University of Southern California Dissertations and Theses

Stimulation strategies to improve efficiency and temporal resolution of epiretinal prostheses

(USC Thesis Other)

Stimulation strategies to improve efficiency and temporal resolution of epiretinal prostheses

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content STIMULATION STRATEGIES TO IMPROVE EFFICIENCY AND
TEMPORAL RESOLUTION OF EPIRETINAL PROSTHESES
by
Navya S. Davuluri
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BIOMEDICAL ENGINEERING)
December 2014
Copyright 2014 Navya S. Davuluri
Epigraph
Whatever the mind of man can conceive and believe, it can achieve
- Napoleon Hill
ii
Acknowledgments
I would not have completed my Ph.D. without the generous help and support of others.
I would like to express my gratitude to my adviser, Dr. James Weiland, for his patience,
motivation, and guidance. I could not have imagined having a better adviser and mentor.
His confidence in my ability to work on a difficult project gave me the strength and
conviction to plough ahead. I would also like to thank the rest of my committee, Drs.
Mark Humayun and Robert Chow, for the valuable advice they provided over the years.
I am very grateful to Aditi Ray and Leanne Chan, who developed the initial frame-
work for my research and spent countless hours training me to perform experiments.
Aditi and Leanne took the time to help me learn the techniques that were fundamental
to my study even after they graduated. I would also like to thank my labmates, past
and present, for their helpful discussions and assistance with my research. Specifically, I
owe thanks to Alice Cho, Andrew Weitz, Ashish Ahuja, Artin Petrossians, Devyani Nan-
duri, Alan Horsager, Vivek Pradeep, Neha Parikh, Tim Nayar, Samantha Cunningham,
Steven Walston, Boshuo Wang, Kiran Nimmagadda, Karthik Murali, Aminat Adebiyi,
Nii Mante, and Alejandra Gonzalez.
iii
I will always be grateful to my family. Their unwavering support and love helped
me get through the most difficult phases of my life. They have been a great source of
encouragement throughout my life.
iv
Table of Contents
Epigraph ii
Acknowledgments iii
List of Figures viii
List of Abbreviations xii
Abstract xiv
Chapter 1: Introduction 1
1.1 Electrical Stimulation of Neurons . . . . . . . . . . . . . . . . . . . . . . . 2
1.1.1 Spherical Cell Stimulation . . . . . . . . . . . . . . . . . . . . . . . 2
1.1.2 Strength Duration Curve . . . . . . . . . . . . . . . . . . . . . . . 3
1.1.3 Fiber Stimulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Gross Anatomy of the Eye . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3 Retina . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3.1 Cellular Organization . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.3.2 Retinal Pigment Epithelium . . . . . . . . . . . . . . . . . . . . . . 9
1.3.3 Photoreceptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.3.4 Horizontal Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.3.5 Bipolar Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.3.6 Amacrine Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.3.7 Ganglion Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.3.8 Glial Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.3.9 Communication within the Retina . . . . . . . . . . . . . . . . . . 16
1.4 Visual Pathway . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.4.1 Human Visual Pathway . . . . . . . . . . . . . . . . . . . . . . . . 17
1.4.2 Rat Retina and Visual Pathway . . . . . . . . . . . . . . . . . . . . 21
1.5 Retinitis Pigmentosa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.6 History of Visual Prostheses . . . . . . . . . . . . . . . . . . . . . . . . . . 28
1.6.1 Primary Visual Cortex Simulation . . . . . . . . . . . . . . . . . . 28
1.6.2 LGN Stimulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
1.6.3 Optic Nerve Stimulation . . . . . . . . . . . . . . . . . . . . . . . . 31
1.6.4 Subretinal Stimulation . . . . . . . . . . . . . . . . . . . . . . . . . 31
v
1.6.5 Epiretinal Stimulation . . . . . . . . . . . . . . . . . . . . . . . . . 32
1.7 Thesis Overview and Structure . . . . . . . . . . . . . . . . . . . . . . . . 32
Chapter 2: Relative Efficiency of Voltage Vs. Current Controlled Pulses
for Epiretinal Stimulation 35
2.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.2 Experimental Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.2.1 Experimental Groups . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.2.2 Electrical Stimulation . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.3.1 In Vivo Experiement Results . . . . . . . . . . . . . . . . . . . . . 39
2.3.2 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.3.3 Power Consumption Results . . . . . . . . . . . . . . . . . . . . . . 43
2.3.4 Comparison of Electrode Properties . . . . . . . . . . . . . . . . . 48
2.3.5 Comparison of Stimilus Artifacts . . . . . . . . . . . . . . . . . . . 48
2.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.4.1 In Vivo Experiments with Standard Pt-Ir . . . . . . . . . . . . . . 51
2.4.2 In Vivo Experiments with High Surface Area Pt-Ir . . . . . . . . . 53
2.4.3 Computational Studies . . . . . . . . . . . . . . . . . . . . . . . . . 53
2.4.4 Power Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
2.4.5 Stimulus Artifact . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Chapter 3: Strategies to Limit Retinal Desensitization Caused by Contin-
uous Electrical Stimulation 56
3.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.2 Experiment Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.3.1 Time-Varying Pulses Limit Desensitization in Normally Sighted Rats 62
3.3.2 Short Duration Pulses Limit Desensitization in Normally Sighted
Rats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.3.3 Time-Varying Pulses Limit Desensitization in Retinal Degenerate
Rats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.3.4 Short-Duration Pulses Limit Desensitization in Retinal Degenerate
Rats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Chapter 4: Optical Imaging of Intrinsic Signals 80
4.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.2 Sources of Intrinsic Signals and Wavelength Dependency . . . . . . . . . . 81
4.2.1 The Blood Volume Component . . . . . . . . . . . . . . . . . . . . 83
4.2.2 The Hemoglobin Oxymetry Component . . . . . . . . . . . . . . . 84
4.2.3 Light Scattering Component . . . . . . . . . . . . . . . . . . . . . 85
4.2.4 Resolving the Different Components of the Intrinsic Signals . . . . 86
4.2.5 Time Course of Intrinsic Signals . . . . . . . . . . . . . . . . . . . 90
4.3 Studies on Visual Cortex Using Intrinsic Signal Imaging . . . . . . . . . . 91
4.4 Intrinsic Signal Imaging and Superior Colliculus . . . . . . . . . . . . . . . 92
vi
4.4.1 Anesthesia and Monitoring Physiological Parameters . . . . . . . . 93
4.4.2 Immobilization of the Superior Colliculus . . . . . . . . . . . . . . 95
4.4.3 Slow-Scan CCD Cameras . . . . . . . . . . . . . . . . . . . . . . . 96
4.4.4 Camera Lenses/Macroscope and Camera Mount . . . . . . . . . . 98
4.4.5 Data Acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
4.4.6 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
4.4.7 Single Condition Maps . . . . . . . . . . . . . . . . . . . . . . . . . 103
Chapter 5: Conclusions and Future Work 104
5.1 Recommendations for Epiretinal Prostheses . . . . . . . . . . . . . . . . . 104
5.1.1 Limiting Desensitization Caused by Continuous Electrical Stimula-
tion of the Retina . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
5.1.2 Effect of Waveform Shape on Stimulation Efficiency . . . . . . . . 107
5.2 Future Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Chapter 6: Methods 110
6.1 Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
6.2 Stimulation Electrode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
6.3 Surgical Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
6.3.1 SC Exposure and Recording Electrode Positioning . . . . . . . . . 112
6.3.2 Stimulation Electrode Insertion . . . . . . . . . . . . . . . . . . . . 113
6.3.3 Positioning Stimulation Electrode via Impedance Sensing . . . . . 114
6.4 Electrical Stimulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
6.5 Data Acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
6.6 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
6.6.1 Measurement of Injected Charge and Quantification of Electrically
Evoked Responses (EERs) . . . . . . . . . . . . . . . . . . . . . . . 116
6.6.2 Signal Strength versus Injected Charge . . . . . . . . . . . . . . . . 117
6.6.3 Power Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . 118
6.7 Computational Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
6.8 Imaging Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
6.9 Imaging Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
References 121
vii
List of Figures
1.1 Intracellular stimulation of a spherical cell - Response to a current step . 2
1.2 Stimulation of a cell in uniform electric field . . . . . . . . . . . . . . . . . 4
1.3 Strength duration curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Cabel model of an axon . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.5 Gross anatomy of the eye . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.6 Light micrograph of a vertical section of the human retina . . . . . . . . . 9
1.7 Visual field of the human eye . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.8 Ganglion cells in the peripheral retina of a rat . . . . . . . . . . . . . . . . 20
1.9 Ganglion cell density in the rat retina . . . . . . . . . . . . . . . . . . . . 21
1.10 Dorsal and lateral views of the skull of a 290g Wistar rat . . . . . . . . . 22
1.11 Cell types identified in the superficial layers of SC . . . . . . . . . . . . . 24
1.12 Topographic arrangement of receptive fields in the left visual field and their
corresponding electrode positions on the contralateral right SC in the rat 25
1.13 Schematic representation of the three stages of retinal degeneration . . . . 27
1.14 Extracranial part of the visual cortex implant . . . . . . . . . . . . . . . . 30
1.15 ARGUS II schematic showing the microelectrode array, inductive coil,
video processing unit (VPU) and a miniature camera . . . . . . . . . . . . 33
2.1 Strength of EERs versus charge - standard Pt-Ir electrode . . . . . . . . . 40
viii
2.2 Strength of EERs versus charge - high surface area Pt-Ir electrode . . . . 41
2.3 Strength of EERs versus charge for RD retina - Standard Pt-Ir electrode . 42
2.4 Membrane dynamics of a retinal ganglion cell model . . . . . . . . . . . . 44
2.5 Simulation of a rat retinal ganglion cell model . . . . . . . . . . . . . . . . 44
2.6 Power consumption versus charge - standard Pt-Ir electrode . . . . . . . . 45
2.7 Power consumption versus charge - high surface Pt-Ir electrode . . . . . . 46
2.8 Power consumption versus charge - both standard and high surface Pt-Ir
electrode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.9 Power consumption versus charge - both standard and high surface Pt-Ir
electrode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.10 Comparison of measured current and voltage between standard Pt-Ir elec-
trode and high surface area Pt-Ir electrode during voltage-controlled and
current-controlled stimulation respectively . . . . . . . . . . . . . . . . . . 49
2.11 Comparison of stimulus artifact generated by the standard recording elec-
trode and modified recording electrode . . . . . . . . . . . . . . . . . . . . 50
2.12 Stimulus artifact strength vs. charge in the stimulus pulse . . . . . . . . . 50
3.1 Stimulation protocol used in control and experimental groups . . . . . . . 59
3.2 Strength-duration curve used to determine the current amplitude for time-
varying pulses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.3 Average normalized EER strength is plotted against time for both control
and time-varying pulse experiments . . . . . . . . . . . . . . . . . . . . . . 63
3.4 Early phase of the average normalized EER strength plotted against time
for both control and time-varying pulse experiments . . . . . . . . . . . . 64
3.5 Late phase of the average normalized EER strength is plotted against time
for both control and time-varying pulse experiments . . . . . . . . . . . . 65
3.6 Average normalized EER strength is plotted against time for both control
and short-duration pulse experiments . . . . . . . . . . . . . . . . . . . . . 66
ix
3.7 Early phase of the average normalized EER strength plotted against time
for both control and short-duration pulse experiments . . . . . . . . . . . 67
3.8 Late phase of the average normalized EER strength is plotted against time
for both control and short-duration pulse experiments . . . . . . . . . . . 68
3.9 S334ter line 3 homozygous rat retina . . . . . . . . . . . . . . . . . . . . . 69
3.10 Average normalized EER strength is plotted against time for both control
and time-varying experiments in RD animals . . . . . . . . . . . . . . . . 69
3.11 Early phase of the average normalized EER strength plotted against time
for both control and time-varying pulse experiments in RD rats. The
shaded regions represent standard error . . . . . . . . . . . . . . . . . . . 70
3.12 Late phase of the average normalized EER strength plotted against time
for both control and time-varying pulse experiments in RD rats . . . . . . 71
3.13 Average normalized EER strength is plotted against time for both control
and short duration pulse experiments in RD animals . . . . . . . . . . . . 72
3.14 Early phase of the average normalized EER strength plotted against time
for both control and short duration pulse experiments in RD rats . . . . . 73
3.15 Late phase of the average normalized EER strength plotted against time
for both control and short duration pulse experiments in RD rats . . . . . 74
4.1 The biphasic timecourse of optical responses . . . . . . . . . . . . . . . . . 85
4.2 Hemoglobin absorption curves . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.3 The “cortical band” and Optical Spectroscopy . . . . . . . . . . . . . . . 89
4.4 Spectral content of light reflectance from a single pixel . . . . . . . . . . . 96
4.5 Spectral content of light reflectance from a single pixel illustrating the
absence from low frequency physiological noise . . . . . . . . . . . . . . . 97
4.6 Images of SC captured using a macroscope . . . . . . . . . . . . . . . . . 98
4.7 SC image captured using a zoom lens from Navitar . . . . . . . . . . . . . 99
4.8 Schematic diagram of the optical imaging of intrinsic signals setup . . . . 100
4.9 Protocol for recording a focal SC response to electrical stimulation with ISI 102
x
4.10 Protocol for recording a focal SC response to electrical stimulation with ISI 102
6.1 An evoked potential recorded from the SC when the retina is electrically
stimulated . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
6.2 An example of the rectangular current-controlled and voltage-controlled
pulses applied to the electrode-retina interface . . . . . . . . . . . . . . . . 116
xi
List of Abbreviations
AMD Age-related Macular Degeneration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
cAMP cyclic Adenosine Mono Phosphate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
CCD Charge-coupled Device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
cGMP cyclic Guanosine Mono Phosphate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17
CNS Central Nervous System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22
EERs electrically evoked responses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35
ERG Electroretinogram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
FDA US Food and Drug Administration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
fMRI Functional Magnetic Resonance Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
GABA Aminobutyric Acid. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17
GCL Ganglion Cell Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
INL Inner Nuclear Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
IPL Inner Plexiform Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
ISI Intrinsic Signal Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
LGN Later Geniculate Nucleus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15
LGND Dorsal Lateral Geniculate Nucleus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22
LGNV Ventral Lateral Geniculate nucleus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
MEAs Multi-electrode Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
NIRS Near-infrared Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
ONL Outer Nuclear Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
PET Positron Emission Tomography. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .79
PPD Paired-pulse depression. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .76
xii
Pt-Ir Platinum-Iridium. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37
RD Retinal Degenerate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .108
RP Retinitis Pigmentosa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
RPE Retinal Pigment Epithelium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
SAP Stratum Album Profundum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
SC Superior Colliculus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
SGI Stratum Griseum Intermediale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
SGP Stratum Griseum Profundum. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23
SGS Stratum Griseum Superficiale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
SO Stratum Opticum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
SSMP Second Sight Medical Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
SZ Stratum Zonale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
VPU Video Processing Unit. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31
xiii
Abstract
Retinitis Pigmentosa (RP) is a degenerative disorder of the retina that begins with night
blindness, leads to tunnel vision and eventually results in complete loss of vision. Physio-
logically, RP results in the complete loss of photoreceptors and rewiring of the inner retinal
neurons. However, electrical stimulation of degenerate retina in RP subjects have shown
that electrical stimulation causes visual perception. Ever since electrical stimulation has
been shown to elicit light perception, neuroprosthetic devices called epiretinal prostheses
are being developed in order to restore some form of functional vision in patients suffering
from RP.
ARGUS II is one such epiretinal implant that has been developed by Second Sight
Medical Products in collaboration with USC. Recent studies with ARGUS II implant
subjects have shown that it does restore some form of functional vision back in RP sub-
jects. However, research also indicated that implant subjects face certain issues when
tested in a clinical setting. The first issue faced by ARGUS II subjects’ deals with
non-functional electrodes. ARGUS II has 60 electrodes and about 45% of them were
non-functional because the charge required to elicit visual percepts using these electrodes
xiv
exceeded the charge density limit for these electrodes. Thus improving stimulation ef-
ficiency may increase the spatial resolution of the implant by increasing the number of
functional electrodes.
In order to address the first issue, I studied the effect of stimulation waveform shape
and stimulus pulse duration on stimulation efficiency. Specifically, I compared the effect
of voltage-controlled and current-controlled pulses of various durations on stimulation
efficiency. I found that when the pulse duration was less than 1 ms, voltage-controlled
pulses were more efficient at stimulating the retina. When the pulse duration was equal
to or greater than 1 ms, current-controlled pulses were more efficient at stimulating the
retina. I also developed a computational model that could be used to test the effect of
pulse duration and shape on stimulation efficiency. In addition to studying the effect of
pulse shape and duration on stimulation efficiency, I tested a novel electrode material
and compared its effect on power consumption. The novel electrode material was a
high surface area Pt-Ir electrode. When the high surface area Pt-Ir electrode was used
for stimulation, it consumed less power to generate stimulus pulses when compared to
the Pt-Ir gray stimulation electrode used in the ARGUS II implant. Reducing power
consumption improves the implant design and prevents running into voltage compliance
issues.
The second issue faced by ARGUS II implant subjects’ was fading of visual percepts
over time even when electrical stimulation was turned on. This reduces the temporal
resolution of the implant. A decrease in the excitability of retina could be one of the
reasons why fading occurs in implant subjects. Many studies have shown that continuous
electrical stimulation causes the retina to desensitize. In our lab, we created an in vivo
xv
model of retinal desensitization and studies strategies to limit desensitization. Specifically,
I studied the effect of pulse trains called time-varying pulses. Time-varying pulses are
defined as a train of pulses where each pulse has a different pulse duration and amplitude
when compared to the preceding pulse. The charge delivered for every pulse is three
times the threshold charge for that pulse width. In addition to time-varying pulses, the
effect of short duration pulses on retinal desensitization was also studied. Studies were
performed in both normally sighted and retinal degenerate rats.
I showed that time-varying pulses limit retinal desensitization in both normally sighted
and retinal degenerate animals. Short duration pulses also limit desensitization in nor-
mally sighted rats but not in retinal degenerate rats. Short duration pulses directly stim-
ulate ganglion cells and thus avoid axonal stimulation. However, short duration pulses
require higher current amplitudes and selective activation of ganglion cells is restricted
to a limited range of stimulus amplitudes. Thus, time-varying pulses might improve the
temporal resolution of the implant.
In addition to addressing the aforementioned issues, I also developed an optical imag-
ing technique called intrinsic signal imaging in order to study maps of functional activity
in superior colliculus (SC), generated by electrical stimulation of the retina. Tradition-
ally, electrophysiological techniques have been used to study electrical stimulation of the
retina. However, such techniques provide limited information about the spatial pattern
of retinal excitation. Intrinsic signal imaging overcomes this limitation as it enables high
spatial resolution imaging of evoked potential activity in SC.
xvi
Chapter 1
Introduction
Vision loss is recognized as one of the most debilitating forms of disability. There are
an estimated 1 million people in US that are legally blind. The number of persons
with low vision is approximated to be an additional 2.5 million (Congdon et al., 2011).
Diseases of aging are the most common cause of vision loss (cataract, glaucoma, AMD)
and low vision. A number of theraupetic approaches including gene therapy, optogenetic
approaches, retinal prosthesis etc. are being investigated in order to address retinal
degeneration that results from these pathologies.
In this chapter, I will provide an overview of the anatomy of the eye, morphology of the
retina, retinal degeneration caused by a genetic disorder retinitis pigmentosa (RP) and
several prostheses that have been developed to restore functional vision in patients with
the retinal degenerative disorders. I will also discuss a few concepts that are important in
the development of retinal prosthesis including electrical stimulation of neurons, human
visual pathway and the visual pathway of animal models used in our experiments.
1
1.1 Electrical Stimulation of Neurons
Neurons are electrically excitable cells that process and transmit information through
chemical and electrical signals. A neuronal membrane consists of metabolically driven
ion pumps and channels that help maintain the voltage gradient across the membrane.
Changes in the transmembrane voltage caused by electrical or chemical stimuli generate
an all-or-none electrochemical pulse called an action potential that travels rapidly along
the axon. The following sections describe mathematical relationships between applied
electrical stimuli and the resulting changes in transmembrane voltage:
1.1.1 Spherical Cell Stimulation
figure 1.1 shows a current step and the response of a spherical cell to the step when the
cell is stimulated intracellularly. The equivalent electrical circuit is also shown.
Figure 1.1: A: Intracellular stimulation of spherical cell. The stimulus current step and the
corresponding trans-membrane voltage are shown as well. Right: The equivalent electrical circuit
for intracellular stimulation of spherical cell.
A current step of magnitude
∘ is applied by the stimulator between the intracellular
and extracellular electrodes. The stimulus current continues indefinitely during time.
2
The current produces a rising trans-membrane voltage,

(solid curve),that does not
have the step waveform of
∘ . Even though the stimulus current
∘ continues on, the rise
of

approaches limiting level

=. Level is called the “strength” of the stimulus.
Of particular interest is the time required to reach a “threshold” voltage level

=
(short lines crossing vm curve at lower right). The

curve is sketched as the response
if membrane resistance is constant.
The trans-membrane potential developed in the cell is given the following equation:

=
∘ (1−
− /
)
ℎ =
For a cell in the presence of a uniform electric field, the transmembrane voltage is
given by the following equation where and are defined as in 1.2:

=
3
2
.
1.1.2 Strength Duration Curve
The relationship of the pulse width to threshold current is given by the strength-duration
equation also called Weiss-Lapicque equation:

ℎ
=

(1−
− /
)
3
Figure 1.2: Stimulation of a cell in uniform electric field. The cell is hyperpolarized on the left
side and depolarized on the right side (Horsch and Dhillon, 2004).

is named Rheobase,
ℎ
is the minimum current required to reach threshold with
stimulation duration . is the time constant. There are two parameters that character-
ize the response of the neuron: rheobase and chronaxie. Rheobase is the minimum current
required to stimulate a neuron with an infinitely long pulse. Chronaxie is the minimum
pulse width required to stimulate a neuron with twice the rheobase current. The strength-
duration curve (figure 1.3) is a plot of the threshold current and the corresponding pulse
duration.
1.1.3 Fiber Stimulation
In case of axonal stimulation, the axon is broken into several components and modeled
as shown in figure 1.4.
4
Figure 1.3: Strength-Duration Curve. The curve shows the combinations of stimulus strength
and stimulus duration that are just sufficient to reach the threshold level. Combinations on side
A of the curve are above threshold and may lead to action potentials, while combinations on side
B are below threshold.
Using Kirchoff’s laws, we get the equation relating trans-membrane voltage and the
current flowing through the axons is:
2

2

2
−

−

=

2

is the transmembrane voltage. is the length constant and is the time constant.
These constants are given by the following equations:
5
Figure 1.4: Cabel model of an axon. Φ

and Φ

are the extracellular and intracellular potentials.

and

are the extracellular and intracellular membrane resistances

and

are the currents
flowing in and out of the membrane.
=

+

=

For steady-state conditions (

= 0) and when there is no stimulus current, the cable
equation reduces to the below. Thus the driving force for membrane perturbation is the
second derivative of transmembrane voltage.
2

2

2
−

= 0
6
1.2 Gross Anatomy of the Eye
Common structural features of all vertebrate eyes include an aperture, the pupil, that
allows light to enter the eye, a circular muscle that controls the size of the pupil termed
the iris, a lens that helps in focusing the light rays, and three different layers including
cornea, sclera and choroid. Light rays enter the eye through the cornea and are focused
to form a real inverted image on a thin sheet called the retina at the back of the eye
(figure 1.5). Sclera is a supportive structure of the eye and the choroid has the blood
vessels that supply essential nutrients for the retina.
1.3 Retina
Retina is a 500um thick tissue lining the back of the eye wall. In the embryonic stage, it
originates from the neural tube and hence is part of the central nervous system. The retina
is functionally organized into two distinct layers: plexiform layers and nuclear layers. The
plexiform layers are intercalated between nuclear layers. There are 2 plexiform layers,
inner and outer, and 3 nuclear layers (inner, outer and ganglion cell layers). Plexiform
layers are where the functional contacts, synapses, are made between neurons. The cell
bodies of these neurons are distributed in the 3 nuclear layers.
There are 5 distinct classes of neurons in the retina (figure 1.6): photoreceptors,
horizontal cells, bipolar cells, amacrine cells and ganglion cells. In addition to the neuronal
cell types, retina also has a few classes of glial cells. The most abundant glial cell type
found in the retina is a Muller cell. Microglia and astrocytes are also present in the retina.
7
Figure 1.5: Saggital horizontal section of the adult human eye. Image from www.webvision.com
1.3.1 Cellular Organization
Photoreceptors are the sensory neurons of the retina. The cell bodies or perikarya of these
neurons are located in the outer nuclear layer (ONL). The inner nuclear layer (INL) is
comprised of cell bodies of horizontal cells, bipolar cells and amacrine cells. The ganglion
cell layer (GCL) consist cell bodies of ganglion cells. Sandwiched between ONL and INL is
the outer plexiform layer. It is comprised of synaptic connections between photoreceptors
and the cells present in INL. The inner plexiform layer (IPL) is intercalated between INL
8
Figure 1.6: Light micrograph of a vertical section of the human retina. Image from www.
webvision.com
and GCL and consists of synaptic connections between cells in INL and ganglion cells
(figure 1.6)
1.3.2 Retinal Pigment Epithelium
In addition to the neurons in the retina, there exists another monolayer of epithelial cells
called the retinal pigment epithelium (RPE). In the vertebrate retina, RPE lies between
the photoreceptor layer and the blood supply of the choroid. The epithelial cells contain
9
a light absorbing pigment and hence the most obvious function of the RPE is to absorb
scattered light in the retina.
1.3.3 Photoreceptors
Photoreceptors are the sensory neurons of the retina that respond to light. Light passes
through the entire thickness of the retina before it reaches the photoreceptors as they
lie at the back of the retina. Photoreceptors contain pigment molecules made out of
opsins that absorb the photons and change configuration that initiates the process that
eventually leads to perception. The pigment bearing molecules, opsins, contain retinal or
vitamin A. In response to photons, retinal changes conformation and in order to regain its
original structure it is recycled in the retinal pigment epithelium layer that coats the back
of the eye. The pigment epithelium layer is a dark layer containing melanin molecules
that absorb any excess photons and prevent them from re-entering the photoreceptor
layer. Most vertebrate retinas have at least two types of receptor cells: rods and cones.
Rods are responsible for dim-light vision and cones for color vision. The shape of their
outer segments can distinguish these two kinds of cells: rods have slender, elongated outer
segments, whereas cones have shorter, wider and tapered outer segments. Rods account
for 95% of the receptor cells in mammalian retina and cones account for the remaining 5%.
Rods and cones are not uniformly distributed in mammalian retinas. In human retina, the
highest density of cones is found in a region called fovea. Fovea is also the region with the
maximal visual acuity. Cells in the foveal region are smaller in size when compared to the
peripheral retina and the density of cells in the fovea is 100 times that in the periphery.
Human retinas have a single type of rods and 3 types of cones. The three types of cones
10
have different visual pigments that allow maximum absorption in red, blue and green
regions of the spectrum. In the absence of light photoreceptors are relatively depolarized
and in response to light they turn off: the Na+ channels are shut closed that results in the
membrane potential to hyperpolarize and inhibit neurotransmitter release. In addition to
morphological differences, rods and cones have different physiological responses to light.
Rods respond to very dim light levels and are found to be much more sensitive than cones.
Cones respond to brighter light levels and enable high acuity color vision. Rods respond
to slower changes whereas cones are sensitive to rapid fluctuations in light levels.
1.3.4 Horizontal Cells
Horizontal cells are post-synaptic to photoreceptors and receive synaptic input from mul-
tiple photoreceptors. In most vertebrate retina, two morphologically distinct types of
horizontal cells exist: axonless cells referred to as A type cells and cells with axons called
B type cells. In mammals, A-type cells only have synaptic connections with cones. H2
is the most well studied A type neuron. This is a trichromatic cell: it responds to three
different wavelengths of light. B type neurons have bushy dendrites that contact several
cones as well. The axon emerges from one of these dendrites and the terminal contacts
several rods. In these horizontal cells, the cone-involved compartment of the neuron is
distinctly segregated from the rod-involved compartment. Based on the physiology of
horizontal cells, they are segregated in to two types: L type and C type. Both cell types
have a resting membrane potential that is close to that of the photoreceptors and they
are depolarized in the dark. L type or luminosity type horizontal cells respond to light
11
with graded hyperpolarizing membrane potentials also termed S-potentials. C type chro-
maticity type horizontal cells have different responses to light based on the wavelength of
the stimulus. They hyperpolarize to light of certain color and depolarize in response to
other wavelengths. In mammals, only L type horizontal cells are present. Another impor-
tant feature of horizontal cells is the presence of electrically connected channels known as
gap junctions. These junctions have high specificity: they only connect cells of the same
type. Gap junctions are formed either between dendrites of neighboring neurons or be-
tween axons terminals. Gap junctions allow the transfer of ions and small molecules. Gap
junctions enhance the receptive field sizes of horizontal cells. The major functional roles
of horizontal cells are modulation of photoreceptor responses to various light conditions
and initiation of the receptive field organization for the inner retinal neurons. Example:
their feed forward connections to bipolar cells aid in defining the surround opponency for
bipolar cell receptive field.
1.3.5 Bipolar Cells
Bipolar cells are categorized into three distinct types based on the synaptic connections
they make with receptor cells. Two of those types contact cones and the other type
contacts rods only. The two-cone related types of bipolar cells make different synaptic
connections with the cones and their axon terminals end in different regions of the inner
plexiform layer. One type of the cone-related bipolar cell has its dendrites make invagi-
nating contacts with the cone terminals. Its axon terminals end in the ON-region of the
inner plexiform layer. This type is called the invaginating bipolar cell. The second type
of cone-related bipolar cell contacts the flattened base of the cone receptor cell and its
12
axon terminals end in the OFF region of the inner plexiform layer. This type is called the
flat bipolar cell. The rod-related bipolar cell extends its dendrites into the invaginations
of the rod terminals and the axon terminals end deep in the inner plexiform layer. The
dendritric arbor of the rod bipolar cells is larger than that of the cone-bipolor cell and one
rod-related bipolar cell contacts several rods. In primate retina, there is a special class
of cone-related bipolar cells called midget bipolar cells that contact only one cone. These
cells are found in the macular region of the retina and are thought to be responsible for
high acuity vision of the central retina. Based on physiology, bipolar cells are divided into
2 categories: those that hyperpolarize in response to light at the center of their receptive
field and those that depolarize in response to center illumination. The former cell type is
called hyperpolarizing or ON-center bipolar cells and the later are called depolarizing or
OFF-center bipolar cells. The receptive field of a bipolar cell has an antagonistic center-
surround organization. In such a receptive field organization, the surround response is
opposed to the center response. Bipolar cells, like receptors and horizontal cells, respond
to light with graded potentials and lack impulse activity.
1.3.6 Amacrine Cells
Amacrine cells are the interneurons of the retina that modulate the visual message pre-
sented to ganglion cells by bipolar cells. Most amacrine cells are axonless although some
exceptions have been notified. Morphologically, 2 varieties of amacrine cells exist: diffuse
amacrine cells and stratified amacrine cells. Diffuse amacrine cells are further divided into
multiple subtypes based on the diameter of their dendritic arbors. Stratified amacrine
cells are so defined based on the extent their dendritic processes penetrate into the inner
13
plexiform layer. IPL in general is divided into 5 strata or sublayers depending on where
the terminals of several cell types end in the layer. Stratified amacrine cells extend their
terminals into either one or multiple sublayers of the IPL. Physiologically, amacrine cells
are divided into 2 types: those that respond to light with sustained membrane depolar-
ization or transient depolarization. In primates, 22 subtypes of amacrine cells have been
identified. The most well-studied type is the AII amacrine cell. This is a narrow field cell
with bistratified morphology. AII amacrine cell predominantly receives input from rod
bipolar cells. It responds to light ON with a transient depolarization and light OFF with
a small hyperpolarization of the membrane.
1.3.7 Ganglion Cells
Ganglion cells are the output neurons of the retina. They collect the electrical message
concerning the visual image from the layers of cells preceding it and relay the information
to higher visual centers in the form of spike trains. The optic nerve bundles all the axons
of the ganglion cells. It contains more than a million nerves that relay information to the
higher visual centers for further processing. These cells have relatively larger diameter
than the other retinal cells. Ganglion cells, like amacrine cells, can be divided into 2 classes
based on morphology: those with diffuse dendritic trees that spread throughout the inner
plexiform layer and those with stratified dendritic trees that spread on one or few levels
of the inner plexiform layer. More than 20 types of ganglion cells have been identified in
vertebrate retinas. They receive excitatory input from bipolar cells and inhibitory input
from amacrine cells. In primates, ganglion cells in the fovea are divided into P type cells
and M type cells. P cells are mostly considered to be midget ganglion cells that project
14
to the Parvocellular layers of the later geniculate nucleus (LGN). M cells project on the
Magnocellular layers of the LGN. These cells receive input from midget bipolar cells and
are responsible for high acuity vision in primates. More recently, a rare sub-population
of ganglion cells have been identified that contains a photopigment called melanopsin.
These cells are photo-sensitive and have far less spatial resolution. They signal changes
in ambient light levels to the brain throughout the day. They are thought of signaling light
for unconscious visual reflexes, such as pupillary constriction, and regulating a number of
daily behavioral and physiological rhythms, collectively called circadian rhythms. Unlike
other retinal neurons that respond in a graded fashion, ganglion cells respond via action
potentials. Based on physiology, ganglion cells like bipolar cells can be divided into ON
and OFF types. ON type ganglion cells respond with a transient burst to light onset. OFF
type ganglion cells respond to light off with a sustained burst of impulses. In addition,
there is another subclass, the ON/OFF ganglion cell that responds with discharge bursts
both at the onset of light and when light is turned off.
1.3.8 Glial Cells
There are three basic types of glial cells found in the human retina: Muller cells, astroglia
and microglia. Muller cells are the principal glial cells in the retina and they form the
architectural support structures of the retina. They are also the limits of the retina at
the outer and inner limiting membrane. Their cell bodies are located in the inner nuclear
layer and project their processes in either direction to the outer limiting membrane and
to the inner limiting membrane. Muller cells have a range of functions that are vital
to the health of retina. They are involved in fueling metabolism in nerve cells, cleaning
15
up neural waste products, protecting neurons from exposure to excess neurotransmitters,
phagocytosis of neuronal debris, controlling homeostatis and in synthesis of retinoic acid
from retinol. Astrocytes are found in the nerve fiber layer of the retina. Their processes
are aligned along the ganglion cell axons. They are not present in the avascular fovea or
ora serrata. Processes of an astrocyte also cover the blood vessels running in and among
the ganglion cells thus making them part of the blood-brain barrier. Similar to Muller
cells, they contain abundant glycogen and they are thought of providing nutritive support
to neurons. In addition, they play a role in maintaining ionic homeostasis in regulating
extracellular potassium levels and metabolism of neurotransmitters like GABA. Microglial
cells are found in every layer in the retina. There appears to be two varieties of microglial
cells: one that enter the retina at early stages of development from the optic nerve and
the other that appear to be blood-borne cells. Both types of microglia are stimulated into
a macrophagic function after trauma to the retina. They also engage in phagocytosis of
the degenerating retinal neurons.
1.3.9 Communication within the Retina
Communication between different neurons in the retina takes place in two different ways:
via the release and uptake of neurotransmitters and via gap junctions. Neurotransmit-
ters are released at the synapse by the presynaptic neuron that then diffuse across the
synaptic cleft to bind a variety of post-synaptic receptors. Glutamate is the neurotrans-
mitter released by neurons in the vertical pathway (photoreceptor -> bipolar cell ->
ganglion cell). Glutamate is an excitatory neurotransmitter and binds to two types of
16
receptors, metabotropic and ionotropic, in the postsynaptic neurons. The classical in-
hibitory neurotransmitter in the retina is gamma aminobutyric acid (GABA). It occurs
in different varieties of amacrine cells and one or two varieties of horizontal cells. In addi-
tion to glutamate and GABA, several other neurotransmitters such as dopamine, glycine,
acetylcholine etc exist in the retina. In addition to chemical neurotransmission, commu-
nication within the retina occurs via gap junctions. Gap junctions are a type of electrical
synapse where the cell membrane of closely positioned neurons is connected. The gap
junction contacts are composed of hemichannels called connexons. These channels can
consist of homomeric or heteromeric connexin subunits, thus, potentially allowing a huge
variety of gap junctions with slightly different properties to exist between various neu-
rons. Gap junctions selectively permeable to different molecules and are modulated by
several molecules such as cAMP and cGMP.
1.4 Visual Pathway
1.4.1 Human Visual Pathway
Visual field is defined as the space that is seen by both eyes without head movement.
Figure 1.7 explains the visual field and how each half of the visual field is projected onto
the retina. The visual field is divided into left and right hemifields based on what each
eye sees. Light entering the eye from each hemifield creates a monocular zone in each
eye and light originating from the central visual field creates a binocular zone in both
eyes. Just like the visual field is divided into two halves, the retinal surface is also divided
into nasal and temporal hemiretina relative to fovea. Each hemifield is projected onto
17
Figure 1.7: Visual field of the human eye. Image taken fromhttp://brain.phgy.queensu.ca/
pare/assets/Central%20Pathways%20handout.pdf
the nasal hemiretina of the ipsilateral eye and temporal hemiretina of the contralateral
eye. The ganglion cell axons bundle to form the optic nerve and extend through the
optic disc. The two optic nerves cross at the optic chiasm and form two optic tracts, so
that the right and left hemifields reach the left and right hemispheres. Each optic tract
looks at the opposite hemifield, combining inputs from the ipsilateral temporal hemiretina
and contralateral nasal hemiretina. The optic tracts project to three major subcortical
regions: the pretectum, the superior colliculus, and the lateral geniculate nucleus. Retinal
18
ganglion cells project directly to the superficial layers of the superior colliculus (SC). Cells
from the superficial layers of SC in turn project to cerebral cortex via pulvinar nucleus of
the thalamus. The superficial layers of SC also receive feedback from the visual cortex.
The deeper layers of the SC respond to auditory and somatosensory stimuli in addition
to visual stimuli. Cells in the deep layers of SC also respond vigorously to the onset of
saccadic eye movements. Retinal ganglion cells that project to the pretectal areas of the
midbrain mediate pupillary light reflexes. The absence of pupillary reflexes is a symptom
of midbrain damage, the region from which the oculomotor nerves originate.
In the primate visual system, about 90% of the retinal ganglion cell axons terminate in
the lateral geniculate nucleus (LGN) of the thalamus. Retinal ganglion cells project onto
the LGN in three distinct pathways: magnocellular pathways, parvocellular pathways
and koniocellular pathways. Midget ganglion cells are considered to be the origin of
the parvocellular pathway and constitute approximately 70% of the total population of
cells that project to the LGN. Parasol ganglion cells are the origin of the magnocellular
pathway and bistratified ganglion cell make up most of the koniocellular pathway (Nassi
and Callaway, 2009). All areas of the retina are not equally represented in the nucleus.
The fovea, the area of the retina with the highest density of ganglion cells, has a relatively
large representation than does the periphery of the retina. The ratio of the area in the
lateral geniculate nucleus or primary visual cortex to the area in the retina representing
one degree of the visual field is called the magnification factor. The retinal ganglion cells
in and near the centrally located fovea are densely packed to compensate for the fact
that the retina’s central area is less than its periphery. Since this physical limitation
does not exist in the LGN and primary visual cortex, the neurons in these regions are
19
Figure 1.8: Ganglion cells in the peripheral retina of a rat. Small, medium and large ganglion
cells are labeled as S, M and L respectively. A glial cell labeled g is shown as well (Fukuda, 1977).
evenly distributed. Thus, connections from the more numerous neurons in the fovea
are distributed over a wide area in the LGN and primary visual cortex. Hence the
magnification factor varies between the foveal and peripheral regions of the retina. The
magnification factor is approximately equal to 4 mm/degree at 2 degrees eccentricity and
declines monotonically to 0.5 mm/degrees at 25 degrees eccentricity (Cowey and Rolls,
1974).
The magnocellular and parvocellular pathways leaving the LGN project to separate
layers of the primary visual cortex. Primary visual cortex is the first point in the visual
pathway where the receptive fields of cells are significantly different from the retinal
neurons. Like the LGN and SC, the primary visual cortex in each cerebral hemisphere
receives information exclusively from the contralateral half of the visual field.
20
Figure 1.9: Ganglion cell density in the rat retina. The lines are called iso-density lines and
the numerals represent number of cells per sq.mm. The approximate center of the central area is
marked with ‘+’. The optic disc (OD), temporal (T), nasal (N), Upper (U), lower (L) directions
are indicated (Fukuda, 1977).
1.4.2 Rat Retina and Visual Pathway
Visible light passes through the cornea and reaches the retina of rats in a way similar
to humans. Rats also have two kinds of photoreceptors: rods and cones. As in humans,
majority of the photoreceptors in rats are rods. In humans, the ratio or rods to cones is
20:1 (Curcio et al., 1990). However, in mice the ration is 100:1 (Rohrer and Crouch,
2006). Humans have trichromatic vision. However, rats have dichromatic vision as they
lack red cones. There are approximately 30 million photoreceptor cells and 100,000 –
120,000 ganglion cells in the rat retina. This higher fiber convergence results in an
increased sensitivity in low light at the expense of visual acuity.
As shown in figure 1.8, ganglion cells in rats can be classified into three groups based
on soma and dendritic tree sizes: small, medium and large: Large (large soma, large
21
Figure 1.10: A-Dorsal and lateral views of the skull of a 290g Wistar rat; B-Drawing of a section
of rat brain cut in a sagittal plane 0.4mm lateral to the midline (Paxinos and Watson, 1998).
dendritic tree), Small (small soma, small dendritic tree) and Medium (small to medium
somata, medium to large dendritic tree) (Huxlin and Goodchild, 1998). Cells within each
group are further divided into sub-classes based on the differences in their dendritic tree
pattern and stratification
The ganglion cell density varies across the retina in rodents . In general, the density
is higher in the upper retina than in the lower one and in the temporal retina than in the
nasal retina. Figure 1.9 shows the density map in a retina. The area with the highest
density is termed ‘central area’. This region is situated in the upper temporal quadrant
about one-third of the way from the optic disc to the retina edge. The figure also shows
isodensity lines where the density of RGCs is the same along the line. The isodensity
lines tend to be more spaced towards the lower nasal quadrant (Fukuda, 1977).
Retinal ganglion cell axons terminate in four midbrain structures in the rat CNS: the
dorsal lateral geniculate nucleus (LGND), the ventral lateral geniculate nucleus (LGNV)
and the Superior Colliculus (SC) and the pretectum (Fukuda and Iwama, 1978; Hale
22
and Sefton, 1978). Functionally, the VLG has been implicated in brightness discrimina-
tion while different regions of the pretectum are involved in optokinetic nystagmus and
pupilloconstrictor light reflex.
Majority of the retinal ganglion cell axons terminate in SC and thus is involved in
processing visual information (Lund, 1969; Lund and Langer, 1974; Tokunaga and Otani,
1976). SC is located 3.5mm rostral and 0.5-1.5mm lateral relative to the interaurel and
midline planes respectively (figure 1.10). The mammalian SC consists of seven distinct
layers: the zonal layer or stratum zonale (SZ), the superficial grey layer or stratum griseum
superficiale (SGS), the optic layer or stratum opticum (SO), the intermediate grey layer
or stratum griseum intermediale (SGI), the intermediate white layer or stratum griseum
profundum (SGP), and the deep white layer or stratum album profundum (SAP). These
layers are grouped into two functional and connectional different units, the superficial (SZ,
SGS, SO) and deep compartments (SGI, SAI, SGP, SAP), on the basis of anatomical and
behavioral data (Sparks and Young, 1989). Figure 1.11 lists the various types of cells
found in the superficial layers of the superior colliculus. Six different types of neurons
were identified in the superficial SC: horizontal, narrow field vertical, wide-field vertical,
piriform, marginal, and stellate cells (Edwards et al., 2002). The presynaptic input to
these cells comes from retinal ganglion cells of the contralateral eye and these cells in turn
project to deeper layers of the SC. Due to the direct connection between the superficial
layers of SC and retinal ganglion cells, this region of SC is of wide interest.
The superficial layer of SC has been shown to have a precise retinotopic map that is
consistent across rat strains with the densest retinal projections found in the SGS layer.
Figure 1.12 shows a topographic arrangement of receptive fields in the left visual field and
23
Figure 1.11: Various cell types identified in the superficial layers of SC (Edwards et al., 2002).
their corresponding electrode positions on the contralateral right SC in the rat (Siminoff
et al., 1966). The superficial layers of SC are exclusively visual in terms of function
and connectivity. In contrast, the deeper layers of SC have anatomical connections with
multiple sensory and motor systems. By the virtue of the physiological characteristics of
its neuronal population, it is uniquely suited for a central role in sensorimotor integration
(Sparks and Young, 1989).
1.5 Retinitis Pigmentosa
Retinitis Pigmentosa (RP) is a family of inherited retinal degenerations involving more
than 180 known genes, with many mutations in the rhodopsin gene (Daiger et al., 2007).
The incidence of RP in US is about 1 in 4000 people (Jones and Marc, 2005). Typical
24
Figure 1.12: Topographic arrangement of receptive fields in the left visual field and their corre-
sponding electrode positions on the contralateral right SC in the rat (Siminoff et al., 1966).
symptoms of RP include night blindness followed by decreasing visual fields, leading to
tunnel vision and eventually total loss of vision in many cases. The clinical hallmarks
of RP are abnormal fundus with bone-spicule deposits, reduced visual fields, abnormal
electroretinogram (ERG) recordings and abnormal blood vessels (Daiger et al., 2007).
Retinitis Pigmentosa results in extensive remodeling of the retina that is broadly classified
into three distinct phases: photoreceptor stress, photoreceptor loss and complex neural
remodeling of the inner retina. The three phases are briefly discussed below:
Phase I: Photoreceptor Stress
Retinal degeneration triggers photoreceptor degeneration that manifests as shortening
of the rod photoreceptor outer segments. The stressed photoreceptors then begin to
deconstruct their synaptic terminals and start sprouting neurites. These neurites form
25
non-conventional synapses and extend their processes as far as the ganglion cell layer.
Often these processes are marked by rhodopsin relocalization in the ganglion cell layer
(Milam et al., 1996). With the onset of sprouting neurites from the rod photoreceptors,
synaptic signaling fails and triggers a range of rewiring events including retraction of
bipolar cell dendrites, switching of synaptic targets by bipolar cells, and anomalous ex-
tension of horizontal cell processes into the inner plexiform layer(Jones and Marc, 2005;
Marc et al., 2007; Strettoi et al., 2002, 2003).
Phase II: Photoreceptor Death
Phase II is marked by the complete loss of photoreceptors due to trophic effects. Glial
cells start removing the debris left over by photoreceptor loss. The signature event of
the phase II is the formation of a glial seal by Muller cell distal processes that seals off
completely the inner retina from the subretinal space and RPE.
Phase III: Neuronal Remodeling
After formation of the glial seal, remodeling of the inner retina begins. Remodeling
includes neuronal death including massive loss of ganglion cells, relocation of all types
of surviving neurons, fragmentation of the IPL lamination, development of neurites from
complex fasciles surrounded by Muller cells and invasion of RPE cells deep into the retina
(Jones et al., 2003; Santos et al., 1997; Stone et al., 1992). Figure 1.13 is a schematic of
the retinal remodeling in its different stages.
26
Figure 1.13: Schematic representation of the three stages of retinal degeneration (Jones and
Marc, 2005).
27
1.6 History of Visual Prostheses
Restoring functional vision in blind people using electrical stimulation has been the goal
of ophthalmologists and vision scientists for more than a century. Earliest known exper-
iments to restore vision with electricity date back to 18th century. A French physician,
Charles Le Roy, began experimenting with electricity in 1755 to influence physiological
function. In one application, Le Roy wound conducting wires around the head of a blind
man. The wires were connected to an array of Leyden jars and 12 shocks were adminis-
tered in the hope that sight would be restored. Along with the pain of the stimulation,
the patient did perceive vivid flashes of light, termed “phosphenes”, and underwent the
treatment several times in the following days (LeRoy, 1755). However, it wasn’t until the
later part of 20th century that scientists and engineers began rigorously investigating the
feasibility of using electrical stimulation to restore visual perception. Earlier attempts
to restore functional vision began with stimulating higher visual centers such as V1 and
LGN. The following is a brief discussion of such attempts at building a visual prosthesis.
1.6.1 Primary Visual Cortex Simulation
After LeRoy’s first attempt at invoking phosphenes in blind patients, a few other studies
confirmed the presence of phosphenes in response to electrical stimulation of the visual
cortex (Foerster, 1929; Krause and Schum, 1931). Foerster demonstrated that awhen
the occipital pole was electrically stimulated, a small spot of light that was directly in
front and motionless was perceived. Krause and Schum further reinforced the generation
of phosphenes in response to electrical stimulation of the left occipital pole in a patient
28
who has been blind for eight years. This showed that the adult visual cortex does not
lose its functional capacity after years of visual input deprivation. These observations
were confirmed by more studies demonstrating that visual sensations could be evoked
by electrical stimulation of the visual cortex (Button and Putnam, 1962; Penfield and
Rasmussen, 1952). Brindley and Lewin’s work in early 1960s was the first true attempt
at a chronic visual prosthesis. They implanted an array of 80 platinum electrodes on
the occipital pole in the right cerebral hemisphere of a 52-year-old patient who has been
left blind due to retinal detachment (Brindley and Lewin, 1968). Figure 1.14 shows
the extracranial part of the implant.When a train of pulses was delivered to any one of
the 80 electrodes, patients perceived small spots of white light (phosphenes) that were
described as “a star in the sky”. Although visual sensations were evoked with electrical
stimulation, this work did not culminate in the restoration of functional vision. Dobelle
conducted similar studies on electrical stimulation of visual cortex in 1974. He implanted
two patients with 64 platinum disc electrodes in a hexagonal array (Dobelle et al., 1974).
Like Brindley, Dobelle carefully evaluated how changes in the parameters of stimulation
lead to changes in the quality of the visual percepts. Although several groups continue
develop cortical prosthesis (Bradley et al., 2005; Normann et al., 1999), the technology
faces a number of hurdles. Significant amount of processing occurs in the retina by the
time the neural signals reach the cortex. This results in more complex phosphenes being
perceived by the patients. Further, surgical complications such as intracranial hemorrhage
limit the development of cortical prosthesis.
29
1.6.2 LGN Stimulation
In addition to stimulating the visual cortex, attempts have been made to develop LGN
based visual prosthesis. Pezaris and Reid have demonstrated that either visually or
electrically stimulating specific receptive fields of the LGN, results in highly localized
and repeatable eye movements to the percept in nonhuman primates (Pezaris and Reid,
2007). Although, it is feasible to stimulate LGN this technology is still in the early days
of development.
Figure 1.14: Extracranial part of the visual cortex implant (Brindley and Lewin, 1968).
30
1.6.3 Optic Nerve Stimulation
With the development of nerve cuff electrodes, it has become feasible to stimulate the
optic nerve in order to evoke visual sensations. Veerat and Brelen have demonstrated that
visual percepts could be generated by implanting a nerve cuff electrode around the optic
nerve in a single blind patient (Brel´ en et al., 2005; Veraart et al., 2004). A major setback
with this approach is that the optic nerve doesn’t maintain retinotopy (Fitzgibbon and
Reese, 1996) making it difficult to spatially map stimuli.
1.6.4 Subretinal Stimulation
In the subretinal approach, the implant sits on the outer surface of the retina in between
the photoreceptor layer and the retinal pigment epithelium. Subretinal implants consist
of a silicon wafer with light sensitive microphotodiodes that convert light into current.
Earlier work in this field involved the implantation of such passive photodiode arrays.
These arrays convert ambient light into small electric currents that in turn were applied
to the retina for stimulation (Chow et al., 2004). Clinical studies using these passive
implants showed that although some improvement in vision was noticed, the current gen-
erated by the array was not sufficient to excite retinal neurons. The slight improvement
in vision was attributed to the release of neurotrophic factors associated with the surgical
implantation. A German group has implemented another subretinal approach using mi-
crophotodiodes. This project, Alpha IMS, utilized an active power source in addition to
the passive photodiodes to stimulate the retina. The microphotodiode array consists of
1500 diodes , each with stimulating electrodes and amplifiers generating a pattern of 38x
40 pixels (Gekeler et al., 2004; Zrenner et al., 2011). Evoked potentials from the cortex
31
were measured from Yucantan micropigs and rabbits implanted with the devie. Clinical
studies with 3 RP subjects have shown that they could locate bright objects on a dark
table. 2 subjects could discern grating pattens and one was able to correctly describe and
name objects (Zrenner et al., 2011).
1.6.5 Epiretinal Stimulation
Although many groups have been working on developing epiretinal stimulation implants(Rizzo
et al., 2003a,b), I will focus on the array (ARGUS II Retinal Prosthesis) developed by
Second Sight Medical Products (SSMP) in collaboration with University of Southern Cal-
ifornia(Caspi et al., 2009; de Balthasar et al., 2008). ARGUS II is the only prosthesis
approved by the FDA for implantation in patients suffering from RP. ARGUS II consists
of a 60-electrode stimulation array with 200 μm disc electrodes, an inductive coil used to
transmit data and power electronics, an external video processing unit (VPU) worn on
a belt and a miniature camera mounted on a pair of glasses (Figure 1.15). The external
receiving coil is fixed to the sclera outside the eye. The electrode array covers a 20-degree
visual field and is tacked to the retina near the macula. The video camera captures a
portion of the visual field and relays the information to the VPU. The VPU processes the
image and converts it into a series of stimulus pulses that are delivered to the electrode
array via the inductive link.
1.7 Thesis Overview and Structure
Studies with ARGUS I and ARGUS II implant subjects’ have indicated that, several
improvements have to be made to the epiretinal prosthesis in order for it to produce
32
Figure 1.15: A schematic showing the microelectrode array and the inductive coil on the left.
The video processing unit (VPU) and a miniature camera are shown as well.
high quality vision in these subjects. Specifically, these studies have demonstrated that
it is still not possible to spatially and temporally control the percept evoked by epiretinal
electrical stimulation. The studies have also demonstrated that higher thresholds limit
the use of all the electrodes on the array without exceeding electrochemical safety issues.
This thesis aims to address these issues using an in vivo model of epiretinal prosthesis to
study effect of stimulus waveform shape on stimulation efficiency and also to develop new
strategies to obtain better control on the temporal characteristics of the retinal ganglion
cells. My specific hypothesis was that by tuning stimulus parameters, continuous electri-
cal stimulation of the retina would result in patterns of ganglion cell activity that better
conform to the stimulus pattern. Indeed, I found that simply modifying the electrical
stimulation pattern from time in-varying pulse in which every pulse had the same ampli-
tude and pulse duration, to time-varying in which every consecutive pulse had varying
duration and amplitude was sufficient to better control the temporal characteristics of
retinal ganglion cells.
In an effort to lower stimulation thresholds, I studied the effect of voltage-controlled
and current-controlled pulses on stimulation efficiency. This study was performed in
vivo and a computational model was also created to study the effect of several waveform
33
shapes on stimulation efficiency. The results of this study are described in chapter 2.
Chapter 3 describes strategies to limit desensitization caused by continuous electrical
stimulation of the retina. Desensitization of the retina could be one of the reasons why the
brightness of phosphenes fades over time and hence it is important to develop strategies
that address this issue. In chapter 4, an imaging technique known as optical imaging of
intrinsic signals has been described. The principles underlying the technique and several
interesting studies that have been performed using the technique have been described. In
addition, I demonstrated the methodology behind imaging intrinsic signals in the superior
colliculus of rats. This technique could be useful in performing chronic studies of electrical
stimulation. Based on all the results, an overall summary and future work has been laid
out in chapter 5. Details of the experimental methods employed throughout the study
have been described in chapter 6.
34
Chapter 2
Relative Efficiency of Voltage Vs. Current Controlled
Pulses for Epiretinal Stimulation
2.1 Background
Retinal prostheses aim to restore functional vision in patients suffering from retinal degen-
eration diseases such as Retinitis Pigmentosa (RP) (Humayun et al., 2012; Rizzo et al.,
2003b; Zrenner et al., 2011). Vision loss in these patients begins with the progressive
degeneration of photoreceptors. As a consequence of the loss of synaptic input from the
photoreceptors, inner retinal circuitry undergoes extensive remodeling (Jones and Marc,
2005). In spite of the photoreceptor loss and subsequent rewiring in the inner retina,
research has shown that blind humans perceive light when the neural retina (including
bipolar cells and ganglion cells) is electrically stimulated, in large part due to the large
number of neurons remaining in these layers(Margolis et al., 2008; Santos et al., 1997).
Retinal prostheses strive to restore vision by electrically stimulating the remaining retinal
neurons.
35
A recent report of a retinal prosthesis clinical trial suggests a need to improve stim-
ulation efficiency. Across 30 patients with retinal implants, only 55% of electrodes could
evoke visual percepts using stimulus below 1 mC/cm
2
, which was the charge density
limit of Platinum-gray, the electrode material used in the implants (Humayun et al.,
2012; Sanders et al., 2007). To increase the number of functional electrodes, more effi-
cient methods of stimulation must be developed, so that visual percepts can be generated
within charge density limits. In general, improvements in stimulus efficiency can lead to
neurostimulation devices that are more effective, smaller, and safer. Non-functional elec-
trodes can many times be attributed to the distance between the electrode and the retina
and a reasonable solution is to simply work towards better positioning of the retinal elec-
trode. However, electrode-positioning issues are difficult to overcome, due to individual
variations in eye size. Indeed, cochlear implants, in spite of decades of development, still
encounter challenges related to consistent electrode positioning in the cochlea. Thus, im-
proving the stimulus efficiency of retinal implants is an important goal with implications
for current and future devices.
Stimulation efficiency has been studied by manipulation of the waveform used for
stimulation. Neural prostheses use either current or voltage rectangular pulses for stim-
ulation. In vivo and computational studies have shown that stimulus parameters such as
pulse duration and waveform shape have an effect on the neural response amplitude and
power efficiency (Fishler, 2000; Foutz and McIntyre, 2010; Jezernik and Morari, 2005,
2010; Kajimoto et al., 2002; Klafter and Hrebien, 1976; Offner, 1946; Sahin and Tie,
2007; Wongsarnpigoon and Grill, 2010). However, the effect of stimulus parameters on
36
the efficiency of retinal stimulation has not been extensively studied. In this study we as-
sessed a number of factors that may contribute to the efficiency of retinal stimulation. We
compared the effect of voltage-controlled and current-controlled pulses on the strength of
neural response in normally sighted and retinal degeneration rats, by measuring electri-
cally evoked responses (EERs) of Superior Colliculus (SC) neurons in response to retinal
stimulation. We also compared two electrode materials: standard platinum-iridium and
high-surface area platinum-iridium (Petrossians et al., 2011). In addition to comparing
the strength of neural response, we measured the power consumed while generating the
voltage-controlled and current-controlled pulses. To further explore the experimental re-
sults, we conducted computational studies to compare the effect of voltage-controlled and
current-controlled pulses on stimulation efficiency.
In addition to using the high-surface area platinum-iridium alloy as a stimulation elec-
trode, we also electroplated the material on a recording electrode to determine if the alloy
helped improve the quality of EERs. The high-surface area platinum-iridium recording
electrode was compared with a standard tungsten electrode. Electrically evoked signals
are often corrupted by electrical stimulus artifact and this often makes the subsequent
analysis of the underlying neural response difficult. This is particularly evident when in-
vestigating short-latency neural response in response to high-rate electrical stimulation.
In the study, we compared the strength of the stimulus artifact generated by the standard
tungsten electrode and that generated by the high-surface area platinum-iridium coated
recording electrode.
37
2.2 Experimental Protocol
2.2.1 Experimental Groups
There were 3 experimental groups. In the 1
st
and 2
nd
experimental groups, 10 and 5
normally sighted rats were used respectively. In the third experimental group 4 RD
rats were used. There were two stimulation electrodes: 1) The first electrode was a
concentric bipolar electrode with a flat tip. 2) The second electrode was similar to the
first one but the tip was modified by electroplating a Pt-Ir alloy on the surface of the
inner electrode. The first electrode is referred to as “uncoated electrode” and the second
electrode is referred to as the “coated electrode”. In the 1
st
and 3
rd
experimental groups,
the uncoated electrode was used as the stimulation electrode. For the 2
nd
experimental
group, the coated stimulation electrode was used.
As part of the stimulus artifact study, there were 2 experimental groups. In the first
and second experimental groups, 5 normally sighted rats per group were used. The stim-
ulation electrode in the first group was the coated electrode and the recording electrode
was an unmodified tungsten electrode. In the second experimental group, the coated
stimulation electrode was used with the modified tungsten electrode for recording.
2.2.2 Electrical Stimulation
Charge-balanced, cathodic first, biphasic current and voltage-controlled pulses were de-
livered to the epiretinal surface across 4 pulse durations (0.3, 0.5, 1 and 2 ms) through
the stimulation electrode. The amplitude of the pulses was chosen such that the charge in
the cathodic phase was between 10 and 60 nC. Stimulus pulses were supra-threshold. For
38
each pulse duration, voltage and current-controlled pulse trains with four different charge
levels were delivered. Thus, 32 stimulus conditions (2 stimulation modes x 4 charge levels
x 4 pulse durations) were applied to the retina per experiment. The interphase interval
was kept constant at 100 μs. The stimulus pulses were delivered at 0.2Hz for all pulse
durations. 25 pulses were delivered for each pulse duration and amplitude in each stim-
ulation mode. The order in which current and voltage-controlled pulses were delivered
was randomized.
2.3 Results
2.3.1 In Vivo Experiement Results
Strength in EERs generated by voltage-controlled pulses and current-controlled pulses
was used to determine stimulation efficiency. In the following figures, signal strength was
plotted against charge. Student t-test was used to analyze the significance of the findings
in the processed data.
Comparison of Signal Strength in EERs Measured in Normally Sighted Rats
The strength of EERs generated by electrical stimulation of the retina with voltage-
controlled and current-controlled pulses is shown in figure 2.1. An uncoated electrode was
used in these experiments. In general, whether current-controlled or voltage-controlled
pulses are more efficient depends on the stimulus pulse width. For 0.3 ms pulse width,
at every charge level current-controlled and voltage-controlled pulses generated EERs
that were not significantly different (p value range: 0.15-0.19). For 0.5 ms pulse width,
39
Figure 2.1: Strength of EERs versus charge. EERs are recorded from SC when rectangular
voltage-controlled and current-controlled pulses stimulate the retina. A standard Pt-Ir electrode
was used to stimulate the retina using rectangular voltage-controlled and current-controlled pulses.
* indicates statistical significance. Error bars indicate standard error.♢ voltage-controlled pulses.
current-controlled pulses. rge levels (p value range: 0.0004-0.0019).
voltage-controlled pulses generated EERs with significantly higher strength than current-
controlled pulses across all charge levels (p value range: 0.02-0.045). For 1 ms pulse
duration, at 3 charge levels (30nC, 40nC and 50nC) current-controlled pulses generated
EERs with significantly higher strength than voltage-controlled pulses (p value range:
0.014-0.021). And for 2 ms pulse width, current-controlled pulses generated EERs with
significantly higher strength than voltage-controlled pulses for all charge levels (p value
range: 0.0004 - 0.0019).
Comparison of Signal Strength in EERs with the Coated Stimulation Electrode
In figure 2.2, the strength in EERs is compared when current-controlled and voltage-
controlled pulses were used to stimulate the retina with the high surface area electrode.
40
Figure 2.2: Strength of EERs versus charge. A high surface area Pt-Ir electrode was used to
stimulate the retina using rectangular voltage-controlled and current-controlled pulses.* indicates
statistical significance. Error bars indicate standard error.♢ voltage-controlled pulses. current-
controlled pulses.
For 0.3 ms pulses, at 1 charge level voltage-controlled pulses generated EERs with sig-
nificantly higher strength than those generated by current-controlled pulses (p value =
0.046). For 0.5 and 1 ms pulses, both current-controlled and voltage-controlled pulses
generated EERs that were not significantly different in strength at every charge (p value
range: 0.07-0.48). 2 ms data is not shown since the range of charge applied in current
controlled and voltage controlled modes did not overlap, and thus a proper comparison
between the two modes could not be made.
Comparison of Signal Strength in EERs Measured in RD Rats
In figure 2.3, the strength in EERs is compared between the two stimulation modes when
the standard Pt-Ir electrode was used for stimulation in retinal degeneration rats. For 0.3
ms pulse width, at every charge level both current-controlled and voltage-controlled pulses
generated EERs that did not differ significantly (p value range: 0.16-0.23). For 0.5 ms
pulses, voltage-controlled pulses generated EERs with significantly higher strength than
current-controlled pulses at the lower charge levels (20 and 30nC, p value range: 0.03-
0.047). For 1 ms duration pulses, at 3 charge levels (30, 40 and 50nC) current-controlled
41
Figure 2.3: Strength of EERs versus charge for RD retina. Rectangular voltage-controlled and
current-controlled pulses stimulated the degenerate retina and EERs were recorded from SC. A
standard Pt-Ir electrode was used for stimulation. * indicates statistical significance. Error bars
indicate standard error.♢ voltage-controlled pulses. current-controlled pulses.
pulses generated EERs with significantly higher strength than voltage-controlled pulses
(p value range: 0.008-0.009). And for 2 ms pulses, current-controlled pulses generated
EERs with significantly higher strength than voltage-controlled pulses at the 3 higher
charge levels (30, 40 and 50nC, p value range: 0.018-0.029).
2.3.2 Simulation Results
The model simulates membrane voltage and gating particles of a cat RGC in response
to stimuli of varying pulse duration in both stimulation modes. The current waveforms
recorded during the in vivo experiments for both stimulation modes were used for intra-
cellular stimulation of the model.
42
Computational Model Responses for Varying Stimuli (Uncoated Electrode)
In figure 2.4, simulation results are shown when current waveforms from the standard Pt-
Ir electrode were used for intracellular stimulation of the model. For stimulus pulses with
0.3 ms duration, the membrane voltage and gating particles had the same trajectory. For
0.5 ms duration pulse, the membrane voltage and gating particles have faster temporal
dynamics with voltage-controlled pulses. The model results were different with longer
pulse widths. For 1 ms and 2 ms pulses, the membrane and gating particles showed a
delayed response with a voltage-controlled pulse.
Computational Model Responses for Varying Stimuli (Coated Electrode)
In figure 2.5, simulation results are shown when current waveforms from the high surface
area electrode were used for intracellular stimulation of the model. The trajectory of the
membrane voltage and gating particles were equivalent for both current-controlled and
voltage-controlled stimuli across all pulse durations. Simulation results for 0.3 ms pulse
duration are shown as representative of all pulse widths.
2.3.3 Power Consumption Results
We compared the power required to generate the current-controlled and voltage-controlled
waveforms. In addition to comparing power requirement between both stimulation modes,
the power requirement between the two electrode types was also compared. Student t-test
was used to analyze the significance of the findings. Note the difference in y-axis range
for all pulse durations. Because power is the rate of energy usage, shorter pulse widths
use energy at a higher rate explaining the difference in y-axis values.
43
Figure 2.4: Membrane dynamics of a retinal ganglion cell model. The current waveforms used
for model simulation were obtained when a standard Pt-Ir electrode was used to stimulate the
retina. red line current-controlled blue line voltage-controlled.
Figure 2.5: Simulation of a rat retinal ganglion cell model. The current waveforms used in the
simulation were obtained when the retina was stimulated with a high surface area Pt-Ir electrode.
red line current-controlled blue line voltage-controlled.
44
Figure 2.6: Power consumption versus charge. The power consumed to generate voltage-
controlled and current-controlled pulses is compared when a standard Pt-Ir electrode was used
for stimulation.* indicates statistical significance. Error bars indicate standard error.♢ voltage-
controlled pulses. current-controlled pulses.
Uncoated Electrode Power Consumption Comparison Between Both Stimulation
Modes
Power required to generate voltage-controlled pulses was significantly lower than that
required to generate current-controlled pulses for 0.3 ms, 1ms and 2ms pulses at lower
charges (p value range: 0.0001 -0.0174). For 0.5 ms pulses, power required was not
significantly different for both stimulation modes (p value> 0.05). Figure 2.6 shows the
comparison between both modes with the uncoated electrode.
Coated Electrode Power Consumption Comparison Between Both Stimulation
Modes
With the coated electrode, power could not be compared at the lowest charge level (20
nC). Voltage-controlled pulses were significantly power efficient at the 30 nC charge level
for 0.5 ms pulse (p value = 0.0001). Voltage-controlled pulses were significantly efficient
45
Figure 2.7: Power consumption versus charge. Power required to generate voltage-controlled and
current-controlled pulses is compared when a high surface area Pt-Ir is used for stimulation. Power
consumption data at the lowest charge could not be compared.* indicates statistical significance.
Error bars indicate standard error.♢ voltage-controlled pulses. current-controlled pulses.
at all charge levels for 1 ms and 2 ms pulses (p value range: 0.0004-0.04). Figure 2.7
shows the power comparison with the coated electrode.
Power Consumption Comparison Between Coated and Uncoated Electrodes
for Current-Controlled Stimulation
We compared power consumption between the two stimulation electrodes in each mode.
Figure 2.8 compares the power required to generate current-controlled pulses when the
coated and uncoated electrode were used for stimulation. Coated electrode consumed
significantly less power to generate current-controlled pulses across all pulse durations (p
value range: 0.0001-0.02).
46
Figure 2.8: Power consumption versus charge. Power required to generate current-controlled
pulses is compared when a standard Pt-Ir electrode and a high surface area Pt-Ir electrode are
used for stimulation. Power consumption data at the lowest charge could not be compared. * indicates statistical significance. Error bars indicate standard error.∘ standard Pt-Ir electrode.
△ high surface area Pt-Ir electrode.
Figure 2.9: Power consumption versus charge. Power required to generate voltage-controlled
pulses is compared when a standard Pt-Ir electrode and a high surface area Pt-Ir electrode are
used for stimulation. Power consumption data at the lowest charge could not be compared. * indicates statistical significance. Error bars indicate standard error.∘ standard Pt-Ir electrode.
△ high surface area Pt-Ir electrode.
47
Power Consumption Comparison Between Coated and Uncoated Electrodes
for Voltage-Controlled Stimulation
Figure 2.9 compares the power required to generate voltage-controlled pulses when the
coated and uncoated electrode were used for stimulation. Coated electrode consumed
significantly less power to generate the voltage-controlled pulses across all durations (p
value range: 0.0002-0.04).
2.3.4 Comparison of Electrode Properties
In figure 2.10, the voltage drop and current are compared between the two electrodes
during current-controlled and voltage-controlled stimulation. The voltage drop across
the electrode-electrolyte interface is lower with the high surface area Pt-Ir electrode for
current pulses of the same amplitude. For voltage-controlled pulses of same amplitude
and duration, the high surface area electrode generates more current.
2.3.5 Comparison of Stimilus Artifacts
Stimulus artifact generated by the standard recording electrode and the high-surface area
platinum-iridium coated recording electrode was compared in figure 2.11. The stimulus
artifact generated by the modified recording electrode was much smaller than that gener-
ated by the standard tungsten recording electrode. Stimulus artifact was also quantified
by calculating the strength within the stimulus artifact. In figure 2.12, the strength of
stimulus artifact was compared for two pulse durations (0.5 ms and 1 ms). As shown
in figure 2.12, the strength of the stimulus artifact generated by the standard tungsten
recording electrode was at least 10 times higher than that generated by the modified
48
Figure 2.10: Comparison of measured current and voltage between standard Pt-Ir electrode
and high surface area Pt-Ir electrode during voltage-controlled and current-controlled stimulation
respectively.
recording electrode. Statistical analysis also indicated that the stimulus artifact gener-
ated by the modified recording electrode was significantly smaller than that generated by
the standard recording electrode (p-value range: 0.00001 – 0.0001).
2.4 Discussion
The key findings of this study are:
• Stimulus waveform efficiency depends on the pulse width: voltage-controlled pulses
generate stronger neural response when the pulses have short duration and current-
controlled pulses generate stronger neural response when the pulses have long du-
ration.
• Models of RGC stimulation show that the temporal dynamics vary based on the
shape and pulse duration of the stimulus waveform and are consistent with the
experimental results.
• Voltage-controlled pulses are power efficient when compared to current-controlled
pulses.
49
Figure 2.11: Comparison of stimulus artifact generated by the standard recording electrode and
modified recording electrode.
Figure 2.12: Stimulus artifact strength vs. charge in the stimulus pulse. Strength of the stimulus
artifact is compared when a standard tungsten recording electrode and a modified recording
electrode were used. The duration of the stimulus pulses were 0.5 ms and 1 ms.
50
• High surface area Pt-Ir electrode reduces power consumption significantly in both
stimulation modes.
• High surface area Pt-Ir recording electrode reduces stimulus artifact significantly.
2.4.1 In Vivo Experiments with Standard Pt-Ir
In general, our results are consistent with other studies that examined pulse duration and
stimulus mode. (Wongsarnpigoon et al., 2010) reported that the stimulation efficiency of
a given waveform depends on the pulse duration and that no given waveform is the most
efficient across all pulse durations. They analyzed the efficiency of several stimuli such
as rising and decaying exponentials, rectangular and ramp waveforms in computational
models and in vivo experiments of cat sciatic nerve excitation. (Goo et al., 2011) also
conducted a similar study comparing current-controlled and voltage-controlled pulses
when stimulating RGCs in an in vitro setup. They compared the threshold charge density
for stimulating RGCs and found that the threshold charge density was lower when voltage-
controlled pulses were used for stimulation. The pulse duration in their study was fixed
at 0.5 ms. Thus, our in vivo results are in agreement with the Goo in vitro study at this
pulse width. Wagenaar et al also compared the effect of current-controlled and voltage-
controlled pulses on neural networks cultured on multi-electrode arrays (MEAs). They
reported that an anodic first biphasic voltage-controlled pulse was efficient in stimulating
the neural networks, but only examined pulse duration below 1 ms. Again, these results
agree with our study, since we found that current-controlled pulses are more efficient
when the duration is greater than or equal to 1ms.
51
As the pulse width increased, the strength in EERs was lower for voltage-controlled
pulses when compared to EERs from current-controlled pulses of equal duration, even
when corrected for the charge delivered to the tissue. To determine the cause of this
behavior, current waveforms with varying pulse widths generated by voltage-controlled
pulses were examined. The voltage step led to high current initially, as the capacitive
double layers behaved as a short circuit (figure 2.10). This high current in the tran-
sient state leads to a higher membrane current and may have resulted in stronger EERs
when voltage-controlled pulses were used. As the double layer charges, current flow de-
creases since impedance increases and voltage is constant. Current below a certain level
(rheobase) can no longer stimulate tissue regardless of pulse duration, so decreasing cur-
rent amplitude at the end of the voltage pulse contributes minimally if at all to the EER.
Also at longer pulse durations, the continuous and consistent delivery of charge in current-
controlled pulses might have caused the inner retinal neurons to generate graded poten-
tials. These graded potentials might have contributed to the increase in EER strength
when long duration current-controlled pulses were used. Another observation is that
stimulation with 0.3 ms duration pulses showed no difference between strength in evoked
potentials generated by voltage-controlled and current-controlled pulses. Examination
of the current waveform from a 0.3 ms voltage pulse shows that the current, while not
uniform, is still in the decay phase and approximates a rectangular current pulse. Thus,
both the current pulse and voltage pulse yield a similar current waveform through the
electrode. Another explanation for the similar neural response is based on the chronaxie
of a retinal ganglion cell. The chronaxie for electrical stimulation of a rodent retinal
ganglion cell is estimated at 0.4 ms (Chan et al., 2011). When the pulse width is smaller
52
than the chronaxie and charge is integrated, the shape of the stimulus waveform may
have a small effect on the EER.
2.4.2 In Vivo Experiments with High Surface Area Pt-Ir
In vivo experiments with the high surface area electrode showed that both stimulation
modes generated EERs of equal strength and hence are equally efficient. The electrode
time constant is greater for high-surface area electrodes, due to greater capacitance (figure
2.1). As a result, constant voltage pulses resulted in a slowly decaying current. Uncoated
electrodes, when a constant voltage was applied, showed a high initial current followed by
a rapid decay (figure 2.10). As the shape of the current waveform was similar to that of a
rectangular current pulse, stimulation with high surface area material may have resulted
in similar neural excitation in SC.
2.4.3 Computational Studies
In order to further examine our in vivo results, we conducted computational studies using
a Hodgkin-Huxley model of cat RGC. We used as input to the RGC model recorded cur-
rent waveforms from both current and voltage stimulation modes. The different shape of
the current waveform in current-controlled pulses and voltage-controlled pulses resulted
in differences in the trajectories of the membrane voltage and gating particles. Moreover,
the temporal dynamics of the RGC model were also dependent on the pulse duration of
the stimulus waveform. With 0.5 ms pulse duration, the model predicted that membrane
voltage changes more rapidly with voltage stimulation compared to current stimulation,
while the opposite was true with pulse duration above 0.5 ms. The model and experiments
53
also agree in cases when the experiments deemed current and voltage stimulation to be
virtually equivalent. (Sahin and Tie, 2007) conducted simulation studies of a mammalian
nerve model and found that the trajectories of membrane voltage and gating particles
vary as a function of the stimulus waveform shape. (Wongsarnpigoon and Grill, 2010;
Wongsarnpigoon et al., 2010) also conducted simulation studies of extracellular stimu-
lation of a population of myelinated axons. They also report that the rates at which
the membrane parameters changed varied with both pulse duration and waveform shape.
As concluded by (Wongsarnpigoon and Grill, 2010; Wongsarnpigoon et al., 2010), the
activity of the membrane parameters does not fully explain why one waveform is more
efficient than the other. Nevertheless, the model shows that the temporal dynamics of
the membrane particles change significantly with stimulus pulse duration and the shape
of the waveform.
2.4.4 Power Efficiency
We found that when the high surface area (coated) Pt-Ir electrode is used for stimulation,
less power is required to generate the stimulus pulses of a given charge when compared
to the uncoated Pt-Ir electrode. Comparison of bode plots for coated and uncoated elec-
trodes demonstrates reduced impedance for coated electrode as a result of the increased
surface area which in turn increases the capacitance of the electrode. The increase in
capacitance can be noted from the longer time constant of the current waveform gener-
ated by the high surface Pt-Ir electrode (figure 2.10). Equation below demonstrates that
impedance modulus is inversely proportional to capacitance.
54
=
1

where is the measured impedance modulus, is the radial frequency and is the
capacitance. By decreasing the impedance of the electrode the measured current increases
while the measured voltage decreases during voltage and current stimulations respectively
(figure 2.10). As a result, the stimulator consumes less power in both voltage-controlled
and current-controlled stimulation modes.
2.4.5 Stimulus Artifact
We found that when the high surface area Pt-Ir alloy was used to modify the standard
tungsten recording electrode, the stimulus artifact generated by the modified recording
electrode was significantly smaller than that generated by the standard recording elec-
trode. As mentioned earlier, the high surface area Pt-Ir alloy reduces the impedance at
the electrode-tissue interface. Lower impedance results in a smaller artifact.
55
Chapter 3
Strategies to Limit Retinal Desensitization Caused by
Continuous Electrical Stimulation
3.1 Background
Psychophysics experiments with epiretinal prosthesis subjects have shown that the bright-
ness of the electrically evoked visual percepts tends to decrease over a period of time with
continuous electrical stimulation (P´ erez et al., 2012). One proposed strategy to improve
outcome for retinal prosthesis subjects’, is to precisely control precisely control the tem-
poral and spatial pattern of ganglion cell activation. Spike control can be accomplished
either by direct activation of the ganglion cell, or through activation of neurons presy-
naptic to the ganglion cells. Direct activation of the ganglion cell has the ability to elicit
spike trains at very high rates (Ahuja et al., 2008; Fried et al., 2005; Sekirnjak et al.,
2006). However, direct activation is also likely to cause incidental activation of passing
axons on the inner retinal surface. This will expand the spatial region of ganglion cell ac-
tivation, and may smear the elicited percept. Alternatively, the activation of presynaptic
neurons is advantageous in that it provides better spatial control over neural activation
56
by avoiding ganglion cell axons. Unfortunately, activation through the synaptic network
limits the ability to control the temporal pattern of ganglion cell spiking. For example,
in response to repetitive stimulation, ganglion cells respond robustly to the first pulse,
but the response decreases for subsequent pulses (Jensen and Rizzo, 2007).
This reduction in ganglion cell excitability is termed desensitization. Ray (2010)
has shown a similar decrease in neural excitability using an in vivo model of epiretinal
prosthesis. One of the causes for retinal desensitization could be increased inhibition from
amacrine cells. (Fried et al., 2005) have shown that in response to electrical stimulation
with long pulses, excitatory and inhibitory currents could be measured in ganglion cells.
These currents indicate that pulses stimulate both bipolar cells and amacrine cells. They
have also shown that prolonged stimulation reduces the excitatory current from bipolar
cells proving that inhibition increases from amacrine cells. This inhibition lasted over 100
ms.
The objective of this study is to present new stimulation strategies to limit desensitiza-
tion. We propose to test the effect of time-varying pulses and short duration pulses (pulse
width < 0.1 ms) on retinal desensitization caused by continuous electrical stimulation.
Time-varying pulses are defined as a train of pulses where each pulse has a different pulse
duration and amplitude when compared to the preceding pulse. The charge delivered
for every pulse is three times the threshold charge for that pulse width. In addition to
time-varying pulses, the effect of short duration pulses was also studied. Different pulse
widths are known to stimulate different populations of neurons (Undurraga et al., 2012;
Weitz, 2013). Microsaccades continuously shift the image on the retina and help prevent
mechanisms of local adaptation and image fading. By stimulating different populations
57
of neurons, the “electrical image” on the retina might shift constantly and help limit
image fading. Hence, time-varying pulses could help limit desensitization. (Weitz, 2013)
have shown that short duration pulses directly activate the ganglion cells. In addition
to directly activating ganglion cells, short duration pulses have been shown to decrease
inhibition from amacrine cells (Fried et al., 2005).
3.2 Experiment Protocol
Figure 3.1 illustrates the experimental protocol. Continuous stimulation consisted of two
types of pulses: test stimuli and probe pulses. Test stimuli were a train of pulses (60
A, 0.5 ms) delivered at 20 Hz for 1 sec. Test stimuli were preceded by a single probe
pulse (0.5 ms, 60 A). During the first part of the experiment (figure 3.1A), probe pulses
interleaved with test stimuli examined the effect of continuous stimulation on electrically
evoked responses (EERs) recorded in the superior colliculus (SC) of rats. The probe pulse
and test stimuli combination was delivered for 2 min and this part of the experiment was
referred to as “stimulation phase”. In the second part of the experiment after the probe
pulse and test stimuli combination ended (figure 3.1B), only probe pulses were delivered
for 3 min to monitor the recovery of the EERs. The second part of the experiment is
referred to as “recovery phase”. Probe pulses were delivered at 0.2 Hz after the continuous
stimulation ended.
Two types of experiments were performed: “control experiments” in which the test
stimuli consisted of stimulus pulses that did not vary in time (all the stimulus pulses had
the same amplitude and pulse duration i.e. 60 A and 0.5 ms) and “time-varying” pulse
58
Figure 3.1: Stimulation protocol used in control and experimental groups.
59
train experiments in which the test stimuli consisted of pulse train where each consecutive
pulse had different amplitude and pulse duration when compared to the preceding pulse.
The pulse duration of the time-varying pulses was randomly chosen from a set of pulse
widths between 0.1 ms and 20 ms. The amplitude of the pulses was chosen such that
it was three times the threshold amplitude for that particular pulse duration. In order
to determine the threshold amplitude, a strength-duration curve was constructed from
3 experiments in addition to those in control and experimental groups (Figure 3.2). In
addition to time-varying pulses, short duration pulses were also tested in the experimental
group. In the short duration pulse experiment, test stimuli consisted of stimulus pulses
with 0.05 ms pulse duration in each phase and the amplitude of the pulses was 1500 A.
The amplitude was three times the threshold amplitude for 0.05 duration pulses.
Control experiments, time-varying pulse and short-duration pulse experiments were
performed in both normally sighted rats (Long-Evans, n = 8 animals) and retinal degen-
erate animals (S334ter line 3 homozygous rats, 400 days post natal, n = 6). For all the
probe pulses, the resulting EERs were recorded. The strength of the EER was calculated
for each probe pulse. Stimulus artifact was excluded from this calculation. In order to
illustrate how the strength of the EER varied over time, the strength of EERs from all
the probe pulses was normalized to the strength of the EER from the first probe pulse.
The normalized EER strength data from all the experiments were averaged and plotted
against time. Standard error was indicated as a shaded region. In all the plots, the nor-
malized EER strength data from 0-120s is from the stimulation phase and from 120-300s
is from the recovery phase.
60
Figure 3.2: Strength-duration curve used to determine the current amplitude for time-varying
pulses.
In order to determine statistical significance, 10 time points were randomly selected
and normalized EER strength data from probe pulses was analyzed at those time points.
The probe pulses were selected from both stimulation phase and recovery phase (5 time
points in each phase). Student t-test was performed to determine statistical significance
between normalized EER strength at a given time from control and time-varying experi-
mental groups. A p-value < 0.05 was considered to indicate significance.
61
3.3 Results
3.3.1 Time-Varying Pulses Limit Desensitization in Normally Sighted
Rats
Normalized EER strength was plotted against time to illustrate the effect of continuous
electrical stimulation on the EER strength. In the following figures, note the difference in
the time scale (x-axis) when the continuous stimulation was on and after the continuous
stimulation ended. The time scale is not linear due to difference in probe pulse frequency
during and after continuous stimulation. Figures 3.3, 3.4 and 3.5 represent average data
from all eight experiments. In figure 3.3, the normalized signal strength is averaged
across all the experiments and is plotted against time. The averaged signal strength was
compared between a control experiment and a time-varying pulse experiment performed
in the same animal. The shaded area in the plot indicates standard error. As is evident in
figure 3.3, the strength of the EERs does not decrease as much when time-varying pulses
were used. The strength of the EERs recovers to pre-stimulation level after continuous
stimulation ends in both time varying and control pulse experiments.
Student t-test was performed to determine statistical significance. Statistical analysis
indicated that the average EER strength was significantly different between the control
and experimental groups (p-values < 0.05). In particular, the normalized strength of
EERs generated by probe pulses in the experimental group during the stimulation phase
was significantly higher than those delivered by probe pulses in the control group. During
the recovery phase, the normalized EER strength data from the experimental group was
statistically equivalent to that from the control group. This indicates that during the
62
Figure 3.3: Average normalized EER strength is plotted against time for both control and
time-varying pulse experiments. The shaded region represents standard error.
stimulation phase, the time-varying pulses do not desensitize the retina as much as the
control pulses. In the recovery phase, the normalized EER strength recovers rapidly when
both control and time-varying pulses were used for stimulation. The range of p-values
for the data points is listed in tables 3.1, 3.2 and 3.3.
A single EER was further divided into early (latency: 3∼ 12 ms) and late (12∼ 40
ms) response. In figure 3.4, the normalized EER strength for early phase was averaged
across all experiments and compared between control and time-varying pulse experiments.
Statistical analysis indicated that the average EER strength in the early phase was sig-
nificantly different between the control and experimental groups (p-values< 0.05). In
particular, the normalized EER strength delivered by probe pulses in the experimental
group during the stimulation phase was significantly higher than those delivered by probe
pulses in the control group. During the recovery phase, the normalized EER strength
63
Figure 3.4: Early phase of the average normalized EER strength plotted against time for both
control and time-varying pulse experiments. The shaded regions represent standard error.
data from the experimental group was statistically equivalent to that from the control
group. This indicates that during the stimulation phase, the time-varying pulses do not
desensitize the early phase response as much when compared to control pulses. The re-
covery phase is statistically equivalent when control and time-varying pulses were used
for stimulation. The p-value range is listed in tables 3.1, 3.2 and 3.3.
In figure 3.5, the normalized signal strength for late phase was compared between
control and time-varying pulse experiments. The late phase response doesn’t desensitize
appreciably when both control and time-varying pulses were used for stimulation. Sta-
tistical analysis indicated that there is no statistical difference between -normalized EER
strength data from control and time-varying experimental groups. The p-value range is
listed in tables 3.1, 3.2 and 3.3.
64
Figure 3.5: Late phase of the average normalized EER strength is plotted against time for both
control and time-varying pulse experiments. The shaded regions represent standard error.
3.3.2 Short Duration Pulses Limit Desensitization in Normally Sighted
Rats
Figures 3.6, 3.7 and 3.8 represent average data from experiments with normally sighted
rats (n=6). In figure 3.6, the normalized signal strength is averaged across all the experi-
ments and is plotted against time. The averaged signal strength was compared between a
control experiment and a short duration pulse experiment performed in the same animal.
The shaded area in the plot indicates standard error. As is evident in figure 3.6, the
strength of the EERs does not decrease as much when short-duration pulses were used.
The strength of the EERs recovers to pre-stimulation level after continuous stimulation
ends in both short duration and control pulse experiments. Statistical analysis indicated
65
Figure 3.6: Average normalized EER strength is plotted against time for both control and
short-duration pulse experiments. The shaded region represents standard error.
that the average EER strength was significantly different between the control and ex-
perimental groups (p-values < 0.05). As with the time-varying pulses, the normalized
strength of EERs generated by probe pulses in the experimental group during the stim-
ulation phase was significantly higher than those delivered by probe pulses in the control
group. In the recovery phase, the normalized EER strength recovers rapidly when both
control and short-duration pulses were used for stimulation. The range of p-values for
the data points is listed in tables 3.1, 3.2 and 3.3.
In 3.7, the normalized EER strength for early phase was averaged across all experi-
ments and compared between control and short-duration pulse experiments. Statistical
analysis indicated that the average EER strength in the early phase was significantly
different between the control and experimental groups (p-values < 0.05). In particular,
the normalized EER strength delivered by probe pulses in the experimental group during
the stimulation phase was significantly higher than those delivered by probe pulses in
66
Figure 3.7: Early phase of the average normalized EER strength plotted against time for both
control and short-duration pulse experiments. The shaded regions represent standard error.
the control group. During the recovery phase, the normalized EER strength data from
the experimental group was statistically equivalent to that from the control group. This
indicates that during the stimulation phase, the short-duration pulses do not desensitize
the early phase response as much when compared to control pulses.
In figure 3.8, the normalized signal strength for late phase was compared between
control and short-duration pulse experiments. The late phase response doesn’t desensitize
appreciably when both control and short-duration pulses were used for stimulation. In
particular, the late phase for the experimental group is quite variable. Statistical analysis
indicated that there is no statistical difference between normalized EER strength data
from control and short-duration experimental groups. The p-value range is listed in tables
3.1, 3.2 and 3.3.
67
Figure 3.8: Late phase of the average normalized EER strength is plotted against time for both
control and short-duration pulse experiments. The shaded regions represent standard error.
3.3.3 Time-Varying Pulses Limit Desensitization in Retinal Degenerate
Rats
Time-Varying pulses were tested in retinal degenerate (RD) rats. S334ter line 3 homozy-
gous rats were used as an RD model. Figure 3.9 is an image of the RD retina.
Figures 3.10, 3.11 and 3.12 represent average data from experiments with RD rats
(n=6). In figure 3.10, the normalized signal strength is averaged across all the experiments
and is plotted against time. The averaged signal strength was compared between a
control experiment and a time-varying pulse experiment performed in the same animal.
The shaded area in the plot indicates standard error. As is evident in figure 3.10, the
strength of the EERs does not decrease as much when time-varying pulses were used. The
strength of the EERs recovers to pre-stimulation level after continuous stimulation ends in
both time-varying and control pulse experiments. Statistical analysis indicated that the
average EER strength was significantly different between the control and experimental
68
Figure 3.9: S334ter line 3 homozygous rat retina (P = 450 days).
Figure 3.10: Average normalized EER strength is plotted against time for both control and
time-varying experiments in RD animals. The shaded region represents standard error.
69
Figure 3.11: Early phase of the average normalized EER strength plotted against time for both
control and time-varying pulse experiments in RD rats. The shaded regions represent standard
error.
groups (p-values < 0.05). The normalized strength of EERs generated by probe pulses in
the experimental group during the stimulation phase was significantly higher than those
delivered by probe pulses in the control group. However, there was no statistical difference
between control and experimental groups in recovery phase. The range of p-values for
the data points is listed in tables 3.1, 3.2 and 3.3.
In figure 3.11, the normalized EER strength for early phase was averaged across all
experiments and compared between control and time-varying pulse experiments in RD
rats. Both control and time-varying pulses desensitize the early phase of the response
to the same degree. Statistical analysis indicated that the average EER strength in the
early phase was not significantly different between the control and experimental groups.
The p-value range is listed in tables 3.1, 3.2 and 3.3. In figure 3.12, the normalized
EER strength for EER late phase was averaged across all experiments and compared
70
Figure 3.12: Late phase of the average normalized EER strength plotted against time for both
control and time-varying pulse experiments in RD rats. The shaded regions represent standard
error.
between control and time-varying pulse experiments. Similar to the early phase, control
and time-varying pulses desensitize the late phase to the same degree. Statistical analysis
indicated that the average EER strength in the late phase was not statistically different
between the control and experimental groups in both stimulation and recovery phases.
The p-value range is listed in tables 3.1, 3.2 and 3.3.
3.3.4 Short-Duration Pulses Limit Desensitization in Retinal Degenerate
Rats
Figures 3.13, 3.14 and 3.15 represent average data from experiments with RD rats (n=6).
In figure 3.13, the normalized signal strength is averaged across all the experiments and
is plotted against time. As is evident in figure 3.13, the strength of the EERs does not
vary as much when short duration pulses were used. The strength of the EERs recovers
to pre-stimulation level after continuous stimulation ends in both short duration and
71
Figure 3.13: Average normalized EER strength is plotted against time for both control and
short duration pulse experiments in RD animals. The shaded region represents standard error.
control experiments. Statistical analysis indicated that the average EER strength was
not significantly different in both stimulation and recovery phases between the control
and experimental groups. The range of p-values for the data points is listed in tables 3.1,
3.2 and 3.3.
In figure 3.14, the normalized EER strength for early phase was averaged across all
experiments and compared between control and short duration pulse experiments in RD
Table 3.1: P-Value Range - Average EER
Case Stimulation Phase Recovery Phase
Normalized Time-Varying 0.0006 - 0.0148 0.2061 - 0.3419
Normalized Short Duration 0.0054 - 0.0213 0.0061 - 0.0442
RD Time-Varying 0.0073 - 0.0203 0.0561 - 0.742
RD Short-Duration 0.0512 - 0.625 0.0628 - 0.812
72
Table 3.2: P-Value Range - Early EER
Case Early EER Stimulation Phase Early EER Recovery Phase
Normalized Time-Varying 0.0002 - 0.0225 0.076 - 0.184
Normalized Short Duration 0.0001 - 0.0312 0.0631 - 0.793
RD Time-Varying 0.0512 - 0.625 0.0512 - 0.625
RD Short-Duration 0.0073 - 0.0603 0.0561 - 0.742
Table 3.3: P-Value Range - Late EER
Case Late EER Stimulation Phase Late EER Recovery Phase
Normalized Time-Varying 0.0862 - 0.6518 0.0776 - 0.8936
Normalized Short Duration 0.0652 - 0.7249 0.163 - 0.8732
RD Time-Varying 0.0628 - 0.812 0.0628 - 0.812
RD Short-Duration 0.0624 - 0.745 0.0613 - 0.532
Figure 3.14: Early phase of the average normalized EER strength plotted against time for both
control and short duration pulse experiments in RD rats. The shaded regions represent standard
error.
73
Figure 3.15: Late phase of the average normalized EER strength plotted against time for both
control and short duration pulse experiments in RD rats. The shaded regions represent standard
error.
rats. Both control and short duration pulses desensitize the early phase of the response
to the same degree. Statistical analysis indicated that the average EER strength in the
early phase was not significantly different between the control and experimental groups.
The p-value range is listed in tables 3.1, 3.2 and 3.3. In figure 3.15, the normalized EER
strength for EER late phase was averaged across all experiments and compared between
control and short-duration pulse experiments. Similar to the early phase, control and
short-duration pulses desensitize the late phase to the same degree. Statistical analysis
indicated that the average EER strength in the late phase was not statistically different
between the control and experimental groups in both stimulation and recovery phases.
The p-value range listed in tables 3.1, 3.2 and 3.3.
74
3.4 Discussion
Two strategies that limit desensitization caused by continuous electrical stimulation of
retina are presented. The two strategies include time-varying pulses and short duration
pulses (< 0.1ms duration). In an in vivo model of retinal desensitization we have shown
that time-varying pulses and short duration pulses limit desensitization and thus could
be used to control precisely the temporal pattern of ganglion cell spiking. The in vivo
model of desensitization has been presented elsewhere (Ray, 2010). Briefly, the model
shows that in response to continuous electrical stimulation of the retina, a depression in
electrically evoked response strength measured from SC could be observed.
Using electrical stimulation to precisely control ganglion cell spiking is one of the
primary goals of an effective stimulation strategy. So far, two general stimulation strate-
gies have been utilized: 1) selectively activating ganglion cells without activating the
synaptic network 2) selectively activating bipolar cells without causing direct activation
of ganglion cells (Greenberg, 1998). With respect to the first approach, it is possible to
selectively activate ganglion cells without eliciting a synaptically mediated response using
epi-retinal stimulation with short duration pulses. (Fried et al., 2005) also showed that
short duration pulses could be used to replicate light-elicited spiking patterns in ganglion
cells. They shown that short-duration pulses consistently elicited spikes at frequencies
lower than 250 Hz, the upper-limit of light elicited spiking responses. This suggested
that precise temporal patterns of spiking, including those patterns elicited by light, could
be generated using programmed arrays of short pulses. (Sekirnjak et al., 2008; Weitz,
2013) have also shown that short-duration pulses could be used to suppress long latency
75
pulses and thus directly activate ganglion cells. The results of the study also indicate
that short duration pulses do not desensitize the retina as much as the pulses used in the
control group (0.5 ms, 60 A). Thus a major advantage of using short duration pulses is
that direct activation of ganglion cells could be achieved at very high rates without any
apparent desensitization. However, it has also been shown that such selective activation
is restricted to a limited range of stimulus amplitudes since increasing the amplitude to a
factor of∼ 2.5 times threshold elicits synaptically mediated spikes. Given that the thresh-
old between individual ganglion cells has been shown to differ by a factor of > 2.5, it is
unlikely that selective activation of large populations of ganglion cells is possible without
producing incidental activation of presynaptic bipolar cells. Another disadvantage of us-
ing direct activation of ganglion cells is that the ganglion cell axons are highly selective
to electric stimulation (Jensen et al., 2003). The incidental activation of passing axons
on the inner retinal surface will expand the region over which neurons are activated and
may smear the elicited percept.
The second strategy is to selectively activate the bipolar cells without activating
the ganglion cell directly. (Weitz, 2013) have shown that long pulses (> 20 ms) and
low frequency sine waves could be used to selectively activate bipolar cells and avoid
synaptically mediated responses. The advantage of this approach is thought to arise
from utilization of existing inner retinal circuitry, presumably creating spiking patterns
that better resemble those that arise under normal physiological conditions. In addition,
avoiding the activation of ganglion cell axons would provide better spatial control over
the pattern of electrical activity. (Weitz, 2013) has shown that using long pulses and low
frequency sine waves, focal stimulation of ganglion cell could be achieved. However, a
76
major disadvantage of eliciting synaptically mediated responses with a retinal prosthesis
is that there is a significant level of desensitization. Using a paired-pulse paradigm,
(Jensen and Rizzo, 2007) have shown that for interstimulus intervals less than 400 ms,
the amplitude of the RGC response to the second pulse in the pair is considerably lower
than that of the first pulse. They have also shown that this paired pulse depression is most
pronounced at very short interpulse intervals. Paired-pulse depression (PPD) has also
been observed at synapses between cones and OFF bipolar cells (DeVries and Schwartz,
1999), rod bipolar cells and AII amacrine cells (Singer and Diamond, 2006), and rods and
second order retinal neurons (ON bipolar, OFF bipolar or horizontal cells) (Rabl et al.,
2006). (Freeman and Fried, 2011) also showed that in response to continuous electrical
stimulation, a pronounced reduction in the number of spikes elicited per pulse could be
observed thus indicating that continuous electrical stimulation suppresses the sensitivity
of ganglion cells. This suppression prevents precise control over the temporal pattern of
elicited spikes.
We propose an alternative strategy for retinal prostheses that overcomes some of the
challenges discussed above (figure 3.3) The stimulation method consists of using time-
varying pulses in which every consecutive pulse has a different pulse duration and am-
plitude when compared to the preceding pulse. The duration of the pulses in the train
were between 0.1 – 20 ms. The amplitude was chosen such that it was three times the
threshold amplitude for that pulse duration. The key to this strategy is the idea that
pulses of different durations preferentially stimulate different retinal neurons of the same
class or different classes, based on field effects. As mentioned earlier, short pulses pref-
erentially activate ganglion cells directly and long pulses preferentially stimulate bipolar
77
cells and thus cause indirect activation of ganglion cells. By using a combination of short,
intermediate and long duration pulses different retinal neurons could be activated. The
assumption behind this strategy is that inner retinal neurons suffer from a prolonged in-
hibition if stimulated electrically under conditions where the surrounding network under
the electrode is being activated as a whole.
This strategy is similar in concept to the stimulation paradigm used by the light
sensitive subretinal implant developed by (Zrenner et al., 2011). Zrenner et al have
reported that in patients implanted with the subretinal implant, images are constantly
visible as a complete entity and do not fade over time as is the case with the ARGUS
II epiretinal implant. They postulate that the difference in the performance between the
two types of implants could be found in involuntary eye movements controlled by the
superior colliculus. Even during fixation, our eyes continuously make slight movements
(slow drifts and Microsaccades up to 50 min of arc and 1 to 3 Hz) that refresh the
image constantly thus changing the activated photoreceptor population – even during
strict fixation (Pritchard, 1961). The light sensitive subretinal implant consists of a
photodiode array that is activated by light. The photodiode array is attached to the
retina and it moves in synchronization with natural eye movement. Thus objects viewed
by patients with this subretinal implant, dynamically activate a range of adjacent pixels on
the array, as eye movements and Microsaccades continuously shift the “electrical image”
on the retina for about 1-3 pixels, thus preventing mechanisms of local adaptation and
image fading. Similar to the mechanism in the subretinal implant, time-varying pulses
might shift the “electrical image” on the retina by stimulating different retinal neurons
in the synaptic network and thus could aid in preventing image fading.
78
Desensitization of the Early and Late Phase Components
Similar to (Freeman and Fried, 2011), we examined the temporal response components
of the EER in order to explore whether the level of desensitization varies as a function of
response latency. A single EER was further divided into early (latency: 3∼ 12 ms) and
late (12∼ 40 ms) response. The phase segregation was based on the fact in in vitro models
of retinal desensitization, spikes generated by RGC stimulation could be classified into
two distinct clusters: early spikes that occurred within 12 ms of stimulation onset and late
spikes that occurred with latencies greater than 12 ms. In our study, when the normalized
EER strength was segregated to early and late phase components, it was observed that
early phase components underwent pronounced desensitization when compared to the late
phase components. This is in contrast with the results observed by (Freeman and Fried,
2011). (Freeman and Fried, 2011) observed that early phase components were relatively
less desensitized and responded more reliably to high pulse rates than the late phase
response. One drawback of our study is that we could not clearly determine whether the
early or late phase EER is the result of early or late activity generated in the retina or
SC neurons. Hence, we cannot clearly compare the variability in the early and late phase
components between our study and those presented by (Freeman and Fried, 2011).
79
Chapter 4
Optical Imaging of Intrinsic Signals
4.1 Background
Several functional brain mapping techniques have been developed over the past 3 decades
which have revolutionized the ability to map activity in the living brain, including
positron emission tomography (PET), functional magnetic resonance imaging (fMRI),
optical imaging, near infrared spectroscopy (NIRS) and transcranial magnetic stimula-
tion. Each modality offers distinct information about functional brain activity and has
certain advantages and limitations. In choosing a functional imaging modality for ex-
periments, it is important to consider a modality’s spatial and temporal resolution, the
etiology of the brain mapping signal, the practicality of the imaging methodology, and
the cost of implementation.
Intrinsic signal imaging (ISI) maps the brain by measuring changes in tissue re-
flectance that are based on intrinsic activity. Physiological changes such as changes
in blood oxymetry, and blood volume increase cause changes in tissue reflectance. These
reflectance changes are utilized in mapping functional brain activity. This method offers
80
a distinct advantage over extrinsic signal imaging, such as dye imaging. Dye imaging
causes toxicity in in vivo preparations and thus alters the physiology of the sample. ISI
does not require any contact with the tissue and thus is suited for chronic studies and
also during neurosurgery in humans when mapping the cortex (Mazziotta et al., 2000).
Activity-related optical changes in tissue reflectance that are associated with electrical
activity or metabolism were first observed by Hill and Keyes (1949) over 50 years. How-
ever, until 1980s intrinsic signals have not been used to map cortical activity (Grinvald
et al., 1986). Optical imaging of intrinsic signals is becoming increasingly popular because
the technique offers both high spatial and temporal resolution. The spatial resolution of
ISI is unparalleled among in vivo imaging techniques (on the order of micrometers), mak-
ing it ideal for studying the fine functional organization of sensory cortices as well as
the physiology of neurovascular coupling at the level of the arteriole, venule, and even
capillaries. Although the temporal resolution of intrinsic imaging is not as great as with
electrophysiological techniques, imaging is commonly performed at video frame rates
(30-75 Hz). This is more than sufficient for imaging the slowly evolving perfusion-related
responses, which peak 3 to 4s after stimulus onset.
4.2 Sources of Intrinsic Signals and Wavelength Dependency
Optical imaging of intrinsic signals does not directly measure neuronal activity Villringer
and Dirnagl (1995). Instead, the activity-related changes in perfusion and metabolism
are detected. Therefore, it is required to understand the hemodynamic response and its
relationship to electrophysiology in order to understand the significance of the results
81
being reported. Moreover, it is critical to understand which aspects of the hemodynamic
response underlie intrinsic signals.
Neuronal activity results in changes in local perfusion and metabolic activity (Roy
and Sherrington, 1890). Increase in blood flow results in the supply of essential nutrients
such as oxygen and glucose to areas that are metabolically active. Other events that
are associated with increased blood flow include regional vasodilation (Ngai et al., 1988),
blood flow changes (Cox et al., 1993; Lindauer et al., 1993), blood volume increases
(Belliveau et al., 1991; Frostig et al., 1990; Narayan et al., 1995), and changes in relative
hemoglobin concentrations (Kwong et al., 1992; LaManna et al., 1987; Malonek and
Grinvald, 1996; Mayhew et al., 1999; Nemoto et al., 1999). Metabolic changes include
increases in local oxygen consumption (Frostig et al., 1990; Malonek and Grinvald, 1996;
Vanzetta and Grinvald, 1999) and glucose utilization (Fox et al., 1988; Sokoloff et al.,
1977). All of these electrophysiological and metabolic changes may contribute to intrinsic
optical changes.
Several studies have shown close spatial connection between neuronal activity and
perfusion-related mapping signals (Villringer and Dirnagl, 1995). In most cases, the
region where intrinsic signals are observed is wider and overspills regions of electrophysi-
ological activity. This phenomenon of “spread” mostly has been observed in the rodent
somatosensory cortex in response to whisker stimulation (Chen-Bee and Frostig, 1996;
Godde et al., 1995; Masino and Frostig, 1996). (Narayan et al., 1995) reported that
intrinsic optical and intravascular fluorescent dye maps overspilled regions of electrophys-
iological activity by about 20%. Optical maps encompass not only the principal barrel
82
but also the adjacent barrels. This may be due to low-level neuronal activity, which
occurs in adjacent barrels in response to stimulation of adjacent, a nonprincipal whisker.
Three major components of optical signals have been identified: blood volume changes,
hemoglobin oxymetry changes, and light scattering (Frostig et al., 1990; Malonek and
Grinvald, 1996; Mayhew et al., 1999; Nemoto et al., 1999).
4.2.1 The Blood Volume Component
Blood volume is the first component of intrinsic signals. This signal originates from
changes in blood volume that is associated with either vasodilation or recruitment of
local capillaries or both in cortically active areas. In order to determine the extent to
which blood volume component contributes to intrinsic signals, (Frostig et al., 1990)
injected a fluorescent dye in rodent cortex. It was observed that the dye was restricted to
the intravascular compartment. When the dye was imaged and corresponding signal was
compared with the intrinsic signal measured at 570nm, both signals were identical. This
indicated that there was increased absorption when the blood volume increased. This led
the authors to conclude that at 570nm, the major component of intrinsic signals primarily
originates from blood volume changes. The study found that the blood volume changes,
however, extend beyond the electrophysiologically active cortical region, suggesting that
the majority of activity-dependent blood volume changes and related intrinsic signals are
not specific to or finely regulated at the level of individual functional domains.
83
4.2.2 The Hemoglobin Oxymetry Component
The second component of the intrinsic signals originates from activity-dependent changes
in hemoglobin oxygen saturation. Several studies have shown that at the site of neuronal
activity, there is a sudden increase in oxygen consumption following neuronal activation
(Frostig et al., 1990; Vanzetta and Grinvald, 1999). This sudden increase in oxygen
consumption is believed to result in an increase in deoxyhemoglobin concentration. This
increase in deoxyhemoglobin concentration is associated with electrophysiologically active
neurons. Certain studies using phosphorescence-quenching dyes have also demonstrated
that functional activation is followed by a brief decrease in oxygen concentration (Vanzetta
and Grinvald, 1999). However, not all groups have observed these initial increases in
oxygen consumption and deoxyhemoglobin concentrations (Kohl et al., 2000; Lindauer
et al., 2001; Vanzetta and Grinvald, 1999).
Intrinsic signals that are measured in the 600 to 630 nm range have two phases that
correspond to deoxyhemoglobin concentration changes: first a decrease in absorption
and then an increase in absorption (from increasing concentration) followed by a more
prolonged decrease in absorption (Figure 4.1). The biphasic phase observed in this wave-
length range is predominately the result of absorption of light by deoxyhemoglobin when
compared to oxyhemoglobin. Other processes such as light scattering and changes in total
hemoglobin also contribute to intrinsic signals at these wavelengths. However, the major
contributor to intrinsic signals in this wavelength range is the change in deoxyhemoglobin
concentration.
84
Figure 4.1: The biphasic timecourse of optical responses are shown in the figure. Following
stimulation, there is initially a focal increase in absorption (at 1 s, interpreted as a decrease in
deoxyhemoglobin and oxygen extraction) followed by a more widespread decrease in absorption
(3.5 – 6.5 s, interpreted as a decrease in deoxyhemoglobin concentration). The second phase of
the signal is related to the BOLD fMRI signal. At 850 nm, neither isoform of hemoglobin absorbs
much light. Instead, the signal is believed to originate from light scattering changes. This signal,
likes at 550 nm, is monophasic but is significantly less intense than the 550 nm response (Vanzetta
and Grinvald, 1999).
4.2.3 Light Scattering Component
The third component of intrinsic signals is changes from light scattering. Light scatter-
ing changes arise from ion and water flux; morphological changes such as expansion or
contraction of vasculature; and blood volume changes (Lou et al., 1987; Narayan et al.,
1995). Light scattering component, although present at all wavelengths, becomes a ma-
jor source of intrinsic signals at wavelength above 630 nm (i.e. at wavelengths with
small hemoglobin absorption). In fact, the light scattering component dominates the
signal in the near infrared region (wavelengths above 750 nm). (Narayan et al., 1995)
demonstrated that at 850 nm, changes in light scattering correlate both spatially and
temporally with changes in blood volume. This is consistent with the fact that erythro-
cytes are the primary scatterers in whole blood; light scattering changes can be induced
either by functional changes in the number of erythrocytes or by functional erythrocytic
85
distortions (Zdrojkowski and Pisharoty, 1970). Light scattering blurs images and expands
area of activity thus introducing a potential confound to intrinsic signal mapping studies
(Orbach and Cohen, 1983).
4.2.4 Resolving the Different Components of the Intrinsic Signals
Oxyhemoglobin and deoxyhemoglobin are the major chromophores that influence intrinsic
signal changes. Both of these chromophores have a unique absorption spectrum. Two dis-
tinct approaches have been developed to determine the etiology of intrinsic signals based
on the differential absorption of light by these two chromophores. The first approach is
referred to as “single-wavelength imaging” and it emphasizes the different physiological
process by imaging at specific wavelengths of light. At such specific wavelengths of light,
the differences between the chromophores are either attenuated or minimized. The second
approach is referred to as “spectroscopic imaging”. Using this approach, multiple wave-
lengths are used to simultaneously image the region of interest. Spectroscopic analysis
is performed in order to determine the contribution of different components of intrinsic
signals.
Single-Wavelength Imaging
Oxyhemoglobin and deoxyhemoglobin absorption spectra are used to select wavelengths
for single wavelength imaging. Thus the time course and spatial distribution of the re-
sponses are wavelength dependent (Figure 4.2) (Hodge et al., 1997; Nemoto et al., 1999).
For example, at 610 nm, the absorption of oxyhemoglobin is negligible compared to that
86
of deoxyhemoglobin (Figure 4.2). Therefore, changes in deoxyhemoglobin are empha-
sized at 610nm (Frostig et al., 1990; Nemoto et al., 1999). Consistent with blood oxygen
level-dependent (BOLD) fMRI studies (Hu et al., 1997; Logothetis et al., 1999; Yacoub
et al., 1999) and oxygen-dependent phosphorescene-quenching dye studies (Vanzetta and
Grinvald, 1999), optical imaging at 610 nm indicates an initial focal increase in deoxy-
hemoglobin (analogous to the “initial dip”) followed by a more global decrease in de-
oxyhemoglobin (Figure 4.1). In contrast, imaging at 550 or 570 nm (isobestic points of
hemoglobin), should emphasize changes in total hemoglobin, since deoxyhemoglobin and
oxyhemoglobin absorb equally at these wavelengths. Accordingly, the time course and
spatial extent of the response at these wavelengths is significantly different from that
observed at 610 nm. Finally, imaging at 850 nm, which is near infrared and at which the
absorption of both hemoglobin chromophores drops dramatically, is dominated by light
scattering changes. Studies comparing intrinsic and intravascular dye suggest that these
light scattering changes correlate well with changes in cerebral blood volume and may
not be directly influenced by hemoglobin levels and oxygenation (Figure 4.2)(Cannestra
et al., 1998; Narayan et al., 1995; OFarrell et al., 2000).
At any particular wavelength, the signal is composed of multiple factors; however, cer-
tain components are emphasized at a particular wavelength. Also, different wavelengths
of light penetrate the cortex to different degrees (longer wavelengths penetrate deeper into
the cortex). Hence, one should exercise caution when interpreting results across wave-
lengths since different cortical volumes may be sampled at different wavelengths (Mayhew
et al., 1999).
87
Figure 4.2: Hemoglobin absorption curves. The absorption curves of oxy- and deoxyhemoglobin
are displayed to demonstrate the differential absorption of light by the primary moieties of
hemoglobin at different wavelengths (Nemoto et al., 1999).
Although intrinsic signals at various wavelengths originate from different sources, they
can all be used for functional mapping. These differences in intrinsic signals across wave-
lengths may be exploited depending on the scientific question at hand. To characterize
the functional architecture of, or “map” the brain, it may be more appropriate to image
at 610 nm. However, to characterize blood volume changes, it would be better to image
at 550 or 570 nm.
Spectroscopic Analysis of Intrinsic Signals
Spectroscopic analysis of intrinsic signals provides more specific determination of the
etiology of intrinsic signals. As part of the analysis, first a spatiospectral image is acquired
from many locations. A spectrophotometer is then used to decompose the image. The
optical imaging apparatus is modified to have two imaging planes in order to retain at
least one spatial dimension from the imaged area. In the first image plane, a narrow slit
isolates a selected line across the cortical surface. This cortical slit (Figure 4.3) is then
projected onto a dispersing grating (in the spectrophotometer), whose grooves are aligned
88
Figure 4.3: The “cortical band” and optical spectroscopy. (A) Image of the cortical surface
with the position of the cortical slit from which spectroscopic data are collected. (B) Image of
the spatiospectral pattern obtained after the gratings in the spectrophotometer disperse the light
transmitted through the cortical slit. The y axis corresponds to the position down the slit, shown
in (A). The x axis in the spatiospectral image corresponds to the wavelength. The absorption
spectra of oxy- and deoxyhemoglobin are superimposed on the image (oxy – solid line, deoxy –
dotted line). The increases in absorption (darker vertical regions) are primarily due to an increase
in the proportion of oxyhemoglobin. The predominant dark horizontal bands reflect the changes
in blood volume in the middle cerebral artery (Mayhew et al., 1999).
parallel to the slit, so that the image of the selected cortical band is dispersed into its
spectral components along the orthogonal axis. Thus, in the second image plane, a two
dimensional image is produced, with one spatial dimension and one spectral dimension,
showing the spectral information about multiple cortical points simultaneously (Figure
4.3). The detector is placed in the second image plane to capture this spatiospectral
image.
Alternatively, in order to retain two dimensions, a filter wheel can be used to acquire
multiwavelength data simultaneously. By interleaving images at different wavelengths
throughout a single trial and by imaging at different wavelengths at the same time point
across different trials, a four-dimensional volume set can be acquired (two spatial dimen-
sions, time, and wavelengths), which can be used for subsequent spectroscopic analysis.
89
The limitation of this approach is that the multiwavelength/spectral data being collected
are not simultaneous. It is also well known that intrinsic signal responses can vary greatly
across trials (Chen-Bee and Frostig, 1996; Masino and Frostig, 1996). This limitation can
be statistically overcome by acquiring a large number of trials so that the variability of
the response is accounted for.
Once spectral data have been obtained, Beer-Lambert Law can be used to extract the
combination of the different chromophores to the intrinsic signal. The spectral changes
are fitted by the known spectra of the oxyhemoglobin and deoxyhemoglobin as well as
cytochromes and light scattering components. The first such study applied a linear com-
ponent analysis to extract the etiology of the intrinsic signals (Malonek and Grinvald,
1996).
4.2.5 Time Course of Intrinsic Signals
As mentioned earlier, intrinsic signal changes at different wavelengths have different time
courses. There are two major patterns of intrinsic signal time courses: monophasic time
course that is characterized by a decrease in reflectance and biphasic timecourse that is
characterized by an initial decrease in reflectance followed by an increase in reflectance.
Up to 590 nm in the visible spectrum, a monophasic pattern is observed in reflectance
changes. A biphasic pattern is observed between 600 and 760 nm. At wavelengths greater
than 760 nm, a monophasic pattern that is similar to that observed at wavelengths less
than 590 nm is observed.
For monophasic responses, responses appear approximately 1 s after stimulus onset,
reach peak between 3 and 4 s after stimulus onset, and return to baseline by approximately
90
8 s (Nemoto et al., 1999). For the biphasic responses, the initial response to stimulation
is usually a little more rapid, appearing within 500 ms, peaking between 1.5 and 2 s
after stimulus onset, reversing phases at about 3 s, peaking in the opposite polarity
approximately 5 s after stimulus onset, and returning to baseline by approximately 10
s. The time course is generally consistent with the “initial dip” in deoxyhemoglobin
concentrations (Cannestra et al., 2001; Malonek and Grinvald, 1996; Mayhew et al., 1999;
Menon et al., 1995; Nemoto et al., 1999).
The time courses of the individual components that contribute to intrinsic signal
changes have also been characterized. Deoxyhemoglobin concentrations increase imme-
diately after stimulus onset, peak between 1 and 2 s after stimulus onset, decrease and
peak between 4 and 6 s after stimulus onset, and return to baseline concentrations.
Oxyhemoglobin concentrations on the other hand are slower to respond to stimulation,
increasing approximately 1 s after stimulus onset, peaking approximately 4 s after stimu-
lus onset, and returning to baseline (monophasic) (Malonek and Grinvald, 1996; Mayhew
et al., 1999; Nemoto et al., 1999).
4.3 Studies on Visual Cortex Using Intrinsic Signal Imaging
The functional architecture of visual cortex has been extensively studied using electro-
physiological techniques. Several cortical features such as orientation preference, ocu-
lar dominance columns, maps of preferential direction of motion and spatial frequency
columns have been identified using electrophysiological techniques. However, ISI has
enabled researchers to establish the accurate layout of these cortical features and also
91
to study the interrelationships between the aforementioned cortical features (Bonhoeffer
and Grinvald, 1991; Grinvald et al., 1986, 1991; Hubener et al., 1997; White et al., 2001).
The main advantage of using ISI is that it allowed visualization of all the aforementioned
features in the same subject at the same imaging time (Bosking et al., 2002; Kim et al.,
1999; Weliky et al., 1996). This in turn facilitated establishing geometric relationships
between different columnar structures. One such important finding came from (Hubener
et al., 1997), where he found that most columnar systems intersect at right angles more
frequently than would be expected in a random arrangement.
ISI has made it possible to visualize the layout of orientation preference and ocular
dominance maps within areas 17 and 18 in a number of species including cat, ferret,
macaque, tree shrew, barn owl and marmoset monkey (Bosking et al., 1997; Grinvald
et al., 1991; Issa et al., 1999; Liu and Pettigrew, 2003; Shtoyerman et al., 2000). It is
also possible to visualize maps of direction of motion preference (Shmuel and Grinvald,
1996; Weliky et al., 1996), hue-selective organization within V2 (Xiao et al., 2003) and
clusters of color selective neurons in V1 (Landisman and Tso, 2002). Recently, retinotopic
maps in macaque monkey V1, owl monkey V3, mouse, tree shrew and cat have also been
reported (Blasdel and Campbell, 2001; Bosking et al., 2000; Kalatsky and Stryker, 2003;
Zepeda et al., 2003).
4.4 Intrinsic Signal Imaging and Superior Colliculus
In rodents, majority of the retinal ganglion cells terminate in Superior Colliculus (SC)
and thus SC is believed to be involved in processing visual information. The presynaptic
92
input to the superficial layers of SC comes from retinal ganglion cells of the contralateral
eye and these cells in turn project to deeper layers of the SC. Due to the direct connection
between the superficial layers of SC and retinal ganglion cells, this region of SC is of wide
interest. The superficial layer of SC has been shown to have a precise retinotopic map
that is consistent across rat strains with the densest retinal projections found in a layer
called stratum griseum superficiale (SGS). Presence of such a retinotopic projection could
be useful in the study of the response of retinal ganglion cells to electrical stimulation.
Electrophysiology has been used as the most common methodology to study retinal output
from SC. However, electrophysiology is invasive and requires positioning electrodes in the
brain that potentially cause damage. Instead, optical imaging of intrinsic signals could be
used as a methodology to study retinal output. Functional maps of SC could be generated
using either electrical or visual stimulation of the retina. The following sections illustrate
the intrinsic signal imaging methodology that was followed in order to create functional
maps of SC. The goal of these experiments is to demonstrate that focal region of SC could
be imaged by electrical stimulation of the retina.
4.4.1 Anesthesia and Monitoring Physiological Parameters
Anesthesia has a strong effect on the coupling between cerebral blood flow and neu-
ronal activity. Therefore, anesthesia must be chosen with great care and levels must
be monitored very closely. The level of anesthesia is especially important since a ma-
jor component of intrinsic signals is related to hemoglobin oxymetry changes. Excessive
anesthesia may adversely alter the ventilation and therefore relative hemoglobin concen-
trations of the subject. (Shtoyerman et al., 2000) demonstrated that if anesthesia is well
93
controlled, intrinsic signals under general anesthesia will have time courses similar to
those observed in awake, behaving animals, although smaller in magnitude. Barbiturates
and gas anesthetics (sevoflurane or isoflurane) work well for optical imaging.
Optical intrinsic signals arise primarily from changes in cerebral blood flow, cerebral
blood volume, hemoglobin oxygenation and tissue scattering. It is therefore crucial to
monitor physiological parameters that may alter cerebral perfusion and metabolism inde-
pendent of functional activity to ensure that any observed changes are not due to changes
in systemic variables. A pulse oxymeter allows noninvasive monitoring of the oxygen sat-
uration and heart rate of the subject. Monitoring the heart rate is critical since a drop
in heart rate will result in a decreased cardiac output and may affect cerebral perfusion.
Similar to monitoring heart rate, it is important to ensure that blood pressure does not
change since this may also impact cerebral blood flow and volume. Finally, it is critical
to monitor core body temperature since thermoregulation is one of the first hemostatic
mechanisms to be compromised during anesthesia. Changes in core body temperature
can significantly alter peripheral vascular resistance and systemic perfusion. Animals
that are not adequately thermo-regulated will frequently fail to produce any optical sig-
nal changes. A self-regulating heating blanket, which monitors the subject’s core body
temperature and adjusts the temperature of the blanket, is highly recommended.
Optical intrinsic signals arise primarily from changes in cerebral blood flow, cerebral
blood volume, hemoglobin oxygenation and tissue scattering. It is therefore crucial to
monitor physiological parameters that may alter cerebral perfusion and metabolism inde-
pendent of functional activity to ensure that any observed changes are not due to changes
94
in systemic variables. A pulse oxymeter allows noninvasive monitoring of the oxygen sat-
uration and heart rate of the subject. Monitoring the heart rate is critical since a drop
in heart rate will result in a decreased cardiac output and may affect cerebral perfusion.
Similar to monitoring heart rate, it is important to ensure that blood pressure does not
change since this may also impact cerebral blood flow and volume. Finally, it is critical
to monitor core body temperature since thermoregulation is one of the first hemostatic
mechanisms to be compromised during anesthesia. Changes in core body temperature
can significantly alter peripheral vascular resistance and systemic perfusion. Animals
that are not adequately thermoregulated will frequently fail to produce any optical sig-
nal changes. A self-regulating heating blanket, which monitors the subject’s core body
temperature and adjusts the temperature of the blanket, is highly recommended.
4.4.2 Immobilization of the Superior Colliculus
To take advantage of the superior spatial resolution of optical imaging, it is critical that
images acquired before and during activation be in identical locations, since the former
will be subtracted from the latter during analysis. The movement of the brain due to
respiration and heartbeat presents a major obstacle. Movement of the brain would present
as low frequency physiological noise and degrades the quality of the intrinsic signal. The
spectral content of light reflectance from a single pixel was plotted in figure 4.4 to illustrate
the presence of low frequency physiological signal. This physiological signal has much
higher amplitude than the intrinsic signal and mars the signal. In order to overcome this
challenge, the superior colliculus is immobilized using 2% agarose solution. Figure 4.5
95
Figure 4.4: Spectral content of light reflectance from a single pixel. The circled region indicates
low frequency physiological signal.
shows the spectral content of a single pixel after superior colliculus has been immobilized
using agarose.
4.4.3 Slow-Scan CCD Cameras
Slow-scan digital CCD (charge-coupled device) cameras offer very good SNR while re-
taining the advantages of high spatial resolution and moderate cost. Several important
parameters may influence image quality, SNR, and acquisition capabilities should be
considered when comparing CCD cameras including shot noise, well capacity etc. The
magnitude of intrinsic signal changes is exceedingly small, on the order of 0.1%. In order
to be able to ascribe statistical and biological significance to such measurements, one
must be able to differentiate these intrinsic signal changes from stochastic fluctuations
in photon emissions. That is, one has to ensure that small changes in reflectance are
96
Figure 4.5: Spectral content of light reflectance from a single pixel illustrating the absence from
low frequency physiological noise. Agarose immobilizes the superior colliculus and eliminates this
noise.
physiological in origin and not caused by statistical fluctuations of the light-emitting pro-
cess. The number of photons that can be attributed to statistical fluctuations equals the
square root of the total number of photons emitted.
The well capacity of a CCD denotes the total number of photons that can be accu-
mulated on 1 pixel before there is charge overflow or saturation. It is important that
well capacities be as large as possible. Well capacities generally range from 300,000 to
700,000; smaller well capacities can also be used but will require a greater number of
trials to be able to detect similar small intrinsic signal changes. Binning could also be
used to increase the effective well capacity. In our study, we have used a CCD camera
from Adimec (model: 1000M) that meets the shot noise and well capacity requirements.
97
Figure 4.6: Images of SC captured using a macroscope. Parts of SC are blurred in A and the
entire SC is blurred when magnified in B.
4.4.4 Camera Lenses/Macroscope and Camera Mount
Traditionally intrinsic signal imaging (ISI) is performed using tandem lens configuration
termed as a macroscope. A macroscope is essentially a microscope, with a low magni-
fication, composed of two “front-to-front” high numerical aperture photographic lenses.
By doing this, the macroscope provides an unusually high numerical aperture compared
to commercial, low-magnification microscope objectives. However, the depth of field is
quite shallow. SC is a curved structure and hence a macroscope blurs the curved regions
of SC. Figure 4.6 illustrates images of SC captured using a macroscope constructed from
two 50mm photographic lenses.
In order to overcome the issue with shallow depth of field, a zoom lens was used to
image SC (Navitar 12X Ultrazoom). This lens provided the required depth of field and
high numerical aperture. Figure 4.7 depicts the images of SC captured with the new lens.
The video camera should be rigidly mounted to a vibration-free support. The ideal
arrangement is to mount the camera to an immobile support system so that the camera
98
Figure 4.7: SC image captured using a zoom lens from Navitar
has no means of moving. This is especially important because many modern imaging
devices have built in cooling devices that may vibrate the camera. By placing the camera
on a firm support, this source of vibration can be eliminated.
4.4.5 Data Acquisition
Experimental Setup
The basic experimental setup is shown in figure 4.8. First, the subject’s head must be
stabilized using a stereotaxic frame. Once SC has been exposed, the brain is illuminated
with light either from LEDs or using light guides. The camera should be positioned over
the SC exposure such that the area of interest is centered within the field of view of the
camera. Images of SC are taken at rest and during activation. The camera digitizes the
images and forwards the data to the data acquisition computer.
99
Figure 4.8: Schematic diagram of the optical imaging of intrinsic signals setup (Vanzetta and
Grinvald, 1999).
Timing and Duration of a Single Data Acquisition
As mentioned earlier, optical imaging of intrinsic signals offers the best combination of
temporal and spatial resolution compared to other brain mapping methods. It is advisable
to design experiments such that a time course can be reconstructed from the data. Since
the time course of intrinsic signals are on the order of seconds, it is often sufficient for
frames to be collected every 250 – 500 ms in order to capture the temporal profile of the
response. However, in the SC imaging experiments, images were captured at a higher
frame rate (50 Hz) in order to construct the temporal profile precisely.
When should imaging begin relative to stimulus onset? Although blood volume
changes do not occur until approximately 1s after stimulus onset, hemoglobin oxymetry
changes may begin within a few hundred milliseconds of neuronal activation (Malonek
and Grinvald, 1996; Mayhew et al., 1999; Nemoto et al., 1999). It is therefore essential
100
that image acquisition begin, at the very latest, simultaneous with stimulus onset. Anal-
ysis of optical images, however, often requires at least one baseline image (i.e., prior to
stimulus onset) in order to determine percentage change in signal from baseline. Some
methodological approaches require an even greater number of baseline frames in order
to be able to estimate and subtract out baseline vascular oscillatory signals. In the SC
imaging experiments, 3 s of baseline images were collected prior to stimulus onset.
How long should images be collected after stimulus onset? Most of the components
that contribute to intrinsic signals return to baseline between 8s and 12 s after stimulus
onset. Therefore, in order to capture the entire temporal profile of the intrinsic signal
response, one should image for approximately 15 s after stimulus onset. The intrinsic
signal response, however, generally peaks within 5 s after stimulus onset. When imaging
SC, images were collected for 12 s after stimulus onset.
Experimental Protocol
The experimental protocol used to demonstrate that a focal region of SC could be imaged
by electrically stimulating the retina is shown in figure 4.9. Background images were
collected for 3 s and after 3 s the retina was electrically stimulated for 4 s. Stimulus
pulses were delivered at 20 Hz. The pulses were cathodic first biphasic pulses and the
duration of each phase was 0.5 ms with 0.1 ms interphase gap. The frame rate was 50
Hz and the size of each frame was 1024* s1024. Each frame was spatially binned (3* 3
binning). While the retina was being stimulated, the camera acquired SC images. After
the end of the electrical pulse train, the camera continued to collect images of SC for
another 8 s. Thus the duration of one trial was 15 s and it was repeated 30 times.
101
Figure 4.9: Protocol for recording a focal SC response to electrical stimulation with ISI.
Figure 4.10: The image on the left shows focal stimulation of SC. The circled area shows SC
response. It has been zoomed and represented in the image on the right. X and Y axes in both
the images represent pixel location.
4.4.6 Data Analysis
In order to obtain activity maps from the SC, images have to be acquired both while the
SC is stimulated and at rest. A “reference image” is then subtracted from and divided
into all subsequent images on a pixel-by-pixel basis in order to determine the percentage
change in reflectance at each pixel at each time point. The reference image is divided
into the difference images in order to normalize for uneven illumination. Using a SC
image prior to stimulus onset as a reference image makes the fewest assumptions about
the functional architecture and physiology of the brain. The advantage of a picture of
the inactive SC is that no assumption is made about the complete set of stimuli that are
required to activate the SC uniformly. The disadvantage of using the blank picture is
102
that it can cause very strong, activity-related blood vessel artifacts in the maps. These
vascular artifacts often overwhelm the fine mapping details.
4.4.7 Single Condition Maps
Single condition maps are calculated by taking the activity map obtained with one par-
ticular stimulus and dividing this image by the reference image. The resulting map then
shows the activation that this particular stimulus causes. In the SC imaging experiments,
single condition maps generated by electrical stimulation of the retina were obtained. The
reference image used in generating these single condition maps is an image of non-active
SC. The stimulus used to generate these maps was electrical stimulation of the retina.
Figure 4.10 is a single condition map generated by stimulating retina.
103
Chapter 5
Conclusions and Future Work
The work presented in this thesis has investigated a number of factors that may contribute
to the efficiency of epiretinal stimulation. In particular, I compared the effect of voltage-
controlled and current-controlled waveforms on the strength of neural response evoked
in the superior colliculus. In addition to studying the effect of waveform shape, I also
developed new stimulation strategies for epiretinal prosthesis that address the issue of
retinal desensitization caused by continuous electrical stimulation of the retina.
5.1 Recommendations for Epiretinal Prostheses
5.1.1 Limiting Desensitization Caused by Continuous Electrical Stimulation
of the Retina
My studies were motivated by the recent finding that in subjects with epiretinal prosthe-
sis, continuous electrical stimulation of the retina causes the brightness of the phosphenes
to quickly reduce over hundreds of milliseconds. Also, in response to stimulation lasting
60 s, several subjects reported a gradual reduction in brightness that decreased for the
104
duration of the stimulus. This gradual decrease in phosphene brightness is termed “de-
sensitization”. (Ray, 2010) generated an in vivo model of desensitization in which the
strength of the evoked potential response to electrical stimulation gradually decreased
over time. The time course of the decrease in evoked potential strength was similar to
that observed in prosthesis subjects. I have also found that using 0.5 ms pulses, close
to the pulse duration used by the prosthesis, desensitization persists as can be seen in
the decline of the evoked potential response. I then focused on developing strategies that
would limit desensitization caused by continuous electrical stimulation of the retina.
Using electrical stimulation to precisely control ganglion cell spiking is one of the pri-
mary goals of an effective stimulation strategy. So far, two general stimulation strategies
have been utilized: 1) selectively activating ganglion cells without activating the synaptic
network 2) selectively activating bipolar cells without causing direct activation of ganglion
cells (Greenberg, 1998). With respect to the first approach, it is possible to selectively
activate ganglion cells without eliciting a synaptically mediated response using epi-retinal
stimulation with short duration pulses. A major advantage of using short duration pulses
is that direct activation of ganglion cells could be achieved at very high rates without any
apparent desensitization. However, it has also been shown that such selective activation
is restricted to a limited range of stimulus amplitudes since increasing the amplitude to a
factor of∼ 2.5 times threshold elicits synaptically mediated spikes (Weitz, 2013). Given
that the threshold between individual ganglion cells has been shown to differ by a factor
of > 2.5, it is unlikely that selective activation of large populations of ganglion cells is
possible without producing incidental activation of presynaptic bipolar cells. Another
disadvantage of using direct activation of ganglion cells is that the ganglion cell axons are
105
highly selective to electric stimulation (Jensen et al., 2003). The incidental activation of
passing axons on the inner retinal surface will expand the region over which neurons are
activated and may smear the elicited percept.
The second strategy is to selectively activate the bipolar cells without activating the
ganglion cell directly. (Weitz, 2013) have shown that long pulses (> 20 ms) and low fre-
quency sine waves could be used to selectively activate bipolar cells and avoid synaptically
mediated responses. The advantage of this approach is though to arise from utilization
of existing inner retinal circuitry, presumably creating spiking patterns that better re-
semble those that arise under normal physiological conditions. In addition, avoiding the
activation of ganglion cell axons would provide better spatial control over the pattern
of electrical activity. However, a major disadvantage of eliciting synaptically mediated
responses with a retinal prosthesis is that there is a significant level of desensitization.
I propose an alternative strategy for retinal prostheses that overcomes some of the
challenges discusses above. The stimulation method consists of using time-varying pulses
in which every consecutive pulse has a different pulse duration and amplitude when
compared to the preceding pulse. The pulse duration of the pulses in the train that
we test were between 0.1 – 20 ms. The amplitude was chosen such that it was three
times the threshold amplitude for that pulse duration. The key to this strategy is the
idea that pulses of different durations preferentially stimulate different retinal neurons.
As mentioned earlier, short pulses preferentially activate ganglion cells directly and long
pulses preferentially stimulate bipolar cells and thus cause indirect activation of ganglion
cells. By using a combination of short, intermediate and long duration pulses different
retinal neurons could be activated.
106
Using time-varying pulse trains, I have demonstrated that it is possible to limit reti-
nal desensitization caused by continuous electrical stimulation. However, it should be
noted that it is difficult to recommend and optimal stimulus waveform without testing
each stimulus pattern in humans. The in vivo model of retinal desensitization is a good
indicator of the temporal characteristics of evoked responses in superior colliculus. How-
ever, this model doesn’t account for the spatial characteristics of phosphene perception.
Therefore, human testing is warranted to determine which stimulus pattern produces the
highest quality vision.
5.1.2 Effect of Waveform Shape on Stimulation Efficiency
A recent report of a retinal prosthesis clinical trial suggests a need to improve stimulation
efficiency. Across 30 patients with retinal implants, only 55% of electrodes could evoke
visual percepts using stimulus below 1 mC/cm
2
, which was the charge density limit of
Platinum-gray, the electrode material used in the implants (Humayun et al., 2012; Sanders
et al., 2007). To increase the number of functional electrodes, more efficient methods of
stimulation must be developed, so that visual percepts can be generated within charge
density limits. Thus, improving the stimulus efficiency of retinal implants is an important
goal with implications for current and future devices.
By comparing voltage-controlled and current-controlled stimulation techniques, I demon-
strated that pulse waveform shape has an effect on stimulation efficiency. I demonstrated
that pulse duration determines whether voltage or current stimulation is more efficient for
activating the retina. These differences are related to the current waveform that is gen-
erated during these two modes and how that current influences the biophysical response
107
properties I also performed a computational study in which I showed how membrane
parameters vary with pulse width and waveform shape. The model will be useful tool
in evaluating the efficiency of other non-conventional waveforms for retinal stimulation.
I also demonstrated that using a high surface area Pt-Ir electrode when compared to
the standard Platinum-gray electrodes used by epiretinal prosthesis could reduce power
consumption.
5.2 Future Experiments
My experiments have shown that time-varying could be used to limit retinal desensi-
tization in the in vivo model; however, some questions remain unanswered. First, the
mechanism by which desensitization occurs remains unclear. Several mechanisms have
been suggested including desensitization of post-synaptic receptors, activation of presy-
naptic receptors, depletion of the readily releasable pool of synaptic vesicles or a change
in the functional state of the exocytotic machinery (Kirischuk et al., 2002; Waldeck et al.,
2000; Zucker and Regehr, 2002). The in vivo model of desensitization is useful in studying
the effect of various stimulation strategies; however, it would be difficult to study underly-
ing mechanisms using this model. It might be easier to study desensitization mechanisms
in an in vitro model since for example, experiments that require using synaptic block-
ers can be easier to perform in an in vitro model. Several studies have suggested that
long-latency responses of RGC’s undergo desensitization when compared to short-latency
responses. In the model that I worked with, it was not quite clear if the responses could
be categorized as short-latency vs. long-latency. It would be interesting to determine if
108
the strength of the evoked responses follow a bimodal distribution. If a bimodal distri-
bution were not observed, then an in vitro model would be better suitable in order to
characterize the response as short latency vs. long latency.
I have also developed a technique to image intrinsic signals in superior colliculus.
Optically imaging of intrinsic signals is a useful tool in studying cortical plasticity. So far,
only acute experiments have been performed to study the effect of continuous electrical
stimulation on the retina. Using optical imaging of intrinsic signals, chronic experiments
could be performed to determine the long-term effects of electrical stimulation on cortical
plasticity.
109
Chapter 6
Methods
6.1 Animals
Normally sighted Long Evans, heterozygous S334ter line 3 [P660-P680] rats and homozy-
gous S334ter line 3 [P480] were used. The heterozygous and homozygous S334ter line 3
rats are referred to as RD (retinal degeneration) rats. Animals were housed in covered
cages and fed a standard rodent diet ad libitum while kept on a 12:12-hour light-dark
cycle animal facility. RD rats were bred in the facility by mating homozygous S334ter
line 3 rats with Copenhagen rats (Charles River, Hollister, CA). Dr. Matthew LaVail,
University of California San Francisco, supplied the homozygous S334ter line 3 breeding
pairs. Since the mutation is dominant, all offspring had one copy of the mutated gene.
All experimental procedures were approved by the Institutional Animal Care and Use
Committee (IACUC) at the University of Southern California.
110
6.2 Stimulation Electrode
The stimulation electrode was a concentric bipolar Pt-Ir electrode (model CBDFG74,
FHC, Bowdoin, ME) with a flat tip. The diameter of the inner pole was 75 m and that
of the outer pole was 300 m. The electrode was used in a monopolar configuration: the
inner pole was used for stimulation and a large surface area platinum needle inserted in the
skin adjacent to the nose was used as the return electrode. The stimulating electrode was
mounted in a 1-ml syringe for handling and attached to a single-axis linear translational
micromanipulator (model NT33-475, Edmund Optics, Barrington, NJ) on a magnetic
based articulating arm.
Modifying the Surface of the Stimulation and Recording Electrodes
High surface area platinum-Iridium films were formed on the standard platinum-iridium
microelectrode and tungsten recording electrode using similar electrodeposition method
that was previously described (Petrossians et al., 2011). A potential sweep technique in
the potential range of E = +0.2V to -0.2V vs. Ag/AgCl at scan rate of 0.2 mV/s was used
for electrodeposition of 60:40% Pt-Ir films. The solution was agitated using an ultrasonic
homogenizer (Misonix, Inc. Newtown, CT, USA) at a frequency of 20kHz to maintain
constant mass transfer during electrodeposition and to maintain the temperature of the
plating solution around 70
∘ C.
111
6.3 Surgical Procedures
All surgeries were performed under general anesthesia induced by intraperitoneal/intramuscular
injection of a cocktail of ketamine (100 mg/kg; Ketaset, Fort Dodge Animal Health, Fort
Dodge, IA) and xylazine (100mg/kg; X-Ject SA, Butler, Dublin, OH) and maintained
by sevoflurane (1% in 100% O2) throughout the entire experiment. Sevoflurane was ad-
ministered through a mask. The animal’s pulse and oxygen saturation were monitored
during the surgical procedures. The body temperature was maintained at 37
∘ C with
a self-regulated heating blanket (model 50-7053-F; Harvard Apparatus, Holliston, MA).
Animals were euthanized after the experiment.
6.3.1 SC Exposure and Recording Electrode Positioning
An example EER is shown in figure 6.1. In rats, axons of 90% of the RGCs synapse onto
the superficial layers of the SC (O’Leary et al., 1986; Simon and O’Leary, 1992), thus
recording from the SC provides a convenient measure of retinal output. In order to access
the SC, the skull was exposed and a craniotomy was performed on the right side ((caudal-
medial corner: ∼ 4 mm caudal and∼ 3 mm lateral to lambda) using a hand-held drill.
The overlying cortex was aspirated approximately 4mm deep from the dura mater until
the SC surface was exposed. Epoxy-coated tungsten microelectrodes (10 MOhms, FHC)
were positioned within the superficial layers of the SC at a depth of 300-350 m from
the SC surface. This recording technique was described in detail in previous work (Chan
et al., 2011; DeMarco et al., 2007; Girman et al., 2005; Kanda et al., 2004; Sagdullaev
et al., 2003; Thomas et al., 2004).
112
Figure 6.1: An evoked potential recorded from the SC when the retina is electrically stimulated
(averaged over 25 trials).
6.3.2 Stimulation Electrode Insertion
The surgical procedure to insert a stimulation electrode into the rat eye was reported
in previous work (Colodetti et al., 2007; Ray et al., 2009). The left eye was dilated
with a few drops each of 1% tropicamide (Tropicacyl, Akorn, Buffalo Grove, IL) and
2.5% phenylephrine (AK-Dilate, Akorn). The dilated eye was proptosed using a small
piece of a surgical glove. Slightly flattening the cornea using a glass coverslip covered
with gel (Goniosol, Gonak) allows focused viewing of the fundus through an operating
microscope. A scleral incision was made using a 25-guage needle near the limbus. The
needle was inserted at a 45
∘ angle with respect to the scleral surface in order to avoid
damaging the lens. The stimulation electrode was inserted through the incision site along
the path made by the needle. The electrode was positioned in the ventral temporal
113
quadrant without contacting the retina. Final positioning of the stimulation electrode to
ensure close proximity utilizes impedance feedback described below.
6.3.3 Positioning Stimulation Electrode via Impedance Sensing
The proximity of the stimulation electrode to the retina was monitored indirectly by
measuring the electrochemical impedance with a commercial potentiostat (Gamry In-
struments, Warminster, PA). The potentiostat consists of a 3-electrode setup to measure
impedance. The inner pole of the stimulation electrode was connected to the working
electrode input. Two platinum needle type electrodes, placed under the skin of the nose
and tail of the rat, were connected to the counter and reference electrode inputs respec-
tively. In all the experiments the electrochemical impedance was maintained between
9.5 kΩ - 10 kΩ when measured at 100 kHz. Further details about this technique were
discussed in previous work (Colodetti et al., 2007; Ray et al., 2009).
6.4 Electrical Stimulation
Charge-balanced, cathodic first, biphasic current and voltage-controlled pulses were de-
livered to the epiretinal surface across 4 pulse durations (0.3, 0.5, 1 and 2 ms) through
the stimulation electrode described above (figure 6.2). The amplitude of the pulses was
chosen such that the charge in the cathodic phase was between 10 and 60 nC. Stimu-
lus pulses were supra-threshold. For each pulse duration, voltage and current-controlled
pulse trains with four different charge levels were delivered. Thus, 32 stimulus conditions
(2 stimulation modes x 4 charge levels x 4 pulse durations) were applied to the retina
per experiment. The interphase interval was kept constant at 100 μs. The stimulus
114
pulses were delivered at 0.2Hz for all pulse durations. 25 pulses were delivered for each
pulse duration and amplitude in each stimulation mode. The order in which current and
voltage-controlled pulses were delivered was randomized. The stimulus pulses were gen-
erated by A-M systems converter (model 2200) that was driven by a voltage pulse from
a programmable analog output card (DataWave Technologies, Berthoud, CO). Limita-
tions in the electronics prevented perfectly rectangular pulses, so we recorded the actual
waveform applied across the electrode and used this waveform in subsequent analysis.
6.5 Data Acquisition
EERs were recorded for each pulse delivered in both voltage and current-controlled modes.
The most sensitive response region within the SC was determined and EERs were recorded
from that region. Details about the determination of the most sensitive region were
discussed in previous work (Chan et al., 2011). Briefly, the recording electrode was
moved in a grid pattern while a standard pulse was applied. The area that responds most
robustly, on visual inspection, is deemed the most sensitive area. In addition to recording
EERs, the current and voltage waveforms (figure 6.2) of the stimulus pulses delivered to
the retina were recorded on an oscilloscope (model 5034B Tektronix TDS). A 20 kΩ sense
resistor was used to record the current and voltage waveforms.
115
Figure 6.2: An example of the rectangular current-controlled and voltage-controlled pulses
applied to the electrode-retina interface. Corresponding voltage drop and current waveforms are
shown as well.

is the solution resistance,

is the double layer capacitance and

is the
polarization resistance.
6.6 Data Analysis
6.6.1 Measurement of Injected Charge and Quantification of Electrically
Evoked Responses (EERs)
For each stimulus pulse, the corresponding current through the electrode-tissue interface
and the voltage drop across the interface were recorded. Charge delivered to the retinal
surface was calculated by integrating the cathodic phase of the current waveform. For
each stimulus condition, 25 stimulus pulses were delivered and the resulting EERs were
recorded. An average EER was obtained and the strength within the average response
was calculated using equation 6.1.
ℎ =

∑︁
=0
(

)
2
(6.1)
116

defines the time window within which the signal strength is calculated. It was the
first 50 ms after the stimulus pulse was delivered. (

) is the amplitude of the EER
measured in V and N is the number of samples with the first 50 ms. Stimulus artifact
was excluded from this calculation.
6.6.2 Signal Strength versus Injected Charge
The amount of charge applied to the tissue and strength within the resulting average EER
were calculated as described above for all stimulus conditions. Signal strength was plotted
against charge and an equation for a line of best fit was obtained using linear regression.
One data set out of 10 experiments in the first experimental group was a significant outlier
and was not included for further analysis (Grubb’s outlier test, p < 0.05). Signal strength
versus injected charge curves were obtained across all pulse durations in both voltage and
current-controlled stimulation modes. Strength in EERs could not be compared directly
as the stimulator did not generate stimulus pulses with the same amount of charge in both
modes. This is due to the fact that impedance varied slightly, thus voltage stimulation
did not produce the same current across experiments. To allow a direct comparison, a
best fit line was calculated for EER strength as a function of charge and comparisons
were made at 4 charge levels (20, 30, 40 and 50 nC) using the best fit line. This processed
EER strength data were plotted against the above determined charge levels in order
to compare signal strength in both stimulation modes. In order to determine if signal
strength in both modes was statistically different from each other for a given charge,
paired t-test was used for statistical analysis and p < 0.05 was considered significant.
117
6.6.3 Power Consumption
The power consumed while generating current-controlled pulses was computed from equa-
tion 6.2.
=
1

(
max
* ∫︁
()) (6.2)
where T is the pulse duration; I(t) is the measured current; dt is the step size.
max
is the maximum voltage measured from the voltage waveform.
max
is used, since a
current source will have a fixed voltage and all current will be produced with that voltage.
The power consumed while generating the voltage-controlled pulses was computed from
equation 6.3.
=
1

(
∫︁
()()) (6.3)
where T is the pulse duration; V(t) is the measured voltage; I(t) is the measured
current; dt is the step size.
As with the signal strength data, power consumption measurements were also fit to
an equation using linear regression. All equations had a good correlation coefficient (R
> 0.9) and were included for further analysis. The power required to generate pulses
with 20, 30, 40 and 50 nC of charge was inter/extrapolated from the best fit equations.
118
Statistical significance was determined from the processed data. A paired t-test was used
for statistical analysis and p < 0.05 was considered significant.
6.7 Computational Model
We performed simulations of a Hodgkin-Huxley computational model to understand the
underlying mechanisms of our findings (Benison G and DW, 2001). The membrane volt-
age and gating particles of a cat RGC were simulated. All simulations were conducted
in Matlab. Current waveforms recorded in our in vivo study while stimulating the retina
in both modes were used as stimulus pulses for the model. The retina was stimulated
extracellularly in our in vivo study; however, the model simulates intracellular stimu-
lation of the retinal ganglion cell. The threshold charge required to generate an action
potential using intracellular stimulation is significantly less than extracellular stimulation.
For in vivo stimulation, a typical threshold charge was about 10 nC, while intracellular
stimulation, via a patch clamp electrode, can elicit action potentials with pulses on the
order of 10 pC (Margolis and Detwiler, 2007). While the mechanisms of stimulation are
slightly different in the two cases, the empirical results support a scaling factor of 0.001,
for comparison of empirical and model data.
6.8 Imaging Setup
The anesthetized rat is held in a stereotactic frame. Illumination is provided by red and
green LEDs stabilized with a regulated power supply in order to minimize the noise. A 10
bit CCD camera is used to capture the light reflected from the SC. To achieve the optimal
119
viewing area and working distance, the camera is fitted with a lens (12X Ultrazoom video
micro lens from Navitar). The camera is focused 300-500 m below the SC surface to
minimize IS contributions from surface blood vessels. A computer controls data collection
and stimulus delivery for a set of trials. The collected data is then analyzed in Matlab.
6.9 Imaging Data Analysis
ISI: In order to analyze the images, fractional changes in reflectance in the image are
computed using the formula:
Δ

=
(stimulation image – reference image)
reference image
(6.4)
where R is the reflectance, stimulation image is the average of frames acquired during
stimulation and reference image is the average of frames acquired in the absence of stim-
ulation. After computing the fractional change in reflectance, functional maps known
as single condition maps can be obtained. These single condition maps are the average
response to each stimulus condition. IS signals tend to be noisy; so nearby pixels are
blurred with median filters to reduce the noise levels.
120
References
Ahuja, A.K., Behrend, M.R, Kuroda, M., Humayun, M.S., and Weiland, J.D. An In Vitro Model
of a Retinal Prosthesis. IEEE Trans Biomed Eng, 55(6):1744–1753, 2008.
Belliveau, J.W., Cohen, M.S., Weisskoff, R.M., Buchbinder, B.R., and Rosen, B.R. Functional
studies of the human brain using high-speed magnetic resonance imaging. Journal of Neu-
roimaging, 1(1):36–41, 1991.
Keizer JBenison G, Chalupa LM and DW, Robinson. Modeling temporal behavior of postnatal
cat retinal ganglion cells. J. theor. Biol., 210:187–199, 2001.
Blasdel, G. and Campbell, D. Functional retinotopy of monkey visual cortex. Journal of Neuro-
science, 21(20):8286–8301, 2001.
Bonhoeffer, T. and Grinvald, A. Iso-orientation domains in cat visual cortex are arranged in
pinwheel-like patterns. Nature, 353(6343):429–431, 1991.
Bosking, W.H., Zhang, Y., Schofield, B., and Fitzpatrick, D. Orientation selectivity and the
arrangement of horizontal connections in tree shrew striate cortex. Journal of Neuroscience,
2112-2127(6):17, 1997.
Bosking, W.H., Kretz, R., Pucak, M.L., and Fitzpatrick, D. Functional specificity of callosal
connections in tree shrew striate cortex. Journal of Neuroscience, 20(6):2346–2359, 2000.
Bosking, W.H., Crowley, J.C., and Fitzpatrick, D. Spatial coding of position and orientation in
primary visual cortex. Nature Neuroscience, 5(9):874–882, 2002.
Bradley, D.C., Troyk, P.R., Berg, J.A., Bak, M., Cogan, S., Erickson, R., Kufta, C., Mascaro,
M., McCreery, D., Schmidt, EM., Towle, VL., and Xu, H. Visuotopic mapping through a
multichannel stimulating implant in primate V1. Journal of NeuroPhysiology, 93:1659–1670,
2005.
Brel´ en, M.E., Duret, F., G´ erard, B., Delbeke, J., and Veraart, C. Creating a meaningful visual
perception in blind volunteers by optic nerve stimulation. Journal of Neural Engineering, 2:
S22–S28, 2005.
Brindley, G.S. and Lewin, W.S. The sensations produced by electrical stimulation of the visual
cortex. Journal of Physiology, 196(2):479–493, 1968.
Button, J. and Putnam, T. Visual responses to cortical stimulation in the blind. J Iowa St Med
Soc, 52:17, 1962.
Cannestra, A.F., Pouratian, N., Shomer, M.H., and Toga, A.W. Refractory periods observed
by intrinsic signal and fluorescent dye imaging. Journal of Neurophysiology, 80(3):1522–1532,
1998.
121
Cannestra, A.F., Pouratian, N., Bookheimer, S.Y., Martin, N.A., Beckerand, D.P., and Toga,
A.W. Temporal spatial differences observed by functional MRI and human intraoperative
optical imaging. Cerebral Cortex, 11(8):773–782, 2001.
Caspi, A., Dorn, J.D., McClure, K.H., Humayun, M.S., Greenberg, R.J., and McMahon, M.J.
Feasibility study of a retinal prosthesis: Spatial vision with a 16-electrode implant. Archives of
Ophthalmology, 127(4):398–401, 2009.
Chan, L.L., Lee, E.J., Humayun, M.S., and Weiland, J.D. Both electrical stimulation thresholds
and SMI-32-immunoreactive retinal ganglion cell density correlate with age in S334ter line 3
rat retina. Journal of Neurophysiology, 105(6):2687–2697, 2011.
Chen-Bee, C.H. and Frostig, R.D. Variability and interhemispheric asymmetry of single-whisker
functional representations in rat barrel cortex. Journal of Neurophysiology, 76(2):884–894, 1996.
Chow, A.Y., Chow, V.Y., Packo, K.H., Pollack, J.S., Peyman, G.A., and Schuchard, R. The
artificial silicon retina microchip for the treatment of vision loss from retinitis pigmentosa.
Archives of Ophthalmology, 122(4):460–469, 2004.
Colodetti, L., Weiland, J.D., Colodetti, S., Ray, A., Seiler, M.J., Hinton, D.R., and Humayun,
M.S. Pathology of damaging electrical stimulation in the retina. Exp Eye Research, 85(1):
23–33, 2007.
Congdon, N., O’Colmain, B., Klaver, CCW., and Klein, R. Causes and Prevalence of Visual
Impairment Among Adults in the United States. Archives of Ophthalmology, 122:477–485,
2011.
Cowey, A. and Rolls, ET. Human cortical magnification factor and its relation to visual acuity.
Experimental Brain Research, 21:447–454, 1974.
Cox, S.B., Woolsey, T.A., and Rovainen, C.M. Localized dynamic changes in cortical blood flow
with whisker stimulation corresponds to matched vascular and neuronal architecture of rat
barrels. Journal of Cerebral Blood Flow and Metabolism, 13(6):899–913, 1993.
Curcio, C.A., Sloan, K.R., Kalina, R.E., and Hendrickson, A.E. Human photoreceptor topography.
Journal of Comparative Neurology, 292(4):497–523, 1990.
Daiger, S.P., Bowne, S.J., and Sullivan, L.S. Perspective on genes and mutations causing retinitis
pigmentosa. Archives of Ophthalmology, 125(2):151–158, 2007.
de Balthasar, C., Patel, S., Roy, A., Freda, R., Greenwald, S., Horsager, A., Mahadevappa, M.,
Yanai, D., McMahon, M.J., and Humayun, M.S. Factors affecting perceptual thresholds in
epiretinal prostheses. Investigative Ophthalmology and Visual Science, 49(6):2303–2314, 2008.
DeMarco, P.J. Jr., Yarbrough, G.L., Yee, C.W., McLean, G.Y., Sagdullaev, B.T., Ball, S.L.,
and McCall, M.A. Stimulation via a subretinally placed prosthetic elicits central activity and
induces a trophic effect on visual responses. Invest Ophthalmol Vis Sci., 48:916–926, 2007.
DeVries, S.H. and Schwartz, E.A. Kainate receptors mediate synaptic transmission between cones
and ’Off’ bipolar cells in a mammalian retina. Nature, 397(6715):157–160, 1999.
Dobelle, W.H., Mladejovsky, M.G., and Girvin, J.P. Artifical vision for the blind: electrical
stimulation of visual cortex offers hope for a functional prosthesis. Science, 183(4123):440–4,
1974.
122
Edwards, M.D., White, A.M., and Platt, B. Characterisation of rat superficial superior colliculus
neurones: firing properties and sensitivity to GABA. Neuroscience, 110(1):93–104, 2002.
Fishler, M.G. Theoretical predictions of the optimal monophasic and biphasic defibrillation wave-
shapes. IEEE Trans Biomed Eng., 47(1):59–67, 2000.
Fitzgibbon, T. and Reese, B.E. Organization of retinal ganglion cell axons in the optic fiber layer
and nerve of fetal ferrets. Vis Neurosci, 13(5):847–61, 1996.
Foerster, O. Beitrage zur pathophysiologie der sehbahn und der spehsphare. J Psychol Neurol,
39:435–463, 1929.
Foutz, T.J. and McIntyre, C.C. Evaluation of novel stimulus waveforms for deep brain stimulation.
Journal of Neural Engineering, 7(6), 2010.
Fox, P.T., Raichle, M.E., Mintun, M.A., and Dence, C. Nonoxidative glucose consumption during
focal physiologic neural activity. Science, 241(4864):462–464, 1988.
Freeman, D.K. and Fried, S.I. Multiple components of ganglion cell desensitization in response
to prosthetic stimulation. Journal of Neural Engineering, 8(1), 2011.
Fried, S.I., Hsueh, H.A., and Werblin, F.S. A method for generating precise temporal patterns of
retinal spiking using prosthetic stimulation. Journal of Neurophysiology, 95(2):970–978, 2005.
Frostig, R.D., Lieke, E.E., Tso, D.Y., and Grinvald, A. Cortical functional architecture and local
coupling between neuronal activity and the microcirculation revealed by in vivo high-resolution
optical imaging of intrinsic signals. Proceedings of the National Academy of Sciences, 87(16):
6082–6086, 1990.
Fukuda, Y. A three-group classification of rat retinal ganglion cells: histological and physiological
studies. Brain Research, 119(2):327–334, 1977.
Fukuda, Y. and Iwama, K. Visual receptive-field properties of single cells in the rat superior
colliculus. Japanese Journal of Physiology, 28(3):385–400, 1978.
Gekeler, F., Kobuch, K., Schwahn, H.N., Stett, A., Shinoda, K., and Zrenner, E. Subretinal
electrical stimulation of the rabbit retina with acutely implanted electrode arrays. Graefes
Arch Clin Exp Ophthalmology, 242(7):587–596, 2004.
Girman, S.V., Wang, S., and Lund, R.D. Time course of deterioration of rod and cone function in
RCS rat and the effects of subretinal cell grafting: a light- and dark-adaptation study. Vision
Research, 45(3):343–354, 2005.
Godde, B., Hilger, T., von Seelen, W., Berkefeld, T., and Dinse, H.R. Optical imaging of rat
somatosensory cortex reveals representational overlap as topographic principle. Neuroreport, 7
(1):24–28, 1995.
Goo, Y.S., Ye, J.H., Lee, S., Nam, Y., Ryu, S.B., and Kim, K.H. Retinal ganglion cell responses
to voltage and current stimulation in wild-type and rd1 mouse retinas. Journal of Neural
Engineering, 8(3), 2011.
Greenberg, R.J. Analysis of electrical stimulation of the vertebrate retina : work towards a retinal
prosthesis. PhD thesis, John Hopkins University, 1998.
Grinvald, A., Lieke, E., Frostig, R.D., Gilbert, C.D., and Wiesel, T.N. Functional architecture of
cortex revealed by optical imaging of intrinsic signals. Nature, 324(6095):361–364, 1986.
123
Grinvald, A., Frostig, R.D., Siegel, S.M., and Bartfeld, E. High-resolution optical imaging of
functional brain architecture in the awake monkey. Proc Natl Acad Science USA, 88(24):
11559–11563, 1991.
Hale, P.T. and Sefton, A.J. A comparison of the visual and electrical response properties of cells
in the dorsal and ventral lateral geniculate nuclei. Brain Research, 153(3):591–595, 1978.
Hodge, C.J., Stevens, R.T., Newman, H., Merola, J., and Chu, C. Identification of functioning
cortex using cortical optical imaging. Neurosurgery, 41(5):1137–1144, 1997.
Horsch, K.W. and Dhillon, G.S. Neuroprosthetics: Theory and Practice. 2004.
Hu, X., Le, T.H., and U˘ gurbil, K. Evaluation of the early response in fMRI in individual subjects
using short stimulus duration. Magnetic Resonance in Medicine, 37(6):877–884, 1997.
Hubener, M., Shoham, D., Grinvald, A., and Bonhoeffer, T. Spatial Relationships among Three
Columnar Systems in Cat Area 17. Journal of Neuroscience, 17(23):9270–9284, 1997.
Humayun, M.S., Dorn, J.D., da Cruz, L., Dagnelie, G., Sahel, J.A., Stanga, P.E., Cideciyan, A.V.,
Duncan, J.L., Eliott, D., Filley, E., Ho, A.C., Santos, A., Safran, A.B., Arditi, A., Del Priore,
L.V., Greenberg, R.J., and Group, Argus II Study. Interim results from the international trial
of Second Sight’s visual prosthesis. Opththalmology, 119(4):779–788, 2012.
Huxlin, K.R. and Goodchild, A.K. Retinal Ganglion cells in the Albino Rat: Revised morpholog-
ical classification. Journal of Comparative Neurology, 385:309–323, 1998.
Issa, N.P., Trachtenberg, J.T., Chapman, B., Zahs, K.R., and Stryker, M.P. The critical period
for ocular dominance plasticity in the Ferret’s visual cortex. Journal of Neuroscience, 19(16):
6965–6978, 1999.
Jensen, R.J. and Rizzo, J.F. Responses of ganglion cells to repetitive electrical stimulation of the
retina. Journal Neural Engineering, 4(1):S1–6, 2007.
Jensen, R.J., Rizzo, J.F., Ziv, O.R., Grumet, A., and Wyatt, J. Thresholds for activation of
rabbit retinal ganglion cells with an ultrafine, extracellular microelectrode. Investigative Oph-
thalmology and Visual Science, 44(8):3533–43, 2003.
Jezernik, S. and Morari, M. Energy-optimal electrical excitation of nerve fibers. IEEE Transac-
tions on Biomedical Engineering, 52(4):740–743, 2005.
Jezernik, S. and Morari, M. Charge and energy minimization in electrical/magnetic stimulation
of nervous tissue. Journal of Neural Engineering, 7(4):046004, 2010.
Jones, B.W. and Marc, R.E. Retinal remodeling during retinal degeneration. Experimental Eye
Research, 81(2):123–137, 2005.
Jones, B.W., Watt, C.B., Frederick, J.M., Baehr, W., Chen, C.K., Levine, E.M., Milam, A.H.,
Lavail, M.M., and Marc, R.E. Retinal remodeling triggered by photoreceptor degenerations.
Journal of Comp. Neurology, 464:1–16, 2003.
Kajimoto, H., Kawakami, N., and Tachi, S. Optimal design method for selective nerve stimulation
and its application to electrocutaneous display. Proc.10th Symp. Haptic Interfaces Virtual
Envir.Teleoperator Syst., pages 303–310, 2002.
Kalatsky, V.A. and Stryker, M.P. New paradigm for optical imaging: temporally encoded maps
of intrinsic signal. Neuron., 38(4):529–545, 2003.
124
Kanda, H., Morimoto, T., Fujikado, T., Tano, Y., Fukuda, Y., and Sawai, H. Electrophysiological
studies of the feasibility of suprachoroidal-transretinal stimulation for artificial vision in normal
and RCS rats. Invest Ophthalmol Vis Sci., 45:560 – 566, 2004.
Kim, D.S., Matsuda, Y., Ohki, K., Ajima, A., and Tanaka, S. Geometrical and topological
relationships between multiple functional maps in cat primary visual cortex. NeuroReport, 10
(12):2515–2522, 1999.
Kirischuk, S., Clements, J.D., and Grantyn, R. Presynaptic and postsynaptic mechanisms underlie
paired pulse depression at single GABAergic boutons in rat collicular cultures. Journal of
Physiology, 543:99–116, 2002.
Klafter, R.D. and Hrebien, L. An in vivo study of cardiac pacemaker optimization by pulse shape
modification. IEEE Transactions on Biomedical Engineering, 23(3):233–239, 1976.
Kohl, M., Lindauer, U., Royl, G., Kuhl, M., Gold, L., Villringer, A., and Dirnagl, U. Physical
model for the spectroscopic analysis of cortical intrinsic optical signals. Physics in Medicine
and Biology, 45(12):3749–3764, 2000.
Krause, F. and Schum, H. Die epiliptischen Erkankungen. American Journal of Psychiatry, 1931.
Kwong, K.K., Belliveau, J.W., Chesler, D.A., Goldberg, I.E., Weisskoff, R.M., Poncelet, B.P.,
Kennedy, D.N., Hoppel, B.E., Cohen, M.S., and Turner, R. Dynamic magnetic resonance
imaging of human brain activity during primary sensory stimulation. Proceedings of National
Academy of Sciences USA, 89(12):5675–5679, 1992.
LaManna, J.C., Sick, T.J., Pikarsky, S.M., and Rosenthal, M. Detection of an oxidizable fraction
of cytochrome oxidase in intact rat brain. American Journal of Physiology, 253(3):477–483,
1987.
Landisman, C.E. and Tso, D.Y. Color processing in macaque striate cortex: relationships to
ocular dominance, cytochrome oxidase, and orientation. Journal of Neurophysiology, 87(6):
3126–3137, 2002.
LeRoy, C. O` u l’on rend compte de quelques tentatives que l’on a faites pour gu´ erir plusieurs
maladies par l’´ electricit´ e. Hist. Acad. Roy. Sciences (Paris), M´ emoires Math. Phys., 60:87–95,
1755.
Lindauer, U., Villringer, A., and Dirnagl, U. Characterization of CBF response to somatosensory
stimulation: model and influence of anesthetics. American Physiological Society, 264(4):1223–
1228, 1993.
Lindauer, U., Royl, G., Leithner, C., K¨ uhl, M., Gold, L., Gethmann, J., Kohl-Bareis, M., Vill-
ringer, A., and Dirnagl, U. No evidence for early decrease in blood oxygenation in rat whisker
cortex in response to functional activation. Neuroimage, 13(6):988–1001, 2001.
Liu, G.B. and Pettigrew, J.D. Orientation mosaic in barn owl’s visual Wulst revealed by optical
imaging: comparison with cat and monkey striate and extra-striate areas. Brain Research, 961
(1):153–158, 2003.
Logothetis, N.K., Guggenberger, H., Peled, S., and Pauls, J. Functional imaging of the monkey
brain. Nature Neuroscience, 2(6):555–562, 1999.
Lou, H.C., Edvinsson, L., and MacKenzie, E.T. The concept of coupling blood flow to brain
function: revision required? Annals of Neurology, 22(3):289–97, 1987.
125
Lund, R.D. Synaptic patterns of the superficial layers of the superior colliculus of the rat.Journal
of Comparative Neurology, 135:179–208, 1969.
Lund, R.D. and Langer, T.P. Synaptic patterns of the superficial layers of the superior colliculus
of the rat. Journal of Comparative Neurology, 158:405–435, 1974.
Malonek, D. and Grinvald, A. Interactions between electrical activity and cortical microcirculation
revealed by imaging spectroscopy: implications for functional brain mapping. Science, 272
(5261):551–554, 1996.
Marc, R.E., Jones, B.W., Anderson, J.R., Kinard, K., Marshak, D.W., Wilson, J.H., Wensel, T.,
and Lucas, R.J. Neural reprogramming in retinal degeneration. Invest Ophthalmol Vis Science,
48(7):3364–3371, 2007.
Margolis, D.J. and Detwiler, P.B. Different Mechanisms Generate Maintained Activity in ON and
OFF Retinal Ganglion Cells. The Journal of Neuroscience, 27(22):5994–6005, 2007.
Margolis, D.J., Newkirk, G., Euler, T., and Detwiler, P.B. Functional stability of retinal ganglion
cells after degeneration-induced changes in synaptic input. Journal of Neuroscience, 28(25):
6526–6536, 2008.
Masino, S.A. and Frostig, R.D. Quantitative long-term imaging of the functional representation
of a whisker in rat barrel cortex. Proceedings of National Academy of Sciences, 93:4942–4947,
1996.
Mayhew, J., Zheng, Y., Hou, Y., Vuksanovic, B., Berwick, J., Askew, S., and Coffey, P. Spectro-
scopic analysis of changes in remitted illumination: the response to increased neural activity in
brain. NeuroImage, 10(3):304–326, 1999.
Mazziotta, J.C., Toga, A.W., and Frackowiak, R.S.J. Brain Mapping: The Disorders. Academic
Press San Diego, 2000.
Menon, R.S., Ogawa, S., Hu, X., Strupp, J.P., Anderson, P., and U˘ gurbil, K. BOLD based
functional MRI at 4 tesla includes a capillary bed contribution: Echo-planar imaging correlates
with previous optical imaging using intrinsic signals. Magnetic Resonance in Medicine, 33(3):
453–459, 1995.
Milam, A.H., Li, Z.Y., Cideciyan, A.V., and Jacobson, S.G. Clinicopathologic effects of the Q64ter
rhodopsin mutation in retinitis pigmentosa. Invest Ophthalmol Vis Sci., 37(5):753–65, 1996.
Narayan, S.M., Esfahani, P., Blood, A.J., Sikkens, L., and Toga, A.W. Functional increases
in cerebral blood volume over somatosensory cortex. Journal of Cerebral Blood Flow and
Metabolism, 15(5):754–765, 1995.
Nassi, JJ. and Callaway, E.M. Parallel processing strategies of the primate visual system. Nat
Rev Neurosci., 10(5):360–372, 2009.
Nemoto, M., Nomura, Y., Sato, C., Tamura, M., Houkin, K., Koyanagi, I., and Abe, H. Analysis
of optical signals evoked by peripheral nerve stimulation in rat somatosensory cortex: dynamic
changes in hemoglobin concentration and oxygenation. Journal of Cerebral Blood Flow and
Metabolism, 19(3):246–259, 1999.
Ngai, A.C., Ko, K.R., Morii, S., and Winn, H.R. Effect of sciatic nerve stimulation on pial
arterioles in rats. American Journal of Physiology, 254(1):133–142, 1988.
126
Normann, R.A, Maynard, E.M., Rousche, P.J, and Warren, D.J. A neural interface for a cortical
vision prosthesis. Vision Research, 39(15):2577–2587, 1999.
OFarrell, A.M., Rex, D.E., Muthialu, A., Pouratian, N., Wong, G.K., Cannestra, A.F., Chen,
J.W., and Toga, A.W. Characterization of optical intrinsic signals and blood volume during
cortical spreading depression. Neuroreport, 11(10):2121–2125, 2000.
Offner, F. Stimulation with minimum power. Journal of Neurophysiology, 9:387–390, 1946.
O’Leary, D.D., Fawcett, J.W., and Cowan, W.M. Topographic targeting errors in the retinocol-
licular projection and their elimination by selective ganglion cell death. The Journal of Neuro-
science, 6(12):3692–705, 1986.
Orbach, H.S. and Cohen, L.B. Optical monitoring of activity from many areas of the in vitro and
in vivo salamander olfactory bulb: A new method for studying functional organization in the
vertebrate central nervous system. Journal of Neuroscience, 3(11):2251–2262, 1983.
Paxinos, G. and Watson, C. The Rat Brain in Stereotaxic Coordinates, Fourth Edition. 1998.
Penfield, W. and Rasmussen, T. The cerebral cortex of man. New York: Macmillan. 1952.
Petrossians, A., Whalen, J.J., Weiland, J.D., and Mansfelda, F. Electrodeposition and character-
ization of thin-film platinum-iridium alloys for biological interfaces. Journal of Electrochemical
Society, 158(5):D269–D276, 2011.
Pezaris, J.S. and Reid, R.C. Demonstration of artificial visual percepts generated through thalamic
microstimulation. Proceedings of the National Academy of Sciences of the United States, 104
(18):7670–7675, 2007.
Pritchard, R.M. Stabilized images on the retina. Sci Am, 204:72–78, 1961.
P´ erez, F.A., Sommerhalder, J., da Cruz, L., Sahel, J.A., Mohand-Said, S., Hafezi, F., and Peliz-
zone, M. Temporal properties of visual perception on electrical stimulation of the retina.
Investigative Ophthalmology and Visual Science, 53(6):2720–31, 2012.
Rabl, K., Cadetti, L., and Thoreson, W.B. Paired-pulse depression at photoreceptor synapses.
The Journal of Neuroscience, 26(9):2555–2563, 2006.
Ray, A. Effect of Continuous Electrical Stimulation on Retinal Structure and Function. PhD
thesis, The University of Southern California, 2010.
Ray, A., Colodetti, L., Weiland, J.D., Hinton, D.R., Humayun, M.S., and Lee, E.J. Immunocyto-
chemical analysis of retinal neurons under electrical stimulation. Brain Research, 1255:89–97,
2009.
Rizzo, JF., Wyatt, J., Loewenstein, J., Kelly, S., and Shire, D. Methods and perceptual thresh-
olds for short-term electrical stimulation of human retina with microelectrode arrays. Invest
Ophthalmol Vis Sci., 44(12):5355–5361, 2003a.
Rizzo, J.F., Wyatt, J., Loewenstein, J., Kelly, S., and Shire, D. Perceptual efficacy of electrical
stimulation of human retina with a microelectrode array during short-term surgical trials.
Investigative Ophthalmology and Visual Science, 44(12):5362–5369, 2003b.
Rohrer, B. and Crouch, R. Rod and Cone Pigment Regeneration in RPE65 -/- Mice. Advances
in Experimental Medicine and Biology, 572:101–107, 2006.
127
Roy, C.S. and Sherrington, C.S. On the Regulation of the Blood-supply of the Brain. Journal of
Physiology, 11(1-2):85–108, 1890.
Sagdullaev, B.T., Aramant, R.B., Seiler, M.J., Woch, G., and McCall, M.A. Retinal
transplantation-induced recovery of retinotectal visual function in a rodent model of retini-
tis pigmentosa. Invest Ophthalmol Vis Sci., 44(4):1686–1695, 2003.
Sahin, M. and Tie, Y. Non-rectangular waveforms for neural stimulation with practical electrodes.
Journal of Neural Engineering, 4(3):227–233, 2007.
Sanders, C.A., Nagler, E.J., Zhou, D.M., and Greenbaum, E. Artificial Sight: Basic Research,
Biomedical Engineering and Clinical Advances. Springer, 2007.
Santos, A., Humayun, M.S., de Juan Jr., E., Greenburg, R.J., Marsh, M.J., Klock, I.B., and
Milam, A.H. Preservation of the inner retina in retinitis pigmentosa: a morphometric analysis.
Archives of Ophthalmology, 115(4):511–515, 1997.
Sekirnjak, C., Hottowy, P., Sher, A., Dabrowski, W., Litke, A.M., and Chichilnisky, E.J. Elec-
trical stimulation of mammalian retinal ganglion cells with multielectrode arrays. Journal of
Neurophysiology, 95(6):3311–3327, 2006.
Sekirnjak, C., Hottowy, P., Sher, A., Dabrowski, W., Litke, A.M., and Chichilnisky, E.J. High-
Resolution Electrical Stimulation of Primate Retina for Epiretinal Implant Design. The Journal
of Neuroscience, 28(17):4446–4456, 2008.
Shmuel, A. and Grinvald, A. Functional Organization for Direction of Motion and Its Relationship
to Orientation Maps in Cat Area 18. The Journal of Neuroscience, 16(21):6945–6964, 1996.
Shtoyerman, E., Arieli, A., Slovin, H., Vanzetta, I., and Grinvald, A. Long-term optical imaging
and spectroscopy reveal mechanisms underlying the intrinsic signal and stability of cortical
maps in V1 of behaving monkeys. Journal of Neuroscience, 20:8111–8121, 2000.
Siminoff, R., Schwassmann, H.O., and Kruger, L. An electrophysiological study of the visual
projection to the superior colliculus of the rat. Journal of Comparitive Neurology, 127(4):
435–444, 1966.
Simon, D.K. and O’Leary, D.D. Development of topographic order in the mammalian retinocol-
licular projection. The Journal of Neuroscience, 12(4):1212–32, 1992.
Singer, J.H. and Diamond, J.S. Vesicle depletion and synaptic depression at a mammalian ribbon
synapse. Journal of Neurophysiology, 95(5):3191–3198, 2006.
Sokoloff, L., Reivich, M., Kennedy, C., Des Rosiers, M.H., Patlak, C.S., Pettigrew, K.D., Saku-
rada, O., and Shinohara, M. THE [14C]DEOXYGLUCOSE METHOD FOR THE MEASURE-
MENT OF LOCAL CEREBRAL GLUCOSE UTILIZATION: THEORY, PROCEDURE, AND
NORMAL VALUES IN THE CONSCIOUS AND ANESTHETIZED ALBINO RAT. Journal
of Neurochemistry, 28(5):897–916, 1977.
Sparks, D.L. and Young, R.T. The deep layers of superior colliculus. The Neurobiology of saccadic
eye movements. Chapter 5. 1989.
Stone, JL., Barlow, WE., Humayun, MS., de Juan, E., and Milam, AH. Morphometric anal-
ysis of macular photoreceptors and ganglion cells in retinas with retinitis pigmentosa. Arch
Ophthalmol., 110(11):1634–1639, 1992.
128
Strettoi, E., Porciatti, V., Falsini, B., Pignatelli, V., and Rossi, C. Morphological and functional
abnormalities in the inner retina of the rd/rd mouse. Journal of Neuroscience, 22(13):5492–
5504, 2002.
Strettoi, E., Pignatelli, V., Rossi, C., Porciatti, V., and Falsini, B. Remodeling of second-order
neurons in the retina of rd/rd mutant mice. Vision Research, 43(8):867–877, 2003.
Thomas, B.B., Seiler, M.J., Sadda, S.R., and Aramant, R.B. Superior colliculus responses to
light - preserved by transplantation in a slow degeneration rat model. Exp Eye Research, 79
(1):29–39, 2004.
Tokunaga, A. and Otani, K. Dendritic patterns of neurons in the rat superior colliculus. Experi-
mental Neurology, 52(2):189–205, 1976.
Undurraga, J.A., Carlyon, R.P., Macherey, O., Wouters, J., and van Wieringen, A. Spread of
excitation varies for different electrical pulse shapes and stimulation modes in cochlear implants.
Hearing Research, 290(1-2):21–36, 2012.
Vanzetta, Ivo. and Grinvald, A. Increased Cortical Oxidative Metabolism Due to Sensory Stim-
ulation: Implications for Functional Brain Imaging. Science Magazine, 286(5444):1555–1558,
1999.
Veraart, C., Duret, F., Brelen, M., and Delbeke, J. Vision rehabilitation with the optic nerve
visual prosthesis. Conf Proc IEEE Eng Med Biol Soc, 6:4163–4, 2004.
Villringer, A. and Dirnagl, U. Coupling of brain activity and cerebral blood flow: basis of func-
tional neuroimaging. Cerebrovasc Brain Metab Rev., 7(3):240–276, 1995.
Waldeck, R.F., Pereda, A., and Faber, D.S. Properties and plasticity of paired-pulse depression
at a central synapse. Journal of Neuroscience, 20:5312–5320, 2000.
Weitz, A. Improving stimulation strategies for epiretinal prostheses. PhD thesis, The University
of Southern California, 2013.
Weliky, M., Bosking, W.H., and Fitzpatrick, D. A systematic map of direction preference in
primary visual cortex. Nature, 379(6567):725–733, 1996.
White, L.E., Coppola, D.M., and Fitzpatrick, D. The contribution of sensory experience to the
maturation of orientation selectivity in ferret visual cortex. Nature, 411:1049–1052, 2001.
Wongsarnpigoon, A. and Grill, W.M. Energy-efficient waveform shapes for neural stimulation
revealed with a genetic algorithm. Journal of Neural Engineering, 7(4), 2010.
Wongsarnpigoon, A., Wock, J.P., and Grill, W.M. Efficiency analysis of waveform shape for
electrical excitation of nerve fibers. IEEE Transactions on Neural Systems and Rehabilitation
Engineering, 18(3):319–328, 2010.
Xiao, Y., Wang, Y., and Felleman, D.J. A spatially organized representation of colour in macaque
cortical area V2. Nature., 421(6922):535–544, 2003.
Yacoub, E., Le, T.H., Ugurbil, K., and Hu, X. Further evaluation of the initial negative response
in functional magnetic resonance imaging. Magnetic Resonance in Medicine, 41(3):436–441,
1999.
Zdrojkowski, R.J. and Pisharoty, N.R. Optical Transmission and Reflection by Blood. IEEE
Trans. Biomed. Eng., 17(2):122–128, 1970.
129
Zepeda, A., Vaca, L., Arias, C., and Sengpiel, F. Reorganization of visual cortical maps after
focal ischemic lesions. Journal of Cerebral Blood Flow and Metabolism, 23(7):811–820, 2003.
Zrenner, E., Bartz-Schmidt, K.U., Benav, H., Besch, D., Bruckmann, A., Gabel, V.P., Gekeler, F.,
Greppmaier, U., Harscher, A., and Kibbel, S. Subretinal electronic chips allow blind patients
to read letters and combine them to words. Proceedings of the Royal Society B, 278(1711):
1489–1497, 2011.
Zucker, R.S. and Regehr, W. G. Short-term synaptic plasticity. Annual Review of Physiology, 64:
355–405, 2002.
130

Abstract (if available)

Abstract Retinitis Pigmentosa (RP) is a degenerative disorder of the retina that begins with night blindness, leads to tunnel vision and eventually results in complete loss of vision. Physiologically, RP results in the complete loss of photoreceptors and rewiring of the inner retinal neurons. However, electrical stimulation of degenerate retina in RP subjects have shown that electrical stimulation causes visual perception. Ever since electrical stimulation has been shown to elicit light perception, neuroprosthetic devices called epiretinal prostheses are being developed in order to restore some form of functional vision in patients suffering from RP. ❧ ARGUS II is one such epiretinal implant that has been developed by Second Sight Medical Products in collaboration with USC. Recent studies with ARGUS II implant subjects have shown that it does restore some form of functional vision back in RP subjects. However, research also indicated that implant subjects face certain issues when tested in a clinical setting. The first issue faced by ARGUS II subjects’ deals with non-functional electrodes. ARGUS II has 60 electrodes and about 45% of them were non-functional because the charge required to elicit visual percepts using these electrodes exceeded the charge density limit for these electrodes. Thus improving stimulation efficiency may increase the spatial resolution of the implant by increasing the number of functional electrodes. ❧ In order to address the first issue, I studied the effect of stimulation waveform shape and stimulus pulse duration on stimulation efficiency. Specifically, I compared the effect of voltage-controlled and current-controlled pulses of various durations on stimulation efficiency. I found that when the pulse duration was less than 1 ms, voltage-controlled pulses were more efficient at stimulating the retina. When the pulse duration was equal to or greater than 1 ms, current-controlled pulses were more efficient at stimulating the retina. I also developed a computational model that could be used to test the effect of pulse duration and shape on stimulation efficiency. In addition to studying the effect of pulse shape and duration on stimulation efficiency, I tested a novel electrode material and compared its effect on power consumption. The novel electrode material was a high surface area Pt-Ir electrode. When the high surface area Pt-Ir electrode was used for stimulation, it consumed less power to generate stimulus pulses when compared to the Pt-Ir gray stimulation electrode used in the ARGUS II implant. Reducing power consumption improves the implant design and prevents running into voltage compliance issues. ❧ The second issue faced by ARGUS II implant subjects’ was fading of visual percepts over time even when electrical stimulation was turned on. This reduces the temporal resolution of the implant. A decrease in the excitability of retina could be one of the reasons why fading occurs in implant subjects. Many studies have shown that continuous electrical stimulation causes the retina to desensitize. In our lab, we created an in vivo model of retinal desensitization and studies strategies to limit desensitization. Specifically, I studied the effect of pulse trains called time-varying pulses. Time-varying pulses are defined as a train of pulses where each pulse has a different pulse duration and amplitude when compared to the preceding pulse. The charge delivered for every pulse is three times the threshold charge for that pulse width. In addition to time-varying pulses, the effect of short duration pulses on retinal desensitization was also studied. Studies were performed in both normally sighted and retinal degenerate rats. ❧ I showed that time-varying pulses limit retinal desensitization in both normally sighted and retinal degenerate animals. Short duration pulses also limit desensitization in normally sighted rats but not in retinal degenerate rats. Short duration pulses directly stimulate ganglion cells and thus avoid axonal stimulation. However, short duration pulses require higher current amplitudes and selective activation of ganglion cells is restricted to a limited range of stimulus amplitudes. Thus, time-varying pulses might improve the temporal resolution of the implant. ❧ In addition to addressing the aforementioned issues, I also developed an optical imaging technique called intrinsic signal imaging in order to study maps of functional activity in superior colliculus (SC), generated by electrical stimulation of the retina. Traditionally, electrophysiological techniques have been used to study electrical stimulation of the retina. However, such techniques provide limited information about the spatial pattern of retinal excitation. Intrinsic signal imaging overcomes this limitation as it enables high spatial resolution imaging of evoked potential activity in SC.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Improving stimulation strategies for epiretinal prostheses

PDF

Manipulation of RGCs response using different stimulation strategies for retinal prosthesis

PDF

Towards a high resolution retinal implant

PDF

Cortical and subcortical responses to electrical stimulation of rat retina

PDF

Effect of continuous electrical stimulation on retinal structure and function

PDF

Electrical stimulation of degenerate retina

PDF

Investigation of the electrode-tissue interface of retinal prostheses

PDF

Understanding the degenerate retina's response to electrical stimulation: an in vitro approach

PDF

Therapeutic electrical stimulation strategies for neuroregeneration and neuroprotection of retinal neurons

PDF

Prosthetic vision in blind human patients: Predicting the percepts of epiretinal stimulation

PDF

Viewing the picture we paint

PDF

Ultrasound neuromodulation and its applications for noninvasive vision restoration

PDF

Electrical stimulation of the orbicularis oculi to restore eye blink

PDF

Characterization of visual cortex function in late-blind individuals with retinitis pigmentosa and Argus II patients

PDF

Optimal electrical stimulation of smooth muscle

PDF

Modeling retinal prosthesis mechanics

PDF

Synaptic second wind: phosphene fading of the Argus II and glutamate secretion in the inner retina

PDF

Neuromorphic motion sensing circuits in a silicon retina

PDF

Dependence of rabbit retinal synchrony on visual stimulation parameters

PDF

Prosthetic visual perception: retinal electrical stimulation in blind human patients

Asset Metadata

Creator Davuluri, Navya S. (author)

Core Title Stimulation strategies to improve efficiency and temporal resolution of epiretinal prostheses

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Biomedical Engineering

Publication Date 12/06/2012

Defense Date 08/15/2014

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag electrical stimulation of retina,OAI-PMH Harvest,Retinal degeneration

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Weiland, James (committee chair), Chow, Robert H. (committee member), Humayun, Mark S. (committee member)

Creator Email navyaswetha@gmail.com,nsdavulu@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-482181

Unique identifier UC11286808

Identifier etd-DavuluriNa-2965.pdf (filename),usctheses-c3-482181 (legacy record id)

Legacy Identifier etd-DavuluriNa-2965.pdf

Dmrecord 482181

Document Type Dissertation

Format application/pdf (imt)

Rights Davuluri, Navya S.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA