spevi lecture 13jan2015 · 16/01/2015' 1' the$hopes$and$challenges$of$a$bionic$eye:$...
TRANSCRIPT
16/01/2015'
1'
The$Hopes$and$Challenges$of$a$Bionic$Eye:$Implica8ons$for$Educa8on$and$Learning'
Dr#Lauren#Ayton#SPEVI#Conference,#13#January#2015#
Bionic'Vision'Australia'
How'Does'Vision'Work?' Common'Causes'of'Vision'Loss'
Corneal'disease'A Keratoconus'
Cataracts'
Glaucoma'
ReEnal'disease'A DiabeEc'reEnopathy'A ReEniEs'pigmentosa'A Macular'degeneraEon'
Neurological'A Stroke'A Tumours'Trauma'
Vision'RestoraEon'OpEons'
1. ReEnal'Prostheses'(“bionic'eyes”)'
2. Stem'cells'
3. Gene'therapy'
4. OptogeneEcs'
InternaEonal'Research'Biohybrid Retinal Implant Tokyo, Japan Intelligent Medical Implants (IMI)
Bonn, Germany (affiliated with Pixium, France)
Retina Implant GmbH Tuebingen/Reutlingen, Germany
Illinois Institute of Technology Chicago, Illinois
3-D Stacked Retinal Prosthesis Tohoku University, Japan
Japanese Artificial Vision Project Osaka, Japan
Okayama University Okayama, Japan
Nano Bioelectrics and Systems Research Center Seoul, South Korea
C-Sight: Chinese Project for Sight Beijing, China
Boston Retinal Implant Project Boston/Cambridge, Massachusetts
CORTIVIS, Alicante, Spain
Ligon Research Center of Vision Detroit, Michigan
Second Sight Medical Products Inc. Sylmar, California
Bionic Vision Australia Melbourne/Sydney/Canberra, Australia
Stanford University, Palo Alto, California (affiliated with Pixium, France)
Neural Rehabilitation Engineering Laboratory Brussels, Belgium
Pixium Paris, France
National Tsing Hua University, Taiwan
Monash Vision Group, Melbourne, Australia
Epi-Ret Aachen, Germany
NanoRetina, Israel
Arizona State University, Phoenix, Arizona
Polystim Neurotechnologies Montreal, Canada
Groups highlighted in red are currently undertaking human clinical trials Map by Joe Rizzo & Lauren Ayton Updated 19 May 2014
CURRENTLY ACTIVE INTERNATIONAL VISION PROSTHESIS GROUPS
16/01/2015'
2'
The'Bionic'Eye' How'Does'It'Work?'
Bionic'Eye'LocaEons'
Retinal implants
Cortical implants (ie. Monash)
LGN implants
ON implants
What'Will'It'Look'Like?'
Phosphene'Vision' Phosphene'Vision'
16/01/2015'
3'
BVA'Prototype'Study'
24$s8mula8ng'electrodes'
No'implanted'electronics'
Materials:'pla8num'and'silicone$
Suprachoroidal$surgical'placement'
3$pa8ents$with$profound'vision'loss'from'reEniEs'
pigmentosa'(light'percepEon'only)$
2$year$study$
The'Prototype'Device'A
B#
Figure 5. Phosphene drawings for each electrode from two patients, obtained using a motion tracking sensor attachedto the patient’s index finger. Drawings are arranged to correspond with their approximate location within the subject’svisual field. Closed shapes were filled in with a grey/white colour unless otherwise noted. P1 (red drawings) exhibiteda range of complex and simple phosphene shapes depending on the proximity of the electrode to the fovea. In contrast,phosphenes for P2 (blue drawings) were generally all blob shapes.Note: phosphene sizes are not to scale and have been adjusted for display purposes. Stimulation parameters: P1 – 500ʅs phase width, 500ʅsinterphase gap, 50pps, 2s, anodic first stimulation at 6dB above threshold; P2 – 148ʅs phase width, 20ʅs interphase gap, 400pps, 2s, anodicfirst stimulation at 4dB above threshold.
PURPOSERetinal prostheses aim to provide functional vision in profoundly vision-impairedpatients using spatiotemporal patterns of electrical stimulation delivered to anelectrode array implanted into the eye. In this study, we hypothesised thatelectrodes implanted into the suprachoroidal space between the choroid and thesclera would produce phosphenes suitable for the representation of visualinformation to blind patients with retinitis pigmentosa.
PSYCHOPHYSICS OF A SUPRACHOROIDAL RETINAL PROSTHESISPeter Blamey1,2, Nicholas Sinclair1, Kyle Slater1, Hugh McDermott1,2, Mohit Shivdasani1,2, Thushara Perera1,
Peter Dimitrov3, Mary Varsamidis3, Lauren Ayton3, Robyn Guymer3 and Chi Luu3
1Bionics Institute, AUSTRALIA 2Department of Medical Bionics, The University of Melbourne, AUSTRALIA 3 Centre for Eye Research Australia, AUSTRALIA
ACKNOWLEDGEMENTS: This research was supported by the Australian Research Council (ARC) through its SpecialResearch Initiative (SRI) in Bionic Vision Science and Technology grant to Bionic Vision Australia (BVA). The BionicsInstitute acknowledges the support it receives from the Victorian Government through its operational infrastructure program.
RESULTSFor each patient, the number of electrodes capable of eliciting a percept varied relative tothe stimulation parameters and electrode configurations (Figure 2). Electrode impedanceswere in the range 15-31 k: (P1), 16-27 k: (P2) and 11-24 k: (P3) for stimulation withbiphasic current pulses with a 500ʅs per phase duration. Thresholds were lowest (down to50nC per phase) when the anodic phase preceded the cathodic phase of each pulse, themonopolar electrode configuration was used, and pulse rates of 200-500 pulses per second(pps) were used. Dynamic ranges were limited by the maximum safe charge density perphase. Phosphene shape, size and position did not vary greatly between return electrodeconfigurations (monopolar and common ground; Figure 3). Electrode recognition variedbetween the patients (Figure 4). Phosphenes varied from quite complex (including lightand dark regions) for electrodes close to the fovea, and became simpler for more peripheralelectrodes (such as grey cloudy convex shapes; Figure 5). The perceived position ofphosphenes in space varied with head position and eye gaze direction. Phosphenes tendedto become larger and/or more intense as charge per phase was increased. Severalphosphene maps using monopolar and common ground electrode configurations wereconstructed from the data and used to create small sets of recognisable stimuli representingnumerals or letters of the alphabet (Figure 6). It was found that the order of stimulation ofelectrodes in an interleaved pattern affected the combined percepts.
CONCLUSIONSA suprachoroidal retinal prosthesis with relatively large electrodes produceddistinct phosphenes when stimulated in monopolar or common groundelectrode configurations, and these phosphenes were used to ‘paint’ distinctiveshapes in blind patients.
METHODSEach patient was implanted with a suprachoroidal electrode array comprised of seventeen600ʅm and three 400ʅm platinum disc electrodes and several return electrode configurations(Figure 1). Pre-operatively, all patients had bare light perception and could not recogniseshapes. A specially designed stimulator and psychophysics test setup were used to measureelectrode impedances, thresholds and dynamic ranges, as well as the perceived shape, size,position and intensity of phosphenes produced by stimulating one electrode at a time. Thesedata were included in several phosphene ‘maps’ suitable for encoding images or creatingcomplex shapes by stimulating multiple electrodes in an interleaved fashion.
SUMMARY OF FINDINGS• Electrical stimulation of individual electrodes produced distinct phosphenes• The shape of the phosphenes was not affected by the return configurations• Electrode recognition varied between patients• Lower thresholds were observed when the anodic phase of stimulation preceded
the cathodic phase• The complexity of phosphenes perceived by the patients varied relative to the
proximity of the electrode to the fovea• Patients could identify basic shapes, letters and numbers
Figure 4. Electrode recognition results for two patients. Each electrode was individually stimulated at 6dB abovethreshold in a random order and the patient was asked to estimate which electrode was used. Plots show normalisedhistograms of the electrodes estimated versus the actual electrodes stimulated across all trials (NS = not seen). P1(left plot) correctly identified the electrode used in 66.7% of trials. P2 (right plot) correctly identified the electrodeused in 33.3% of trials. Note that due to the electrode layout, an error of 5 places indicates that the estimatedelectrode was horizontally adjacent to the actual electrode (see Figure 1B).Stimulation parameters: P1 – Monopolar configuration, 500ʅs phase width, 500ʅs interphase gap, 50pps, 2s, cathodic phase first;P2 – Monopolar configuration,148ʅs phase width, 20ʅs interphase gap, 400pps, 2s, anodic phase first.
Commercial Relationships Disclosure: Peter Blamey: Bionics Institute – Code P (Patent), Cochlear Limited – Code P(Patent); Robyn Guymer: Novartis Advisory Board – Code C (Consultant), Bayer Advisory Board – Code C (Consultant),Novartis – Code R (Recipient).
PATIENTSThree retinitis pigmentosa (RP) patients comprised this study (Table 1).
PROGRAM # 1044
24chPP1 24chPP2 24chPP3Gender Female Male Male
Age 53yo 50yo 63yoEye condition RP RP - syndromic RP
Current level of vision Light perception only Light perception only Light perception onlyYears of blindness Approx. 20 years Approx. 8 – 10 years Approx. 20 years
Primary mobility aid Guide dog Guide dog Guide dog
24chPP1 24chPP2 24chPP3
Monopolar500µs PW, 500µs IPG, 50pps
19/19 (CF)20/20 (AF)
18/20 (AF) No perceptsobtained
Common Ground500µs PW, 500µs IPG, 50pps
19/19 (CF)19/19 (AF)
18/20 (AF) No perceptsobtained
Hexagonal500µs PW, 500µs IPG, 50pps
4/8 (CF)7/8 (AF)
1/8 (AF) Not tested
Monopolar500µs PW, 500µs IPG, 500pps
Not tested 20/20 (AF) Not tested
Monopolar148µs PW, 20µs IPG, 400pps
Not tested 17/20 (AF) Not tested
Monopolar Ganged Pairs500µs PW, 500µs IPG, 500pps
Not tested Not tested 9/9 (AF)
Monopolar Ganged Pairs200µs PW, 200µs IPG, 200pps
Not tested Not tested 9/9 (AF)
Figure 2. Electrode yield for different electrodeconfigurations. The table (A) shows the number ofelectrodes that were capable of eliciting a visualpercept using various stimulation parameters andelectrode configurations. The active (red) andreturn (black) electrodes used for each electrodeconfiguration are shown in (B). For P1, reliableand consistent percepts were obtained using astimulation rate of 50pps. For P2, whilst perceptswere obtained using 50pps, they mainly consistedof only onset and offset flashes. Increasing thestimulation rate to 400pps or 500pps producedpersistent and reliable percepts. For P3, highstimulation rates and ganged pairs of electrodeswere required in order to obtain visual percepts.Narrow pulse widths were also tested with P2 andP3 to allow stimulation on multiple electrodes to beinterleaved in time at high rates to form patterns(see Figure 6).PW = phase width, IPG = interphase gap, CF = cathodicphase first, AF = anodic phase first. Stimulus duration was 2sin all cases.
Monopolar Ganged Pair
Common Ground
Hexagonal
Monopolar
B)
A)
Figure 3. Phosphene drawings comparing monopolar and common ground return configurations at 2dB abovethreshold for several electrodes. Drawings were obtained from P1 using a motion tracking sensor attached to thepatient’s index finger. Phosphenes produced using monopolar stimulation (M; purple) were very similar to thoseproduced with common ground stimulation (CG; green). Offsets in position can be attributed to minor changes in eyeand head position between stimuli.Stimulation parameters: P1 – 500ʅs phase width, 500ʅs interphase gap, 50pps, 2s, cathodic first stimulation.
Figure 6. Percepts as drawn by P2 for each patternof electrodes stimulated in an interleaved fashion.Drawings were obtained using a motion trackingsensor attached to the patient’s index finger. Aboveeach percept is the description given by the patient.Inset shows the electrodes (corresponding to visualspace) that were used for each pattern (Asterisk (*)denotes the position of E1 on the electrode array).
Figure 1. (A) The suprachoroidalelectrode array has 33 platinum discelectrodes (30 u 600µm; 3 u 400µm)and 2 large return electrodes (2mmdiameter). Note that the electrodeson the outer edges (top, bottom andright edge) of the implant are shortedtogether to form an extra ‘guard’return, thus giving 20 individualstimulating electrodes (B).
A)1 6 11 16
2 7 12 17
3 8 13 18
4 9 14 19
5 10 15 20
B)C## 21a 21b 21c 21d
21e
21f
21g
21h
21i
21j 21m 21l 21k Percutaneous#connector#
Intraocular#array#
Helical#lead#wire#
Third#(remote)#return#electrode#
Silicone#array#
Pla<num#s<mula<ng#electrodes#2#x#return#electrodes#
D##
E##
Figure 5. Phosphene drawings for each electrode from two patients, obtained using a motion tracking sensor attachedto the patient’s index finger. Drawings are arranged to correspond with their approximate location within the subject’svisual field. Closed shapes were filled in with a grey/white colour unless otherwise noted. P1 (red drawings) exhibiteda range of complex and simple phosphene shapes depending on the proximity of the electrode to the fovea. In contrast,phosphenes for P2 (blue drawings) were generally all blob shapes.Note: phosphene sizes are not to scale and have been adjusted for display purposes. Stimulation parameters: P1 – 500ʅs phase width, 500ʅsinterphase gap, 50pps, 2s, anodic first stimulation at 6dB above threshold; P2 – 148ʅs phase width, 20ʅs interphase gap, 400pps, 2s, anodicfirst stimulation at 4dB above threshold.
PURPOSERetinal prostheses aim to provide functional vision in profoundly vision-impairedpatients using spatiotemporal patterns of electrical stimulation delivered to anelectrode array implanted into the eye. In this study, we hypothesised thatelectrodes implanted into the suprachoroidal space between the choroid and thesclera would produce phosphenes suitable for the representation of visualinformation to blind patients with retinitis pigmentosa.
PSYCHOPHYSICS OF A SUPRACHOROIDAL RETINAL PROSTHESISPeter Blamey1,2, Nicholas Sinclair1, Kyle Slater1, Hugh McDermott1,2, Mohit Shivdasani1,2, Thushara Perera1,
Peter Dimitrov3, Mary Varsamidis3, Lauren Ayton3, Robyn Guymer3 and Chi Luu3
1Bionics Institute, AUSTRALIA 2Department of Medical Bionics, The University of Melbourne, AUSTRALIA 3 Centre for Eye Research Australia, AUSTRALIA
ACKNOWLEDGEMENTS: This research was supported by the Australian Research Council (ARC) through its SpecialResearch Initiative (SRI) in Bionic Vision Science and Technology grant to Bionic Vision Australia (BVA). The BionicsInstitute acknowledges the support it receives from the Victorian Government through its operational infrastructure program.
RESULTSFor each patient, the number of electrodes capable of eliciting a percept varied relative tothe stimulation parameters and electrode configurations (Figure 2). Electrode impedanceswere in the range 15-31 k: (P1), 16-27 k: (P2) and 11-24 k: (P3) for stimulation withbiphasic current pulses with a 500ʅs per phase duration. Thresholds were lowest (down to50nC per phase) when the anodic phase preceded the cathodic phase of each pulse, themonopolar electrode configuration was used, and pulse rates of 200-500 pulses per second(pps) were used. Dynamic ranges were limited by the maximum safe charge density perphase. Phosphene shape, size and position did not vary greatly between return electrodeconfigurations (monopolar and common ground; Figure 3). Electrode recognition variedbetween the patients (Figure 4). Phosphenes varied from quite complex (including lightand dark regions) for electrodes close to the fovea, and became simpler for more peripheralelectrodes (such as grey cloudy convex shapes; Figure 5). The perceived position ofphosphenes in space varied with head position and eye gaze direction. Phosphenes tendedto become larger and/or more intense as charge per phase was increased. Severalphosphene maps using monopolar and common ground electrode configurations wereconstructed from the data and used to create small sets of recognisable stimuli representingnumerals or letters of the alphabet (Figure 6). It was found that the order of stimulation ofelectrodes in an interleaved pattern affected the combined percepts.
CONCLUSIONSA suprachoroidal retinal prosthesis with relatively large electrodes produceddistinct phosphenes when stimulated in monopolar or common groundelectrode configurations, and these phosphenes were used to ‘paint’ distinctiveshapes in blind patients.
METHODSEach patient was implanted with a suprachoroidal electrode array comprised of seventeen600ʅm and three 400ʅm platinum disc electrodes and several return electrode configurations(Figure 1). Pre-operatively, all patients had bare light perception and could not recogniseshapes. A specially designed stimulator and psychophysics test setup were used to measureelectrode impedances, thresholds and dynamic ranges, as well as the perceived shape, size,position and intensity of phosphenes produced by stimulating one electrode at a time. Thesedata were included in several phosphene ‘maps’ suitable for encoding images or creatingcomplex shapes by stimulating multiple electrodes in an interleaved fashion.
SUMMARY OF FINDINGS• Electrical stimulation of individual electrodes produced distinct phosphenes• The shape of the phosphenes was not affected by the return configurations• Electrode recognition varied between patients• Lower thresholds were observed when the anodic phase of stimulation preceded
the cathodic phase• The complexity of phosphenes perceived by the patients varied relative to the
proximity of the electrode to the fovea• Patients could identify basic shapes, letters and numbers
Figure 4. Electrode recognition results for two patients. Each electrode was individually stimulated at 6dB abovethreshold in a random order and the patient was asked to estimate which electrode was used. Plots show normalisedhistograms of the electrodes estimated versus the actual electrodes stimulated across all trials (NS = not seen). P1(left plot) correctly identified the electrode used in 66.7% of trials. P2 (right plot) correctly identified the electrodeused in 33.3% of trials. Note that due to the electrode layout, an error of 5 places indicates that the estimatedelectrode was horizontally adjacent to the actual electrode (see Figure 1B).Stimulation parameters: P1 – Monopolar configuration, 500ʅs phase width, 500ʅs interphase gap, 50pps, 2s, cathodic phase first;P2 – Monopolar configuration,148ʅs phase width, 20ʅs interphase gap, 400pps, 2s, anodic phase first.
Commercial Relationships Disclosure: Peter Blamey: Bionics Institute – Code P (Patent), Cochlear Limited – Code P(Patent); Robyn Guymer: Novartis Advisory Board – Code C (Consultant), Bayer Advisory Board – Code C (Consultant),Novartis – Code R (Recipient).
PATIENTSThree retinitis pigmentosa (RP) patients comprised this study (Table 1).
PROGRAM # 1044
24chPP1 24chPP2 24chPP3Gender Female Male Male
Age 53yo 50yo 63yoEye condition RP RP - syndromic RP
Current level of vision Light perception only Light perception only Light perception onlyYears of blindness Approx. 20 years Approx. 8 – 10 years Approx. 20 years
Primary mobility aid Guide dog Guide dog Guide dog
24chPP1 24chPP2 24chPP3
Monopolar500µs PW, 500µs IPG, 50pps
19/19 (CF)20/20 (AF)
18/20 (AF) No perceptsobtained
Common Ground500µs PW, 500µs IPG, 50pps
19/19 (CF)19/19 (AF)
18/20 (AF) No perceptsobtained
Hexagonal500µs PW, 500µs IPG, 50pps
4/8 (CF)7/8 (AF)
1/8 (AF) Not tested
Monopolar500µs PW, 500µs IPG, 500pps
Not tested 20/20 (AF) Not tested
Monopolar148µs PW, 20µs IPG, 400pps
Not tested 17/20 (AF) Not tested
Monopolar Ganged Pairs500µs PW, 500µs IPG, 500pps
Not tested Not tested 9/9 (AF)
Monopolar Ganged Pairs200µs PW, 200µs IPG, 200pps
Not tested Not tested 9/9 (AF)
Figure 2. Electrode yield for different electrodeconfigurations. The table (A) shows the number ofelectrodes that were capable of eliciting a visualpercept using various stimulation parameters andelectrode configurations. The active (red) andreturn (black) electrodes used for each electrodeconfiguration are shown in (B). For P1, reliableand consistent percepts were obtained using astimulation rate of 50pps. For P2, whilst perceptswere obtained using 50pps, they mainly consistedof only onset and offset flashes. Increasing thestimulation rate to 400pps or 500pps producedpersistent and reliable percepts. For P3, highstimulation rates and ganged pairs of electrodeswere required in order to obtain visual percepts.Narrow pulse widths were also tested with P2 andP3 to allow stimulation on multiple electrodes to beinterleaved in time at high rates to form patterns(see Figure 6).PW = phase width, IPG = interphase gap, CF = cathodicphase first, AF = anodic phase first. Stimulus duration was 2sin all cases.
Monopolar Ganged Pair
Common Ground
Hexagonal
Monopolar
B)
A)
Figure 3. Phosphene drawings comparing monopolar and common ground return configurations at 2dB abovethreshold for several electrodes. Drawings were obtained from P1 using a motion tracking sensor attached to thepatient’s index finger. Phosphenes produced using monopolar stimulation (M; purple) were very similar to thoseproduced with common ground stimulation (CG; green). Offsets in position can be attributed to minor changes in eyeand head position between stimuli.Stimulation parameters: P1 – 500ʅs phase width, 500ʅs interphase gap, 50pps, 2s, cathodic first stimulation.
Figure 6. Percepts as drawn by P2 for each patternof electrodes stimulated in an interleaved fashion.Drawings were obtained using a motion trackingsensor attached to the patient’s index finger. Aboveeach percept is the description given by the patient.Inset shows the electrodes (corresponding to visualspace) that were used for each pattern (Asterisk (*)denotes the position of E1 on the electrode array).
Figure 1. (A) The suprachoroidalelectrode array has 33 platinum discelectrodes (30 u 600µm; 3 u 400µm)and 2 large return electrodes (2mmdiameter). Note that the electrodeson the outer edges (top, bottom andright edge) of the implant are shortedtogether to form an extra ‘guard’return, thus giving 20 individualstimulating electrodes (B).
A)1 6 11 16
2 7 12 17
3 8 13 18
4 9 14 19
5 10 15 20
B)
Figure 5. Phosphene drawings for each electrode from two patients, obtained using a motion tracking sensor attachedto the patient’s index finger. Drawings are arranged to correspond with their approximate location within the subject’svisual field. Closed shapes were filled in with a grey/white colour unless otherwise noted. P1 (red drawings) exhibiteda range of complex and simple phosphene shapes depending on the proximity of the electrode to the fovea. In contrast,phosphenes for P2 (blue drawings) were generally all blob shapes.Note: phosphene sizes are not to scale and have been adjusted for display purposes. Stimulation parameters: P1 – 500ʅs phase width, 500ʅsinterphase gap, 50pps, 2s, anodic first stimulation at 6dB above threshold; P2 – 148ʅs phase width, 20ʅs interphase gap, 400pps, 2s, anodicfirst stimulation at 4dB above threshold.
PURPOSERetinal prostheses aim to provide functional vision in profoundly vision-impairedpatients using spatiotemporal patterns of electrical stimulation delivered to anelectrode array implanted into the eye. In this study, we hypothesised thatelectrodes implanted into the suprachoroidal space between the choroid and thesclera would produce phosphenes suitable for the representation of visualinformation to blind patients with retinitis pigmentosa.
PSYCHOPHYSICS OF A SUPRACHOROIDAL RETINAL PROSTHESISPeter Blamey1,2, Nicholas Sinclair1, Kyle Slater1, Hugh McDermott1,2, Mohit Shivdasani1,2, Thushara Perera1,
Peter Dimitrov3, Mary Varsamidis3, Lauren Ayton3, Robyn Guymer3 and Chi Luu3
1Bionics Institute, AUSTRALIA 2Department of Medical Bionics, The University of Melbourne, AUSTRALIA 3 Centre for Eye Research Australia, AUSTRALIA
ACKNOWLEDGEMENTS: This research was supported by the Australian Research Council (ARC) through its SpecialResearch Initiative (SRI) in Bionic Vision Science and Technology grant to Bionic Vision Australia (BVA). The BionicsInstitute acknowledges the support it receives from the Victorian Government through its operational infrastructure program.
RESULTSFor each patient, the number of electrodes capable of eliciting a percept varied relative tothe stimulation parameters and electrode configurations (Figure 2). Electrode impedanceswere in the range 15-31 k: (P1), 16-27 k: (P2) and 11-24 k: (P3) for stimulation withbiphasic current pulses with a 500ʅs per phase duration. Thresholds were lowest (down to50nC per phase) when the anodic phase preceded the cathodic phase of each pulse, themonopolar electrode configuration was used, and pulse rates of 200-500 pulses per second(pps) were used. Dynamic ranges were limited by the maximum safe charge density perphase. Phosphene shape, size and position did not vary greatly between return electrodeconfigurations (monopolar and common ground; Figure 3). Electrode recognition variedbetween the patients (Figure 4). Phosphenes varied from quite complex (including lightand dark regions) for electrodes close to the fovea, and became simpler for more peripheralelectrodes (such as grey cloudy convex shapes; Figure 5). The perceived position ofphosphenes in space varied with head position and eye gaze direction. Phosphenes tendedto become larger and/or more intense as charge per phase was increased. Severalphosphene maps using monopolar and common ground electrode configurations wereconstructed from the data and used to create small sets of recognisable stimuli representingnumerals or letters of the alphabet (Figure 6). It was found that the order of stimulation ofelectrodes in an interleaved pattern affected the combined percepts.
CONCLUSIONSA suprachoroidal retinal prosthesis with relatively large electrodes produceddistinct phosphenes when stimulated in monopolar or common groundelectrode configurations, and these phosphenes were used to ‘paint’ distinctiveshapes in blind patients.
METHODSEach patient was implanted with a suprachoroidal electrode array comprised of seventeen600ʅm and three 400ʅm platinum disc electrodes and several return electrode configurations(Figure 1). Pre-operatively, all patients had bare light perception and could not recogniseshapes. A specially designed stimulator and psychophysics test setup were used to measureelectrode impedances, thresholds and dynamic ranges, as well as the perceived shape, size,position and intensity of phosphenes produced by stimulating one electrode at a time. Thesedata were included in several phosphene ‘maps’ suitable for encoding images or creatingcomplex shapes by stimulating multiple electrodes in an interleaved fashion.
SUMMARY OF FINDINGS• Electrical stimulation of individual electrodes produced distinct phosphenes• The shape of the phosphenes was not affected by the return configurations• Electrode recognition varied between patients• Lower thresholds were observed when the anodic phase of stimulation preceded
the cathodic phase• The complexity of phosphenes perceived by the patients varied relative to the
proximity of the electrode to the fovea• Patients could identify basic shapes, letters and numbers
Figure 4. Electrode recognition results for two patients. Each electrode was individually stimulated at 6dB abovethreshold in a random order and the patient was asked to estimate which electrode was used. Plots show normalisedhistograms of the electrodes estimated versus the actual electrodes stimulated across all trials (NS = not seen). P1(left plot) correctly identified the electrode used in 66.7% of trials. P2 (right plot) correctly identified the electrodeused in 33.3% of trials. Note that due to the electrode layout, an error of 5 places indicates that the estimatedelectrode was horizontally adjacent to the actual electrode (see Figure 1B).Stimulation parameters: P1 – Monopolar configuration, 500ʅs phase width, 500ʅs interphase gap, 50pps, 2s, cathodic phase first;P2 – Monopolar configuration,148ʅs phase width, 20ʅs interphase gap, 400pps, 2s, anodic phase first.
Commercial Relationships Disclosure: Peter Blamey: Bionics Institute – Code P (Patent), Cochlear Limited – Code P(Patent); Robyn Guymer: Novartis Advisory Board – Code C (Consultant), Bayer Advisory Board – Code C (Consultant),Novartis – Code R (Recipient).
PATIENTSThree retinitis pigmentosa (RP) patients comprised this study (Table 1).
PROGRAM # 1044
24chPP1 24chPP2 24chPP3Gender Female Male Male
Age 53yo 50yo 63yoEye condition RP RP - syndromic RP
Current level of vision Light perception only Light perception only Light perception onlyYears of blindness Approx. 20 years Approx. 8 – 10 years Approx. 20 years
Primary mobility aid Guide dog Guide dog Guide dog
24chPP1 24chPP2 24chPP3
Monopolar500µs PW, 500µs IPG, 50pps
19/19 (CF)20/20 (AF)
18/20 (AF) No perceptsobtained
Common Ground500µs PW, 500µs IPG, 50pps
19/19 (CF)19/19 (AF)
18/20 (AF) No perceptsobtained
Hexagonal500µs PW, 500µs IPG, 50pps
4/8 (CF)7/8 (AF)
1/8 (AF) Not tested
Monopolar500µs PW, 500µs IPG, 500pps
Not tested 20/20 (AF) Not tested
Monopolar148µs PW, 20µs IPG, 400pps
Not tested 17/20 (AF) Not tested
Monopolar Ganged Pairs500µs PW, 500µs IPG, 500pps
Not tested Not tested 9/9 (AF)
Monopolar Ganged Pairs200µs PW, 200µs IPG, 200pps
Not tested Not tested 9/9 (AF)
Figure 2. Electrode yield for different electrodeconfigurations. The table (A) shows the number ofelectrodes that were capable of eliciting a visualpercept using various stimulation parameters andelectrode configurations. The active (red) andreturn (black) electrodes used for each electrodeconfiguration are shown in (B). For P1, reliableand consistent percepts were obtained using astimulation rate of 50pps. For P2, whilst perceptswere obtained using 50pps, they mainly consistedof only onset and offset flashes. Increasing thestimulation rate to 400pps or 500pps producedpersistent and reliable percepts. For P3, highstimulation rates and ganged pairs of electrodeswere required in order to obtain visual percepts.Narrow pulse widths were also tested with P2 andP3 to allow stimulation on multiple electrodes to beinterleaved in time at high rates to form patterns(see Figure 6).PW = phase width, IPG = interphase gap, CF = cathodicphase first, AF = anodic phase first. Stimulus duration was 2sin all cases.
Monopolar Ganged Pair
Common Ground
Hexagonal
Monopolar
B)
A)
Figure 3. Phosphene drawings comparing monopolar and common ground return configurations at 2dB abovethreshold for several electrodes. Drawings were obtained from P1 using a motion tracking sensor attached to thepatient’s index finger. Phosphenes produced using monopolar stimulation (M; purple) were very similar to thoseproduced with common ground stimulation (CG; green). Offsets in position can be attributed to minor changes in eyeand head position between stimuli.Stimulation parameters: P1 – 500ʅs phase width, 500ʅs interphase gap, 50pps, 2s, cathodic first stimulation.
Figure 6. Percepts as drawn by P2 for each patternof electrodes stimulated in an interleaved fashion.Drawings were obtained using a motion trackingsensor attached to the patient’s index finger. Aboveeach percept is the description given by the patient.Inset shows the electrodes (corresponding to visualspace) that were used for each pattern (Asterisk (*)denotes the position of E1 on the electrode array).
Figure 1. (A) The suprachoroidalelectrode array has 33 platinum discelectrodes (30 u 600µm; 3 u 400µm)and 2 large return electrodes (2mmdiameter). Note that the electrodeson the outer edges (top, bottom andright edge) of the implant are shortedtogether to form an extra ‘guard’return, thus giving 20 individualstimulating electrodes (B).
A)1 6 11 16
2 7 12 17
3 8 13 18
4 9 14 19
5 10 15 20
B)
Figure 5. Phosphene drawings for each electrode from two patients, obtained using a motion tracking sensor attachedto the patient’s index finger. Drawings are arranged to correspond with their approximate location within the subject’svisual field. Closed shapes were filled in with a grey/white colour unless otherwise noted. P1 (red drawings) exhibiteda range of complex and simple phosphene shapes depending on the proximity of the electrode to the fovea. In contrast,phosphenes for P2 (blue drawings) were generally all blob shapes.Note: phosphene sizes are not to scale and have been adjusted for display purposes. Stimulation parameters: P1 – 500ʅs phase width, 500ʅsinterphase gap, 50pps, 2s, anodic first stimulation at 6dB above threshold; P2 – 148ʅs phase width, 20ʅs interphase gap, 400pps, 2s, anodicfirst stimulation at 4dB above threshold.
PURPOSERetinal prostheses aim to provide functional vision in profoundly vision-impairedpatients using spatiotemporal patterns of electrical stimulation delivered to anelectrode array implanted into the eye. In this study, we hypothesised thatelectrodes implanted into the suprachoroidal space between the choroid and thesclera would produce phosphenes suitable for the representation of visualinformation to blind patients with retinitis pigmentosa.
PSYCHOPHYSICS OF A SUPRACHOROIDAL RETINAL PROSTHESISPeter Blamey1,2, Nicholas Sinclair1, Kyle Slater1, Hugh McDermott1,2, Mohit Shivdasani1,2, Thushara Perera1,
Peter Dimitrov3, Mary Varsamidis3, Lauren Ayton3, Robyn Guymer3 and Chi Luu3
1Bionics Institute, AUSTRALIA 2Department of Medical Bionics, The University of Melbourne, AUSTRALIA 3 Centre for Eye Research Australia, AUSTRALIA
ACKNOWLEDGEMENTS: This research was supported by the Australian Research Council (ARC) through its SpecialResearch Initiative (SRI) in Bionic Vision Science and Technology grant to Bionic Vision Australia (BVA). The BionicsInstitute acknowledges the support it receives from the Victorian Government through its operational infrastructure program.
RESULTSFor each patient, the number of electrodes capable of eliciting a percept varied relative tothe stimulation parameters and electrode configurations (Figure 2). Electrode impedanceswere in the range 15-31 k: (P1), 16-27 k: (P2) and 11-24 k: (P3) for stimulation withbiphasic current pulses with a 500ʅs per phase duration. Thresholds were lowest (down to50nC per phase) when the anodic phase preceded the cathodic phase of each pulse, themonopolar electrode configuration was used, and pulse rates of 200-500 pulses per second(pps) were used. Dynamic ranges were limited by the maximum safe charge density perphase. Phosphene shape, size and position did not vary greatly between return electrodeconfigurations (monopolar and common ground; Figure 3). Electrode recognition variedbetween the patients (Figure 4). Phosphenes varied from quite complex (including lightand dark regions) for electrodes close to the fovea, and became simpler for more peripheralelectrodes (such as grey cloudy convex shapes; Figure 5). The perceived position ofphosphenes in space varied with head position and eye gaze direction. Phosphenes tendedto become larger and/or more intense as charge per phase was increased. Severalphosphene maps using monopolar and common ground electrode configurations wereconstructed from the data and used to create small sets of recognisable stimuli representingnumerals or letters of the alphabet (Figure 6). It was found that the order of stimulation ofelectrodes in an interleaved pattern affected the combined percepts.
CONCLUSIONSA suprachoroidal retinal prosthesis with relatively large electrodes produceddistinct phosphenes when stimulated in monopolar or common groundelectrode configurations, and these phosphenes were used to ‘paint’ distinctiveshapes in blind patients.
METHODSEach patient was implanted with a suprachoroidal electrode array comprised of seventeen600ʅm and three 400ʅm platinum disc electrodes and several return electrode configurations(Figure 1). Pre-operatively, all patients had bare light perception and could not recogniseshapes. A specially designed stimulator and psychophysics test setup were used to measureelectrode impedances, thresholds and dynamic ranges, as well as the perceived shape, size,position and intensity of phosphenes produced by stimulating one electrode at a time. Thesedata were included in several phosphene ‘maps’ suitable for encoding images or creatingcomplex shapes by stimulating multiple electrodes in an interleaved fashion.
SUMMARY OF FINDINGS• Electrical stimulation of individual electrodes produced distinct phosphenes• The shape of the phosphenes was not affected by the return configurations• Electrode recognition varied between patients• Lower thresholds were observed when the anodic phase of stimulation preceded
the cathodic phase• The complexity of phosphenes perceived by the patients varied relative to the
proximity of the electrode to the fovea• Patients could identify basic shapes, letters and numbers
Figure 4. Electrode recognition results for two patients. Each electrode was individually stimulated at 6dB abovethreshold in a random order and the patient was asked to estimate which electrode was used. Plots show normalisedhistograms of the electrodes estimated versus the actual electrodes stimulated across all trials (NS = not seen). P1(left plot) correctly identified the electrode used in 66.7% of trials. P2 (right plot) correctly identified the electrodeused in 33.3% of trials. Note that due to the electrode layout, an error of 5 places indicates that the estimatedelectrode was horizontally adjacent to the actual electrode (see Figure 1B).Stimulation parameters: P1 – Monopolar configuration, 500ʅs phase width, 500ʅs interphase gap, 50pps, 2s, cathodic phase first;P2 – Monopolar configuration,148ʅs phase width, 20ʅs interphase gap, 400pps, 2s, anodic phase first.
Commercial Relationships Disclosure: Peter Blamey: Bionics Institute – Code P (Patent), Cochlear Limited – Code P(Patent); Robyn Guymer: Novartis Advisory Board – Code C (Consultant), Bayer Advisory Board – Code C (Consultant),Novartis – Code R (Recipient).
PATIENTSThree retinitis pigmentosa (RP) patients comprised this study (Table 1).
PROGRAM # 1044
24chPP1 24chPP2 24chPP3Gender Female Male Male
Age 53yo 50yo 63yoEye condition RP RP - syndromic RP
Current level of vision Light perception only Light perception only Light perception onlyYears of blindness Approx. 20 years Approx. 8 – 10 years Approx. 20 years
Primary mobility aid Guide dog Guide dog Guide dog
24chPP1 24chPP2 24chPP3
Monopolar500µs PW, 500µs IPG, 50pps
19/19 (CF)20/20 (AF)
18/20 (AF) No perceptsobtained
Common Ground500µs PW, 500µs IPG, 50pps
19/19 (CF)19/19 (AF)
18/20 (AF) No perceptsobtained
Hexagonal500µs PW, 500µs IPG, 50pps
4/8 (CF)7/8 (AF)
1/8 (AF) Not tested
Monopolar500µs PW, 500µs IPG, 500pps
Not tested 20/20 (AF) Not tested
Monopolar148µs PW, 20µs IPG, 400pps
Not tested 17/20 (AF) Not tested
Monopolar Ganged Pairs500µs PW, 500µs IPG, 500pps
Not tested Not tested 9/9 (AF)
Monopolar Ganged Pairs200µs PW, 200µs IPG, 200pps
Not tested Not tested 9/9 (AF)
Figure 2. Electrode yield for different electrodeconfigurations. The table (A) shows the number ofelectrodes that were capable of eliciting a visualpercept using various stimulation parameters andelectrode configurations. The active (red) andreturn (black) electrodes used for each electrodeconfiguration are shown in (B). For P1, reliableand consistent percepts were obtained using astimulation rate of 50pps. For P2, whilst perceptswere obtained using 50pps, they mainly consistedof only onset and offset flashes. Increasing thestimulation rate to 400pps or 500pps producedpersistent and reliable percepts. For P3, highstimulation rates and ganged pairs of electrodeswere required in order to obtain visual percepts.Narrow pulse widths were also tested with P2 andP3 to allow stimulation on multiple electrodes to beinterleaved in time at high rates to form patterns(see Figure 6).PW = phase width, IPG = interphase gap, CF = cathodicphase first, AF = anodic phase first. Stimulus duration was 2sin all cases.
Monopolar Ganged Pair
Common Ground
Hexagonal
Monopolar
B)
A)
Figure 3. Phosphene drawings comparing monopolar and common ground return configurations at 2dB abovethreshold for several electrodes. Drawings were obtained from P1 using a motion tracking sensor attached to thepatient’s index finger. Phosphenes produced using monopolar stimulation (M; purple) were very similar to thoseproduced with common ground stimulation (CG; green). Offsets in position can be attributed to minor changes in eyeand head position between stimuli.Stimulation parameters: P1 – 500ʅs phase width, 500ʅs interphase gap, 50pps, 2s, cathodic first stimulation.
Figure 6. Percepts as drawn by P2 for each patternof electrodes stimulated in an interleaved fashion.Drawings were obtained using a motion trackingsensor attached to the patient’s index finger. Aboveeach percept is the description given by the patient.Inset shows the electrodes (corresponding to visualspace) that were used for each pattern (Asterisk (*)denotes the position of E1 on the electrode array).
Figure 1. (A) The suprachoroidalelectrode array has 33 platinum discelectrodes (30 u 600µm; 3 u 400µm)and 2 large return electrodes (2mmdiameter). Note that the electrodeson the outer edges (top, bottom andright edge) of the implant are shortedtogether to form an extra ‘guard’return, thus giving 20 individualstimulating electrodes (B).
A)1 6 11 16
2 7 12 17
3 8 13 18
4 9 14 19
5 10 15 20
B)
OpEcal'Coherence'Tomography'
silicone(pla+num(electrode(
re+na(
choroid(
Psychophysics'
O&M'and'ADL'TesEng' Likely'Benefits'of'Bionic'Eyes'–'Short'Term''
• Light'percepEon'
• Object'detecEon'and'avoidance'
• Visual'certainty'
• Social'cues'
• Enjoyment'
16/01/2015'
4'
Possible'Benefits'of'Bionic'Eyes'–'Long'Term''
• Higher'resoluEon'
• Reading'large'print?'
• Face'recogniEon?'
Vision'RestoraEon'OpEons'
1. ReEnal'Prostheses'(“bionic'eyes”)'
2. Stem'cells'
3. Gene'therapy'
4. OptogeneEcs'
Vision'RestoraEon'OpEons'
1. ReEnal'Prostheses'(“bionic'eyes”)'
2. Stem'cells'
3. Gene'therapy'
4. OptogeneEcs'
Vision'RestoraEon'OpEons'
1. ReEnal'Prostheses'(“bionic'eyes”)'
2. Stem'cells'
3. Gene'therapy'
4. OptogeneEcs'
Likely'Benefits'–'Short'Term''
• Beneficial'to'people'with'very'specific'causes'of'vision'loss'(geneEc)'
• More'“natural”'restoraEon'of'vision'
• Promising'iniEal'trials'
Likely'Benefits'–'Long'Term''
• Likely'to'be'beneficial'to'people'with'earlier'stages'of'vision'loss'
• Unlikely'to'be'of'benefit'for'profound'vision'loss'
16/01/2015'
5'
NeuroplasEcity'
• William'James'–'1890'
• PlasEcity'='intrinsic'property'of'the'nervous'system'throughout'life'
• Evidence'that'the'vision'related'areas'of'the'occipital'cortex'are'reAwired'in'a'compensatory'cross'modal'manner'in'people'who'are'blind'
NeuroplasEcity'
NeuroplasEcity'
• In'experiments,'five'days'of'blindfolding'is'enough'to'show'changes'in'auditory'and'tacEle'mapping'in'the'visual'cortex'
• This'means'it'is'likely'that'the'corEcal'connecEons'were'already'present,'and'are'just'“unmasked”'
NeuroplasEcity'
• Likely'to'improve'outcomes'in'vision'restoraEon'intervenEons'(as'has'been'shown'in'cochlear'implants)'
• Possible'avenue'for'other'therapies'
• Important'area'of'research'
How'to'Keep'Updated'
• Contact'us'directly:'[email protected]''
• www.arEficialvision.org'
• Low'vision'support'agencies'
Acknowledgements'ARC'funding'and'Victorian'Government'infrastructure'support'
The'research'team:'
'
ReEna'Australia,'Blind'CiEzens'AssociaEon'Australia,'Guide'Dogs'Australia,'Vision'Australia'
'
Our'research'parEcipants'(and'their'families'and'friends)'