Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Manipulation of RGCs response using different stimulation strategies for retinal prosthesis
(USC Thesis Other)
Manipulation of RGCs response using different stimulation strategies for retinal prosthesis
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
Manipulation of RGCs Response
Using Different Stimulation Strategies for Retinal Prosthesis
By
Yao-Chuan Chang
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 2017
Copyright 2017 Yao-Chuan Chang
i
Epigraph
When you know a thing, to hold that you know it; and when you do not know a thing, to allow
that you do not know it - this is knowledge.
- Confucius
ii
Acknowledgement
Completing my Ph.D. journey would not have been possible without the generous support and help
from others. I would like to express my deepest gratitude to my advisor Dr. James Weiland, for his
excellent guidance, understanding, patience, and providing me with an extraordinary atmospheres
for doing research. I am not sure many graduate students are given the opportunity to develop their
own individuality and self-sufficiency by being allowed to work with such independence. I was so
fortunate to have advisor who always made effective discussion and reasonable requirements so I
can complete dissertation within a relatively short time frame while maintaining work life balance.
For everything you’ve done for me, Dr. Weiland, I thank you. I also owe many thanks to Dr. Robert
Chow who teach me lots of background across different fields, so I can gain the knowledge of
biology and science rapidly though I was trained as an engineer. I would also like to thank the rest
of my committee, Drs. Mark Humayun, Scott Fraser, and Andrew Weitz, for the valuable advice
they provided over the years.
I am very grateful to Steven Walston, who developed the initial framework for my research and
spent countless hours training me to perform experiments. Steven was a brilliant and patient
experimental mentor help me learn several techniques and concepts that became the solid
foundation of my study. I would also like to thank my labmates at University of Southern California,
past and present, for their helpful discussions and assistance with my research. Specifically, I owe
thanks to Lan Yue, Yi Zhang, Nancy Lee, Artin Petrossians, Navya Davuluri, Boshuo Wang, Kiran
Nimmagadda, Karthik Murali, Nii Mante, Aminat Adebiyi, Mort Arditti, Curtis Lee, Chris Girard,
Alejandra Gonzalez, Fernando Gallardo, Lina Flores, Ellis Troy, Doris Lee, Noelle Stiles, Sahar
Elyahoodayan, Jack Whalen, Yuan Zhao, Victoria Wolseley, and Sonia Lin. I would also like to
thank my labmates at University of Michigan,Ann Arbor, These people include Aaron Reifler,
Dorsa Haji Ghaffari, Faranak Pourdanesh, Yoree Chung.
iii
I also thank my parents, Der-Ming Chang and Yu-Lan Hsu, for their faith in me and allowing me
to be as ambitious as I wanted. It was under their watchful eye that I gained so much drive and an
ability to tackle challenges head on. They have been a source of encouragement throughout my life.
Last, and most importantly, I would like to thank my wife Yi-Chen Wu. Her support,
encouragement, quiet patience and unwavering love were undeniably the bedrock upon which the
past eight years of my life have been built. Her tolerance of my occasional vulgar moods is a
testament in itself of her unyielding devotion and love.
iv
Table of Contents
Epigraph .......................................................................................................................................... i
Acknowledgement ......................................................................................................................... ii
Table of Contents .......................................................................................................................... iv
List of Figures .............................................................................................................................. vii
List of Tables ................................................................................................................................. ix
List of Abbreviations ..................................................................................................................... x
CHAPTER I: Introduction ........................................................................................................... 1
1.1 Retina..................................................................................................................................... 1
1.2 Retinal Degenerative Diseases .............................................................................................. 7
1.1.1 Retinis Pigmentosa ......................................................................................................... 7
1.1.2 Age-related Macular Degeneration ................................................................................ 8
1.3 Components of Retinal Prosthesis ......................................................................................... 9
1.3.1 Packaging ....................................................................................................................... 9
1.3.2 Electronics Modules ..................................................................................................... 13
1.3.3 Electrode ....................................................................................................................... 19
1.3.4 Camera and Video Processing ...................................................................................... 22
1.4 Types of Retinal Prosthesis ................................................................................................. 25
1.4.1 Epiretinal Prostheses .................................................................................................... 25
1.4.2 Subretinal Prostheses .................................................................................................... 29
1.4.3 Suprachoroidal prostheses ............................................................................................ 34
1.5 Retinal Neurons in Response to Electrical Stimulation ....................................................... 37
1.5.1 Direct/Indirect Activation of RGCs.............................................................................. 37
1.5.2 Selective Stimulation .................................................................................................... 38
1.5.3 Stimulation encoding .................................................................................................... 40
1.5.4 Stimulation efficiency .................................................................................................. 41
1.6 Calcium Imaging ................................................................................................................. 42
1.6.1 History .......................................................................................................................... 42
1.6.2 Delivery method ........................................................................................................... 47
1.7 Summary of Introduction .................................................................................................... 52
1.8 Thesis Overview .................................................................................................................. 53
v
CHAPTER II: GCaMP Expression Characterized Using Fundus Imaging .......................... 55
Abstract ..................................................................................................................................... 55
2.1 Introduction ......................................................................................................................... 56
2.2 Methods ............................................................................................................................... 58
2.2.1 Overview ...................................................................................................................... 58
2.2.2 Animal .......................................................................................................................... 58
2.2.3 Funduscope ................................................................................................................... 59
2.2.4 Fundus Imaging ............................................................................................................ 61
2.2.5 Imaging Processing ...................................................................................................... 61
2.2.6 Viral Vector .................................................................................................................. 63
2.2.7 Electrical Stimulation ................................................................................................... 64
2.2.8 Calcium Imaging .......................................................................................................... 65
2.3 Results ................................................................................................................................. 66
2.3.1 System Performance ..................................................................................................... 66
2.3.2 Long-term Expression Tracking ................................................................................... 69
2.3.3 In vitro Calcium Imaging Validation ........................................................................... 73
2.4 Discussion and Conclusion.................................................................................................. 76
2.4.1 Head motion and optical blurring ................................................................................. 76
2.4.2 Photobleaching of fluorophore ..................................................................................... 76
2.4.3 Cytomorbidity .............................................................................................................. 77
2.4.4 Limitations of Calcium imaging ................................................................................... 77
CHAPTER III: Strategies for Selective Activation of Retinal Ganglion Cells. ..................... 79
Abstract ..................................................................................................................................... 79
3.1 Introduction ......................................................................................................................... 80
3.2 Methods ............................................................................................................................... 83
3.2.1 Overview ...................................................................................................................... 83
3.2.2 Animal .......................................................................................................................... 83
3.2.3 Virus-transduced Calcium Indicator............................................................................. 84
3.2.4 Intravitreal AAV Injection ........................................................................................... 84
3.2.5 Calcium imaging .......................................................................................................... 85
3.2.6 Electrical Stimulation ................................................................................................... 87
3.2.7 Spatial Threshold Mapping .......................................................................................... 88
3.2.8 Statistical Analysis ....................................................................................................... 89
3.3 Results ................................................................................................................................. 91
vi
3.3.1 GCaMP Expression Profile .......................................................................................... 91
3.3.2 Calcium Transient ........................................................................................................ 93
3.3.3 Duration Manipulation ................................................................................................. 95
3.3.4 Phase Manipulation ...................................................................................................... 99
3.3.5 Waveform Manipulation ............................................................................................ 103
3.4 Discussion and Conclusion................................................................................................ 107
3.4.1 Short Duration Pulse for Direct RGC Activation ....................................................... 107
3.4.2 Symmetric Anodic-first Pulse for Localized RGC responses .................................... 108
3.4.3 Asymmetric Anodic-first Pulse for RGCs Thresholds Manipulation ........................ 108
3.4.4 Limitations of Calcium Imaging ................................................................................ 109
3.4.5 Conclusion .................................................................................................................. 110
CHAPTER IV: Current Steering to Alter RGC Responses .................................................. 112
Abstract ................................................................................................................................... 112
4.1 Introduction ....................................................................................................................... 113
4.2 Methods ............................................................................................................................. 115
4.2.1 Overview .................................................................................................................... 115
4.2.2 Current Steering ......................................................................................................... 115
4.2.3 Data Analysis ............................................................................................................. 117
4.3 Results ............................................................................................................................... 119
4.4 Discussion and Conclusion................................................................................................ 125
4.4.1 Current Steering to Create Virtual Electrode ............................................................. 125
4.4.2 Current Steering to Change Electrical field ................................................................ 126
4.4.3 Conclusion .................................................................................................................. 129
CHAPTER V: Conclusions and Future Work........................................................................ 130
5.1 Recommendations for Epiretinal Prostheses ..................................................................... 130
5.2 Future Work ...................................................................................................................... 132
5.2.1 In vivo Calcium Imaging ............................................................................................ 132
5.2.2 Light Stimulation VS. Electrical Stimulation ............................................................. 133
CHAPTER VI: References ....................................................................................................... 134
vii
List of Figures
Figure 1.1: Structure of the eye ........................................................................................... 1
Figure 1.2: Light sensitivity of retinal photoreceptors in the human retina ................... 2
Figure 1.3: Morphologies of bipolar cells ........................................................................... 4
Figure 1.4: Sample affected visual field for RP and AMD patients ................................. 7
Figure 1.5: Hermetic packaging schemes ......................................................................... 11
Figure 1.6: High density feedthrough technology ............................................................ 13
Figure 1.7: Chip microphotograph of the epiretinal stimulator .................................... 17
Figure 1.8: Exemplary system and verification platform for epiretinal prostheses ..... 17
Figure 1.9: Geometries of 3D recessed electrode ............................................................. 22
Figure 1.10: An intraocular camera in place of the crystalline lens ................................ 23
Figure 1.11: Epiretinal prostheses ...................................................................................... 29
Figure 1.12: Subretinal prostheses ...................................................................................... 33
Figure 1.13: Suprachoroidal prostheses ............................................................................. 36
Figure 1.14: Spike thresholds for several stimulation configurations ............................. 39
Figure 1.15 Crystal structure of GCaMP .......................................................................... 46
Figure 1.16 AAV pseudotype vectors ................................................................................. 52
Figure 2.1: Fundus imaging system schematic diagram .................................................. 60
Figure 2.2: Alignment the of endoscope with the mouse eye .......................................... 61
Figure 2.3: Calculation of normalized fluorescence intensity ......................................... 63
Figure 2.4: Map of pAAV2-CAG-GCaMP ....................................................................... 64
Figure 2.5: Representative fundus images ....................................................................... 67
Figure 2.6: Fundus images with different exposure setting ............................................ 68
Figure 2.7: Spatial resolution analysis of individual cells ............................................... 69
Figure 2.8: Fundus images for transgenic YFP mouse ................................................... 70
Figure 2.9: Fundus images for GCaMP6f expression ..................................................... 71
viii
Figure 2.10: Number of observeral RGCs.......................................................................... 73
Figure 2.11: Calcium transients of electrically activated RGCs ...................................... 74
Figure 3.1: Calcium imaging experimental setup ............................................................ 86
Figure 3.2: Stimulation protocol ....................................................................................... 88
Figure 3.3: Correlation tests between responses .............................................................. 90
Figure 3.4: Retinal whole mounts of WT and RD mice .................................................. 93
Figure 3.5: Calcium imaging of electrically activated RGCs ......................................... 94
Figure 3.6: Normalized ccalcium transients ..................................................................... 95
Figure 3.7: Spatial threshold maps for duration variation ............................................. 97
Figure 3.8: Strength duration curve for somatic and axonal activations ...................... 98
Figure 3.9: Threshold comparison for somatic and axonal activations ......................... 99
Figure 3.10: Spatial threshold maps for phase manipulation......................................... 101
Figure 3.11: Effective area analysis for phase manipulation .......................................... 102
Figure 3.12: Spatial threshold maps for waveform manipulation ................................ 104
Figure 3.13: Correlations of response patterns ................................................................ 106
Figure 4.1: Current steering experimental setup .......................................................... 116
Figure 4.2: Interface board for current steering experiment ....................................... 117
Figure 4.3: Spatial threshold maps for current steering (region 1) ............................. 120
Figure 4.4: Spatial threshold maps for current steering (region 2) ............................. 121
Figure 4.5: Standard orientation for responsive area analysis..................................... 123
Figure 4.6: Responsive area analysis for different return electrode setups ................ 123
Figure 4.7: Percentage of somatic versus axonal activation for current steering ....... 125
Figure 4.8: Electrical field simulation for current steering .......................................... 129
ix
List of Tables
Table 1.1: Non-viral DNA transfection methods ........................................................... 50
Table 2.1: Camera setting for the fundus imaging system............................................ 60
Table 2.2: Statistical results of responsive RGCs .......................................................... 75
Table 3.1: Electrical stimulation waveform parameters ............................................... 88
x
List of Abbreviations
AAV adeno-associated virus ...................................................................................................51
AM acetoxymethyl ................................................................................................................48
AMD age-related macular degeneration .................................................................................... 7
BAPTA 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid ...........................................43
BC bipolar cell ......................................................................................................................78
CaM calmodulin ......................................................................................................................44
CBA chicken β-actin ...............................................................................................................63
CCD charge-coupled device ....................................................................................................62
cGMP cyclic guanosine monophosphate .................................................................................... 3
CMOS complementary metal-oxide semiconductor ...................................................................16
DBS deep brain stimulation ...................................................................................................121
DNA deoxyribonucleic acid .....................................................................................................48
EGTA ethylene glycol-bis(2-aminoethylether) -N,N,N',N'-tetraacetic acid ...............................43
EMCCD electron-multiplied charge-coupled device .....................................................................65
FDA food and drug administration ..........................................................................................26
FRET fluoresncence resonance energy tranfer ..........................................................................44
GABA gamma-Aminobutyric acid .............................................................................................. 3
GCL ganglion cell layer............................................................................................................ 1
GECI genetically encoded calcium indicator ...........................................................................44
GFP green fluorescence protein ..............................................................................................42
GMP guanosine monophosphate .............................................................................................. 3
GPi globus pallidus ..............................................................................................................122
IC integrated circuit … . … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … .15
INL inner nuclear layer ............................................................................................................ 1
IPL inner plexiform layer ....................................................................................................... 1
IR infrared ............................................................................................................................27
ITO indium tin oxide ..............................................................................................................64
LCA Leber congenital amaurosis ............................................................................................32
LED light-emitting diode .........................................................................................................65
LN linear-nonlinear ..............................................................................................................40
MEA multielectrode array ........................................................................................................55
MPDA microphotodiode array ....................................................................................................30
NIR near-infrared Spectroscopy .............................................................................................32
OAT organic anion transporter ................................................................................................49
ONL outer nuclear layer ........................................................................................................... 1
xi
OPL outer plexiform layer ....................................................................................................... 1
PDE phosphodiesterase ..........................................................................................................81
PPN pedunculopontine nucleus ............................................................................................122
RD retinal degeneration ........................................................................................................89
RF radio frequency ..............................................................................................................13
RGC retinal ganglion cell ........................................................................................................ 5
RNA ribonucleic acid .............................................................................................................49
ROI region of interest ............................................................................................................62
RP retinis pigmentosa ........................................................................................................... 7
RPE retinal pigment epithelium ............................................................................................... 7
SLAM simultaneous localization and mapping ..........................................................................23
SNR signal to noise ratio ........................................................................................................42
STN subthalamic nucleus .....................................................................................................122
UV ultraviolet .......................................................................................................................43
Vim ventro-intermediate nucleus.........................................................................................122
VTA volume of tissue activated ...........................................................................................121
WPRE woodchuck hepatitis virus post-transcriptional regulatory element ...............................63
WT wild-type ........................................................................................................................58
YFP yellow fluorescent protein .............................................................................................45
1
CHAPTER I: Introduction
1.1 Retina
The retina is a thin, light-sensitive layer of tissue located on the back of the eye. Light entering
the eye will be first refracted by the cornea and the lens, further forming a focused light rays that
incident onto the retina. The light information is converted to corresponding biological signals until
it passes through all transparent layers in the retina and finally reaches the photoreceptors on the
rear surface of the retina. After the phototransduction, the neural signals are back-propagated and
preprocessed from the posterior to anterior of retina, through photoreceptor layer, outer nuclear
layer (ONL), outer plexiform layer (OPL), inner nuclear layer (INL), inner plexiform layer (IPL),
retinal ganglion cell layer (GCL), and ultimately the nerve fiber layer (Figure 1.1).
Figure 1.1: Cross section of the adult human eye and retina. Light enters through the cornea, is
focused by the lens, and travels through all the layers before reaching the photoreceptors.
Photoreceptors convert the incoming light to neural signals, which are propagated and preprocessed
2
through each layer of the retina to the ganglion cells, and finally conveyed to the nerve fiber layer
and optic nerve. Image modified from Webvision, http://webvision.med.utah.edu/.
The photoreceptors consist of two types of cells: cones and rods. Rod receptors mainly operate
at low light level so are mostly responsible for night vision. The spectral sensitivity of rods is
centered at approximately 500nm (Figure 1.2). In human retina, there are roughly 92 million rod
photoreceptors distributed outside the fovea, maximally at the parafoveal region. (Curcio et al.,
1990) Cones operate at room light and daylight levels and mediate color vision. The number of
cone types greatly varies by species. Primates have three types of cones – short (blue), medium
(green), and long (red) - each detects photons of different but overlapping wavelengths, spanning
the visual spectrum (400-700 nm) (Figure 1.2). The human retina contains roughly 5 million cones
which are concentrated primarily inside the fovea (Curcio et al., 1990).
Figure 1.2: Normalized light sensitivity of retinal photoreceptors in the human retina. Maximal
sensitivity of opsins are 498nm for rhodopsin (black), 420nm for S-opsin (blue), 534nm for M-
opsin (green), and 564 for L-opsin (red) respectively. Image from Wikiwand,
http://www.wikiwand.com/en/Photoreceptor_cell.
3
The machinery of phototransductions for both rods and cones is similar. When photoreceptors
receive light, photons are absorbed by G protein-coupled receptor called opsin, leading to a
conformational change. Once activated, opsin initiates a secondary messenger cascade through
regulatory protein transducin, which activates phosphodiesterase, thereby leading to break-down
of cyclic guanosine monophosphate (cGMP) to GMP. Decreased concentration of cGMP causes
cGMP-gated sodium channels to close which then results in the hyperpolarization of the membrane
(Pugh and Lamb, 1993). Since each step in the cascade process contributes to amplification of the
signal, the photoreceptors, especially rods, are capable of detecting a single photon of light in a
dark environment (Rieke and Baylor, 1998).
Bipolar and horizontal cell located in ONL layer are the two groups of cells synaptically
connected with photoreceptors. In human retina, at least three types of horizontal cells that are
identified to mediate color opponency through an inhibitory network (Ahnelt and Kolb, 1994). In
a negative feedback loop, they are hyperpolarized by lessening release of glutamate from
photoreceptors in response to light, and therefore secrete less gamma-aminobutyric acid (GABA),
thus leading to depolarization of nearby photoreceptors. (Demb and Singer, 2015) They also
synapse onto bipolar cells in a feedforward inhibitory network that may partially contribute of the
center-surround response (Yang and Wu, 1991).
Bipolar cells are the primary neurons that relay the neural signals from photoreceptors to
ganglion cells; therefore, their dendrites terminate within the OPL, their somas are located in the
INL, and their axons descend into IPL. There are more than 10 types of bipolar cells in the
mammalian retina (Figure 1.3), and at least 9 distinct forms of cone bipolar cells and one rod bipolar
cell were identified in human (Euler et al., 2014). Depending on the polarity of their light response,
bipolar cells can be mainly divided into ON-center and OFF-center cells. Generally, when light
strikes photoreceptors, ON cells become active in response to less release glutamate, whereas OFF
bipolar cells hyperpolarize. Furthermore, bipolar cells can also be labelled as chromatic or
4
achromatic based on their presynaptic contact with cone or rod types, respectively (Figure 1.3).
Typically, the mammalian IPL can be subdivided into five strata of equal thickness where bipolar
cells form synapses with ganglion cells. Generally, ON bipolar cell terminals terminate within the
half of the IPL closer to ganglion cell layer, while OFF bipolar cell axon terminals are found in the
other half (Figure 1.3). Bipolar cells with terminals closer to the edges of the IPL typically exhibit
sustained response to light for both ON-and OFF-types, whereas those located in the center of the
IPL exhibit more transient responses (Baden et al., 2013). The information in dendritic morphology
as well as the shape and IPL stratification level of a bipolar cell’s terminal system serve together
as the best indicator of cell type identity (Euler et al., 2014).
Figure 1.3: Morphologies of the 12 types of cone bipolar cells and the rod bipolar cell in the
mouse, which are arranged according to their IPL stratification level. Some of the functional
differences between bipolar cell types are indicated below this schematic. Image modified from
(Euler et al., 2014).
Similar to bipolar cells, amacrine cells are also mainly located in the INL, though some cells
somas lies along the border of the IPL or retinal ganglion cell layer (Vaney, 1990). Amacrine cells
are predominantly responsible for integrating, modulating and interposing a temporal domain to
the visual information presented to ganglion cell through release of inhibitory neurotransmitters
5
GABA or glycine (Hartveit, 1999). They have also been found to synapse back to rod bipolar cells
as reciprocal input. Based on their dendritic field diameters, amacrine cells can be categorized as
narrow-field (around 70 µm), medium-field (around 170 µm), and wide-field (around 350 µm).
Amacrine cells function differently, according to the size of dendritic arbor. Narrow field amacrine
cells contribute to vertical communication across retina and form functional subunits with the
ganglion cells in the receptive field. The interaction between local subunits and their overlapping
subunits enable certain types of retinal ganglion cell (RGC) to detect a small spot or movement in
the visual field. Medium field amacrine cells also communicate vertically among different retinal
levels. Since the size of their dendritic fields is similar to RGCs, they possibly allow blurring of
ganglion cell visual field, though their overall function remains unknown. Wide field amacrine
cells have not been studied thoroughly due to their sparse distribution across the retinal; however,
the coverage of dendritic field suggests that their primary functions is lateral communication within
a layer (Masland, 2012).
Ganglion cell are the last stage of neurons in retina that collectively summarize the neural inputs
from bipolar and amacrine cells and transmit the signals to higher visual processing center in brain.
While their somas reside in the retinal ganglion cells layer, the dendrites stratify into the IPL and
the axons converge at optic disk to form optic nerve. Similar to bipolar cells, depending on the
types of inputs, they have center-surround receptive field with varying spatial coverage, and can be
mainly identified as ON or OFF cells on the basis of their light responses (excluding some with
special function, such as pupillary light reflex). In primates, there are two major classes of ganglion
cell, the P-cells (Parvocellular, or P pathway) and M-cells (Magnocellular, or M pathway) each
include ON and OFF types. P-cells have relatively smaller receptive field and slow conduction
velocity, and respond to changes in color with great contrast, whereas M-cells have much
larger receptive fields and fast conduction velocity, and can respond to low-contrast stimuli but are
insensitive to changes in color. (Kandel, 2013) The distribution of ganglion cells varies greatly
6
between different species, even in mammals. Common experimentally-used rodents such as mice
and rat have only one layer of ganglion cells, while primates and animals have a fovea structure in
the retina with as many as 6 layers of ganglion cells located in parafoveal region. The fovea enables
high acuity vision, since the inner retina and ganglion cells are displaced (into the parafovea) and
light is relatively unimpeded prior to contacting the photoreceptors.
7
1.2 Retinal Degenerative Diseases
Degeneration of retinal neurons, especially photoreceptors, is a major cause of permanent
visual impairment. Retinitis pigmentosa (RP) and age-related macular degeneration (AMD) are two
prevalent degenerative diseases of retina that lead to significant visual loss (Hartong et al., 2006,
Gehrs et al., 2010). Epidemiologic studies have showed that the RP afflicts both children and young
adults while AMD occurs predominantly in the elderly (Wong et al., 2011). Together, they account
for millions of cases of blindness worldwide. Figure 1.4 demonstrates the visual loss patterns for
both disease.
Figure 1.4: Sample affected visual field for RP and AMD patients. (A) Visual field of subject
with normal sight. (B) Central vision loss of AMD patients at early stage. (C) Peripheral vision
loss of RP patients at early stage. Images adapted from (Yue et al., 2016).
1.1.1 Retinis Pigmentosa
RP is identified by a progressive degeneration of rod photoreceptors beginning in the peripheral
retina, causing early symptoms such as loss of peripheral and night vision at early stage (Hartong
et al., 2006). With disease progress, degeneration of rods is followed by deterioration of cones and
retinal pigment epithelium (RPE), ultimately resulting in visual functional decline from tunnel
vision to complete blindness. RP is estimated to affect 1.5 million individuals around the world
(den Hollander et al., 1999) and many factors contribute to its occurrence; it can be syndromic,
secondary to other systemic diseases, or a genetic disorder (Daiger et al., 2014, Hartong et al., 2006).
8
A wide range of hereditary genetic mutations that encode the protein related to phototransduction
have been associated with RP phenotype (Daiger et al., 2014).
1.1.2 Age-related Macular Degeneration
AMD degeneration occurs primarily in the macula, which located in the center of retina, affects
primarily the cone photoreceptors, and leads to blurred central vision. As the disease progresses,
the affected region may grow larger and the blurred spot might ultimately turn into a scotoma. With
the aging population, AMD has been estimated to affect 3 million people in United States and 196
million people worldwide respectively in 2020 (Klein et al., 2011, Wong et al., 2014). AMD can
mainly be categorized into two forms: the dry form and wet form. The dry (atrophic) type affects
approximately 80%-90% of individuals with AMD. It can be identified by drusen, small white or
yellowish deposits, forming under the retina in the macula and ultimately causing photoreceptor to
degeneration. The wet form account for the other 10-15% of AMD occurrence and mainly can be
attributed to abnormal neovascularization from choroidal vessels underneath the macula. These
new blood vessels are fragile, tend to break, bleed, and leak fluid, thus damaging the macula and
resulting in a rapid and severe loss of central vision.
9
1.3 Components of Retinal Prosthesis
Several retinal prostheses have been tested in clinical trials and two systems have regulatory
approval, but many technical challenges need to be resolved to enable a long-lasting, high
resolution device. Microelectronics cannot survive long-term exposure to water and ions inside the
body, thus new materials or processes are required for forming thin and robust isolation barriers to
protect the electronics. To support the increased number of electrodes, more efficient electronic
circuits with safe levels of power consumption are needed for parallel stimulation of 100s of
individual contacts. The electrode array, as the main interface between implant and retina, can be
improved from many aspects. The electrode material must support higher charge density due to the
reduction of electrode size and safety concerns with current materials. To improve the attachment
as well as alleviate damage, the electrode substrates need to be flexible for close fit to the curvature
of the retina. In addition, electrical stimulation patterns can be further optimized for ideal visual
perception and power consumption reduction. Finally, the usage of state-of-art camera technology
and advanced algorithms for video processing has great potential to compensate for the limited
visual task performance shown by patients using lower resolution devices.
1.3.1 Packaging
Perhaps the most important technology for any medical implant is the hermetic packaging. This
requires a set of materials be used to form a barrier that is virtually impervious to penetration by
water or ions. Since a perfect barrier for an infinitely long time is not possible, instead leak rate
specifications are developed based on empirical and theoretical models, too detailed to review here,
but see (Vanhoestenberghe and Donaldson, 2013) and (Jiang and Zhou, 2010). These specifications
ensure that a high percentage (greater than 99%) of implanted devices will last for decades in the
body. Generally, encapsulation and hermetic enclosure are two commonly used methods that were
described by Donaldson (Vanhoestenberghe and Donaldson, 2013) (Figure 1.5).
10
1.3.1.1 Encapsulation
Encapsulation relies on using a conformal layer of material(s) to coat the electronics (Figure
1.5: Top). Metals and ceramics are the best water barriers, but depositing thin films of these
materials in a conformal manner is technically challenging. Polymers, such as silicone, can be
applied conformally, but have poor water vapor transmission rates (i.e. water vapor readily passes
through silicone). However, good protection can be achieved if the polymer coating is tightly fit to
the surface of the electronics, and if water cannot condense on the surface after penetrating the
polymers. One of the drawbacks of this strategy is the requirement for a level of surface cleanliness
which is extremely difficult to achieve in practice (Vanhoestenberghe and Donaldson, 2013). If the
chip surface is not free of particulates or if there is the presence of voids in the conformal coating,
then water can condense at these imperfections and ultimately lead to failure of electronics. As
evidence, some reports shows that the retinal prosthesis devices encapsulated with polymers can
only sustain for 1-2 years, thus suggesting process development must be improved before
encapsulation become a viable approach.
Most of the current research on improving the performance of encapsulation focuses on
replacing polymer-based film with different materials. A multi-layer, multi-material film composed
of both diamond-like coating and metal films has been proposed as an alternative to polymer
coatings (Weiland et al., 2013). The structure demonstrates robust ion barrier properties and good
conformality around the corner and edge of the coating, though long-term testing has not been
completed. Ultra-nano-crystalline diamond film has also been used to form an inert and thin coating
layer; however, a better dielectric material might be needed to embed into the layer due to the
insufficient insulating ability at high voltage (Xiao et al., 2006). Recently, amorphous silicon
carbide (a-SiCx:H or a-SiC) has been used by several groups, due to its excellent dielectric property,
resistance to degradation, and capability of deposition when fabricating (Cogan et al., 2003, Sharma
et al., 2012). However, the deposition rate for SiC is relatively slow (0.2 to 0.5 µm/hr) at low
11
temperature (< 400°C), and the compressive stress of a-Sic intrinsic (0.2 to 0.3 GPa) might limit
the maximum film thickness (< 5 µm) to guarantee good adhesion and prevent from device
distortion. In summary, thin-film encapsulants remain a research topic that has tremendous
potential, but no clear solution at this time.
Figure 1.5: Hermetic packaging schemes. Top: Encapsulation uses conformal coating coat over an
electronic chip for protection, but to date no coating technology has proven adequate for long-term
implantation. Bottom: An electronics module is placed inside an enclosure, which includes a
feedthrough platform with conductors and a case or lid. The feedthrough and case are sealed
together.
1.3.1.2 Enclosure
Hermetic enclosures demonstrate relatively robust performance thus almost all clinically
approved implantable neurostimulation devices adopt this approach. Traditionally, the titanium or
ceramic cases (like a shell) are accompanied by a feedthrough (a substrate with isolated conductors
channeling the electrical signals across enclosure); combined, the case and feedthrough form a
complete enclosure (Jiang and Zhou, 2010) (Figure 1.5: Bottom). However, compared with the
enclosed electronics, case thickness and feedthrough conductor spacing (pitch) enlarge the size of
implants significantly. For example, the Argus II uses this style of packaging and the implant size
12
is determined largely by the size requirement for the feedthrough substrate that has 60 independent
stimulus channels. The Argus II hermetic enclosure represents a great engineering accomplishment,
specifically, a 10X reduction in volume and 3X increase in independent channels, compared with
the previous state-of-the-art, which was the cochlear implant. Yet, contrast the dimensions of the
feedthrough features (100s of microns), with the minimum size of other electronic components
used in implants (integrated circuit pads can be made 50 µm diameter, electrodes can have features
below 10 µm, and integrated transistor sizes are less than 0.1 µm), and it is clear that a packaging
technology defines the size of the implant, and relatively large available technology limits further
size reduction of the Argus II and other implants that use enclosures. Functionally, the number of
channels will be limited, thereby limiting the best possibly visual acuity. Additionally, the present
size does occasionally lead to conjunctival erosion, which may occur less with a smaller implant.
Therefore, improving hermetic packaging technology is critical for both improving resolution and
increasing biocompatibility.
Advanced processing techniques for enclosures have been developed to increase the density of
conducting channels. Suaning et al. proposed a method that places patterned platinum foil to form
lines between two alumina sheets (a ceramic commonly used in medical implants) with gaps filled
with alumina particles suspended in viscous liquid (Suaning et al., 2006). Subsequent high-
temperature operation promotes crystal growth in alumina, thus fusing the sheets. To connect with
the internal electronics, holes corresponding to the bond pads are drilled through one of the alumina
sheets. Schuettler et al. achieved 360 channels in a feedthrough with dimension less than 25 mm
2
by a screen printing approach (Schuettler et al., 2010). Another group fabricated high density
feedthrough with a stack of alumina layer and platinum wire in interlocking pattern.(Gill et al.,
2013) (Figure 1.6). After proper heat temperature operation and compression, the transverse
cutting section normal to the wires results in a patterned grid of platinum conductors in order, with
13
helium leak rate less than 8×10
-11
mbar-l/s, which is generally considered an acceptable leak rate,
though this must also consider the implant volume.
Figure 1.6: High density feedthrough technology. Front view shows patterned grid of platinum
conductors embedded in alumina substrate. Image from (Gill et al., 2013).
1.3.2 Electronics Modules
The electronics used for retinal prostheses are varied and encompass several subspecialties of
electrical engineering. The connection between the camera and the electrode array constitutes a
signal chain. The most complex signal chains involves the following steps: 1) conversion of video
data into digital data format 2) processing of the digital camera data to convert brightness detected
in a particular region into a level of stimulation to be applied to the retina 3) encoding of the
stimulation levels for all electrodes into a serial data stream 4) generation of a wireless, radio
frequency (RF) transmission signal that carries both power and the serial data stream to the implant
5) RF energy transmission between a pair of well-aligned inductive coils, one external and one
implanted 6) power recovery and data decoding from the received RF signal 7) generation of
stimulation pulses based on the camera input. Steps 1-4 are done in a wearable external system,
14
steps 6, 7 are done in the implant and step 5 links the two. The signal chain can be simplified, in
some ways, by placing the light sensing element inside the eye, which will eliminate some of the
steps above, but add complexity to the implant design, which in this case must have electronics for
photon detection as well as stimulus current generation.
For those systems using an external camera, both camera and supporting hardware do not
represent significant technical challenges. Rapid advances in cell phone camera quality and
miniaturization benefit the external system by making available camera hardware that meets the
specifications in terms of pixel density. In contrast, greater design challenges remain for the
wireless transfer of data and power and for the design of efficient stimulator electronics. Reviews
of these areas are listed below.
1.3.2.1 Telemetry
Wireless transmission of data and power has been adopted for most retinal prosthetic systems,
due to the requirement for small size and continuously refreshed data. Devices that operate on
batteries, such as deep brain stimulators, are implanted in the upper trunk thus have room to
accommodate batteries, whereas the orbit is space constrained. Retinal implants must be constantly
fed with the latest camera data, so a wireless data link is required. Deep brain stimulators run on
pre-set stimulation parameters and only update data during a clinic visit using special external
programming hardware.
Since the wireless power and data transmission has the potential to interfere, design of retinal
prostheses should consider power and data together. However, a fundamental conflict exists when
choosing the optimal wireless transmission frequency. Data and power are often transmitted using
the same radio frequency signal. Data is encoded by modulating the signals amplitude or phase.
The implant electronics recovers both power and data from this signal, for implant operation. For
advanced retinal prosthesis, the data rate should be on the order of megabits/sec and the carrier
frequency to reliably support such a data rate is typically set to an order of magnitude higher, say
15
10MHz (Zhou et al., 2008). Yet, to maintain the power transmission efficiency, the frequency need
to be set under 10MHz due to signal attenuation across tissue and internal AC-DC conversion
(Wang et al., 2006).
Multi-band transmitters which separate the frequencies for power and data individually have
been regarded as one of the most effective approaches for retinal prosthesis RF transmission. In
this approach, different transmitter and receiver coils were designed and embedded in the device
for power and data links (Chen et al., 2013a). The power signal requires larger coils and operates
at lower frequency for efficient power transmission, whereas the data signal can be transmitted
through a smaller coils at higher frequency, since the data signal demands lower power. Although
the efficiency can be dramatically improved, such a system requires careful design to prevent
interference or crosstalk between the links. Previous study has shown that a 256 channel retinal
prosthesis system with dual band telemetry can be realized on the bench top, thus indicating that
the technique might be feasible in a clinical device (Chen et al., 2013a).
Optical transmission of both power and data have also been studied for other retinal implants.
For a subretinal prosthesis, a optobionics microphotodiode array was proposed to directly convert
incident light to electric power, but the generated power is not sufficient to activate neurons on the
retina (Palanker et al., 2005). The other optical application was presented by Gross et al. who
combined the infrared optical data link with inductive power link (Gross et al., 1999). This design
was ultimately included in the IMI epiretinal prosthesis. The optical data link can achieve a data
rate of 200Kb/sec.
1.3.2.2 Integrated circuit
Depending on the type of retinal implant, generally, the integrated circuit (IC) needs to perform
multiple functions, including AC-DC conversion (if using inductive power), data demodulation (if
using wireless data transmission), digital control, analog voltage or current stimulus, and reverse
telemetry (from the implant to an external system) for diagnostic information about the implant.
16
Among them, the design of voltage drivers is challenging since both positive and negative voltages
are needed for supplying adequate anodic and cathodic current for charge balanced stimulation.
The size and power consumption are other competing requirement for the stimulator chip.
Liu et al. have designed, fabricated, and tested several generations of multichannel stimulator
IC for retinal prostheses. The most recent contribution of the group is to completely embed the data
demodulation, timing controlled rectification (power efficient AC-DC conversion), digital control,
and 256-channel independent stimulus drivers into a compact system (Chen et al., 2010) (Figure
1.7). In particular, the timing-controlled rectifier eliminates a great amount of power loss, when
compared to a typical diode-based rectifier, during AC-DC conversion. IC area is reduced with the
advance of circuit-under-pad layout which utilizes multiple metal layers in the process to place
bond pads over the circuitry. This platform was validated as functional in a benchtop end-to-end
system validation (Chen et al., 2013a) with real-time visual feedback and wireless power/data links
(Figure 1.8).
Ortmanns and colleagues have created a 232 channel stimulator chip which was manufactured
in 0.35 µm complementary metal–oxide–semiconductor (CMOS) fabrication (Ortmanns et al.,
2006). The overall size of the chip is less than 5x5 mm and its high-voltage feature allows high
stimulus current through small electrodes. They also constructed a charge-balancing scheme to
balance the current sources without the need for large capacitors on every output, thus reducing the
accumulated charge on electrodes which might ultimately results in neural injury and saving space
on the overall implant size. The balance pulse will be triggered when recorded electrode voltages
between pulses exceed ±50 mV.
17
Figure 1.7: Chip microphotograph of the 256-channel epiretinal stimulator manufactured by Liu
and colleagues. The chip has integrated power conditioning and data decode, and 256 independent
output channels. Image from (Chen et al., 2010).
Figure 1.8: The exemplary system diagram of the verification platform for epiretinal prostheses
with multi-band approaches. Image from (Chen et al., 2013a).
A chip with 256 channels has been developed in 0.18 µm CMOS, as part of a subretinal
prosthesis (Shire et al., 2012). Similar to other epiretinal applications, the chip contains power/data
18
telemetry modules and reverse low-rate data link for monitoring the electrode voltage on any
outputs and transmissions. This chip is not designed to reside underneath the retina, rather the
proposed system uses a microelectrode array as a retinal interface. The drivers are designed with
high voltage components allowing of a wide range of stimulus current and can be configured as
sources or sinks arbitrarily for current steering.
Another retinal stimulator with 96 channels has been designed for allowing current steering
between local sources and sinks which presumably can facilitate focused stimulation to the intended
site (Dommel et al., 2009). The voltage drivers of this chip are implemented by specific high-
voltage transistors which are fabricated in 0.35 µm process, whereas the remaining chip feature are
operated with low-voltage transistors for space and power saving. For localized stimulus, the chip
contains special switch design that allows any electrode to be configured as either a center current
source or within a group of 6 surrounding sinks forming a hexagonal pattern near the source. This
concept has been tested on a test chip which accommodates 14 hexagonal electrode mosaic with
switching and two current drivers. The other feature of this chip is to short all electrode for
dissipating accumulated charge during the period between stimuli. This is another approach to
ensure charge balance without the need for large capacitors on each output.
A high-density 512-channel retinal stimulator chip has been developed for supporting higher
resolution retinal prostheses (Monge et al., 2013). A novel feature of this chip is auto-calibration
circuitry on the output, which can be used to improve stimulation precision and eliminate charge
accumulation. The use of 65-nm transistors in the design allows an increase in output channels as
well as a reduction in chip size to 4.5 × 3.1 mm
2
, which has the potential to be installed entirely
inside the eye if proper packaging techniques are used. The whole chip is operated at ±2.5 V to
reduce power consumption. However, with this choice of a low voltage supply, the output current
range might be limited.
19
1.3.3 Electrode
With the advance of retinal prostheses, such as the Argus II and Alpha-IMS, electrode
technology, especially in term of fabrication, has been improved significantly for implantable
bioelectronics. Conventional implants, such as deep brain stimulator cochlear implants, only
require hand-made electrode assemblies, featuring several platinum contacts supported by a
polymer substrate. In contrast, due to the requirement for a comparatively high number of densely
spaced contacts, to match the 2-D retinal structure, photolithography and micromachining
techniques have been used to fabricate these arrays. In addition to using advanced manufacturing
techniques, the electrode materials have been improved to compensate for limited electrode size
and allow adequate charge injection capability. Argus II and Alpha-IMS take advantage of state-
of-art platinum gray and titanium nitride, respectively. Both are superior to bulk platinum, which
is typically used for neurostimulation. Compared with 0.1-0.35 mC/cm
2
charge injection capability
for platinum (Rose and Robblee, 1990), the platinum gray and titanium nitride has pushed the limits
to 1mC/cm
2
and 0.9mC/cm
2
(Zhou et al., 2013, Weiland et al., 2002). The electrode array substrate
serves to hold in place the electrodes. Since this serves as a mechanical interface with the retina,
care must be taken to avoid a design that will damage the retina, which is a very delicate tissue.
Most devices use polymer substrates that have an integrated cable and leads to connect the
electronics with the electrode array. In contrast, the Alpha-IMS has electrodes patterned directly
on the silicon IC. This has the advantage of avoiding complex routing schemes to connect the
electronics and the array, but the silicon IC is rigid compared to a polymer array, and can potentially
damage the retina.
Visual field and acuity is another fundamental problem facing retinal prostheses. The structure
of human retina is a roughly 2.5cm diameter hemisphere attaching to the back of the eye, spanning
approximate 60° nasal and superior field, about 70° inferior field, as well as 90° temporal field
(Barton and Benatar, 2003). To completely cover the entire retina, an electrode array with π×(d/2)
2
20
or almost 5cm
2
dimension is required. For an epiretinal prosthesis, the array size is mostly limited
by the incision that can be safely made in the implant surgery (less than 5mm). A wide field
foldable array that can be inserted through a 5mm eye incision has been proposed by Ameri et al
(Ameri et al., 2009). This approach can potentially provide a visual field of 34° once it expands in
eye, compared with 19° for Argus II and 11° for Alpha-IMS. Based on this concept, subsequent
development has been tested through long-term implantation in animal eyes to validate its
feasibility (Zhang et al., 2013). On the other hand, subretinal implant size is mainly limited by the
risk of retinal detachment attributed to the insertion of an array underneath the retina. With the
increase of subretinal array size, a higher risk of detaching the entire retina can be expected. Retinal
detachment has disastrous and irreversible consequences for retina health, so the surgery process
for the subretinal implant Alpha-IMS includes a silicone oil injection to mitigate the likelihood of
retinal detachment (Stingl et al., 2013b).
Visual acuity is used to measure the spatial resolution of a visual system. For people with
normal visual acuity (20/20 vision), two points separated by 1 arcmin, equivalent to 4.5 µm of
retina, can be resolved. Based on this finding, the electrode pitch (the distance between the centers
of two adjacent electrodes), size, and electrode-retina contact would contribute the perception of
retinal implant users. The best reported visual grating acuity for Argus II and Alpha-IMS user are
20/1260 and 20/546 respectively which roughly matches the theoretical limits of electrode spacing
(525 µm and 70 µm) (Yue et al., 2016). Practically, since a threshold amount of charge is required
to create a perception of light, the electrode size cannot be selected arbitrary small (since the amount
of safe charge is reduced with electrode area). The contact between retina tissue and electrode is
another factor for acuity, since the retina, along with the surrounding physiological vitreous, can
be regarded as inhomogeneous conductive medium. The vitreous (or saline after vitrectomy) is
more conductive than the retina and underlying tissue. Thus, there exists the scenario that stimulus
current, from an electrode away from the retina, preferentially passes through the vitreous, parallel
21
to the retina and disperses current significantly. Special recessed electrodes have been developed
for constraining the electric field or increasing selectivity (Suesserman et al., 1991, Wilke et al.,
2011b). For the subretinally implanted recessed 3D electrode, the bipolar cell has further been
found to migrate into the cavity of electrode, thus inducing a more selective and robust neural
interface although device removal would be more complicated (Butterwick et al., 2009, Djilas et
al., 2011) (Figure 1.9). Some groups use localized return electrodes near the stimulating electrode
to confine the stimulus current and achieve focal activation of neurons (Palanker et al., 2005).
Recently, Habib et al. (2013) tested the performance of a hexagonal electrode array that has one
stimulating electrode surrounded by 6 “guard ” (return) electrodes through a whole mount rabbit
retina model. The results show that the difference of RGC thresholds inside and outside of the hex
guards can be enlarged to 2 fold higher, implying the ability for localized stimulation (Habib et al.,
2013). However, since the proximity between stimulating and return electrode caused current
shunting though more conductive saline rather than retina, the required current threshold is elevated
as the stimulation efficiency declines compared with simple monopolar configuration (which uses
a distant return electrode and forces current to pass through the retina).
22
Figure 1.9: Geometries of 3D recessed electrode, fabricated electrode array, endoscopic control of
the subretinal implant position, and neuronal integration of the P23H rat retina in a 3D retinal
implant well. (D) Blue DAPI nuclear staining shows the cell nuclei integrating the well. Red GFAP
immunolabeling shows the absence of a major retinal gliosis at the site of integration in retinal
implant well. Green Goα immunolabeling visualize ON bipolar cells integrated into the implant
well (arrows). Images reproduced from (Djilas et al., 2011).
1.3.4 Camera and Video Processing
State of the art cell phone cameras have vastly superior resolution compared with the retinal
implant electrode array; therefore, little engineering research is needed on camera technology,
except for incorporating depth information to aid in detection of important obstacles. However,
since the visual perception integrates not only the signal from the retina, but also the motion of
eye/head and proprioception, camera positioning becomes a critical issue for prosthesis user to
obtain a more natural perception. Video processing algorithms have potential to improve the
performance of retinal prostheses through selectively highlighting important areas (nearby objects)
and attenuating irrelevant background scene (far in distance).
1.3.4.1 Camera Positioning
The Alpha-IMS has imaging capability positioned underneath the retina. The generated vision,
eye and head coordinates naturally since the imaging components are located inside the eye, thus
move along with the eye. In contrast, the cameras of epiretinal implants are usually externally
installed on a glasses frame, so head movement is needed to refresh an image. Although no
controlled comparison has been studied in patients, the head-mounted camera seems to be less
instinctive and long-term training may be required for adaptation.
23
1.3.4.2 Novel Camera and video processing
The potential of depth cameras has been generally validated in the computer vision field and
in commercial products such as the Microsoft Kinect. Some studies have extended this technology
to the retinal prosthesis conferring the ability to capture the relative distance between user and
objects. McCarthy et al. has established a visual representation that augments intensity of regions
of the current scene, based on the local variation of surface originated from the RGB and depth
information (McCarthy et al., 2013). Their results show that potentially important obstacles such
as a low wall near a sidewalk can be accentuated and presented in the simulated prosthetic vision
paradigm. Moreover, the depth information can be used as input to simultaneous localization and
mapping (SLAM) algorithms which construct a map of an unknown environment while
simultaneously keeping track of the location of user (Pradeep et al., 2009). This concept might
ultimately be transformed into guidance clue to help retinal prosthesis users to pass through a
complicated environment.
Figure 1.10: An intraocular camera positioned in place of the crystalline lens, similar to what is
done in cataract surgery when an intraocular lens is implanted. Such a device will allow scanning
of the visual field with eye movement, thus create more natural vision.
For epiretinal implants, a camera that can directly replace the role of the crystalline lens might
be a solution for more natural vision. Nasiatka et al. has proposed an intraocular camera that can
24
be implanted in the position of the crystalline lens (Nasiatka et al., 2010) (Figure 1.10). Due to the
limitation of space, a small cylindrical hermetic capsule about 3 mm in diameter and 4 mm long
was designed to fit in the place and an aspherical lens with extremely short back focal length (~
1mm) was used to form image on the wide-dynamic rage CMOS sensor. With the power and control
circuit behind the sensors, the visual information can be delivered to the stimulation chip, further
driving the related stimulation patterns on the electrode array. Simulation of the intraocular camera
suggests that with the control-ability of eye movement, the user can perform visual tasks in a more
natural manner (Weiland and Humayun, 2014).
25
1.4 Types of Retinal Prosthesis
Although state-of-art retinal prosthesis systems share similar basic architectures and
components, their actual designs are diverse, mostly depending on the locations of the implanted
electrode array and the complexity of the electronics. Three main types of retinal implants have
been developed: epiretinal prostheses, anchored to the inner surface of retina; subretinal prostheses,
embedded between retina and RPE/choroid; and suprachoroidal prostheses, implanted between
choroid and sclera.
1.4.1 Epiretinal Prostheses
Epiretinal prostheses, with its electrode array sitting on the innermost layer of retina, have
several advantages in terms of surgery, heat dissipation, and distance to the target neurons. Since
the array contacts the retina on the inner surface with vitreous cavity nearby, the surgical procedure
is relatively easier and risks of mechanical damage to the retina are greatly reduced. The fluid in
the vitreous cavity can functionally serve as an extra heat sink that enhances the dissipation of heat
generated by the electronics of the implant, lowering the thermal risks of tissue for chronic use
(Opie et al., 2012). Moreover, as mentioned in previous section, as RGCs are the main targeted
neurons, the proximity between the electrode-tissue interface makes direct electrical stimulation
much easier, thus potentially lowering the power consumption. Drawbacks of epiretinal prosthesis
include the difficulty of forming the proximate interface between array and retina, as well as the
non-ideal elongated visual percepts due to activation of the RGC axons; however, the former might
be ameliorated by the redesign of flexible electrode array whereas the latter is highly possibly
improved through using different stimulation patterns.
Owing to the advantages outweighing the disadvantages, a large share of the efforts in retinal
prosthesis development has been directed towards the epiretinal type. Leading groups include
Second Sight Medical Products (SSMP) Inc. in Sylmar, California, USA, Intelligent Medical
26
Implants (IMI) GmbH (acquired by Pixium Vision SA., France) and EpiRet GmbH in Germany.
Devices from these three groups all adopt external cameras with wireless power and data
transmission to avoid transcutaneous wires.
Two generations of devices, Argus I and Argus II, have been developed and released by SSMP.
Argus I was approved as the first-generation epiretinal prosthesis for an investigational clinical trial
by the United States Food and Drug Administration (FDA). With the studies on the Argus I
demonstrating the function and safety of long-term electrical stimulation, the more advanced Argus
II retinal implant was developed, then evaluated in a multi-center clinical trial, which resulted in
European Union approval (CE Mark) in 2011 and FDA market approval in 2013. Argus I was
adapted from a cochlear implant as an experimental device, while Argus II was designed as a retinal
prosthesis intended for commercial use.
Both Argus I and II prostheses share similar operational system design, consisting of a
miniature camera mounted on a pair of glasses, a wearable external video processing unit (VPU),
a pair of extraocular and an intraocular coil for data and power transmission, as well as an electrode
array that is mounted on retina using a unique tack (Figure 1.11A, B). The array of Argus I and
Argus II contains 16 (4 × 4) and 60 (4 × 10) electrodes, respectively (Figure 1.11D). The Argus I
electrodes are designed with two sizes: 260 µm and 520 µm in diameter, whereas the Argus II
electrodes have been reduced in size to 200 µm. Compared with Argus I, the Argus II array not
only achieve higher electrode (pixel) density for increasing spatial resolution of perception, but
also expand its covering area to accommodate a greater visual angle – an estimated increase from
10◦×10◦ for Argus I to 11◦×19◦ for Argus II.
The Argus I was implanted in 6 subjects blinded by RP between 2002 and 2004 (de Balthasar
et al., 2008, Mahadevappa et al., 2005, Horsager et al., 2009). All patients reported that they were
able to perceive light when the device was activated, even after suffering blindness for several years
already. The safety of the device was confirmed in every patient, except one subject who had to
27
have the device removed due to an unrelated health problem (Chader et al., 2009). Further
functional testing demonstrates that each patient had the ability to perform simple visual tasks, such
as detecting objects, counting and discriminating them (Yanai et al., 2007). More practical purposes,
such as following the cross walk line when crossing the street, have been successfully demonstrated
in some subjects.
Between 2007 and 2009, a total of 30 subjects (29 with RP and 1 with choroideremia) received
the Argus II implant in the United States and Europe (Ho et al., 2015). All patients have been able
to perceive light percepts and perform visual spatial and motion tasks after a short period of training
without any adverse events. With the system on, more than 90% of the subjects performed a spatial-
motor task (to locate and touch a white square on a black computer screen) with greater accuracy
and repeatability than with the system off (Ahuja et al., 2011). The best reported visual acuity was
improved from 20/3244 of Argus I to 20/1262 of Argus II. Most patients demonstrated the ability
to identify large high-contrast letters with fairly high accuracy with the device on compared with it
off. However, scanning was always required and the time spent on task was relatively long (tens of
seconds) (daCruz et al., 2010).
The Intelligent Medical Implants (IMI) retinal implant share similar modules as SSMP’s Argus
implants, consisting of a visual interface, a pocket processor and a retina stimulator (Hornig et al.,
2007). The main difference between two devices is the wireless transmission; instead of using a
pair of RF transceiver coil for all transmission, IMI takes advantage of two wireless links, including
RF transmission for power and an infrared (IR) optical link for data. The IR transmitter, consisting
of IR LEDs, is embedded in the visual interface which is positioned in front of the eye. The IR
receiver sits intraocular, as part of the implant. One advantage of using the optical link is that the
transmission can be actively controlled by simply opening or closing the eyelid, as an instinctive
response in normally sighted people. The prototype IMI contains 49 iridium oxide electrodes that
are connected via a cable to the retina stimulator electronics encapsulated by polyimide (Figure
28
1.11D). Since non-hermetic packaging was used, the device was considered a temporary,
experimental implant and not a commercial grade implant.
Between 2003 and 2004, 20 subjects with advanced RP were tested with the IMI’s device, of
which 19 reported light perception, though generated visual percepts were reported in different
shapes, sizes, colors and brightness in response to activation of one electrode (Hornig et al., 2007).
The threshold charge measured from 15 patients fell within the safe limit of charge capacity of the
iridium oxide electrodes and remained relatively stable over time (Keseru et al., 2012, Richard et
al., 2009). Test results showed that the patients were able to distinguish phosphenes elicited by
nearby electrodes and recognize simple shapes, such as a horizontal bar, when multiple electrodes
were activated (Richard et al., 2007). In 2016, Pixium Vision, which acquired IMI technology,
launched a clinical trial to assess the safety and effectiveness of IRIS II, an updated 150-electrode
epiretinal device, in 10 patients with retinal dystrophy. Although the technical details of the IRIS
II system are not released yet, several distinguishing features have been unveiled, including an
increased number of electrodes compared with the existing epiretinal systems, an explantable
design for future electrode array removal, and a smart imaging acquisition system that only captures
changes of the visual scene to avoid temporal redundancy.
EPIRET3 is the latest-generation retinal implant model from Epi-Ret. Instead of placing the
electronics on the eye wall, EPIRET3 fits entire device inside the eyeball, excluding the camera
and image processor. This completely wireless design eliminates the need for using a transscleral
(across the eye) cable that links the electrode array with the stimulator. The electrode array contains
25 penetrating iridium oxide electrodes each 100 μm in diameter and 25 μm in height, and fixation
relies on 2 tacks onto the macula (Roessler et al., 2009, Klauke et al., 2011). The entire implant is
encapsulated by silicone to protect the electronics and coated with parylene C to ensure
biocompatibility.
29
Figure 1.11: Epiretinal prostheses. (A) External module of the Argus II system. (B) Implant part
of Argus II system. (C) Electrode array of the Argus I (left) and Argus II (right) implants, each
containing 16 and 60 electrodes respectively. (D) System Schematic of the IMI system and the
implant part prototype. The stimulating electrode array is labeled by red square. Image from (Yue
et al., 2016)
EPIRET3 implants were tested in 6 patients with RP in 2006 and left in place only for 4 weeks
based on the study design (Klauke et al., 2011). Studies show that the stimulation thresholds
required to evoke percepts were within the electrochemically safe limit. The subjects reported that
they were able to discriminate between simple stimulation patterns, such as a circle or a line. Due
to the protruding electrodes, relatively lower thresholds can be expected due to the improved
contacts with the neurons on retina. A 2-year follow-up study in 5 subjects after the removal of the
device indicate that no major structural disruption or alteration occurred in the eye except for
moderate gliosis at the tacks (which were left in place) (Menzel-Severing et al., 2012).
1.4.2 Subretinal Prostheses
Subretinal prostheses sit between the RPE and degenerated photoreceptor layer, thus mainly
targeting the neurons located in middle sections of retina (e.g. bipolar cells). Advantages of this
30
application include the retention of the neural processing that occurs at the inner plexiform layer
and avoidance of direct activation of RGC axons that causes distortion of the visual perception.
Owing to the location of the implant, the surgical operation for subretinal devices is considerably
more difficult because of the limited subretinal space as well as having to detach the retina and/or
cut across the highly vascular choroid; however, it avoids tack fixation since the implant can be
held in place through pressure (intravitreal silicone oil tamponade is needed against retinal
detachment) (Stingl et al., 2013a). Compared to epiretinal implantation where the vitreous flow and
choroidal perfusion help heat dissipation and diffusion of nutrients, subretinal implantation may
hinder the fluid flow between the retina and the choroid; however, whether this obstruction could
lead the atrophy of the retinal tissues around the implant area in the long term is still under debate
(Rizzo and Wyatt, 1997, Peachey and Chow, 1999, Sailer et al., 2007).
Two approaches for subretinal devices have been developed, including one that stays with a
standard electrode array and the other that uses a microphotodiode array (MPDA). The former
approach is similar to epiretinal implant from the system perspective in the sense that images are
acquired and processed by a wearable system, and the implanted electrode array only functions as
a slave current source under the command of the external system. On the other hand, a MPDA
directly transforms light into corresponding electrical pulse based on local luminance level. As
mentioned in the previous chapter, since the imager is located in eye rather than fixed on the head,
the user can use eye movements to perform visual tasks in a more natural manner.
The Artificial Ailicon Retina (ASR) developed by Optobionics was the first subretinal implant
to be evaluated in an FDA-approved clinical trial. The ASR is a silicon-based chip, 2mm in
diameter and 25 μm thick. This device contains approximately 5000 microphotodiodes, each
coupled to a 9 × 9 μm iridium oxide electrode and functions passively without external power and
wires (Chow et al., 2004).
31
ASR was reported to be implanted and tested in more than in 10 RP patients and 6 of them
have been followed up to > 7 years (Chow et al., 2010). Although the safety and partially restored
visual function have been observed in most of the patients, most of the improvements were found
in areas far from the implant site (Chow et al., 2004, Chow et al., 2010). Instead of electrically
activating retinal neurons, the improvements have been mainly attributed to the neurotrophic effects
of the implant that rescued or preserved the damaged retinal tissues. Further inspection of ASR’s
output capacity shown that the microphotodiode can only generated current at the nA level under
ambient light environment, whereas activation of neurons requires current with µA scale amplitude
(Palanker et al., 2005). The result showed that the current prosthesis design is inadequate to reliably
drive prosthetic vision if solely relying on the natural incident light. Therefore, the research
direction for technological variants has concentrated on amplifying photocurrents in the later
MPDA devices.
The Alpha IMS, developed by Retina Implant AG in Germany, improved the design of the
MPDA for light detection and current generation by including externally powered circuits to
amplify the photocurrents. To supply its power, two forms of circuits have been developed for
power transmission, including wired and wireless. In the investigational device, power is supplied
by a percutaneous cable that crosses the skin positioned behind the ear, while in the commercial
device (Alpha-IMS), the percutaneous cable is substituted by a subdermal power module for
wireless transmission (Wilke et al., 2011a, Stingl et al., 2013b). The MPDA chip consists of about
1500 microphotoiodes that connect to differential amplifiers independently, whose output is
cascade-coupled to a square-shaped titanium nitride (TiN) electrode (50 × 50 μm) (Figure 1.12 A).
The size of chip is roughly 3 × 3 mm and 70 μm thick, estimated to cover a visual angle of 11◦×11◦.
The clinical trial interim report of the commercial version of Alpha-IMS was recently published
(Stingl et al., 2015b). Since 2009, 29 patients with hereditary eye disease, including 25 with RP
and 4 with cone-rod dystrophy have been implanted with the Alpha-IMS. Within 10 months
32
postoperative observation, 21 participants reached the primary experimental endpoint, of which 13
participants showed significant restoration of visual function used in daily life. The report further
showed that 3 patients were capable of letter reading in which visual angle was subtended to up to
10◦. The best reported visual acuity of Alpha-IMS was 20/546 Landolt C acuity (LCA), which
roughly matched the theoretical limit determined by the 70 μm pitch between neighboring
electrodes (Stingl et al., 2015b). On the other hand, this study indicated the decreasing tendency of
restored visual function over time in a number of patients, mainly due to the technical failure of
implants. One of the common issue causing failure in early devices was intraorbital cable breakage
which can be attributed the mechanical stress induced by eye movements. With the introduction of
parabulbar loop design, the stress has been greatly reduced and breakage is not the top reason of
device breakdown. Alternatively, the non-hermetic packaging is another unsolved major problem
leading to the implant failure, as the corrosion of the chip gradually leads to malfunction of the
electronics. Although the modified encapsulation technology has demonstrated encouraging results
in animal studies and is now under assessment in the ongoing clinical trial, long-term implantation
in patients still yet to be realized (Stingl et al., 2015b, Stingl et al., 2013b).
Instead of using an external integrated circuit to amplify the simulation current as in the Alpha-
IMS, Palanker et al. developed a photovoltaic retinal prosthesis that adopts optical amplification
through high intensity near infrared laser (NIR, 880-915 nm) on goggles within established safety
margins (Palanker et al., 2005, Mathieson et al., 2012). With the laser projection system mounted
on video goggle, the ambient visual scene is first converted to laser pulses, further projected onto
the corresponding subretinal photodiode array to elicit electrical stimuli. Compared with the
conventional MPDA, this design can increase the generated photocurrent by a factor of ~1000
which theoretically is more than sufficient to drive functional retinal neural stimulation. For each
element in the array, an iridium oxide electrode is located in the center surrounded by 2 or 3
photodiodes in series to increase the dynamic range of the charge injection (Figure 1.12B).
33
However, the safety and efficiency of this device in human subjects is still unknown since it has
yet to enter clinical testing.
Figure 1.12: Subretinal prostheses. (A) Prototype of the Alpha-IMS implant. The overview of the
implant (top) and the zoom-in view of the microphotodiode array (MPDA) with additional 4 × 4
Tin electrode array attached to the end (bottom). The MPDA part consists of 1500 photodiodes on
the area of 3 × 3 mm. (B) MPDA of the photovoltaic prosthesis developed by the Palanker group,
with zoom-in view of a single stimulating element with 3 photodiodes in series. (C) Prototype of
the 256 channel Boston retinal implant. The concept of the device mounted on the eye with the
secondary coil surrounding the cornea (left) and electrodes bonded to the feedthrough of the
hermetic package. Image from (Yue et al., 2016).
The Boston Retinal Implant group has designed a subretinal device with a markedly different
approach, as its prototype was derived from their earlier epiretinal prosthesis experience. Rather
than use photodiodes as the trigger of stimulation, this group uses a passive electrode array. With
the advance of fabrication techniques in electronic system and hermetic sealing, the new generation
device has 256 channels, compared to 15 electrodes in their first version. To support the increased
34
number of independently driven channels, the investigator improved the telemetry system with
higher data transmission rate and power capacity, through moving the transceiver secondary coils
from the temporal region of the eye to the anterior, surrounding the cornea (Figure 1.12C). The
relocation provides larger space for a larger coil for better inductive coupling efficiency with
external primary coil (Kelly et al., 2011, 2013). The other feature of this device is the design of
high-count 256 ceramic feedthroughs in a titanium case, which was evaluated to support lifetime
of 5-10 years in carefully controlled physiological environments (Shire et al., 2014, Kelly et al.,
2013). This device is currently at pre-clinical stage.
1.4.3 Suprachoroidal prostheses
Compared with the epiretinal and subretinal implant, suprachoroidal prostheses have an
electrode array relatively distant from the retina since the array is located between the choroid and
the sclera. With the separation, the risks of retinal damage from surgery and the implants have been
potentially reduced. Moreover, thermal dissipation is less of a concern for this type of device since
the abundance of the passing blood vessels at the choroid layer can effectively carry away the extra
generated heat from the chip (Parver et al., 1983, Hadjinicolaou et al., 2015). However, the
increased distance between the electrode and target tissue is expected to lead to a lengthened current
diffusion path, elevated perceptual threshold and worse spatial resolution (Yamauchi et al., 2005).
Typically, a return electrode is placed in the anterior part of the eye, such as the cornea or vitreous
cavity, to guarantee the current steering can flow through the retina.
A Semichronic Suprachoroid transcleral (STS) prosthesis developed by Fujikadoet al. was
implanted in two RP patients and tested in 2011 (Fujikado et al., 2011). The implant was positioned
in a scleral pocket (6×5 mm) formed by cutting a flap in the sclera and its electronic metal case
located subdermal behind the ears (Morimoto et al., 2011). Within the stimulation chip, 49 platinum
35
electrodes of 500-µm diameter formed an array while 9 of them were active for testing (Figure
1.13A). In a recent study, 4-6 out of these 9 electrodes had the capability to elicit perceivable
phosphenes within the current limit and the selected charge levels were universally greater
compared with the epiretinal device. However, this device did enhance the patients’ ability to
localize objects and allow them to perform reaching task. Four weeks follow-up observation
suggests that the implantation causes no retinal detachment nor hemorrhage, but long-term
performance requires further investigation.
Bionic Vision Australia (BVA) Consortium, composed of multiple researchers from different
Australian institutes, is currently designing several prototypes of suprachoroidal and epiretinal
implants to achieve wide-view and high resolution vision. For a suprachoroidal device, an
intraocular array is fabricated using silicon substrate with 33 platinum stimulating electrodes (400-
µm and 600-µm diameter in sizes) and 2 large return electrodes (2000 µm diameter) (Figure 1.13B).
While applying stimulation, the 13 stimulating electrodes on the outer ring can be selectively
shorted together to function as a large common ground. Moreover, an extra electrode was implanted
subcutaneously behind the ear to serve as an additional extraocular return (Ayton et al., 2014a,
Shivdasani et al., 2014). Altogether, this implant consists of 20 individual addressable stimulating
and 4 optional return electrodes. Similar to cochlear implants, this device communicates with an
external controller using a percutaneous plug which connects to the intraocular electrode array
through a helical lead wire. As mentioned by the investigators, the plug will be replaced by wireless
communication systems in the next device generation.
From 2012 to 2014, the suprachoroidal prototype was implanted in 3 RP patients for a Phase I
clinical trial. Post-surgical monitoring up to 2 years shows that the device remains functional and
stable in the suprachoroidal space with no significant retinal edema or atrophy. In
electrophysiological testing, phosphenes can be reliably evoked in all three subjects with charge
density (158 µC/cm
2
for 600-µm electrodes and 237 µC/cm
2
for 400-µm electrode) within safety
36
limit for chronic stimulation of platinum electrodes (350 µC/cm
2
). (Ayton et al., 2014b) Similar to
what has been found on epiretinal implants, the threshold for generating perceivable percepts
depends on retina-electrode distance (de Balthasar et al., 2008). An increase in threshold
corresponding to the increase in the measure of the retina-electrode distance has been found.
Among all return configurations, the monopolar pattern with 1 large distant return electrode has
showed highest efficiency. In visual function test, all 3 patients were able to perform light spot
localization task with accuracy better than chance level, and the best reported LCA can reach
2.62longMAR (20/8397) (Ayton et al., 2014b).
Figure 1.13: Suprachoroidal prostheses. (A) The STS implant. The system contains the stimulating
array with the remote return electrode (left), and the zoom-in view of the multielectrode array with
49 stimulating electrodes (right). (B) The BVA implant consists of intraocular array, helical lead
wire, percutaneous connector and 1 remote return electrode. The zoom-in view of the
multielectrode array which contain 33 stimulating electrodes and 2 return electrodes (right). Image
from (Yue et al., 2016).
37
1.5 Retinal Neurons in Response to Electrical
Stimulation
The visual perception generated by retinal prostheses is primarily determined by how the retinal
neurons respond to external electrical stimulation. Many studies have measured the
electrophysiological response of retinal neurons, especially RGCs, since they are the most
accessible neurons in a degenerated retina. Various electrical stimulation strategies have been
developed to improve the performance of visual perception from different perspectives.
Theoretically, since elicited visual percepts are based on the types of activated neurons, selective
electrical stimulation of specific neurons can alter the actual perception. Moreover, with about 20
identified different circuit network mosaics in the retina for extracting distinct visual information
(Dacey, 1999), the stimulation patterns that can produce firing patterns specific to each mosaic
have been investigated to generate more natural input to the visual cortex. The stimulation
efficiency has been also addressed to manipulate the threshold of neurons or alleviate phosphene
fading in response to continuous stimulation.
1.5.1 Direct/Indirect Activation of RGCs
The final output of the retina is retinal ganglion cells (RGCs) that collectively transmit
preprocessed visual information to the brain, and those neurons are the main targets for retinal
prosthetic research since the visual percepts are mostly determined by their activation patterns. The
RGC neurons can be activated directly by sufficient depolarization of the RGC membrane or
indirectly via synaptic transmission from activated bipolar cells (BP). Instead of firing one action
potential per pulse for direct stimulation, synaptically activated RGCs tend to fire a burst of spikes,
which most likely resemble the natural response that RGC would generate in reaction to light
38
stimulation (Jensen et al., 2005). While the elicited neural activities of RGCs from indirect
stimulation might be ideal, one major drawback for this stimulation strategy is that the sensitivity
of the RGCs to repetitive electrical stimulation progressively decreases (Ahuja et al., 2008,
Freeman and Fried, 2011, Jensen and Rizzo, 2007). This phenomenon has been attributed to
amacrine cell inhibition the ganglion cell response through refinement of bipolar cell synaptic
processing, but the intrinsic mechanism is still unidentified. Since desensitization has strong
correlation with artificial percept fading found in patients (Jensen and Rizzo, 2007, Freeman and
Fried, 2011, Im and Fried, 2016), some advanced stimulation strategies have been proposed to
alleviate this problem (see Section 1.5.4).
On the other hand, although direct stimulation can only activate one or two action potentials of
RGCs, studies have shown that most subtypes of RGCs are capable of following stimulation with
high frequency. In rabbit retina, brisk-transient ganglion cells could reliably follow pulse rates up
to 600 pulses per second (pps). ON-OFF direction selective ganglion cells can follow up to
approximately 200 pps, whereas local edge detector ganglion cells can follow rates of around 100
pps (Cai et al., 2011). These findings suggest that direct stimulation might be able to generate firing
patterns in RGCs with similar natural neural coding evoked by presynaptic bipolar cells when
coupled with retina encoder (see Section 1.5.3), However, since direct stimulation of RGCs might
possibly lead to activation of axons and produce to streak-like percepts, novel direct stimulation
paradigms that can selectively activate soma preferentially, while avoiding activation of axons are
extremely desirable (see Section 1.5.2).
1.5.2 Selective Stimulation
Manipulation the duration of stimulation pulses is commonly used to selectively activate
neurons in the retina. Studies have shown that RGCs are more easily excited by short duration pulse
39
(<150 µs) whereas BPs tend to respond preferentially to longer pulse width (Margalit and Thoreson,
2006, Fried et al., 2006a, Freeman et al., 2010). For ideal retinal implants, each electrode should
only activate nearby RGCs, thus forming small round visual percepts. However, Argus II epiretinal
implants patients reported that most evoked percepts by single electrodes were elongated and
aligned with the estimated axon paths in the retina, suggesting the activation of axon bundles
(Nanduri et al., 2012, Nanduri et al., 2011). A similar phenomenon has also been validated through
electrically elicited responses in salamander and rodent model using calcium imaging techniques
(Behrend et al., 2009, Weitz et al., 2013). To prevent streak-like perceptions, longer duration pulse
has been used to produce more confining retinal responses to the site of the electrode through
indirect stimulation (Freeman et al., 2010, Weitz et al., 2015). Chichilnisky et al. demonstrated the
manipulation of spatial patterns of current injection generated by high density multi-electrode can
be used to selectively evoke action potential firing from individual ganglion cells, thus increasing
the spatial resolution of stimulation (Sekirnjak et al., 2006, Jepson et al., 2013) (Figure 1.14). In
subretinal stimulation, the anodic-leading current pulse has been shown to result in differential
thresholds for ON and OFF cells which might further be applied on targeted stimulation of ON/OFF
signaling pathways (Jensen and Rizzo, 2006).
Figure 1.14: Spike thresholds for several stimulation configurations. Region adjacent to the center
recording electrode (R) on the hexagonal array is shown, with filled circles indicating active
electrodes for stimulation and open circles denoting unused electrodes. All active electrodes were
simultaneously stimulated at the same current amplitude. In addition to the standard setup (left), 3
alternative configurations were tested: all neighboring electrodes, 3 electrodes, and a single
neighboring electrode. Threshold currents are shown as averages over 8 cells and indicate lowest
value found when multiple electrode combinations were tested. Corresponding charge densities
40
were 0.09, 0.22, 0.25, and 0.32 mC/cm
2
. Electrodes were 60μm apart. The figure is redrawn from
(Sekirnjak et al., 2006).
1.5.3 Stimulation encoding
In addition to targeting a specific cell type, any electrical stimulation protocol that seeks to
create a natural visual signal must recreate the typical spatiotemporal response. Ideally, such an
approach could manipulate perception in terms of luminance, contrast, shape, color and motion.
Electrophysiological studies have shown that the response of RGCs can be predicted through a
linear-nonlinear (LN) cascade model with a difference-of-Gaussians or wavelet-shaped
characteristic spatial profile (Frishman et al., 1987, Meister and Berry, 1999, Van Rullen and
Thorpe, 2001). Based on that idea, Bomash et al. have developed an adaptable LN model for a wide
range of stimuli to more precisely investigate population encoding (Bomash et al., 2013). Eckmiller
et al. proposed an encoder which models receptive field properties of primate RGCs as individual
spatiotemporal filters (Eckmiller et al., 2005). For each filter, the parameters were trained through
machine learning algorithms iteratively supervised with subject feedback to mimic the retinal
information processing. The visual inputs from an asynchronous contrast sensor was also used for
reconstructing the spatiotemporal properties of several RGC groups, and possibly outperforms the
frame-based inputs in term of temporal precision (Lorach et al., 2012). Recently, involuntary eye
movement including microsaccades, drifts and tremor were incorporated in retinal models. The
result showed that the visual sensitivity to both spatial and temporal changes in luminance can be
ameliorated (Olmedo-Payá et al., 2015). Yet, none of above conceptualized methods has been
successfully applied on real retinal prostheses mainly due to the constraints on the electrode size
versus the retinal cell size. It is not yet possible to control single RGC or bipolar cells with a retinal
prosthesis, which is necessary for the biological models to be effective.
41
1.5.4 Stimulation efficiency
A more pertinent concern for clinical retinal implant is stimulus efficiency. Using lower
stimulus intensity can extend the battery life of the external system (which supplies power for the
implant). For the epiretinal implant Argus II, adding an interphase gap between the symmetric
cathodic-first biphasic can effectively reduce the required electrical stimulation thresholds (Weitz
et al., 2014). Perceptions fade when high stimulus rates are used, but if stimulation pulses are
applied at too low a rate, the percept will flicker and be less useful. For subretinal Alpha-IMS
implants, the working frequency for stimulation is typically set to 5-20 Hz and some patients claim
that the lower frequency option helps alleviating phosphene fading (Stingl et al., 2015a). Another
fading-resistant strategy proposed by Freeman and Fried includes interleaving longer duration
desensitizing pulses (1 ms) between the short duration (200 µs) stimulating pulse trains delivered
as a function of incoming luminance (Freeman and Fried, 2011). The desensitizing pulses are
proposed to diminish the synaptically mediated response which is believed to cause desensitization,
thus allowing luminance-driven pulses to elicit 1-2 spike per pulse through direct activation of RGC.
Until now, this strategy has not been experimentally validated.
42
1.6 Calcium Imaging
Calcium imaging is a well-established optical method for monitoring neural activity. As
neurons fire action potential, the membrane depolarization leads to the influx of Ca
2+
ion into the
cytoplasm through voltage or neurotransmitter-gated ion channels, therefore offering an indirect
measurement or neural activation. Studies have demonstrated that the calcium transient change
corresponds to the spiking activity of neuron, thus serving as a functional tool for multiple single-
neuron activity recording application (Lohmann et al., 2002, Smetters et al., 1999, Wong, 1998,
Looger and Griesbeck, 2012). Although the calcium dynamics within the cytoplasm are relatively
slow compared with the rate of spiking, which might limit the temporal resolution, the slow nature
of calcium transients also improves signal quality in terms of overall signal-to-noise ratio (SNR)
(Looger and Griesbeck, 2012).
1.6.1 History
Aequorin is the first documented calcium indicator which is a blue luminescent protein
extracted from the jellyfish Aequorea aequorea (Shimomura et al., 1962). In 1967, researchers used
this protein to measure calcium transients in single muscle fibers (Ridgway and Ashley), which
was regarded as the first experiment of calcium imaging to record neural activity. Cloning of
Aequorin and its mutants was proposed in 1985, ultimately expanding its use (Inouye et al., 1985).
Different from most of fluorescent probes, aequorin requires no optical excitation for observation.
However, major rawbacks of aequorin include low light output and one-time use, ultimately
limiting its use, so more advanced calcium indicators have been successively developed.
A major advance in calcium indicators was made by Roger Tsien, who was later awarded the
Nobel Prize in chemistry in 2008 for his contribution with green fluorescent protein (GFP). The
43
first small-molecule calcium dye was developed by Tsien’s group in the 1980s from the novel UV-
excitable Ca
2+
chelator BAPTA, which is originated from EGTA with better calcium:magnesium
selectivity (10-fold differences). Other advantages of BAPTA includes faster kinetics for binding
and dissociation, less pH sensitivity, and generation of sizeable fluorescence signal. All of them
made BAPTA is the preferred chelator at that moment (Cobbold and Rink, 1987, Tsien, 1980).
BAPTA was not used for calcium imaging but for Ca buffering.
Quin-2 was the first biologically compatible calcium indicator synthesized by Tsien (1980)
with shifted excitation band centered at 339-nm, since shorter wavelength UV is unsuitable for
biological use due to possible UV-induced cell damage or severe interference from cell proteins
and nucleotides. Once Quin-2 binds to calcium, it exhibits 5-6 fold increase in emission
fluorescence for 339-nm excitation. Moreover, when excited at 370-nm light, on the contrary, it
exhibits decrease in fluorescence. The dual-wavelength excitation characteristic could allow Quin-
2 to serve as a tool for ratiometric [Ca
2+
] i measurements (Cobbold and Rink, 1987), though the
emission at 370 nm is so weak that this is not commonly done. Quin-2 is also the first calcium
probe providing measurement of [Ca
2+
] i in many important cell types; however, the quantum yield
(percentage of incident photons are absorbed) is relative low compared with other state-of-art
indicators (Cobbold and Rink, 1987).
In 1985, two more UV-excitable calcium sensors that are still in use today, indo-1 and fura-2,
were developed by Tsien’s group. Compared with Quin2, these two dyes provide a 30-fold increase
in terms of emission fluorescence (Grynkiewicz et al., 1985). Both sensors are also capable of
ratiometric measurement. While binding to calcium ion, indo-1 exhibits a shift in emission
wavelength while fura-2 performs a shift in absorption spectra. Moreover, lower Ca
2+
affinities
than Quin-2 for both dyes facilitate observation of Ca
2+
kinetics due to faster off rates and less
calcium buffering (Grynkiewicz et al., 1985, Takahashi et al., 1999).
44
As mentioned previously, ultraviolet light is not an ideal source for excitation since it
potentially damages the cells; therefore, the indicators that can be excited within visible wavelength
spectrum are highly required. The first generation of these dyes, including rhod-2 (552/581 nm
excitation/emission) and fluo-3 (506/526 nm excitation/emission), were further synthesized by
Tsien’s group. Upon binding to calcium ion, rhod-2 and fluo-3 exhibits 15- and 40-fold increase of
fluorescence respectively (Minta et al., 1989). However, since both dyes do not perform shift in
excitation or emission spectrum which limit them to use as non-ratiometric indicators.
Owing to the needs for better imaging SNR and different experimental setting, several calcium
dyes have been developed and improved from various perspectives, such as the range of
excitation/emission wavelength, calcium affinities, molecular weights, fluorescent brightness.
Oregon-Green BAPTA-1 and fluo-4 are two of the most prevalent dyes since they are among the
brightest indicators operating within visible wavelength. However, the usage of calcium dyes is
still mostly constrained in acute imaging experiment because of their natural limitations: leakage,
photobleaching, compartmentalization, and lack of cell-type specificity.
Genetically encoded calcium indicators (GECIs) has been developed to overcome many of the
problems associated with synthetic dyes. Most of GECIs consist of one or two fluorescent
molecules, calcium binding domain, and a binding peptide. Cameleon was the first GECI also
proposed by Tsien and his colleagues (Miyawaki et al., 1997). The sensor consists of two GFP
mutants with a different emission wavelength (cyan and yellow). The fluorophore are bridged by
the well-known calcium binding protein calmodulin (CaM) and the CaM-binding M13 peptide of
myosin light chain kinase. Binding of calcium ion makes CaM wrap around the M13 domain,
causing the two fluorophores to move closer to each other. The shrinkage of distance increases the
fluorescence resonance energy transfer (FRET) between the donor (cyan) and acceptor (yellow)
fluorophores, thus intensifying the brightness. Through measuring the change of the fluorescence
45
intensity ratio between the two interested bands, the calcium concentration can be accurately
estimated and obtained.
Despite the advantages mentioned above, the 1
st
generation of cameleon exhibits relatively
small fluorescence ratio change (< 2), resulting insufficient sensitivity of calcium ion for most of
applications (Takahashi et al., 1999). Later generations of FRET-based GECIs, such as YC3.60,
D3cpv, and TN-XXL, basically were derived from the original cameleons through incorporating
different fluorophores and /or calcium-binding proteins. Although their Ca
2+
sensitivity have been
improved, their performance still lags behind of the best synthetic dyes (Mank and Griesbeck,
2008).
A major breakthrough of calcium indicator occurred with development of single-wavelength
GECIs, as an alternative to FRET-based GECIs. Again, it was Tsien’s group that discovered that
the β-barrel around the GFP chromophore allows insertion of entire proteins without quenching
fluorescence in 1999 (Baird et al., 1999, Mank and Griesbeck, 2008). Based on this finding, first
single-fluorophore GECI, camgaroo-1, was synthesized through inserting CaM into yellow
fluorescent protein (YFP) (Baird et al., 1999). However, this indicator was never widely adopted
due to its low Ca
2+
affinity and high sensitivity to ambient pH and temperature (Griesbeck et al.,
2001). Despite those drawbacks, the design paved the way to future generations of single-
wavelength GECIs.
GCaMP is one of the significant achievement of GECIs as single-wavelength sensors which
incorporates a GFP fluorophore and the CaM-M13 calcium binding
complex (Nakai et al., 2001).
Similar to many other calcium dyes, the 1
st
generation of GCaMP was released with defects such
as pH and temperature sensitivity, dim fluorescence at resting states, as well as nonlinear bleaching
(Mank and Griesbeck, 2008). However, consistent iterative improvements through randomly
selected site-directed mutagenesis ameliorated these problems one by one and resulted in multiple
generations of GCaMP variants (GCaMP1.6, GCaMP2, GCaMP3, GCaMP5, and GCaMP6), as
46
well as different choices in term of excitation/emission wavelength (i.e., red and blue) (Zhao et al.,
2011).
Figure 1.15: Crystal structure of GCaMP in Ca
2+
-bound state. The indicator is composed of a
circularly permuted GFP linked to CaM and the CaM-binding M13 peptide of myosin light chain
kinase. Binding of calcium ion makes CaM wrap around the M13 domain and further blocks solvent
access to the cpFGP chromophore, causing the molecule become brighter.
GCaMP3 was the first GECI to rival the capability of synthetic calcium dyes with its improved
imaging SNR and photostability (Borghuis et al., 2011, Tian et al., 2009). When the sensor is
transduced and well-expressed in neurons, with appropriate imaging system it can reliably report
three or more action potentials in a short burst (Borghuis et al., 2011). A newer version of calcium
sensor, GCaMP5, engineered from GCaMP3 has increased the dynamic range of the fluorescence
response, the Ca
2+
saturated brightness and affinity, thus leading to a 2-3 fold SNR improvement
(Akerboom et al., 2012). The latest GCaMP variants, GCaMP6, were designed and evaluated
47
directly in neuron. The ultrasensitive protein calcium sensor is capable of reporting single action
potentials with near 100% accuracy in neuronal somata and orientation-tuned synaptic calcium
transient in individual dendritic spines. GCaMP6 is the first GECI to outperform the best existing
synthetic indicators, in terms of sensitivity, dynamic range, and kinetic speed (Chen et al., 2013c).
1.6.2 Delivery method
Calcium indicator can be delivered into cell populations via multiple methods. Several factors,
including indicator type, cell type, and the desired length of imaging period together contribute to
the selection of appropriate method. In this section, the calcium loading technique related to retina
will be reviewed.
1.6.2.1 Synthetic Dye Loading
Due to the hydrophobic tails on the phospholipid bilayer structure, calcium dyes, as hydrophilic
molecules, cannot directly pass through the membrane. Accordingly, researchers have developed
different procedures for disrupting cell membrane, thus creating temporary opening for the dye to
enter.
Microinjection is one of the most common dye-loading method which involves the use of
intracellular or patch pipettes for injecting dye into cytoplasm through penetrating the membrane
(Wu et al., 2004). As an established technique around 1970’s for injection of liquid substances,
early calcium imaging experiment relied on this method for delivering calcium indicators. However,
the laborious labeling process limits the possibility for multiple cell application, such as RGC
populations.
Electroporation and biolistics are two techniques used to deliver the reporters through
disrupting cell membrane. Electroporation uses electrical pulse applying large electric field (kV/m)
to create temporary pores in cell membranes. After the procedure, a recovery period is usually
48
required for the disrupted membrane to reseal and recover from induced shock and swelling. For
in vivo retina research, early attempts for using electroporation to load calcium dye into RGCs only
resulted in relatively sparse labeling (Yu et al., 2009). However, with more refined procedure, a
recent study shown it is capable of staining virtually the entire RGC layer of retina in adult mice
(Briggman and Euler, 2011). This method is still not considered an ideal method for RGC
application since the amacrine cells, which constitute roughly 50-60% of cells in rodent ganglion
cell, were also labeled indiscriminately (Jeon et al., 1998, Perry, 1981). However, higher specificity
of retinal neurons can be achieved when replacing calcium dye with genetically encoded calcium
indicators driven by viral vector (Borghuis et al., 2011).
Biolistics method relies on a “gene gun” which serves to propel dye-coated particles or
exogenous deoxyribonucleic acid (DNA) to living cells (Kettunen et al., 2002). This tool is capable
of accelerating the target particles so that they can easily penetrate cell membrane with high
velocity. Biolistics has been used for labeling the photoreceptors and RGCs in adult mice with
nanoparticles coated with indicator and resulted in relatively sparse and not cell-specific staining
(Roizenblatt et al., 2006). However, similar to electroporation approach, specificity can be
improved when using genetically encoded calcium indicators as sensors (Bleckert et al., 2013).
Other than the methods that involve cell membrane disruption, non-invasive dye-loading
approach have also been developed. Many calcium dyes have been reformatted in acetoxymethyl
(AM) ester forms. The extra lipophilic AM ester is designed to cap the hydrophilic carboxylic acids
appearing on the indicator, thus making the dyes become neutral and free to permeate cell
membranes. After entering the cell, endogenous esterases will cleave AM group, thus giving the
place to charged compound and reactivating the dye (Paredes et al., 2008).
Loading indicators in AM form is relatively simple compared with previously mentioned
methods since it only involves bath incubation of the cells or tissue with the dye. One problem of
this approach is that it fails to specifically label desired type of cells. Moreover, the dye cannot
49
sustainably stay inside the cells of in most of adult mammalian, including retina (Yuste and Katz,
1991, Wong and Oakley, 1996, Wong, 1998). Strongly negative charged compounds tend to be
pumped out of cells by organic anion transporters (OATs) (extrusion), thus limiting the observation
period of calcium imaging application (Behrend et al., 2009).
One possible method to improve dye with higher resistance to extrusion is to increase its size
via dextran conjugation. Dextran are hydrophilic polysaccharides which are relatively nontoxic and
inert to the cells. Multiple dextran-conjugated calcium dyes with a variety of molecular weight are
commercially available on market. However with the increased size of molecule, those dyes are
generally not trivial to load due to limited permeability.
For retina research, the dextran-conjugate calcium dyes has been successfully loaded to RGCs
with retrograde loading technique. The first implementation of this method relies on using a dye-
laden syringe needle to make axons cut so that the RGC on the isolated rabbit retina can uptake the
dye. Over a rough two hour period, the dye will finally reach the soma of RGC through natural
diffusion. Even though the method is useful, it is not sufficient since only a small percentage of
RGCs can be labeled (Baldridge, 1996, Hartwick et al., 2004).
An advanced retrograde dye loading method has been developed by our group. Through
applying calcium dye to the cut optic nerve with appropriate incubation time, most of ganglion cells
in the isolated retina can be selectively labeled (Behrend et al., 2009). However, the approach is
not practical for mammalian retina since the fluorescence intensity in soma becomes weaker over
time, presumably due to active extrusion by OATs.
1.6.2.2 Expression of Genetically Encoded Calcium Indicators
The other main category of calcium sensors is genetically encoded calcium indicators (GECIs),
which relies on delivery the designed foreign encoded DNA into target cells. With high negative
charge of DNA or ribonucleic acid (RNA), the cell do not readily uptake naked gene segments
50
since they tend to be repelled by anionic cell membrane. After entering cytoplasm, DNA must be
transported to nucleus so can be translated to corresponding protein, before degraded by
endogenous nucleases (Patil et al., 2005). To achieve the goal, several DNA delivery methods have
been developed.
Typically, DNA delivery methods can be classified as non-viral or viral. Non-viral approaches
achieve transfection through electrical, mechanical or chemical means, which are less harmful to
the cell but relatively ineffective compared to viral approaches. Viral methods rely on the use of
recombinant viruses, such as viral vector, for transduction. Although they have been regarded as
the most effective means to deliver genetic materials, the drawbacks include immunogenicity and
toxicity which must be overcome through appropriate experimental control and vector design (Luo
and Saltzman, 2000).
Approach Method
Electrical Electroporation
Mechanical
Microinjection
Biolistics
Sonoporation
Chemical
Artificial lipids
Cationic polymers
Table 1.1: Non-viral DNA transfection methods (adapted from Luo and Saltzman, 2000).
Appropriate selection of DNA delivery methods is of critical importance to accomplish
successful expression in targeted cells. For retina study, electroporation, generation of transgenic
animal, or viral vectors are the most commonly used methods for gene delivery. Studies have
showed that electroporation can transfect RGC populations with naked plasmid DNA in vivo
(Dezawa et al., 2002) and in vitro (Garcia-Frigola et al., 2007), and results in limited non-RGC
specific labeling pattern. Higher specificity of electroporation in On-type bipolar cells can be
achieved when using viral vector of GECIs, and it can resolve the incompatibility of cell surface
51
receptors with mismatched viral capsid serotype (Borghuis et al., 2011). Transgenic animals are
also used to express GECIs in neuronal population. For example, the Pvalb-2A-Cre:Ai38 line has
been produced through cross breeding to have GCaMP3 expressed in RGCs, horizontal cells, and
Müller glia (Zariwala et al., 2012). However, generating a mouse line that exclusively expresses
gene in RGCs is challenging owing to the lack of a pan-ganglion-cell-specific premotor, thus
limiting the possibility of transgenic mouse (Feng et al., 2000). Furthermore, establishing a stable
transgenic lines requires multiple generations of animals which limits the productivity to match the
throughput of new GECIs. Viral vectors, on the other hand, overcome these limitations in terms of
specificity and productivity, especially in the retina research (Borghuis et al., 2011, Hellstrom et
al., 2009).
Recombinant adeno-associated virus (AAV) is the most frequently used vector for retinal gene
transduction (Grimm and Kay, 2003). AAV is a small (25-nm), non-enveloped virus that packages
a linear single-stranded DNA of 4.7kb genome (Daya and Berns, 2008). The attractive properties
of AAV include lack of toxicity and pathogenicity, as well the ability to achieve long-term gene
expression. With the advance of genetic engineering techniques, twelve naturally occurring AAV
serotypes (AAV1-AAV12) and over 100 variants that target different tissue tropisms are now
available (Li et al., 2009). Through cross-packaging one serotype’s vector genome into another
serotype’s capsid procedure, called pseudotyping, the tropism of vectors can be further altered to
target various cell or tissue (Heilbronn and Weger, 2010). Moreover, The transduction properties
and immune neutralization issues can be improved by site-directed mutagenesis (Kwon and
Schaffer, 2008, Petrs-Silva et al., 2011). The limited packing capacity of recombinant AAV is
roughly 4.7 kb which is more than sufficient for the regular size of GECIs.
Studies have demonstrated that AAV vectors have superior capability to target different retinal
cell types respectively (Grimm and Kay, 2003). Through designs of various combination of
promoters and serotypes, five major neuron classes (photoreceptors, horizontal cells, ON-type
52
bipolar cells, amacrine cells, ganglion cells) in mouse retina can be successfully labeled by
GCaMP3, though some non-ideal labeling occurred in horizontal cells (< 30%) for AAV2/1-
SYN1_GCaMP3 designed to target RGCs specifically (Borghuis et al., 2011). Based on the desired
transduction regions, subretinal (outer retina) and intravitreal (inner retina) injection techniques
have been developed and shown to be able to achieve the goal. One of the latest GECI incorporated
with GCaMP5G has demonstrated the ability to specifically target about 85% of RGCs which has
been further validated by double-labeling of tracer dye retrograde loaded from optic never cut
(Weitz et al., 2013).
Figure 1.16: Pseudotyping of AAV vectors by cross-packaging one serotype’s AAV2 vector
genome into the capsids of the other serotypes. Those vector with altered tropism improve
transduction efficiency in certain types of tissues. Image from (Heilbronn and Weger, 2010).
1.7 Summary of Introduction
Retina is one of the complicated nervous system responsible for visual information processing
and relay, through several layers of neurons performing early stage processes such as illuminance
adjustment, movement and edge detection, and color filtering. Degeneration of retinal
53
photoreceptors results in vision loss but no effective treatment can reverse the progress so far.
Retinal prosthesis designed to partially restore visual function by stimulating the remaining neurons
have been regarded as one of the most promising solutions.
Despite the advance made for different types of retinal implants, from hardware to software,
technical challenges and unknown electrophysiological properties of retinal neurons still limit the
performance of this medical electronics. One of the most important goal is to improve the visual
acuity in prosthesis patient which still fall behind normal vision and acuity is believed to be
improved through multiple approaches. In terms of hardware, the increase number and density of
electrode with refined coating materials might possibly exhibit better spatial resolution. The other
approach with great potential is to identify smarter stimulation strategies that can create a more
ideal focal response of RGCs nearby the stimulating electrode through selective activation of
neurons.
To test possible stimulation paradigms, a method to simultaneously measure neuron activities
of multiple RGCs in response external electrical stimulation is required. Genetically encoded
calcium indicators (GECIs) have been widely used for observing neural activity, because they
enable repeated measurement of many neurons in parallel at single-cell resolution. Of all deliver
methods, viral transduction provides the greatest cellular specificity in the retina. However, since
viral transduction of GECI is known to vary with time and an optimal time window of expression
exists. Therefore, a suitable tool for tracking GECI genetic expression in vivo following
transduction is highly desirable.
1.8 Thesis Overview
The purpose of this thesis can be mainly divided into thus two-fold. First we developed a
noninvasive imaging approach based on a custom-modified low-cost and simple fundus system that
enabled us to monitor and characterize in vivo bright-field and fluorescence retinal image of mouse
54
model, based on the concepts proposed by Paques et al. (Paques et al., 2007) and Schejter et al
(Schejter et al., 2012). To monitor the genetic expression of selected GECI GCaMP6f (Chen et al.,
2013b) transduced in mouse RGCs by adeno-associated viral vector (AAV), we perform long-term
tracking following intravitreal injection and find an optimal window of expression (Chapter 2).
Moreover, based on the determined window and established in vitro calcium imaging technique,
we aim to attain control of RGCs responses by using different stimulation paradigms. The goals is
to address three specific problems arose during clinical testing with Argus I and II patients:
1. Elongated Phosphenes: Finding stimulation patterns that can selectively activate RGCs somas
nearby electrode, thus forming more ideal round shape percept theoretically (Chapter 3).
2. High thresholds: Finding stimulation patterns that can reduce the threshold of RGCs, thus
potentially reduce the power consumption and improve the device efficiency (Chapter 3).
3. Limited electrode number: Applying current steering method by using local return electrode
to produce virtual electrode effect or alter the RGCs responses (Chapter 4).
The significance of my findings is evaluated in the context, along with suggestion of future
experiment (Chapter 5).
55
CHAPTER II: GCaMP Expression Characterized
Using Fundus Imaging
Abstract
Virus-transduced calcium indicators are effective reporters of neural activity, offering the
advantage of cell-specific labeling. To track the time dependence of in vivo expression levels of
genetically encoded calcium indicators (GECIs) in rodent retina, we developed a noninvasive
imaging approach based on a custom-modified, low-cost and simple fundus system that enabled us
to monitor and characterize in vivo bright-field and fluorescence retinal image. The system clearly
resolves individual retinal ganglion cells (RGCs) and axons. RGC fluorescence intensity and
number of observable fluorescent cells show a consistent rising trend from week 1 to week 3 after
viral injection, indicating a uniform increase of GCaMP6f expression. At defined time points, we
prepared wholemount retina mounted on a transparent multielectrode array (MEA) and used
calcium imaging to identify the optimal time for studying the responsiveness of RGCs to external
electrical stimulation. The results show that the fluorescence-endoscopy fundus system is a
powerful and widely accessible tool for evaluating in vivo fluorescence reporter expression.
56
2.1 Introduction
Genetically encoded calcium indicators (GECIs) have been widely used for observing neural
activity in multiple systems because they enable repeated measurement of many cells in parallel at
single-cell resolution. For example, some genetically encoded activity reporters were developed
for achieving functional imaging of hippocampal place cells during virtual navigation (Dombeck
et al., 2010), obtaining tuning curves of orientation-selective neurons in visual cortex (Mank et al.,
2008), optimizing the stimulation parameters of retinal prostheses (Weitz et al., 2013), or
monitoring responses in targeted retinal cell populations during visual information processing
activity with relatively high spatial resolution (Borghuis et al., 2011). GECIs can be delivered to
cells via electroporation, biolistics, viral vector transduction, or generation of transgenic animals
(Zariwala et al., 2012). Of these methods, viral transduction provides the greatest cellular
specificity in the retina (Weitz et al., 2013, Borghuis et al., 2011). However, viral transduction of
GECIs is known to vary with time and an optimal time window of expression exists (Chen et al.,
2013c, Weitz et al., 2013). Before this window, underexpression leads to small fluorescence
changes that are difficult to measure. Past this window, overexpression renders cells unresponsive.
Therefore, a suitable tool for tracking GECI expression in vivo following transduction is highly
desirable to allow experimental determination of the best time window.
Several systems have been reported for microscopic fundus imaging in rodents, including
scanning laser ophthalmoscopes incorporating adaptive optics (Roorda et al., 2002, Gray et al.,
2006, Leung et al., 2008) and commercial fundus systems (Koronyo-Hamaoui et al., 2011, Dalkara
et al., 2012). Applications range from observing autofluorescence in retinal pigment epithelial cells
(Gray et al., 2006), identifying RGC bodies and axons labeled with dye (Leung et al., 2008),
imaging curcumin-labeled Aβ plaques in Alzheimer's disease mice retina (Koronyo-Hamaoui et al.,
2011), or evaluating retinal transduction following neonatal intravascular administration of virus
57
vectors (Dalkara et al., 2012). Two-photon microscopy has also been used for monitoring calcium
transients of retinal bipolar cells modulated by visual stimulation (Dreosti et al., 2011). Although
those systems allow fluorescent fundus imaging, the high cost of these systems is a barrier to wide-
spread use.
Compared with these relatively complex and expensive imaging techniques, digital fundus
imaging in fluorescence mode, has been considered a cost-effective method for fluorescein
angiograms (Bernardes et al., 2011). Fundus autofluorescence imaging, which focuses on the
fluorescent properties of pigments in the retina, has been used to diagnose various disease processes
clinically (Schmitz-Valckenberg et al., 2008). Some customized topical endoscopy funduscopes
for rodents have been demonstrated for the screening of the ciliary body and retinal vessels (Paques
et al., 2007) or monitoring gene expression in transduced neurons (Schejter et al., 2012). Such
systems offer the advantage of user-friendly procedures that facilitate the training and reduce
execution time needed to perform the examination.
We report here the development and validation of a custom endoscope-based fundus system
for monitoring and characterizing in vivo brightfield and fluorescence retinal images, based on low-
cost adaptations of a simple fundus imaging system similar to that proposed by Paques et al.
(Paques et al., 2007) and Schejter et al (Schejter et al., 2012). To monitor the genetic expression of
GECI GCaMP6f (Chen et al., 2013b) transduced in mouse RGCs by adeno-associated viral vector
(AAV), we performed long-term tracking following intravitreal injection and identified an optimal
window of expression. The in vivo imaging data was then compared to in vitro calcium imaging
using a retinal wholemount preparation.
58
2.2 Methods
2.2.1 Overview
Adult mice (C57BL6/J) receiving an intravitreal injection of an AAV vector encoding a GECI
(AAV2-CAG-GCaMP6f) were used to perform fundus imaging and calcium imaging experiments.
One group of retinas was imaged in vivo by our fundus imaging system from week 1 to week 5 post
injection at 4-day intervals. The other group of retinas was dissected at defined time points to
identify the optimal time following AAV transduction for performing in vitro calcium imaging
experiments, which are designed to determine the neurophysiological properties of RGCs
expressing GCaMP6f. All procedures were approved by the Institutional Animal Care and Use
Committee (IACUC) and the Institutional Biosafety Committee (IBC) at the University of Southern
California.
2.2.2 Animal
Two strains of animals were used. To assess inter-experiment differences with in vivo
monitoring, the Thy1-YFP-H line transgenic wild-type (WT) YFP-expressing mice developed by
Feng et al. were selected since they have consistently, sparsely-labeled RGCs (< 10%), suitable for
visualizing individual cells without overlap (Feng et al., 2000). These cells should have stable
expression of YFP. For tracking the GECI expression at different time points, adult female and
male mice C57BL6/J (Jackson Lab, Bar Harbor, Maine), aged postnatal day 60-180, were used for
both in vivo monitoring and in vitro assessment.
59
2.2.3 Funduscope
The fundus imaging system is illustrated in Figure 2.1A, with divided excitation and emission
pathways to facilitate fluorescence imaging of the retina at single-cell resolution. For the emission
pathway (Figure 2.1A, C), an endoscope with a 5-cm otoscope and 3-mm outer diameter (1218 AA,
Karl Storz, Tuttlingen, Germany) was positioned in front of the camera. Assembled with a step-
down ring, the manufactured adaptor was used to connect the digital camera (D5100 with AF-S
VR Micro-Nikkor 105mm f/2.8G lens, Nikon, Tokyo, Japan) with the endoscope, providing the
ability to install optical emission filters. For the excitation pathway (Figure 2.1A, B), a xenon lamp
(LH-M100CB-1, Nikon, Tokyo, Japan) was used as the light source to generate collimated light.
Depending on the chosen mode, the light was projected through the excitation filter or neutral
density filter on to a custom-made optical fiber connector that transmitted the light source to the
endoscope by means of a commercial optic fiber cable (495NA, Karl Storz, Tuttlingen, Germany).
For fluorescence imaging, the illumination power in the band of interest (centered on λ = 480nm)
was calibrated to 6mW by bench top optical power meter (1936-R, Newport, Irvine, CA). This
power setting ensures that the average brightness of the obtained fluorescent fundus images and the
SNR are acceptable, while remaining in the range of permissible exposures for ocular safety (Delori
et al., 2007).
Fluorescence images were obtained using a 469 nm excitation filter (MF469-35, Thorlabs,
Newton, NJ) placed in the removable filter holder on the light source breadboard (Figure 2.1B),
and a 535 nm emission filter (MF525-39, Thorlabs, Newton, NJ) positioned in the adaptor in front
of the camera lens. Brightfield and green-field images were obtained with and without the emission
filter, respectively, and in both cases the excitation filter was replaced with a neutral density filter
(NE06B-A, Thorlabs, Newton, NJ) attenuating the illumination power by 75%. The camera settings
are listed in Table 2.1.
60
Figure 2.1: (A) System schematic diagram. 1: light source, 2: convex lens, 3: excitation filter, 4:
objective lens, 5: optic fiber, 6: endoscope, 7: in vivo eye, 8: adaptor, 9: emission filter, 10: lens of
camera, 11: camera. (B) Excitation part of the system, including compartments 1 through 5 in (A).
(C) Emission part of the system, including compartments 6 through 11 in (A).
Parameters Settings
Image resolution fine (4928×3264)
Focus manual operating mode
Aperture F 3.3
ISO 3200
Focal length 105 mm
Shutter speed 1/8 sec (bright and green field)_
8 - 25 sec (fluorescence)
Table 2.1: Camera setting for the proposed fundus imaging system.
61
2.2.4 Fundus Imaging
For in vivo fundus imaging, the mice were anesthetized using a mixture of ketamine (80mg/kg)
and xylazine (10mg/kg). The pupil was dilated with 0.5% tropicamide and 2.5% phenylephrine
hydrochloride. Topical tetracaine hydrochloride was applied for local corneal anesthesia. The
animals were placed on the adjustable platform while the eye was positioned such that it lightly
touched the endoscope tip. A drop of saline (NaCl 0.9%) was used to keep the eye hydrated and
optically coupled to the endoscope. To make images comparable at different time points, the edge
of dilated pupil was arranged perpendicular to the endoscope tip (Figure 2.2A), and the optic disc
was aligned at the center of camera under green-field mode (Figure 2.2B) before performing the
fluorescent data collection.
Figure 2.2: Alignment between the tip of endoscope and mouse eye. (A) Manipulate the platform
as well as the angle of the head so that the eye can barely touch the endoscope tip, with the objective
parallel to the dilated pupil. (B) Alignment of the center of image to optic disc (red square dot on
the camera screen).
2.2.5 Imaging Processing
For the fluorescence intensity analysis, we used MATLAB (The MathWorks, Natick, MA) for
post-processing. To eliminate the temporal and spatial variation caused by the slight variations in
the optical pathway of the eye or the alignment between the eye and the tip of endoscope, the green-
62
field intensity was used for normalization. The green channel was first separated and extracted from
the raw RGB fluorescence image, and 10 consistently observable cells expressing GCaMP6f
fluorescence across different time points were randomly selected from the fundus images. The
normalized fluorescence intensity was calculated by the average of fluorescence intensity within
each cell divided by the baseline green field reflection taken from an identical region of interest
(ROI) (eq. (1)). For each time points, 6 to 8 trials of image pairs were taken to compute the final
value.
(1) ROI in the pixel of number where
) , (
1
) , (
1
N
y x f
N
y x f
N
I
I
I
ROI
field green
ROI
t fluorescen
baseline
t fluorescen
normalized
To count cells, we used ImageJ (National Institutes of Health, Bethesda, MD). The green
channel was first separated from the raw fluorescence data, and the background subtraction tool
was applied with a rolling ball radius of 40 pixels. After removing the background, the inverted
image threshold was adjusted so that distinct cell contours could be clearly observed. To eliminate
the noise generated by the CCD camera, the noise correction function, including outlier removal
and despeckle, were used in order to eliminate the hot and dead pixels. The images were further
processed by binary function fill holes and watershed functions to fill in the nucleus and partition
the overlapping cells. Finally, the analyze particles function was used to count the number of
observable cells, with the size set to 500-infinity. Figure 2.3 shows representative fundus images
and selected ROIs.
63
Figure 2.3: Calculation of normalized fluorescence intensity of individual cell. Locally selective
fluorescence intensity was normalized by the green field intensity from identical region of interest.
(A) Group of cells selected from fluorescence fundus image. (B) Baseline captured from identical
location of relevant green field image.
2.2.6 Viral Vector
The pGFP plasmid (Klein et al., 2002, Wu et al., 2003) was selected as the backbone for the
GECI-containing vector. The original GFP sequence was removed and replaced with the GECI
GCaMP6f (Chen et al., 2013b) to create pAAV-CAG-GCaMP6f. A CMV enhancer, chicken β-
actin promoter (CBA promoter), exon, and intron were collectively used to form a ubiquitously
strong CAG promotor, located upstream of the cassette encoding GCaMP (Niwa et al., 1991). To
enhance protein translation, woodchuck hepatitis virus posttranscriptional regulatory element
(WPRE) was placed downstream of the transgene (Loeb et al., 1999). The entire cassette was
flanked by AAV2 inverted terminal repeats. Figure 2.4 shows the complete sequence. Recombinant
AAV vectors were produced by the two-plasmid co-transfection method (Zolotukhin et al., 1999,
Weitz et al., 2013). Final concentration of AAV2-CAG-GCaMP6f was 2.6 x 10
12
vector genomes
per milliliter. Viral stock was consistently diluted to 1.04 x 10
12
with balanced salt solution before
injection to eliminate the factor of concentration.
64
Figure 2.4: Map of pAAV2-CAG-GCaMP. The CMV enhancer, chicken β-actin promoter (CBA
promoter), exon, and intron collectively form the CAG promoter. Woodchuck hepatitis virus
posttranscriptional regulatory element (WPRE) is placed downstream of the GCaMP6f transgene
to increase protein translation. The cassette is flanked by adeno-associated virus type 2 (AAV2)
inverted terminal repeats (TR).
2.2.7 Electrical Stimulation
Transparent MEAs were fabricated in a class 100 cleanroom. Arrays were patterned by
selective etching of indium tin oxide (ITO) on no. 1 cover glass substrates (Vaculayer, Mississauga,
Ontario, Canada). A dual-insulation layer, including SU-8 epoxy photoresist and silicon nitride,
was formed atop the ITO. The insulation layer was removed to create 200-µm-diameter disk
electrodes. These dimensions are within the range of present day retinal prostheses.
Electrical stimuli consisted of 50-sec trains of charge-balanced, biphasic current rectangular
pulses with 5-ms duration per phase at 20 pulses per second to produce a robust calcium transient.
Voltage stimuli were generated from a computer-controlled stimulus generator (STG-2008, Multi
Channel Systems, Reutlingen, Germany) and fed through a custom voltage-to-current converter. A
customized interface circuit board was used to relay the signal to the designated electrode. A
platinum wire encircling the top of the recording chamber served as the return electrode. The
amplitude of pulse train was selected to be 1.2 times the maximum threshold of RGCs in the region
of interest, where the threshold set as 5% increase in calcium fluorescence ( ∆ 𝐹 / 𝐹 ).
65
2.2.8 Calcium Imaging
Virus-injected mice were euthanized at times corresponding to times at which fundus imaging
was conducted, in order to assay the neurophysiological responses as a function of post-injection
time of RGCs expressing GCaMP6f. After anesthesia with ketamine/xylazine, the mice were
rapidly decapitated and the treated eye was enucleated and hemisected with Vannas spring scissors.
To flatten the retina, 4 cuts were made, from periphery to center, to create quadrants of near equal
size. Vitreous was gently peeled from the retina surface with fine forceps to allow for a tight
interface between the retina and MEA.
After removal from the eye cup, the retina was mounted on a porous membrane (cat. No.
JVWP01300; Millipore) and placed on the transparent MEA chamber with ganglion cell side facing
down. Images of calcium fluorescence were acquired through an inverted epifluorescence
microscope. Fluorescence excitation was provided by a super bright cool white light-emitting diode
(LED). Excitation and emission light were filtered through a commercial filter set (GFP-4050A,
Semrock, Rochester, NY) for GCaMP6f. Images were viewed through a Nikon (Tokyo, Japan) Plan
Apo 0.75-numerical aperture (NA) ×20 objective and captured by an electron-multiplied charge-
coupled device (EMCCD) camera. (iXon, Andor Technology, Belfast, Northern Ireland).
For superfusion, bicarbonate-buffered Ames’ Medium (Sigma-Aldrich, St. Louis, MO) was
used in all procedures. Media was supplemented with penicillin-streptomycin to prevent bacterial
growth, equilibrated with 5% CO2-95% O2 gas, and adjusted to pH 7.4 and 280 mOsm. During the
course of each experiment, the retina was continuously superfused at a flow rate of 4–5 ml/min and
a temperature of 33°C.
66
2.3 Results
2.3.1 System Performance
Brightfield, green-field and fluorescence fundus images were acquired from virus-transduced
GCaMP6f mice. Figure 2.5(A) illustrates a representative brightfield image in which the blood
vessels radiating from the center of optic nerve can be visualized clearly through the funduscope
system. Moreover, the axon bundles of RGCs located within nerve fiber layer can be observed near
the center when higher illumination power is used. The white/grey region on the right side of the
view faithfully depicts the needle track of the needle used for intravitreal injection of the virus
(indicated by yellow arrow). In the fluorescence images, axons and individual RGCs can be
visualized (Figure 2.5C, D) and appeared to be relatively well resolved when compared to the
previous fluorescence endoscope-based fundus imaging work [8]. Tests of different exposure time
settings showed that 25 sec is required to observe GCaMP6f-expressing cells located at the
peripheral regions of fundus in fluorescence mode, especially for the period when GCaMP6f is
sparsely expressed (Figure 2.6). However, the extended exposure might also lead to the imaging
blurring caused by breathing artifact or other head motion. The fluorescence intensity profile of
individual GCaMP6f RGCs shown in Figure 2.7 demonstrate the spatial resolutions of an individual
cell remain relatively consistent across different exposure settings, even when the high exposure
setting was used.
67
Figure 2.5: Representative fundus images in a C57 mouse. (A) Bright field image. (B) Green-field
image. (C) Fluorescence image. (D) Magnification of (C). The yellow arrows in (A) and (B)
indicate the track of virus injection needle.
68
Figure 2.6: Representative fundus images under different exposure setting of virus-induced
GCaMP6f in mouse RGCs at 15 day post injection. From (A) to (E) are 8s, 10s, 15s, 25s
respectively.
69
Figure 2.7: Evaluation of the ability of the system to resolve individual cells. Top row: Zoomed in
normalized fluorescent fundus images of GCaMP6f-expressing RGCs under different exposure
time settings. Red lines illustrate the path along which intensity profiles were taken. Bottom row:
Intensity profiles taken along the cross-sections of soma, fitted with Gaussian curves. The standard
deviation of each Gaussian curve is listed (unit: pixel).
2.3.2 Long-term Expression Tracking
The long-term tracking of the fundus of transgenic YFP-expressing mice is shown in Figure
2.8. Generally, a similar fluorescence pattern can be observed consistently at different time points
in terms of YFP-RGCs, though the background and overall fluorescence slightly declined over time.
This experiment also demonstrates that since identical field of view cannot be obtained for different
time points, due to the slightly different alignment for each measurement, a normalization
procedure relative to the baseline green-field is necessary for quantitatively analyzing the change
of fluorescent intensity.
70
Figure 2.8: Representative fundus images demonstrating the expression of transgenic YFP in
mouse RGCs. From (A) to (E) are 7, 11, 15, 19, 23 days corresponding to the time points for the
experimental GCaMP6f group.
71
Figure 2.9: Representative fundus images demonstrating the long-term expression of virally-
transduced GCaMP6f in mouse RGCs. From (A) to (E) are 11, 15, 19, 23, 27 days post injection,
respectively.
72
Representative fluorescence images for different stages of virus-induced GCaMP6f expression
are shown in Figure 2.9. At 1 to 1.5 week post injection, only a few RGCs could be distinguished
from the relatively weak background. The number of observable cells, as well as the fluorescence
intensity of each cell, both demonstrated a consistent rising trend with time. In addition, there was
a gradual increase in background intensity at late stages, which may be due to increased scattering
seen with enhanced fluorescence intensity.
Figure 2.10 shows the statistical analysis for the change in RGC fluorescence intensity and cell
number across different experimental and control subjects (n=5 and n=1), respectively. In the
intensity analysis, each curve represents the averaged normalized intensity from 10 different RGCs.
The results indicate that RGCs share a similar trend of rising intensity from week 1 to week 2, post
injection. However, for most of the cases, the fluorescence intensity reached a plateau or decreased
slightly at 2-3 weeks. In contrast, for the control transgenic YFP retina, the normalized intensity
was relatively unchanged or slightly decreased over time, as observed in Figure 2.8. The only
dramatically declining case (subject: FT-3c) was later found to be attributable to the deterioration
of the crystalline lens. The average slope of the normalized fluorescence intensity before and after
week 2 are 0.2997 (n=5) and 0.0990 (n=4, one animal removed due to deterioration of crystalline
lens) respectively, with p<0.05. The significant change of fluorescence intensity in terms of slope
can serve as an effective indicator for determining the optimal time window.
As for the number of fluorescent cells, a progressively increasing trend can also be identified
in GCaMP6f-injected animals, which is to be contrasted with the stable number of cells in the
transgenic YFP mice. Both findings support a uniform increase in both fluorescence within a cell
and the number of cells expressing GCaMP6f.
73
Figure 2.10: Normalized intensity in selective RGCs and number of observable RGCs in the fundus
field of view across 5 GCaMP6 expressing and 1 control YFP subjects. For each time point, 6-8
trials of data were averaged to demonstrate the tendency of fluorescent expression.
2.3.3 In vitro Calcium Imaging Validation
To test whether the images obtained by fundus imaging could be used to establish optimal
experimental time windows for neurophysiological experiments, virus-injected mice were used for
in vitro calcium imaging at time points coinciding with in vivo imaging time points. As shown in
Figure 2.11, burst electrical stimulation with an ITO electrode was applied to elicit calcium
74
transient fluorescence responses in GCaMP6f-expressing RGCs in the field of view. The difference
between the post-stimulus and baseline images show that the responsive cells can be located and
correctly labeled in the field of view.
Figure 2.11: Electrical stimulation activates RGCs as revealed through large changes in GCaMP6f
fluorescence intensity in well expressed region in vitro. (A) Before Stimulation, cells are at baseline
fluorescence. (B) Original calcium imaging in response to external stimuli from the ITO electrode
in the middle, the electrically activated cells became visibly brighter. (C) Image subtraction of (B)
from (A) baseline highlights the responding cells in the field of view. (D) Normalized changes in
fluorescence ( ΔF/F) for responsive and non-responsive RGCs expressing GCaMP6f in response
to the 20 Hz 5 ms duration biphasic symmetric pulse train stimuli. The red and black arrows indicate
the onset and offset of the stimulation pulse train.
Table 2. Responsive cells (in vitro)
Time post injection Expression Ratio
(mean)
Expression Ratio
(std)
# of RGCs w/
GCaMP6f
2 weeks 29.60% ±6.38% 43.3
3 weeks 86.97% ±3.83% 86.3
≥4 weeks 69.24% ±5.00% 94.0
75
Table 2.2: Statistical results for the ratio of responsive RGC versus total cell expressing GCaMP6f
in the region of interest. The data was acquired from 3-4 well-expressed regions of each retina.
Since the expression is highly non-uniform across the retina, we selected 3 to 4 regions where
expression is relatively higher in terms of number of GCaMP-expressing cells as the region of
interest (Table 2.2, column 4). This sampling strategy is also used for in other retinal calcium
imaging experiment because more cells with expression allow more data to be collected in a single
experiment. The statistical results shown in Table 2 indicate that the ratio for the responsive RGC
versus the total observable cells was highest at week 3 post injection. For earlier time points, there
were few observable cells and the responsiveness rate of those cells was also extremely low. If the
incubation time was longer than 3 weeks, although the number of GCaMP6f-expressing cells was
high, the responsive ratio decreased significantly from week 3.
The in vitro calcium imaging results correspond to the in vivo fundus imaging results shown in
Figure 2.9 and 2.10. In most cases, the fluorescence intensity within RGCs began to increase around
1 to 1.5 week and started to plateau or decrease slightly around 3 week, which parallels the trend
of in vitro neurophysiological responsiveness. The optimal time window for the proposed virus-
transduced calcium indicator can thus be determined in terms of the normalized fluorescence
intensity trend. When the curve reaches a plateau or demonstrates decreasing trend with significant
reduction of slope, the GCaMP6f-expressing retina has optimal expression the for electrical
stimulation experiment.
76
2.4 Discussion and Conclusion
We have presented an in vivo fundus system that enables high-resolution digital fundus imaging
in the mouse eye. The system is built of off-the-shelf components. Additionally, the long-term
tracking experiment demonstrates the system can serve as a tool for obtaining consistent, cell-
resolved fluorescent images, and this information can be used to monitor level of GECI expression,
thus greatly reducing the time to evaluate transduction outcomes. Moreover, the in vitro calcium
imaging shows that the fluorescence fundus imaging can be used to determine the best time window
for neurophysiological studies of transduced RGCs.
2.4.1 Head motion and optical blurring
Most of fundus imaging acquisition techniques require head fixation of animals to eliminate
the motion caused by breathing, thus improving the spatial resolution. However, in our experiments,
the imaging quality did not improve significantly when the mouse head was fixed. Tests of different
exposure time settings for GCaMP6f expressing RGCs show fluorescence intensity profiles with
similar standard deviation for those cells consistently appearing near the optic disc. (Figure 2.6, 2.7)
This observation suggests the effective resolution is primarily limited by optical point spread
function, instead of movement-related blurring. Instead, offline image processing techniques such
as those applied in super resolution image reconstruction (Sung Cheol et al., 2003) and
deconvolution algorithms (Swedlow, 2013) possibly can be used to enhance the resolution.
2.4.2 Photobleaching of fluorophore
Previous studies have shown that the fluorescence intensity decreases during illumination with
a high intensity light source, due to photobleaching of the fluorophores (Patterson and Piston, 2000,
Morgan and Pugh, 2013). In our experiments, the power and the exposure time were set to 6mW
77
and 25sec, to limit illumination exposure to a level that still ensured high signal-to-noise ratio
(SNR). In our transgenic YFP control experiments, we observed a slight reduction of background
and overall fluorescence at late-stages. For GCaMP6f-transduced RGCs, on the other hand,
fluorescence showed a net rising trend overall, most likely due to continued expression of
GCaMP6f.
2.4.3 Cytomorbidity
Our findings demonstrate that fluorescence intensity, measured in vivo by imaging the fundus,
generally reaches a plateau around 3-4 weeks post injection. At the time of this plateau, the
responsive ratio decreases in vitro. Further investigation indicates that some of neurons with
aberrant responses at that stage are those with GCaMP fluorescence appearing in the nucleus, while
the others display relatively stable baseline fluorescence, resembling the results on cytomorbidity
proposed in one of the first studies on GCaMP6 (Chen et al., 2013b) and previous generations of
GCaMP (Weitz et al., 2013). Thus, virus-induced cytomorbidity most likely contributes to the in
vitro reduction of the responsive ratio as well as the in vivo fluorescent intensity plateau at late
times.
2.4.4 Limitations of Calcium imaging
One drawback to calcium imaging in studies of WT retina is that the excitation light causes
bleaching of photoreceptors, which limits the applicability for studies involving light stimulation.
To minimize photobleaching at photoreceptors while imaging calcium dynamics at inner retinal
neurons, some groups have used two-photon (infrared) excitation which can be used to selectively
excite fluorescence only in inner retinal neurons. For our studies, however, which rely on
extracellular electrical stimulation (the method of stimulation with epiretinal prosthetic implants,
such as the Argus II) of retinal neurons, even strong excitation light does not inhibit RGC or bipolar
78
neuron responsiveness. Therefore, calcium imaging is still a valuable tool for investigating the
spatial distribution of inner retinal neuron responsiveness in response to electrical stimulation.
79
CHAPTER III: Strategies for Selective Activation
of Retinal Ganglion Cells.
Abstract
Retinal prosthetic implants have shown potential to restore partial vision to patients blinded by
retinitis pigmentosa or dry age-related macular degeneration, via a camera-driven multielectrode
array that electrically stimulates surviving retinal neurons. Commercial epi-retinal prostheses
mostly use charge-balanced symmetric cathodic-first biphasic pulses to depolarize retinal ganglion
cells (RGCs) and bipolar cells (BCs), resulting in the perception of light in blind patients. However,
previous clinical study for patients with Argus II epiretinal implants reported most percepts evoked
by single electrode stimulation were elongated and aligned with estimated axon path of retinal
ganglion cells, suggesting the activation of axon bundles. In this project, using an established
genetically encoded calcium indicator (GECI), we performed in vitro calcium imaging for different
stimulation paradigms, focusing primarily on short duration pulse that can avoid axonal stimulation
and selectively activate targeted RGC soma. The findings support the possibility to manipulate the
responses of RGCs through varying the stimulation waveform, thus potentially forming more ideal
shape perception with higher spatial resolution in future retinal prosthesis design.
80
3.1 Introduction
Retinitis pigmentosa (RP) and dry age-related macular degeneration (AMD) (Stanga et al.,
2016) are two prevalent degenerative diseases of retina that lead to significant visual impairment
or blindness. Prior clinical testing demonstrated that applying electrical stimulation to a
degenerated retina can elicit visual percepts (Humayun et al., 1999). Based on these findings,
several retinal prostheses have been developed and two systems have regulatory approval with the
best reported 20/1260 and 20/546 visual grating acuity respectively (Yue et al., 2016). Long term
human testing studies of patients implemented with those devices show improved object
discrimination and localization, mobility, and, in some patients, the ability to recognize simple
patterns such as letters (Humayun et al., 2012, Zrenner et al., 2011). With the Argus II epiretinal
implants, although subjects are only capable for reading large letters at a rate slower than normal
reading, the improvements are significant considering that these RP subjects demonstrated bare or
no light perception prior to implementation.
One of the most critical challenges of retinal implants is to achieve ideal shape perception. That
is, each electrode should only activate nearby retinal cells, thus forming small round visual percepts.
Integrating multiple percepts generated from different electrodes, the complex perception of shape
such as a letter can be created. However, clinical studies of patients with Argus II epiretinal
implants reported most evoked percepts by single electrode were elongated and aligned with
estimated axon path of retinal ganglion cells (RGCs), suggesting the activation of axon bundles
(Nanduri et al., 2012). Spatial responses consistent with axonal stimulation have also been
measured using calcium imaging techniques and in vitro rodent retina (Weitz et al., 2013).
Stimulation strategies that avoid axonal stimulation and decrease the threshold of targeted neurons
may significantly improve prosthetic vision in terms of spatial resolution.
81
The final output of the retina coms from RGCs. Those neurons are a main targets for retinal
prosthetic research since the visual percepts are mostly determined by their activation patterns. The
RGC neurons can be activated directly by sufficient depolarization of the RGC membrane or
indirectly via synaptic transmission from activated bipolar cells (BP) (Greenberg et al., 1999, Shah
et al., 2006, Fried et al., 2006b, Sekirnjak et al., 2006). Direct stimulation has the advantages of
precise control of retinal output, thus providing superior temporal resolution. Advantages of
indirect stimulation include the retention of the neural processing that occurs at the inner plexiform
layer and avoiding direct activation of axons of passage. Studies have showed the BPs tend to
respond preferentially to longer pulse width (25 Hz sinusoidal) (Freeman et al., 2010) and more
localized responses of RGCs mainly resulted from indirect stimulation can be achieve with 25-ms
duration pulses (Weitz et al., 2015). However, desensitization of the RGC responses to continuous
indirect stimulation argues against this approach. It is well documented that, depending on the
stimulation frequency, sensitivity of the electrically-evoked RGCs progressively decreased with
the repeated indirect stimulation (Freeman and Fried, 2011, Jensen and Rizzo, 2007). This
desensitization in the cellular response is observed from multiple animal models through electrical
recording and is believed to be highly correlated with the percept fading (Stronks and Dagnelie,
2014, Ahuja et al., 2008) reported by patients. Therefore, the possibility for direct stimulation of
RGC cell bodies, which are more easily excited by short duration pulse (< 150 µs) (Fried et al.,
2006b), is worth further investigation to alleviate phosphene fading in response to continuous
stimulation.
To test different potential stimulation parameters, we developed a virus-transduced GECI
GCaMP6f and performed calcium imaging to record the neural activity from RGCs at single cell
resolution in wholemount retinas while applying stimulation through a microelectrode array (MEA)
with transparent indium tin oxide electrodes. The results suggest that the electrical stimulation
thresholds and response of RGCs can be manipulated through pulse duration, phase, and profile of
82
pulses to form more confined pattern nearby the working electrode and achieve reduction of RGC
threshold.
83
3.2 Methods
3.2.1 Overview
Adult mice (C57BL6/J) receiving an intravitreal injection of an AAV vector encoding a GECI
(AAV2-CAG-GCaMP6f) were used to perform calcium imaging experiments. Based on the
established in vitro calcium imaging and electrophysiological mouse animal model, we tested
different stimulation paradigms that can avoid axonal stimulation, manipulate the thresholds in
response to continuous stimulation. All procedures were approved by the Institutional Animal Care
and Use Committee (IACUC) and the Institutional Biosafety Committee (IBC) at the University of
Southern California and University of Michigan, Ann Arbor.
3.2.2 Animal
For preliminary studies, the wide-type C57BL6/J mice (aged between 1-3 month) purchased
from Jackson Lab were used to preform calcium imaging experiment, aiming to find the stimulation
parameters that can selectively active RGCs nearby electrode. The results were further validated
through B6.CXB1-Pde6brd10/J mice (aged between1-3 month). The strain line has mutations in
phosphodiesterase (PDE) that are associated with some types of retinitis pigmentosa and night
blindness. Homozygous animal for retinal degeneration 10 (rd10) exhibit sclerotic retinal vessels
at 4 weeks of age, progressive retinal outer nuclear layer degeneration beginning at 16 days, and a
progressive decline in rod and cone ERG a- and b- waves. Although similar to rd1, the rd10
phenotype has a later onset and delayed retinal degeneration which provide a better experimental
model for retinitis pigmentosa.
84
3.2.3 Virus-transduced Calcium Indicator
Recombinant AAV2-CAG-GCaMP6f was constructed with pGFP plasmid where the original
GFP sequence was removed and inserted with the GECIs GCaMP6f, following a procedure similar
to that used for another plasmid used in our lab with GCaMP5G (Weitz et al., 2013). Before
GCaMP6f cDNA, the CMV enhancer, chicken β-actin promoter (CBA promoter), exon, and intron
were collectively used to form ubiquitously strong CAG promoter. To enhance protein translation,
woodchuck hepatitis virus posttranscriptional regulatory element (WPRE) was placed downstream
of the transgene. The entire cassette was flanked by AAV2 inverted terminal repeats. Recombinant
AAV vector was produced at the University of Florida Vector Core by the two-plasmid co-
transfection method (Weitz et al., 2013). Final concentration of AAV2-CAG-GCaMP6f was 2.6 ×
10
12
vector genomes per milliliter (vg/ml). To limit the viral expression and cytomorbidity, viral
stock was diluted to 1.04 × 10
12
vg/ml with balanced salt solution before injection. With this scale
of concentration, a previous study with older-generation AAV2-CAG-GCaMP5G shown that the
same type virus has high specificity with rodent RGCs after intravitreal injection (~85%) (Weitz et
al., 2013).
3.2.4 Intravitreal AAV Injection
Animals were anesthetized with intraperitoneal injection of a mixture of ketamine (80mg/kg)
and xylazine (10mg/kg). The pupil was dilated with 1% tropicamide and 2.5% phenylephrine
hydrochloride. Topical tetracaine hydrochloride was applied for local corneal anesthesia. For
intravitreal injection, a pilot hole near the cornea (0.5-1mm posterior to corneal limbus) was first
made using a 30-gauge needle through sclera, choroid, and retina. A 5µL microliter syringe
(Hamilton) with a blunt 32-gauge needle was inserted through the pilot hole and slowly injected
1µL of AAV2-CAG-GCaMP6f into the vitreous on top of retina over a roughly 30-sec period. After
85
injection, the needle remain in place for another 30-sec and withdrawn with slow pace to prevent
leakage. Antibiotic eye ointment was applied at the end for helping wound sealing and preventing
infection.
3.2.5 Calcium imaging
Virus-transduced mice were euthanized at 3-4 weeks post injection, which was determined by
a customized fundus imaging system to be the best time window for optimal GCaMP6f expression
in RGCs (Chapter 2) (Chang et al., 2016, Chang et al., 2017). After anesthesia with
ketamine/xylazine, the mice were rapidly decapitated and the treated eye was enucleated and
hemisected with Vannas spring scissors. To flatten the retina while still attached the posterior
eyecup, 4 cuts were made, from periphery to center, to create quadrants with near equal size.
Vitreous was gently peeled from the retina surface with fine forceps to allow for a tight interface
between the retina and MEA. After removal from the eye cup, the retina was mounted on a porous
membrane (cat. No. JVWP01300; Millipore) held by titanium ring and placed on the transparent
MEA chamber with ganglion cell side facing down. The wholemount retina was imaged using a
customized up-down microscope (Olympus, Center Valley, PA). Fluorescence excitation was
provided by a super bright cool white light-emitting diode (LED) by means of inverted optical path.
Excitation and emission light were filtered through a commercial filter set (49002 - ET - EGFP
(FITC/Cy2), Chroma Technology Corp, Bellows Falls, VT) for GCaMP6f. Images were viewed
through an Olympus (Center Valley, PA) UPLFLN 0.3-numerical aperture (NA) x10 objective and
captured by an electron-multiplied charge-coupled device (EMCCD) camera. (iXon 897, Andor
Technology, Belfast, Northern Ireland) at 10Hz (75% exposure duty cycle). For superfusion, the
bicarbonate-buffered Ames’ Medium (Sigma-Aldrich, St. Louis, MO) was used in all procedures.
Media was supplemented with penicillin-streptomycin to prevent bacterial growth, equilibrated
with 5% CO2 - 95% O2 gas, and adjusted to pH 7.4 and 280 mOsm. During the course of each
86
experiment, the retina was continuously superfused at a flow rate of 4–5 ml/min and a temperature
of 33 °C.
Figure 3.1: Calcium imaging experimental setup. (A) Recording chamber contains a transparent
microelectrode array that delivers current-controlled electrical stimulation. An inverted
microscope equipped with an EMCCD camera is juxtaposed below the recording chamber. (B) An
inverted and upright microscope with the EMCCD camera installed at bottom left port. (C) MEA
recording chamber and interface board, with glass pipes for superfusion. (D) Heat controller, MCS
stimulator, and the other end of interface board.
87
3.2.6 Electrical Stimulation
Transparent MEAs were fabricated in the W. M. Keck Photonics Laboratory at USC, a class
100 cleanroom. Arrays were patterned by selective etching of indium tin oxide (ITO) on no. 1 cover
glass substrates (Vaculayer, Mississauga, Ontario, Canada). A dual-insulation layer, including SU-
8 epoxy photoresist and silicon nitride, was formed atop the ITO. The insulation layer was removed
to create 200 µm diameter disk electrodes, in a 6×10 pattern with 500 µm electrode pitch. These
dimensions are within the range of present day retinal prostheses.
Electrical stimuli consisted of charge balanced, biphasic, square current pulse waveform.
Current stimuli were generated from a computer-controlled stimulus generator (STG-4008 – 1.6mA,
Multi Channel Systems, Reutlingen, Germany) and fed through a custom capacitive isolation
circuitry to prevent DC leak current. A customize interface circuit board was used to relay the signal
to the designated electrode. A platinum wire encircling the top of the recording chamber served as
the current return electrode (submerged into the perfusion solution).
All stimuli were repetitive to evoke a burst of spikes and generate a detectable calcium transient
(Weitz et al., 2013). Three types of rectangular pulses, including symmetric cathodic-first,
symmetric anodic-first, and asymmetric anodic-first were applied to the retina, with individual
pulse durations from 40 µs to 4 ms at 120 Hz depending on different experimental setting. The
asymmetric biphasic pulse consists of an anodic first phase of relatively long duration and low
amplitude and a cathodic second phase of short duration and high amplitude. Table 3.1 provides
the pulse width selection of cathodic phase for each waveforms. The ratio is computed by dividing
the longer phase width by the shorter phase width. For each waveform and duration, stimulation
protocols were designed so that current amplitude progressively increase from subthreshold to
suprathreshold. A total of 10 amplitudes were used, each lasting 5 secs with 20 secs interval in
between to bring calcium transient back to baseline. No stimuli were delivered during the first and
last 5 secs of each protocol. The overall stimulation protocol is shown in Figure 3.2.
88
Table 3.1: Electrical stimulation waveform parameters.
Figure 3.2: Stimulation protocol. Stimuli were delivered 10 times on 5-seconds intervals with
monotonically increasing amplitudes and 20-seconds resting intervals in between. Each stimulus
was a burst of rectangular pulses with symmetric cathodic-first, symmetric anodic-first, or
asymmetric anodic-first pulse paradigm designed to evoke a burst of spikes and generate a
detectable calcium transient.
3.2.7 Spatial Threshold Mapping
For each stimulation paradigm, the calcium imaging of RGCs nearby the working electrode is
recorded at 10 fps. Based on the stimulation paradigm, the images during simulation (2-3 secs after
pulse train initiation) were extracted and the baseline image was subtracted (baseline obtained 1
sec before stimulation initiation), so the calcium transient of responsive RGCs can be detected. For
each pixel, the obtained transient ( ΔF) was further normalized with respect to baseline (F) to
89
calculate the percentage of fluorescent change ( ΔF/F). With proper threshold selection to remove
noise (> 15%), processed spatial and temporal information in the region of interest can form the
spatial threshold maps.
3.2.8 Statistical Analysis
The 2-D mapping information can further be used to perform activated region analysis which
directly corresponds to shape of visual percepts and spatial acuity. Based on the geometry of
neurons with respect to the working electrode, the evoked RGCs can further be directly associated
with somatic activation and axonal activation respectively, depending on the relative distance to
the center of working electrode. We define the responses occurred directly above the working
electrode as somatic activation region (forming ideal visual percept), whereas those occurring
outside two times the radius of electrode as axonal activation region (forming elongated shape
percept).
Statistical analysis for activated neurons within the two defined regions has been used show
the ability to form localized response nearby the electrode for different stimulation patterns.
Moreover, response patterns with varying pulse width of stimulation was used to establish the
strength duration curves for targeted regions, thus offering the possibility for optimizing the
stimulation parameters. For responsive area analysis, the effective regions under different
thresholds were computed using the geometry information of threshold mapping (Summation of
activated neurons on pixel basis). All post-processing steps and analysis of calcium imaging were
analyzed by MATLAB (The Mathworks, Natick, MA) on pixel-by-pixel basis. We also performed
the correlation analysis of normalized 2-D calcium imaging transient between responses elicited by
different stimulation waveforms. For waveform with same current amplitude or charge injection,
this index can be used to test the efficiency of each applied stimulation pattern.
90
Figure 3.3: Illustration of correlation tests between responses. Example of correlations of response
patterns between 20 times ratio asymmetric anodic-first pulse with different delivered charges and
control symmetric cathodic-first pulses with fixed delivered charge (60 µA) that generally leads to
localized response.
91
3.3 Results
3.3.1 GCaMP Expression Profile
Representative retinal whole mounts for wild-type (WT) and retinal degeneration (RD) mice
are shown in Figure 3.4. For WT retina, the GCaMP expression exhibiting green fluorescence is
generally abundant across the four quadrants. Typically, the baseline fluorescent intensities within
ganglion cells bodies were greater than their axons, so the dynamic of calcium transients of somas
can be observed through the transparent superficial never fiber layer. The brightest region in the
middle of retina represents the convergence of axon bundles to form the optic nerve. Due to the
dominating fluorescence of condensed axons, individual RGCs could not be resolved and data was
not collected from the middle region. Compared with WT retina, the RD retina has fewer number
of RGCs with GCaMP expression and exhibit a more non-uniform distribution. The baseline
fluorescent intensity within RGC soma is usually dimmer than WT, especially for those located at
the relatively dark region (Figure 3.4 Bottom). The decrease fluorescence might be attributed to
deterioration of RGC layer of RD model. Experiment also showed that the only RGC at lighter
region can response to external electrical stimulation. Since the expression window of the viral
vector has been optimized (Chapter 2), for most of RGCs across animals, fluorescence remained
predominantly localized to cell cytoplasms without major overexpression and cytomorbidity issue.
92
93
Figure 3.4: Retinal whole mounts of a wild-type (WT) (Top) and RD (Bottom) adult mice
transduced with AAV2-CAG-GCaMP6f (3 weeks postinjection). The retina was mounted on a 6 ×
10 rectangular multielectrode array with a homemade retaining ring and imaged with an inverted
fluorescence microscope. The baseline fluorescent image shows that GCaMP6f indicators are well
expressed throughout the ganglion cell layer (GCL) for WT and most regions (with lighter but
frosted texture background) for RD. The mosaic was created by stitching together 101 and 78 ×10
images respectively. The dark circles are 200-mdiameter indium tin oxide (ITO) electrodes.
3.3.2 Calcium Transient
To maximize the number of observable neuron, through our experiment, we selected regions
of interests by choosing a location of the active electrode where expression is greater in terms of
94
number of GCaMP-expressing cells, since the expression was highly non-uniform across the retina.
This same sampling strategy was also used for in other retinal calcium imaging experiment
(Chapter 2). The averaged baseline and post-stimulation 2-D calcium images are shown in Figure
3.5. The responding neurons as well as responsive area can be highlighted by the normalized of
fluorescent difference ( ΔF/ F) in the 2-D imaging mapping. Some RGCs that appear very bright in
the baseline are non-responsive (no change in fluorescence detected) so are eliminated by the
normalization process.
Normalized changes in fluorescence in response to stimulation for two randomly selected RGC
expressing GCaMP6f are shown in Figure 3.6. The initiation of each stimulus is indicated by the
red arrow. The result clearly show that increasing stimulus amplitude leads to larger calcium
transient, which corresponds to stronger RGC activation. In addition, the chosen 20 secs resting
interval between stimuli is sufficient for the calcium fluorescence to return back to baseline for next
stimulus, which is important for capturing time-invariant results and preventing electroporation of
cells.
Figure 3.5: Electrical stimulation activates RGCs as revealed through large changes in GCaMP6f
fluorescence intensity in well expressed region in vitro. (A) Before Stimulation, cells are at baseline
fluorescence (F). (B) Original calcium imaging in response to external stimuli from the ITO
electrode in the middle. (C) Image normalized of fluorescent difference ( ΔF / F) highlights the
responding cells in the field of view. The yellow arrows indicate that some RGC with very bright
baseline fluorescence did not generate detectable calcium transients. Background subtraction
eliminated the non-responding cells (as shown in (C)).
95
Figure 3.6: Normalized changes in fluorescence ( ΔF/ F) for two RGCs expressing GCaMP6f in
response to the 120 Hz 40 µs duration biphasic symmetric pulse train stimuli with different
amplitude of stimuli, from subthreshold to suprathreshold. Each stimulus sustained for 5 secs with
20 secs inter-stimulus interval for calcium level returning back to normal. The red arrows indicate
the onset of each stimulation pulse train.
3.3.3 Duration Manipulation
Figure 3.7 shows a set of representative spatial threshold maps for pulse widths ranging from
0.04 to 4 ms (40 µs, 80 µs, 120 µs, 0.5 ms, 1 ms, 4 ms) with symmetric cathodic-first waveform at
120 Hz (minimum temporal resolution for STG-4008 MCS system is 20 µs), for WT and RD retinas
respectively. As mentioned in Section 3.2.8, the responses were classified as somatic or axonal
activations based on if the soma showing fluorescence change was inside the light blue contour
(location of electrode) or outside the green contour (two times the radius of electrode) respectively.
The corresponding threshold for each pixel (corresponding current amplitude) is illustrated with
color map, where lower and higher thresholds are represented with more reddish and whitish colors,
as shown in the scale color bar. Results demonstrate that all pulses ≤ 150 µs duration activate fewer
axon bundles and result in more focused response near the working electrode if proper stimulation
current amplitude is selected (around 20-30 µA for the region shown in Figure 3.7 Top). On the
contrary, pulses ≥ 0.5 ms activate axons in greater numbers, causing more non-selective elongated
96
responses extending away from the optic disc at threshold. The stripe shaped black shadow inside
the activation patterns was resulted from the blood vessel occlusion on top of retinal ganglion cell
layer. Similar finding for varying pulse durations have been also shown using RD model, where
the number of responsive cells within the ROI is much less due to the progressive deterioration of
retina.
The strength duration curves (as shown in Figure 3.8) for somatic and axonal responses from
multiple ROIs (20 regions, 5 WT animals and 15 regions, 5 RD animals) were also fitted using
decaying exponential model for all pulse durations of symmetric cathodic-first pulses in Table 3.1.
𝑦 = 𝑎𝑒
− 𝑡 𝜏 ⁄
+ 𝑏
where parameters of fits were optimized to minimize the sum of squared error. For both WT and
RD retinas, we can observe that the largest percentage difference of average soma and axon
thresholds occurred when short duration pulse (≤ 150 µs, especially 40 µs) was used, indicating
that the selective activation of soma and axon can be more easily achieved due to larger
manipulation ranges of current amplitude which is important for retinal prosthesis owing to limited
resolution scale of current amplitude settings. These differences were statistically significant ( 𝑃 <
0.05, two-sample t-tests). Figure 3.9 shows the average thresholds of somatic and axonal activation
for WT and RD retinas. Results demonstrate that the RGC somas of RD retina have slightly lower
threshold on average for all pulses ≤ 150 µs duration ( 62.56 µA(WT) and 62.27 µA(RD) for 40
µs, 26.07 µA(WT) and 25.21 µA(RD) for 80 µs, 16.64 µA(WT) and 16.24 µA(RD) for 120 µs),
but higher threshold on average for all pulse ≥ 0.5ms (4.39 µA(WT) and 4.471 µA(RD) for 0.5 ms,
3.106 µA(WT) and 3.164 µA(RD) for 1 ms, 1.539 µA(WT) and 1.582 µA(RD) for 4 ms).
97
Figure 3.7: Spatial threshold maps for symmetric cathodic-first pulses with different duration in
WT (Top) and RD (Bottom) retina. The color bar shows thresholds in terms of current amplitude
(µA). The blue and green contours represent the location of electrode and two times the radius of
electrode respectively, defining the somatic activation region (inside the blue circle) axonal
activation region (outside the green circle).
98
Figure 3.8: The strength duration curve plotted in terms of current amplitude for somatic and
axonal activations respectively, from 20 regions of 5 WT mice (Top) and 15 regions of 5 RD mice
(Bottom). The number listed near each error bar shows the percentage difference between mean
soma and axon thresholds. Results indicated that stimulation with short duration pulse has greatest
threshold difference, implying largest stimulation strength manipulation ranges for somatic versus
axonal activations. ** means 𝑃 < 0.01, and * means 𝑃 < 0.05.
99
Figure 3.9: Threshold comparison for somatic and axonal activations with different pulse durations
of WT and RD retina.
3.3.4 Phase Manipulation
Responses of RGCs generated by representative short (40 µs) and long duration (0.5 ms)
symmetric cathodic-first pulses versus anodic-first pulses (reverse) with identical current amplitude
are demonstrated in Figure 3.10. The spatial patterns of threshold maps of both WT and RD retinas
demonstrate the symmetric anodic-first produces more localized RGC responses compared with
cathodic-pulse, especially with 40 µs pulses, though the required current amplitudes are universally
higher. The increased thresholds of most RGCs is highly possibly related to the sharp
hyperpolarization caused by high amplitude anodic phase before depolarization, thus conditioning
the membrane to be more inactive. While less pronounced, a similar phenomenon can also be
observed for pulses with 0.5ms duration.
Comparison of responsive area statistical analysis from multiple ROIs for different paradigms
using the mean soma and axon threshold for stimulation shows the same tendency, indicating the
anodic-first pulse with extremely short duration can effectively avoid axonal activation and confine
100
the response to more nearby electrode location (Figure 3.11). A possible explanation for this finding
is that the sharp anodic phase can effectively inhibit the following cathodic-driven depolarization
of axonal bundles across the electrode, but do not have a strong inhibitory effect on RGC somas.
101
Figure 3.10: Spatial threshold maps for symmetric cathodic-first and anodic-first pulses with
representative short duration (40 µs) and long duration (4 ms) in WT (Top) and RD (Bottom) retina.
Identical setting as Figure 3.7.
102
Figure 3.11: The responsive area statistical analysis for symmetric cathodic-first and anodic-first
pulses with representative short duration (40 µs) and long duration (4 ms) from WT and RD mice
(23 regions of 6 WT animals and 15 regions of 5 RD animals). The strength of stimulation is
selected using the mean soma and axon threshold for stimulation respectively. The blue and yellow
bar demonstrate the percentage of somatic and axonal activated region, which clearly show that the
103
symmetric anodic-first pulse with short duration can significantly confine the response area nearby
electrode, thus activating RGC somatically and potentially forming more ideal visual percept.
3.3.5 Waveform Manipulation
Figure 3.12 shows the threshold maps of WT (Top) and RD retina (Bottom) for asymmetric
anodic-first pulse with different ratios, symmetric cathodic-first, and anodic-first pulse for
comparison. Results highly support that the threshold of RGCs can be manipulated by asymmetric
waveform selection, especially with 20 or 10 times ratio. Through extending the duration of anodic
phase with sub-threshold amplitude that slightly hyperpolarizes the membrane potential of RGCs,
the RGC cells become conditioned to be more sensitive to stimulation, thus amplifying the
depolarizing effect of the following cathodic pulse. However, when the ratio was reduced, as
expected, the generated response patterns are close to the symmetric anodic-first pulse, which
produces more localized but higher-threshold maps. In addition, the threshold maps suggest that
semi-localized response can be selectively activated if proper stimulation current amplitude is used
(10-20 µA for both representative WT and RD retinas). Correlations between response patterns
driven by asymmetric anodic-first pulsed with different delivered charges and control symmetric
cathodic-first pulses reach maximum value at 50% and 66.6% of charge for WT and RD animals
respectively, suggesting that the charge consumption can be significantly reduced (save about 87%
and 77% power consumption calculated by Ohm’s law: ∫ 𝐼 2
𝑅 𝑑𝑢 𝑟 𝑎𝑡 𝑖 𝑜 𝑛 ) to generate similar pattern
as targeted symmetric cathodic-first pulses do.
104
Figure 3.12: Spatial threshold maps for asymmetric anodic-first pulse (consistent with cathodic
phase duration 40 µs) with different ratios and two controls (symmetric cathodic-first and anodic-
first pulses), for representative WT (Top) and RD (Bottom) retina respectively. Data suggest that
the thresholds for RGCs can be significantly reduced through asymmetric anodic-first pulse with
phase ratio greater than 10 times. If the current amplitude is precisely control, the local response
nearby electrode can still be achieved with the special design waveform. On the contrary, when
105
using phase ratio lower than 5 times, the responses become more localized but the thresholds of
RGCs are universally increase. This finding is consistent with symmetric anodic-first (reverse)
pulse since the short duration and high amplitude anodic phase hyperpolarize the membrane sharply,
instead of making it hypersensitive.
106
Figure 3.13: Correlations of response patterns between 20-times ratio asymmetric anodic-first
pulse with different delivered charges and control symmetric cathodic-first pulses with fixed
delivered charge (60 µA) that generally leads to localized response (Top: WT; Bottom: RD). The
averaged correlation plot across multiple regions identifies the required amount of charge for 20-
times ratio asymmetric anodic-first pulse to produce similar response as control symmetric
cathodic-first pulse (23 regions of 6 WT animals and 15 regions of 5 RD animals).
107
3.4 Discussion and Conclusion
We have presented a calcium imaging technique accompanied with transparent MEA platform
that provides the unique ability to visualize spatial patterns of RGC activation to electrical
stimulation in real time. The responses of RGC support the possibility to manipulate the activation
of RGCs through varying the stimulation durations, phases, and waveforms. The proposed
waveforms have great potential to improve the retinal prosthetic implants through forming more
ideal shape perception from different perspective.
3.4.1 Short Duration Pulse for Direct RGC Activation
Various stimulation strategies have been used to improve the RGCs response and selective
activation of RGCs and BPs is possible by manipulation of stimulus duration. Studies have shown
that more localized responses of RGCs can be achieved using indirect stimulation with long
duration pulses (Weitz et al., 2015). However, clinical testing demonstrate that thresholds gradually
rose during the test session and some patients reported fading of brightness of percepts when using
ms-scale pulse, implying that the RGCs were becoming desensitized (Zrenner et al., 2011). More
evidence from animal studies reveals that indirect stimulation of RGCs originated from BP
stimulation desensitize quickly (Suzuki et al., 2004, Freeman and Fried, 2011, Jensen et al., 2009).
The mechanisms behind this phenomenon still remain unknown, however electrophysiology
studies suggest multiple factors, including amacrine cell inhibition (Freeman and Fried, 2011),
depletion of synaptic vesicles (Jensen and Rizzo, 2007), and BP calcium channels inactivation (Hu
et al., 2009). Short duration pulse overcomes this complicated issue by primarily targeting RGCs
directly which have the capability to follow high rates of electrical stimulation (Cai et al., 2011).
Our results further show that the RGC somas can be selectively activated with extremely short
duration pulse (≤150 µs) with proper current amplitude, thus reducing the axonal activation which
108
corresponds to the elongated non-ideal percept reported patient. However, the range of stimulus for
selective activation of soma vs. axons is small (~25% of threshold).
3.4.2 Symmetric Anodic-first Pulse for Localized RGC responses
Symmetric biphasic cathodic-first pulse is the most commonly used stimulation waveforms for
neural implants for charge balance and efficiency. The cathodic and anodic phase delivered in the
specific order are designed to depolarize and repolarize the neuron respectively. In this project, we
found that the symmetric anodic-first pulse has potential to create more localized response nearby
working electrode, thus suggesting better selection of somatic activation. As a trade-off, the overall
RGCs thresholds for symmetric anodic-first are increased, as shown in recent research using MEA
recording on RGC responses in retinal degeneration (rd1) mice (Ahn et al., 2015). However, this
stimulation pattern is still worth further exploration in clinical study if the increased thresholds of
RGCs still remain within charge safety limit. Based on our findings, this phase alternation method
has the greatest potential to form localized RGC activation which implies high spatial resolution of
retinal prosthesis.
3.4.3 Asymmetric Anodic-first Pulse for RGCs Thresholds
Manipulation
Anode break excitation was first reported in the middle of last century to excite frog leg and
squid axon model with hyperpolarization current (Frankenhaeuser and Widen, 1956, Guttman and
Hachmeister, 1972). When the small hyperpolarizing current was consistently applied, the potential
across the neuron membrane first falls which is followed by a conditioned reduction of threshold
required for action potential. With the termination of the hyperpolarizing current, the membrane
109
potential rapidly increases. This findings agree with the Hodgkin-Huxley (HH) computational
model (Hodgkin and Huxley, 1952). The hyperpolarization increases the probability of Na channels
in “resting state” (i.e. able to be activated). Removal of hyperpolarization quickly increases the
number of sodium channels in “open state” and allows enough positive charges to enter the neuron
to offset the efflux of positive charges from potassium channels (which have slower dynamics).
The difference in Na and K dynamics leads to further depolarization and finally an action potential
with sufficient depolarization.
The design of asymmetric anodic-first pulse is inspired by anode break excitation. Instead of
using 2
nd
sharp anodic phase to repolarize the neuron and retain charge balance after cathodic phase
depolarization, the 1
st
anodic phase with extended duration and sub-threshold amplitude can
effectively hyperpolarize and condition the membrane slightly to be easily excitable (relative to
equilibrium), the following cathodic phase can generate an AP with less energy. We also find that
the RGC responses can be properly controlled to be localized nearby the active electrode. This
result validates that the selective activation of RGC somas can still may be possible with the special
stimulation profile design. With proper adjustment and modification of the proposed pulse
waveform, the same stimulation strategy also has great potential in terms of threshold manipulation
for other neural stimulation application, such as spinal cord (Miller et al., 2016) or peripheral nerves
(Araujo et al., 2015).
3.4.4 Limitations of Calcium Imaging
Virus vector is considered to be ideal for calcium indicator delivery for retina application owing
to the ease of delivery and production (Weitz et al., 2013). However, in practice, GCaMP-induced
cytomorbidity, suggested to emerge from overexpression and interaction of the sensor CaM and/or
M13 motifs with endogenous proteins (Hasan et al., 2004), limits the time span of the calcium
110
imaging experiment that can be performed. With the validation of a customized in vivo fundus
imaging system, we found that the optimal expression window for proposed virus-transduced
calcium indicator is around 3~4 weeks post injection without exhibiting abnormal cellular
physiology which is in agreement with the other studies (Borghuis et al., 2011, Chen et al., 2013c).
The precise control of viral vector expression constrain the degree of cytomorbidity and the same
in vivo fundus imaging characterization is suggested before proceeding with experiments of other
mammalian species or mouse strain.
The other main drawback to calcium imaging in studies of WT retina is that the excitation light
causes bleaching of photoreceptors, further limiting the applicability for studies involving light
stimulation due to lack of ability of phototransduction. To minimize photobleaching at
photoreceptors while imaging calcium dynamics at inner retinal neurons, some groups have used
two-photon excitation with infrared wavelength to selectively excite fluorescence in inner retinal
neurons (Stosiek et al., 2003). For our studies, however, which rely on extracellular electrical
stimulation (the method of stimulation with epiretinal prosthetic implants, such as the Argus II) of
retinal neurons, even strong excitation light does not inhibit RGC or bipolar neuron responsiveness.
3.4.5 Conclusion
In summary, we demonstrate an advanced imaging techniques that can detects the RGCs
response in large scale with cellular level resolution. The activation pattern is a good reference for
producing perceptions with retinal prostheses, since RGCs response directly corresponds to the
visual perceptions. Experimental results showed the focal response can be achieved with relative
short duration (<150 µs) pulse, and can be improved by changing the phase order. The RGCs
threshold can further be significantly reduced through the special design of asymmetric anodic-first
111
pulse. Our findings provide strong evidence to support development of high-resolution and low-
power consumption retinal prostheses.
112
CHAPTER IV: Current Steering to Alter RGC
Responses
Abstract
Retinal prosthetic implants are one of the most promising treatment for partial vision
restoration in retinal degenerative diseases. While increasing the number and density of electrodes
seems like a good approach to improve spatial resolution, electrodes of a certain size are required
for safety considerations. Current steering, as an effective method to alter the electrical field around
active electrode, has been proposed as an alternative approach to improve spatial resolution
perception for retinal prosthesis. In this project, based on the established calcium imaging in
Chapter 3, we tested current steering stimulation through using adjacent local active and return
electrodes on a transparent indium tin oxide microelectrode array (MEA). The results suggest that
the RGCs response evoked by relative longer duration pulse can be successfully altered by current
steering, but not as effectively for short duration pulse. These results also provide further evidence
that the sources of RGCs response for long and short duration pulses might be different.
113
4.1 Introduction
Retinitis pigmentosa (RP) (Humayun and de Juan, 1998) and dry age-related macular
degeneration (AMD) (Stanga et al., 2016) are two leading diseases that causes blindness. The
prevalence of both diseases together result in approximately 200 million people with visual
impairment worldwide around 2020 (Wong et al., 2014, Hartong et al., 2006). There is no cure for
either AMD or RP and the current therapies, such as nutritional supplements (Richer et al., 2004),
anti-VEGF (vascular endothelial growth factor) and laser (Heier et al., 2012, Fine et al., 2000),
mostly aim to slow down the cell death and the concomitant vision loss. Novel treatments such as
new drugs, gene therapy and optogenetic-based treatment are still under investigation.(Beltran et
al., 2012, Ku et al., 2016) Therefore, retinal prosthetic implants designed to partially restore visual
function by stimulating the remaining neurons have been regarded as one of the most promising
solutions.
Several retinal prostheses have been developed and two systems, including Argus II implant
(Second Sight Medical Produces, Inc.) and Alpha IMS implant (Retina Implant AG), have
regulatory approval, with the best reported 20/1260 and 20/546 visual grating acuity respectively
(Humayun et al., 2012, Stingl et al., 2015a). With the present design, the patients implanted with
those devices can regain orientation and mobility to a certain degree, and in some patients, the
ability to recognize simple patterns such as letters.(Humayun et al., 2012, Zrenner et al., 2011,
Stingl et al., 2013b) Although the results from clinical studies seemed to be assuring, more
sophisticated visual tasks such as sentence reading and human face recognition require higher
spatial resolution of visual perception.
Visual acuity is used to measure the spatial resolution of a visual system. For people with
normal visual acuity (20/20 vision), two points separated by 1 arcmin, equivalent to 4.5 µm of
retina, can be resolved (Yue et al., 2016). Based on this information, the electrode pitch (the
114
distance between the centers of two adjacent electrodes) and size will determine the visual acuity
possible with retinal implants (Palanker et al., 2005). Other factors, such as proximity of the
electrode array to the retina and retinal and cortical processing will also contribute to the perception
experienced by retinal implant users. However, the electrode size cannot be selected arbitrarily
small since the amount of safe charge capacity is reduced with electrode area (Weiland and
Humayun, 2014). In addition, the ideal electrode-electrolyte interface of implanted electrode array
that can perfectly fit the curvature of the eye is hard to be formed due the difficulty of fabrication.
Current steering has been applied in modern neuroprosthetics. The position of active and return
electrodes will influence the shape of the electric field around electrode. In neural stimulation, it
has been used to either redirect the current to excite different areas or focus the current, thus
narrowing the effective area (Bonham and Litvak, 2008). For cochlear implant recipients, this
stimulation strategy has been found to create additional pitch percepts (Firszt et al., 2007). In central
nerve system such as brain and spinal cord, balancing current flow through adjacent electrodes can
alter the activated groups of neurons (Butson and McIntyre, 2008, Hegarty, 2012). Simulation using
finite-element retinal model in response to current steering has shown that the different selections
of stimulating and return electrodes can effectively alter the charge density distribution (Savage et
al., 2012).
In this project, we performed calcium imaging using virus-transduced GECI GCaMP6f to
record the neural activity from RGCs at single cell resolution in wholemount retinas while applying
current steering stimulation using adjacent local active and return electrodes on a transparent
indium tin oxide microelectrode array (MEA). The results suggest that the activation pattern of
RGCs can be altered by current steering when using longer duration pulses (4 ms), but not short
duration pulses (40 µs).
115
4.2 Methods
4.2.1 Overview
All experiment were conducted in adult mice (C57BL6/J) receiving an intravitreal injection of
an AAV vector encoding a GECI (AAV2-CAG-GCaMP6f). Based on the established in vitro
calcium imaging and electrophysiological mouse animal model shown in Chapter 3, we altered the
return electrode for both long and short duration stimulation from multiple region of interests (ROIs)
with different axon orientation. All procedures were approved by the Institutional Animal Care and
Use Committee (IACUC) and the Institutional Biosafety Committee (IBC) at the University of
Southern California and University of Michigan, Ann Arbor.
4.2.2 Current Steering
Several combination of active and return electrodes were used to perform the current steering
experiments. For control remote electrode, a platinum wire coated with platinum iridium encircled
on the top middle of the incubation chamber to simulate the placement of Argus II return electrode
(on top of sclera, other side of retina). This return electrode was mounted around 5 mm from the
active electrode to serve as a remote return (10 times greater than the pitch between electrodes).
Next to the stimulating electrode, one of 4 (or 3 on the edge of array) local adjacent electrodes were
used as a local return electrode to test the current steering method. Figure 4.1 shows the relative
positions of the active, remote and local return electrode. The transparent MEA array is composed
of 200 µm diameter disk electrodes, in a 6×10 pattern with 500 µm electrode pitch. These
dimensions are within the range of present day epiretinal prostheses.
116
Figure 4.1: Current steering experimental setup. The retina was whole-mounted with ganglion cell
layer side faced down in transparent MEA chamber. The remote return electrode was positioned
above the chamber on top of the photoreceptor layer side. For local return electrode, one of the 4
(or 3 on the edge of array) adjunct electrodes was used to serve as a local ground. (Drawing not to
scale)
Electrical stimuli are composed of a series of symmetric biphasic square current pulses with 40
µs or 4 ms pulse durations at 120 Hz to ensure charge balance. For each duration, stimulation
protocols were designed as Chapter 3 that current amplitude progressively increase from
subthreshold to suprathreshold. A total of 10 amplitudes were used, each lasting 5 secs to generate
a detectable calcium transient with 20 secs interval in between to bring calcium transient back to
baseline. No stimuli were delivered during the first and last 5 secs of each protocol. For each region
of interest (ROI), a repeated remote return electrode control experiment has been performed at the
end of testing to make sure the response pattern is nearly time invariant. Current stimuli were
generated from a computer-controlled stimulus generator (STG-4008 – 1.6mA, Multi Channel
Systems, Reutlingen, Germany) and fed through a custom capacitive isolation circuitry for
preventing constant leaking current. A customize interface circuit board shown in the Figure 4.2
was used to relay the stimulation signal to the designated electrode and switch between the remote
and corresponding local return electrodes.
117
Figure 4.2: Interface board for current steering experiment. The 60 channels correspond to the 60
electrodes on 6 × 10 pattern.
4.2.3 Data Analysis
The same imaging processing procedure for producing spatial threshold mapping in Chapter 3
has also been used here. Based on the geometry of activated neurons with respect to the working
electrode, the evoked RGCs can further be directly associated with somatic activation and axonal
activation respectively, depending on the relative distance to the center of working electrode. We
define the responses occurring directly above the working electrode as somatic activation region
(forming ideal visual percept), whereas those occurring outside two times the radius of electrode
as axonal activation region (forming elongated shape percept).
For current steering setups, statistical analysis for activated neurons within the two defined
regions has been applied to explore the effective region. The responsive areas for various current
steering setups under the different thresholds were computed using geometry information of spatial
mapping on pixel-by-pixel basis (Summation of activated neurons on pixel basis). All post-
118
processing steps and analysis of calcium imaging were analyzed by MATLAB (The Mathworks,
Natick, MA).
119
4.3 Results
Responses of RGCs generated by representative short (40 µs) and long duration (4 ms)
symmetric cathodic-first pulses with different return electrode setups are demonstrated in Figure
4.3 and 4.4. The left part of the figure shows the corresponding RGC axon bundles on top of the
active electrode, where the convergent point represent the position of optic disk. The relative
location of the return electrode versus the active electrode are illustrated on top of each spatial
threshold mapping. For this two examples, since the selected electrodes are located on the edge of
the MEA, the current steering methods have been only tested on three local return electrodes. A
control remote electrode experiment were done followed by all the current steering methods to
eliminate the factors of fluorescence photobleaching or desensitization resulted from continuous
stimulation.
Results show that the responses elicited by short duration pulse (40 µs), especially those nearby
the active electrode, were not strongly influenced by current steering. The change of return
electrodes just slightly altered the response pattern in terms of reduction of response cell and
increase of overall threshold. Compared with short duration pulse, there were significant attenuated
response patterns for long duration pulse (4 ms) when using different current steering methods.
With the identical stimulation current range, only a few RGCs can generate detectable calcium
transients, with the increase of thresholds.
120
Figure 4.3: Spatial threshold maps for different return electrode setups with 40 µs (Top) and 4 ms
(Bottom) symmetric cathodic-first pulses (Animal 2, region 5). The relative position between the
corresponding RGC axon bundles on top of the active electrode as well as the optic disc are shown
on the left side of each set of maps. The color bar shows thresholds in terms of current amplitude
(µA). The blue and green contours represent the location of electrode and two times the radius of
electrode respectively, defining the somatic activation region (inside the blue circle) axonal
activation region (outside the green circle).
121
Figure 4.4: Spatial threshold maps for different return electrode setups with 40 µs (Top) and 4 ms
(Bottom) symmetric cathodic-first pulses (Animal 2, region 4). The relative position between the
corresponding RGC axon bundles on top of the active electrode as well as the optic disc are shown
on the left side of each set of maps. The color bar shows thresholds in terms of current amplitude
(µA). Identical setting as Figure 4.4.
122
To do the statistical analysis, all ROIs have first been re-oriented by their axon bundles so that
results can be summed up and compared, as shown in Figure 4.5. In the 5 ROIs, all spatial threshold
maps are adjusted so that the axon bundles share similar distribution. The axons passing the active
electrode retain horizontally distributed and converge at right hand side at the optic disc. Along the
horizontal line, the electrodes located at downstream and upstream of axon bundles are named as
“Central (closer to the optic disc)” and “Peripheral (further from the optic disc)” return respectively.
The other two local returns are seated 90º with respect to the horizontal line, named as “Up” and
“Down” based on the relative position with the active electrode. Since only 2 out of 5 regions
contain the data with “Peripheral” electrode (upstream of axon bundle), we did not analyze this
current steering setup statistically owing to insufficient number of neurons ( due to 3 out of 5 active
electrode selected from the edge of array).
Comparison of responsive area from multiple ROIs (2 animals, 5 regions) for different current
steering electrodes using the maximum current amplitude (100 µA for 40 µs and 2 µA for 4 ms)
for stimulation shows the same tendency in terms of responsive neurons (Figure 4.6). With the long
duration pulses, the evoked responses for both somatic and axonal activation dramatically decrease
in terms of pixel. While less pronounced, a similar phenomenon can also be observed for pulses
with 0.5 ms duration. Figure 4.7 further demonstrate the percentage of responsive area for different
current steering electrodes using the mean of soma and axon thresholds for stimulation. Results
show that the percentages of somatic versus axonal activation for short duration pulses were not
altered greatly by switching from remote to local returns. However, for long duration pulses, there
are dramatic increase of somatic activation when using “Up” and “Down” local return electrodes
which forms relatively perpendicular electrical field with respect to the orientation of axon bundles.
On the contrary, the return electrode, “Central”, along with the axons bundles produce similar
activation ratio as the control remote electrode.
123
Figure 4.5: Standard orientation for responsive area analysis. The axon bundles are oriented so that
they extend horizontally across the active electrode and converge at right hand side at the optic disc.
The electrode located at downstream and upstream of the axons are named as “Central” and
“Peripheral”, whereas the other two adjacent are seated vertically, named as “Up” and “Down”
based on the relative position with the active electrode
124
Figure 4.6: The responsive area statistical analysis for different return electrode setups with 40 µs
(Top) and 4 ms (Bottom) symmetric cathodic-first pulses when applying maximum current
injection amplitude (100 µA for 40 µs and 2 µA for 4 ms) (5 regions of 2 WT animals). The green
and yellow bars demonstrate the somatic and axonal activated region in terms of pixels for different
current steering setups.
125
Figure 4.7: The responsive area statistical analysis for different return electrode setups with 40 µs
(Top) and 4 ms (Bottom) symmetric cathodic-first pulses (5 regions of 2 WT animals). The strength
of stimulation is selected using the mean soma and axon threshold for stimulation respectively. The
blue and yellow bar demonstrate the percentage of somatic and axonal activated region, which
clearly show that the current steering with “Up” and “Down” electrodes for long duration pulse (4
ms) can significantly confine the response area nearby electrode.
4.4 Discussion and Conclusion
4.4.1 Current Steering to Create Virtual Electrode
Current steering methods have been used to control the neuron activities through altering the
distribution of electrical field. In deep brain stimulation (DBS) realm, study has shown that the
balancing current flow through adjacent cathodes increased the volume of tissue activated (VTA)
magnitude, relative to monopolar stimulation, and allowing the possibility to sculpt the shape of
the VTA to fit a given anatomical target (Butson and McIntyre, 2008). Study of cochlear implant
also demonstrated that current steering involves the simultaneous delivery of different weighted
current on adjacent electrodes can alter the stimulation sites between the contacts and generated
varying pitch (Firszt et al., 2007). In suprachoroidal retinal implant, current steering has been used
126
to alter the cortical activation maps in response to the retinal stimulation in cat model (Dumm et
al., 2014). For those examples, since the new generated responses were similar to the centroid shifts
seen when using physical electrode, it has been named as “virtual electrode” to describe the idea.
Distance between the neuron and tissue might be the main factor that our study did not show
similar effect as virtual electrode which shifts the response to targeted region. For DBS in
Parkinson’s disease, there are currently four possible target sites in the brain, including internal
segment of the globus pallidus (GPi), the subthalamic nucleus (STN), the pedunculopontine
nucleus (PPN), and a subdivision of the thalamus referred to as ventro-intermediate nucleus (Vim),
and usually the size of targeted region is in mm scale (Johnson et al., 2008). A computational model
of cochlear implant showed that the electrode-neuron distances are normally set within the range
of 0.5 - 1.3mm, depending on the position of the selected electrode on the array (Goldwyn et al.,
2010). Suprachoroidal prostheses have electrode array implanted between choroid and sclera.
When compared with the epiretinal and subretinal prostheses, suprachoroidal implants are
relatively distant from the retina (around 400 µm) (Barriga-Rivera et al., 2017) and result in
elevated perceptual threshold and worsened spatial resolution (Yamauchi et al., 2005). In our
calcium imaging experiment, since the vitreous of retina was mostly remove for forming better
electrode-tissue contact, the actual distance between electrode and RGCs were generally less
(around 20-50 µm) than the real Argus II retinal implant, which has the mean electrode-retina
distance around 180 µm (Ahuja et al., 2013). Under this circumstances, electrode field did not
change significantly nearby the active electrode, so the RGCs response on top of the electrode was
not shifted a lot (Figure 4.8).
4.4.2 Current Steering to Change Electrical field
Although the “virtual electrode” phenomenon has not been observed with current steering in
our study, the RGCs response attenuated significantly in terms of effective area and thresholds,
especially for long duration pulse. Studies have showed that RGCs are more easily excited by short
127
duration pulse (<150 µs) whereas BPs tend to respond preferentially to longer pulse width.
(Margalit and Thoreson, 2006, Fried et al., 2006a, Freeman et al., 2010, Weitz et al., 2015).
Therefore, based on those finding, we postulate that the electrical field formed by current steering,
especially close to BP layer, possibly has been altered significantly.
Electrical field simulation for different current steering setups using homogeneous retinal
model are shown in Figure 4.8 (Schmid et al., 2014). The MEA contains 50 × 50 µm electrodes
with 70 pitch in quadratic arrangement. Comparisons of the density of electrical field lines and
equipotential lines nearby the electrode (z < 50 µm) do not demonstrate obvious distribution change
between remote return (Figure 4.8A) with dipole electric (Figure 4.8B). However, with the increase
of the vertical distance, especially when is z > 50 µm, electrical field lines actually distributed more
sparsely with current steering (dipole) than remote return. This simulation explains why current
steering effects are only noted in long duration stimulation. Since long duration pulses mainly target
BPs, which are relatively far from the electrode, the current steering method significantly changes
the electrical field, thus causing the attenuation of RGCs response that is mostly evoked through
indirect stimulation. In practice, our calcium imaging setting is probably more close to the model
shown in Figure 4.8C, since we applied a thin layer of Ame’s buffer before mounting the retina to
prevent the adhesion between tissue and MEA array. In this case, the equipotential lines distribution
is even much closer to the electrode for dipole model which leads to greater electrical field
differences with the increase of vertical distance from electrode.
128
129
Figure 4.8: Electrical field simulation with different return electrode usage and condition.
Electrical field line (or current lines) are shown as full (red) lines, equipotential lines as broken
(blue) lines, which having a spacing of 0.1 V. (A) Only the center electrode of the array is activated,
with return electrode at infinite. (B) Two adjacent electrodes forms an electric dipole with tissue
touching the chip. The electric current injected is equal to the one injected in (A). (C) There is a 20
µm layer of saline with assumed tenfold conductivity between the chip and the tissue. (Adapted
from (Schmid et al., 2014))
The change of electrical field for long duration pulses also influence the percentage of RGC
somatic versus axonal response. When the electrical field generated by dipole electric is
perpendicular to the RGC nerve fiber (Up, Down electrode as return in our case), the evoked
response became more localized compared with parallel situation. As electrical field is parallel to
the axons, a larger difference of voltage gradient along fiber bundles will be produced, thus
facilitating axonal activation. Similar finding in terms of fiber excitation thresholds for bipolar
stimulating electrodes has also been identified using retinal modeling (Grumet et al., 2000), where
the stimulation along fiber result in lower threshold.
4.4.3 Conclusion
In summary, we demonstrate that current steering method can alter the RGCs response
compared with using remote return electrode. Instead of leading to shift of centroid of response site,
known as “virtual electrode”, the altered electrical field resulted from current steering significantly
attenuates the RGCs response in terms of effective area and threshold, especially for long duration
pulses. This results imply the sources of RGCs response caused by long and short duration pulses
might be different.
130
CHAPTER V: Conclusions and Future Work
In this project, I proposed a new fundus imaging system that can evaluate in vivo fluorescence
reporter expression, so the optimal expression time window for the updated virus-transduced
calcium indicators, AAV2-CAG-GCaMP6f, can be determined. To perform calcium imaging
recoding with external stimulation to retina, I developed and established an up-down microscope
platform with customized transparent MEAs array which allow us to visualize RGC neurons
activity in response to different stimulation paradigms at cellular resolution in real time. Through
the in vitro imaging system, I have demonstrated several stimulation strategies with great potential
to improve retinal prostheses. Through leveraging the pulse width, phase, and waveform, I was able
to evoke focal responses with reduction of RGCs threshold, thus supporting the possibility for
development of high-resolution and low-power consumption retinal prostheses. My work also
showed that the current steering method can be used to alter the electrical field of stimulation and
provide further evidence that the sources of RGCs response depends on pulse duration.
5.1 Recommendations for Epiretinal Prostheses
An ideal retinal implant with high spatial resolution should form small round visual percept
nearby active electrode, through selective evoke the neuron somas. However, many clinical studies
for Argus II epiretinal prosthesis reported most evoked percepts by single electrodes were elongated
and aligned with estimated axon path in the retina, suggesting the activation of axon bundles
(Nanduri et al., 2012, Nanduri et al., 2011). Long-duration stimuli can be used to selectively target
BPs, thus indirectly activate RGCs and avoiding axonal stimulation (Weitz et al., 2015). However,
the disadvantages of long-duration pulses are that they limit the maximal stimulation rate and
require more charge (Geddes, 2004). In addition, depending on the stimulation frequency,
131
sensitivity of the electrically-evoked RGCs has shown progressively decreasing trend with the
repeated indirect stimulation in multiple animal models (Freeman and Fried, 2011, Jensen and
Rizzo, 2007).
Direct response of RGCs evoked by short pulses can follow at high frequencies with less charge.
The high frequency feature makes it possible to encode of visual information in terms of modulation
of RGCs spike rate. Moreover, since one of the challenging issue for current epiretinal prosthesis
is to generate sensible percepts while remaining within electrochemical safety limits, the less
charge feature make the short duration pulse option more appropriate until more flexible electrode
substrates that can form close coupling between retina and MEA to be developed.
Through spatial threshold mapping of responses evoked by different durations of stimuli, I
found that short duration pulses (≤ 120 µs) can generate focal activation pattern of RGCs. The
phase manipulation by making anodic phase ahead can produce more localized RGC responses
nearby the active electrode, with overall increase of RGC thresholds as trade-off. The RGCs
threshold can further be significantly reduced through the special design of asymmetric anodic-first
pulse with proper duration ratio (> 10 times) between two phases. The findings support the
possibility to manipulate the responses of RGCs through varying the stimulation parameters, thus
potentially forming more ideal shape perception with higher spatial resolution but low power
consumption in future retinal prosthesis design, though it might need to be further validated in
clinical study. Future retinal prostheses should have the capability to use flexible, asymmetric
pulses.
132
5.2 Future Work
5.2.1 In vivo Calcium Imaging
Although in vitro calcium imaging with wholemount retina is a useful tool to investigate the
activation of RGCs in response to electrical stimulation, the model is still not complete enough to
simulate the real retinal implant, especially for its chronic feature. Compared with nearly flat
wholemount retina, the actual retina seated inside the eyecup make it difficult to attach the MEA
perfectly to form close electrode-tissue interface, thus introducing the varying electrical field
distribution and thresholds for percepts when applying identical stimuli. Studies also demonstrated
that chronic stimulation might result in electrically induced retinal damage and contribute to the
change of neural activities (Colodetti et al., 2007, Cohen et al., 2011, Gonzalez-Calle and Weiland,
2016), although these studies used stimulation levels greater than what is typically used for retinal
stimulation in patients. The actual location of retinal implant might rotate and reposition after the
device being installed (Seider and Hahn, 2015). None of the above issues can be addressed by in
vitro calcium imaging method considering the long-lasting characteristic of retinal implant.
In vivo calcium imaging has been realized through two photon laser scanning microscopy and
allowed for deep tissue imaging in the brain (Lutcke et al., 2010, Grewe et al., 2010). A number of
novel fiber-optic-based two-photon micro endoscopes have been recently developed (Helmchen et
al., 2001; Gobel et al., 2004), which makes the technique capable of both recording physiological
activity in vivo and also investigating pathophysiology of disease states. Recently, in vivo two-
photon imaging system of the mouse retina has been proposed which enables functional calcium
imaging of repeated retinal responses to light stimulation using the genetically encoded indicator,
GCaMP6s (Bar-Noam et al., 2016). It would be interesting to use the system to image the patterns
of RGCs excited by electrical stimuli, and the effect of chronic implant to the neural activities.
133
5.2.2 Light Stimulation VS. Electrical Stimulation
In this project, GECIs have demonstrated its ability for measuring the RGCs neural activities
in response to external electrical stimulation. Another possibility to use this tool for detecting RGCs
response while applying light stimulation (Borghuis et al., 2011). I have shown that the localized
RGCs response corresponding to ideal percept can be achieved through proper selection of
stimulation parameters in WT retina; however, whether the response is identical or similar to the
actual percept evoked by normal light stimulation still remain unknown. Comparison between
RGCs activation patterns elicited by electrical and light stimuli with same power will be interesting
and possibly provide more in-depth information about the how retina transform and process light
into neural signals.
The calcium imaging system in this project excite GCaMP with blue light bandwidth which
causes bleaching of photoreceptors, further limiting the applicability for studies involving light
stimulation. This can be improved by several optical methods that have been already well
established. The light sheet fluorescence microscopy which allows only a thin slice (usually a few
hundred nanometers to a few micrometers) of the sample illuminated perpendicularly to the
direction of observation (Truong et al., 2011) can selectively excite RGC layer without bleaching
of the photoreceptors (Icha et al., 2016). In addition, as mentioned in previous section, infrared
two-photon excitation of RGCs can effectively prevent photoreceptors from exposure to visible
light, thus remaining the excitability (Stosiek et al., 2003).
134
CHAPTER VI: References
AHN, K. N., AHN, J. Y., KIM, J. H., CHO, K., KOO, K. I., SENOK, S. S. & GOO, Y. S. 2015. Effect of
Stimulus Waveform of Biphasic Current Pulse on Retinal Ganglion Cell Responses in
Retinal Degeneration (rd1) mice. Korean J Physiol Pharmacol, 19, 167-75.
AHNELT, P. & KOLB, H. 1994. Horizontal cells and cone photoreceptors in human retina: a Golgi-
electron microscopic study of spectral connectivity. J Comp Neurol, 343, 406-27.
AHUJA, A. K., BEHREND, M. R., KURODA, M., HUMAYUN, M. S. & WEILAND, J. D. 2008. An in vitro
model of a retinal prosthesis. IEEE Trans Biomed Eng, 55, 1744-53.
AHUJA, A. K., DORN, J. D., CASPI, A., MCMAHON, M. J., DAGNELIE, G., DACRUZ, L., STANGA, P.,
HUMAYUN, M. S., GREENBERG, R. J. & ARGUS, I. I. S. G. 2011. Blind subjects implanted
with the Argus II retinal prosthesis are able to improve performance in a spatial-motor
task. Br J Ophthalmol, 95, 539-43.
AHUJA, A. K., YEOH, J., DORN, J. D., CASPI, A., WUYYURU, V., MCMAHON, M. J., HUMAYUN, M.
S., GREENBERG, R. J. & DACRUZ, L. 2013. Factors Affecting Perceptual Threshold in Argus
II Retinal Prosthesis Subjects. Transl Vis Sci Technol, 2, 1.
AKERBOOM, J., CHEN, T. W., WARDILL, T. J., TIAN, L., MARVIN, J. S., MUTLU, S., CALDERON, N. C.,
ESPOSTI, F., BORGHUIS, B. G., SUN, X. R., GORDUS, A., ORGER, M. B., PORTUGUES, R.,
ENGERT, F., MACKLIN, J. J., FILOSA, A., AGGARWAL, A., KERR, R. A., TAKAGI, R., KRACUN,
S., SHIGETOMI, E., KHAKH, B. S., BAIER, H., LAGNADO, L., WANG, S. S., BARGMANN, C. I.,
KIMMEL, B. E., JAYARAMAN, V., SVOBODA, K., KIM, D. S., SCHREITER, E. R. & LOOGER, L.
L. 2012. Optimization of a GCaMP calcium indicator for neural activity imaging. J
Neurosci, 32, 13819-40.
AMERI, H., RATANAPAKORN, T., UFER, S., ECKHARDT, H., HUMAYUN, M. S. & WEILAND, J. D.
2009. Toward a wide-field retinal prosthesis. J Neural Eng, 6, 035002.
ARAUJO, T., CANDEIAS, R., NUNES, N. & GAMBOA, H. 2015. Evaluation of Motor Neuron
Excitability by CMAP Scanning with Electric Modulated Current. Neurosci J, 2015,
360648.
AYTON, L. N., BLAMEY, P. J., GUYMER, R. H., LUU, C. D., NAYAGAM, D. A., SINCLAIR, N. C.,
SHIVDASANI, M. N., YEOH, J., MCCOMBE, M. F., BRIGGS, R. J., OPIE, N. L., VILLALOBOS,
J., DIMITROV, P. N., VARSAMIDIS, M., PETOE, M. A., MCCARTHY, C. D., WALKER, J. G.,
BARNES, N., BURKITT, A. N., WILLIAMS, C. E., SHEPHERD, R. K., ALLEN, P. J. & BIONIC
VISION AUSTRALIA RESEARCH, C. 2014a. First-in-human trial of a novel suprachoroidal
retinal prosthesis. PLoS One, 9, e115239.
AYTON, L. N., BLAMEY, P. J., GUYMER, R. H., LUU, C. D., NAYAGAM, D. A., SINCLAIR, N. C.,
SHIVDASANI, M. N., YEOH, J., MCCOMBE, M. F., BRIGGS, R. J., OPIE, N. L., VILLALOBOS,
J., DIMITROV, P. N., VARSAMIDIS, M., PETOE, M. A., MCCARTHY, C. D., WALKER, J. G.,
BARNES, N., BURKITT, A. N., WILLIAMS, C. E., SHEPHERD, R. K., ALLEN, P. J. &
CONSORTIUM, B. V. A. R. 2014b. First-in-human trial of a novel suprachoroidal retinal
prosthesis. PLoS One, 9, e115239.
BADEN, T., BERENS, P., BETHGE, M. & EULER, T. 2013. Spikes in mammalian bipolar cells support
temporal layering of the inner retina. Curr Biol, 23, 48-52.
BAIRD, G. S., ZACHARIAS, D. A. & TSIEN, R. Y. 1999. Circular permutation and receptor insertion
within green fluorescent proteins. Proc Natl Acad Sci U S A, 96, 11241-6.
BALDRIDGE, W. H. 1996. Optical recordings of the effects of cholinergic ligands on neurons in
the ganglion cell layer of mammalian retina. J Neurosci, 16, 5060-72.
135
BAR-NOAM, A. S., FARAH, N. & SHOHAM, S. 2016. Correction-free remotely scanned two-
photon in vivo mouse retinal imaging. Light Sci Appl, 5, e16007.
BARRIGA-RIVERA, A., GUO, T., YANG, C. Y., ABED, A. A., DOKOS, S., LOVELL, N. H., MORLEY, J. W.
& SUANING, G. J. 2017. High-amplitude electrical stimulation can reduce elicited
neuronal activity in visual prosthesis. Sci Rep, 7, 42682.
BARTON, J. J. S. & BENATAR, M. 2003. An Introduction to Perimetry and the Normal Visual Field.
Field of Vision: A Manual and Atlas of Perimetry. Totowa, NJ: Humana Press.
BEHREND, M. R., AHUJA, A. K., HUMAYUN, M. S., WEILAND, J. D. & CHOW, R. H. 2009. Selective
labeling of retinal ganglion cells with calcium indicators by retrograde loading in vitro. J
Neurosci Methods, 179, 166-72.
BELTRAN, W. A., CIDECIYAN, A. V., LEWIN, A. S., IWABE, S., KHANNA, H., SUMAROKA, A.,
CHIODO, V. A., FAJARDO, D. S., ROMAN, A. J., DENG, W. T., SWIDER, M., ALEMAN, T. S.,
BOYE, S. L., GENINI, S., SWAROOP, A., HAUSWIRTH, W. W., JACOBSON, S. G. & AGUIRRE,
G. D. 2012. Gene therapy rescues photoreceptor blindness in dogs and paves the way
for treating human X-linked retinitis pigmentosa. Proc Natl Acad Sci U S A, 109, 2132-7.
BERNARDES, R., SERRANHO, P. & LOBO, C. 2011. Digital ocular fundus imaging: a review.
Ophthalmologica, 226, 161-81.
BLECKERT, A., PARKER, E. D., KANG, Y., PANCAROGLU, R., SOTO, F., LEWIS, R., CRAIG, A. M. &
WONG, R. O. 2013. Spatial relationships between GABAergic and glutamatergic synapses
on the dendrites of distinct types of mouse retinal ganglion cells across development.
PLoS One, 8, e69612.
BOMASH, I., ROUDI, Y. & NIRENBERG, S. 2013. A virtual retina for studying population coding.
PLoS One, 8, e53363.
BONHAM, B. H. & LITVAK, L. M. 2008. Current focusing and steering: modeling, physiology, and
psychophysics. Hear Res, 242, 141-53.
BORGHUIS, B. G., TIAN, L., XU, Y., NIKONOV, S. S., VARDI, N., ZEMELMAN, B. V. & LOOGER, L. L.
2011. Imaging light responses of targeted neuron populations in the rodent retina. J
Neurosci, 31, 2855-67.
BRIGGMAN, K. L. & EULER, T. 2011. Bulk electroporation and population calcium imaging in the
adult mammalian retina. J Neurophysiol, 105, 2601-9.
BUTSON, C. R. & MCINTYRE, C. C. 2008. Current steering to control the volume of tissue
activated during deep brain stimulation. Brain Stimul, 1, 7-15.
BUTTERWICK, A., HUIE, P., JONES, B. W., MARC, R. E., MARMOR, M. & PALANKER, D. 2009.
Effect of shape and coating of a subretinal prosthesis on its integration with the retina.
Exp Eye Res, 88, 22-9.
CAI, C., REN, Q., DESAI, N. J., RIZZO, J. F., 3RD & FRIED, S. I. 2011. Response variability to high
rates of electric stimulation in retinal ganglion cells. J Neurophysiol, 106, 153-62.
CHADER, G. J., WEILAND, J. & HUMAYUN, M. S. 2009. Artificial vision: needs, functioning, and
testing of a retinal electronic prosthesis. Prog Brain Res, 175, 317-32.
CHANG, Y.-C., WALSTON, S. T., CHOW, R. H. & WEILAND, J. D. 2016. In vivo characterization of
genetic expression of virus-transduced calcium indicators in retinal ganglion cells using a
low-cost funduscope. 38th Annual International Conference of the IEEE Engineering in
Medicine and Biology Society. Orlando, FL.
CHANG, Y.-C., WALSTON, S. T., CHOW, R. H. & WEILAND, J. D. 2017. GCaMP expression in retinal
ganglion cells characterized using a low-cost fundus imaging system. Journal of Neural
Engineering, 14, 056018.
136
CHEN, K., LO, Y. K., YANG, Z., WEILAND, J. D., HUMAYUN, M. S. & LIU, W. 2013a. A system
verification platform for high-density epiretinal prostheses. IEEE Trans Biomed Circuits
Syst, 7, 326-37.
CHEN, K., YANG, Z., HOANG, L., WEILAND, J., HUMAYUN, M. & LIU, W. 2010. An Integrated 256-
Channel Epiretinal Prosthesis. IEEE Journal of Solid-State Circuits, 45, 1946-1956.
CHEN, T.-W., WARDILL, T. J., SUN, Y., PULVER, S. R., RENNINGER, S. L., BAOHAN, A., SCHREITER,
E. R., KERR, R. A., ORGER, M. B., JAYARAMAN, V., LOOGER, L. L., SVOBODA, K. & KIM, D.
S. 2013b. Ultrasensitive fluorescent proteins for imaging neuronal activity. Nature, 499,
295-300.
CHEN, T. W., WARDILL, T. J., SUN, Y., PULVER, S. R., RENNINGER, S. L., BAOHAN, A., SCHREITER,
E. R., KERR, R. A., ORGER, M. B., JAYARAMAN, V., LOOGER, L. L., SVOBODA, K. & KIM, D.
S. 2013c. Ultrasensitive fluorescent proteins for imaging neuronal activity. Nature, 499,
295-300.
CHOW, A. Y., BITTNER, A. K. & PARDUE, M. T. 2010. The artificial silicon retina in retinitis
pigmentosa patients (an American Ophthalmological Association thesis). Trans Am
Ophthalmol Soc, 108, 120-54.
CHOW, A. Y., CHOW, V. Y., PACKO, K. H., POLLACK, J. S., PEYMAN, G. A. & SCHUCHARD, R. 2004.
The artificial silicon retina microchip for the treatment of visionloss from retinitis
pigmentosa. Arch Ophthalmol, 122, 460-469.
COBBOLD, P. H. & RINK, T. J. 1987. Fluorescence and bioluminescence measurement of
cytoplasmic free calcium. Biochem J, 248, 313-28.
COGAN, S. F., EDELL, D. J., GUZELIAN, A. A., PING LIU, Y. & EDELL, R. 2003. Plasma-enhanced
chemical vapor deposited silicon carbide as an implantable dielectric coating. J Biomed
Mater Res A, 67, 856-67.
COHEN, E., AGRAWAL, A., CONNORS, M., HANSEN, B., CHARKHKAR, H. & PFEFER, J. 2011. Optical
coherence tomography imaging of retinal damage in real time under a stimulus
electrode. J Neural Eng, 8, 056017.
COLODETTI, L., WEILAND, J. D., COLODETTI, S., RAY, A., SEILER, M. J., HINTON, D. R. &
HUMAYUN, M. S. 2007. Pathology of damaging electrical stimulation in the retina. Exp
Eye Res, 85, 23-33.
CURCIO, C. A., SLOAN, K. R., KALINA, R. E. & HENDRICKSON, A. E. 1990. Human photoreceptor
topography. J Comp Neurol, 292, 497-523.
DACEY, D. M. 1999. Primate retina: cell types, circuits and color opponency. Prog Retin Eye Res,
18, 737-63.
DACRUZ, L., COLEY, B., CHRISTOPHER, P., MERLINI, F., WUYYURU, V., SAHEL, J. A., STANGA, P.,
FILLEY, E., DAGNELIE, G. & GROUP, A. I. I. S. 2010. Patients Blinded by Outer Retinal
Dystrophies Are Able to Identify Letters Using the Argus TM II Retinal Prosthesis System.
Investigative Ophthalmology & Visual Science, 51, 2023-2023.
DAIGER, S. P., BOWNE, S. J., SULLIVAN, L. S., BLANTON, S. H., WEINSTOCK, G. M., KOBOLDT, D.
C., FULTON, R. S., LARSEN, D., HUMPHRIES, P., HUMPHRIES, M. M., PIERCE, E. A., CHEN,
R. & LI, Y. 2014. Application of next-generation sequencing to identify genes and
mutations causing autosomal dominant retinitis pigmentosa (adRP). Adv Exp Med Biol,
801, 123-9.
DALKARA, D., BYRNE, L. C., LEE, T., HOFFMANN, N. V., SCHAFFER, D. V. & FLANNERY, J. G. 2012.
Enhanced gene delivery to the neonatal retina through systemic administration of
tyrosine-mutated AAV9. Gene Ther, 19, 176-81.
DAYA, S. & BERNS, K. I. 2008. Gene therapy using adeno-associated virus vectors. Clin Microbiol
Rev, 21, 583-93.
137
DE BALTHASAR, C., PATEL, S., ROY, A., FREDA, R., GREENWALD, S., HORSAGER, A.,
MAHADEVAPPA, M., YANAI, D., MCMAHON, M. J., HUMAYUN, M. S., GREENBERG, R. J.,
WEILAND, J. D. & FINE, I. 2008. Factors affecting perceptual thresholds in epiretinal
prostheses. Invest Ophthalmol Vis Sci, 49, 2303-14.
DELORI, F. C., WEBB, R. H., SLINEY, D. H. & AMERICAN NATIONAL STANDARDS, I. 2007.
Maximum permissible exposures for ocular safety (ANSI 2000), with emphasis on
ophthalmic devices. J Opt Soc Am A Opt Image Sci Vis, 24, 1250-65.
DEMB, J. B. & SINGER, J. H. 2015. Functional Circuitry of the Retina. Annual Review of Vision
Science, 1, 263-289.
DEN HOLLANDER, A. I., TEN BRINK, J. B., DE KOK, Y. J., VAN SOEST, S., VAN DEN BORN, L. I., VAN
DRIEL, M. A., VAN DE POL, D. J., PAYNE, A. M., BHATTACHARYA, S. S., KELLNER, U.,
HOYNG, C. B., WESTERVELD, A., BRUNNER, H. G., BLEEKER-WAGEMAKERS, E. M.,
DEUTMAN, A. F., HECKENLIVELY, J. R., CREMERS, F. P. & BERGEN, A. A. 1999. Mutations
in a human homologue of Drosophila crumbs cause retinitis pigmentosa (RP12). Nat
Genet, 23, 217-21.
DEZAWA, M., TAKANO, M., NEGISHI, H., MO, X., OSHITARI, T. & SAWADA, H. 2002. Gene transfer
into retinal ganglion cells by in vivo electroporation: a new approach. Micron, 33, 1-6.
DJILAS, M., OLES, C., LORACH, H., BENDALI, A., DEGARDIN, J., DUBUS, E., LISSORGUES-BAZIN, G.,
ROUSSEAU, L., BENOSMAN, R., IENG, S. H., JOUCLA, S., YVERT, B., BERGONZO, P., SAHEL,
J. & PICAUD, S. 2011. Three-dimensional electrode arrays for retinal prostheses:
modeling, geometry optimization and experimental validation. J Neural Eng, 8, 046020.
DOMBECK, D. A., HARVEY, C. D., TIAN, L., LOOGER, L. L. & TANK, D. W. 2010. Functional imaging
of hippocampal place cells at cellular resolution during virtual navigation. Nat Neurosci,
13, 1433-40.
DOMMEL, N. B., WONG, Y. T., LEHMANN, T., DODDS, C. W., LOVELL, N. H. & SUANING, G. J.
2009. A CMOS retinal neurostimulator capable of focussed, simultaneous stimulation. J
Neural Eng, 6, 035006.
DREOSTI, E., ESPOSTI, F., BADEN, T. & LAGNADO, L. 2011. In vivo evidence that retinal bipolar
cells generate spikes modulated by light. Nat Neurosci, 14, 951-2.
DUMM, G., FALLON, J. B., WILLIAMS, C. E. & SHIVDASANI, M. N. 2014. Virtual electrodes by
current steering in retinal prostheses. Invest Ophthalmol Vis Sci, 55, 8077-85.
ECKMILLER, R., NEUMANN, D. & BARUTH, O. 2005. Tunable retina encoders for retina implants:
why and how. J Neural Eng, 2, S91-S104.
EULER, T., HAVERKAMP, S., SCHUBERT, T. & BADEN, T. 2014. Retinal bipolar cells: elementary
building blocks of vision. Nat Rev Neurosci, 15, 507-19.
FENG, G., MELLOR, R. H., BERNSTEIN, M., KELLER-PECK, C., NGUYEN, Q. T., WALLACE, M.,
NERBONNE, J. M., LICHTMAN, J. W. & SANES, J. R. 2000. Imaging neuronal subsets in
transgenic mice expressing multiple spectral variants of GFP. Neuron, 28, 41-51.
FINE , S. L., BERGER , J. W., MAGUIRE , M. G. & HO , A. C. 2000. Age-Related Macular
Degeneration. New England Journal of Medicine, 342, 483-492.
FIRSZT, J. B., KOCH, D. B., DOWNING, M. & LITVAK, L. 2007. Current steering creates additional
pitch percepts in adult cochlear implant recipients. Otol Neurotol, 28, 629-36.
FRANKENHAEUSER, B. & WIDEN, L. 1956. Anode break excitation in desheathed frog nerve. J
Physiol, 131, 243-7.
FREEMAN, D. K., EDDINGTON, D. K., RIZZO, J. F., 3RD & FRIED, S. I. 2010. Selective activation of
neuronal targets with sinusoidal electric stimulation. J Neurophysiol, 104, 2778-91.
FREEMAN, D. K. & FRIED, S. I. 2011. Multiple components of ganglion cell desensitization in
response to prosthetic stimulation. J Neural Eng, 8, 016008.
138
FRIED, S. I., HSUEH, H. A. & WERBLIN, F. S. 2006a. A Method for Generating Precise Temporal
Patterns of Retinal Spiking Using Prosthetic Stimulation. Journal of Neurophysiology, 95,
970-978.
FRIED, S. I., HSUEH, H. A. & WERBLIN, F. S. 2006b. A method for generating precise temporal
patterns of retinal spiking using prosthetic stimulation. J Neurophysiol, 95, 970-8.
FRISHMAN, L. J., FREEMAN, A. W., TROY, J. B., SCHWEITZER-TONG, D. E. & ENROTH-CUGELL, C.
1987. Spatiotemporal frequency responses of cat retinal ganglion cells. J Gen Physiol, 89,
599-628.
FUJIKADO, T., KAMEI, M., SAKAGUCHI, H., KANDA, H., MORIMOTO, T., IKUNO, Y., NISHIDA, K.,
KISHIMA, H., MARUO, T., KONOMA, K., OZAWA, M. & NISHIDA, K. 2011. Testing of
semichronically implanted retinal prosthesis by suprachoroidal-transretinal stimulation
in patients with retinitis pigmentosa. Invest Ophthalmol Vis Sci, 52, 4726-33.
GARCIA-FRIGOLA, C., CARRERES, M. I., VEGAR, C. & HERRERA, E. 2007. Gene delivery into mouse
retinal ganglion cells by in utero electroporation. BMC Dev Biol, 7, 103.
GEDDES, L. A. 2004. Accuracy limitations of chronaxie values. IEEE Trans Biomed Eng, 51, 176-81.
GEHRS, K. M., JACKSON, J. R., BROWN, E. N., ALLIKMETS, R. & HAGEMAN, G. S. 2010.
Complement, age-related macular degeneration and a vision of the future. Arch
Ophthalmol, 128, 349-58.
GILL, E. C., ANTALEK, J., KIMOCK, F. M., NASIATKA, P. J., MCINTOSH, B. P., TANGUAY, A. R. J. &
WEILAND, J. D. 2013. High-Density Feedthrough Technology for Hermetic Biomedical
Micropackaging. MRS Online Proceedings Library Archive, 1572, mrss13-1572-ss05-08 (6
pages).
GOLDWYN, J. H., BIERER, S. M. & BIERER, J. A. 2010. Modeling the electrode-neuron interface of
cochlear implants: effects of neural survival, electrode placement, and the partial
tripolar configuration. Hear Res, 268, 93-104.
GONZALEZ-CALLE, A. & WEILAND, J. D. 2016. Evaluation of Effects of Electrical Stimulation in the
Retina with Optical Coherence Tomography. Conf Proc IEEE Eng Med Biol Soc, 2016,
6182-6185.
GRAY, D. C., MERIGAN, W., WOLFING, J. I., GEE, B. P., PORTER, J., DUBRA, A., TWIETMEYER, T. H.,
AHAMD, K., TUMBAR, R., REINHOLZ, F. & WILLIAMS, D. R. 2006. In vivo fluorescence
imaging of primate retinal ganglion cells and retinal pigment epithelial cells. Opt Express,
14, 7144-58.
GREENBERG, R. J., VELTE, T. J., HUMAYUN, M. S., SCARLATIS, G. N. & DE JUAN, E., JR. 1999. A
computational model of electrical stimulation of the retinal ganglion cell. IEEE Trans
Biomed Eng, 46, 505-14.
GREWE, B. F., LANGER, D., KASPER, H., KAMPA, B. M. & HELMCHEN, F. 2010. High-speed in vivo
calcium imaging reveals neuronal network activity with near-millisecond precision. Nat
Methods, 7, 399-405.
GRIESBECK, O., BAIRD, G. S., CAMPBELL, R. E., ZACHARIAS, D. A. & TSIEN, R. Y. 2001. Reducing
the environmental sensitivity of yellow fluorescent protein. Mechanism and
applications. J Biol Chem, 276, 29188-94.
GRIMM, D. & KAY, M. A. 2003. From virus evolution to vector revolution: use of naturally
occurring serotypes of adeno-associated virus (AAV) as novel vectors for human gene
therapy. Curr Gene Ther, 3, 281-304.
GROSS, M., BUSS, R., KOHLER, K., SCHAUB, J. & JAGER, D. Optical signal and energy transmission
for a retina implant. [Engineering in Medicine and Biology, 1999. 21st Annual
Conference and the 1999 Annual Fall Meetring of the Biomedical Engineering Society]
BMES/EMBS Conference, 1999. Proceedings of the First Joint, 1999 1999. 476 vol.1.
139
GRUMET, A. E., WYATT, J. L., JR. & RIZZO, J. F., 3RD 2000. Multi-electrode stimulation and
recording in the isolated retina. J Neurosci Methods, 101, 31-42.
GRYNKIEWICZ, G., POENIE, M. & TSIEN, R. Y. 1985. A new generation of Ca2+ indicators with
greatly improved fluorescence properties. J Biol Chem, 260, 3440-50.
GUTTMAN, R. & HACHMEISTER, L. 1972. Anode break excitation in space-clamped squid axons.
Biophys J, 12, 552-63.
HABIB, A. G., CAMERON, M. A., SUANING, G. J., LOVELL, N. H. & MORLEY, J. W. 2013. Spatially
restricted electrical activation of retinal ganglion cells in the rabbit retina by hexapolar
electrode return configuration. J Neural Eng, 10, 036013.
HADJINICOLAOU, A. E., MEFFIN, H., MATURANA, M. I., CLOHERTY, S. L. & IBBOTSON, M. R. 2015.
Prosthetic vision: devices, patient outcomes and retinal research. Clin Exp Optom, 98,
395-410.
HARTONG, D. T., BERSON, E. L. & DRYJA, T. P. 2006. Retinitis pigmentosa. Lancet, 368, 1795-809.
HARTVEIT, E. 1999. Reciprocal synaptic interactions between rod bipolar cells and amacrine cells
in the rat retina. J Neurophysiol, 81, 2923-36.
HARTWICK, A. T., LALONDE, M. R., BARNES, S. & BALDRIDGE, W. H. 2004. Adenosine A1-receptor
modulation of glutamate-induced calcium influx in rat retinal ganglion cells. Invest
Ophthalmol Vis Sci, 45, 3740-8.
HASAN, M. T., FRIEDRICH, R. W., EULER, T., LARKUM, M. E., GIESE, G., BOTH, M., DUEBEL, J.,
WATERS, J., BUJARD, H., GRIESBECK, O., TSIEN, R. Y., NAGAI, T., MIYAWAKI, A. & DENK,
W. 2004. Functional fluorescent Ca2+ indicator proteins in transgenic mice under TET
control. PLoS Biol, 2, e163.
HEGARTY, D. 2012. Spinal cord stimulation: the clinical application of new technology.
Anesthesiol Res Pract, 2012, 375691.
HEIER, J. S., BROWN, D. M., CHONG, V., KOROBELNIK, J. F., KAISER, P. K., NGUYEN, Q. D.,
KIRCHHOF, B., HO, A., OGURA, Y., YANCOPOULOS, G. D., STAHL, N., VITTI, R., BERLINER,
A. J., SOO, Y., ANDERESI, M., GROETZBACH, G., SOMMERAUER, B., SANDBRINK, R.,
SIMADER, C., SCHMIDT-ERFURTH, U., VIEW & GROUPS, V. S. 2012. Intravitreal
aflibercept (VEGF trap-eye) in wet age-related macular degeneration. Ophthalmology,
119, 2537-48.
HEILBRONN, R. & WEGER, S. 2010. Viral vectors for gene transfer: current status of gene
therapeutics. Handb Exp Pharmacol, 143-70.
HELLSTROM, M., RUITENBERG, M. J., POLLETT, M. A., EHLERT, E. M., TWISK, J., VERHAAGEN, J. &
HARVEY, A. R. 2009. Cellular tropism and transduction properties of seven adeno-
associated viral vector serotypes in adult retina after intravitreal injection. Gene Ther,
16, 521-32.
HO, A. C., HUMAYUN, M. S., DORN, J. D., DA CRUZ, L., DAGNELIE, G., HANDA, J., BARALE, P. O.,
SAHEL, J. A., STANGA, P. E., HAFEZI, F., SAFRAN, A. B., SALZMANN, J., SANTOS, A., BIRCH,
D., SPENCER, R., CIDECIYAN, A. V., DE JUAN, E., DUNCAN, J. L., ELIOTT, D., FAWZI, A.,
OLMOS DE KOO, L. C., BROWN, G. C., HALLER, J. A., REGILLO, C. D., DEL PRIORE, L. V.,
ARDITI, A., GERUSCHAT, D. R., GREENBERG, R. J. & GROUP, A. I. S. 2015. Long-Term
Results from an Epiretinal Prosthesis to Restore Sight to the Blind. Ophthalmology, 122,
1547-54.
HODGKIN, A. L. & HUXLEY, A. F. 1952. A quantitative description of membrane current and its
application to conduction and excitation in nerve. J Physiol, 117, 500-44.
HORNIG, R., ZEHNDER, T., VELIKAY-PAREL, M., LAUBE, T., FEUCHT, M. & RICHARD, G. 2007. The
IMI retinal implant system. Artificial Sight. Springer.
140
HORSAGER, A., GREENWALD, S. H., WEILAND, J. D., HUMAYUN, M. S., GREENBERG, R. J.,
MCMAHON, M. J., BOYNTON, G. M. & FINE, I. 2009. Predicting visual sensitivity in retinal
prosthesis patients. Invest Ophthalmol Vis Sci, 50, 1483-91.
HU, C., BI, A. & PAN, Z. H. 2009. Differential expression of three T-type calcium channels in
retinal bipolar cells in rats. Vis Neurosci, 26, 177-87.
HUMAYUN, M. S. & DE JUAN, E., JR. 1998. Artificial vision. Eye (Lond), 12 ( Pt 3b), 605-7.
HUMAYUN, M. S., DE JUAN, E., JR., WEILAND, J. D., DAGNELIE, G., KATONA, S., GREENBERG, R. &
SUZUKI, S. 1999. Pattern electrical stimulation of the human retina. Vision Res, 39, 2569-
76.
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. & ARGUS, I. I. S. G. 2012. Interim
results from the international trial of Second Sight's visual prosthesis. Ophthalmology,
119, 779-88.
ICHA, J., SCHMIED, C., SIDHAYE, J., TOMANCAK, P., PREIBISCH, S. & NORDEN, C. 2016. Using Light
Sheet Fluorescence Microscopy to Image Zebrafish Eye Development. Journal of
Visualized Experiments : JoVE, 53966.
IM, M. & FRIED, S. I. 2016. Temporal properties of network-mediated responses to repetitive
stimuli are dependent upon retinal ganglion cell type. J Neural Eng, 13, 025002.
INOUYE, S., NOGUCHI, M., SAKAKI, Y., TAKAGI, Y., MIYATA, T., IWANAGA, S., MIYATA, T. & TSUJI,
F. I. 1985. Cloning and sequence analysis of cDNA for the luminescent protein aequorin.
Proc Natl Acad Sci U S A, 82, 3154-8.
JENSEN, R. J. & RIZZO, J. F., 3RD 2006. Thresholds for activation of rabbit retinal ganglion cells
with a subretinal electrode. Exp Eye Res, 83, 367-73.
JENSEN, R. J. & RIZZO, J. F., 3RD 2007. Responses of ganglion cells to repetitive electrical
stimulation of the retina. J Neural Eng, 4, S1-6.
JENSEN, R. J., ZIV, O. R. & RIZZO, J. F. 2005. Responses of rabbit retinal ganglion cells to electrical
stimulation with an epiretinal electrode. J Neural Eng, 2, S16-21.
JENSEN, R. J., ZIV, O. R., RIZZO, J. F., 3RD, SCRIBNER, D. & JOHNSON, L. 2009. Spatiotemporal
aspects of pulsed electrical stimuli on the responses of rabbit retinal ganglion cells. Exp
Eye Res, 89, 972-9.
JEON, C. J., STRETTOI, E. & MASLAND, R. H. 1998. The major cell populations of the mouse
retina. J Neurosci, 18, 8936-46.
JEPSON, L. H., HOTTOWY, P., MATHIESON, K., GUNNING, D. E., DABROWSKI, W., LITKE, A. M. &
CHICHILNISKY, E. J. 2013. Focal electrical stimulation of major ganglion cell types in the
primate retina for the design of visual prostheses. J Neurosci, 33, 7194-205.
JIANG, G. & ZHOU, D. D. 2010. Technology Advances and Challenges in Hermetic Packaging for
Implantable Medical Devices. In: ZHOU, D. D. & GREENBAUM, E. (eds.) Implantable
Neural Prostheses 2. Springer New York.
JOHNSON, M. D., MIOCINOVIC, S., MCINTYRE, C. C. & VITEK, J. L. 2008. Mechanisms and targets
of deep brain stimulation in movement disorders. Neurotherapeutics, 5, 294-308.
KANDEL, E. 2013. Principles of Neural Science, Fifth Edition, McGraw-Hill Education.
KELLY, S. K., SHIRE, D. B., CHEN, J., DOYLE, P., GINGERICH, M. D., COGAN, S. F., DROHAN, W. A.,
BEHAN, S., THEOGARAJAN, L., WYATT, J. L. & RIZZO III, J. F. 2011. A hermetic wireless
subretinal neurostimulator for vision prostheses. IEEE Trans Biomed Eng, 58, 3197-205.
KELLY, S. K., SHIRE, D. B., CHEN, J., GINGERICH, M. D., COGAN, S. F., DROHAN, W., ELLERSICK, W.,
KRISHNAN, A., BEHAN, S., WYATT, J. L. & RIZZO, J. F. Developments on the Boston 256-
141
channel retinal implant. Multimedia and Expo Workshops (ICMEW), 2013 IEEE
International Conference on, 2013. IEEE, 1-6.
KESERU, M., FEUCHT, M., BORNFELD, N., LAUBE, T., WALTER, P., ROSSLER, G., VELIKAY-PAREL,
M., HORNIG, R. & RICHARD, G. 2012. Acute electrical stimulation of the human retina
with an epiretinal electrode array. Acta Ophthalmol, 90, e1-8.
KETTUNEN, P., DEMAS, J., LOHMANN, C., KASTHURI, N., GONG, Y., WONG, R. O. & GAN, W. B.
2002. Imaging calcium dynamics in the nervous system by means of ballistic delivery of
indicators. J Neurosci Methods, 119, 37-43.
KLAUKE, S., GOERTZ, M., REIN, S., HOEHL, D., THOMAS, U., ECKHORN, R., BREMMER, F. &
WACHTLER, T. 2011. Stimulation with a wireless intraocular epiretinal implant elicits
visual percepts in blind humans. Invest Ophthalmol Vis Sci, 52, 449-55.
KLEIN, R., CHOU, C. F., KLEIN, B. E., ZHANG, X., MEUER, S. M. & SAADDINE, J. B. 2011. Prevalence
of age-related macular degeneration in the US population. Arch Ophthalmol, 129, 75-80.
KLEIN, R. L., HAMBY, M. E., GONG, Y., HIRKO, A. C., WANG, S., HUGHES, J. A., KING, M. A. &
MEYER, E. M. 2002. Dose and promoter effects of adeno-associated viral vector for
green fluorescent protein expression in the rat brain. Exp Neurol, 176, 66-74.
KORONYO-HAMAOUI, M., KORONYO, Y., LJUBIMOV, A. V., MILLER, C. A., KO, M. K., BLACK, K. L.,
SCHWARTZ, M. & FARKAS, D. L. 2011. Identification of amyloid plaques in retinas from
Alzheimer's patients and noninvasive in vivo optical imaging of retinal plaques in a
mouse model. Neuroimage, 54 Suppl 1, S204-17.
KU, C. A., HARIPRASAD, S. M. & PENNESI, M. E. 2016. Gene Therapy Trial Update: A Primer for
Vitreoretinal Specialists. Ophthalmic Surg Lasers Imaging Retina, 47, 6-12.
KWON, I. & SCHAFFER, D. V. 2008. Designer gene delivery vectors: molecular engineering and
evolution of adeno-associated viral vectors for enhanced gene transfer. Pharm Res, 25,
489-99.
LEUNG, C. K., LINDSEY, J. D., CROWSTON, J. G., JU, W. K., LIU, Q., BARTSCH, D. U. & WEINREB, R.
N. 2008. In vivo imaging of murine retinal ganglion cells. J Neurosci Methods, 168, 475-8.
LI, C., HIRSCH, M., DIPRIMIO, N., ASOKAN, A., GOUDY, K., TISCH, R. & SAMULSKI, R. J. 2009.
Cytotoxic-T-lymphocyte-mediated elimination of target cells transduced with
engineered adeno-associated virus type 2 vector in vivo. J Virol, 83, 6817-24.
LOEB, J. E., CORDIER, W. S., HARRIS, M. E., WEITZMAN, M. D. & HOPE, T. J. 1999. Enhanced
expression of transgenes from adeno-associated virus vectors with the woodchuck
hepatitis virus posttranscriptional regulatory element: implications for gene therapy.
Hum Gene Ther, 10, 2295-305.
LOHMANN, C., MYHR, K. L. & WONG, R. O. 2002. Transmitter-evoked local calcium release
stabilizes developing dendrites. Nature, 418, 177-81.
LOOGER, L. L. & GRIESBECK, O. 2012. Genetically encoded neural activity indicators. Curr Opin
Neurobiol, 22, 18-23.
LORACH, H., BENOSMAN, R., MARRE, O., IENG, S. H., SAHEL, J. A. & PICAUD, S. 2012. Artificial
retina: the multichannel processing of the mammalian retina achieved with a
neuromorphic asynchronous light acquisition device. J Neural Eng, 9, 066004.
LUO, D. & SALTZMAN, W. M. 2000. Synthetic DNA delivery systems. Nat Biotechnol, 18, 33-7.
LUTCKE, H., MURAYAMA, M., HAHN, T., MARGOLIS, D. J., ASTORI, S., ZUM ALTEN BORGLOH, S.
M., GOBEL, W., YANG, Y., TANG, W., KUGLER, S., SPRENGEL, R., NAGAI, T., MIYAWAKI,
A., LARKUM, M. E., HELMCHEN, F. & HASAN, M. T. 2010. Optical recording of neuronal
activity with a genetically-encoded calcium indicator in anesthetized and freely moving
mice. Front Neural Circuits, 4, 9.
142
MAHADEVAPPA, M., WEILAND, J. D., YANAI, D., FINE, I., GREENBERG, R. J. & HUMAYUN, M. S.
2005. Perceptual thresholds and electrode impedance in three retinal prosthesis
subjects. IEEE Trans Neural Syst Rehabil Eng, 13, 201-6.
MANK, M. & GRIESBECK, O. 2008. Genetically encoded calcium indicators. Chem Rev, 108, 1550-
64.
MANK, M., SANTOS, A. F., DIRENBERGER, S., MRSIC-FLOGEL, T. D., HOFER, S. B., STEIN, V.,
HENDEL, T., REIFF, D. F., LEVELT, C., BORST, A., BONHOEFFER, T., HUBENER, M. &
GRIESBECK, O. 2008. A genetically encoded calcium indicator for chronic in vivo two-
photon imaging. Nat Methods, 5, 805-11.
MARGALIT, E. & THORESON, W. B. 2006. Inner retinal mechanisms engaged by retinal electrical
stimulation. Invest Ophthalmol Vis Sci, 47, 2606-12.
MASLAND, R. H. 2012. The tasks of amacrine cells. Vis Neurosci, 29, 3-9.
MATHIESON, K., LOUDIN, J., GOETZ, G., HUIE, P., WANG, L., KAMINS, T. I., GALAMBOS, L., SMITH,
R., HARRIS, J. S., SHER, A. & PALANKER, D. 2012. Photovoltaic Retinal Prosthesis with
High Pixel Density. Nat Photonics, 6, 391-397.
MCCARTHY, C., FENG, D. & BARNES, N. Augmenting intensity to enhance scene structure in
prosthetic vision. Multimedia and Expo Workshops (ICMEW), 2013 IEEE International
Conference on, 15-19 July 2013 2013. 1-6.
MEISTER, M. & BERRY, M. J., 2ND 1999. The neural code of the retina. Neuron, 22, 435-50.
MENZEL-SEVERING, J., LAUBE, T., BROCKMANN, C., BORNFELD, N., MOKWA, W., MAZINANI, B.,
WALTER, P. & ROESSLER, G. 2012. Implantation and explantation of an active epiretinal
visual prosthesis: 2-year follow-up data from the EPIRET3 prospective clinical trial. Eye
(Lond), 26, 501-9.
MILLER, J. P., ELDABE, S., BUCHSER, E., JOHANEK, L. M., GUAN, Y. & LINDEROTH, B. 2016.
Parameters of Spinal Cord Stimulation and Their Role in Electrical Charge Delivery: A
Review. Neuromodulation, 19, 373-84.
MINTA, A., KAO, J. P. & TSIEN, R. Y. 1989. Fluorescent indicators for cytosolic calcium based on
rhodamine and fluorescein chromophores. J Biol Chem, 264, 8171-8.
MIYAWAKI, A., LLOPIS, J., HEIM, R., MCCAFFERY, J. M., ADAMS, J. A., IKURA, M. & TSIEN, R. Y.
1997. Fluorescent indicators for Ca2+ based on green fluorescent proteins and
calmodulin. Nature, 388, 882-7.
MONGE, M., RAJ, M., NAZARI, M. H., CHANG, H. C., ZHAO, Y., WEILAND, J. D., HUMAYUN, M. S.,
TAI, Y. C. & EMAMI, A. 2013. A fully intraocular high-density self-calibrating epiretinal
prosthesis. IEEE Trans Biomed Circuits Syst, 7, 747-60.
MORGAN, J. I. & PUGH, E. N., JR. 2013. Scanning laser ophthalmoscope measurement of local
fundus reflectance and autofluorescence changes arising from rhodopsin bleaching and
regeneration. Invest Ophthalmol Vis Sci, 54, 2048-59.
MORIMOTO, T., KAMEI, M., NISHIDA, K., SAKAGUCHI, H., KANDA, H., IKUNO, Y., KISHIMA, H.,
MARUO, T., KONOMA, K., OZAWA, M., NISHIDA, K. & FUJIKADO, T. 2011. Chronic
implantation of newly developed suprachoroidal-transretinal stimulation prosthesis in
dogs. Invest Ophthalmol Vis Sci, 52, 6785-92.
NAKAI, J., OHKURA, M. & IMOTO, K. 2001. A high signal-to-noise Ca(2+) probe composed of a
single green fluorescent protein. Nat Biotechnol, 19, 137-41.
NANDURI, D., FINE, I., GREENBERG, R. J., HORSAGER, A., BOYNTON, G. M. & WEILAND, J. D.
Predicting the percepts of electrical stimulation in retinal prosthesis subjects.
Computational and Systems Neuroscience Meeting, 2011.
143
NANDURI, D., FINE, I., HORSAGER, A., BOYNTON, G. M., HUMAYUN, M. S., GREENBERG, R. J. &
WEILAND, J. D. 2012. Frequency and amplitude modulation have different effects on the
percepts elicited by retinal stimulation. Invest Ophthalmol Vis Sci, 53, 205-14.
NASIATKA, P. J., MCINTOSH, B. P., STILES, N. R. B., HAUER, M. C., J.D. WEILAND, HUMAYUN, M.
S. & A.R. TANGUAY, J. An Intraocular Camera for Provision of Natural Foveation in
Retinal Prostheses. Neural Interfaces Conference, 2010 Long Beach, CA.
NIWA, H., YAMAMURA, K. & MIYAZAKI, J. 1991. Efficient selection for high-expression
transfectants with a novel eukaryotic vector. Gene, 108, 193-9.
OLMEDO-PAYÁ, A., MARTÍ NEZ-ÁLVAREZ, A., CUENCA-ASENSI, S., FERRÁNDEZ, J. M. &
FERNÁNDEZ, E. 2015. Modeling the role of fixational eye movements in real-world
scenes. Neurocomputing, 151, Part 1, 78-84.
OPIE, N. L., BURKITT, A. N., MEFFIN, H. & GRAYDEN, D. B. 2012. Heating of the eye by a retinal
prosthesis: modeling, cadaver and in vivo study. IEEE Trans Biomed Eng, 59, 339-45.
ORTMANNS, M., UNGER, N., ROCKE, A., GEHRKE, M. & TIETDKE, H. J. A 0.1mm/sup 2/, Digitally
Programmable Nerve Stimulation Pad Cell with High-Voltage Capability for a Retinal
Implant. 2006 IEEE International Solid State Circuits Conference - Digest of Technical
Papers, 6-9 Feb. 2006 2006. 89-98.
PALANKER, D., VANKOV, A., HUIE, P. & BACCUS, S. 2005. Design of a high-resolution
optoelectronic retinal prosthesis. J Neural Eng, 2, S105-20.
PAQUES, M., GUYOMARD, J. L., SIMONUTTI, M., ROUX, M. J., PICAUD, S., LEGARGASSON, J. F. &
SAHEL, J. A. 2007. Panretinal, high-resolution color photography of the mouse fundus.
Invest Ophthalmol Vis Sci, 48, 2769-74.
PAREDES, R. M., ETZLER, J. C., WATTS, L. T., ZHENG, W. & LECHLEITER, J. D. 2008. Chemical
calcium indicators. Methods, 46, 143-51.
PARVER, L. M., AUKER, C. R. & CARPENTER, D. O. 1983. Choroidal blood flow. III. Reflexive
control in human eyes. Arch Ophthalmol, 101, 1604-6.
PATIL, S. D., RHODES, D. G. & BURGESS, D. J. 2005. DNA-based therapeutics and DNA delivery
systems: a comprehensive review. AAPS J, 7, E61-77.
PATTERSON, G. H. & PISTON, D. W. 2000. Photobleaching in two-photon excitation microscopy.
Biophys J, 78, 2159-62.
PEACHEY, N. S. & CHOW, A. Y. 1999. Subretinal implantation of semiconductor-based
photodiodes: progress and challenges. J Rehabil Res Dev, 36, 371-6.
PERRY, V. H. 1981. Evidence for an amacrine cell system in the ganglion cell layer of the rat
retina. Neuroscience, 6, 931-44.
PETRS-SILVA, H., DINCULESCU, A., LI, Q., DENG, W. T., PANG, J. J., MIN, S. H., CHIODO, V.,
NEELEY, A. W., GOVINDASAMY, L., BENNETT, A., AGBANDJE-MCKENNA, M., ZHONG, L.,
LI, B., JAYANDHARAN, G. R., SRIVASTAVA, A., LEWIN, A. S. & HAUSWIRTH, W. W. 2011.
Novel properties of tyrosine-mutant AAV2 vectors in the mouse retina. Mol Ther, 19,
293-301.
PRADEEP, V., MEDIONI, G. & WEILAND, J. Visual loop closing using multi-resolution SIFT grids in
metric-topological SLAM. Computer Vision and Pattern Recognition, 2009. CVPR 2009.
IEEE Conference on, 20-25 June 2009 2009. 1438-1445.
PUGH, E. N., JR. & LAMB, T. D. 1993. Amplification and kinetics of the activation steps in
phototransduction. Biochim Biophys Acta, 1141, 111-49.
RICHARD, G., HORNIG, R., KESERÜ , M. & FEUCHT, M. 2007. Chronic epiretinal chip implant in
blind patients with retinitis pigmentosa: long-term clinical results. Invest Ophthalmol Vis
Sci, 48, 666-666.
144
RICHARD, G., KESERUE, M., HORNIG, R., POST, N. & ZEITZ, O. 2009. Long-Term Stability of
Stimulation Thresholds Obtained From a Human Patient With a Prototype of an
Epiretinal Retina Prosthesis. Investigative Ophthalmology & Visual Science, 50, 4580-
4580.
RICHER, S., STILES, W., STATKUTE, L., PULIDO, J., FRANKOWSKI, J., RUDY, D., PEI, K., TSIPURSKY,
M. & NYLAND, J. 2004. Double-masked, placebo-controlled, randomized trial of lutein
and antioxidant supplementation in the intervention of atrophic age-related macular
degeneration: the Veterans LAST study (Lutein Antioxidant Supplementation Trial).
Optometry, 75, 216-30.
RIEKE, F. & BAYLOR, D. A. 1998. Origin of reproducibility in the responses of retinal rods to single
photons. Biophys J, 75, 1836-57.
RIZZO, J. F. & WYATT, J. 1997. Review: Prospects for a Visual Prosthesis. Neuroscientist, 3, 251-
262.
ROESSLER, G., LAUBE, T., BROCKMANN, C., KIRSCHKAMP, T., MAZINANI, B., GOERTZ, M., KOCH,
C., KRISCH, I., SELLHAUS, B., TRIEU, H. K., WEIS, J., BORNFELD, N., ROTHGEN, H.,
MESSNER, A., MOKWA, W. & WALTER, P. 2009. Implantation and explantation of a
wireless epiretinal retina implant device: observations during the EPIRET3 prospective
clinical trial. Invest Ophthalmol Vis Sci, 50, 3003-8.
ROIZENBLATT, R., WEILAND, J. D., CARCIERI, S., QIU, G., BEHREND, M., HUMAYUN, M. S. &
CHOW, R. H. 2006. Nanobiolistic delivery of indicators to the living mouse retina. J
Neurosci Methods, 153, 154-61.
ROORDA, A., ROMERO-BORJA, F., DONNELLY III, W., QUEENER, H., HEBERT, T. & CAMPBELL, M.
2002. Adaptive optics scanning laser ophthalmoscopy. Opt Express, 10, 405-12.
ROSE, T. L. & ROBBLEE, L. S. 1990. Electrical stimulation with Pt electrodes. VIII.
Electrochemically safe charge injection limits with 0.2 ms pulses. IEEE Trans Biomed Eng,
37, 1118-20.
SAILER, H., SHINODA, K., BLATSIOS, G., KOHLER, K., BONDZIO, L., ZRENNER, E. & GEKELER, F.
2007. Investigation of thermal effects of infrared lasers on the rabbit retina: a study in
the course of development of an active subretinal prosthesis. Graefes Arch Clin Exp
Ophthalmol, 245, 1169-78.
SAVAGE, C. O., KIRAL-KORNEK, F. I., TAHAYORI, B. & GRAYDEN, D. B. Can electric current steering
be used to control perception of a retinal prosthesis patient. 2012 Annual International
Conference of the IEEE Engineering in Medicine and Biology Society, Aug. 28 2012-Sept.
1 2012 2012. 3013-3016.
SCHEJTER, A., TSUR, L., FARAH, N., REUTSKY-GEFEN, I., FALICK, Y. & SHOHAM, S. 2012. Cellular
Resolution Panretinal Imaging of Optogenetic Probes Using a Simple Funduscope. Transl
Vis Sci Technol, 1, 4.
SCHMID, E. W., FINK, W. & WILKE, R. 2014. Operational challenges of retinal prostheses. Med
Eng Phys, 36, 1644-55.
SCHMITZ-VALCKENBERG, S., HOLZ, F. G., BIRD, A. C. & SPAIDE, R. F. 2008. Fundus
autofluorescence imaging: review and perspectives. Retina, 28, 385-409.
SCHUETTLER, M., ORDONEZ, J. S., SILVA SANTISTEBAN, T., SCHATZ, A., WILDE, J. & STIEGLITZ, T.
2010. Fabrication and test of a hermetic miniature implant package with 360 electrical
feedthroughs. Conf Proc IEEE Eng Med Biol Soc, 2010, 1585-8.
SEIDER, M. I. & HAHN, P. 2015. Argus II retinal prosthesis malrotation and repositioning with
intraoperative optical coherence tomography in a posterior staphyloma. Clin
Ophthalmol, 9, 2213-6.
145
SEKIRNJAK, C., HOTTOWY, P., SHER, A., DABROWSKI, W., LITKE, A. M. & CHICHILNISKY, E. J. 2006.
Electrical stimulation of mammalian retinal ganglion cells with multielectrode arrays. J
Neurophysiol, 95, 3311-27.
SHAH, H. A., MONTEZUMA, S. R. & RIZZO, J. F., 3RD 2006. In vivo electrical stimulation of rabbit
retina: effect of stimulus duration and electrical field orientation. Exp Eye Res, 83, 247-
54.
SHARMA, A., RIETH, L., TATHIREDDY, P., HARRISON, R., OPPERMANN, H., KLEIN, M., TOPPER, M.,
JUNG, E., NORMANN, R., CLARK, G. & SOLZBACHER, F. 2012. Evaluation of the packaging
and encapsulation reliability in fully integrated, fully wireless 100 channel Utah Slant
Electrode Array (USEA): Implications for long term functionality. Sens Actuators A Phys,
188, 167-172.
SHIMOMURA, O., JOHNSON, F. H. & SAIGA, Y. 1962. Extraction, purification and properties of
aequorin, a bioluminescent protein from the luminous hydromedusan, Aequorea. J Cell
Comp Physiol, 59, 223-39.
SHIRE, D. B., ELLERSICK, W., KELLY, S. K., DOYLE, P., PRIPLATA, A., DROHAN, W., MENDOZA, O.,
GINGERICH, M., MCKEE, B., WYATT, J. L. & RIZZO, J. F. 2012. ASIC design and data
communications for the Boston retinal prosthesis. Conf Proc IEEE Eng Med Biol Soc,
2012, 292-5.
SHIRE, D. B., JONES, W. K., KARBASI, A., BEHAN, S., GINGERICH, M., KELLY, S., WYATT, J. L. &
RIZZO, J. F. 2014. Advanced Hermetic Feedthrough and Packaging Technology for the
Boston Retinal Prosthesis. Invest Ophthalmol Vis Sci, 55, 1835-1835.
SHIVDASANI, M. N., SINCLAIR, N. C., DIMITROV, P. N., VARSAMIDIS, M., AYTON, L. N., LUU, C. D.,
PERERA, T., MCDERMOTT, H. J., BLAMEY, P. J. & BIONIC VISION AUSTRALIA, C. 2014.
Factors affecting perceptual thresholds in a suprachoroidal retinal prosthesis. Invest
Ophthalmol Vis Sci, 55, 6467-81.
SMETTERS, D., MAJEWSKA, A. & YUSTE, R. 1999. Detecting action potentials in neuronal
populations with calcium imaging. Methods, 18, 215-21.
STANGA, P. E., JALIL, A., TSAMIS, E., PAPAYANNIS, A., DORN, J. D., GREENBERG, R. J. & MCGUIRE,
W. Argus II® electronic epiretinal prosthesis in advanced dry AMD: safety and feasibility
study and preliminary functional results. The Annual Meeting of the Association for
Research in Vision and Ophthalmology, 2016 Seattle WA.
STINGL, K., BACH, M., BARTZ-SCHMIDT, K. U., BRAUN, A., BRUCKMANN, A., GEKELER, F.,
GREPPMAIER, U., HORTDORFER, G., KUSNYERIK, A., PETERS, T., WILHELM, B., WILKE, R.
& ZRENNER, E. 2013a. Safety and efficacy of subretinal visual implants in humans:
methodological aspects. Clin Exp Optom, 96, 4-13.
STINGL, K., BARTZ-SCHMIDT, K. U., BESCH, D., BRAUN, A., BRUCKMANN, A., GEKELER, F.,
GREPPMAIER, U., HIPP, S., HORTDORFER, G., KERNSTOCK, C., KOITSCHEV, A.,
KUSNYERIK, A., SACHS, H., SCHATZ, A., STINGL, K. T., PETERS, T., WILHELM, B. &
ZRENNER, E. 2013b. Artificial vision with wirelessly powered subretinal electronic
implant alpha-IMS. Proc Biol Sci, 280, 20130077.
STINGL, K., BARTZ-SCHMIDT, K. U., BESCH, D., CHEE, C. K., COTTRIALL, C. L., GEKELER, F.,
GROPPE, M., JACKSON, T. L., MACLAREN, R. E., KOITSCHEV, A., KUSNYERIK, A.,
NEFFENDORF, J., NEMETH, J., NAEEM, M. A., PETERS, T., RAMSDEN, J. D., SACHS, H.,
SIMPSON, A., SINGH, M. S., WILHELM, B., WONG, D. & ZRENNER, E. 2015a. Subretinal
Visual Implant Alpha IMS--Clinical trial interim report. Vision Res, 111, 149-60.
STINGL, K., BARTZ-SCHMIDT, K. U., BESCH, D., CHEE, C. K., COTTRIALL, C. L., GEKELER, F.,
GROPPE, M., JACKSON, T. L., MACLAREN, R. E., KOITSCHEV, A., KUSNYERIK, A.,
NEFFENDORF, J., NEMETH, J., NAEEM, M. A. N., PETERS, T., RAMSDEN, J. D., SACHS, H.,
146
SIMPSON, A., MANDEEP, S. S., WILHELM, B., WONG, D. & ZRENNER, E. 2015b. Subretinal
Visual Implant Alpha IMS –Clinical trial interim report. Vision Res, 111, 149-160.
STOSIEK, C., GARASCHUK, O., HOLTHOFF, K. & KONNERTH, A. 2003. In vivo two-photon calcium
imaging of neuronal networks. Proc Natl Acad Sci U S A, 100, 7319-24.
STRONKS, H. C. & DAGNELIE, G. 2014. The functional performance of the Argus II retinal
prosthesis. Expert Rev Med Devices, 11, 23-30.
SUANING, G. J., LAVOIE, P., FORRESTER, J., ARMITAGE, T. & LOVELL, N. H. 2006. Microelectronic
retinal prosthesis: III. a new method for fabrication of high-density hermetic
feedthroughs. Conf Proc IEEE Eng Med Biol Soc, 1, 1638-41.
SUESSERMAN, M. F., SPELMAN, F. A. & RUBINSTEIN, J. T. 1991. In vitro measurement and
characterization of current density profiles produced by nonrecessed, simple recessed,
and radially varying recessed stimulating electrodes. IEEE Transactions on Biomedical
Engineering, 38, 401-408.
SUNG CHEOL, P., MIN KYU, P. & MOON GI, K. 2003. Super-resolution image reconstruction: a
technical overview. IEEE Signal Processing Magazine, 20, 21-36.
SUZUKI, S., HUMAYUN, M. S., WEILAND, J. D., CHEN, S. J., MARGALIT, E., PIYATHAISERE, D. V. &
DE JUAN, E., JR. 2004. Comparison of electrical stimulation thresholds in normal and
retinal degenerated mouse retina. Jpn J Ophthalmol, 48, 345-9.
SWEDLOW, J. R. 2013. Quantitative fluorescence microscopy and image deconvolution. Methods
Cell Biol, 114, 407-26.
TAKAHASHI, A., CAMACHO, P., LECHLEITER, J. D. & HERMAN, B. 1999. Measurement of
intracellular calcium. Physiol Rev, 79, 1089-125.
TIAN, L., HIRES, S. A., MAO, T., HUBER, D., CHIAPPE, M. E., CHALASANI, S. H., PETREANU, L.,
AKERBOOM, J., MCKINNEY, S. A., SCHREITER, E. R., BARGMANN, C. I., JAYARAMAN, V.,
SVOBODA, K. & LOOGER, L. L. 2009. Imaging neural activity in worms, flies and mice with
improved GCaMP calcium indicators. Nat Methods, 6, 875-81.
TRUONG, T. V., SUPATTO, W., KOOS, D. S., CHOI, J. M. & FRASER, S. E. 2011. Deep and fast live
imaging with two-photon scanned light-sheet microscopy. Nat Methods, 8, 757-60.
TSIEN, R. Y. 1980. New calcium indicators and buffers with high selectivity against magnesium
and protons: design, synthesis, and properties of prototype structures. Biochemistry, 19,
2396-404.
VAN RULLEN, R. & THORPE, S. J. 2001. Rate coding versus temporal order coding: what the
retinal ganglion cells tell the visual cortex. Neural Comput, 13, 1255-83.
VANEY, D. I. 1990. Chapter 2 The mosaic of amacrine cells in the mammalian retina. Progress in
Retinal Research, 9, 49-100.
VANHOESTENBERGHE, A. & DONALDSON, N. 2013. Corrosion of silicon integrated circuits and
lifetime predictions in implantable electronic devices. J Neural Eng, 10, 031002.
WANG, G., LIU, W., SIVAPRAKASAM, M., ZHOU, M., WEILAND, J. D. & HUMAYUN, M. S. 2006. A
dual band wireless power and data telemetry for retinal prosthesis. Conf Proc IEEE Eng
Med Biol Soc, 1, 4392-5.
WEILAND, J. D., ANDERSON, D. J. & HUMAYUN, M. S. 2002. In vitro electrical properties for
iridium oxide versus titanium nitride stimulating electrodes. IEEE Trans Biomed Eng, 49,
1574-9.
WEILAND, J. D. & HUMAYUN, M. S. 2014. Retinal prosthesis. IEEE Trans Biomed Eng, 61, 1412-
24.
WEILAND, J. D., KIMOCK, F. M., YEHODA, J. E., GILL, E., MCLNTOSH, B. P., NASIATKA, P. J. &
TANGUAY, A. R. Chip-scale packaging for bioelectronic implants. Neural Engineering
(NER), 2013 6th International IEEE/EMBS Conference on, 6-8 Nov. 2013 2013. 931-936.
147
WEITZ, A. C., BEHREND, M. R., AHUJA, A. K., CHRISTOPHER, P., WEI, J., WUYYURU, V., PATEL, U.,
GREENBERG, R. J., HUMAYUN, M. S., CHOW, R. H. & WEILAND, J. D. 2014. Interphase
gap as a means to reduce electrical stimulation thresholds for epiretinal prostheses.
Journal of neural engineering, 11, 016007-016007.
WEITZ, A. C., BEHREND, M. R., LEE, N. S., KLEIN, R. L., CHIODO, V. A., HAUSWIRTH, W. W.,
HUMAYUN, M. S., WEILAND, J. D. & CHOW, R. H. 2013. Imaging the response of the
retina to electrical stimulation with genetically encoded calcium indicators. J
Neurophysiol, 109, 1979-88.
WEITZ, A. C., NANDURI, D., BEHREND, M. R., GONZALEZ-CALLE, A., GREENBERG, R. J.,
HUMAYUN, M. S., CHOW, R. H. & WEILAND, J. D. 2015. Improving the spatial resolution
of epiretinal implants by increasing stimulus pulse duration. Sci Transl Med, 7, 318ra203.
WILKE, R., GABEL, V. P., SACHS, H., BARTZ SCHMIDT, K. U., GEKELER, F., BESCH, D., SZURMAN, P.,
STETT, A., WILHELM, B., PETERS, T., HARSCHER, A., GREPPMAIER, U., KIBBEL, S., BENAV,
H., BRUCKMANN, A., STINGL, K., KUSNYERIK, A. & ZRENNER, E. 2011a. Spatial resolution
and perception of patterns mediated by a subretinal 16-electrode array in patients
blinded by hereditary retinal dystrophies. Invest Ophthalmol Vis Sci, 52, 5995-6003.
WILKE, R. G., MOGHADAM, G. K., LOVELL, N. H., SUANING, G. J. & DOKOS, S. 2011b. Electric
crosstalk impairs spatial resolution of multi-electrode arrays in retinal implants. J Neural
Eng, 8, 046016.
WONG, I. Y., POON, M. W., PANG, R. T., LIAN, Q. & WONG, D. 2011. Promises of stem cell
therapy for retinal degenerative diseases. Graefes Arch Clin Exp Ophthalmol, 249, 1439-
48.
WONG, R. O. 1998. Calcium imaging and multielectrode recordings of global patterns of activity
in the developing nervous system. Histochem J, 30, 217-29.
WONG, R. O. & OAKLEY, D. M. 1996. Changing patterns of spontaneous bursting activity of on
and off retinal ganglion cells during development. Neuron, 16, 1087-95.
WONG, W. L., SU, X., LI, X., CHEUNG, C. M., KLEIN, R., CHENG, C. Y. & WONG, T. Y. 2014. Global
prevalence of age-related macular degeneration and disease burden projection for 2020
and 2040: a systematic review and meta-analysis. Lancet Glob Health, 2, e106-16.
WU, C. C., REILLY, J. F., YOUNG, W. G., MORRISON, J. H. & BLOOM, F. E. 2004. High-throughput
morphometric analysis of individual neurons. Cereb Cortex, 14, 543-54.
WU, K., KLEIN, R. L., MEYERS, C. A., KING, M. A., HUGHES, J. A., MILLARD, W. J. & MEYER, E. M.
2003. Long-term neuronal effects and disposition of ectopic preproNGF gene transfer
into the rat septum. Hum Gene Ther, 14, 1463-72.
XIAO, X., WANG, J., LIU, C., CARLISLE, J. A., MECH, B., GREENBERG, R., GUVEN, D., FREDA, R.,
HUMAYUN, M. S., WEILAND, J. & AUCIELLO, O. 2006. In vitro and in vivo evaluation of
ultrananocrystalline diamond for coating of implantable retinal microchips. J Biomed
Mater Res B Appl Biomater, 77, 273-81.
YAMAUCHI, Y., FRANCO, L. M., JACKSON, D. J., NABER, J. F., ZIV, R. O., RIZZO, J. F., KAPLAN, H. J.
& ENZMANN, V. 2005. Comparison of electrically evoked cortical potential thresholds
generated with subretinal or suprachoroidal placement of a microelectrode array in the
rabbit. J Neural Eng, 2, S48-56.
YANAI, D., WEILAND, J. D., MAHADEVAPPA, M., GREENBERG, R. J., FINE, I. & HUMAYUN, M. S.
2007. Visual Performance Using a Retinal Prosthesis in Three Subjects With Retinitis
Pigmentosa. American Journal of Ophthalmology, 143, 820-827.e2.
YANG, X. L. & WU, S. M. 1991. Feedforward lateral inhibition in retinal bipolar cells: input-output
relation of the horizontal cell-depolarizing bipolar cell synapse. Proc Natl Acad Sci U S A,
88, 3310-3.
148
YU, J., DANIELS, B. A. & BALDRIDGE, W. H. 2009. Slow excitation of cultured rat retinal ganglion
cells by activating group I metabotropic glutamate receptors. J Neurophysiol, 102, 3728-
39.
YUE, L., WEILAND, J. D., ROSKA, B. & HUMAYUN, M. S. 2016. Retinal stimulation strategies to
restore vision: Fundamentals and systems. Prog Retin Eye Res, 53, 21-47.
YUSTE, R. & KATZ, L. C. 1991. Control of postsynaptic Ca2+ influx in developing neocortex by
excitatory and inhibitory neurotransmitters. Neuron, 6, 333-44.
ZARIWALA, H. A., BORGHUIS, B. G., HOOGLAND, T. M., MADISEN, L., TIAN, L., DE ZEEUW, C. I.,
ZENG, H., LOOGER, L. L., SVOBODA, K. & CHEN, T. W. 2012. A Cre-dependent GCaMP3
reporter mouse for neuronal imaging in vivo. J Neurosci, 32, 3131-41.
ZHANG, Y., RAUEN, S. L., KOSS, M., CALLE, A., DINIZ, B., SWENSON, S., MARKLAND, F. S., UFER,
S., ECKHARDT, H., M.S. HUMAYUN & WEILAND, J. D. Wide-Field Retinal Prosthesis with
three dimensional, contoured, silicone/polyimide substrates. IEEE Neural Engineering
conference, 2013 San diego, CA.
ZHAO, Y., ARAKI, S., WU, J., TERAMOTO, T., CHANG, Y. F., NAKANO, M., ABDELFATTAH, A. S.,
FUJIWARA, M., ISHIHARA, T., NAGAI, T. & CAMPBELL, R. E. 2011. An expanded palette of
genetically encoded Ca(2)(+) indicators. Science, 333, 1888-91.
ZHOU, D. D., DORN, J. D. & GREENBERG, R. J. The Argus II Retinal Prosthesis System: An
Overview. IEEE ICME conference, 2013 San Jose, CA.
ZHOU, M., YUCE, M. R. & LIU, W. 2008. A Non-Coherent DPSK Data Receiver With Interference
Cancellation for Dual-Band Transcutaneous Telemetries. IEEE Journal of Solid-State
Circuits, 43, 2003-2012.
ZOLOTUKHIN, S., BYRNE, B. J., MASON, E., ZOLOTUKHIN, I., POTTER, M., CHESNUT, K.,
SUMMERFORD, C., SAMULSKI, R. J. & MUZYCZKA, N. 1999. Recombinant adeno-
associated virus purification using novel methods improves infectious titer and yield.
Gene Ther, 6, 973-85.
ZRENNER, E., BARTZ-SCHMIDT, K. U., BENAV, H., BESCH, D., BRUCKMANN, A., GABEL, V. P.,
GEKELER, F., GREPPMAIER, U., HARSCHER, A., KIBBEL, S., KOCH, J., KUSNYERIK, A.,
PETERS, T., STINGL, K., SACHS, H., STETT, A., SZURMAN, P., WILHELM, B. & WILKE, R.
2011. Subretinal electronic chips allow blind patients to read letters and combine them
to words. Proc Biol Sci, 278, 1489-97.
Abstract (if available)
Abstract
Retinal prosthetic implants have shown potential to restore partial vision to patients blinded by retinitis pigmentosa or dry age-related macular degeneration, via a camera-driven multielectrode array that electrically stimulates surviving retinal neurons. To analyze the effectiveness of these implants, single-unit or multielectrode recordings of neurons from isolated animal retina are commonly used. However, such electrical recording methods are strongly affected by stimulation artifact and limited in terms of the spatial patterns of retinal activation. Virus-transduced calcium indicators are effective reporters of neural activity, offering the advantage of cell-specific labeling. To track the time dependence of in vivo expression levels of genetically encoded calcium indicators (GECIs) in rodent retina, we developed a noninvasive imaging approach based on a custom-modified, low-cost and simple fundus system that enabled us to monitor and characterize in vivo bright-field and fluorescence retinal image, further evaluating in vivo fluorescence reporter expression. ❧ Commercial epi-retinal prostheses mostly use charge-balanced symmetric cathodic-first biphasic pulses to depolarize retinal ganglion cells (RGCs) and bipolar cells (BCs), resulting in the perception of light in blind patients. However, previous clinical study for patients with Argus II epiretinal implants reported most percepts evoked by single electrode stimulation were elongated and aligned with estimated axon path of retinal ganglion cells, suggesting the activation of axon bundles. Based on the established in vitro calcium imaging and electrophysiological animal model, we performed in vitro calcium imaging for different stimulation paradigms, focusing primarily on short duration pulse with different types of waveform and current steering method that can avoid axonal stimulation and manipulate the thresholds of targeted RGC somas. The findings support the possibility to manipulate the responses of RGCs through varying the stimulation waveform or return electrodes, thus potentially forming more ideal shape perception with higher spatial resolution in future retinal prosthesis design.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Improving stimulation strategies for epiretinal prostheses
PDF
Towards a high resolution retinal implant
PDF
Stimulation strategies to improve efficiency and temporal resolution of epiretinal prostheses
PDF
Understanding the degenerate retina's response to electrical stimulation: an in vitro approach
PDF
Viewing the picture we paint
PDF
Cortical and subcortical responses to electrical stimulation of rat retina
PDF
Electrical stimulation of degenerate retina
PDF
Modeling retinal prosthesis mechanics
PDF
Saliency based image processing to aid retinal prosthesis recipients
PDF
Synaptic second wind: phosphene fading of the Argus II and glutamate secretion in the inner retina
PDF
An in vitro model of a retinal prosthesis
PDF
Prosthetic vision in blind human patients: Predicting the percepts of epiretinal stimulation
PDF
PET study of retinal prosthesis functionality
PDF
Investigation of the electrode-tissue interface of retinal prostheses
PDF
Effect of continuous electrical stimulation on retinal structure and function
PDF
Dependence of rabbit retinal synchrony on visual stimulation parameters
PDF
Next generation neural interfaces for neural modulation and prosthesis
PDF
Prosthetic visual perception: retinal electrical stimulation in blind human patients
PDF
Ultrasound neuromodulation and its applications for noninvasive vision restoration
PDF
Therapeutic electrical stimulation strategies for neuroregeneration and neuroprotection of retinal neurons
Asset Metadata
Creator
Chang, Yao-Chuan
(author)
Core Title
Manipulation of RGCs response using different stimulation strategies for retinal prosthesis
School
Viterbi School of Engineering
Degree
Doctor of Philosophy
Degree Program
Biomedical Engineering
Publication Date
10/23/2017
Defense Date
09/08/2017
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
Biomedical Engineering,calcium imaging,electrophysiology,multielectrode arrays,neural engineering,neural stimulation,OAI-PMH Harvest,retina,retinal ganglion cells,retinal prosthesis
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Weiland, James (
committee chair
), Chow, Robert (
committee member
), Fraser, Scott (
committee member
), Humayun, Mark (
committee member
), Weitz, Andrew (
committee member
)
Creator Email
charles9412029@gmail.com,yaochuac@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c40-446234
Unique identifier
UC11263886
Identifier
etd-ChangYaoCh-5811.pdf (filename),usctheses-c40-446234 (legacy record id)
Legacy Identifier
etd-ChangYaoCh-5811.pdf
Dmrecord
446234
Document Type
Dissertation
Rights
Chang, Yao-Chuan
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
Tags
calcium imaging
electrophysiology
multielectrode arrays
neural engineering
neural stimulation
retina
retinal ganglion cells
retinal prosthesis