University of Southern California Dissertations and Theses

Effect of continuous electrical stimulation on retinal structure and function

(USC Thesis Other)

Effect of continuous electrical stimulation on retinal structure and function

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content EFFECT OF CONTINUOUS ELECTRICAL STIMULATION ON RETINAL
STRUCTURE AND FUNCTION

by

Aditi Ray

A Dissertation Presented to the
FACULTY OF USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BIOMEDICAL ENGINEERING)

August 2010

Copyright 2010 Aditi Ray
ii
Dedication

In loving memory of my grandmother, Smt. Pravarani Mallick
(11.16.1927 – 02.19.2007).

iii
Acknowledgements

“PhD is more about the journey than the end”, reminded my thesis advisor Dr.
Weiland during the long months of bad experiments and unexpected results. I owe my
deepest gratitude to him for his support and encouragement these past six years. He has
been a great mentor and role model and impressed upon me the importance of
maintaining a balance between professional and personal life. His confidence in my
ability to work on a project that took many unexpected and often difficult turns gave me
the strength and conviction to plough ahead.
I would like to thank all my committee members. My conversations with Dr.
Humayun helped me realize that my true interest lay in taking knowledge and applying it
towards helping others. Long discussions with Dr. Loeb helped me navigate successfully
through my research. During the countless hours spent in the basement of Biegler Hall,
Dr. Lee taught me everything I know about retinal histology, for which I wish to thank
her. I would also like to thank Dr. Hinton and Dr. Mansfeld who helped me shape my
research into an interesting scientific story.
My fellow colleagues at the Bioelectronics Research Lab including Alice, Leanne,
Neha, Devyani, Vivek, Tim, Andrew, Ashish, Brooke, Sam, Nick, Navya, Samantha,
Alan and Mathew made my days in lab enjoyable. Their enthusiasm and zest for high
quality scientific research inspired me to raise my expectations of myself. Special regards
to Jack, one of the first people in lab that I worked with and who trained me in
electrochemical techniques. I would also like to thank all the members of the Visual
iv
Processing Lab including Dr. Grzywacx, Susmita, Junkwan, Xiwu, Joaquin, Nadav,
Yerina, Gerry and Arvind who made me feel a part of their lab these past 3 years. Thanks
to Doris, Diana and Christian, I will always have fond memories of our trips to DC every
year.
I wish to thank all my friends who made sure that my life these past few years
remained balanced. I would especially like to thank Bhavna, Prateek, Ashmita, Arpitha,
Amit, Gaurav, Anurupa, Jassi, Pramod and Vaibhav who kept life filled with fun.
I will always be grateful to my parents and sister. Though thousands of miles
away their unwavering support and love helped me through the most difficult phases of
my life. The presence of my two closest friends Saryu and Avinash by my side every step
of the way gave me the strength to push myself further. I cherish their continued presence
in my life.

v
Table of Contents

Dedication ii
Acknowledgements iii
List of Tables viii
List of Figures ix
Abbreviations xii
Abstract xvi
Chapter 1: Introduction
1.1 Electrode-Electrolyte Interface 1
1.1.1 Mechanism of charge injection 4
1.1.2 Electrode characterization
1.1.2.1 Electrode potential 6
1.1.2.2 Polarizable and non-polarizable electrodes 9
1.1.2.3 Measurement of impedance 9
1.1.2.4 Surface reactions and potential limits 13
1.1.2.5 Voltage response 18
1.1.3 Overview of extracellular stimulation 19
1.1.4 Safe stimulation of tissue
1.1.4.1 Mechanisms of neural injury 27
1.1.4.2 Parameters for safe stimulation 29
1.2 Anatomy and Physiology
1.2.1 Structure and function of retina 32
1.2.2 Central visual pathway 44
1.3 Visual Prosthesis 51
1.4 Goal of present study 56

Chapter 2: Morphological Changes in Retina due to Retinal Degeneration in
Transgenic Animal Model
2.1 Background 58
2.2 Experimental Protocol 60
2.3 Results
2.3.1 Photoreceptor degeneration 60
2.3.2 Glial reaction 63
2.3.3 Changes in horizontal cells 64
2.3.4 Changes in rod bipolar cells 65
2.3.5 Cone bipolar cells 66
vi
2.3.6 Amacrine cells 67
2.3.7 Quantification of modifications of horizontal and rod
bipolar cells in RD retinas 69
2.4 Discussion 71

Chapter 3: Electrode-Retina Interface
3.1 Characterization of stimulation electrode
3.1.1 Background 75
3.1.2 Experimental protocol 80
3.1.3 Results
3.1.3.1 Impedance spectroscopy 80
3.1.3.2 Cyclic voltammogram 81
3.1.3.3 Voltage response 83
3.1.4 Conclusion 85
3.2 Electrode positioning
3.2.1 Background 85
3.2.2 Experimental protocol 88
3.2.3 Results
3.2.3.1 Electrode impedance as a function of distance
from the retina 89
3.2.3.2 Presence of electrically evoked potentials in the
SC as a function of electrode impedance 91
3.2.3.3 Retinal damage due to mechanical contact
between electrode and retina 93
3.2.4 Discussion 94

Chapter 4: Morphological changes due to electrode contact and continuous
stimulation
4.1 Background 96
4.2 Experimental protocol 100
4.3 Results
4.3.1 Gross morphology 101
4.3.2 Expression of synaptic vesicle proteins (SV2A and SV2B) 104
4.3.3 Abnormal processes of horizontal and rod bipolar cells 105
4.3.4 Double immunofluorescence for SV2B with PKC alpha 107
4.3.5 No morphological changes in inner retina 108
4.3.6 Glial expression 109
4.4 Discussion 110
vii
Chapter 5: Elevation in excitation threshold due to continuous stimulation
5.1 Background 113
5.2 Experimental protocol 117
5.3 Results
5.3.1 Retinal morphology
5.3.1.1 Histology 119
5.3.1.2 Glial reaction 124
5.3.2 Electrically evoked response in SC
5.3.2.1 Threshold 125
5.3.2.2 Response feature analysis 126
5.4 Discussion 128

Chapter 6: Temporal dynamics of excitation threshold increase
6.1 Background 132
6.2 Experimental protocol 132
6.3 Results
6.3.1 Effect of stimulation duration 133
6.3.2 Effect of 1s stimulation 136
6.4 Discussion 141

Chapter 7: Summary
7.1 Key findings and future work 145
7.2 Implications for a retinal prosthesis 148

Chapter 8: Methods
8.1 Animal model 150
8.2 Surgical procedure 150
8.3 Stimulation electrode and retinal stimulation 151
8.4 Electrochemical techniques – impedance and cyclic voltammetry 152
8.5 Superior colliculus surgery and recording electrode insertion 153
8.6 Electrically evoked potential recording and analysis 153
8.7 Tissue preparation 154
8.8 Hematoxylin staining 155
8.9 Immunocytochemistry 155
8.10 Statistics 158
Bibliography 159

viii
List of Tables

Table 5.1 Experimental groups 118
Table 8.1 Antibody list 156 

ix
List of Figures

Figure 1.1 Equivalent circuit model of the electrode-electrolyte interface 12
Figure 1.2 Cyclic voltammogram of platinum 14 
Figure 1.3 Cyclic voltammograms from reversible to irreversible domain 18 
Figure 1.4 Example of strength-duration graph 25
Figure 1.5 Example of stimulus-response graph 26 
Figure 1.6 Charge vs charge density graph for safe stimulation 31 
Figure 1.7 Horizontal section of the right human eye 34 
Figure 1.8 Schematic of the cell types and layers of retina 35 
Figure 1.9 Rod and cone density in human retina 38 
Figure 1.10 Schematic of rat brain 47 
Figure 1.11 Images of superior colliculus cells 50 
Figure 1.12 Retinotopic mapping in the superior colliculus 50 
Figure 1.13 Schematic of epi- and subretinal implants in the eye 54 
Figure 2.1 Micrographs of TOPRO-3 stained retinas 62  
Figure 2.2 Micrographs of TOPRO-3 and S-opsin labelled retinas 63 
Figure 2.3 Micrographs of GFAP labelled retinas 64 
Figure 2.4 Micrographs taken of calbindin D-28K labelled retinas 65 
Figure 2.5 Micrographs of PKCα labelled retinas 66 
Figure 2.6 Micrographs of recoverin labelled retinas 67  
Figure 2.7 Micrographs of GABA and parvalbumin labelled retinas 68 
x
Figure 2.8 Wholemount retinas labelled with calbindin and PKCα 70 
Figure 3.1 Impedance spectrum of stimulation electrode 81
Figure 3.2 Cyclic voltammograms of stimulation electrode 82 
Figure 3.3 Voltage response curves across electrode-tissue interface 84
Figure 3.4 Impedance plots of stimulation electrode 90
Figure 3.5 Impedance vs distance graph 91
Figure 3.6 Impedance vs threshold graph 92
Figure 3.7 H & E stained experimental retinas for low and high impedances 93 
Figure 4.1 Micrographs of hematoxylin stained experimental retinas 103 
Figure 4.2 Micrographs of SV2A and SV2B labelled experimental retinas 105
Figure 4.3 Micrographs of calbindin and PKC labelled experimental retinas 107 
Figure 4.4 Double exposure of SV2B and PKC labelled experimental retinas 108
Figure 4.5 Micrographs of Glyt-1 and GAD65 labelled experimental retinas 109
Figure 4.6 Micrographs of GFAP labelled experimental retinas 110 
Figure 5.1 Micrographs of hematoxylin stained experimental retinas 120 
Figure 5.2 Micrographs of major retinal cell types in experimental retinas 123 
Figure 5.3 Micrographs of GFAP labelled experimental retinas 124 
Figure 5.4 Graph illustrating threshold change factor 126 
Figure 5.5 Example of electrically evoked potential 127 
Figure 5.6 Graphs illustrating response strength vs stimulus current 128 
Figure 6.1 Schematic of experimental protocol 134 
Figure 6.2 Response strength vs stimulus duration at 1.5 times threshold 135 
xi
Figure 6.3 Response strength vs stimulation duration at suprathreshold 135 
Figure 6.4 Schematic of experimental protocol 136 
Figure 6.5 Graphs illustrating effect of 1s stimulation in normal retina 138 
Figure 6.6 Graphs illustrating effect of 1s stimulation in degenerate retina 139 
Figure 6.7 Graph illustrating effect of 1s stimulation applied every 7s 140 
Figure 6.8 Micrographs of hematoxylin stained experimental retinas 140 

xii
List of Abbreviations

Ach Acetylcholine
AIROF Anodic Iridium Oxide Films
cAMP cyclic Adenosine Monophosphate
cGMP cyclic Guanosine Monophosphate
CMOS Complimentary metal-oxide semiconductor
CNS Central Nervous System
CNTs Carbon Nanotubes
CPE Constant Phase Element
CV Cyclic Voltammetry
DBS Deep Brain Stimulation
DLG Dorsal Lateral Geniculate
DpG Deep Gray
DpW Deep White
EAD Early Axonal Degeneration
EEPs Electrically Evoked Potentials
EIS Electrochemical Impedance Spectroscopy
FA Fluorescein Angiography
FITC-PNA Fluorescein-conjugated peanut agglutinin
GABA γ-aminobutyric acid
xiii
GAD 65 Glutamic-acid decarboxylase
GCL Ganglion Cell Layer
GFAP Glial Fibrillary Acidic Protein
Glyt-1 Glycine Transporter-1
GOα G-protein
IC Inferior Colliculus
IHP Inner Helmholtz Plane
In W Intermediate White
InG Intermediate Gray
INL Inner Nuclear Layer
IPL Inner Plexiform Layer
LGN Lateral Geniculate Nucleus
LTD Long-Term Depression
MAP1 Microtubule-Associated Protein 1
MCP1 Monocyte Chemoattractant Protein 1
NMDA N-methyl-D-aspartate
OHP Outer Helmholtz Plane
ONL Outer Nuclear Layer
Op Optic
OPL Outer Plexiform Layer
OPN Olivary Pretectal Nucleus
xiv
P Postnatal
PBS Phosphate-buffered saline
PEDOT poly (3,4-ethylenedioxythiophene)
PKCα Alpha isoform of protein kinase C
PPD Paired-Pulse Depression
PSTHs Post-Stimulus Time Histograms
RCS Royal College of Surgeons
RD retinal degenerate
RGC Retinal ganglion cell
RGFs Response Growth Functions
RHE Reversible Hydrogen Electrode
RP Retinitis Pigmentosa
RPE Retinal Pigment Epithelium
SANR Short Acting Neuronal Refractory
SC Superior Colliculus
SCN Suprachiasmatic Nucleus
SGS Stratum Griseum Superficiale
SIDNE Stimulation Induced Depression in Neuronal Excitability
SIROF Sputtering the oxide films onto a substrate from an iridium target
SO Stratum Opticum
SSMP Second Sight Medical Products
xv
STS Suprachoriodal Transretinal Stimulation
SuG Superficial Gray
SZ Stratum Zonale
TENS Transcutaneous Electrical Nerve Stimulation
TIROF Thermal decomposition of layers of Iridium salts
TUNEL Terminal deoxynucleotidyl transferase (TdT)-mediated dUTP-biotin
Nick-End Labelling

VLG Ventral Lateral Geniculate
Z Zonal

xvi
Abstract

Electrical stimulation of the central nervous system albeit an unnatural way, has
been found to be an effective way of causing neuronal excitation. Retinal prosthesis is an
example of such a neuroprosthesis that strives to provide vision to people suffering from
Retinitis Pigmentosa and Age-related Macular Degeneration. In these diseases, the
photoreceptors in the retina undergo a progressive degeneration leaving the remaining
retinal neurons relatively intact. It is by electrically stimulating these retinal neurons that
a retinal prosthesis aims to elicit visual percepts.
In addition to providing effective stimulation, such a device has to do so in a
manner that is safe for both the device and surrounding biological environment.
Numerous studies have been performed to assess the safety limits of electrical stimulation
of neurons. However, majority of these studies have been performed in structures other
than the retina. The few safety studies performed in the retina have looked at the effect of
brief pulses of stimulation. Hence, the work presented in this thesis investigates the effect
of prolonged stimulation of the retina on both the structure and function of the retina.
One-hour long epiretinal stimulation was performed in an in vivo animal model
along with the recording of electrically evoked responses in the superior colliculus and
retinal histology. Results presented in this thesis demonstrate that the retina is capable of
tolerating continuous stimulation at charge densities higher than the safe limit of platinum
even when delivered at high stimulus frequencies. However, such a stimulation regime
causes a decrease in the electrically evoked responses in the superior colliculus resulting
xvii
in an elevation of threshold of excitation. This temporary desensitization was found to
occur within the first tens of seconds of stimulation after which the responses attained a
steady state level. Observations made during this research points towards some form of
adaptation by the retina in response to continuous electrical stimulation.
The work presented in this thesis is aimed towards the development of high-
resolution epiretinal prosthesis. This research is one of the first to systematically
investigate the effect of continuous stimulation from a safety perspective and will
hopefully help in the design of safe and efficacious stimulus protocols for retinal
prosthesis recipients.

1
Chapter 1
Introduction

Neuroprostheses strive to restore lost functionality by electrically stimulating
target cells. With increasing popularity of neural stimulation as an alternate treatment or
therapy, applications include visual prostheses, upper and lower limb prostheses for
spinal cord injury, bladder prostheses and deep brain stimulation for Parkinson’s disease,
dystonia and epilepsy. All such prostheses need to not only provide adequate stimulation
but also need to do so without causing any damage to the target tissue or to the
electrodes. Also as progress is being made in the field of microelectrode fabrication,
there is an increasing thirst for developing and implanting electrodes as small as single
cells in an effort to closely mimic bodily functions. All these in turn place severe
engineering requirements on many aspects of the neuroprosthetic system and in many
cases become specific to the application in question. The present chapter provides an
overview of the different components in a neuroprosthetic system. These include the
electrode-electrolyte interface along with neural stimulation basics and anatomy and
physiology of the visual system.

1.1 Electrode-Electrolyte Interface
A major component of a neuroprosthesis is the electrode-tissue interface.
Implanted electrodes are placed in close apposition with the target neural structure and
are required to provide focal stimulation that is safe for both the electrodes and the neural
2
structure. Also, in many cases, the implanted electrodes are required to function over the
lifetime of the implant recipients. Thus, in order to aid the development of
neuroprostheses it is important to study this interface in detail. The following section
provides an overview of the electrode-electrolyte interface with reference to requirements
for neural stimulation. The material presented has been derived mainly from three
sources: Principles of Neural Science by Kandel, Schwartz and Jessell (Kandel et al.,
2000), Electrochemical Methods: Fundamental and Applications by Bard and Faulkner
(Bard and Faulkner, 2004) and Electrical Stimulation of Excitable Tissue: design of
efficacious and safe protocols by Merrill (Merrill et al., 2005).
Whenever a metal electrode is placed in an electrolyte, thermodynamic processes
operate to bring the two phases in electrochemical equilibrium. This causes attraction
between the charge carriers in the two phases leading to the formation of a net potential
across the interface. This interface is popularly known as the electrical double layer with
the principal charge carriers in the metal phase being the electrons and those in the
electrolyte being the ions. The importance of this interface lies in the fact that for any
neural excitation to take place, current has to flow through tissue. Hence, a critical part of
understanding and controlling stimulation through metal electrodes lies in understanding
the different electrochemical processes that take place at the electrode-electrolyte
interface.
The finite separation of charge that leads to the formation of the electrical double
layer has several manifestations. One reason for charge redistribution at the interface is
ions in the electrolyte combining with the electrode. This leads to a net transfer of
3
electrons between the two phases causing a plane of charge at the metal electrode that is
opposed by a plane of charge in the electrolyte. Other reasons for the formation of the
double layer include the specific adsorption of certain chemical species and preferential
orientation of polar molecules such as water. The solution side of the double layer is
composed of several layers. The inner layer called the Helmholtz or Stern layer consists
of solvent molecules and some other species such as specifically adsorbed ions or
molecules. The locus of electrical centres of the specifically adsorbed ions defines the
inner Helmholtz plane (IHP) while the locus of centres of the nearest solvated ions
defines the outer Helmholtz plane (OHP). The solvated ions are said to be non-
specifically adsorbed as their interaction with the charged metal is independent of the
chemical properties of the ions. These ions are distributed in the three dimensional region
called the diffuse layer extending from the OHP into the bulk solution. The thickness of
the diffuse layer is dependent upon the total ionic concentration of the solution.
The metal electrode-solution interface has been shown to behave like a capacitor
with a finite amount of charge residing in a very thin layer (<0.1Å) on the metal surface
(excess or deficiency of electrons). In the solution side, the charge is made up of excess
anions or cations residing close to the electrode surface. Although only one water
molecule thick, at any given potential, the double layer is characterized by its double
layer capacitance C
dl
(10-40 µC/cm
2
).

4
1.1.1 Mechanism of Charge Injection
Before proceeding into understanding the basics of neural stimulation and
electrode characterization, it is worth noting the different terminologies assigned to the
electrodes employed, which vary depending upon the experimental conditions. For
electrochemical characterization, a three-electrode system is employed where the
electrode of interest is referred to as the working electrode, while the other two are called
the counter and reference electrodes. For neural stimulation, a two-electrode system is
employed where current is applied to the tissue through the stimulating electrode while
the return electrode is used to complete the electrical circuit. Neural stimulation can be
further subdivided into monopolar and bipolar stimulation. Monopolar stimulation uses a
small stimulating electrode and a relatively large return electrode while in bipolar
stimulation, two similarly sized electrodes are used as the source and sink in an effort to
focus the current to small regions. Electrode size range from 0.2 mm such as those used
in retinal prostheses and cochlear implants to 1.5 mm DBS electrodes (Benabid et al.,
1996; Huang and Shepherd, 1999; Mahadevappa et al., 2005; Lee et al., 2007).
Measurements may contain a third electrode termed the reference electrode required for
measuring precisely controlled electrical potentials. At equilibrium (no current), the
potential of the system remains constant and is typically referred to as the open-circuit
potential. Net electrochemical processes begin to take place as soon as the potential is
forced away from equilibrium and resulting current begins to flow through the system.
Charge transfer across the interface takes place through two primary mechanisms viz.,
Faradaic and non-Faradaic reactions.
5
Faradaic and Non-Faradaic Reactions
Non-Faradaic processes include redistribution of the charge at the electrode-
electrolyte interface and do not involve any net transfer of charge species across the
interface. If charge injection is achieved through only non-Faradaic reactions, i.e.
charging and discharging the double-layer capacitance, then the electrode-electrolyte
interface can be modelled as a simple capacitor, viz. the double-layer capacitor C
dl
. If the
total amount of charge transferred is small then the transferred charge can be recovered
by simply reversing the polarity of the applied pulse or by discharging the capacitor. In
addition to charging-discharging of the double layer capacitance, charge injection can
also be achieved by Faradaic processes such as oxidation-reduction reactions. These
reactions involve the transfer of electrons between the two phases of the reaction and
unlike the capacitive mechanism, may or may not be completely reversible in nature. In
case of reactions in which at least one of the chemical species is surface bound, the
reaction is completely reversible under steady state conditions. Such reactions are limited
by the available surface area of electrode and the amount of species adsorbed onto the
interface. However, reactions that do not involve at least one surface bound species, have
no mechanism to force the reaction to be reversible in the steady state. Charge balancing
in the Faradaic regimen is most often achieved via multiple partially reversible reactions
that result in the release of one or more possibly cytotoxic chemical substances in the
surrounding tissue. Below are examples of some reversible and irreversible Faradaic
reactions.

6
Reversible Faradaic reactions:

€
Pt +H
2
O⇔PtO +2H
+
+2e
−
(1.1)

€
Pt +H
+
+e
−
⇔Pt−H (1.2)

€
Pt +H
2
O +e
−
⇔Pt−H +OH
−
(1.3)
Irreversible Faradaic reactions:

€
Pt +4Cl
−
⇒ PtCl
4
[ ]
−2
+2e
−
(1.4)

€
2H
2
O +2e
−
⇒H
2
↑+2OH
−
(1.5)

€
2H
2
O⇒O
2
↑+4H
+
+4e
−
(1.6)

1.1.2 Electrode Characterization
In order to depolarize neurons or to record biological potentials, an interface is
required between the body and the electronic apparatus. This interface is called the
biopotential electrode. Biopotential electrodes deal with challenges different from
electrodes used in other systems. First the electrode material has to be biocompatible i.e.
non-toxic to the body and second it has to have the ability to serve as a transducer. As
mentioned in the preceding section the current in the electrode is carried by electrons
while in the electrolyte it is carried by ions, thus transduction between the charge carriers
is needed.

1.1.2.1 Electrode Potential
When a metal is brought into contact with a solution, a net rearrangement of
charge occurs at the interface leading to a loss of neutrality of charge at the interface. As
7
a result, the electrolyte in the immediate vicinity of the electrode is at a potential different
from the rest of the solution. This difference in potential is called the half-cell potential
and is determined by many different parameters such as the type of metal, the type and
concentration of ions in the solution, temperature etc. This half-cell potential is also
referred to as the electrode interfacial potential. It is not possible to measure this potential
without utilizing a second electrode. However, the second electrode would then create an
interface of its own with the electrolyte thus making it impossible to separate the two
resulting potentials from each other. To overcome this, electrochemical cells are
evaluated in their entirety, generally composed of a working electrode and a reference
electrode separated by the electrolyte. Thus a cell’s potential is defined as the potential of
the working electrode vs. the reference electrode.

Consider the reaction between a metal electrode and a redox couple in the electrolyte: -
€
O +ne
−
↔R
The equilibrium potential for any electrochemical cell can be calculated using the Nernst
equation: -

€
E
x
=
RT
nF
ln
X
[ ]
o
X
[ ]
i
(1.7)

where, [X]
o
and [X]
i
are the concentrations of the species, R is the gas constant, T is the
absolute temperature (Kelvin), F is Faraday’s constant and n is the number of electrons
transferred. For the electrochemical cell above, if the concentration of both species in
8
solution is equal then the potential of the cell will equilibrate to its formal potential E
0
.
For unequal concentrations, using the Nernst equation, the equilibrium potential for the
electrochemical cell is: -

€
E
eq
=E
0
+
RT
nF
ln
O
[ ]
R
[ ]
(1.8)

In the absence of any net current, the measured cell potential is called the open-
circuit potential, which again is the sum of the two interfacial potentials. Now if instead a
current is present, then the observed potential is different from the equilibrium potential.
This is due to the polarization of the electrode. The difference between the observed
potential and the equilibrium potential is known as the overpotential η.

€
η =E−E
eq
(1.9)
Three basic mechanisms contribute to overpotential: ohmic, concentration and
activation overpotentials. Ohmic overpotential is due to the electrolyte resistance which
leads to a voltage drop across the solution during the passage of current between the
electrodes. Concentration overpotential occurs due to changes in the distribution of ions
at the electrode-electrolyte interface. Activation overpotential occurs due to charge
transfer processes involved during oxidation-reduction reactions that are not completely
reversible. The net overpotential is simply a sum of the three mechanisms.

9
1.1.2.2 Polarizable and Non-Polarizable Electrodes
For ideally polarizable electrodes, no actual charge crosses the electrode-
electrolyte interface during current flow. Instead, during current flow, redistribution of
ions occurs at the interface thus exhibiting capacitor like properties. As a result the
overpotential is dominated by the concentration overpotential. One material that
approaches an ideal polarizable electrode is titanium nitride, where charge injection takes
place through capacitive charging-discharging processes. Noble metals such as platinum
also behave as polarizable electrodes but over a limited range of voltages. Ideally non-
polarizable electrodes on the other hand are the ones in which current passes freely
between the electrode-electrolyte interface and hence causes no overpotential. Electrodes
such as silver-silver chloride and saturated calomel come closest to behaving as non-
polarizable electrodes. These electrodes are best used as reference electrodes during
measurement of electrode potential as there is no change in voltage across their interface
during current flow. However, it is essential to note that in reality no electrode behaves
either as ideally polarizable or ideally non-polarizable. Electrodes come closest to ideal
characteristics only over a limited range of voltages.

1.1.2.3 Measurement of Impedance
Electrochemical impedance spectroscopy (EIS) has been used successfully to
characterize the electrode-electrolyte interface. Specifically for neuroprostheses
employing current stimulation, impedance measurement techniques have been employed
to test the efficacy of neural stimulation.
10
As it is not possible to control the tissue properties of the target system, efforts are
made instead to control the electrode design in order to allow safe and effective
stimulation. To achieve this, equivalent circuit models of the electrode-electrolyte
interface have been developed and the model parameters can be estimated by EIS
techniques. The first ever model was proposed by Warburg in 1899. He modeled the
interface as a polarization resistance in series with a polarization capacitance. This would
produce a straight vertical line on the complex plane plots (Z imaginary vs. Z real).
However, for solid electrodes it was often observed that the straight vertical line had an
angle less than 90°. Thus, the electrode impedance consisted of a polarization resistance
in series with complex impedance exhibiting frequency dependency. Fricke first showed
the phenomenon of constant phase angle. The impedance associated with it is termed as
the constant phase element (CPE). CPE is thought to arise from surface inhomogeneities
and slow reaction kinetics (Brug et al., 1984). Mathematically, CPE is represented as: -

φ
ω) (
1
j T
Z
CPE
= (1.10)
Where T is a constant in F cm
-2
s
-1
and φ is related to the angle of rotation of a straight
capacitive line on the complex plane plots. The CPE is often used to represent a “leaky
capacitor” and only when φ = 1, T = C
dl
, a purely capacitive behavior is obtained (Lasia,
2002). Equation (1.10) can be used to describe a pure resistor for φ = 0 and a pure
inductor for φ = -1. However, these models did not include the passage of direct current
through the interface including the popular Randles model. It was in 1968 that Geddes
and Baker showed via numerous impedance-frequency curves in saline and living
11
subjects the importance of the Faradaic resistance to account for the passage of direct
current through the interface. Their model consisted of the Warburg elements
(polarization resistance and capacitance) in parallel with the Faradaic resistance (Geddes,
1997). Since then numerous versions of equivalent circuit models have been described
for the electrode-electrolyte interface depending upon the specific application. For neural
stimulation, the model most often employed consists of a constant phase element to
represent the interface capacitance shunted by a charge transfer resistance (R
CT
) in series
with the solution resistance (R
s
). Some models also include a Warburg element along
with the charge transfer resistance that comes into play at low frequencies during
diffusion-controlled processes (Fig. 1.1). Using different versions of this model studies
have been done to characterize different electrode materials and their surfaces (Weiland
and Anderson, 2000; Germain et al., 2004; Franks et al., 2005). As platinum is the most
widely used electrode material for biomedical applications, groups have focused on
extensively characterizing its properties. Frank et al. used EIS techniques to compare 3
electrode materials geared towards biomedical applications: platinum, platinum black and
titanium nitride (Franks et al., 2005).

12

Figure 1.1 Equivalent circuit model of the electrode-electrolyte Rs: solution
resistance; R
CT
: charge transfer resistance; Z
W
: Warburg element; Z
CPE
: constant
phase element.

The electrochemical impedance theory describes the response of a system to an
alternating current or voltage input as a function of frequency. The basic approach of EIS
is to apply small amplitude sinusoidal current or voltage perturbations to the electrodes
and measure the system’s voltage or current response, respectively. In most studies of
bioelectrode stimulation, a sinusoidal voltage signal (5 ~ 10 mV) is used as the excitation
signal and the resulting current is measured as the response of the system (potentiostatic
EIS). Typically the single-sine technique is used where the excitation signal is applied at
discrete frequencies and the resulting response signal is measured at each frequency to
develop the impedance spectrum.
The impedance profiles of microelectrodes assist in developing circuit model
analogues of the electrode-tissue system models, such as the one described in the
preceding paragraphs. Values of the different circuit elements can be estimated from the
impedance measurements, which in turn provide some indication on electrode
performance. The profiles are viewed either through the ‘Nyquist plot’ or the ‘Bode plot’
and corresponding model parameters can be estimated. At high frequencies, the
13
impedance of the equivalent circuit (Fig. 1.1) becomes almost entirely dominated by the
solution resistance R
s
while at low frequencies the resistance of the electrochemical
reaction R
CT
also comes into play. The solution resistance has long been shown to have
an inverse relationship with the radius of a disc electrode. However, recent work by
Ahuja et al. suggests that this dependence may not hold for all frequencies (Ahuja et al.,
2008). Their work showed that the electrode impedance does scale with radius but only in
the high frequency regime (∼ 100 kHz), whereas at lower frequencies (∼ 10 Hz) it scales
with the area of the electrode. Thus, only the electrode edge contributes at higher
frequencies due to the primary current distribution while at lower frequencies, a
secondary current distribution comes into play that drives the current to the centre of the
disk leading to an area dependence. They also showed that for microelectrodes of radii
less than 50 µm, the area dependence is exhibited even at relatively higher frequencies
due to the decreased RC time constant and double layer charging of the electrodes at
these frequencies.

1.1.2.4 Surface Reactions and Potential Limits
Cyclic voltammetry (CV) falls under the class of voltammetric methods where the
electrode potential is controlled and the resulting current is measured. In voltammetric
methods, solutes in contact with the electrode undergo oxidation or reduction reactions
producing current at the electrode surface that is measured. In case of cyclic
voltammetry, the applied potential is linearly varied with time (cycled) while the
resulting current is measured. The applied potential has a triangular waveform with
14
negative and positive turn-around potentials. Since, in a cyclic voltammogram, the range
of applied potential is quite large, the measured current aids in understanding the reaction
mechanisms available during stimulation. CV can characterize the potential at which the
reaction proceeds maximally, the reaction kinetics, and the reversibility of the reaction,
all of which are critical to determining if this reaction can be safely used to transfer
charge to tissue. Platinum by far has been the most well studied electrode material (figure
1.2) but with increasing demands of neural stimulation treatment strategies, focus has
shifted towards analyzing and characterizing other candidate electrode materials such as
iridium oxide, titanium nitride etc.

Figure 1.2 Cyclic voltammogram of poly crystalline platinum in 1M KOH at
scan rate of 100 mV/s exhibits all the different processes involved during the
cathodic and anodic direction (Hamann et al., 1998). The potential scale is referred
to a reversible hydrogen electrode (RHE) in the same solution.

15
For microelectrode characterization in neural stimulation applications, CV plots
are used to study a number of important parameters associated with the safe and effective
charge-injection at the electrode-tissue interface. Some of these have been described
below.

Voltage limits
Electrodes used for chronic stimulation of tissue are typically restricted to operate
within the “water window”, the term given for the potential range between hydrogen
evolution potential (negative) and oxygen evolution potential (positive).

€
2H
2
O +2e
−
→H
2
↑+2OH
−
(1.11)

€
2H
2
O→O
2
↑+4H
+
+4e
−
(1.12)
Cyclic voltammetry is used to determine these voltage limits, which are material
and solution dependent. During neural stimulation only reversible reactions are employed
for charge injection, to avoid causing damage to either the electrode or tissue. For
example, Robblee at al. showed that activation of iridium by repetitive cycling of
potential between 0.05 and 1.45 V versus RHE (reversible hydrogen electrode) increased
the charge storage capacity of the electrode (Robblee et al., 1983). More recently, cyclic
voltammograms of IrOx and TiN in PBS exhibited a water window in the voltage range
of -0.7V to 0.8V for IrOx and -0.6V to 0.8V for TiN versus a saturated calomel electrode
(Weiland et al., 2002a). Note that the water window does not change in width, but can
shift depending on the reference potential.

16
Charge-injection Mechanism
In a CV plot, the presence of peaks indicates electrochemical reactions occurring
at the electrode-electrolyte interface along with the charging-discharging of the double
layer capacitance. As an example, the CV plot of platinum exhibits distinct peaks
associated with the different surface reactions such as hydrogen-atom plating (Fig. 1.2).
Voltammograms of IrOx exhibit distinct peaks indicating reduction-oxidation reactions
involving transfer of electrons across the interface, along with current flow due to
capacitive charging-discharging (Robblee et al., 1983; Weiland et al., 2002a; Cogan et
al., 2006). On the other hand, the CV traces of TiN show no distinct peaks indicating that
the current flow is dominated by the capacitive charging-discharging mechanism
(Weiland et al., 2002a). Also, from the nature of peaks, the type of reaction that is
occurring can be determined. For example along with the oxidation-reduction reaction
peaks, the presence of additional peaks indicates existence of other electro-active
substances.

Charge-storage capacity
An important parameter for neural stimulation is the charge-storage capacity of
the electrode. This is determined by integrating the area under either the cathodic or
anodic sweep in the CV plot within the water window. The value obtained indicates the
maximum charge that can be injected via reversible surface processes by an electrode.
This is usually expressed in terms of charge density limit of the electrode. As an example,
the charge storage capacity of activated iridium oxide has been reported to range from 10
17
mC/cm
2
to 240 mC/cm
2
depending upon the thickness of the film (Stieglitz, 2004).
However, it should be noted that this is the maximum possible charge capacity of the
electrode. The actual amount of charge injection that can be achieved during neural
stimulation is usually only a fraction of this charge storage capacity and depends upon
factors such as the thickness and morphology of the film, specific reactions of the redox
material, pulse duration, etc.

Reversibility of reaction
Whether the electrochemical reaction occurring is reversible or irreversible in
nature can be determined from the cyclic voltammogram of the electrode. All chemical
reactions, including reactions occurring at the electrode-electrolyte interface, proceed at a
finite rate. The reversibility of a reaction is thus governed by the rate of electron transfer
and surface concentrations. In a perfectly reversible reaction, the cathodic peak height is
equal to the anodic peak height and the reversible half-wave potential will lie exactly
midway between the peaks. However, as the reaction becomes more and more
irreversible, the cathodic peak height no longer remains equal to the anodic peak height
and the separation between the peaks increase (Fig. 1.3). This situation can occur at high
scan rates where due to slow reaction kinetics, the voltammogram changes from
reversible to irreversible shape.
18

Figure 1.3 Transition of cyclic voltammograms from reversible to irreversible
domain. Case 1: reversible reaction with equal cathodic and anodic peak heights.
Case 2: transition from reversible to irreversible reaction with increasing separation
between cathodic and anodic peaks. Case 3: Irreversible reactions indicated by large
separation between the cathodic and anodic peaks. Modified from Montenegro et
al. (Montenegro et al., 1991).

1.1.2.5 Voltage Response
While charge storage estimates acquired from CVs give the maximum charge
value that the electrode in question can store without causing hydrolysis, the actual
amount that is injected during current stimulation is quite different. Hence, in order to get
a comprehensive picture of how the electrode will behave during active stimulation, one
must study the voltage response developed during current controlled stimulation.
Whenever a current pulse is applied across the electrode-electrolyte or electrode-tissue
interface, a voltage develops across the interface. This voltage waveform is characterized
by an initial potential step that is known by several different terms such as ohmic drop, iR
drop, access voltage and series resistance. Irrespective of its name, this initial voltage
19
drop results from the ohmic losses in the system due to resistance of the electrolyte or
tissue and is modelled as the solution resistance (R
s
) in equivalent circuit model (Fig.
1.1). These iR losses do not contribute to the potential difference across the interface that
drives the charge across the interface. Hence, before performing analysis of the potential
transients, it is essential to subtract these losses from the total voltage response. Based on
the net potential across the electrode for a given pulse amplitude and duration, estimates
of the actual charge injection capacities can be made. Also, by monitoring the voltage
drop across the electrode, the safe charge injection limits can be estimated for voltage
drops that do not exceed the water window of the electrode.

1.1.3 Overview of Extracellular Stimulation
The bilipid layer membrane separates the intracellular region of the cell from the
extracellular environment and acts as a barrier to the movement of ions between these
two regions. It plays a crucial role in determining which ions are allowed to pass through
and hence has the important properties of specificity and selectivity. The membrane also
includes two specialized regions, the afferent region at which the neuron receives the
signal and the efferent region at which the neuron sends the signal.
All cells have a resting transmembrane potential (from hereon referred to as
membrane potential) with the interior of the cell negative with respect to the exterior of
the cell. This membrane potential is dependent on the concentration of the ionic species
such that the equilibrium potential of each ion differs from the membrane potential. In
general, the ions of interest are K
+
(potassium), Na
+
(sodium) and Cl
-
(chloride). At rest,
20
concentration of K
+
ions is higher inside giving it a negative equilibrium potential
compared to the membrane potential. This gradient tends to move the ions out of the cell.
The concentration of Na
+
ions on the other hand is higher outside than inside the cell
giving it a positive equilibrium potential, which causes them to move into the cell. At
rest, the membrane acts as a barrier and is less permeable to Na
+
ions compared to K
+

ions. The concentration ratios of these ions are maintained by ionic pumps that force the
movement of each of the ions in opposite directions thus maintaining a constant charge
separation across the membrane and keeping the cell at its resting membrane potential. A
typical value of the resting membrane potential is -60 mV measured inside the cell with
reference to the outside. Using equation 1.7, the equilibrium potential associated with
each of the ions is: -

€
E
x
=
RT
zF
ln
[X]
o
[X]
i
(1.13)

where, z is the valence of the ion. Although the membrane potential is dependent upon
the ionic fluxes, it is not equal to any of their equilibrium potentials. Instead, the
membrane potential of the cell is determined by the concentrations of the ions inside and
outside the cell along with the ease with which each of the ions can cross the membrane,
i.e. on the conductivity and permeability of the membrane to the specific ions. The
Goldman equation describes quantitatively the dependence of the membrane potential at
steady state on ionic concentration and permeability (P): -

€
V
m
=
RT
F
ln
P
K
[K
+
]
o
+P
Na
[Na
+
]
o
+P
Cl
[Cl
−
]
o
P
K
[K
+
]
i
+P
Na
[Na
+
]
i
+P
Cl
[Cl
−
]
i
(1.14)
21

As mentioned previously, the cell membrane is selectively permeable to certain
ionic species. This is possible due to the presence of ion channels that are pore-like
structures spanning across the membrane. As an example, the potassium channels remain
open causing a leak of K
+
ions out of cell, contributing to the negative membrane
potential. During neuronal signalling, the membrane potential rapidly changes in
response to some stimulus. The stimulus may be external or internal, but the stimulus
changes the membrane permeability and thus membrane potential. Cell activation occurs
when this initial change in potential results in membrane potential becoming more
positive (depolarization). This leads to the opening of voltage-gated sodium ion channels
causing a further positive increase in membrane potential. Initially the cell’s response is
proportional to the stimulus strength, i.e. the cell responds as a graded potential. Once the
membrane potential crosses threshold, the cell responds by generating an action potential
that propagates down the cell’s axon all the way to its axon terminals. The axon terminals
in turn connect to other cells (through synapses) thereby activating them and thus
initiating a signalling cascade. The action potential is described as an all-or-none
phenomenon, i. e., once initiated it will actively propagate down the axon irrespective of
the presence of the initial stimulus. Typically, action potentials last for about a
millisecond after which the cells return to their resting state through the inactivation
(closing) of voltage-gated sodium channels and activation of voltage-gated potassium
channels. These two mechanisms have longer time constants compared to sodium
activation but work together to bring the cell back to its resting membrane potential.
22
Action potentials are followed by a brief period of diminished excitability termed as the
refractory period. The refractory period can be further divided into two phases. The
absolute refractory period occurs immediately after the action potential and during this
period, the cell cannot be excited even with a very large stimulus. This phase is followed
by the relative refractory period during which the cell can be excited however with a
much larger stimulus than that required normally to reach threshold. These periods of
refractoriness are caused by the residual inactivation of Na
+
channels and increased
opening of K
+
channels and typically last for only few milliseconds.
Electrical stimulation of excitable tissue generates action potentials that in turn
initiate neuronal signalling and enable partial restoration of lost functionality in sensory
or motor systems. This process requires the extracellular region to be driven more
negative by applying a rapid negative charge injection via an extracellular stimulating
electrode. For the simplest case of stimulation, a single electrode is placed near the
excitable tissue and the electrode is driven as a cathode causing the outside of the
membrane to become more negative. This causes the membrane potential to become
positive thus leading to a net reduction in the membrane potential (depolarizing the
membrane). If on the other hand, the stimulating electrode is driven as an anode, then it
will cause the outside of the cell to become more positive than the inside thus causing the
membrane potential to become more negative. This will lead to a net increase in the
membrane potential causing the membrane to hyperpolarize. Since a current generator
must have a source and a sink, during extracellular stimulation, a second electrode is
required for the current loop to be complete. This second electrode is usually called the
23
return electrode and based upon its size and position can cause a number of different
events to occur. If the return electrode is much larger than the stimulating electrode, then
the current density is highest at the stimulating electrode causing excitation of neurons
near it. However, if the return electrode is similar in size as the stimulating electrode,
then the current density at both sites will be the same and hence neuronal excitation can
occur at both sites. In this case, during cathodic stimulation, the neurons in close
proximity to the stimulating electrode are depolarized while those underneath the return
electrode are hyperpolarized. In some cases this hyperpolarization may be large enough
to suppress an action potential initiated near the electrode (anodic surround block)
(Rattay, 1989; Plonsey and Barr, 1991). On the other hand, if anodic stimulation is
employed then the neurons near the stimulating electrode will be hyperpolarized while
those near the return electrode will be depolarized. In this case, the action potentials are
initiated in regions distant from the electrode known as virtual cathodes. The
depolarization that occurs through anodic stimulation is about a seventh to a third of that
accomplished through cathodic stimulation although this depends upon the electrode
position (Rattay, 1999). Thus, cathodic stimulation requires less current to cause a cell to
cross threshold and initiate action potentials. The stimulation protocols described above
may be effective at selectively activating one population of neurons without activating
neighbouring neurons.
Activation thresholds are usually defined in terms of the amount and duration of
the current pulse needed to cause the excitation. Another way to define excitation
thresholds is in terms of the applied charge that is simply a product of amplitude and
24
duration. Since in neural stimulation, currents applied are in the range of microamps and
are applied typically for a few milliseconds, the charge delivered ranges from a few
microcoulombs to a few nanocoulombs. By far, the best known law of stimulation is the
one by Lapicque that relates the threshold current (I) required for stimulation to the
duration (d) of the applied pulse (Lapicque, 1909). He introduced the tissue specific
excitability parameter called the chronaxie (c) and defined it as the pulse duration that
required twice the rheobase current (b). Here, rheobase current is defined as the threshold
current (I) for very long pulses. Mathematically, b is the limit of I, as pulse duration goes
to infinity. The Lapicque law for stimulation is: -

€
I =b(1+c/d) (1.15)

Based on the above equation, strength-duration curves can be plotted to
graphically illustrate the relationship between the three parameters I, d and c. The
strength-duration curve is an essential tool in all types of studies where electrical
stimulation of excitable tissue is employed. Studies have shown how different parameters
can be calculated from these curves including charge and energy-duration relationships
(Geddes and Bourland, 1985). Although numerous studies illustrate chronaxie values of
different excitable tissues, the accuracy of the measurements can be affected by factors
such as the electrode characteristics, tissue inhomogeneity, stimulus waveform etc.
(Geddes, 1999, 2004). Studies in motor nerves and different types of muscle have shown
the dependence of chronaxie on different parameters such as temperature and location of
electrodes (Geddes, 1999).
25

Figure 1.4 Strength-duration graph illustrating threshold current required to
elicit response at different pulse durations. Rheobase current = b; chronaxie = c.
Modified from Geddes et al. (Geddes, 2004).

Another way to define the relationship between stimulus strength and excitation is
through amplitude-intensity function. This is typically used where the response is an
evoked potential and generates a plot of the stimulus strength at fixed pulse duration
against the amplitude of the evoked response. It helps in determination of true threshold
by simply extrapolating the curve to intersect the x-axis. Amplitude-intensity functions
are useful because neural prostheses typically operate above threshold to provide a range
of sensation or activation. Finally for the case of single units, analysis methods such as
post-stimulus time histograms (PSTHs) are employed that sort the individual spikes
based on their latencies. More sophisticated analyses of a mixture of action potentials
produced by multiple cells involve grouping the individual spikes based on their
individual waveform characteristics.

26

Figure 1.5 Representative graph illustrating the gradual increase in response
amplitude as the stimulus strength is increased. The amplitude of response is usually
measured in microvolts (µV) while the applied stimulus amplitude is usually in
microamps (µA).

It is worthwhile to mention that some neuroprosthetic systems such as deep brain
stimulation (DBS) and transcutaneous electrical nerve stimulation (TENS) use voltage
source to deliver charge in order to cause neuronal excitation (Benabid et al., 1996;
Perlmutter and Mink, 2006; Lee et al., 2007; Li and Mogul, 2007; Walsh et al., 2009).
This method of stimulation involves a direct connection of a voltage source between the
working and counter electrode, thereby controlling the net potential between the two
electrodes. In this mode of stimulation, the current is maximum at the beginning of the
applied voltage pulse due to the charging of the double layer capacitances of the two
electrodes. With a longer duration pulse, the applied voltage maintains a steady state
Faradaic current. Although this mode of stimulation requires simpler circuitry, there are
several disadvantages compared to current-controlled stimulation. Given the fact that
neuronal excitation is related to the amount of charge delivered, the main disadvantage of
27
using this method is that there is no direct control on the amount of current delivered.
Thus, a change in resistance anywhere in the conduction path or a change in tissue
properties will change the actual amount of current delivered and influence stimulation
efficacy.

1.1.4 Safe Stimulation of Tissue
A neural stimulation system that is not properly designed can cause damage to the
tissue or to the electrode itself. For such a system to be successful, it must elicit the
required neuronal excitation without causing any damage to the biological system.
Electrode shape, size and material along with stimulus pulse parameters need to be
carefully chosen to meet the requirements of the system. Extensive work has been done
in defining the role of all the different parameters that determine the safety limit of the
tissue and electrode.

1.1.4.1 Mechanisms of Neural Injury
There are several mechanisms that may cause neural injury; they are broadly
categorized into two main classes. The first mechanism of damage is associated with the
electrochemical processes through which the stimulus current is injected into the target
tissue. Damage is induced due to formation of toxic electrochemical reaction products
during stimulation at a rate greater than what can be tolerated by the physiological
system. As described in the section on electrode characterization, toxic products of
electrochemical reactions are generated once the safe limits of the electrode are exceeded.
28
Studies have shown that neural damage can be induced due to the generation of these
electrochemical reaction products. The threshold for neural damage has been shown to be
higher for the case of monophasic stimulus pulses as compared to biphasic stimulus
pulses (Lilly et al., 1952; Lilly et al., 1955; Mortimer et al., 1970; Pudenz et al., 1975).
An early study by McCreery and colleagues showed that continuous stimulation
of the cortex using Faradaic (platinum) and capacitor (anodized sintered tantalum
pentoxide) electrodes caused similar type of neural damage (McCreery et al., 1988). This
led them to conclude that at the charge densities applied in their study, neural damage
was mainly induced by the passage of stimulus current rather than electrochemical
reaction products as those would be formed only by the platinum (Faradaic) electrodes.
Thus, a second mechanism of neural injury is associated with the excitation of tissue
(McCreery, 2004) and is multi-factorial and complex. Different factors contributing to
damage via this second mechanism have been put forward. Out of the different factors,
one involves the metabolic stresses induced on the tissue causing a transient or permanent
elevation of neurotransmitter release (excitotoxic effect). Work done by Agnew et al. has
shown that prolonged electrical stimulation of cat cerebral cortex at defined charge
density and charge per phase levels can lead to neural damage in the form of shrunken,
hyperchromic neurons (Agnew et al., 1993). However, on the application of MK-801, a
non-competitive N-methyl-D-aspartate (NMDA) receptor antagonist, only slight to no
neural damage was observed depending upon the MK-801 dosage applied prior to
prolonged stimulation. Thus the study showed that prolonged electrical stimulation can
cause neural injury via glutamate-induced excitotoxic damage through the NMDA
29
receptor pathway. Support for this theory also comes from studies showing that blocking
of NMDA receptors appear to provide some protection to neurons in the hippocampus
and neocortex from ischemia or hypoxia (Ashton et al., 1986; Meldrum, 1995). Other
factors include large depolarizations and hyperpolarizations induced by the voltage
gradient (membrane electroporation). Electroporation has been shown to cause cellular
damage due to dielectric breakdown of the bilipid membrane (Needham and Hochmuth,
1989; Krauthamer et al., 1991; Tovar and Tung, 1992) and can also cause damage to the
voltage-dependent gating mechanisms of ion channels (Chen et al., 2006).
Electroporation of neural tissue has also shown to render the neurons refractory to
subsequent electrical stimulation (Bonnot et al., 2005). Another phenomenon known as
spreading depression has been shown in both retina and cortex (Leão, 1944; Gouras,
1958). The process is characterized by an expanding wave of depression of spontaneous
electrical activity that travels across the cortex and retina. It causes profound depression
of the neurons and also causes negative shifts in the resting membrane potential.
Although the process is spontaneous, it can be induced by trauma and weak electrical
stimulation.

1.1.4.2 Parameters for Safe Stimulation
One of the well-established principles of neural stimulation is to achieve charge
balancing during stimulation between the different phases of the stimulus pulse. This was
first reported by Lilly in 1961 and ensures that the total net charge during stimulation at
the electrode-tissue interface is zero (Lilly, 1961). If charge balancing is not
30
accomplished, then a net accumulation of charge will ultimately lead to the rise of
electrode potentials to levels where water hydrolysis will start. For monophasic
stimulation, charge balancing is accomplished by the use of a blocking capacitor that
slowly discharges after the application of the pulse. Although charge balancing ensures
that there is no net accumulation of charge, it does not guarantee safety. Such stimulus
waveforms may momentarily exceed the established safety limits of total charge, charge
density or electrode potential. Classically, safety limits for neural stimulation have been
divided into two broad categories: -
1. Neural damage limits dictated by the ability of biological tissue to withstand
electric current without any permanent changes to the neural tissue.
2. Electrochemical limits based on the ability of the electrode to store or dissipate
electric charge without exceeding the water window, outside of which formation
of harmful products start.

While neural injury limits are defined in terms of both charge density and charge
per phase, electrochemical limits are defined in terms of charge density only. Charge
density is simply the total charge per unit area of electrode and determines the magnitude
of the depolarization or hyperpolarization induced in the neurons and axons close to the
electrode. Charge per phase is the amount of charge injected during each phase of the
stimulus pulse and determines the distance over which the applied stimulation can
activate the neurons i.e. the number of neurons activated. McCreery and colleagues have
shown that charge density and charge per phase act synergistically to determine the safe
31
or unsafe levels of stimulation (McCreery et al., 1990). They showed that neural damage
is induced with low charge per phase but high charge density, as is often the case for
microelectrodes. Based on these data delineating the boundary between safe and unsafe
charge injection for different charge and charge density levels, Shannon (Shannon, 1992)
developed the following empirical relationship: -

€
log(D) =k−log(Q) (1.16)

where, D is the charge density in µC/cm
2
/phase and Q is the charge per phase in
µC/phase. The equation describes a family of lines for different values of k. The line for
which k = 1.5 describes combinations of charge density and charge per phase values for
which no damage was observed. Merrill et al. have graphically summarized the work of
both studies and also included results of other studies assessing safety of neural
stimulation (Fig. 1.6).

Figure 1.6 Charge (Q) versus charge density (Q/A) for safe stimulation. Different
symbols indicate results of different studies (Merrill et al., 2005).
32
Along with charge density and charge per phase, other stimulus parameters such
as frequency of stimulation, duration etc. have been found to play an important role in
determining the presence or absence of neural damage. McCreery and colleagues
demonstrated the effect of stimulus frequency as a parameter in causing injury during
peripheral nerve stimulation (McCreery et al., 1995). Their study showed that continuous
stimulation of the cat sciatic nerve for 8 hours over 3 days causes the myelin sheath to
collapse into the axonal space leading to early axonal degeneration (EAD). The threshold
of neural injury decreased with increasing stimulus pulse frequency.
Most of the aforementioned studies have employed single electrode stimulation.
However, a recent study (McCreery et al., 2002) found that in the case of multi-electrode
stimulation, both sequential and simultaneous stimulation at levels previously found to be
safe create transient depression in the resulting neural response. One theory put forward
by the authors is the creation of overlapping electric fields that cause certain neurons to
be driven at rates higher than what is actually being delivered. The authors dubbed the
observed effect “SIDNE” (stimulation induced depression in neuronal excitability).

1.2 Anatomy and Physiology
1.2.1 Structure and Function of Retina
All vertebrate eyes are based on a common structural plan in which rays of light
enter through the cornea and are focussed by both the cornea and lens to form a real
inverted image on a thin sheet called the retina at the back of the eye (Fig. 1.7).
Historically the retina was viewed simply as a sheet of film analogous to the film in a
33
camera. However, it is now known that the retina is a neuroepithelial tissue comprising of
light-sensitive neurons, neurons involved in visual processing and the retinal pigment
epithelium. Embryologically derived from the neural tube, the retina is part of the central
nervous system (CNS) and thus has been characterized by Dowling as an approachable
part of the brain (Dowling, 1987). The retina has been divided into two parts based on the
major cells types and functions. Rodieck described the retina to be comprised of the
retinal epithelium, which is a single layer of columnar epithelial cells and remaining
retina as the sensory retina comprised of neurons and glia (Rodieck, 1973). Another way
of dividing the retina is into a sensory component consisting of the light-sensitive
photoreceptors cells involved in the phototransduction process and the neural retina
consisting of neurons that carry out the first steps in processing of visual information.
None of the aforementioned terminologies change the striking feature of retina, which is
the layered organization of cell bodies and processes into distinct layers with the depth
within the retina determining the synaptic connections. Light rays pass through the entire
retina before reaching the pigment molecules of photoreceptors to excite. The transduced
signal is then passed on the neurons in the inner retina and ultimately sent to higher
centres of visual processing. The outer or distal retina refers to the scleral side while inner
or proximal retina refers to the vitreal side of retina.

34

Figure 1.7 Horizontal section of the right human eye (Rodieck, 1973)

Functionally, the retina consists of three layers of neurons (Fig. 1.8). The outer
nuclear layer (ONL) comprises of the cell bodies of all photoreceptors, both rods and
cones. The inner nuclear layer (INL) is comprised of cell bodies of horizontal cells,
bipolar cells and amacrine cells. In addition, cell bodies of radial glial cells (Müller cells)
are also found in this layer. The ganglion cell layer (GCL) comprises of cell bodies of
displaced amacrine cells and ganglion cells. Information from the retina is transmitted
35
from this layer to the brain via the optic nerve. The three neuronal layers are separated by
two synaptic or plexiform layers, which consist of vast majority of fine dendrites and
synapses. The outer plexiform layer (OPL) lies between the ONL and INL and consists of
synaptic connections formed between the photoreceptors and neurons found in the INL.
The inner plexiform layer (IPL) lies between the INL and GCL and comprises of synaptic
connections between the neurons found in the INL and ganglion cells.

Figure 1.8 Schematic of the different cell types and retinal layers (A);
Hematoxylin & Eosin (H & E X 640) stained retinal section exhibiting the ten
distinct histological zones of the retina; VB = vitreous body (Wheater et al., 1982).

Neural communication in the retina takes place either via chemical
neurotransmission or via gap junctions. Chemical neurotransmission involves the release
36
of a neurotransmitter from the presynaptic neuron, which then diffusses across a narrow
synaptic cleft to bind to a variety of postsynaptic receptors. The neurotransmitters and
receptors are often linked to voltage-gated ion channels, the opening and closing of which
produces electrical changes in the cells. In the retina, the dominant neurotransmitter is
glutamate. In addition, other amino acid neurotransmitters such as GABA (γ-
aminobutyric acid), glycine and acetylcholine (ACh) play an important role in neural
signalling. In the retina, in addition to chemical neurotransmission, many neurons have
been found to be connected directly via electrical synapse known as gap junctions. These
are narrow gaps between closely apposed cell membranes and are composed of docked
hemichannels of connexons. Each connexon is built from six identical protein subunits
called connexins surrounding a central pore. It is this central pore that forms an
intracellular channel between the connected cells and allows the passage of ions and
small molecules. Different gap junctions are selectively permeable to different molecules
and are differentially modulated by various signalling molecules such as cAMP and
cGMP.

Retinal Pigment Epithelium
The retina consists of a variety of cell types. The cells of the retinal pigment
epithelium (RPE) contain pigment particles that absorb light not caught by the
photoreceptors. The RPE is involved in numerous other functions such as they assist in
the phagocytosis of rod outer segments and act as an electrical barrier by blocking the
37
passage of ions intracellularly. The RPE also provides a steady stream of a vital
molecule, retinal or vitamin A, required in the phototransduction process (Kolb, 2003).

Photoreceptors
Mammalian retina consists of two types of photoreceptors of which rods account
for 95% of all photoreceptors. Rods are numerous densely packed cells with slender outer
segments specialized for high sensitivity in dark or starlight conditions. Cones on the
other hand account for only about 5% of all photoreceptors and are larger with tapering
segments located in the top row of the ONL. Most mammalian species are dichromatic
containing medium (green) and short (blue) wavelength sensitive cones while humans
and primates are trichomatic containing an additional long wavelength sensitive cone
(red). They provide high acuity, colour vision during daylight when there is an abundance
of photons. Together rods and cones cover a wide range of intensity of around 10 log
units from the darkest night to bright sunlight. In most mammalian retinas cones are not
uniformly distributed and are often concentrated in specialized regions. In human retina,
highest density of cones of around 150,000 per mm
2
can be found at the fovea outside of
which it falls fairly rapidly into relatively uniform distribution. The peak rod density can
be found in a ring around the fovea at about 18 degrees from the foveal pit (Fig. 1.9). The
density at the fovea is approximately 100 times that of the periphery and is the region of
maximum visual acuity. Analogous to the fovea, in species such as cats and dogs, highest
density of cones are found in the area centralis while in the retinas of rabbits, squirrels
and turtles, found in a long horizontal strip called the visual streak. Together rods and
38
cones produce graded responses to photons of light modulated around the mean light
level. When a photon is absorbed, visual pigment is activated and a cascade of
biochemical events is triggered. This known as phototransduction and the transduced
light signal is sent to neurons postsynaptic to photoreceptors.

Figure 1.9 Density of rods and cones in human retina (Rodieck, 1973).

Horizontal cells
Horizontal cells are laterally extensive interneurons that are located in the outer
row of the INL and respond to diffuse light with a large hyperpolarization. In most
mammalian retinas, two types of horizontal cells can be found: axon-less and axon-
bearing. Each horizontal cell receives input from many cones and provides the necessary
feedback to create centre-surround organization of bipolar cells.

39
Bipolar cells
Bipolar cells are excitatory interneurons with glutamate as the neurotransmitter
and receive input from photoreceptors and transmit their signal to the inner retina. There
are around 10 types of cone bipolar cells and a single type of rod bipolar cell. Cone
bipolar cells are further divided into OFF cone bipolar cells and ON cone bipolar cells.
OFF cone bipolar cells contact every cone within their dendritic field at basal synapses.
The terminals of these cells ramify in the top half, sublamina a, of the IPL where they
synapse with OFF ganglion cells. Like cones, they are hyperpolarized by light and their
responses are modulated by ionotropic glutamate receptors. On the other hand, ON cone
bipolar cells are depolarized by light with their responses modulated by metabotropic
glutamate receptors. The dendrites of ON cone bipolar cells invaginate the cone pedicles
and approach the synaptic ribbons in a central position. The terminals of these cells
ramify in the bottom part, sublamina b, of the IPL where they synapse with ON ganglion
cells. In primate retina, the central retina is dominated by midget bipolar cells, both ON
and OFF. These cells have very small dendritic fields and receive inputs from single
cones and make output to single midget ganglion cells. There is only one
morphologically distinct type of rod bipolar cell. Rod bipolar cell dendrites receive input
from many rods resulting in a high convergence that contributes to the sensitivity of the
rod pathway. Each rod bipolar cell gives rise to a long slender axon that descends to
sublamina 5 of the IPL. There are only ON type of rod bipolar cells, which is consistent
with the presence of metabotropic glutamate receptors and stratification in the inner part
of the IPL. Rod bipolar cells usually do not make direct contact with ganglion cells and
40
instead pass on their output to amacrine cells which in turn relay the output to the
ganglion cell via the cone pathway.

Amacrine cells
Amacrine cells are mostly inhibitory interneurons located primarily in the inner
row of the INL. Some displaced amacrine cells are also found in the GCL. In fact in
mouse retina more than half the cells in the ganglion cell layer are displaced amacrine
cells. Two major connections are made by these cells: feedback inhibition to bipolar cell
terminals and feed-forward inhibition to ganglion cells. Amacrine cell types range from
22 to 30 different morphological types based on gross morphology, dendritic field size
and depth of stratification in the IPL. The most numerous type is the AII amacrine cell
which accounts for around 11% of the total. It has distinct bistratified morphology with
inputs received from rod bipolar cells and output coupled via gap junctions to ON cone
bipolar cells. On the whole, amacrine cells are involved in feedback inhibition, surround
inhibition, some forms of adaptation, signal averaging and noise reduction (Massey,
2005).

Ganglion cells
Output from the retina is relayed to higher order centres of visual processing via
ganglion cells. They receive inputs from bipolar cells via ribbon synapses and from
amacrine cells via conventional synapses and gap junctions. Like bipolar cells, there are
ON and OFF ganglion cells receiving inputs from ON and OFF bipolar cells respectively.
41
In addition, there are a few ON/OFF ganglion cells such as the directionally selective
type that have a bistratified morphology. Axons of ganglion cells run along the vitreal
surface and gather to form the optic nerve. Unlike other retinal neurons that respond in a
graded fashion, ganglion cells respond via action potentials, which are an all or none
phenomenon. Information about the visual scene is encoded in the spiking rate that is
relayed to higher visual centres via the optic nerve. The site at which the optic nerve
leaves the eye is known as the blind spot. This is because this region has no
photoreceptors and consequently any image falling on this region will not be seen. The
blind spot is always nasal to the fovea. To date, around 20 different types of ganglion
cells have been identified based on physiological and anatomical criteria such as size,
response, receptive field, conduction velocity, stratification, coverage, central projections
etc. In addition, a particular type of ganglion cell called the midget ganglion cell is found
in the central region of primate retina. They receive input from midget bipolar cells and
are involved in providing high visual acuity. More recently, a set of ganglion cells were
found to express an additional visual pigment called melanopsin. These melanopsin
ganglion cells have been thought to be responsible for the light reflexes in blind mice.
They project to the suprachiasmatic nucleus (SCN) that is involved in maintaining the
circadian rhythm. Some projections were also found in the olivary pretectal nucleus
(OPN), which is involved in controlling pupillary reflexes.

42
Glial cell
There are roughly twice as many glial cells as there are neurons in the central
nervous system. In the retina, Müller cells, a type of glial cells, form the primary
scaffolding of the retina. The cell bodies of Müller cells lie in the INL while their
processes extend to and from the inner to outer limiting membranes. Müller cell
processes fill most of the extracellular space in the retina. Along with structural role,
these cells play a nutritive role and are also involved in maintaining the glutamate
homeostasis. In addition to Müller cells, three other types of glial cells are present in the
retina: retinal astrocytes, perivascular glia and microglia. Retinal astrocytes are sparse in
the retina and are present in the IPL, optic fibre layer and optic disk with their form
depending upon their location. Perivascular glia surround the capillaries and together
with astrocytes are claimed to insulate the capillaries from the neural tissue. Microglia in
the retina are the same as those found in the brain and are known to proliferate and
migrate when destruction of retinal tissue has occurred.

Rat retina
The structure and function of the rat eye shares remarkable similarity to that of
human and primate eye. Light enters through the cornea and is focussed by the lens,
which occupies large part of the interior, and is processed by the various layers of the
retina. Like primate retina, rodent retina is also rod dominated with a rod to cone ratio of
100:1 (Rohrer and Crouch, 2006). However, rats have a dichromatic (blue and green
cones) as compared to trichomatic (blue, green and red cones) vision of humans. Overall
43
there are approximately 30 million photoreceptors and around 100 – 120 thousand
ganglion cells in the rat retina resulting in a high convergence ratio of around 300:1
(Sefton and Dreher, 1985; Mayhew and Astle, 1997). Thus the rat retina has almost a
three times higher convergence ratio compared to human retina that helps it to
compensate for low acuity by having greater sensitivity in the scotopic range. The
ganglion cell density in rat retina ranges from 1000 to 6000 cells/mm
2
with minimal
differences in the size of dendritic trees and receptive field centres in the different
regions. Thus, unlike primate retina, the rat retina lacks a region analogous to the fovea
and area centralis to provide high visual acuity. Visual resolution in the rat is
substantially lower than primates, with the rat’s eye capable of optically resolving 12
minutes of visual angle (Sefton and Dreher, 1985).
Despite some of these differences, the rat retina is quite similar to primate retina.
Visual information is processed in the three different neural layers separated by the two
plexiform layers. Extensive work has been done to identify major subtypes of the six
different retinal cell classes. Three types of monostratified and one type of bistratified
ganglion cells have been identified. The ganglion cells have been classified based on the
size of their somas and dendritic fields into large soma, large dendritic field (RG
A
, type I,
class I), small soma, small dendritic field (RG
B
, type II, class II) and small-to-medium
soma but large dendritic field (RG
C
, type III, class III) ganglion cells (Sefton and Dreher,
1985; Sun et al., 2002). These three different types of ganglion cells correspond to three
groups of axons with different conduction velocities (Sefton and Swinburn, 1964).
Studies have estimated the optic nerve of adult albino rats to contain about 100, 000
44
axons, virtually all being myelinated (Sefton and Dreher, 1985). Due to these features of
rat retina, and the presence of disease models analogous to those found in humans, the rat
retina has been a popular model to study and understand visual processing in both normal
and degenerated retina.

1.2.2 Central Visual Pathway
Human Visual Pathway
The visual field is the view seen by the two eyes without any head movement.
The surface of the retina has been divided into regions relative to the fovea. The nasal
hemiretina lies medial while temporal hemiretina lies lateral to the fovea. Each half is
further divided into dorsal (or superior) and ventral (or inferior). When light originates
from the central part of the visual field, it enters both eyes creating a binocular zone. In
addition, light originating in each half of the visual field enters one eye creating a
monocular zone. Ganglion cell axons form the optic bundle and extend through the optic
disc. At the optic chiasm, the axons from the nasal half of each retina cross over to the
opposite side of the brain (contralateral projection) while the axons from the temporal
hemiretinas project to the same side of the brain (ipsilateral projection). In this
arrangement, the left optic tract carries information about the right visual hemifield while
the right optic tract carries information about the left visual hemifield. This separation is
maintained in all projections to subcortical visual nuclei.
Retinal ganglion cell axons stream towards the optic disc where they become
myelinated and are bundled into the left and right optic tracts. The optic tract projects to
45
three main subcortical regions: the pretectum, the superior colliculus and the lateral
geniculate nucleus. Retinal ganglion cell axons project directly to the superficial layers of
the superior colliculus (SC) where they form a map of the contralateral visual field. The
cells in the SC in turn project to the thalamus and onwards to a broad area of the cerebral
cortex thus forming an indirect pathway from the retina to the cortex. Cells in the SC are
involved in the control of saccadic eye movements. In addition, cells in the deeper layers
also respond to auditory and somatosensory stimuli. Retinal ganglion cell axons that
project directly to the pretectum in the midbrain are involved in mediating pupillary light
reflexes. Absence of such reflexes is a symptom of damage to the midbrain, the region
from which oculomotor nerves originate. In humans, 90% of the retinal ganglion cells
axons project to the lateral geniculate nucleus (LGN). Ganglion cell axons project in an
orderly manner to the LGN so that within each LGN there is a retinotopic representation
of the contralateral half of the visual field. However, all areas of the retina are not
represented equally with the fovea having a relatively lager representation. As the central
area of retina is less than its periphery, numerous neurons from the fovea are distributed
over a wide area in the LGN. The ratio of area in the LGN and primary visual cortex to
the area in retina representing one degree of visual field is called the magnification factor.
Visual acuity changes by a factor of 40 between fovea and peripheral retina. This drop in
visual acuity has been used to estimate magnification factor of human visual cortex
(Rovamo and Virsu, 1979). The magnification factor has been found to be linked with the
decrease in retinal ganglion cell density with eccentricity (Perry and Cowey, 1985;
Wassle et al., 1990). The magnification factor in humans is approximately equal to 4
46
mm/degree at 2 degrees eccentricity and declines monotonically to 0.5 mm/degree at 25
degrees eccentricity (Cowey and Rolls, 1974).

Rat visual Pathways
Most retinal ganglion cell axons in rat retina project to the SC but many axons
branch to supply both the dorsal lateral geniculate (DLG) and the SC. Afferent inputs in
the DLG arise from a diverse source while efferent projections are restricted to occipital
cortex and visual thalamic reticular nucleus. Additionally, about 8% and 13% of retinal
ganglion cell axons branch to project to the ventral lateral geniculate (VLG) and the
pretectum respectively. Functionally, the VLG has been implicated in brightness
discrimination while different regions of the pretectum are involved in optokinetic
nystagmus and pupilloconstrictor light reflex.
In mammals, the superior colliculus (SC) is presumed to be the homologue of the
optic tectum in other vertebrates. Although, retinal axons are known to terminate in this
region, it does not constitute the principle visual projection. In the rat however, the SC is
the major target of retinal ganglion cells and is thus involved in processing visual
information (Lund, 1969; Langer and Lund, 1974). It falls in an area with coordinates
around 3.5 mm rostral and 0.5-1.5 mm lateral relative to the interaurel and midline planes
respectively (Fig. 1.10).

47

Figure 1.10 A – Dorsal and lateral views of the skull of a 290g Wistar rat; B –
Drawing of a section of rat brain cut in a sagittal plane 0.4 mm lateral to the midline
(Paxinos and Watson, 1998).

48
Studies have implicated the rat SC to be involved in several visual localizing
functions, visual orienting responses, directing eye movements and locomotor
exploration (Sefton and Dreher, 1985). However, unlike human and primate retina, the rat
SC has a uniform retinal projection with a very low magnification factor ranging from
approximately 0.02 mm/degree in the medial-lateral axis to 0.04 mm/degree in the
posterior-anterior axis (Siminoff et al., 1966). Its horizontally laminated organization is
divided into superficial, intermediate and deep parts based upon cell types, efferent and
afferent projections. While the superficial layers are almost exclusively involved in
processing visual information, cells in the deeper layers can be activated by visual,
somatosensory and auditory stimuli. The superficial layer has been further subdivided
into four main layers: zonal (Z) or stratum zonale (SZ), superficial gray (SuG) or stratum
griseum superficiale (SGS), optic (Op) or stratum opticum (SO). The intermediate layers
have been subdivided into the intermediate gray (InG) or stratum griseum intermediale
and the intermediate white (In W) or stratum album intermediale while the deep layers of
the SC have been divided into deep gray (DpG) or stratum griseum profundum and deep
white (DpW) or stratum album profundum. It has been shown that the superficial layers
are innervated by retinal axons arising in the contralateral eye and the cells in these layers
in turn project to deeper layers of the SC. However, a direct retinal projection has been
shown to terminate in the intermediate layers of some mammals including rat. Due to the
direct connection of retinal ganglion cells with the superficial layers of the SC, this
region of SC has been of wide interest. Studies have shown that superficial layers of the
SC maintain a precise retinotopic map that is consistent across rat strains with the densest
49
retinal projections found in the SGS. A more recent study has shown that the superficial
part of the SGS (SGS1) almost exclusively consists of retinal projections with few, if any
cortical inputs and thus is involved in sophisticated analysis of visual information
(Girman and Lund, 2007). Deeper sublamina of SGS and SO receive inputs from visual
cortical areas along with afferents from SGS1 and retinal ganglion cells. Early studies
identified a 3 to 4 group classification of cell types in the superficial layers of the SC via
optic nerve stimulation (Fukuda et al., 1978). These studies classified cells in each of the
superficial layers based upon conduction velocities of the optic nerve fibres innervating
them. More recently, six major cell types have been identified in the superficial layers of
the SC: stellate, narrow- and wide-field, vertical, horizontal, piriform and marginal
(Edwards et al., 2002). No distinct properties were observed among the major cell types
although all were subject to strong inhibitory modulation. Spike frequencies ranged from
10 to 70 Hz.

50

Figure 1.11 Images of six different neuronal cell types identified via Lucifer
Yellow staining in the superficial layers of the superior colliculus. Scale bar = 15
µm. (Figure has been modified from Edwards et al., 2002).

Figure 1.12 Topographical arrangement of receptive fields mapped in the left
visual field and their corresponding electrode positions on the contralateral right
superior colliculus in the rat (Siminoff et al., 1966).

51
1.3 Visual Prosthesis
The idea of a visual prosthesis developed in the early 1900s with Foerster’s work
demonstrating the presence of a small spot of light in response to electrical stimulation of
the occipital cortex (Foerster, 1929). This spot of light termed phosphene was found to
correspond to the anatomical location stimulated. Study done by Kraus & Schum (1931)
further reinforced the idea that stimulating the occipital pole in a patient that had been
blind for eight years could generate well-defined, localized sensations of light (Krause
and Schum, 1931). These early studies demonstrated an important finding that despite
long-term visual deprivation, phosphenes could be generated via electrical stimulation.
Work in this direction continued with observations made by Penfield that visual
sensations could be produced by stimulating the cortex (Penfield and Rasmussen, 1950).
However, the regions stimulated were observed to be mainly those lying outside the
striate (also known as primary visual cortex or V1). It was in the late 1960s when
Brindley & Lewin first showed the presence of visual percepts via electrical stimulation
of the primary visual cortex (Brindley and Lewin, 1968). They implanted 80 platinum
electrodes (square of side 0.8 mm) on the occipital pole in the right cerebral hemisphere
in a 52 year old left blind due to retinal detachment. While surface stimulation of the
visual cortex clearly demonstrated the generation of phosphenes, it did not culminate in
restoration of useful vision. A major drawback found with surface stimulation was the
requirement of relatively large currents to elicit phosphenes, which in turn limited the
size of electrodes. In some cases, such large currents were observed to induce seizures.
52
Other studies showed that such surface stimulation may or may not create phosphene
flicker (Dobelle and Mladejovsky, 1974).
To address these issues, the next period in development of a cortically based
visual prosthesis was focussed on using small clusters of microelectrodes implanted in
the visual cortex. Work done in human subjects (Bak, 1989; Bak et al., 1990; Schmidt et
al., 1996) and non-human primates (Bradley et al., 2005) demonstrated focal stimulation
of cortical neurons and provided basic information on the nature of such percepts. More
recent work in this field has been focussed towards increasing the density of the
microelectrode array to be implanted in the visual cortex. In addition to implanting in the
visual cortex, attempts have been made to develop an LGN based visual prosthesis
(Pezaris and Reid, 2007). Although considerable progress has been made in developing a
cortical-based visual prosthesis, it has not yet resulted in a clinical visual prosthesis
system. Some of the challenges in this approach are requirement of complex signal
processing and dangers associated with surgical implantation.
Another approach that has been investigated is the stimulation of the optic nerve,
which is a bundle of retinal ganglion cell axons. Optic nerve stimulation became feasible
with the development of cuff electrodes for nerve fibre recording. The idea of an optic
nerve implant was developed by Veraart and colleagues (Veraart et al., 2003) and
currently it being pursued by a group in China. However, as the optic nerve does not
maintain spatial organization, the major difficulty with this approach remains in
preserving spatial mapping of the stimulus.
53
In the 1990s, when investigators demonstrated the relative preservation of inner
retinal neurons (Milam et al., 1998; Humayun et al., 1999b), the development of a retinal
prosthesis began to take shape. Back in the 1960s, Potts & Inoue (1969) had
demonstrated that phosphenes could be elicited in blind patients via corneal stimulation
(Potts and Inoue, 1969). Today numerous groups worldwide are involved in developing a
retinal prosthesis. Currently three main approaches are being investigated. In the case of
epiretinal implants, the design includes an extraocular component to capture the image
and an intraocular array of electrodes to stimulate the retina. These electrodes are
positioned on the epiretinal surface thus closest to the retinal ganglion cells, the output of
the retina. Early work done by Humayun et al. showed that epiretinal stimulation in
patients with end-stage retinitis pigmentosa (RP) could elicit distinct percepts from
adjacent electrodes (Humayun et al., 1999a). Although epiretinal implants have numerous
advantages over a cortically based visual prosthesis, it still has certain drawbacks. As the
electrodes bypass most of the inner retinal neurons, the device may require more
sophisticated image processing. Also, one of the major concerns with epiretinal implants
has been in optimal placement of the electrode array on the epiretinal surface.
54

Figure 1.13 Schematic illustrating epiretinal and subretinal implants in the eye.

The subretinal implants on the other hand have taken a slightly different approach
in the development of a retinal prosthesis. In this case the active or passive elements of
the device are positioned in the subretinal space (between RPE and retina). These
microphotodiodes are envisaged to play the role of photoreceptors. Early work in this
field involved the implantation of passive photosensitive diode arrays that absorbed
ambient light and the corresponding electrical signal was transmitted to the retinal
neurons (Chow and Peachey, 1998; Chow et al., 2004). Although initial studies reported
some improvement in vision, this was later attributed to possible neurotrophic factors
associated with surgical implantation (Pardue et al., 2005). Since the current generated by
such microphotodiodes was found to be inadequate to stimulate retina neurons, more
recent work has included an external power source along with the microphotodiodes in a
subretinal implant (Zrenner, 2002; Gekeler et al., 2004; Palanker et al., 2005). However,
complications due to surgical implantation still remain a major drawback. Also, the
55
presence of a thick glial seal between the retina and RPE along with substantial
remodelling of bipolar cells in late-stage RP may pose a problem in signal transmission
between the device and retina.
A third approach in the development of retinal prosthesis is the suprachoriodal
transretinal stimulation (STS). In this approach, the stimulating electrodes are positioned
on the suprachoriodal space while the return electrode is placed in the vitreous. Work
done by Tano and colleagues have shown the feasibility of this approach in animal
experiments (Kanda et al., 2004; Fujikado et al., 2007; Nishida et al., 2009). Although
less invasive compared to the other two approaches, this design requires higher stimulus
currents.
Despite a large number of groups worldwide involved in developing a retinal
prosthesis, only few have made it to the stage of clinical trials. In the US, Second Sight
Medical Products (SSMP) have conducted a long-term clinical trial with a 16-electrode
version of their epiretinal implant and are currently involved in a 60-electrode clinical
trial. In Germany, at present there are three companies pursuing clinical trials of retinal
prostheses. IMI GmbH has reported implantation of four subjects with their 49-electrode
epiretinal array that is designed to be functional for 18 months. EPIRET3 reported the
implantation of a 25-electrode system in six subjects for four weeks and Retina Implant
AG have reported implantation in eight subjects for four weeks.

56
1.4 Goal of Present Study
The present study aims to determine the effect of continuous epiretinal
stimulation on both the morphology of stimulated retinal neurons and on the resulting
evoked response in the higher order centres (superior colliculus). This study has been
undertaken in an effort to aid the development of a retinal prosthesis employing epiretinal
stimulation.
At present, clinical trials of retinal prostheses involve test subjects suffering from
RP. The retina of such patients is far from normal. In order to decide on an appropriate
animal model for the present investigation, the study begins with a description of the
morphological changes that occur in a transgenic animal model of RP (chapter 2). Since
massive changes in the retina were observed due to retinal degeneration, a normal animal
model was chosen for majority of the study. This choice was made in order to avoid any
confusion between stimulation-induced and degeneration-induced changes in the retina.
In an effort to examine only stimulation-induced changes without any mechanical effect
of electrode tip on retina, a method to position the electrode at a safe distance from the
retina was developed. This method of electrode positioning along with a brief
characterization of the stimulation electrode is described in chapter 3. Chapter 4 describes
acute experiments that show moderate to severe retinal damage due to pressure exerted
by electrode tip on retina and accompanying high intensity stimulation. Chapter 5
describes the effect of continuous stimulation on both retinal morphology and threshold
of evoked response in the SC with proper stimulation electrode placement. Chapter 6
describes experiments that were undertaken in an effort to understand possible
57
mechanism(s) that may be at play during continuous epiretinal stimulation. Based on all
the results, an overall summary and future work has been laid out in chapter 7. Details of
experimental methods employed throughout the study have been described in chapter 8.

58
Chapter 2
Morphological Changes in Retina due to Retinal Degeneration in
Transgenic Animal Model

2.1 Background
Genetic mutations occurring in either photoreceptors or pigment epithelium often
lead to retinal degeneration characterized by photoreceptor dysfunction and death. Out of
these retinal degenerative diseases, one of the best characterized is retinitis pigmentosa
(RP) (Bird, 1995), which has an incidence of 1 in every 4000 humans (Bunker et al.,
1984). Different treatment strategies being explored for RP include prosthetic implants
(Zrenner, 2002; Weiland et al., 2005; Fujikado et al., 2007), gene therapies (Nour and
Naash, 2003; Auricchio and Rolling, 2005; Bi et al., 2006), and retina and stem cell
transplants (Klassen et al., 2004; Das et al., 2005; Thomas et al., 2006b; Thomas et al.,
2006a). The presence of animal models with diseases homologous or analogous to those
in humans provides an opportunity to explore treatment strategies that may ultimately be
tested in the clinic (LaVail, 1981; Chader, 2002). Several such models are popular for the
retina. They include naturally occurring rat and mouse models, such as Royal College of
Surgeons (RCS) rats, which exhibit a defective MERTK gene in the pigment epithelium
(D'Cruz et al., 2000). These models also include the retinal degenerate (rd) mouse model
exhibiting a rod-phosphodiesterase gene mutation (McLaughlin et al., 1993). With
increasing efforts directed towards developing treatment strategies for retinal
59
degenerative diseases, understanding and characterizing the degeneration profile of these
animal models becomes essential.
Retinal degeneration with photoreceptor dysfunction ultimately leads to
remodeling of the inner retina. As an example, studies performed in both the faster-
degenerating rd1 and the slower-degenerating rd10 mice have shown abnormal changes
in the inner retinal neurons (Strettoi et al., 2003; Gargini et al., 2007). Both models show
similar alteration patterns to other models of photoreceptor disease (Kolb and Gouras,
1974; Marc and Jones, 2003; Cuenca et al., 2005; Jones and Marc, 2005; Wang et al.,
2005), with the rd1 model exhibiting a more aggressive mutation than the rd10 model.
To date, most studies involving degenerate models have focused on naturally
occurring models of RP (Machida et al., 2000; Strettoi et al., 2003; Suzuki et al., 2004;
DeMarco et al., 2007; Gargini et al., 2007; Jensen and Rizzo, 2008; Wang et al., 2008).
These models serve as invaluable tools for understanding the dynamics and mechanisms
of retinal degeneration. However, a drawback of the naturally occurring models is that
most express recessive forms of RP. In addition, these models express pigment-
epithelium defects rather than the rhodopsin gene mutation, which accounts for more than
25% of human RP cases (Colley et al., 1995; Hartong et al., 2006). Towards this end,
genetically engineered models have been developed that closely mimic the expression
pattern of the disease in humans. Examples of such transgenic models are the S334ter and
P23H rats that express a rhodopsin mutation found in human RP patients. Due to this
advantage over other models of retinal degeneration, a few retina transplant studies have
been performed on the S334ter animal model (Sagdullaev et al., 2003; Thomas et al.,
60
2006b; Thomas et al., 2006a). Currently, little information is available on secondary
changes that occur in the S334ter retinal degeneration model (Liu et al., 1999a). In an
effort to aid the ongoing research into developing therapies for retinal degenerative
diseases, present work focuses on examining the patterns of alterations observed in the
retina of the S334ter-line-3 model. Much of this work has already been published
elsewhere (Ray et al., 2010). This model shall be referred to as the RD model in the rest
of the thesis.

2.2 Experiment Protocol
In this study, retinas from RD rats were collected at postnatal (P) days P11, P15,
P21, P60, P90, P180, and P600 (N = 7 for each stage). Controls were age-matched
Sprague Dawley rats (N = 7 for each stage). Immunohistochemical analysis was
performed on retinas at each stage to identify morphological changes due to retinal
degenerations.

2.3 Results
2.3.1. Photoreceptor Degeneration
As the S334ter-line-3 rat is an outer-retinal degeneration model, the first step
involved investigating the gross morphology of the retina. For this, TOPRO-3 staining
was carried out on vertical sections of P11, P15, P21, P60, P90, and P180 RD rat retinas.
A progressive decrease in the thickness of the ONL was observed in TOPRO-3 stained
retinas (Fig. 2.1 — blue). The thickness of ONL showed no statistically significant
61
difference at P11 between RD (70.29 µm; N=13) and control (82.26 µm; N=9) (p=0.07).
At P15, reduced thickness of ONL was observed in the central retina as compared to the
periphery (Fig. 2.1c, d). The P21 RD retinas showed three or four layers of ONL (Fig.
2.1e). In contrast, at P60, only a single layer of photoreceptors was left. Finally, at P90,
very few photoreceptors could be seen (Fig. 2.1f). These photoreceptors were identified
as leftover cones by double labelling TOPRO-3 nuclei with Short-wavelength sensitive
cone opsin (S-opsin -- Fig. 2.2) and FITC-PNA (data not shown). The ONL first thinned
in the centre, followed 3-5 days later in the periphery (P15 centre vs. P15 peripheral).
Positive labelling of TUNEL (terminal deoxynucleotidyl transferase (TdT)-mediated
dUTP-biotin nick-end labelling) (green) in the nuclear layers indicated degenerating
retinal cells undergoing apoptosis.
62

Figure 2.1 Light micrographs taken from vertical cryostat sections processed for
TOPRO-3 staining (blue) in normal (a) and RD (b- h) retinas. Apoptotic marker for
cell death (TUNEL) appears in green. The largest number of TUNEL positive cells
is seen at P15 in the ONL, both in the central (c) and peripheral (d) RD retinas. Few
TUNEL positive cells can be seen in the INL at later stages of degeneration (g, h).
ONL - outer nuclear layer, OPL - outer plexiform layer, INL - inner nuclear layer;
IPL - inner plexiform layer IPL, GCL - ganglion cell layer. Scale bar = 50 µm.

63

Figure 2.2 Confocal micrographs of vertical sections labelled with TOPRO-3 (a)
and S-opsin (b) in P90 RD retina. Double exposure of TOPRO-3 and S-opsin
labelled cells show that the remaining photoreceptors in the ONL are leftover cones
(c – arrows). Scale bar = 50 µm.

2.3.2. Glial Reaction
Degeneration causes a typical glial reaction involving the upregulation of the glial
fibrillary acidic protein (GFAP) by Müller cells (Roque and Caldwell, 1990; DiLoreto et
al., 1995; Tanihara et al., 1997). One can visualize this glial reaction by means of an
antibody directed against the intermediate filament protein GFAP. Although the highest
number of apoptotic cells in the ONL occurred at P15, the glial reaction began at P21,
increasing thereafter (Fig. 2.3). Moreover, weak glial sealing composed of Müller cell
processes was observed in the outer retina from P60 RD onwards. This glial seal became
more evident as the disease progressed (Figure 2.3e-f) and is similar to that has been
reported in other studies involving retinal degeneration (Li et al., 1995; Milam et al.,
1998; Jones et al., 2003; Jones and Marc, 2005).

64

Figure 2.3 Light micrographs taken from vertical vibratome sections processed
for GFAP immunoreactivity at six developmental stages of RD retinas (a-f). Up-
regulation of GFAP in Müller cells of the RD retina can be seen from P21 onwards
(b). Glial sealing in the outer retina is evident at P180 (e) and at P600 (f). Scale bar =
50 µm.

2.3.3 Changes in Horizontal Cells
Next the effect of degeneration on horizontal cells was examined. Antibodies
against calbindin stain horizontal cells entirely, while antibodies against neurofilaments
label their axonal complexes (Peichl and Gonzalez-Soriano, 1994; Haverkamp and
Wässle, 2000; Reese et al., 2005). Calbindin was used to view the entire horizontal cell
plexus at different stages of retinal degeneration in the RD retina (Fig. 2.4). Horizontal-
65
cell morphology remained unchanged until P15 (Fig. 2.4a). At P21 (Fig. 2.4c), the
processes of horizontal cells in central retina appeared attenuated at the OPL compared to
that of P21 control retina (Fig. 2.4b). The gradual attenuation of the fine plexus became
more evident at P60, forming a weak immunoreactivity at the OPL (Fig. 2.4d - arrows).

Figure 2.4 Light micrographs taken from vertical vibratome sections processed
for calbindin D-28K immunoreactivity in normal (B) and RD (A, C and D) retinas
at different developmental stages. Gradual attenuation of horizontal-cell labelled
processes can be seen from P21 onwards (C and D - arrow). Immunoreactive
pattern is the same for P60 RD retina fixed for 15 min (E). Scale bar = 50 µm.

2.3.4 Changes in Rod Bipolar Cells
In the mammalian retina, one can stain rod bipolar cells with antibodies against
the alpha isoform of protein kinase C (PKCα) (Haverkamp and Wassle, 2000). At P15
RD retina, dendrites of rod bipolar cells showed apical processes towards the ONL (Fig.
2.5A, E). This expression was similar in P21 control retina (Fig. 2.5B, F). At P21, the
dendrites of rod bipolar cells were retracted or distributed laterally in the OPL of RD
retina (Fig. 2.5C, G). The retraction of dendritic tree was more severe as the disease
progressed. At P60, rod bipolar cell bodies appeared to be devoid of any processes (Fig.
2.5D and H). This maybe on account of reabsorption of the processes due to lack of any
66
contacts with photoreceptor cells. However, no changes in the terminals of these cells
were observed.

Figure 2.5 Light micrographs taken from vertical vibratome sections processed
for PKC immunoreactivity in normal (B, F) and RD (A, C, D, E, G, H) retinas.
Dendrites of rod bipolar cells project laterally in the OPL of RD retinas (G – arrow)
in contrast to the apical processes in normal retinas (F – arrow) at P21. Misplaced
cells with no processes appear at P60 (H – arrowhead). Scale bar = 50 µm. Bottom
pictures are at higher magnification of the same stage of normal (F) and RD (E, G
and H) retinas. Scale bar = 20 µm.

2.3.5 Cone Bipolar Cells
In the rat retina, antibody against recoverin has been shown to label
photoreceptors and two types of cone bipolar cells (Milam et al., 1993). Vertical sections
of the RD retinas labelled with anti-recoverin did not exhibit any obvious abnormalities
in ON and OFF cone bipolar cells up to P60 (Fig. 2.6A, B, C, F). However, at P90, the
dendritic trees showed retraction and flat form in the OPL. In addition, recoverin
67
labelling showed disappearance of dendritic trees from cone bipolar cell bodies. Although
recoverin stained both bipolar cells and photoreceptors, the dendrites of cone bipolar cells
could be clearly observed (Fig. 2.6G).

Figure 2.6 Light micrographs taken from vertical vibratome sections processed
for recoverin immunoreactivity. Labelled cone bipolar cells and photoreceptors can
be seen at different developmental stages of RD retinas (A-G). No change occurs in
cone bipolar cell morphology until P60. At P90, dendritic retraction takes place in
the OPL (G – arrow). Moreover, some cells lose their processes (G – arrowhead).
Scale bar = 50 µm. Bottom pictures are taken at higher magnification of the same
stage of RD retinas (E – G). Scale bar = 20 µm.

2.3.6 Amacrine Cells
At all ages examined, changes in amacrine cells were found only in late stages of
degeneration. In the mammalian retina, most amacrine cells contain either glycine or γ-
aminobutyric acid (Vardi and Auerbach, 1995). Therefore, vertical sections of RD retinas
were labelled with antisera against GABA and parvalbumin. Parvalbumin is a specific
68
marker for the AII amacrine cell (an important kind of glycinergic amacrine cell) in the
mammalian retina (Haverkamp and Wassle, 2000; Lee et al., 2004). The general
morphology and immunoreactive pattern of AII and GABAergic amacrine cells of RD
retinas (Fig. 2.7A, D) was similar to those of control retinas up to P180 (Fig. 2.7G, H).
However, at P180, misplaced GABA (Fig. 2.7B) and parvalbumin (Fig 2.7E)
immunoreactive cells were observed in the outer portion of the INL. This misplacement
also occurred in P600 RD retinas (Fig. 2.7C, F).

Figure 2.7 Light micrographs taken from vertical vibratome sections processed
for GABA in normal (G) and RD (A – C) retinas, and Parvalbumin in normal (H)
and RD (D– F) retinas. No change in cellular morphology is apparent at P90 (A, E).
Displacements of GABAergic (B, C – arrows) and AII amacrine cells (E, F – arrows)
occur in the INL at P180 and P600. Scale bar = 50 µm.
69
2.3.7 Quantification of Modifications of Horizontal and Rod Bipolar Cells in RD
Retinas
Figure 2.8 shows example of whole mounts processed for calbindin (a-d) and
PKC alpha (f-i) immunoreactivities. These whole mount photos were from the central
and peripheral part of P60-and P90-RD retinas. Horizontal cell density was observed to
undergo a drastic reduction in the central regions of RD retinas at P60 (360 ± 70
cells/mm
2
) and P90 (350 ± 70 cells/mm
2
). These densities were found to be statistically
significantly lower than that of P60 normal retina (670 ± 90 cells/mm
2
; p = 0.001). A
decrease in cell density was also observed in the peripheral regions. The densities in P60
(550 ± 70 cells/mm
2
) and P90 (550 ± 70 cells/mm
2
) RD retinas were lower than normal at
P60 (630 ± 70 cells/mm
2
) (Fig. 8e). In order to confirm that the reduction in cell density
was due to cell death and not due to loss of immunoreactivity, the cell densities of all
TOPRO-3 stained cells in the outer part of the INL was measured. The cell density
measurement was made on the peripheral region at the same focal plane as calbindin and
PKC immunoreactive cells in the outer part of the INL. For these measurements, P60-
normal (n=2), and P60-RD (n=2) whole mounts were used that were double labelled with
either calbindin and TOPRO-3 or PKCα and TOPRO-3. There was a clear reduction in
the number of TOPRO-3 labelled cells in the INL between P60 normal (23541 ± 600
cells/mm
2
) and P60 RD (13637± 400 cells/mm
2
) retinas.

70

Figure 2.8 Wholemount images of RD retinas processed for Calbindin
immunoreactivity in P60 retinas (a-central; b-peripheral) and P90 retinas (c-
central; d-peripheral) and for PKC immunoreactivity in P60 (f-central; g-
peripheral) and P90 (h-central; i-peripheral) retinas. Density of horizontal (e) and
rod bipolar (j) cells at P60 normal, and P60 and P90 RD retinas. Cellular densities
were analyzed for both central and peripheral regions of each retina. A significant
reduction in cell density took place for both horizontal and rod bipolar cells in the
central region of RD retinas at P60, with further reduction at P90 in the case of rod
bipolar cells. A decrease in cell density also occurred in peripheral regions of RD
retinas at P60 for both horizontal and rod bipolar cells. Data presented as mean ±
s.d. Scale bar = 50 µm.

PKC immunoreactivity also revealed a marked reduction in the density of rod
bipolar cells at P60 in central RD retina (Fig. 2.8j). The density fell to about 50% (4500 ±
500 cells/mm
2
) at P60 (p = 0.003) and to about 30% (2900 ± 400 cells/mm
2
) at P90 (p =
0.0006). In normal retina at P60, the density was 8500 ± 500 cells/mm
2
. The density in
71
the periphery was also reduced in degenerated retinas at P60 (6000 ± 700 cells/mm
2
) and
P90 (4800 ± 400 cells/mm
2
). In P60 normal retinas, the density was 7100 ± 200
cells/mm
2
which was statistically significantly higher than P90 RD (p = 0.0003) retinas.

2.4 Discussion
Retinal degeneration due to photoreceptor loss leads to remodeling of the retina.
Numerous studies in different animal models of retinal degeneration have shown that
remodeling of retina ensues independent of the initial molecular targets of retinal
degeneration. Retinal remodeling takes place in three distinct stages with each stage
characterized by specific events such as photoreceptor loss, Müller cell hypertrophy,
cellular migration etc (Jones et al., 2003; Jones and Marc, 2005; Marc et al., 2007). In the
present study, morphological alterations that occur in the S334ter-line-3 (RD) transgenic
rat were investigated. As seen in other models of retinal degeneration, the RD model
exhibits typical pattern of photoreceptor degeneration followed by secondary
modifications. Bipolar and horizontal cells are the first cells to undergo cellular
remodeling. At later stages of degeneration, changes in amacrine cells also occur.
The RD model exhibits degeneration timeline that is slower than the rd1 mouse
model but slightly faster than the P23H-line-1 model. The onset of photoreceptor
degeneration begins around P15 compared to rd1 model where it begins at around P10
(Strettoi et al., 2002), thus starting right after synaptogenesis (Horsburgh and Sefton,
1987). Similar to other models of photoreceptor degeneration, the RD model studied here
exhibits a clear centre to periphery gradient, with the rods undergoing rapid degeneration
72
followed later on by cones (Strettoi et al., 2003; Cuenca et al., 2004; Cuenca et al., 2005;
Barhoum et al., 2008). At P90, only a few scattered photoreceptor cells are observed that
are found to be positive for S-opsin and FITC-PNA. These cells are hypothesized to be a
part of remaining cones in the degenerating retina as has been seen in other models of
retinal degeneration (LaVail et al., 1997; Gargini et al., 2007). Degenerating
photoreceptors undergo apoptotic cell death as illustrated by TUNEL-positive cells in the
ONL of retinas at P15. Moreover, an earlier study has shown the photoreceptor
degeneration in the RD model to be associated with a caspase-3 dependent mechanism
(Liu et al., 1999a).
Parallel to photoreceptor degeneration, secondary changes in RD retinas are
observed to begin around P21. At that time, horizontal-cell processes appear thinner and
less complex, with changes becoming progressively severe with age. However, no
sprouting is observed in the RD model of retinal degeneration. This model thus differs
from the RCS rat and rd mouse models, where horizontal cells appear to send their
processes deep into the inner retina (Cuenca et al., 2005; Wang et al., 2005; Barhoum et
al., 2008).
Cell density of horizontal cells undergoes a marked reduction at P60, reaching
about 50% of that in the normal, central retina. Concomitant to changes in horizontal
cells, rod bipolar cells exhibit classical signs of remodeling by retracting their dendrites
in the OPL. With progressive degeneration, some rod bipolar cells appear to be devoid of
any processes. The density of these cells also shows a marked reduction, falling to half
the number at P60 in central retina. The decrease in density continues until P90, when it
73
reaches about one-third that of the normal retina in the central region. As seen in other
models of retinal degeneration, cone bipolar cells also show changes in response to
retinal degeneration, but at a much slower rate (Strettoi et al., 2002; Cuenca et al., 2004;
Cuenca et al., 2005; Wang et al., 2005; Barhoum et al., 2008). For the ages examined
here, discontinuity of the dendritic plexus in the OPL was noted around P90. In
comparison, similar changes observed in the rd10 mouse model occurred around P60
(Gargini et al., 2007). Other studies have shown the recruitment of the rod pathway by
the cones in absence of rods (Peng et al., 2000; Peng et al., 2003; Strettoi et al., 2004). In
these studies, formation of ectopic synapses between surviving cones and rod bipolar
cells has been shown. Although the present study did not look into detail the formation of
such synapses, it may be safe to assume that due to similar pattern of degeneration,
formation of such synapses may occur at late stages of degeneration. Glial reaction
begins around P21 and increases with degeneration. Glial seal between the neural retina
and the pigment epithelium forms from P60 onwards. This is similar to what has been
described as a signature event of phase 2 of retinal remodeling (Jones and Marc, 2005).
Finally, as observed in other models of retinal degeneration, amacrine cells remained
unchanged till late stages of degeneration at which time cellular misplacement was
observed (Jones et al., 2003).
Thus, the RD model studied here exhibits a slower time course of retinal
degeneration compared to the rd1 mouse. This slow pace of the RD model proves to be
advantageous in studying vision-rescue strategies as the overlap of cell death with retinal
synaptogenesis in other models creates certain limitations. However, slower RD
74
degeneration may also have a downside. The progression of secondary changes in the RD
retina such as sprouting of cellular processes may pose difficulties in restoring normal
vision. Despite widespread morphological changes in the retina, studies have shown the
ability of ganglion cells in retinal degenerate models to preserve their intrinsic firing
properties (Jensen and Rizzo, 2008; Margolis et al., 2008; Stasheff, 2008). Other studies
have exhibited the capability of higher order centres of visual processing to respond to
electrical stimulation in such retinal degenerate animal models (DeMarco et al., 2007;
Chan et al., 2008) and also in clinical trials of retinal prosthesis (Caspi et al., 2009;
Horsager et al., 2009b). Furthermore, studies have shown preservation of ganglion cell
morphology (Mazzoni et al., 2008). The preservation of morphology and intrinsic
properties, and the response of higher centres to retinal stimulation together show that
vision rescue strategies geared towards stimulating the remaining retinal neurons may
prove to be successful.

75
Chapter 3
Electrode-Retina Interface

3.1 Characterization of Stimulation Electrode
3.1.1 Background
An ideal candidate for electrode material for neural stimulation is one which is
biocompatible, mechanically stable to surgical implantation, maintains its electrical and
mechanical properties for the entire duration of use and is able to support the charge-
injection requirements without inducing damage to itself or to the target tissue. Studies
have been done to identify parameters that govern the efficacy and safety of electrode
materials. Subsequent studies have involved determination of values of these parameters
for materials that are most commonly used as electrodes for neural stimulation.
Platinum and its alloys with iridium is the most widely used electrode material for
neural stimulation. Being a noble metal, it is highly resistant to corrosion and hence
suitable for chronic implantations. As platinum is a soft metal it is sometimes alloyed
with iridium that is hard and brittle to improve hardness. The electrochemistry of
platinum has been well studied along with its charge storage and injection capacities.
Along with double layer charging, charge injection can occur through the reversible
adsorption of hydrogen onto the platinum surface (H-atom plating) responsible for the
pseudocapacity of platinum. Brummer and Turner studied the underlying mechanisms
during charge injection through platinum electrodes and its alloys (Brummer and Turner,
76
1975; Brummer and Turner, 1977a, b) and found that these chemically reversible
processes can provide charge injection up to 300-350 µC/cm
2
in simulated cerebrospinal
fluid (Brummer and Turner, 1977a). In practice, the safe charge injection limit of
platinum depends upon a variety of factors such as the pulse duration, current density and
geometry of the electrode surface. For rectangular pulses 0.2 ms in duration, the safe
charge injection limit with platinum was found to range from 50 to 150 µC/cm
2
(Rose
and Robblee, 1990). Some studies have attempted to increase the electrochemical safe
charge injection limit of platinum by increasing the real surface area of the electrode by
roughening and have shown varying degrees of success (Whalen et al., 2005; Greenbaum
et al., 2006).
Iridium oxide belongs to the category of electrodes that are termed as valence
change oxides. The oxide layer can be formed in three different ways. Anodic iridium
oxide films (AIROF) are produced through repetitive potential cycling of the bulk metal
between 0.0 and 1.5 V versus a reversible hydrogen electrode in an acid or buffered
neutral electrolyte (Robblee et al., 1983; Robblee and Rose, 1990). The activated iridium
is highly resistant to dissolution and corrosion and exhibit charge storage capacities
ranging from 10 to 240 mC/cm
2
(Stieglitz, 2004). This charge storage capacity depends
upon the thickness of the film and even moderate activation can lead to high values.
However, during neural stimulation, only a fraction of this charge can actually be used.
Weiland et al. found the reversible charge injection limits of AIROF to be about 4
mC/cm
2
, which is greater than platinum and some other metals used for neural
stimulation (Weiland et al., 2002a). Beebe et al. showed charge injection limits of about 2
77
mC/cm
2
for biphasic pulses and 3.5 mC/cm
2
for monophasic pulses, 0.2 ms in duration
with activated iridium wire electrodes (Beebe and Rose, 1988). Iridium oxide films can
also be formed by thermal decomposition of layers of iridium salts (TIROF) or by
reactively sputtering the oxide films onto a substrate from an iridium target (SIROF).
Iridium oxide films on the whole have exhibited poor stability during chronic stimulation
regimes. However, recent work on SIROF shows improvement in in vitro stability during
long-term pulsing (Cogan et al., 2009). In a separate study, Weiland et al. found the
metal-tissue interface to be altered after chronic stimulation using thin film iridium oxide
electrodes implanted in guinea pig cortex. They observed that current pulsing within safe
limits increased the impedance at low frequencies (<100 Hz) after one or two days of
stimulation and found the impedance change to correspond to a reduction in the charge
storage capacity (Weiland and Anderson, 2000). Other studies have also found iridium
oxide electrodes to delaminate under high current pulsing with deposits in the
surrounding tissue (Cogan et al., 2004).
Capacitive electrodes are ideal for neural stimulation as they do not involve any
reactions for charge injection and hence do not have to deal with problem of irreversible
reactions. However, these electrodes still have to be operated within the water window in
order to avoid hydrolysis. The metal is insulated from the solution by a thin layer of
dielectric material that must be able to withstand the electric fields without any
significant DC leakage. Materials in this group that have been found to be safe are
anodized tantalum (Ta/Ta
2
O
5
), anodized titanium (Ti/TiO
2
), thin films of barium titanate
(BaTiO
3
) and sputtered deposited titanium nitride (TiN). While anodized tantalum was
78
found to have higher charge storage capacities than anodized titanium and thin films of
barium titanate, titanium nitride was found to have charge storage capacities of 23
mC/cm
2
when combined with CMOS technology to develop microcolumnar structures
(Stieglitz, 2004). However, the injectable charge limit of titanium nitride was found to be
about 0.87mC/cm
2
for microelectrodes while for Ta/Ta
2
O
5
to be around 0.1-0.2 mC/cm
2

for large electrodes (Robblee and Rose, 1990). Hence, capacitive electrodes though safer
than electrodes employing Faradaic reactions, have in general lower charge injection
capabilities when operating within the water window.
Carbon nanotubes are also part of the capacitive electrode category exhibiting
interesting electrochemical and mechanical properties. They are about five times stronger
than steel and yet can be bent and twisted without breaking them. Recent work has shown
them as potential electrode material for neural stimulation. Wang et al. developed
vertically aligned multiwalled carbon nanotubes (CNTs) using catalytic thermal vapour
deposition system (Wang et al., 2006). They tested the properties of the CNTs and found
that CNTs have a higher charge injection limit of 1-1.6 mC/cm
2
after some surface
treatment had been performed. Also, continuous pulsing did not degrade the properties of
the CNTs. They also found these carbon nanotubes to be capable of causing neuronal
excitation in embryonic rat hippocampal neurons. With its precise control of size,
geometry and location by lithographic patterning of the catalyst and high charge injection
capabilities without any Faradaic reactions, carbon nanotubes may be an answer to the
requirements of neuroprostheses employing localized chronic neural stimulation.
79
However, CNTs generally are formed at very high temperatures, making them
incompatible with most batch electrode processes.
Conductive polymers are one of the more recent members to the family of
electrode materials for neural stimulation applications. Quite a few recent studies
illustrate the feasibility of electrochemically polymerizing polypyrrole, polythiophene
and their derivatives from aqueous solutions and depositing them on microelectrodes
(Cui et al., 2001a; Cui et al., 2001b; Cui et al., 2003; Cui and Martin, 2003; Kim et al.,
2004; Yang and Martin, 2004; Richardson-Burns et al., 2007b; Richardson-Burns et al.,
2007a). Some of these studies have also shown that these polymers can successfully be
incorporated with cell adhesion molecules, growth factors etc. to further enhance their
properties. With its superior electrochemical stability and biocompatibility, poly (3,4-
ethylenedioxythiophene) or more commonly known as PEDOT may be well suited for
chronic neural interfaces. Recent work suggests that PEDOT coatings can be deposited
over platinum electrodes and be used for chronic neural stimulation (Cui and Zhou,
2007). The impedance of PEDOT coated electrodes was found to be lower than the bare
platinum electrodes with corresponding lower voltage excursion to applied current pulses
in PBS. However, the stability of the PEDOT coated electrodes under chronic stimulation
regimes was found to depend largely upon the thickness of the coating that can be
controlled through deposition time. Physical degradation and changes in microstructure
of the film have been suggested as possible modes of failure. Hence, more work needs to
be done to make these polymers successful electrode materials for chronic neural
stimulation.
80
Thus the absolute values of parameters of safety for a stimulation electrode
material vary widely depending upon the electrode material. The experiments described
in this chapter were undertaken to characterize the safety limits of the stimulation
electrode used throughout this work.

3.1.2 Experimental Protocol
Three new stimulation electrodes were used for electrochemical characterization.
Impedance measurements were performed before and after cyclic voltammograms. These
measurements were performed in 0.1M PBS. In addition, voltage response to a range of
stimulus current levels in both PBS and in vivo rat eye was measured.

3.1.3 Results
3.1.3.1 Impedance Spectroscopy
Prior to using the electrodes for stimulation, an impedance spectrum was acquired
across wide range of frequencies (100 kHz – 100 mHz). Although the characteristic bend
in the magnitude is only slightly apparent at the highest frequencies, the phase value can
be observed to approach zero and slightly flatten out before dipping again. The dip may
be due to parasitic capacitance. The magnitude of impedance at 100 kHz was
approximately 4 kΩ.
81

Figure 3.1 Impedance spectrum of stimulation electrode in 0.1M PBS. Graph
illustrates both magnitude and phase of impedance spectrum between frequency
range of 100 mHz – 100 kHz.

3.1.3.2 Cyclic Voltammogram
As the stimulation electrode is not made of pure platinum but has about 20% of
iridium, cyclic voltammograms were generated in 0.1M PBS in order to estimate the
voltage limits where water hydrolysis was evident. Figure 3.2 illustrates the CV of the
stimulation electrode across different scan limits. At scan limits ranging from -3.5V to
2.5V, the CV clearly exhibited evolution of hydrogen and oxygen gas (Fig. 3.2B). This
corresponded to presence of a bubble at the electrode tip. On the other hand no hydrolysis
was evident at the more typical scan limits employed for pure platinum (Fig. 3.2A).
82

Figure 3.2 CV trace of Pt-Ir stimulation electrode in 0.1M PBS across scan limits
of -0.8V to 0.8V (A) and from -3.5V to 2.5V (B).

A
B
83
3.1.3.3 Voltage Response
In order to ascertain whether the applied stimulus levels would exceed the water
window of the stimulation electrode, voltage response of the electrode-solution interface
was analyzed. A range of current levels (10 µA – 100 µA) was applied to electrode in
0.1M PBS and in in vivo rat eye (degenerated retina). Figure 3.3 illustrates the voltage
that developed across the interface in response to three current levels (10, 60 and 100
µA). The voltage drop was slightly higher in the case of in vivo measurement compared
to PBS. The largest drop was associated with the highest applied current level was
approximately 2 V in the cathodic phase (Fig. 3.3B).
84

Figure 3.3 Voltage response across electrode-solution (A) and electrode-retina
(B) interface at different stimulus currents (10 µA, 60 µA and 100 µA).

A
B
85
3.1.4 Conclusion
Electrochemical characterization of the stimulation electrode exhibits features
slightly different than pure platinum. This may be due to presence of about 20% iridium
in the electrode. Also the largest voltage drop observed in vivo for the stimulation
electrode was within the voltage limits of water hydrolysis. Thus it can be assumed that
no hydrolysis products were generated during the stimulation protocols used in the rest of
the study.

3.2 Electrode Positioning
3.2.1 Background
Bioelectrical impedance measurement techniques have been used to characterize
properties of electrode-tissue interface in many different applications. In neural implants,
impedance measurement provides an important tool for assessing reactive changes at the
surface of the implant. Studies have shown that chronic implantation of neural probes in
the brain and the cochlea lead to reactive changes in the tissue and cells immediately
surrounding the electrode due to implantation (Xu et al., 1997; Liu et al., 1999b; Haberler
et al., 2000; Szarowski et al., 2003; Biran et al., 2005; Williams et al., 2007; Mercanzini
et al., 2009). Microglial activation, astrocytic reaction and extracellular matrices can all
form a reactive tissue encapsulation layer around the neural probe which in turn limits the
ability of the implanted probe to record neural signals reliably over extended periods of
time. Impedance measurement techniques have also been employed in in vitro and in vivo
conditions to characterize both the electrode and the surrounding environment (Weiland
86
and Anderson, 2000; Weiland et al., 2002a; Cui et al., 2003; Cui and Zhou, 2007; Shah et
al., 2007; Williams et al., 2007; Lempka et al., 2009; McConnell et al., 2009; Mercanzini
et al., 2009). They have also been used to determine tissue properties for different
pathologies and also to locate nerves for correct electrode placement (Prokhorov et al.,
2002; Micco and Richter, 2006; Bhati et al., 2009; Sauter et al., 2009).
For neuroprostheses aiming to restore lost functionality via electrical stimulation,
most often employ microelectrodes in order to limit the amount of tissue stimulated, thus
potentially creating finer resolution. However, this increases the electrode impedance,
which in turn might lead to a large voltage drop across the electrode-tissue interface for a
given stimulus-threshold. If the voltage drop exceeds the water window then it might lead
to hydrolysis. Also, large voltage drops put additional constraints on the power
requirements of the device.
One way to lower the excitation threshold and limit the spatial selectivity of stimulation
is to control the electrode-tissue distance. Auditory brainstem implant studies have shown
that for electrodes further away from the target neurons, higher stimulus strength was
required to elicit the required neural response indicating a higher threshold (Shannon et
al., 1997). Studies in the past have shown that the amount of current required to activate
neurons depends upon the distance between the electrode and the target neurons (Stoney
et al., 1968; Ranck, 1975; Tehovnik, 1996). Although there is evidence that this current-
distance relationship varies from linear to cubic, it has been best described by a quadratic
equation (equation 3.1). The earliest study done to evaluate threshold current-distance
relationship was in the sixties in pyramidal tract neurons (Stoney et al., 1968). This study
87
showed that excitation threshold changed with the square of distance between the
electrode at a particular depth and the minimal threshold. Based on these studies, an
equation was developed to describe the relationship between the excitation threshold I
th

and the minimal or absolute threshold I
R
for different distances R between the electrode
and the neural elements.

€
I
th
=I
R
+k⋅R
2
(3.1)
where k is defined as the current-distance constant and describes the excitability of the
neurons for different values of R. More recent computational models have been
developed to estimate the extent of neural activation based on the current-distance
relationship (Mahnam et al., 2008).
In the case of retinal prostheses too, studies have shown impedance to be
negatively correlated with increasing thresholds (Mahadevappa et al., 2005; de Balthasar
et al., 2008). As time after implantation increased, the electrode array was observed to lift
off the retinal surface in some test subjects. This in turn led to an increase in the
excitation thresholds and corresponding decrease in the impedance of the electrodes.
The retinal implant studies suggest that one way to sense the electrode-retina
proximity is by measuring the impedance at this interface. The electrode-electrolyte
interface has been a subject of interest ever since Volta in 1800 showed it to be a source
of electrical potential. Numerous circuit models have been designed to depict this
interface as closely as possible (Dymond, 1976; Bates and Chu, 1992; Geddes, 1997;
Franks et al., 2005). As explained in section 1.1.2.3, in its simplest form, the electrode-
electrolyte interface has been described as a parallel combination of a double layer
88
capacitance and resistance associated with the faradaic processes, in series with the
resistance of the target tissue. At high frequencies, where the capacitive double layer acts
as a short circuit, the impedance is dominated by the tissue resistance.
The present study focuses on using impedance measurement to sense the
proximity of the electrode to the epiretinal surface in in vivo condition. This method has
been developed as part of investigations involving epiretinal stimulation for retinal
prostheses.

3.2.2 Experiment Protocol
For all the experiments, adult Copenhagen rats were used. Initial experiments
involved insertion of stimulation electrode in the left eye and measurement of impedance
across a wide range of frequencies. In subsequent experiments, electrically evoked
responses were recorded in the right SC over a range of impedance values at a single
frequency. Finally, a few experiments were performed where the stimulation electrode
was held at either low or high impedance for a period of 1 hr in vivo. No recording was
carried out in the SC. The animals were given a post surgical recovery period of 3 days
after which the eyes were enucleated and prepared for histopthalogical examination to
assess the presence of any damage.

89
3.2.3 Results
3.2.3.1. Electrode Impedance as a Function of Distance from the Retina
In order to assess the feasibility of using impedance as a tool to sense proximity, a
comparison between the impedance spectrum of the electrode in PBS (in vitro setup) and
in vivo (N = 2) was performed. The in vivo impedance spectrum was seen to closely
match that obtained in vitro (Fig. 3.4). Impedance magnitude traces showed a resistive
behaviour at the high frequency range of the spectrum (Fig. 3.4A). This corresponded
well with the phase data that also showed it to be approaching 0° at the highest frequency
of 100 kHz (Fig. 3.4B). Based on this observation next, the sensitivity of impedance
value to the separation between the electrode and the retina at the single frequency of 100
kHz was tested (N = 3). After inserting into the eye, the electrode was advanced from the
periphery to a point close to the retina at which the impedance value was approximately
equal to 8 kΩ. This location was referenced as the point of no separation between the
electrode and retina (Fig. 3.5). From here on the impedance was observed to slowly
decrease as the electrode was advanced away from the retina in steps of 20 µm. Although
the reference electrode (in the tail) was positioned far from the working electrode, the
impedance magnitude was still sensitive to the change in resistance associated with the
vitreous and retina as a function of distance between electrode and retina. This is because
in the case of microelectrodes, the voltage drop due to solution resistance is concentrated
in the region near the electrode (Newman, 1966).
90

Figure 3.4 Graph illustrating impedance magnitude (A) and phase (B) in 0.1M
PBS (in vitro) and in rat eye (in vivo).

A
B
91

Figure 3.5 Graph illustrates the relationship between magnitude of impedance
and distance between electrode tip and retina.

As an additional measure, the sensitivity of impedance value to the separation
between the electrode and the retina at a low frequency of 1 Hz was tested. However, the
impedances measured at the different separations were less sensitive compared to the
impedances measured at high frequency (100 kHz) (data not shown).

3.2.3.2. Presence of Electrically Evoked Potentials in the SC as a Function of
Electrode Impedance
Next, experiments were conducted to determine the presence of robust electrically
evoked potentials (EEP) in the SC over different impedance ranges (N=3). For this, first
the area of lowest threshold in the SC was located. This area corresponded to the
92
retinotopically-mapped area stimulated by the electrode in the eye. After this, the
threshold of excitation was measured for different values of impedances. These
corresponded to different positions of the electrode’s tip relative to the retina. Figure 3.6
illustrates the dependence of threshold on electrode impedance. The threshold dropped at
a faster rate until an impedance of 8 kΩ was reached. Increasing the impedance above 8
kΩ further lowered the threshold but at a lower rate. The lowest threshold was observed
for impedance of 16 kΩ at which point a visible dimple made by the electrode on the
retina could be observed. The threshold current required to elicit EEPs at an impedance
of 8 kΩ corresponded to a charge density (0.03 ± 0.03 mC/cm
2
) that was within the safe
limit of platinum.

Figure 3.6 Graph illustrates the relationship between impedance and EEP
threshold for three separate experiments. Threshold is observed to decrease with
increasing impedance.
93
3.2.3.3. Retinal Damage due to Mechanical Contact between Electrode and Retina
Finally, experiments were conducted to determine whether there was any damage
to the retina at different impedance ranges. For this, the electrode was positioned to
correspond to either 8 kΩ (low impedance group; N=2) or >16 kΩ (high impedance
group; N=2) and held in position for 1 hour. The impedance was monitored continuously
and maintained at the required level during the 1-hour period. Histopathological
examination revealed focal damage to the retina in the case of high impedance group.
The damage included disorganization of underlying retinal layers and in one case, some
tearing of the epiretinal surface (Fig. 3.7B). No such damage was observed in the case of
low impedance group (Fig. 3.7A).

Figure 3.7 Hematoxylin & Eosin stained vertical sections of retinas from low
impedance (A) and high impedance (B) experiments. Retinal section from high
impedance exhibit severe disorganization of retinal layers due to pressure exerted
by electrode while retinal section from low impedance group exhibits normal
morphology. ONL: outer nuclear layer, INL: inner nuclear layer, GCL: ganglion
cell layer. Scale bar = 50 µm.

94
3.2.4 Discussion
This study demonstrates the feasibility of using impedance measurement as a tool
to sense proximity of the stimulation electrode to the retina. As the electrode was
advanced towards the retina, a corresponding increase in the magnitude of impedance and
a decrease in the threshold of electrically evoked response in the SC was observed. This
finding is similar to that reported in other studies (Tehovnik, 1996; Jensen et al., 2003;
Shah et al., 2007). Also, in the case of retinal prosthesis patients, imaging studies coupled
with impedance measurement has shown perceptual thresholds to correlate with the
distance between the electrode array and retina (de Balthasar et al., 2008).
One point of difference from other studies using impedance measurement is in the
choice of frequency for measurement. While in most other studies, impedance
measurement is carried out at 1 kHz, results in the present study showed that impedance
was most sensitive to proximity to retina only at a high frequency value of 100 kHz. This
corresponds with equivalent circuit models of the electrode-tissue/solution interface,
where at high frequencies, the impedance is observed to be predominantly governed by
the tissue/solution resistance.
Although threshold values were observed to decrease with increasing impedances,
at high impedance values, it increased the probability of causing retinal damage due to
accidental contact between electrode tip and retina. Also the aim of the study was to
position the electrode such that no damage due to contact occurred along with focal
stimulation and not to position it for lowest possible threshold. Due to these reasons, an
impedance of 8 kΩ was chosen as the best value to conduct all experiments. The
95
threshold for eliciting a response in SC at this impedance range is within the range
reported in similar studies. Also, the threshold charge density is within the safety limit of
the stimulation electrode.
A drawback of the current study is that no direct measurement could be made
between impedance, threshold and distance between electrode and retina. A reason for
this is that in order to make threshold measurements in the SC, it is critical to maintain
the impedance during that period. However, due to breathing movements made by the
animal, slight adjustment of electrode is needed to maintain the impedance. Separate
experiments were conducted that showed impedance increases with decrease in electrode-
retina separation. Experiments conducted in the future should focus on finding a
quantitative relationship between the measured impedance and the distance between the
electrode tip and retina. One way to find this relationship would be via OCT
measurement of the electrode-retina interface at different values of impedance.

96
Chapter 4
Morphological changes due to electrode contact and continuous
stimulation

4.1 Background
As mentioned in preceding sections, neuroprostheses employing electrical
stimulation have to operate within a safe and effective window of neural stimulation.
Parameters that govern safety of neural stimulation are total charge, charge density,
stimulus frequency, duration of stimulation and electrode arrangement and placement.
Studies have shown that electrical stimulation conducted at intensities outside the
safety limit of the stimulation electrode can lead to neural damage. A study by Agnew et
al. showed that when chronic intracortical stimulation was conducted in the cat
sensorimotor cortex, neural damage could be induced based upon the stimulus intensity
(Agnew et al., 1986). They used two different types of electrodes: a platinum electrode
with 30% iridium and activated iridium electrodes. Neural damage was observed with the
Pt-Ir electrode at charge density level of 3.2 mC/cm
2
while did not detect any damage at
this level of stimulation with activated Ir electrodes. They also observed electrode
dissolution to have begun at levels lower than the highest stimulation level in the case of
Pt-Ir electrodes. While a minimal amount of electrode dissolution was observed in the
case of activated Ir electrodes. Thus, the study showed that neural damage can be created
when electrodes are operated outside their safety limits. While remaining within the
97
safety limits of electrode is crucial in order to avoid damage due to products of
electrochemical reactions, studies have shown neural damage to be created even when
stimulation is performed within these safety limits. Chronic stimulation of cat cochlear
nucleus has shown to cause tissue injury near the microelectrode tips at levels three times
the excitation threshold (3 nC/phase) (McCreery et al., 1994). When the stimulation level
was increased to 6 nC/phase or higher, damage extended to a larger area (200 µm from
the microelectrode tip). A positive correlation was observed between the severity of
neural damage and applied charge per phase. The charge density values applied were
within the safety limits of the activated iridium electrodes used in the study. Chronic
stimulation of peripheral nerves within safety limits of stimulation electrodes has shown
to cause early axonal degeneration (EAD) characterised by the collapse of the myelin
sheath into the axoplasmic space (Agnew et al., 1989; McCreery et al., 1995). They
showed that continuous low frequency stimulation even at high stimulus amplitudes did
not induce neuronal damage. Thus, these studies exhibit the possibility of increasing the
margin of safety by reducing the stimulus frequency to the absolute minimum required in
clinical applications. Numerous other studies have shown that chronic stimulation can
induce neural damage with the type and severity depending upon the target tissue
properties and combination of stimulus parameters employed (Agnew et al., 1975;
Pudenz et al., 1975).
Along with histological damage, chronic electrical stimulation has also been
shown to cause elevation in excitation thresholds and metabolic changes in the biological
environment. Prolonged stimulation of the cat cerebral cortex at charge densities ranging
98
from 10 to 300 µC/cm
2
lead to an increase in the potassium and calcium ionic
concentrations. In some cases the increase in ionic concentrations was only transient
without any accompanying histological damage. However, as the stimulus intensity was
increased, the greater episodic fluctuations in the ionic concentrations was observed
along with neural damage in the form of shrunken neurons, widespread edema and
swollen axons and dendrites (Agnew et al., 1983). In a separate study, it led to an
increase deposition of calcium ions in the mitochondria of several cells and in
postsynaptic dendrites (Agnew et al., 1979).
So far, studies in the retina have been focussed on understanding the response
properties of retinal cells to electrical stimulation. Extensive data can be found on firing
patterns of retinal cells in response to both electrical and visual stimulation in in vitro
(Jensen et al., 2005b; Sekirnjak et al., 2006) and in vivo (Gekeler et al., 2004; Baig-Silva
et al., 2005; Sachs et al., 2005) preparations. Although different authors have reported
varying activation threshold values (correlating to their respective electrode designs and
animal models), together they provide strong evidence that electrical stimulation can
elicit visual responses. However, most of these studies have looked at the retinal
response to either single or brief train of pulses with only a few studies dedicated towards
looking at the effect of more long-term stimulation. One such study was performed in
rabbits via suprachoriodal-transretinal stimulation (STS) (Nakauchi et al., 2007). In this
study, 100 µm diameter Pt wire electrode was placed in the suprachoriodal space with the
return electrode positioned in the vitreous and continuous stimulation for 1 hr at 20 Hz
was carried out. They found the threshold for safe charge to increase approximately
99
linearly with increasing pulse duration while threshold for safe current to decrease
logarithmically with increase in pulse duration. Absolute values of damage threshold
varied depending upon the method of detection. Damage threshold detected through
fluorescein angiography (FA) for 0.5 ms pulse duration was found to be 0.6 mA. This
translates to threshold charge density of 1.3 mC/cm
2
for damage. In the case of
ophthalmoscopic observation, damage threshold was 1mA of current for 0.5 ms pulse
duration, which translates to a charge density approximately 2.12 mC/cm
2
. They
observed severe retinal damage when the applied current was around 1.5 mA (3.2
mC/cm
2
). At this level of stimulation, greater damage to the inner retina was observed
compared to the outer retina. In all cases of damage observed, the applied charge
densities were at least four times the safe charge density limit of platinum (0.35 mC/cm
2
).
With the return electrode placed on the vitreous side, majority of current can be assumed
to flow into the retina. At such high amplitude and charge density levels, it is not
surprising that severe retinal damage was observed. In the case of epiretinal stimulation,
most of the histological damage observed to date has been associated with electrode array
placement and accompanying retinal tack placement. Chronic implantation of passive
electrode arrays has shown retinal disorganization, some hyperpigmentation of the RPE,
retinal folding only near the site of tack insertion (Majji et al., 1999; Walter et al., 1999;
Guven et al., 2006). Some localized thinning was found in areas that were contacted
directly by the edge of the electrode array. Retinal area under the electrode array itself
was found to be normal. To date, very few studies involving epiretinal stimulation have
ventured into the domain of chronic stimulation. Studies in canine retina (normal and
100
RCD1) have shown that chronic epiretinal stimulation at low charge density levels of
0.05 and 0.1 mC/cm
2
do not cause any histological damage to the underlying retina
(Weiland et al., 2002b; Guven et al., 2005). In these studies, the retina was chronically
stimulated for up to 120 days (8-10 hrs per day). This chronic stimulation regime was
found to have no effect on the electroretinograms and visual evoked potentials. The
charge density levels applied in these studies have been found to be capable of evoking
cortical potentials and have been used in clinical testing of retinal prosthesis.
Thus, the present study was designed in an effort to understand the effect of
continuous epiretinal stimulation. Specifically, this study was designed to create damage
with a view towards understanding the histological consequences of applying charge
densities (~ 1 mC/cm
2
) that exceeded the safety limit of the stimulation electrode. The
stimulus frequency (100 Hz) applied was also much higher than that used in current
clinical applications in order to apply a high amount of total charge. Also, in this study
the effect of direct electrode contact on the retina was also studied in order to distinguish
the type and degree of damage created by it from that created by electrical stimulation.
Much of this work has already been published elsewhere (Ray et al., 2009).

4.2 Experiment Protocol
Twenty normal adult Long Evans rats were used with 4 to 5 rats per experimental
group. Single Pt-Ir microelectrode was inserted into left eye of each animal and
continuous epiretinal stimulation was performed for 1-hr. After cessation of stimulation,
101
animals received post surgical recovery period of 14 days after which both eyes were
enucleated and prepared for histopathological examination.
There were four experimental groups. In group 1, stimulation electrode was
placed in the vitreous and was held passively for 1-hr without any stimulation (0.0
µC/phase). In group 2, the stimulation electrode was placed near the retina and stimulated
for 1 hr at 0.09 µC/phase. In group 3, the stimulation electrode was placed in direct
contact with the retina but no stimulation was performed during the 1-hr period (0.0
µC/phase). In group 4, the stimulation electrode was placed in direct contact with the
retina and stimulated for 1-hr at 0.09 µC/phase. Charge-balanced, biphasic current pulses
1 ms long were delivered at 100 Hz during the 1-hr period.

4.3 Results
4.3.1 Gross Morphology
To begin, first the area contacted directly by the electrode was identified as 2.5
mm away from the optic disc towards the nasal direction in each retina. The adjacent
area was defined as the central retina area that was 1.5 mm away from the optic disc
towards the nasal side. For experimental groups in which the electrode did not directly
contact the retina, same regions as described above were used in order to facilitate
comparisons across groups. No change in the thicknesses of the retinal layers was
observed between the normal and experimental groups in the central retina area adjacent
to electrode contact (ONL, 38 ± 1 µm; OPL, 10 ± 2 µm; INL, 21 ± 1 µm; IPL, 48 ± 1 µm;
GCL, 15 ± 1 µm - Fig. 4.1A-E, M). In addition, no change was observed in the
102
thicknesses of retinal layers in the peripheral retinas of normal and experimental groups
(ONL, 22 ± 1 µm; OPL, 5 ± 1 µm; INL, 15 ± 1 µm; IPL, 25 ± 1 µm; GCL, 10 ± 1 µm --
Fig. 4.1F-J, N). However, changes in the thicknesses of the ONL and OPL in Groups 3
(Fig. 4.1K, O) and 4 (Fig. 4.1L, P) were observed in areas directly contacted by the
electrode (ONL, 6 ± 3 µm (p < 0.0001); OPL, 4 ± 1 µm (p < 0.0001)). In these areas the
thicknesses of INL (21 ± 1 µm), IPL (48 ± 1 µm) and GCL (15 ± 1 µm) remained
unchanged. Also, the extent of lesion observed in group 3 and 4 retinas was measured. A
statistically significantly greater extent of damage was observed in group 4 retinas (380 ±
18 µm) as compared to group 3 retinas (205 ± 16 µm, p < 0.0001) (Fig. 4.1R). Hence,
mechanical pressure alone, and mechanical pressure with high charge stimulation, both
can cause severe damage to the retinal area directly contacted by the electrode.

103

Figure 4.1 Photographs taken from 5-µm-thick vertical sections processed by
hematoxylin staining. Hematoxylin staining in normal (A, F) and experimental
groups (Group 1 - B, G; Group 2 - C, H; Group 3 -D, I; Group 4 - E, J) showed no
detectable changes in the thicknesses of electrode-adjacent area of retina (A-E, M --
1.5 mm away from the optic disc towards the nasal side). This held for the peripheal
(F-J, N) of the retinas. However, lesion areas were noted in the directly electrode-
contacted area of retina (K, L-- ≈ 2.5 mm away from the optic disc towards the nasal
side). The thickness of the ONL and OPL in Group 3 and 4 was significantly
reduced in retinal areas directly contacted by the electrode (O, P--the symbol *
represents p < 0.0001).

104
4.3.2 Expression of Synaptic Vesicle Proteins (SV2B and SV2B)
Next, despite a lack of gross morphological change, whether the synaptic
properties of retinal neurons were affected by the experimental conditions in electrode-
adjacent retinal areas was tested. For this SV2A (Fig. 4.2A), present in conventional
synapses in the IPL and cone terminals (Wang et al., 2003)) and SV2B (Fig. 4.2H-N)
present only in the ribbon-synapse-containing terminals of bipolar cells, and of rods and
cones (Wang et al., 2003) immunoreactivity was tested to determine whether they
showed normal expression patterns in experimental retinas. In normal (Fig. 4.2A, H),
group 1 (Fig. 4.2B, I), and group 2 (Fig. 4.2C, J) retinas, SV2A and SV2B
immunoreactivities were restricted to the OPL and IPL. In contrast, in group 3 (Fig.
4.2D, K) and group 4 (Fig. 4.2E, L) retinas, SV2A and SV2B-immunoreactive puncta
were dispersed in the ONL (Fig. 4.2F, G, M, N). Although abnormal expression of
SV2A and SV2B was observed in the outer retina, the expression pattern in the IPL was
normal.
105

Figure 4.2 Photographs of vertical sections of electrode-adjacent retinal areas
processed for SV2A (A-G) and SV2B (H-N) immunoreactivity in normal (A, H) and
experimental groups (Group 1 – B, I; Group 2 – C, J; Group 3 – D, K; Group 4 – E,
L). A-C: Labeling for SV2A was present in synaptic terminals in the OPL and IPL.
D, E: Labeling for SV2A was dispersed in ONL. H-L: Labeling for SV2B was
present in the OPL and IPL. K, L: Labeling for SV2B was dispersed in ONL. Scale
bar = 50 µm. F, G: Higher-power microphotographs of the same retinal field as in D
and E. M, N: Higher-power microphotographs of the same retinal field as in K and
L. Scale bar = 20 µm.

4.3.3 Abnormal Processes of Horizontal and Rod Bipolar Cells
Since an abnormal expression of SV2A and SV2B was found in the outer retinas
in groups 3 and 4, therefore, whether the experimental conditions also altered processes
of horizontal or rod bipolar cells was tested in electrode-adjacent retinal areas (Fig. 4.3).
106
In normal (Fig. 4.3A), group 1 (Fig. 4.3B), and group 2 (Fig. 4.3C) retinas, the somata of
calbindin-immunoreactive cells were strongly stained and found at the outer margin of
the INL. Moreover, calbindin-immunoreactivity was also present in the OPL. Similar
results were apparent in group 3 (Fig. 4.3D) and group 4 (Fig. 4.3E) retinas. In addition
to this, calbindin antibodies revealed a sprouting of processes from horizontal cells,
oriented towards the ONL (Fig. 4.3F, G). Also retinas in these groups exhibited
abnormal horizontal-cell somas that were not round in shape. The dendritic arbours of rod
bipolar cells in the Group 3 (Fig. 4.3K) and 4 (Fig. 4.3L) retinas were abnormally
oriented toward the middle of the ONL (Fig.4. 3M, N).
107

Figure 4.3 Photographs of vertical sections of electrode-adjacent retinal areas
processed for calbindin (A-G) and PKCα (H-N) immunoreactivity in normal (A, H)
and experimental groups (Group 1 – B, I; Group 2 – C, J; Group 3 – D, K; Group 4
– E, L). A-C: Labeling for calbindin was present in horizontal cells and OPL. D, E:
Labeling for calbindin showed sprouting processes oriented toward the ONL. H-L:
Labeling for PKCα was present in the rod bipolar cells. K, L: Dendritic trees of rod
bipolar cells was present in middle of the ONL. Scale bar = 50 µm. F, G: Higher-
power microphotographs of the same retinal field as in D and E. M, N: Higher-
power microphotographs of the same retinal field as K and L. Scale bar = 20 µm.

4.3.4 Double Immunofluorescence for SV2B with PKC alpha

To determine whether the abnormal, radially extended processes of rod bipolar
cells were coincident with the abnormal distribution of SV2B in the ONL, double-
labeling experiments were performed with antisera against SV2B and PKCα (Fig. 4.4).
In normal retinas (Fig. 4.4A), apical dendrites were confined to the OPL, where the
photoreceptor terminals are present. In contrast, in group 3 (Fig. 4.4B), the apical
dendrites of rod bipolar cells were extended into the ONL, thus appearing in close
apposition to the SV2B immunoreactivity. These results suggest that the apical dendrites
108
of rod bipolar cells follow the abnormal distribution of SV2B into the middle of the ONL.

Figure 4.4 Confocal micrographs of a vertical retinal section of electrode-
adjacent retinal areas processed for SV2B (green) and PKCα (red)
immunoreactivity in normal (A) and Group 3 (B) retinas. A: Double exposure
shows close contact between SV2B and PKCα at the OPL (arrows). B: Double
exposure shows close contact between SV2B and PKCα at the OPL (arrow) and
middle of the ONL (arrowheads). Scale bar = 10 µm.

4.3.5 No Morphological Changes in Inner Retina
In the mammalian retina, most amacrine cells contain either glycine or γ-
aminobutyric acid (GABA -- (Vardi and Auerbach, 1995)). Therefore, labelling with
antisera against the GABA synthesizing enzyme, glutamic-acid decarboxylase (GAD65),
and glycine transporter-1 (Glyt-1 – (Haverkamp and Wässle, 2000)) was performed to
probe for abnormalities in adjacent retinal areas (Fig. 4.5). No changes in the overall
morphology and expression pattern of Glyt-1 (Fig. 4.5A-E) and GAD65 (Fig. 4.5F-J)
immunoreactive amacrine cells were observed. These results were consistent with the
normal SV2A expression pattern that we found in the IPL for all groups (Fig. 4.2).
109

Figure 4.5 Photographs of vertical sections of electrode-adjacent retinal areas
processed for Glyt-1 (A-E) and GAD65 (F-J) immunoreactivity in normal (A, F) and
experimental groups (Group 1 – B, G; Group 2 – C, H; Group 3 – D, I; Group 4 – E,
J). A-E: Labeling for Glyt-1 was present in numerous amacrine cells in the INL and
processes in the IPL. F-J: Labeling for GAD65 was present in numerous amacrine
cells in the INL and processes in the IPL. Expression pattern of Glyt-1 and GAD65
was consistent in normal and experimental conditions. Scale bar = 50 µm.
4.3.6 Glial Expression
In addition to the major neuronal types in the retina, Müller cells, the major glial
cells were also examined in both normal and experimental retinas. GFAP upregulation is
known to occur in response to stress or injury in the retina. In normal (Fig. 4.6A) and
Group 1 (Fig. 4.6B) central retinas, only weak GFAP immunoreactivity was observed,
specifically in the Müller cell endfeet located in the nerve-fibre layer. However,
pronounced GFAP immunoreactivity was observed in the inner retina of group 2 (Fig.
4.6C). In group 3 (Fig. 4.6D) and group 4 (Fig. 4.6E) retinas, GFAP expression
increased greatly. This increase was due to gliosis in Müller cells (DiLoreto et al., 1995;
Eisenfeld et al., 1984; Roque & Caldwell, 1990; Tanihara et al., 1997). This was evident
110
as GFAP immunoreactivity was no longer restricted to the nerve-fibre layer but instead
extended throughout the length of the Müller cells (Fig. 4.6D, E).

Figure 4.6 Photographs of vertical sections of the rat retina processed for GFAP
immunoreactivity in normal (A) and experimental groups (Group 1 - B; Group 2 -
C; Group 3 - D; Group 4 - E). A, B: Weak GFAP immunoreactivity was observed
at the nerve fiber layer. C: GFAP immunoreactivity was pronounced in the inner
retina of the Group 2. D, E: GFAP expression was no longer limited to the retinal
nerve fiber layer or inner retina, but extended throughout the whole length of the
Müller cells. Scale bar = 50 µm.

4.4 Discussion
The present study demonstrates that retinal damage can be induced by direct
contact of electrode tip with the retina. The extent of damage is statistically significantly
larger when additional continuous stimulation is performed at 0.09 µC/phase. This charge
per phase translates to a charge density of approximately 1 mC/cm
2
for the stimulation
electrode geometry. Studies in the cortex have shown that continuous stimulation at this
level of charge density induces neural damage. Retinal damage included disorganization
of the layers underneath the area contacted directly by the electrode.
Although the gross morphology of retinal areas adjacent to that contacted by the
electrode was normal, abnormal distribution of SV2A and SV2B was observed in the
outer retinas of Groups 3 and 4 in the electrode-adjacent areas. This abnormal expression
111
pattern of synaptic vesicle proteins in the outer retina may be due to retraction of
photoreceptor terminals. The retraction of rod photoreceptor synapses is a feature that is
evident in other overt insults of retinal tissues such as detached retinas (Fisher et al.,
2005), and in diseases such as retinitis pigmentosa (Cuenca et al., 2005). Double
labelling showed very close contact between the dendritic trees of rod bipolar cells and
photoreceptor terminals in the middle of the ONL. Similar changes in rod bipolar cell
dendrites have been noted in age-related macular degeneration (Sullivan et al., 2007),
detached retinas (Fisher et al., 2005), and in retinitis pigmentosa (Cuenca et al., 2005)
due to retraction of photoreceptors terminals. Thus, for a retinal prosthesis, it is
important to place the electrode array without exerting any pressure on the retina as this
can cause damage in areas directly underneath the electrode array as well as areas
adjacent to it.
On the other hand, GAD65- and Glyt-1-immunoreactive amacrine cells did not
appear to be affected by any of the experimental conditions in electrode-adjacent areas. In
addition, no change in GCL thickness was observed. Despite electrode placement on the
ganglion cell side, and the high charge stimulation, ganglion cells seemed unaffected.
This suggests that an epiretinal electrode array placed on a degenerate retina may be
tolerated well although this would have to be tested. It is somewhat surprising that
although mechanical pressure was applied on the ganglion cell side, the inner retinas
were not severely affected. Instead morphological changes were observed in the outer
retina (Fig. 4.1K, L). In a number of retinal disorders, including macular degeneration,
retinal detachment, diabetic retinopathy, retinopathy of prematurity, and retinitis
112
pigmentosa (RP), photoreceptors are highly vulnerable and undergo apoptosis (Cook et
al., 1995; Arroyo et al., 2005). Recently, it has been noted that Monocyte
chemoattractant protein 1 (MCP1) plays a critical role in mediating photoreceptor
apoptosis in an experimental model of retinal detachment. Retinal detachment led to
increased MCP1 expression in the Müller glia and increased CD11b
macrophage/microglia in the detached retina (Nakazawa et al., 2006). As increased
GFAP expression was observed in Group 3 and Group 4 retinas, it is conceivable that
MCP1 levels may have increased in the glia to cause cell death in photoreceptor cells in
Group 3 and Group 4 in areas directly contacted by the electrode.
Another interesting finding of the study was that no retinal damage was observed
when continuous stimulation at 0.09 µC/phase was performed without direct retinal
contact. A possible reason for this could be that as the electrode was placed some
distance away from the retina the current could have been shunted away by the vitreous.
Another reason may be that the vitreous might act as a good source for removing toxic
products generated by the electrochemical reactions taking place at the interface. This
may imply that without direct retinal contact, high charge stimulation may be well
tolerated by the retina. However, it still needs to be tested whether similar results will be
observed when the electrode is placed at an optimal distance from the retina for focal
retinal stimulation to take place without damage due to direct contact.
113
Chapter 5
Elevation in Excitation Threshold due to Continuous Stimulation

5.1. Background
Electrical stimulation via microelectrodes causes excitation of a localized
population of neurons thereby allowing selective stimulation that is otherwise not
possible with macroelectrodes. In the preceding section, it was shown that chronic
stimulation at sufficiently high levels of charge density, charge/phase and stimulus
frequency can cause irreversible tissue damage to central and peripheral nervous system.
In many cases of such chronic stimulation with microelectrodes, a depression in the
neuronal excitability has been observed to precede or occur in the absence of
histopathological damage (Agnew et al., 1986; McCreery et al., 1986). For about 80
years, it has been known that short periods of electrical stimulation at rates sufficiently
high to ensure that stimuli occur within the neuron’s relative refractory period cause a
reduction in neural excitability or ‘neural fatigue’ (Forbes and Rice, 1929). Since then
studies have shown that such a depression in nerve excitability is dependent upon the
duration and rate of evoked response (Gasser, 1935; Katz, 1939; Bowman and McNeal,
1986).
Comprehensive work done in the field of auditory prostheses has demonstrated a
strong correlation between depression in neuronal excitability and stimulus frequency. In
addition, charge per phase and not charge density was found to be strongly correlated to
the depression of neuronal excitability. These studies brought to light an interesting
114
observation that chronic stimulation at levels that did not cause histological damage still
caused a depression in the neural excitability. This depression was reversible in nature
with its recovery time dependent upon factors such as charge per phase, frequency and
duration (Shepherd et al., 1983; Tykocinski et al., 1995; Huang and Shepherd, 1999,
2000). Based on their findings, McCreery and colleagues coined the term ‘SIDNE’ or
‘Stimulation Induced Depression of Neuronal Excitability’ to represent this
phemomenon. They found that chronic stimulation for 4 hrs at 200 Hz with a single
microelectrode in cat cochlear nucleus can cause SIDNE even at low to moderate levels
of charge per phase (McCreery et al., 1992). They found no evidence of any histological
damage, which correlated well with their observation that the induced depression was
transient in nature. They found that the severity of SIDNE and the time to recover
depended upon the applied charge per phase. In a subsequent study they attempted to
characterize the effect of such a chronic stimulation of the cochlear nucleus on the
evoked potentials recorded in the inferior colliculus (IC) (McCreery et al., 1997). They
found that 7 hrs of continuous stimulation at rates between 100 – 500 Hz and current
levels only slightly above threshold can cause a depression in the neuronal excitability.
They generated response growth functions (RGFs) at a lower frequency of 50 Hz before
and after the chronic stimulation at high frequency. The RGFs represented the amplitude
of the first potential of the recorded evoked response in the IC. As the latency of the first
potential was very short (~ 1 ms), they assumed that this potential represented direct as
opposed to transsynaptic activation. An interesting observation made in this study was
that the slope of the RGFs was steeper after the chronic stimulation at current levels
115
above the applied current level for chronic stimulation. This indicated that neurons that
could earlier be excited by the applied current level were now in relative rather than
absolute refractory. With a more intense current level, they could be recruited after the
chronic stimulation. From this McCreery and colleagues concluded that SIDNE differed
from long-term depression (LTD) as it did not involve changes in synaptic efficacy. One
important finding of the study was the SIDNE could be reduced by rotating the stimulus
over several microelectrodes thus, reducing the duty cycle at each. Based on this they
extended the study to include chronic stimulation for 7 hrs at 250 Hz and 50% duty cycle
for 10 – 21 days (McCreery et al., 2000). In addition to generating RGFs at a lower
frequency before and after the chronic stimulation, they also included RGFs generated
during the first or last hour of chronic stimulation at high frequency. They called the
former non-embedded RGFs while the latter embedded RGFs. As observed in the
previous studies, chronic stimulation caused a decrease in the non-embedded RGFs when
stimulated at 250 Hz. In addition, they noted a decrease in the embedded RGFs and
called this depression ‘SANR’ or ‘Short Acting Neuronal Refractory’. Both SIDNE and
SANR were found to be strongly dependent upon the stimulus frequency.
An important part of the result in the above studies is that the depression in
neuronal excitability was always observed at stimulus levels close to threshold. At higher
amplitudes, no appreciable change in response was observed. One can thus infer that
chronic stimulation at levels close to threshold and sufficiently high stimulus rates,
although do not cause any histological damage, cause a depression in the excitability of
neurons closest to the stimulation electrode.
116
In the development of retinal prostheses, parameters such as total charge/phase,
charge density and stimulus frequency play an important role in determining the
technology used for the implantable system. With increasing efforts towards developing
high-resolution prostheses, feasibility of encoding brightness and size of visual percepts
via amplitude and stimulus frequency is being explored (Greenwald et al., 2009;
Horsager et al., 2009a; Horsager et al., 2009b). As progress is being made in providing
patients with retinal prostheses, it can be envisaged that in the near future, patients would
be allowed to use these devices for longer periods of time. Retinal ganglion cells have a
mean spike rate of around 14 Hz (Margolis and Detwiler, 2007; Stasheff, 2008) but are
known to be capable to following stimulus rates as high as 300 Hz (Sekirnjak et al.,
2006). Therefore it is important to study the effect of such high frequency stimulation in
the resulting evoked response. Only a few studies to date have attempted to analyze the
effect of prolonged stimulation of the retina on the resulting evoked response (Jensen and
Rizzo, 2007; Cohen, 2009; Ryu et al., 2009). Moreover, these studies were carried out in
in vitro preparation of retina and only brief periods of electrical stimulation were
performed in order to understand the effect of stimulus frequency on the resulting evoked
response. Similar to results in the cochlear nucleus, a study by Cohen demonstrated a
positive correlation between the applied charge per phase and depression and subsequent
recovery of light-evoked responses in rabbit retina to 1 min long epiretinal stimulation at
50 Hz (Cohen, 2009). In addition, ganglion cell firing was observed to diminish over the
1 min period of continuous stimulation indicating some form of depression. He found that
the charge density to create depression was higher than threshold charge density of retinal
117
ganglion cells. Also, higher depression was observed to be induced with the electrode
placed closer to the retina. This is in concordance of a study showing that stimulation
thresholds are lowest when the stimulation electrode appears to be in contact with the
retinal surface (Jensen et al., 2005a). While the aforementioned study employed 1 min
long stimulation, an earlier study found pairs of biphasic stimulus pulses were capable of
creating a depression in the amplitude and number of spikes of the resulting RGC
responses (Jensen and Rizzo, 2007). A subsequent application of more biphasic pulses
only created a moderate depression. Depression was found to increase with increasing
stimulus frequency (1.5 – 65 Hz). In addition, a sharp decline in the amplitude of
electrically evoked response was observed after 25 Hz while the cells were unable to
follow light stimulation at 40 Hz. A support for this finding can be found in the fact that
the critical flicker frequency of rabbits is around 26 – 27 Hz.
While providing valuable insight into the effect of continuous stimulation of the
retina, these studies further exemplify the need for a more comprehensive study to look at
the effects not only at the level of the retina but also in higher order visual processing
centres. Towards this end, the present study is dedicated towards understanding and
evaluating the effect of continuous epiretinal stimulation on both retinal morphology and
resulting electrically evoked potentials (EEPs) in the superior colliculus (SC).

5.2 Experiment Protocol
Normal Copenhagen rats were used for all experiments. One-hour epiretinal
stimulation was performed and resulting EEPs were recorded in the SC. The stimulation
118
electrode was positioned close to retina via the impedance measurement technique
described in section 3.2. For experiments evaluating retinal morphology, only epiretinal
stimulation was performed following which animals were given a post surgical recovery
period of either 3, 7 or 14 days. Subsequently the eyes were enucleated and prepared for
histopathological and immunocytochemical analysis. For experiments evaluating the
effect on the evoked response in the SC, EEPs were generated before and after 1-hr
epiretinal stimulation. Animals were sacrificed immediately after completion of the
experiment. Table 5.1 lists the experimental groups and parameters followed in each.

Experimental
group
Charge
density
(mC/cm
2
)
Stimulus
Frequency
Data Collected (Number of
experiments, n)
Normal NA NA Immunohistochemistry (n=2)
Control NA NA Immunohistochemistry (n=6), SC
recording (n=4)
Low-frequency 0.68 2 Immunohistochemistry (n=8)
Mid-frequency 0.1 20 Immunohistochemistry (n=4), SC
recording (n=4)
High-
frequency
0.68 300 Immunohistochemistry (n=4), SC
recording (n=5)

Table 5.1 Experimental groups with experimental parameters and data
collected for each group.

119
5.3 Results
5.3.1 Retinal Morphology
Initial investigation in the study involved analyzing the retinal morphology after 1- hour
epiretinal stimulation in all the experimental groups.

5.3.1.1 Histology
Hematoxylin staining of stimulated retina across all the different experimental
groups showed no disruption of the different retinal layers. Comparison between normal
and experimental groups showed that the integrity of individual retinal layers had been
maintained under the different stimulation conditions (Fig. 5.1). The only disruption
noted was at the site of electrode insertion.
120

Figure 5.1 Photographs of 25 µm thick retinal sections processed with
Hematoxylin staining. Hematoxylin staining showed gross morphological changes
due to any of the experimental conditions across entire length of retina. A- Normal,
B - Control, C - High Frequency. Higher magnification insets exhibit no layer
disorganization in any of the experimental groups. Scale bar = 50 µm.
121
In addition to gross histology, individual retinal cell morphology was also
examined for all the different experimental groups. In control and experimental group
retinas, the dendrites of rod bipolar cells were observed at the outer plexiform layer
(OPL) similar to that observed in normal retina (Fig. 5.2a-c) via PKCα immunoreactivity.
The terminals of these cells could be found as punctate staining in the inner part of the
inner plexiform layer (IPL). In addition to this, normal morphology of ON bipolar cells
was observed through labelling with an antibody directed against the G-protein (GOα)
(Fig. 5.2g-i). GOα has been shown to be expressed in ON bipolar cells of cat and primate
retina (Haverkamp and Wässle, 2000). Next, the effect of stimulation on the morphology
of horizontal cells was examined. The processes of horizontal cell could be viewed at the
OPL while cell bodies were present in the outer part of the INL in both normal and
experimental retinas (Fig. 5.2d-f). In addition to assessing the morphology of different
cell types found in the retina, expression of synaptic vesicle proteins was also performed.
SV2B expression pattern was examined and found to be restricted to the OPL and IPL in
normal and experimental retinas (Fig. 5.2j-l). Both GABA and glycine labelled amacrine
cells were observed to form two rows of cell bodies at the inner margin of the INL in
normal and experimental retinas (Fig. 5.2m-r). Finally, as the stimulation electrode was
closest to retinal ganglion cells, their morphology was assessed through antibodies
directed against microtubule-associated protein 1 (MAP1) (Okabe et al., 1989). Retinal
ganglion cell bodies were restricted to the ganglion cell layer and processes at the IPL in
normal and experimental retinas (Fig. 5.2s-u). All the results described above are based
on the analysis of stimulated retinas with a 3-day post stimulation recovery period. In
122
order to ensure that the post stimulation recovery period had no effect on retinal
morphology, analysis of stimulated retinas in the same fashion as described above was
carried out for two additional recovery periods within the low-frequency and control
group. Similar results as described above were observed for recovery period of 7 and 14
days (data not shown) indicating that recovery period did not have an effect on the lack of
any morphological alteration due to stimulation.

123

Figure 5.2 Photographs of 25 µm thick vertical retinal sections processed for
PKCα (a-c), Calbindin (d-f), GOα (g-i), SV2B (j-l), Glycine (m-o), GABA (p-r) and
MAP1 (s-u) immunoreactivity for normal, control and high-frequency group retinas
exhibiting normal morphology of the major retinal cells and synaptic vesicles. Scale
bar = 50 µm.
124
5.3.1.2 Glial Reaction
In normal, healthy retina, GFAP immunoreactivity was restricted to the nerve
fibre layer (Fig. 5.3a). However, weak gliosis was observed in the control group retina
(Fig. 5.3b), while a more pronounced gliosis was evident in the high frequency
stimulation group retinas (Fig. 5.3c) over the entire retina. Müller cell processes were
observed to span across the entire length of the retina – inner to outer. Similar to that
observed in the high frequency stimulation group, pronounced gliosis was observed in
both low and mid frequency stimulation group retinas (data not shown). In addition,
GFAP upregulation observed for stimulated retinas analyzed after post stimulation
recovery periods of 7 and 14 days were found to be similar to those analyzed after 3 days
(data not shown).

Figure 5.3 Photographs of 25 µm thick vertical sections processed for GFAP
immunoreactivity after 3d recovery period. For normal retina, GFAP
immunoreactivity was observed only in astrocytes in the nerve fibre layer (a). Weak
GFAP immunoreactivity was observed in control retina (b) while a more
pronounced immunoreactivity was observed in high-frequency group retina (c).
Scale bar = 50 µm.

125
5.3.2 Electrically Evoked Response in SC
5.3.2.1 Threshold
Separate experiments were carried out using the same stimulation parameters to
evaluate the effect of 1-hr stimulation on the threshold of EEPs in the SC (Table 5.1).
Comparison of threshold values before and after 1-hr stimulation for control, mid and
high frequency groups was carried out by computing a threshold ratio for each
experiment. The threshold ratio was computed by dividing the EEP threshold post-
stimulation by EEP-threshold pre-stimulation. In 8 out of 9 cases of 1-hr stimulation at
mid and high frequency, the threshold ratio was observed to be greater than 1 indicating
an elevation in the EEP threshold due to the 1-hr stimulation (Fig. 5.4). Comparison of
the absolute values of threshold between pre and post stimulation demonstrated that the
elevation in threshold due to stimulation was statistically significant (p=0.02). On the
other hand, a similar comparison of absolute values of threshold in control group of
experiments demonstrated that the changes observed in EEP threshold after the 1-hr
period were not statistically significant (p=0.6).
126

Figure 5.4 Threshold change in EEPs across control, mid-frequency and high-
frequency groups. Threshold ratio = EEP Threshold After/EEP Threshold Before.
Increase in threshold was observed for all cases of stimulation. Dashed line
represents no change in threshold.

5.3.2.2 Response Feature Analysis
Pre and post-stimulation EEPs (average of 50 traces) were compared to examine
whether the same response features were present after the 1-hr stimulation period. A
representative EEP with the different features is illustrated in figure 5.5. In all but one
experiments, a prepotential as described to precede a large negative wave in juxtazonal
potentials was observed (McIlwain, 1978). The prepotential (P1) had an average latency
(5.04 ± 1.4 ms) and was clearly separated from the stimulus artifact. Although there was
a change in latency of N1-N3 for all stimulation and control experiments after the 1-hr
period, this difference was not statistically significant (p > 0.05). Similarly, the valley-to-
127
peak amplitude of each of the individual components (N1-N3) of the EEPs was found to
be different after 1-hr stimulation or control period. However, these differences were
found to be statistically insignificant (p > 0.05).

Figure 5.5 Example of electrically elicited potential (EEP) in the superior
colliculus (SC). Trace illustrates small prepotential (P1) followed by negative
deflections (N1-N3) in the recorded EEP. Shaded area represents stimulus artifact.

In addition to analyzing the pattern of response at threshold, the power of EEPs
elicited at different stimulus levels was compared before and after the 1-hr period. For
this, the total area of EEP elicited (single trace) was computed at increasing stimulus
amplitude levels and were plotted against the stimulus current level (stimulus response
curves). In the case of mid-frequency group, a reduction in the power of EEPs was
observed in three out of the four cases. This reduction was observed more at stimulus
levels close to threshold (Fig. 5.6). However, this result was observed in only one out of
128
the five cases of high-frequency group experiments. In the case of control group, none of
the experiments exhibited a reduction in the power of EEP across the different stimulus
levels.

Figure 5.6 Graphs illustrating the computed power of EEPs generated at
increasing levels of stimulus current. A decrease in the EEP power can be observed
in the mid (A) and high (B) frequency group at lower stimulus levels. No such clear
decrease in EEP power was observed in control group (C).

5.4. Discussion
The results in the present study demonstrate that one-hour epiretinal stimulation
without direct contact with retina does not cause any structural changes in the retina.
However an upregulation of GFAP was observed in stimulated tissue. This result is
consistent with the results presented in chapter 4 where group 2 retinas did not exhibit
129
any morphological changes after 1-hour stimulation at high charge and frequency. In both
cases, there was no direct contact between the electrode tip and retina. However, in group
2, the electrode was placed in a somewhat qualitative manner away from the retina so it
was unclear whether enough current was reaching the retinal surface to cause neuronal
excitation. On the other hand, in the present case, by using impedance to sense the
proximity of the electrode to retina, the electrode could be positioned reliably in the same
fashion in each experiment without causing any damage due to direct contact. At the
same time, recording in the SC exhibited presence of robust EEPs at threshold values
reported in other studies. This proves that in the present study, retinal cells were
stimulated under each of the experimental conditions. Despite this, under the most severe
stimulation parameters tested (high frequency group), the major retinal cells exhibited a
normal morphology. Hence, it can be concluded that 1-hour stimulation at charge density
higher than the safe limit of platinum and at high stimulus frequency does not induce any
morphological changes in the retina.
On the other hand, in all but one case of 1-hr stimulation, a statistically significant
increase in the threshold of evoked response in the SC was observed. The choice of
stimulus frequencies was based upon those used currently in clinical experiments
(Nanduri et al., 2008; Horsager et al., 2009b) and as the highest stimulus frequency that
retinal ganglion cells have been reported to be capable of following (Fried et al., 2006;
Sekirnjak et al., 2006). The occurrence of threshold elevation in the absence of any
morphological damage has been observed in chronic stimulation studies in the CNS
(Agnew et al., 1986; McCreery et al., 1986; McCreery et al., 1997; McCreery et al.,
130
2002). However, it is still unclear what the mechanism behind such depression in
neuronal response is. A study performed in the auditory nerve showed a decrease in the
amplitude of electrically evoked auditory brainstem response following prolonged
stimulation at high stimulus frequencies (Huang and Shepherd, 2000). The study showed
this depression to be present in the case of standard platinum electrodes with small active
surface area. They found no such change when they used instead platinum electrodes
with high active surface area. In the present study, the stimulation electrode has a surface
area (75 µm diameter disk) much smaller than the one used in the above study (0.3 mm
square band). Hence this could be a possible reason for the elevation in threshold of EEP.
In addition to measurement of EEP threshold, the pattern of response elicited pre
and post 1-hr stimulation was also compared. Although there were some changes in the
absolute values of the amplitude and latencies of each of the individual peaks of the
EEPs, these changes were not found to be statistically significant. Overall, a similar
pattern of activity could be evoked after 1-hr stimulation albeit at a higher stimulus
intensity. This implies that the stimulated cells are in relative rather than absolute
refractory and that given enough current can still be excited. Also, in some of the
experiments, stimulus response curves generated before and after 1-hr stimulation
exhibited that neurons closest to the stimulation electrode were affected more severely.
Other studies have shown similar results, where a similar pattern of activity could be
evoked at higher stimulus intensities after prolonged stimulation (McCreery et al., 1997;
McCreery et al., 2002). Although the present study did not test whether the observed
depression was transient in nature, due to the lack of any morphological change, it can
131
perhaps be assumed that the depression observed may be temporary rather than
permanent. Future studies in this field need to be directed towards determining possible
mechanism(s) behind this depression.
Thus, epiretinal stimulation for 1-hr at threshold and high charge densities and
low and high stimulus frequencies can induce an elevation in the threshold of evoked
response in the SC without any morphological changes in the retina. The experimental
conditions tested in the present study represent a far from natural stimulation paradigm
while in actual use the amplitude will be constantly varying. Hence, these experiments
were carried out to test a first order experiment in a well-controlled fashion. Despite this,
the results of the study have significant implications in the design of a retinal prosthesis,
as it implies that continuous usage of the device at constant amplitude and frequency
levels may lead to a depression in the resulting neural response. It is possible that this
depression may be a protective mechanism to avoid permanent damage to the stimulated
structure. Whether this depression occurs at the level of the retina or SC or both remains
at present unclear. Hence, along with an additional parameter of total duration of
stimulation, time-varying stimuli should also be explored during development of efficient
stimulation protocols. Time-varying stimuli have been used in the case of auditory
prosthesis to simulate a moderately noisy acoustic environment (McCreery et al., 2000)
and perhaps can be used to better encode a natural visual environment.
132
Chapter 6
Temporal Dynamics of Excitation Threshold Increase

6.1 Background
The results presented in section 5 have demonstrated that prolonged stimulation
of the retina induces depression in neuronal excitability. This depression was observed to
occur at both threshold and high stimulus intensity levels when applied at rates higher
than the natural firing rate of retinal ganglion cells. However, stimulus threshold was
only evaluated before and after 1 hour of stimulation. The experiments described in this
chapter make intermediate measurements to determine the time course of depression.
Other investigators have demonstrated similar results. Cohen (2009) has shown epiretinal
stimulation at 50 Hz to decrease light evoked responses and also to affect the spike rates
of retinal ganglion cells. Jensen et al. (2007) showed a decrease in retinal ganglion cell
response following a train of stimulus pulses delivered at stimulus rates ranging from 1.5
to 65 Hz.

6.2 Experiment Protocol
For the following set of experiments normal Copenhagen (n=7) and retinal
degenerate (RD) (n=2) rats were used. Continuous stimulation was performed at 60 µA,
20 Hz for brief durations. After each application of continuous stimulation, EEPs were
recorded at 1.5 times threshold. The strength of EEPs was quantified by calculating the
133
total area of response. The strength of EEPs was monitored towards the end of each
experiment in order to determine if there was any change in the response in the absence
of continuous stimulation. Two types of stimuli were used. Test stimuli was a train of
pulses (60 µA, 0.5 ms) delivered at 20 Hz for different durations (1s to 30min). Each test
stimuli was preceded by a probe. The probe was either a single stimulus pulse or a train
of 10 stimulus pulses (applied at 0.2 Hz) with amplitude equal to 1.5 times pre stimulus
threshold. During the first part of the experiment, test stimuli interleaved with probes
examined the effect of continuous stimulation on EEPs recorded in the SC. After a brief
(20s) pause during which the programme was loaded, only probe pulses were applied to
monitor the recovery of EEPs in the absence of any continuous stimulation.

6.3 Results
6.3.1 Effect of Stimulation Duration
As previous results showed an elevation in the threshold of EEPs after 1-hr of
stimulation, experiments were carried out to determine whether the total duration of
stimulation had an effect (n=3). EEPs were recorded after application of train of test
stimuli of different durations (train length = 30s, 1min, 5min, 10min, 20min, 30min) in
rapid succession. Prior to applying each train of test stimuli, a train of 10 probe pulses
was applied. After the application of the last test stimuli (train length = 30 min), only
probe pulses were applied every 10min to monitor the recovery of the EEPs in absence of
continuous stimulation. Figure 6.1 illustrates the experimental protocol followed here.

134

Figure 6.1 The experimental protocol followed to determine the effect duration
of stimulation on the strength of EEPs recorded. After a brief pause (20s) only the
probe was applied.

The EEP strength was observed to decrease after 30s of continuous stimulation
(Fig. 6.2-inset). The strength of EEP did not decrease appreciably on further application
of increasing duration (Fig. 6.2). After the application of 30 min of stimulation, the EEPs
were monitored every 10 mins. Some recovery in the strength was observed. Along with
recording EEPs at 1.5 times threshold, EEPs were also recorded at 60 µA (n=5). No
obvious change was observed in the strength of the EEPs generated at 60 µA (Fig. 6.3).

135

Figure 6.2 Graph illustrates the effect of short bursts of continuous stimulation
(60 µA, 20 Hz) for different durations (30s, 1min, 5min, 10min, 20min and 30min)
on the strength of EEPs elicited at 1.5 times threshold for three separate
experiments (normal). After the last stimulation of 30mins, the strength of EEP was
monitored every 10mins. Inset shows the EEP strength during the initial few
minutes. Data has been normalized to the first data point for each experiment.

Figure 6.3 Graph illustrates the effect of short bursts of continuous stimulation
(60 µA, 20 Hz) for different durations (30s, 1min, 5min, 10min and 20min) on the
strength of EEPs elicited at 60 µA (n=5). Data has been normalized to the first data
point for each experiment.
136
6.3.2 Effect of 1s Stimulation
Based on the above result, continuous stimulation (60 µA, 20 Hz) for 1s was
applied and EEPs recorded at 1.5 times threshold after each application (n=2). The total
duration of application was varied from 30s to 5mins. In this set of experiments, train of
test pulses was applied for a train length equal to 1s. A single probe pulse was recorded
prior to the application of the test pulses. After a delay of 0.5s, the combination of probe
pulse followed by train of test pulses was applied and this was repeated for different
durations of stimulation (Fig. 6.4A). Following this stimulation regime, single probe
pulses were applied every 3s for different durations to monitor the recovery of the EEPs
(Fig. 6.4B).

Figure 6.4 Experimental protocol followed to investigate the effect of short
duration of continuous stimulation.
A
B
137
A single EEP was divided into an early (latency: 3 ~ 12 ms) and late (12 ~ 40 ms)
response. The strength of EEP was observed to decrease within the first 30s. Subsequent
application of continuous stimulation did not cause a further decrease in EEP strength.
Once the stimulation was stopped, a rapid recovery of EEPs was observed. This trend
was more obvious in the case of early response as compared to the late response. The
experimental protocol was repeated using old RD rats (~700 days old, n=2). A similar
decrease in the strength of EEPs was observed in the first 20s after which there was no
further decrease (Fig. 6.6). Similar to that observed in normal rats, a rapid recovery of
EEP strength was observed once the stimulation was stopped.

138

Figure 6.5 Representative graphs from single experiment illustrating the effect of
1s stimulation (60 µA, 20Hz) for a total duration of 30s (A), 1min (B), 2min (C),
3min (D) and 5min (E) on the strength of EEPs elicited at 1.5 times threshold. After
a brief pause of 20s, the strength of EEPs was monitored without any 1s stimulation.
Data has been normalized to the first data point in each graph.

A B
C D
E
139

Figure 6.6 Graphs illustrate the effect of 1s stimulation (60 µA, 20 Hz) for a
duration of 30s (A), 1min (B), 2min (C) and 3min (D) on the strength of EEPs
generated at 1.5 times threshold in RD animal. After each of the stimulation
duration, the EEP strength was monitored without any 1s stimulation. Data has
been normalized to the first data point in each graph.

In addition to the above experiments, in one animal (RD), each of the 1s long
stimulus trains was applied every 7 seconds (train interval) apart. Figure 6.7 shows the
total strength of the EEP during the 10 min application of the stimulus trains followed by
a 2 min monitor only period. As can be observed, there is a decline in the EEP strength
however, not as rapidly as observed in the preceding graphs. Similar to preceding results,
the EEP strength recovered once the stimulation was stopped.

A B
C D
140

Figure 6.7 Graph illustrates the effect of 1s stimulation trains (60 µA, 20 Hz)
applied every 7s for a total duration of 10 min, on the strength of the EEPs
generated at 1.5 times threshold in RD animal. The EEP strength was monitored
without any continuous stimulation for duration of 2 min. Data has been normalized
to the first data point.

For one RD animal used in the present investigation, the stimulated retina was
prepared for histological examination. Figure 6.8 (B) illustrates the morphological
changes due to retinal degeneration in the stimulated retina. As can be observed, virtually
no ONL is present and some disorganization of the INL can be observed. Despite these
changes the effect of continuous stimulation on the EEPs recorded in the SC was similar
to that of normal retina.

Figure 6.8 Hematoxylin stained sections of normal (A) and degenerate (B) retina.
Scale bar = 50 µm.

141
6.4 Discussion
The results of the present study exhibit that continuous epiretinal stimulation
causes a depression in the EEP strength in the SC over a short duration of time.
Subsequent application of stimulus pulses does not cause any further decrease in the EEP
strength. Once the continuous stimulation is stopped, a relatively rapid recovery of EEP
strength is observed. The duration and severity of depression and subsequent recovery of
EEP strength thus seem independent of the total duration of continuous stimulation.
A number of different reasons may account for the temporary desensitization of
the retina due to continuous electrical stimulation. One reason for this desensitization
could be due increased inhibition from amacrine cells. Fried and colleagues showed that
longer stimulus pulses (≥ 1ms) elicited large excitatory and inhibitory inputs to retinal
ganglion cells (Fried et al., 2006). They found that the inhibition lasted for almost 100 ms
and hence hypothesized that this long lasting inhibition may interfere with stimulation at
higher stimulus frequencies. They confirmed this hypothesis by showing a decline in
excitatory input current from bipolar cells with increasing stimulus frequencies. Thus
they concluded that the reduction in excitation could be due to direct inhibition by
amacrine cells on ganglion cells or indirectly via bipolar cells. Since the stimulus pulse
width used in the present study was close to that used by Fried et al., it is possible that
during continuous stimulation, a similar increase in inhibitory input from amacrine cells
took place. As shown by Fried and colleagues, one way to reduce this inhibition is by
using short stimulus pulses (≤ 200 µs).
142
Another reason could be due to the presence of residual presynaptic calcium ions.
In cell cultures developed from neonatal rat superior colliculus, it was observed that
application of a pair of stimulus pulses induced a depression after the application of the
second stimulus pulse (Kirischuk et al., 2002). This form of depression has been termed
paired-pulse depression (PPD), the severity and subsequent recovery of which is
dependent upon the inter-stimulus interval. The fastest recovery observed was around 86
ms. One reason behind this depression was found to be dependent upon extracellular
concentration of Ca
2+
ions. At elevated levels of Ca
2+
ions, greater depression was
observed on the application of the second stimulus pulse. Since in the present study,
continuous retinal stimulation was performed and resulting EEPs were recorded in the
SC, there may have been elevated levels of residual presynaptic Ca
2+
ions in the retina
and/or SC that may have led to a temporary desensitization.
A third reason for this temporary desensitization could be some form of
adaptation by the retina in response to continuous electrical stimulation. Adaptive
mechanisms in the visual system have been well studied and are known to operate across
the visual hierarchy, from the retina to the visual cortex (Clifford et al., 2007; Kohn,
2007). Adaptation provides a mean for the nervous system to efficiently encode widely
varying inputs as encountered in everyday visual scenes into the limited range of neural
responses. The most well studied form of adaptation in the visual system is the one that
involves adjustment to the wide range of light intensities. Light and dark adaptation has
been showed to operate at the level of photoreceptors. In addition to photoreceptors, other
retinal neurons also adapt to variations in illumination by changing their gain and
143
response time course. While adaptation to mean intensity levels has been shown to be
present in all retinal neurons including photoreceptors, temporal contrast adaptation has
been shown to come into play at the level of bipolar cells. Studies in the retina have
shown that contrast sensitivity of retinal ganglion cells is altered before and after
adaptation (Chander and Chichilnisky, 2001; Kim and Rieke, 2001; Baccus and Meister,
2002). In the present study, although single units were not recorded, the overall strength
of the EEPs was observed to decline in the first 20-30s of continuous stimulation. This
effect was more pronounced in the early component of the EEPs compared to the late
component. It may be safe to assume the early component of the EEPs to include direct
activation from retinal ganglion cells and hence one can perhaps assume some adaptive
mechanisms initiated by them to play a role in the decline of EEP strength in response to
continuous stimulation. Another interesting feature is that the change in response was
observed in the SC. Adaptive mechanisms, especially those associated with contrast
adaptation have been shown to occur at various neural structures in the visual system.
Soloman et al. (2004), showed changes in contrast sensitivity in the LGN of primates
(Solomon et al., 2004). They showed this sensitivity to be particularly strong after
adaptation with gratings presented at high temporal frequency (25 cycles/s). This is an
interesting feature for the results presented here. The present study shows a decrease in
the EEP strength recorded in the SC in response to stimuli presented at 20 Hz. A final
common feature between the case at hand and adaptation is the time scale over which the
effects take place. Changes in neural response after adaptation have been shown to occur
as brief as tens of milliseconds to a more prolonged effect of many seconds. In the
144
present study, the EEP strength was observed to decline within 30s of continuous
stimulation exhibiting similarities with visual adaptation. Also, in visual adaptation, how
quickly the visual system adjusts to the environment may depend upon how frequently
the environment changes. A similar scenario can be viewed in the case of continuous
stimulation. The time scale over which the EEP strength decreases may depend upon how
quickly the trains of high frequency stimuli are applied to the retina. This is best
exhibited in figure 6.7, where the high frequency trains (1s long) were applied every
seven seconds as compared to every two seconds.
Finally, results observed in normal retina were similar to those observed in
degenerated retina. Kim and Rieke (2001) showed that in addition to the effect on input
currents to ganglion cells, increasing the temporal contrast had a greater effect on the
spike generation thus concluding the presence of some intrinsic adaptive mechanism in
retinal ganglion cells. The degenerated retina used in the present study showed
widespread changes due to degeneration with an almost no ONL present. Despite these
changes, a similar reduction in the EEP strength was observed in response to continuous
stimulation. Thus one can perhaps assume similar intrinsic mechanisms of retinal
ganglion cells to be at play in response to continuous electrical stimulation in degenerated
retinas.

145
Chapter 7
Summary

The work presented in this thesis has investigated the effect of continuous
stimulation of the retina on the retinal morphology and resulting evoked response in the
superior colliculus.

7.1 Key Findings and Future Work
The first investigation described in this thesis demonstrated that mechanical
contact between electrode tip and retina can cause damage and disruption to the
underlying retinal layers. This damage was limited to the area directly underneath the
electrode tip. However when high intensity electrical stimulation accompanied the
mechanical contact, the damage spread to a much larger area. The difference between the
extent of damage observed in the two cases was found to be statistically significant. This
could imply that after initial damage due to mechanical contact, the retina is more
sensitive to any additional onslaught. Also, interestingly severe damage was observed in
the outer retina as compared to the inner retina even though the electrode was positioned
on the epiretinal surface. This may indicate that photoreceptors seem to be more
vulnerable to damage due to mechanical contact and electrical stimulation. This could
also be due to the fact that the ONL is avascular, making the cells in the ONL more
vulnerable to mechanical pressure. Also as the photoreceptors are juxtaposed against the
146
RPE and choroid that are more rigid structures, it may cause the maximum stress to
develop across the ONL thus damaging the photoreceptors the most.
In any event, this result has important implications for the case of retinal
degeneration. As shown in chapter 2, a transgenic animal model of retinal degeneration
was studied at different ages to investigate the morphological changes that occur in the
retina. By P90, only a few scattered photoreceptors were observed. One implication of
this could be that due to the absence of the outer nuclear layer at advanced stages of
degeneration, the retina may not be as sensitive to either mechanical contact and/or
electrical stimulation. Clinically, degenerated retina is difficult to detach (personal
communication: Mark S. Humayun). Alternatively, due to reduction in the overall
thickness of the retina, it may make the retinal tissue more fragile and thus highly
sensitive to the slightest pressure exerted by the electrode tip.
To combat the effect of mechanical contact between the electrode tip and retina,
in this study, a method to sense the proximity to the retinal surface was developed. By
measuring the impedance of the interface at a high frequency (100 kHz), the electrode
could be positioned close to the retina without causing any damage due to contact. At the
same time the low threshold electrically evoked potentials (EEPs) could be recorded in
the superior colliculus (SC) when the electrode was positioned at impedance
approximately equal to 8 kΩ. This technique may be useful in general retinal physiology
experiments or even in vitreo-retinal surgical instrumentation.
With this method of positioning, continuous epiretinal stimulation was carried out
at different combinations of charge density and stimulus frequency for a period of 1 hour.
147
No morphological changes were observed even at the most severe level of stimulation
employed in the study. Similar results have been observed during clinical testing of
retinal prosthesis recipients where brief periods of stimulation at levels (~ 0.6 mC/cm
2
)
higher than the safety limit for chronic stimulation with platinum did not cause any
adverse reaction (Humayun et al., 2003). This suggests that the retina is more sensitive to
mechanical contact as opposed to continuous electrical stimulation and in fact is tolerant
to high levels of stimulation. However, an elevation in the EEP threshold was observed
after 1 hour of continuous stimulation. Post stimulation, the response was in general weak
and required a statistically significant increase in current to meet the threshold criteria.
On closer examination, the depression in the response was observed to occur in
the first tens of seconds of continuous stimulation. Additional stimulation did not cause
any further changes in the response strength. This change in neural response in the
absence of any morphological changes in the retina implies some mechanism of
adaptation might be at play. These may include adaptive mechanisms that are intrinsic to
the retina, changes in extracellular ion concentrations and changes in strength of synaptic
inhibition. However, the investigation of possible mechanisms is an important
consideration for future study but may be better studied in a more appropriate model of
human visual cortex, possibly through fMRI studies in implanted patients.
As recipients of retinal prostheses will be those with advanced stages of retinal
degeneration, it was important to explore whether the results observed in normal retina
were the same as those in degenerate retina. A subset of experiments was performed in
animals suffering from advanced stages of retinal degeneration. A rapid decline in the
148
EEP response was observed in the retinal degenerate animals in response to continuous
epiretinal stimulation as seen in the case of normal animals. Thus, it can perhaps be
concluded that similar results will be observed in the degenerate retinas as seen in the
present study for normal retina.

7.2 Implications for a Retinal Prosthesis
The results presented in this thesis have two main implications for the design of a
retinal prosthesis.
First, the failure to create any morphological changes from continuous epiretinal
stimulation implies that it may be possible to use small electrodes with high charge
density stimulation, a necessary but in no way a sufficient requirement for a high-
resolution prosthesis.
Second, the elevation in the threshold of evoked response in the absence of any
morphological change in the retina leads one to speculate whether this elevation is due to
some form of adaptation to continuous electrical stimulation. Also, the rapid decline in
the evoked response to continuous stimulation and an almost equally rapid recovery lends
further support to a more adaptive mechanism at play. Anecdotal evidence from clinical
trials of retinal prosthesis recipients has shown an elevation in perceptual threshold after
the end of a long (approximately 1hr) testing session. Subjects have also reported
percepts to be initially bright that progressively get dimmer with prolonged usage. Some
evidence of temporary desensitization to continuous stimulation at high frequency (45
Hz) has been reported for retinal prosthesis recipients (Nanduri et al., 2007). This
149
perceptual fading with electrical stimulation has also been shown in the case of
intracortical stimulation of the visual cortex (Schmidt et al., 1996). The study
demonstrated that phosphenes could be prolonged by interrupting the stimulation.
In light of present findings and those observed in the clinic, future testing
protocols in both humans and animals should be performed once the visual system has
been allowed to adapt to the artificial stimulation. Also, it may be worthwhile to
investigate the time course of adaptation to electrical stimulation for retinal prosthesis
recipients (Schmidt et al., 1996). This may further aid in the design of efficient stimulus
protocols for retinal prostheses especially for a prolong usage of the device.
The work presented in this thesis is the first study that has attempted to look at the
effect of continuous electrical stimulation of the retina in an in vivo animal model.
Although a temporary desensitization occurs due to continuous electrical stimulation, the
results of the present study are encouraging and exhibit that the retina is capable of
tolerating stimulation at charge density levels higher than the safety limit of platinum.

150
Chapter 8
Methods

8.1 Animal Model
Normal Copenhagen rats, 2~3 months old (Harlan, Indianapolis, IN, USA) and
Long Evans rats, ~5 months old (Charles River Laboratories, MA, USA) were used.
Additionally, a transgenic animal model of RP, the S334ter-line3 rat was also used. The
third line of albino Sprague-Dawley rats homozygous for the truncated murine opsin gene
(stop codon at residue 334; S334ter-3) was obtained from M.M. LaVail (University of
California, San Francisco, CA). Homozygous S334ter-3 breeding pairs were mated with
normal Copenhagen rats to produce offspring heterozygous for the S334ter transgene that
were subsequently used in this study. All animals were maintained on a daily 12 h
light/dark cycle and fed standard rodent diet ad libitum. All procedures were in
conformance with the Guide for Care and Use of Laboratory Animals (National Institutes
of Health). The University of Southern California Institutional Animal Care and Use
Committee reviewed and approved all procedures.

8.2 Surgical Procedure
All animals received an initial bolus of anaesthesia prepared from a cocktail of
ketamine (100 mg/kg; KETASET, Fort Dodge, IA, USA) and xylazine (20 mg/kg; X-
Ject SA, Butler, Dublin, OH, USA) administered through an intraperitoneal injection.
151
After being positioned on a stereotaxic apparatus (Kopf Instruments, CA, USA), animals
received sevoflourane (1% in 100% O
2
) through a gas-inhalant mask for the entire
duration of the experiment. Blood pressure and heart rate was monitored throughout the
experiment via a sensor clipped to the hind leg (Heska, Loveland, CO, USA). Body
temperature was maintained at 37°C through a rectal thermometer connected to a self-
regulating electric heating blanket (model 50-7053-F; Harvard Apparatus, Holliston MA,
USA). Animals were euthanized through an intra-cardiac injection of sodium
pentobarbital (30 mg/kg; Butler, Dublin, OH, USA). For stimulation only experiments,
topical antibiotic was applied to the eye and metacam and dexamethasone was
administered via intramuscular and intraperitoneal injections respectively. The health of
the animal was monitored during the entire recovery period and subsequent injections of
metacam were administered as and when required.

8.3 Stimulation Electrode and Retinal Stimulation
Epiretinal stimulation was carried out using a concentric bipolar electrode (Model
CBDFG74, FHC Inc., Bouidain, ME, USA). The Pt-Ir inner pole with a flat-tip, 75 µm
diameter was used for stimulation. The 300 µm diameter stainless steel outer pole was
used as the return electrode for producing focal stimulation in order to record EEPs in the
SC. In order to study the effect of continuous stimulation, the retina was stimulated
continuously for 1-hr using the inner pole as the stimulation electrode and a needle
electrode inserted in the skin next to the nose as the return electrode. This monopolar
configuration was employed in an effort to mimic the stimulation protocol followed with
152
the retina prostheses that have a remote ground. During the 1-hr stimulation period,
cathodic-first, charge-balanced biphasic pulses were delivered to the stimulation
electrode. Each phase of the pulse was 0.5 ms long with a 0.1 ms delay between the two
phases. Biphasic voltage pulses were generated using a stimulus generator (STG 2004,
Multi Channel Systems, Reutlingen, Germany) that was then fed to an analogue voltage-
to-current stimulus isolator (Model 2200, A-M Systems, Carlsborg, WA, USA). The
output of the stimulus isolator was a train of charge-balanced current pulses that was
delivered to the stimulation electrode via blocking capacitors. The output of the isolator
was periodically checked on an oscilloscope to ensure that there was no DC offset and
that the output delivered was within the compliance limit of the isolator.

8.4 Electrochemical Techniques – Impedance and Cyclic Voltammetry
For in vivo measurements, the stimulating electrode (described above) was
inserted into the eye. The inner pole of the electrode was used as the working electrode
while needle electrodes inserted in the skin next to the nose and in the tail of the animal
were used as counter and reference electrodes respectively.
For in vitro measurement of the electrode, measurements were carried out in 0.1M
PBS at room temperature. The inner pole of the electrode was used as the working
electrode while a large platinum electrode was used as the counter electrode. A third Ag-
AgCl electrode was used as the reference electrode.
All measurements for impedance were carried out using a commercial potentiostat
(FAS1 potentiostat, Gamry Instruments Inc., PA, USA). For all measurements, 10 mV
153
sine wave was applied to the working electrode with DC held at the electrode’s open
circuit potential. For impedance spectrum measurements, the signal was applied over a
range of frequencies starting from 100 kHz to 1 Hz. For single frequency measurements,
the signal was applied at 100 kHz at intervals of 0.01s.

8.5 Superior Colliculus Surgery and Recording Electrode Insertion
A craniotomy on the right skull was performed (caudal-medial corner: ~4mm
caudal and ~3mm lateral to lambda) using a handheld driller ((Dremel, Robert Bosch
Tool Corporation, Mount Prospect, IL, USA). The cortex was aspirated till the SC
surface was exposed (∼ 4mm deep from dura matter). Epoxy-coated tungsten
microelectrodes (10MΩ, FHC Inc., Bouidain, ME, USA) were positioned within the
superficial layers of the SC where majority of the retinal ganglion cell axons terminate
(Langer and Lund, 1974; Girman and Lund, 2007). The microelectrode was carefully
lowered into the cavity till a characteristic noise was heard right after penetrating the pia
matter. This noise is produced by the juxtazonal potentials that are known to be retina-
driven responses found in the SC (McIlwain, 1978). Data presented here comprises of
recording and analysis of these potentials.

8.6 Electrically Evoked Potential Recording and Analysis
In order to locate the lowest threshold site in the SC, the retina was initially
stimulated over a wide range of current amplitude (1, 2, 5, 10, 20, 30, 40, 50, 60, 70, 80,
90 and 100 µA). The site which produced a response at stimulus level ≤ 20 µA was
154
investigated in finer detail. Here, in order to better estimate the EEP threshold, 4-5
current levels 2 µA apart were applied and an algorithm was used to process the data to
determine the threshold. Threshold was defined as the stimulus level that elicited an
electrically evoked response that was 5 times the baseline activity and was present 75%
of the total number of trials (n=50). After determining the threshold, the recording was
suspended during which the retina was stimulated continuously for 1-hr. The recording
electrode was held in position during the 1-hr period. Immediately after the 1-hr period,
recording was resumed at the same location and threshold was determined in the same
manner as described above.
To better understand the effect of continuous stimulation on the EEPs recorded, a
custom program (SciWorks,Ver. 6.0, DataWave Technology) was designed to analyse the
individual features of the response at threshold before and after the 1-hr stimulation. Each
of the EEPs were divided into four potentials: P1 corresponded to the initial prepotential
right after stimulus onset and clearly separated from the stimulus artifact, N1
corresponded to the first largest negative deflection, N2 and N3 corresponded to the
following two potentials (Fig. 5.5). Parameters that were measured included latency of
onset of each potential, valley-to-peak amplitude and the area of each peak.

8.7 Tissue preparation
Animals were deeply anaesthetized (described above) and the left eye was
enucleated. The anterior segments of eyes were dissected and the eyecups were
immersion fixed in 4% paraformaldehyde in 0.1M phosphate buffer (PB), ph 7.4 for 2 hr.
155
Following fixation, eyecups were transferred into 30% sucrose for overnight in 4°C. For
cryostat sections, eyecups were embedded in OCT embedding medium (Tissue-Tek,
Elkhart, IN, USA), then quickly frozen in liquid nitrogen and subsequently sectioned
along the vertical meridian on a cryostat at a thickness of 25µm. Sections were collected
on gelatin-coated slides for immunostaining.

8.8 Hematoxylin Staining
For hematoxylin staining, every 10
th
slide was used within each sample to analyze
the gross histology. Slides were rinsed in tap water for 5 min to remove OCT and then
dipped in hematoxylin for 5 min. Slides were then rinsed in tap water, dehydrated in
alcohol, cleared in xylene and then mounted in xylene-based medium (Richard-Allen
Scientific, Kalamazoo, MI, USA).

8.9 Immunocytochemistry
Fluorescence immunocytochemistry was carried out on the cryostat sections that
were incubated in 10% normal donkey serum in PBS for 1-hr in order to block all the
non-specific binding sites. Sections were then incubated overnight with primary antibody.
Antibody list with dilution and product information has been listed in Table 8.1. Each
antiserum was diluted in PBS containing 0.5% Triton X-100 at 4ºC. Retinas were washed
in PBS for 45 min (3 x 15 min) and afterwards incubated for 2 h at room temperature in
either carboxymethylindocyanine (Cy3)-conjugated affinity-purified donkey anti-rabbit
IgG (Jackson Immuno Laboratories, West Grove, PA, dilution 1:500) or Alexa 488 anti-
156
goat IgG (Molecular Probes, Eugene, OR, dilution 1:300). The sections were washed for
30 min with 0.1M PB and coverslipped with Vectashield mounting medium (Vector
Labs, Burlingame, CA).
Antiserum Immunogen Source Dilution
Goat polyclonal Anti blue-
sensitive opsin (S-opsin
N-terminus of blue
sensitive opsin of human
origin
Santa Cruz Biotechnology Inc.
Santa Cruz, CA (SC-14363)
1:500
Rabbit polyclonal Anti-
Calbindin D-28K
Recombinant mouse
calbindin
Millipore, Temecula, CA.
(AB1778)
1:1000
Rabbit polyclonal Anti-
GABA
Rabbit using GABA-BSA
as the immunogen
Sigma, St. Louis, MI (A2052)

1:3000
Rabbit polyclonal Anti-
glial fibrillary acidic
protein (GFAP)
Purified from human brain

Sigma, St. Louis, MI (G9269)

1:500
Rabbit polyclonal Anti-
parvalbumin
Rat muscle parvalbumin Swant, Bellinzona,
Switzerland (PV-28)
1:1500
Fluorescein-peanut
Agglutinin (FITC-PNA)
Arachis hypogaea
(peanuts)
Vector Labs, Burlingame, CA
(FL-1071)
1:500
Rabbit polyclonal Anti-
protein kinase C (PKC)
Carboxy terminus of PKC
alpha of human origin
Santa Cruz Biotechnology Inc.
Santa Cruz, CA (sc-208)
1:1000
Rabbit polyclonal Anti-
recoverin
Recombinant human
recoverin
Chemicon international,
Temecula, CA (AB5585)
1:1500

Table 8.1 Antibody list
157

Antiserum Immunogen Source Dilution
Rabbit polyclonal Anti-
GAD65
Human GAD65 from
baculovirus infected cells
Chemicon international,
Temecula, CA (AB5082)
1:1000
Goat polyclonal anti-glycine
transporter-1 (Glyt-1)
Carboxy-terminus of Glyt1
of rat origin
Chemicon international,
Temecula, CA (AB1770)
1:8000
Rabbit polyconal Anti-
SV2A and SV2B
Synthetic peptides
corresponding to residues 2-
17 (human and rat
respectively) coupled to
key-hole limpet hemocyanin
via N-terminal added
cysteine reside
Synaptic Systems, Goettingen,
Germany (119-002 & 119-
102)
1:2000
Mouse monoclonal Anti-
GOα
Bovine Brian GOα Chemicon international,
Temecula, CA (MAB3073)
1:5000
Mouse monoclonal Anti-
microtubule associated
protein (MAP1)
Rat brain microtubule
associated protein
Chemicon international,
Temecula, CA (MAB362)
1:1000
Rat polyclonal Anti-glycine Paraformaldehyde conjugate
of glycine amino acid
Gift from Dr. David Pow 1:5000

Table 8.1 Antibody list continued

Retinal sections were analyzed using a confocal microscope (LSM 510, Zeiss,
NY, USA). Immunofluorescence images were processed in Zeiss LSM-PC software. The
brightness and contrast was adjusted using Adobe Photoshop 7.0 (Adobe Systems,
158
Mountain View, CA). All Photoshop manipulations (brightness and contrast) were
carried out equally across sections.

8.10 Statistics
For all statistical measurements, a two-sided paired t-test or independent t-test was
performed using Microsoft Excel (version 11.5.6). A significance level of 0.05 was
chosen for all tests.

159
Bibliography

Agnew WF, Yuen TG, McCreery DB (1983) Morphologic changes after prolonged
electrical stimulation of the cat's cortex at defined charge densities. Exp Neurol 79:397-
411.

Agnew WF, Yuen TG, Pudenz RH, Bullara LA (1975) Electrical stimulation of the brain.
IV. Ultrastructural studies. Surg Neurol 4:438-448.

Agnew WF, Yuen TG, McCreery DB, Bullara LA (1986) Histopathologic evaluation of
prolonged intracortical electrical stimulation. Exp Neurol 92:162-185.

Agnew WF, McCreery DB, Yuen TG, Bullara LA (1989) Histologic and physiologic
evaluation of electrically stimulated peripheral nerve: considerations for the selection of
parameters. Ann Biomed Eng 17:39-60.

Agnew WF, McCreery DB, Yuen TG, Bullara LA (1993) MK-801 protects against
neuronal injury induced by electrical stimulation. Neuroscience 52:45-53.

Agnew WF, Yuen TG, Bullara LA, Jacques D, Pudenz RH (1979) Intracellular calcium
deposition in brain following electrical stimulation. Neurol Res 1:187-202.

Ahuja AK, Behrend MR, Whalen JJ, Humayun MS, Weiland JD (2008) The dependence
of spectral impedance on disc microelectrode radius. IEEE Trans Biomed Eng 55:1457-
1460.

Ashton D, Reid K, Willems R, Wauquier A (1986) N-methyl-D-aspartate and hypoxia
induced Ca2+-changes in the CA1 region of the hippocampal slice. Brain Res 385:185-
188.

Auricchio A, Rolling F (2005) Adeno-associated viral vectors for retinal gene transfer
and treatment of retinal diseases. Curr Gene Ther 5:339-348.

Baccus SA, Meister M (2002) Fast and slow contrast adaptation in retinal circuitry.
Neuron 36:909-919.

Baig-Silva MS, Hathcock CD, Hetling JR (2005) A preparation for studying electrical
stimulation of the retina in vivo in rat. J Neural Eng 2:S29-38.

Bak M (1989) [Therapeutic problems of patients with diabetes mellitus after total
pancreatectomy]. Pol Arch Med Wewn 81:137-143.
160
Bak M, Girvin JP, Hambrecht FT, Kufta CV, Loeb GE, Schmidt EM (1990) Visual
sensations produced by intracortical microstimulation of the human occipital cortex. Med
Biol Eng Comput 28:257-259.

Bard AJ, Faulkner LR (2004) Electrocehmical Methods: Fundamentals and Applications,
Second Edition: John Wiley & Sons.

Barhoum R, Martinez-Navarrete G, Corrochano S, Germain F, Fernandez-Sanchez L, de
la Rosa EJ, de la Villa P, Cuenca N (2008) Functional and structural modifications during
retinal degeneration in the rd10 mouse. Neuroscience 155:698-713.

Bates JB, Chu YT (1992) Electrode-electrolyte interface impedance: experiments and
model. Ann Biomed Eng 20:349-362.

Beebe X, Rose TL (1988) Charge injection limits of activated iridium oxide electrodes
with 0.2 ms pulses in bicarbonate buffered saline. IEEE Trans Biomed Eng 35:494-495.

Benabid AL, Pollak P, Gao D, Hoffmann D, Limousin P, Gay E, Payen I, Benazzouz A
(1996) Chronic electrical stimulation of the ventralis intermedius nucleus of the thalamus
as a treatment of movement disorders. J Neurosurg 84:203-214.

Bhati CS, Silva MA, Wigmore SJ, Bramhall SR, Mayer DA, Buckels JA, Neil DA,
Murphy N, Mirza DF (2009) Use of bioelectrical impedance analysis to assess liver
steatosis. Transplant Proc 41:1677-1681.

Bi A, Cui J, Ma YP, Olshevskaya E, Pu M, Dizhoor AM, Pan ZH (2006) Ectopic
expression of a microbial-type rhodopsin restores visual responses in mice with
photoreceptor degeneration. Neuron 50:23-33.

Biran R, Martin DC, Tresco PA (2005) Neuronal cell loss accompanies the brain tissue
response to chronically implanted silicon microelectrode arrays. Exp Neurol 195:115-
126.

Bird AC (1995) Retinal photoreceptor dystrophies LI. Edward Jackson Memorial
Lecture. Am J Ophthalmol 119:543-562.

Bonnot A, Mentis GZ, Skoch J, O'Donovan MJ (2005) Electroporation loading of
calcium-sensitive dyes into the CNS. J Neurophysiol 93:1793-1808.

Bowman BR, McNeal DR (1986) Response of single alpha motoneurons to high-
frequency pulse trains. Firing behavior and conduction block phenomenon. Appl
Neurophysiol 49:121-138.
161
Bradley DC, Troyk PR, Berg JA, Bak M, Cogan S, Erickson R, Kufta C, Mascaro M,
McCreery D, Schmidt EM, Towle VL, Xu H (2005) Visuotopic mapping through a
multichannel stimulating implant in primate V1. J Neurophysiol 93:1659-1670.

Brindley GS, Lewin WS (1968) The sensations produced by electrical stimulation of the
visual cortex. J Physiol 196:479-493.

Brug GJ, Van Den Eeden ALG, Sluythers-Rehbach M, Suythers JH (1984) The analysis
of electrode impedances complicated by the presence of a constant phase element.
Journal of electroanalytical chemistry 176:275-295.

Brummer SB, Turner MJ (1975) Electrical stimulation of the nervous system: The
principle of safe charge injection with noble metal electrodes. Bioelectrochemistry and
Bioenergetics 2:13-25.

Brummer SB, Turner MJ (1977a) Electrical stimulation with Pt electrodes: II-estimation
of maximum surface redox (theoretical non-gassing) limits. IEEE Trans Biomed Eng
24:440-443.

Brummer SB, Turner MJ (1977b) Electrical stimulation with Pt electrodes: I-a method
for determination of "real" electrode areas. IEEE Trans Biomed Eng 24:436-439.

Bunker CH, Berson EL, Bromley WC, Hayes RP, Roderick TH (1984) Prevalence of
retinitis pigmentosa in Maine. Am J Ophthalmol 97:357-365.

Caspi A, Dorn JD, McClure KH, Humayun MS, Greenberg RJ, McMahon MJ (2009)
Feasibility study of a retinal prosthesis: spatial vision with a 16-electrode implant. Arch
Ophthalmol 127:398-401.

Chader GJ (2002) Animal models in research on retinal degenerations: past progress and
future hope. Vision Res 42:393-399.

Chan LH, Ray A, Thomas BB, Humayun MS, Weiland JD (2008) In vivo study of
response threshold in retinal degenerate model at different degenerate stages. Conf Proc
IEEE Eng Med Biol Soc 1:1781-1784.

Chander D, Chichilnisky EJ (2001) Adaptation to temporal contrast in primate and
salamander retina. J Neurosci 21:9904-9916.

Chen W, Zhongsheng Z, Lee RC (2006) Supramembrane potential-induced
electroconformational changes in sodium channel proteins: a potential mechanism
involved in electric injury. Burns 32:52-59.

162
Chow AY, Peachey NS (1998) The subretinal microphotodiode array retinal prosthesis.
Ophthalmic Res 30:195-198.

Chow AY, Chow VY, Packo KH, Pollack JS, Peyman GA, Schuchard R (2004) The
artificial silicon retina microchip for the treatment of vision loss from retinitis
pigmentosa. Arch Ophthalmol 122:460-469.

Clifford CW, Webster MA, Stanley GB, Stocker AA, Kohn A, Sharpee TO, Schwartz O
(2007) Visual adaptation: neural, psychological and computational aspects. Vision Res
47:3125-3131.

Cogan SF, Guzelian AA, Agnew WF, Yuen TG, McCreery DB (2004) Over-pulsing
degrades activated iridium oxide films used for intracortical neural stimulation. J
Neurosci Methods 137:141-150.

Cogan SF, Troyk PR, Ehrlich J, Plante TD, Detlefsen DE (2006) Potential-biased,
asymmetric waveforms for charge-injection with activated iridium oxide (AIROF) neural
stimulation electrodes. IEEE Trans Biomed Eng 53:327-332.

Cogan SF, Ehrlich J, Plante TD, Smirnov A, Shire DB, Gingerich M, Rizzo JF (2009)
Sputtered iridium oxide films for neural stimulation electrodes. J Biomed Mater Res B
Appl Biomater 89:353-361.

Cohen ED (2009) Effects of high-level pulse train stimulation on retinal function. J
Neural Eng 6:035005.

Colley NJ, Cassill JA, Baker EK, Zuker CS (1995) Defective intracellular transport is the
molecular basis of rhodopsin-dependent dominant retinal degeneration. Proc Natl Acad
Sci U S A 92:3070-3074.

Cowey A, Rolls ET (1974) Human cortical magnification factor and its relation to visual
acuity. Exp Brain Res 21:447-454.

Cuenca N, Pinilla I, Sauve Y, Lund R (2005) Early changes in synaptic connectivity
following progressive photoreceptor degeneration in RCS rats. Eur J Neurosci 22:1057-
1072.

Cuenca N, Pinilla I, Sauve Y, Lu B, Wang S, Lund RD (2004) Regressive and reactive
changes in the connectivity patterns of rod and cone pathways of P23H transgenic rat
retina. Neuroscience 127:301-317.

Cui X, Wiler J, Dzaman M, Altschuler RA, Martin DC (2003) In vivo studies of
polypyrrole/peptide coated neural probes. Biomaterials 24:777-787.

163
Cui X, Lee VA, Raphael Y, Wiler JA, Hetke JF, Anderson DJ, Martin DC (2001a)
Surface modification of neural recording electrodes with conducting
polymer/biomolecule blends. J Biomed Mater Res 56:261-272.

Cui XT, Zhou DD (2007) Poly (3,4-ethylenedioxythiophene) for chronic neural
stimulation. IEEE Trans Neural Syst Rehabil Eng 15:502-508.

Cui XY, Martin DC (2003) Electrochemical deposition and characterization of poly(3,4-
ethylenedioxythiophene) on neural microelectrode arrays. Sensors and Actuators B:
Chemical 89:92-102.

Cui XY, Hetke J, F, Wiler JA, Anderson DJ, Martin DC (2001b) Electrochemical
deposition and characterization of conducting ploymer polypyrrole/pss on multichannel
neural probes. Sensors and Actuators A: Phys 93:8-18.

D'Cruz PM, Yasumura D, Weir J, Matthes MT, Abderrahim H, LaVail MM, Vollrath D
(2000) Mutation of the receptor tyrosine kinase gene Mertk in the retinal dystrophic RCS
rat. Hum Mol Genet 9:645-651.

Das AM, Zhao X, Ahmad I (2005) Stem cell therapy for retinal degeneration: retinal
neurons from heterologous sources. Semin Ophthalmol 20:3-10.
de Balthasar C, Patel S, Roy A, Freda R, Greenwald S, Horsager A, Mahadevappa M,
Yanai D, McMahon MJ, Humayun MS, Greenberg RJ, Weiland JD, Fine I (2008) Factors
affecting perceptual thresholds in epiretinal prostheses. Invest Ophthalmol Vis Sci
49:2303-2314.

DeMarco PJ, Jr., Yarbrough GL, Yee CW, McLean GY, Sagdullaev BT, Ball SL, McCall
MA (2007) Stimulation via a subretinally placed prosthetic elicits central activity and
induces a trophic effect on visual responses. Invest Ophthalmol Vis Sci 48:916-926.

DiLoreto DA, Jr., Martzen MR, del Cerro C, Coleman PD, del Cerro M (1995) Muller
cell changes precede photoreceptor cell degeneration in the age-related retinal
degeneration of the Fischer 344 rat. Brain Res 698:1-14.

Dobelle WH, Mladejovsky MG (1974) Phosphenes produced by electrical stimulation of
human occipital cortex, and their application to the development of a prosthesis for the
blind. J Physiol 243:553-576.

Dowling JE (1987) The retina: an approachable part of the brain. Cambridge
Massachusetts and London England: Belknap Press of Harvard University.

Dymond AM (1976) Characteristics of the metal-tissue interface of stimulation
electrodes. IEEE Trans Biomed Eng 23:274-280.
164
Edwards MD, White AM, Platt B (2002) Characterisation of rat superficial superior
colliculus neurones: firing properties and sensitivity to GABA. Neuroscience 110:93-104.

Fisher SK, Lewis GP, Linberg KA, Verardo MR (2005) Cellular remodeling in
mammalian retina: results from studies of experimental retinal detachment. Prog Retin
Eye Res 24:395-431.

Foerster O (1929) Beitrage zur Pathophysiologie der Sehbahn und der Spehsphare.
Journal of Psychology and Neurology 39:435-463.

Forbes A, Rice LH (1929) Quantitative studies of the nerve impulse IV. Fatigue in
peripheral nerve. American Journal of Physiology 90:119-145.

Franks W, Schenker I, Hierlmann A (2005) Impedance characterization and modeling of
electrodes for biomedical applications. IEEE Trans Biomed Eng 52:1295-1302.

Fried SI, Hsueh HA, Werblin FS (2006) A method for generating precise temporal
patterns of retinal spiking using prosthetic stimulation. J Neurophysiol 95:970-978.

Fujikado T, Morimoto T, Kanda H, Kusaka S, Nakauchi K, Ozawa M, Matsushita K,
Sakaguchi H, Ikuno Y, Kamei M, Tano Y (2007) Evaluation of phosphenes elicited by
extraocular stimulation in normals and by suprachoroidal-transretinal stimulation in
patients with retinitis pigmentosa. Graefes Arch Clin Exp Ophthalmol 245:1411-1419.

Fukuda Y, Suzuki DA, Iwama K (1978) A four group classification of the rat superior
collicular cells responding to optic nerve stimulation. Jpn J Physiol 28:367-384.

Gargini C, Terzibasi E, Mazzoni F, Strettoi E (2007) Retinal organization in the retinal
degeneration 10 (rd10) mutant mouse: a morphological and ERG study. J Comp Neurol
500:222-238.

Gasser HS (1935) Changes in nerve potentials produced by rapidly repeated stimuli and
their relation to the responsiveness of nerve stimulation. American Journal of Physiology
96:35-50.

Geddes LA (1997) Historical evolution of circuit models for the electrode-electrolyte
interface. Ann Biomed Eng 25:1-14.

Geddes LA (1999) Chronaxie. Australasian Physical & Engineering Sciences in
Medicine 22:13 - 17.

Geddes LA (2004) Accuracy limitations of chronaxie values. IEEE Trans Biomed Eng
51:176-181.
165
Geddes LA, Bourland JD (1985) The strength-duration curve. IEEE Trans Biomed Eng
32:458-459.

Gekeler F, Kobuch K, Schwahn HN, Stett A, Shinoda K, Zrenner E (2004) Subretinal
electrical stimulation of the rabbit retina with acutely implanted electrode arrays. Graefes
Arch Clin Exp Ophthalmol 242:587-596.

Germain PS, Pell WG, Conway BE (2004) Evaluation and origins of the differences
between double-layer capacitance behaviour at Au-metal and oxidized Au surfaces.
Electrochemica Acta 49:1775-1788.

Girman SV, Lund RD (2007) Most superficial sublamina of rat superior colliculus:
neuronal response properties and correlates with perceptual figure-ground segregation. J
Neurophysiol 98:161-177.

Gouras P (1958) Spreading depression of activity in amphibian retina. Am J Physiol
195:28-32.

Greenbaum E, Sanders C, Zhou D (2006) Dynamic interactions at retinal prosthesis
electrode interface. In: Association for research in vision and ophthalmology. Fort
lauderdale, Florida.

Greenwald SH, Horsager A, Humayun MS, Greenberg RJ, McMahon MJ, Fine I (2009)
Brightness as a function of current amplitude in human retinal electrical stimulation.
Invest Ophthalmol Vis Sci 50:5017-5025.

Guven D, Weiland JD, Maghribi M, Davidson JC, Mahadevappa M, Roizenblatt R, Qiu
G, Krulevitz P, Wang X, Labree L, Humayun MS (2006) Implantation of an inactive
epiretinal poly(dimethyl siloxane) electrode array in dogs. Exp Eye Res 82:81-90.

Guven D, Weiland JD, Fujii G, Mech BV, Mahadevappa M, Greenberg R, Roizenblatt R,
Qiu G, Labree L, Wang X, Hinton D, Humayun MS (2005) Long-term stimulation by
active epiretinal implants in normal and RCD1 dogs. J Neural Eng 2:S65-73.

Haberler C, Alesch F, Mazal PR, Pilz P, Jellinger K, Pinter MM, Hainfellner JA, Budka
H (2000) No tissue damage by chronic deep brain stimulation in Parkinson's disease. Ann
Neurol 48:372-376.

Hamann CH, Hamnett A, Vielstich W (1998) Methods for the study of
electrode/electrolyte interface: Wiley-VCH.

Hartong DT, Berson EL, Dryja TP (2006) Retinitis pigmentosa. Lancet 368:1795-1809.
Haverkamp S, Wassle H (2000) Immunocytochemical analysis of the mouse retina. J
Comp Neurol 424:1-23.
166
Haverkamp S, Wässle H (2000) Immunocytochemical analysis of the mouse retina. J
Comp Neurol 424:1-23.

Horsager A, Greenberg RJ, Fine I (2009a) Spatiotemporal interactions in retinal
prosthesis subjects. Invest Ophthalmol Vis Sci.

Horsager A, Greenwald SH, Weiland JD, Humayun MS, Greenberg RJ, McMahon MJ,
Boynton GM, Fine I (2009b) Predicting visual sensitivity in retinal prosthesis patients.
Invest Ophthalmol Vis Sci 50:1483-1491.

Horsburgh GM, Sefton AJ (1987) Cellular degeneration and synaptogenesis in the
developing retina of the rat. J Comp Neurol 263:553-566.

Huang CQ, Shepherd RK (1999) Reduction in excitability of the auditory nerve
following electrical stimulation at high stimulus rates. IV. Effects of stimulus intensity.
Hear Res 132:60-68.

Huang CQ, Shepherd RK (2000) Reduction in excitability of the auditory nerve
following electrical stimulation at high stimulus rates: V. Effects of electrode surface
area. Hear Res 146:57-71.

Humayun MS, de Juan E, Jr., Weiland JD, Dagnelie G, Katona S, Greenberg R, Suzuki S
(1999a) Pattern electrical stimulation of the human retina. Vision Res 39:2569-2576.

Humayun MS, Prince M, de Juan E, Jr., Barron Y, Moskowitz M, Klock IB, Milam AH
(1999b) Morphometric analysis of the extramacular retina from postmortem eyes with
retinitis pigmentosa. Invest Ophthalmol Vis Sci 40:143-148.

Humayun MS, Weiland JD, Fujii GY, Greenberg R, Williamson R, Little J, Mech B,
Cimmarusti V, Van Boemel G, Dagnelie G, de Juan E (2003) Visual perception in a blind
subject with a chronic microelectronic retinal prosthesis. Vision Res 43:2573-2581.

Jensen RJ, Rizzo JF, 3rd (2007) Responses of ganglion cells to repetitive electrical
stimulation of the retina. J Neural Eng 4:S1-6.

Jensen RJ, Rizzo JF, 3rd (2008) Activation of retinal ganglion cells in wild-type and rd1
mice through electrical stimulation of the retinal neural network. Vision Res 48:1562-
1568.

Jensen RJ, Ziv OR, Rizzo JF, 3rd (2005a) Thresholds for activation of rabbit retinal
ganglion cells with relatively large, extracellular microelectrodes. Invest Ophthalmol Vis
Sci 46:1486-1496.

167
Jensen RJ, Ziv OR, Rizzo JF (2005b) Responses of rabbit retinal ganglion cells to
electrical stimulation with an epiretinal electrode. J Neural Eng 2:S16-21.

Jensen RJ, Rizzo JF, 3rd, Ziv OR, Grumet A, Wyatt J (2003) Thresholds for activation of
rabbit retinal ganglion cells with an ultrafine, extracellular microelectrode. Invest
Ophthalmol Vis Sci 44:3533-3543.

Jones BW, Marc RE (2005) Retinal remodeling during retinal degeneration. Exp Eye Res
81:123-137.

Jones BW, Watt CB, Frederick JM, Baehr W, Chen CK, Levine EM, Milam AH, Lavail
MM, Marc RE (2003) Retinal remodeling triggered by photoreceptor degenerations. J
Comp Neurol 464:1-16.

Kanda H, Morimoto T, Fujikado T, Tano Y, Fukuda Y, Sawai H (2004)
Electrophysiological studies of the feasibility of suprachoroidal-transretinal stimulation
for artificial vision in normal and RCS rats. Invest Ophthalmol Vis Sci 45:560-566.

Kandel ER, Schwartz JH, Jessell TM (2000) Principles of Neural Science, Fourth
Edition: McGraw-Hill.

Katz B (1939) Nerve excitation by high-frequency alternating current. J Physiol 96:202-
224.

Kim DH, Abidian M, Martin DC (2004) Conducting polymers grown in hydrogel
scaffolds coated on neural prosthetic devices. J Biomed Mater Res A 71:577-585.

Kim KJ, Rieke F (2001) Temporal contrast adaptation in the input and output signals of
salamander retinal ganglion cells. J Neurosci 21:287-299.

Kirischuk S, Clements JD, Grantyn R (2002) Presynaptic and postsynaptic mechanisms
underlie paired pulse depression at single GABAergic boutons in rat collicular cultures. J
Physiol 543:99-116.

Klassen H, Sakaguchi DS, Young MJ (2004) Stem cells and retinal repair. Prog Retin
Eye Res 23:149-181.

Kohn A (2007) Visual adaptation: physiology, mechanisms, and functional benefits. J
Neurophysiol 97:3155-3164.

Kolb H (2003) How the Retina Works. In: American Scientist, pp 28-35.

Kolb H, Gouras P (1974) Electron microscopic observations of human retinitis
pigmentosa, dominantly inherited. Invest Ophthalmol 13:487-498.
168
Krause F, Schum H (1931) Die epiliptischen Erkankungen.

Krauthamer V, Bekken M, Horowitz JL (1991) Morphological and electrophysiological
changes produced by electrical stimulation in cultured neuroblastoma cells.
Bioelectromagnetics 12:299-314.

Langer TP, Lund RD (1974) The upper layers of the superior colliculus of the rat: a Golgi
study. J Comp Neurol 158:418-435.

Lapicque L (1909) Definition experimentale de l'excitabilite. Comptes Rendus de
I'Academie des sciences 67:280-283.

Lasia A (2002) Electrochemical Impedance Spectroscopy and its Application. In: Modern
Aspects of Electrochemistry (Conway BE, Bockris JOM, White RE, eds): Kluwer
Academic Publishers.

LaVail MM (1981) Analysis of neurological mutants with inherited retinal degeneration.
Friedenwald lecture. Invest Ophthalmol Vis Sci 21:638-657.

LaVail MM, Matthes MT, Yasumura D, Steinberg RH (1997) Variability in rate of cone
degeneration in the retinal degeneration (rd/rd) mouse. Exp Eye Res 65:45-50.

Leão AAP (1944) Spreading depression of activity in the cerebral cortex. Journal of
Neurophysiology 7:359.

Lee EJ, Kim HJ, Lim EJ, Kim IB, Kang WS, Oh SJ, Rickman DW, Chung JW, Chun MH
(2004) AII amacrine cells in the mammalian retina show disabled-1 immunoreactivity. J
Comp Neurol 470:372-381.

Lee JY, Deogaonkar M, Rezai A (2007) Deep brain stimulation of globus pallidus
internus for dystonia. Parkinsonism Relat Disord 13:261-265.

Lempka SF, Miocinovic S, Johnson MD, Vitek JL, McIntyre CC (2009) In vivo
impedance spectroscopy of deep brain stimulation electrodes. J Neural Eng 6:46001.

Li Y, Mogul DJ (2007) Electrical control of epileptic seizures. J Clin Neurophysiol
24:197-204.

Li ZY, Possin DE, Milam AH (1995) Histopathology of bone spicule pigmentation in
retinitis pigmentosa. Ophthalmology 102:805-816.

Lilly JC (1961) Injury and excitation by electric currents: the balanced pulse-pair
waveform: Hogg Foundation for Mental Health.
169
Lilly JC, Austin GM, Chambers WW (1952) Threshold movements produced by
excitation of cerebral cortex and efferent fibers with some parametric regions of
rectangular current pulses (cats and monkeys). J Neurophysiol 15:319-341.

Lilly JC, Hughes JR, Alvord EC, Jr., Galkin TW (1955) Brief, noninjurious electric
waveform for stimulation of the brain. Science 121:468-469.

Liu C, Li Y, Peng M, Laties AM, Wen R (1999a) Activation of caspase-3 in the retina of
transgenic rats with the rhodopsin mutation s334ter during photoreceptor degeneration. J
Neurosci 19:4778-4785.

Liu X, McCreery DB, Carter RR, Bullara LA, Yuen TG, Agnew WF (1999b) Stability of
the interface between neural tissue and chronically implanted intracortical
microelectrodes. IEEE Trans Rehabil Eng 7:315-326.

Lund RD (1969) Synaptic patterns of the superficial layers of the superior colliculus of
the rat. J Comp Neurol 135:179-208.

Machida S, Kondo M, Jamison JA, Khan NW, Kononen LT, Sugawara T, Bush RA,
Sieving PA (2000) P23H rhodopsin transgenic rat: correlation of retinal function with
histopathology. Invest Ophthalmol Vis Sci 41:3200-3209.

Mahadevappa M, Weiland JD, Yanai D, Fine I, Greenberg RJ, Humayun MS (2005)
Perceptual thresholds and electrode impedance in three retinal prosthesis subjects. IEEE
Trans Neural Syst Rehabil Eng 13:201-206.

Mahnam A, Hashemi SM, Grill WM (2008) Computational evaluation of methods for
measuring the spatial extent of neural activation. J Neurosci Methods 173:153-164.

Majji AB, Humayun MS, Weiland JD, Suzuki S, D'Anna SA, de Juan E, Jr. (1999) Long-
term histological and electrophysiological results of an inactive epiretinal electrode array
implantation in dogs. Invest Ophthalmol Vis Sci 40:2073-2081.

Marc RE, Jones BW (2003) Retinal remodeling in inherited photoreceptor degenerations.
Mol Neurobiol 28:139-147.

Marc RE, Jones BW, Anderson JR, Kinard K, Marshak DW, Wilson JH, Wensel T,
Lucas RJ (2007) Neural reprogramming in retinal degeneration. Invest Ophthalmol Vis
Sci 48:3364-3371.

Margolis DJ, Detwiler PB (2007) Different mechanisms generate maintained activity in
ON and OFF retinal ganglion cells. J Neurosci 27:5994-6005.
170
Margolis DJ, Newkirk G, Euler T, Detwiler PB (2008) Functional stability of retinal
ganglion cells after degeneration-induced changes in synaptic input. J Neurosci 28:6526-
6536.

Massey SC (2005) Functional Anatomy of the Mammalian Retina: Elsevier.

Mayhew TM, Astle D (1997) Photoreceptor number and outer segment disk membrane
surface area in the retina of the rat: stereological data for whole organ and average
photoreceptor cell. J Neurocytol 26:53-61.

Mazzoni F, Novelli E, Strettoi E (2008) Retinal ganglion cells survive and maintain
normal dendritic morphology in a mouse model of inherited photoreceptor degeneration.
J Neurosci 28:14282-14292.

McConnell GC, Butera RJ, Bellamkonda RV (2009) Bioimpedance modeling to monitor
astrocytic response to chronically implanted electrodes. J Neural Eng 6:55005.

McCreery DB (2004) Tissue reaction to electrodes: the problem of safe and effective
stimulation of neural tissue. In: Neuroprosthetics: Theory and Practice (Horch KWaD,
G.S, ed): World Scientific Publishing Co. Pte. Ltd.

McCreery DB, Bullara LA, Agnew WF (1986) Neuronal activity evoked by chronically
implanted intracortical microelectrodes. Exp Neurol 92:147-161.

McCreery DB, Yuen TG, Bullara LA (2000) Chronic microstimulation in the feline
ventral cochlear nucleus: physiologic and histologic effects. Hear Res 149:223-238.

McCreery DB, Agnew WF, Bullara LA (2002) The effects of prolonged intracortical
microstimulation on the excitability of pyramidal tract neurons in the cat. Ann Biomed
Eng 30:107-119.

McCreery DB, Agnew WF, Yuen TG, Bullara LA (1988) Comparison of neural damage
induced by electrical stimulation with faradaic and capacitor electrodes. Ann Biomed Eng
16:463-481.

McCreery DB, Agnew WF, Yuen TG, Bullara L (1990) Charge density and charge per
phase as cofactors in neural injury induced by electrical stimulation. IEEE Trans Biomed
Eng 37:996-1001.

McCreery DB, Yuen TG, Agnew WF, Bullara LA (1992) Stimulation with chronically
implanted microelectrodes in the cochlear nucleus of the cat: histologic and physiologic
effects. Hear Res 62:42-56.

171
McCreery DB, Yuen TG, Agnew WF, Bullara LA (1994) Stimulus parameters affecting
tissue injury during microstimulation in the cochlear nucleus of the cat. Hear Res 77:105-
115.

McCreery DB, Agnew WF, Yuen TG, Bullara LA (1995) Relationship between stimulus
amplitude, stimulus frequency and neural damage during electrical stimulation of sciatic
nerve of cat. Med Biol Eng Comput 33:426-429.

McCreery DB, Yuen TG, Agnew WF, Bullara LA (1997) A characterization of the
effects on neuronal excitability due to prolonged microstimulation with chronically
implanted microelectrodes. IEEE Trans Biomed Eng 44:931-939.

McIlwain JT (1978) Cat superior colliculus: extracellular potentials related to W-cell
synaptic actions. J Neurophysiol 41:1343-1358.

McLaughlin ME, Sandberg MA, Berson EL, Dryja TP (1993) Recessive mutations in the
gene encoding the beta-subunit of rod phosphodiesterase in patients with retinitis
pigmentosa. Nat Genet 4:130-134.

Meldrum BS (1995) Excitatory amino acid receptors and their role in epilepsy and
cerebral ischemia. Ann N Y Acad Sci 757:492-505.

Mercanzini A, Colin P, Bensadoun JC, Bertsch A, Renaud P (2009) In vivo electrical
impedance spectroscopy of tissue reaction to microelectrode arrays. IEEE Trans Biomed
Eng 56:1909-1918.

Merrill DR, Bikson M, Jefferys JG (2005) Electrical stimulation of excitable tissue:
design of efficacious and safe protocols. J Neurosci Methods 141:171-198.

Micco AG, Richter CP (2006) Electrical resistivity measurements in the mammalian
cochlea after neural degeneration. Laryngoscope 116:1334-1341.

Milam AH, Dacey DM, Dizhoor AM (1993) Recoverin immunoreactivity in mammalian
cone bipolar cells. Vis Neurosci 10:1-12.

Milam AH, Li ZY, Fariss RN (1998) Histopathology of the human retina in retinitis
pigmentosa. Prog Retin Eye Res 17:175-205.

Montenegro I, Queiros MA, Daschbach JL (1991) Review of voltammetric methods for
the study of electrode reactions: Kluwer Academic Publishers.

Mortimer JT, Shealy CN, Wheeler C (1970) Experimental nondestructive electrical
stimulation of the brain and spinal cord. J Neurosurg 32:553-559.
172
Nakauchi K, Fujikado T, Kanda H, Kusaka S, Ozawa M, Sakaguchi H, Ikuno Y, Kamei
M, Tano Y (2007) Threshold suprachoroidal-transretinal stimulation current resulting in
retinal damage in rabbits. J Neural Eng 4:S50-57.

Nanduri D, Humayun MS, Greenberg RJ, McMahon MJ, Weiland JD (2008) Retinal
prosthesis phosphene shape analysis. Conf Proc IEEE Eng Med Biol Soc 2008:1785-
1788.

Nanduri D, Weiland JD, Horsager A, HUumayun MS, Greenberg RJ, Fine I (2007)
Selective Adaptation Using Electrical Stimulation in Humans. In: Association for
Research in Vision and Ophthalmology. Fort Lauderdale, Florida: Investigative
Ophthalmology & Visual Science.

Needham D, Hochmuth RM (1989) Electro-mechanical permeabilization of lipid
vesicles. Role of membrane tension and compressibility. Biophys J 55:1001-1009.

Newman J (1966) Resistance for Flow of Current to a Disk. Journal of the
Electrochemical Society 113:501-502.

Nishida K, Kamei M, Kondo M, Sakaguchi H, Suzuki M, Fujikado T, Tano Y (2009)
Efficacy of Suprachoroidal-Transretinal Stimulation in Rabbit Model of Retinal
Degeneration. Invest Ophthalmol Vis Sci.

Nour M, Naash MI (2003) Mouse models of human retinal disease caused by expression
of mutant rhodopsin. A valuable tool for the assessment of novel gene therapies. Adv Exp
Med Biol 533:173-179.

Okabe S, Shiomura Y, Hirokawa N (1989) Immunocytochemical localization of
microtubule-associated proteins 1A and 2 in the rat retina. Brain Res 483:335-346.

Palanker D, Vankov A, Huie P, Baccus S (2005) Design of a high-resolution
optoelectronic retinal prosthesis. J Neural Eng 2:S105-120.

Pardue MT, Phillips MJ, Yin H, Sippy BD, Webb-Wood S, Chow AY, Ball SL (2005)
Neuroprotective effect of subretinal implants in the RCS rat. Invest Ophthalmol Vis Sci
46:674-682.

Paxinos G, Watson C (1998) The Rat Brain in Stereotaxic Coordinates, Fourth Edition:
Academic Press.

Peichl L, Gonzalez-Soriano J (1994) Morphological types of horizontal cell in rodent
retinae: a comparison of rat, mouse, gerbil, and guinea pig. Vis Neurosci 11:501-517.

Penfield W, Rasmussen T (1950) The Cerebral Cortex of Man. New York: Macmillan.
173
Peng YW, Hao Y, Petters RM, Wong F (2000) Ectopic synaptogenesis in the mammalian
retina caused by rod photoreceptor-specific mutations. Nat Neurosci 3:1121-1127.

Peng YW, Senda T, Hao Y, Matsuno K, Wong F (2003) Ectopic synaptogenesis during
retinal degeneration in the royal college of surgeons rat. Neuroscience 119:813-820.

Perlmutter JS, Mink JW (2006) Deep brain stimulation. Annu Rev Neurosci 29:229-257.

Perry VH, Cowey A (1985) The ganglion cell and cone distributions in the monkey's
retina: implications for central magnification factors. Vision Res 25:1795-1810.

Pezaris JS, Reid RC (2007) Demonstration of artificial visual percepts generated through
thalamic microstimulation. Proc Natl Acad Sci U S A 104:7670-7675.

Plonsey R, Barr RC (1991) Functional Neuromuscular Stimulation. In: Bioelectricity: A
Quantitative Aprroach: Plenum Press.

Potts AM, Inoue J (1969) The electrically evoked response (EER) of the visual system.
II. Effect of adaptation and retinitis pigmentosa. Invest Ophthalmol 8:605-612.

Prokhorov E, Llamas F, Morales-Sanchez E, Gonzalez-Hernandez J, Prokhorov A (2002)
In vivo impedance measurements on nerves and surrounding skeletal muscles in rats and
human body. Med Biol Eng Comput 40:323-326.

Pudenz RH, Bullara LA, Dru D, Talalla A (1975) Electrical stimulation of the brain. II.
Effects on the blood-brain barrier. Surg Neurol 4:265-270.

Ranck JB, Jr. (1975) Which elements are excited in electrical stimulation of mammalian
central nervous system: a review. Brain Res 98:417-440.

Rattay F (1989) Analysis of models for extracellular fiber stimulation. IEEE Trans
Biomed Eng 36:676-682.

Rattay F (1999) The basic mechanism for the electrical stimulation of the nervous
system. Neuroscience 89:335-346.

Ray A, Colodetti L, Weiland JD, Hinton DR, Humayun MS, Lee EJ (2009)
Immunocytochemical analysis of retinal neurons under electrical stimulation. Brain Res
1255:89-97.

Ray A, Sun GJ, Chan L, Grzywacz NM, Weiland J, Lee EJ (2010) Morphological
alterations in retinal neurons in the S334ter-line3 transgenic rat. Cell Tissue Res 339:481-
491.
174
Reese BE, Raven MA, Stagg SB (2005) Afferents and homotypic neighbors regulate
horizontal cell morphology, connectivity, and retinal coverage. J Neurosci 25:2167-2175.

Richardson-Burns SM, Hendricks JL, Martin DC (2007a) Electrochemical
polymerization of conducting polymers in living neural tissue. J Neural Eng 4:L6-L13.

Richardson-Burns SM, Hendricks JL, Foster B, Povlich LK, Kim DH, Martin DC
(2007b) Polymerization of the conducting polymer poly(3,4-ethylenedioxythiophene)
(PEDOT) around living neural cells. Biomaterials 28:1539-1552.

Robblee LS, Rose TL (1990) Electrochemical guidlines for selection of protocols and
electrode materials for neural stimulation. In: Neural prostheses: Fundamental studies
(Agnew WF, McCreery DB, eds): Prentice hall.

Robblee LS, Lefko JL, Brummer SB (1983) Activated Ir: An electrode suitable for
reversible charge injection in saline solution. Journal of Electrochemical Society
130:731-732.

Rodieck RW (1973) The Vertebrate Retina: principles of structure and function. San
Francisco: W.H. Freman and Company.

Rohrer B, Crouch R (2006) Rod and cone pigment regeneration in RPE65-/- mice. Adv
Exp Med Biol 572:101-107.

Roque RS, Caldwell RB (1990) Muller cell changes precede vascularization of the
pigment epithelium in the dystrophic rat retina. Glia 3:464-475.

Rose TL, Robblee LS (1990) Electrical stimulation with Pt electrodes. VIII.
Electrochemically safe charge injection limits with 0.2 ms pulses. IEEE Trans Biomed
Eng 37:1118-1120.

Rovamo J, Virsu V (1979) An estimation and application of the human cortical
magnification factor. Exp Brain Res 37:495-510.

Ryu SB, Ye JH, Lee JS, Goo YS, Kim CH, Kim KH (2009) Electrically-evoked Neural
Activities of rd1 Mice Retinal Ganglion Cells by Repetitive Pulse Stimulation. Korean J
Physiol Pharmacol 13:443-448.

Sachs HG, Gekeler F, Schwahn H, Jakob W, Kohler M, Schulmeyer F, Marienhagen J,
Brunner U, Framme C (2005) Implantation of stimulation electrodes in the subretinal
space to demonstrate cortical responses in Yucatan minipig in the course of visual
prosthesis development. Eur J Ophthalmol 15:493-499.
175
Sagdullaev BT, Aramant RB, Seiler MJ, Woch G, McCall MA (2003) Retinal
transplantation-induced recovery of retinotectal visual function in a rodent model of
retinitis pigmentosa. Invest Ophthalmol Vis Sci 44:1686-1695.

Sauter AR, Dodgson MS, Kalvoy H, Grimnes S, Stubhaug A, Klaastad O (2009) Current
threshold for nerve stimulation depends on electrical impedance of the tissue: a study of
ultrasound-guided electrical nerve stimulation of the median nerve. Anesth Analg
108:1338-1343.

Schmidt EM, Bak MJ, Hambrecht FT, Kufta CV, O'Rourke DK, Vallabhanath P (1996)
Feasibility of a visual prosthesis for the blind based on intracortical microstimulation of
the visual cortex. Brain 119 ( Pt 2):507-522.

Sefton AJ, Swinburn M (1964) Electrical activity of lateral geniculate nucleus and optic
tract of the rat. Vision Res 4:315-328.

Sefton AJ, Dreher B (1985) Visual System: Academic Press.

Sekirnjak C, Hottowy P, Sher A, Dabrowski W, Litke AM, Chichilnisky EJ (2006)
Electrical stimulation of mammalian retinal ganglion cells with multielectrode arrays. J
Neurophysiol 95:3311-3327.

Shah S, Hines A, Zhou D, Greenberg RJ, Humayun MS, Weiland JD (2007) Electrical
properties of retinal-electrode interface. J Neural Eng 4:S24-29.

Shannon RV (1992) A model of safe levels for electrical stimulation. IEEE Trans Biomed
Eng 39:424-426.

Shannon RV, Moore JK, McCreery DB, Portillo F (1997) Threshold-distance measures
from electrical stimulation of human brainstem. IEEE Trans Rehabil Eng 5:70-74.

Shepherd RK, Clark GM, Black RC (1983) Chronic electrical stimulation of the auditory
nerve in cats. Physiological and histopathological results. Acta Otolaryngol Suppl
399:19-31.

Siminoff R, Schwassmann HO, Kruger L (1966) An electrophysiological study of the
visual projection to the superior colliculus of the rat. J Comp Neurol 127:435-444.

Solomon SG, Peirce JW, Dhruv NT, Lennie P (2004) Profound contrast adaptation early
in the visual pathway. Neuron 42:155-162.

Stasheff SF (2008) Emergence of sustained spontaneous hyperactivity and temporary
preservation of OFF responses in ganglion cells of the retinal degeneration (rd1) mouse. J
Neurophysiol 99:1408-1421.
176
Stieglitz T (2004) Electrode materials for recording and stimulation: World Scientific.

Stoney SD, Jr., Thompson WD, Asanuma H (1968) Excitation of pyramidal tract cells by
intracortical microstimulation: effective extent of stimulating current. J Neurophysiol
31:659-669.

Strettoi E, Mears AJ, Swaroop A (2004) Recruitment of the rod pathway by cones in the
absence of rods. J Neurosci 24:7576-7582.

Strettoi E, Porciatti V, Falsini B, Pignatelli V, Rossi C (2002) Morphological and
functional abnormalities in the inner retina of the rd/rd mouse. J Neurosci 22:5492-5504.

Strettoi E, Pignatelli V, Rossi C, Porciatti V, Falsini B (2003) Remodeling of second-
order neurons in the retina of rd/rd mutant mice. Vision Res 43:867-877.

Sullivan RK, Woldemussie E, Pow DV (2007) Dendritic and synaptic plasticity of
neurons in the human age-related macular degeneration retina. Invest Ophthalmol Vis Sci
48:2782-2791.

Sun W, Li N, He S (2002) Large-scale morophological survey of rat retinal ganglion
cells. Vis Neurosci 19:483-493.

Suzuki S, Humayun MS, Weiland JD, Chen SJ, Margalit E, Piyathaisere DV, de Juan E,
Jr. (2004) Comparison of electrical stimulation thresholds in normal and retinal
degenerated mouse retina. Jpn J Ophthalmol 48:345-349.

Szarowski DH, Andersen MD, Retterer S, Spence AJ, Isaacson M, Craighead HG, Turner
JN, Shain W (2003) Brain responses to micro-machined silicon devices. Brain Res
983:23-35.

Tanihara H, Hangai M, Sawaguchi S, Abe H, Kageyama M, Nakazawa F, Shirasawa E,
Honda Y (1997) Up-regulation of glial fibrillary acidic protein in the retina of primate
eyes with experimental glaucoma. Arch Ophthalmol 115:752-756.

Tehovnik EJ (1996) Electrical stimulation of neural tissue to evoke behavioral responses.
J Neurosci Methods 65:1-17.

Thomas BB, Aramant RB, Sadda SR, Seiler MJ (2006a) Retinal transplantation. A
treatment strategy for retinal degenerative diseases. Adv Exp Med Biol 572:367-376.

Thomas BB, Arai S, Ikai Y, Qiu G, Chen Z, Aramant RB, Sadda SR, Seiler MJ (2006b)
Retinal transplants evaluated by optical coherence tomography in photoreceptor
degenerate rats. J Neurosci Methods 151:186-193.
177
Tovar O, Tung L (1992) Electroporation and recovery of cardiac cell membrane with
rectangular voltage pulses. Am J Physiol 263:H1128-1136.

Tykocinski M, Shepherd RK, Clark GM (1995) Reduction in excitability of the auditory
nerve following electrical stimulation at high stimulus rates. Hear Res 88:124-142.

Vardi N, Auerbach P (1995) Specific cell types in cat retina express different forms of
glutamic acid decarboxylase. J Comp Neurol 351:374-384.

Veraart C, Wanet-Defalque MC, Gerard B, Vanlierde A, Delbeke J (2003) Pattern
recognition with the optic nerve visual prosthesis. Artif Organs 27:996-1004.

Walsh DM, Howe TE, Johnson MI, Sluka KA (2009) Transcutaneous electrical nerve
stimulation for acute pain. Cochrane Database Syst Rev:CD006142.

Walter P, Szurman P, Vobig M, Berk H, Ludtke-Handjery HC, Richter H, Mittermayer
C, Heimann K, Sellhaus B (1999) Successful long-term implantation of electrically
inactive epiretinal microelectrode arrays in rabbits. Retina 19:546-552.

Wang K, Fishman HA, Dai H, Harris JS (2006) Neural stimulation with a carbon
nanotube microelectrode array. Nano Lett 6:2043-2048.

Wang MM, Janz R, Belizaire R, Frishman LJ, Sherry DM (2003) Differential distribution
and developmental expression of synaptic vesicle protein 2 isoforms in the mouse retina.
J Comp Neurol 460:106-122.

Wang S, Lu B, Lund RD (2005) Morphological changes in the Royal College of
Surgeons rat retina during photoreceptor degeneration and after cell-based therapy. J
Comp Neurol 491:400-417.

Wang S, Lu B, Girman S, Holmes T, Bischoff N, Lund RD (2008) Morphological and
functional rescue in RCS rats after RPE cell line transplantation at a later stage of
degeneration. Invest Ophthalmol Vis Sci 49:416-421.

Wassle H, Grunert U, Rohrenbeck J, Boycott BB (1990) Retinal ganglion cell density and
cortical magnification factor in the primate. Vision Res 30:1897-1911.

Weiland JD, Anderson DJ (2000) Chronic neural stimulation with thin-film, iridium
oxide electrodes. IEEE Trans Biomed Eng 47:911-918.

Weiland JD, Anderson DJ, Humayun MS (2002a) In vitro electrical properties for iridium
oxide versus titanium nitride stimulating electrodes. IEEE Trans Biomed Eng 49:1574-
1579.
178
Weiland JD, Liu W, Humayun MS (2005) Retinal prosthesis. Annu Rev Biomed Eng
7:361-401.

Weiland JD, Fujii G, Mahadevappa M, Greenberg R, Tameesh M, Guven D, de Juan E,
Humayun MS (2002b) Chronic Electrical Stimulation of the Canine Retina. Proceedings
of the Second Joint EMBS/BMES Conference IEEE:2.

Whalen JJ, Weiland JW, Searson P (2005) Electrochemical deposition of platinum from
aqueous ammonium hexachloroplatinate solution. Journal of electrochemical society
152:C738-743.

Wheater PR, Burkitt HG, Daniels VG (1982) Functional Histology. London and New
York: Churchill Livingstone.

Williams JC, Hippensteel JA, Dilgen J, Shain W, Kipke DR (2007) Complex impedance
spectroscopy for monitoring tissue responses to inserted neural implants. J Neural Eng
4:410-423.

Xu J, Shepherd RK, Millard RE, Clark GM (1997) Chronic electrical stimulation of the
auditory nerve at high stimulus rates: a physiological and histopathological study. Hear
Res 105:1-29.

Yang JY, Martin DC (2004) Microporous conducting polymers on neural microelectrode
arrays II. Physical characterization. Sensors and Actuators A: Physical 113A:204-211.

Zrenner E (2002) The subretinal implant: can microphotodiode arrays replace
degenerated retinal photoreceptors to restore vision? Ophthalmologica 216 Suppl 1:8-20;
discussion 52-23.

Abstract (if available)

Abstract Electrical stimulation of the central nervous system albeit an unnatural way, has been found to be an effective way of causing neuronal excitation. Retinal prosthesis is an example of such a neuroprosthesis that strives to provide vision to people suffering from Retinitis Pigmentosa and Age-related Macular Degeneration. In these diseases, the photoreceptors in the retina undergo a progressive degeneration leaving the remaining retinal neurons relatively intact. It is by electrically stimulating these retinal neurons that a retinal prosthesis aims to elicit visual percepts.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Electrical stimulation of degenerate retina

PDF

Towards a high resolution retinal implant

PDF

Cortical and subcortical responses to electrical stimulation of rat retina

PDF

Manipulation of RGCs response using different stimulation strategies for retinal prosthesis

PDF

Stimulation strategies to improve efficiency and temporal resolution of epiretinal prostheses

PDF

Saliency based image processing to aid retinal prosthesis recipients

PDF

PET study of retinal prosthesis functionality

PDF

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

PDF

Modeling retinal prosthesis mechanics

PDF

Improving stimulation strategies for epiretinal prostheses

PDF

Investigation of the electrode-tissue interface of retinal prostheses

PDF

Viewing the picture we paint

PDF

Electrical stimulation approaches to restoring sight and slowing down the progression of retinal blindness

PDF

Electrical stimulation of the orbicularis oculi to restore eye blink

PDF

Therapeutic electrical stimulation strategies for neuroregeneration and neuroprotection of retinal neurons

PDF

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

PDF

Prosthetic visual perception: retinal electrical stimulation in blind human patients

PDF

An in vitro model of a retinal prosthesis

PDF

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

PDF

Intraocular and extraocular cameras for retinal prostheses: effects of foveation by means of visual prosthesis simulation

Asset Metadata

Creator Ray, Aditi (author)

Core Title Effect of continuous electrical stimulation on retinal structure and function

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Biomedical Engineering

Publication Date 08/06/2010

Defense Date 04/14/2010

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

Tag electrical stimulation,immunohistochemistry,impedance,OAI-PMH Harvest,retina,Retinal degeneration,superior colliculus

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Weiland, James D. (committee chair), Hinton, David R. (committee member), Humayun, Mark S. (committee member), Lee, Eun Jin (committee member), Loeb, Gerald E. (committee member), Mansfeld, Florian B. (committee member)

Creator Email aditiray@gmail.com,aditiray@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m3297

Unique identifier UC1312409

Identifier etd-Ray-3727 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-372801 (legacy record id),usctheses-m3297 (legacy record id)

Legacy Identifier etd-Ray-3727.pdf

Dmrecord 372801

Document Type Dissertation

Rights Ray, Aditi

Type texts

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

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu