stimulating electrode in position. The electrode was 90% Pt/10% Ir concentric bipolar (FHC, Bowdain, ME), with a diameter of 75-micron (Model CBDFG75) or 250-micron (Custom). The stimulating electrode was held by a three axis translational stage (Model 4044 M Parker DAEDAL) mounted on a magnetic based articulating arm. For the 75-micron electrode, the center pole was used for stimulation. A needle electrode was placed on the head for current return. (Figure 1)
The stimulating electrode was placed through the valved trocar and advanced inside the eye until it became visible in the fundus view of the OCT system. The scan line was adjusted to be along the length of the electrode tip to allow simultaneous scanning of the electrode tip and the retina. The micromanipulator was used to advance the electrode towards the retina for better positioning. (Figure 2) Charged-balanced, cathodic first, biphasic current pulses were applied to the retina. Different pulse widths, pulse amplitudes, frequencies, stimulation times, and electrode sizes were used during these experiments (Table 1). OCT images were acquired every 2 minutes during and after stimulation. Retinal thickness was measured to assess retinal damage.
Eighteen retinal regions were imaged in thirteen different animals. The electrode was placed on the inferior temporal region, close to the visual streak. (Figure 2).
OCT images were taken as the electrode was advanced towards the retina. After insertion into the eye, the electrode was advanced from the periphery to a point where the electrode tip was visible in the fundus view of the OCT system. The electrode was placed closed to the epiretinal surface but not touching. The micromanipulator was used to advance the electrode until it was near the retina.
The stimulating electrode was placed inside the left eye 59.25 μm ± 27.26 μm (mean, SD) away from the retina. The distance between the electrode and the retina was measured using the calipers function of the OCT analysis software as shown in figure 2.
A fundus image and OCT images were taken before stimulation as control. During the stimulation period, images were acquired every two minutes for a period of 30 minutes. After stimulation, the stimulating electrode was moved away from the retina and images were acquired every 5 minutes for a period of 15 minutes. The effects of the different charge densities applied to the retina are shown on Figure 4. OCT images are shown right before starting stimulating, 20 minutes into the stimulation and 30 minutes after stimulation. Images shown are for three different charge densities: 0.92 mC/cm2, 1.22 mC/cm2 and 1.63 mC/cm2 applied with a 250 μm diameter platinum/iridium electrode using 1 ms pulses.
When using the 250 μm diameter electrode and stimulating for 30 min (frequency: 333 Hz, charge density: 1.63 mC/cm2) we observed an increase in retinal thickness from 160.3 μm ± 2.05 μm (mean, SD) to 190 μm ± 0.81μm (mean, SD), a 19.52% ± 1.86% (mean, SD) increase compared to baseline. The region of increased thickness extended laterally for 1.27 mm ± 0.19 mm (mean, SD), roughly centered about the position of the stimulating electrode. When using a 250 μm diameter electrode and stimulating for 30 min (frequency: 333 Hz, charge density: 1.22 mC/cm2) we observed an increase in retinal thickness from 154.3 μm ± 7.04 μm (mean, SD) to 179.67 μm ± 0.47 μm (mean, SD), a 16.66 % ± 5.49% (mean, SD) increase compared to baseline. The region of increased thickness extended laterally for 0.56 mm ± 0.009 mm (mean, SD).
Experiments using the 75 μm diameter electrode with any settings listed in Table 2 and using the 250 μm diameter electrode and charge densities of 0.92 and 1.02 mC/cm2 showed no change in retinal thickness.
Figure 3 shows a summary of the OCT measurements of retinal thickness versus current density. Values shown are the averages of all the animals studied per group.
A two-way analysis of variance with Holm-Sidak post-hoc testing for multiple pairwise comparisons showed no significant difference between retinal thicknesses from different groups (1.63,1.22 and 0.92 mC/cm2) before stimulation, and a significant difference between retinal thicknesses from the same groups after stimulation (Table 2). Similarly, the difference between retinal thicknesses for the 0.92 mC/cm2 group pre and post stimulation was not statistically significant (P=1); but there was a significant difference between retinal thicknesses pre and post stimulation for groups 1.22 mC/cm2 (P<0.001) and 1.63 mC/cm2 (P<0.001).
Electrical stimulation at very high charge densities causes swelling of the retina within minutes of stimulation. The stimulus levels at which swelling was noted is significantly higher than parameters used generally for stimulation in humans in terms of rate and charge density, but in limited cases high currents  or high rates  have been used in humans, on a short term basis in the context of controlled psychophysical testing. Other studies have linked stimulation rate and/or duty cycle to neural damage . Rate may be an important factor for retinal stimulation safety as well.
We developed a new method that allows us to study the effects of electrical stimulation on the retina in an in vivo model during the stimulation period. We used an in vivo model because it allows us to study the retina response in its natural conditions. This technique will also allow study of long-term effects of retinal stimulation, since it is potentially a survival surgery.
We will like to thank Lina Flores, Fernando Gallardo and Mort Arditti for technical support.
*Research supported by NEI EY022931 and Research to Prevent Blindness.
A. Gonzalez-Calle, Department of Biomedical Engineering, Viterbi School of Engineering, University of Southern California, Los Angeles, CA 90089, USA.
J.D. Weiland, Department of Biomedical Engineering, Viterbi School of Engineering, University of Southern California, Los Angeles, CA 90089, USA and Department of Ophthalmology, Keck School of Medicine, University of Southern California, Los Angeles, CA 90033, USA.