|
|
||||||||
1From the Department of Ophthalmology, Harvard Medical School and the Massachusetts Eye and Ear Infirmary, Boston, Massachusetts; the 2Center for Innovative Visual Rehabilitation, Veterans Administration Hospital, Boston, Massachusetts; the 3Department of Electrical Engineering and Computer Science and the 4Research Laboratory of Electronics, Massachusetts Institute of Technology, Cambridge, Massachusetts; and the 5Nanofabrication Laboratory, Cornell University, Ithaca, New York.
| Abstract |
|---|
|
|
|---|
METHODS. Five volunteers with severe retinitis pigmentosa and one with normal vision were studied. A companion paper in this issue provides details on demographics, visual function, surgical methods, general stimulation strategy, and data analysis. Volunteers were awake during surgery while a 10-µm-thick, microfabricated electrode array was placed on the retina. The array was connected to extraocular current sources that delivered charges to 50-, 100-, and 400-µm-diameter electrodes. Negative control trials were randomly included. Perceptual quality was judged by the similarity between the form of stimulation and perception (i.e., accuracy) and the reproducibility of responses.
RESULTS. Only 1 of 40 control tests yielded a false-positive result. On average, volunteers 3, 5, and 6 reported percepts that matched the stimulation pattern 48% and 32% of the time for single- and multiple-electrode trials, respectively. Two-point discrimination in the best cases may have been achieved in two blind subjects using (center-to-center) electrode separation of 600 and 1960 µm. Reproducibility was achieved 66% of the time in the blind subjects. By comparison, in the normal-sighted subject, perceptual form was reported accurately 57% of the time, with 82% reproducibility, and two-point discrimination may have been achieved in one trial with 620-µm electrode spacing and in two trials each with 1860- and 2480-µm electrode spacing. In subjects 5 and 6, perceptual size was inconsistently related to the charge, although relatively large differences in charge (median: 0.55 microcoulombs [µC]) between two trials produced differently sized percepts. Longer stimuli did not produce rounder percepts.
CONCLUSIONS. Single percepts induced by single-electrode stimulation were relatively small, but the form of percepts, especially after multielectrode stimulation, often did not match the stimulation pattern, even in a normal-sighted volunteer. Reproducible percepts were more easily generated than those that matched the stimulation pattern.
| Methods |
|---|
|
|
|---|
|
|
|
| Results |
|---|
|
|
|---|
Hypotheses
Hypothesis 1.
Blind subjects will report a single, small percept after stimulation through one electrode at or slightly above threshold.
With a 100- or 400-µm electrode, 185 percepts were elicited from volunteers 3, 5, and 6. A small percept was reported in 1 (6%) of 17, 38 (35%) of 109, and 50 (85%) of 59 trials, respectively (Table 2) . On average, the hypothesis was satisfied 48% of the time. Of responses not consistent with the hypothesis: (1) volunteer 3 mostly reported faint, flashing images and on four occasions a long line; (2) volunteer 5 reported a line 5% of the time and clusters of two or three small images 60% of the time; (3) volunteer 6 reported a line 5% of the time and clusters of two or three small images 10% of the time.
|
Eighty-four trials tested this hypothesis, which for subjects 3, 5, and 6 was satisfied 55%, 21%, and 29% (average 32%) of the time, respectively (Table 2) . Given these limited results, only three letter-recognition (T or L) trials were attempted, none of which yielded the anticipated responses.
Hypothesis 3.
Two-point discrimination can be achieved in a blind subject by delivering electrical stimulation through two electrodes.
In experiment 3, driving eight large electrodes (600-µm spacing) induced perception of a line of four distinct images or similar percept three of three times. Stimulation of one column of four large electrodes yielded no response, but driving the other four electrodes again induced perception of a line of four images. We illuminated the eye and saw a tilted array, with only one edge contacting the retina, which presumably accounted for the perception of four images, rather than eight (Fig. 1) .
|
|
|
The last three blind subjects reported similar images 3 (100%) of 3, 35 (57%) of 62 and 27 (79%) of 34 times, respectively, when stimulation was unchanged between two trials (average 66%; Table 2 ).
Hypothesis 5.
Accurate percepts can be induced more frequently in normal-sighted than in blind volunteers.
In the single-electrode (100 or 400 µm) trials, the normal subject met our expectation 8 (57%) of 14 times. By this measure, subject 6, legally blind for 15 years, performed better (Table 2) . The normal subject reported reproducible percepts 82% (9/11) of the time versus the blind subjects 66% (hypothesis 4).
In multiple-electrode trials, the normal subject reported "accurate" percepts 43% of the time (versus 32% for blind volunteers: hypothesis 2). A subset of these trials tested two-point discrimination (for one example, see Fig. 5 ). Across these and other similar trials (n = 14), at best, she distinguished stimulation separated by 620 µm one time, and in two of two cases each she distinguished electrode separation of 1860 and 2480 µm. See hypothesis 3 for comparison to blind patients.
|
In one analysis, perceptual size was recorded per stimulus charge. There were 52 and 33 trials with single electrodes for the last two subjects, respectively. To permit uniform comparison of charge density, the smaller number of trials with the 100-µm electrode was excluded, leaving 40 and 30 trials, respectively, with the 400-µm electrode. In subject 5, the median charge that yielded a pea- or dime-sized percept was identical (1.4 microcoulombs [µC]). In subject 6, the median charges that produced a pea-, dime-, or quarter-sized percept were 0.4, 0.8, and 1.1 µC, respectively, which is consistent with the hypothesis.
In another analysis, we assessed whether percepts enlarged or shrank in pairs of trials (not necessarily sequential) in which charge was the only variable. Subjects 5 and 6 satisfied the hypothesis only 29% of the time (Table 4) . However, trials that satisfied the hypothesis had a median difference in charge of 0.55 µC versus 0.24 µC for those that did not.
|
Only single, round, or elongated percepts induced by stimulation through one electrode were considered; 108 responses (from subjects 4, 5, and 6) met these criteria. Five (100%) of 5 percepts by subject 4, 47 (63%) of 74 by subject 5, and 25 (86%) of 29 by subject 6 were round. Across all durations, a round percept was reported 2.5 times more than an elongated percept (Table 5) . This hypothesis was rejected.
|
In experiment 5, a "circle" was reported four of four times when columnar electrodes (oriented along axons) were driven, whereas orthogonal stimulation produced multiple percepts (circle and lines) five of five times. In experiment 6, four trials using columnar stimulation yielded an elongated, curved ("banana"-shaped) percept three times and a spot "larger than a quarter" once (Fig. 6) , whereas four trials across axons yielded a "round" percept three times and a line of percepts once (all with 250 µA).
|
| Discussion |
|---|
|
|
|---|
Our hypotheses were designed to address the ability of blind subjects with retinitis pigmentosa to report basic form perception (hypotheses 14), perceptual differences between the normal-sighted volunteer and blind subjects (hypothesis 5), and perceptual effects of various stimulus parameters (hypotheses 68). Our results are both encouraging and sobering.
Hypothesis 1 tested whether stimulation through one electrode would yield single, small percepts. The hypothesis was satisfied 48% of the time over 185 trials. Percepts that were too large were uncommon errors. Much more commonly (i.e., 60% of the time in volunteer 5), multiple percepts were reported, the explanation for which is unknown, although a similar phenomenon occurs with visual cortical stimulation.15 16
For hypothesis 2, we studied percepts generated by multiple-electrode stimulation. Here, less success (32% vs. 48% for single electrode trials) was achieved in producing percepts that matched the stimulation pattern. Candidate explanations include anatomic and physiological disease of the retina and visual cortex secondary to chronic blindness17 18 19 20 21 ; our ignorance of effective stimulation strategies; interaction of electrical fields from adjacent electrodes; and insufficient learning opportunity for the subjects. Hypothesis 5 eliminates the first consideration because our normal subject performed less well than blind subject 6, which indicates that factors other than blindness hindered our outcomes.
Hypothesis 3 produced the least optimistic results. At best, two-point discrimination may have been achieved by subject 3 with electrode spacing of 600 µm and by subject 5 with electrode spacing of 1960 (but not 2480) µm. Yet, hypothesis 4 revealed relatively good reproducibility. This suggests that seemingly aberrant responses, especially seeing multiple images when one electrode is driven, are not random. Unchanging factors, such as our methods of stimulation or retinal or cortical disease, rather than subjective factors, probably accounted for a substantial fraction of responses that did not match the stimulation pattern.
Hypotheses 6, 7, and 8 explored perceptual effects of various stimulus paradigms. In hypothesis 6, we presumed that higher charges would enlarge the electrical field and hence the percept. Mixed results were obtained. With one analysis, volunteer 6 but not volunteer 5 satisfied the hypothesis. In a second analysis, relatively large differences in charge (median:
0.55 µC) between two trials yielded larger percepts.
The motivation to test hypothesis 7 derived from Greenberg22 who reported that longer duration stimuli (
0.5 ms) preferentially activate bipolar neurons and from Weiland et al.,9 who suggested that activation of the middle retina produces round percepts. We tested this hypothesis with nearly 100 trials and discovered that round percepts were equally or more frequently reported at durations that were considerably shorter and longer than Greenbergs benchmark (Table 5) . Our finding does not discount Greenbergs in vitro observations, for which we have some supportive evidence.23
Hypothesis 8 was tested because we assumed that activation of multiple electrodes along axons would be more likely to activate those axons. In experiment 6, this orientation generated elongated percepts, which is consistent with the hypothesis. In experiment 5, the orthogonal orientation generated multiple percepts, which suggests that in this configuration each electrode had a higher probability of producing an individual percept. The differences in outcome between these two patients suggest that stimulation strategies of a prosthesis may have to be customized to achieve desired percepts in individual patients.
By comparison to our results, Humayun et al.24 reported resolution of 1.5° of visual angle in a patient with light perception vision, despite the variable positioning of electrode(s) that must have occurred with their handheld technique. Moving a needle electrode by hand through the vitreous cavity provides the advantage of being able to survey a wide area of retina for points of low threshold. In five of six experiments, we also used a handheld approach as a screening technique to be certain that volunteers would see percepts in response to electrical stimulation near the retina before the introduction of an electrode array, which required additional surgical steps (see companion paper14 for more information). At best, we may have achieved resolution of 2.25° to 4.50° with our electrode array in contact with the retina. Further, our patients often did not report percepts that matched the stimulation pattern and frequently described multiple percepts when one electrode was driven, neither of which was reported by Humayun et al.24 The results from normal volunteers are equally disparate. The two subjects in Weiland et al.9 reported football-sized, dark percepts every time the normal retina was stimulated near threshold.9 Over 43 trials, our normal-sighted patient never reported darkness, and all percepts were considerably smaller than a football. Use of different stimulation frequencies and other methodological differences, insofar as they can be gleaned, may account for some differences in outcomes.
In summary, volunteers who have been legally blind for many years can see percepts induced by electrical stimulation of the retina. The single percepts were relatively small, which offers hope of generating a montage of such percepts to create useful images. However, the form of percepts, especially with multielectrode stimulation, often did not match the stimulation pattern. The lack of a better outcome in our normal-sighted patient suggests that retinal degeneration alone does not explain the limited results in our blind patients and emphasizes the need to learn effective stimulation methods. Nonetheless, even simple images, if reproducible, could help severely blind patients.
Acute testing provides useful insights into strategies for creating vision, but probably underestimates what could be achieved with permanently implanted devices, which offer opportunity for learning (by patients and researchers) and neural plasticity. Indeed, Humayun et al.25 have reported a learning effect for a patient who had received a chronic implant.
| Acknowledgements |
|---|
| Footnotes |
|---|
Submitted for publication August 13, 2002; revised November 18, 2002, and May 21 and June 26, 2003; accepted June 30, 2003.
Disclosure: J.F. Rizzo III (P); J. Wyatt (P); J. Loewenstein, None; S. Kelly, None; D. Shire (P)
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked "advertisement" in accordance with 18 U.S.C.
1734 solely to indicate this fact.
Corresponding author: Joseph F. Rizzo, III, Massachusetts Eye and Ear Infirmary, 243 Charles Street, Boston, MA 02114; joseph_rizzo@meei.harvard.edu.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
C. de Balthasar, S. Patel, A. Roy, R. Freda, S. Greenwald, A. Horsager, M. Mahadevappa, D. Yanai, M. J. McMahon, M. S. Humayun, et al. Factors Affecting Perceptual Thresholds in Epiretinal Prostheses Invest. Ophthalmol. Vis. Sci., June 1, 2008; 49(6): 2303 - 2314. [Abstract] [Full Text] [PDF] |
||||
![]() |
Y. Okawa, T. Fujikado, T. Miyoshi, H. Sawai, S. Kusaka, T. Mihashi, Y. Hirohara, and Y. Tano Optical Imaging to Evaluate Retinal Activation by Electrical Currents Using Suprachoroidal-Transretinal Stimulation Invest. Ophthalmol. Vis. Sci., October 1, 2007; 48(10): 4777 - 4784. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. Scholz Perspectives on: Materials Aspects for Retinal Prostheses Journal of Bioactive and Compatible Polymers, September 1, 2007; 22(5): 539 - 568. [Abstract] [PDF] |
||||
![]() |
T. Matsuo and N. Morimoto Visual acuity and perimacular retinal layers detected by optical coherence tomography in patients with retinitis pigmentosa Br. J. Ophthalmol., July 1, 2007; 91(7): 888 - 890. [Abstract] [Full Text] [PDF] |
||||
![]() |
J Salzmann, O P Linderholm, J-L Guyomard, M Paques, M Simonutti, M Lecchi, J Sommerhalder, E Dubus, M Pelizzone, D Bertrand, et al. Subretinal electrode implantation in the P23H rat for chronic stimulations Br. J. Ophthalmol., September 1, 2006; 90(9): 1183 - 1187. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. R. Montezuma, J. Loewenstein, C. Scholz, and J. F. Rizzo III Biocompatibility of materials implanted into the subretinal space of yucatan pigs. Invest. Ophthalmol. Vis. Sci., August 1, 2006; 47(8): 3514 - 3522. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. Sekirnjak, P. Hottowy, A. Sher, W. Dabrowski, A. M. Litke, and E. J. Chichilnisky Electrical Stimulation of Mammalian Retinal Ganglion Cells With Multielectrode Arrays J Neurophysiol, June 1, 2006; 95(6): 3311 - 3327. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. Dagnelie, D. Barnett, M. S. Humayun, and R. W. Thompson Jr Paragraph text reading using a pixelized prosthetic vision simulator: parameter dependence and task learning in free-viewing conditions. Invest. Ophthalmol. Vis. Sci., March 1, 2006; 47(3): 1241 - 1250. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. I. Fried, H. A. Hsueh, and F. S. Werblin A Method for Generating Precise Temporal Patterns of Retinal Spiking Using Prosthetic Stimulation J Neurophysiol, February 1, 2006; 95(2): 970 - 978. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. P. Fornos, J. Sommerhalder, B. Rappaz, A. B. Safran, and M. Pelizzone Simulation of Artificial Vision, III: Do the Spatial or Temporal Characteristics of Stimulus Pixelization Really Matter? Invest. Ophthalmol. Vis. Sci., October 1, 2005; 46(10): 3906 - 3912. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. J. Jensen, O. R. Ziv, and J. F. Rizzo III Thresholds for Activation of Rabbit Retinal Ganglion Cells with Relatively Large, Extracellular Microelectrodes Invest. Ophthalmol. Vis. Sci., April 1, 2005; 46(4): 1486 - 1496. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. Palanker, P. Huie, A. Vankov, R. Aramant, M. Seiler, H. Fishman, M. Marmor, and M. Blumenkranz Migration of Retinal Cells through a Perforated Membrane: Implications for a High-Resolution Prosthesis Invest. Ophthalmol. Vis. Sci., September 1, 2004; 45(9): 3266 - 3270. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. I. Loewenstein, S. R. Montezuma, and J. F. Rizzo III Outer Retinal Degeneration: An Electronic Retinal Prosthesis as a Treatment Strategy Arch Ophthalmol, April 1, 2004; 122(4): 587 - 596. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. F. Rizzo III, J. Wyatt, J. Loewenstein, S. Kelly, and D. Shire Methods and Perceptual Thresholds for Short-Term Electrical Stimulation of Human Retina with Microelectrode Arrays Invest. Ophthalmol. Vis. Sci., December 1, 2003; 44(12): 5355 - 5361. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |