Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Multi-region recordings from the hippocampus with conformal multi-electrode arrays
(USC Thesis Other)
Multi-region recordings from the hippocampus with conformal multi-electrode arrays
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
MULTI-REGION RECORDINGS FROM THE HIPPOCAMPUS WITH CONFORMAL
MULTI-ELECTRODE ARRAYS
Dissertation by
Huijing Xu
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
BIOMEDICAL ENGINEERING
University of Southern California
Los Angeles, California
A DISSERTATION PRESENTED TO THE FACULTY OF THE USC GRADUATE SCHOOL
VITERBI SCHOOL OF ENGINEERING
DECEMBER 2019
ii
© Copyright by
Huijing Xu
December 2019
iii
To Life and Death
iv
Table of Contents
Chapter 1 Introduction and Background ......................................................................................... 1
1.1 Motivation ........................................................................................................................ 1
1.2 The hippocampus ............................................................................................................. 2
1.3 Wireless Implantable Bioelectronic System for Neuroscience Research (WIBSNR) ..... 4
1.4 Hippocampal Prosthesis ................................................................................................... 5
1.5 Recording from the Brain ................................................................................................. 7
1.6 Objectives ......................................................................................................................... 9
1.7 References ...................................................................................................................... 10
Chapter 2 Development of the Triple-Region Microwire Electrodes Array ................................ 13
2.1 Design of the Triple-Region Microwire Electrodes Array............................................. 14
2.1.1 Brain atlas measurements ....................................................................................... 15
2.1.2 Histological measurement of the depth of cell body layers .................................... 15
2.1.3 Relative depths of multiple sub-regions of the hippocampus calculated from neural
recordings .............................................................................................................................. 18
2.1.4 The triple-region microwire electrodes array ......................................................... 24
2.2 in vivo Evaluation of the Triple-Region Microwire Electrodes Array ........................... 25
2.2.1 Recording with the triple-region microwire electrodes array .................................. 26
2.2.2 Limitation of the triple-region microwire electrodes array ..................................... 27
2.3 Summary ........................................................................................................................ 27
2.4 References ...................................................................................................................... 28
Chapter 3 Development of the Parylene-Based Multi-Electrode Array ....................................... 30
3.1 Design of the Parylene Multi-Electrode Array .............................................................. 31
3.2 Fabrication and Electrical Packaging ............................................................................. 33
3.2.1 Optimization of fabrication procedure ................................................................... 34
3.2.2 Final fabrication scheme ......................................................................................... 38
3.2.3 Electrical packaging ............................................................................................... 40
3.3 Development of Array Insertion Techniques ................................................................. 41
3.3.1 Insertion shuttles ..................................................................................................... 41
3.3.2 Dissolvable brace for insertion support .................................................................. 42
v
3.4 Summary ........................................................................................................................ 44
3.5 References ...................................................................................................................... 45
Chapter 4 Evaluation of the Parylene Multi-Electrode Array ...................................................... 47
4.1 Benchtop Testing of the Parylene Multi-Electrode Array ............................................. 47
4.1.1 Electrode impedance ............................................................................................... 47
4.1.2 in vitro testing of the insertion strategy .................................................................. 49
4.2 Acute, in vivo Evaluation of the Parylene Multi-Electrode Array ................................. 50
4.2.1 in vivo insertion of the Parylene multi-electrode array ........................................... 51
4.2.2 Acute, in vivo recording with the Parylene multi-electrode array .......................... 53
4.3 Chronic, in vivo Evaluation of the Parylene Multi-Electrode Array .............................. 58
4.3.1 Chronic fixation of the Parylene multi-electrode array and the package ............... 59
4.3.2 Long-term recording with the Parylene multi-electrode array ............................... 61
4.3.3 Life-time testing ..................................................................................................... 66
4.3.4 Immune responds to the Parylene multi-electrode array ........................................ 67
4.4 Comparison between Microwire Electrodes Arrays and the Parylene Multi-Electrode
Array ........................................................................................................................................ 73
4.4.1 Comparison of signals recorded under acute preparations ..................................... 74
4.4.2 Comparison of signal quality under chronic preparations ...................................... 76
4.4.3 Literature comparison ............................................................................................. 76
4.5 Summary ........................................................................................................................ 78
4.6 References ...................................................................................................................... 78
Chapter 5 Neuroscience Applications of Conformal, Multi-Electrode Arrays ............................. 80
5.1 Recording of Place Cells from Multiple, Hippocampal Sub-Regions ........................... 82
5.1.1 Hippocampal place cells ......................................................................................... 83
5.1.2 Recording from free moving animals ..................................................................... 84
5.1.3 Place cells in individual sub-regions of the hippocampus ...................................... 86
5.1.4 Place cells simultaneously recorded from multiple sub-regions of the hippocampus
............................................................................................................................................... 88
5.2 Neural Response to Manipulations of Visual Stimulus ................................................. 88
5.2.1 Novel cue card ........................................................................................................ 89
5.2.2 Novel objects .......................................................................................................... 90
vi
5.2.3 Cue card rotation .................................................................................................... 91
5.3 Summary ........................................................................................................................ 95
5.4 References ...................................................................................................................... 96
Chapter 6 Conclusion and Future Directions ................................................................................ 98
6.1 Development of Conformal, Multi-regional Neural Interfaces ...................................... 98
6.2 Multi-regional Recordings from Behaving Animals .................................................... 100
6.3 References .................................................................................................................... 102
vii
Acknowledgments
First, I would like to thank my advisors, Professor Dong Song and Professor Theodore W.
Berger for providing me with the opportunity and resource to work on this exciting project. This
work and thesis would not have been made possible without their continuous guidance and
encouragement. Their vision of the future of neuroengineering and their passion and dedication to
their work inspired me to pursue the work presented in this thesis.
I also would like to give special thanks to my dissertation committee members, Professor
Ellis Meng and Professor Hossein Hashemi for their great support. They are not just committee
members but also close collaborators. The work presented in this thesis cannot be achieved without
their groups’ hard works and great contributions. Specially, I would like to thank Dr. Ahuva W.
Hirschberg who provided fabrication and benchtop testing of the Parylene multi-electrode array
and participated basically in every step of the develop and evaluation of the Parylene array.
I also would like to thank NSF INSPIRE and NIH BRAIN for their great support to this
project.
In addition, I would like to thank all the support from previous and current lab members,
including but not limit to Drs. Jean-Marie Bouteiller, Min-Chi Hsiao, Ude Lu, Rosa Chan, Viviane
S. Ghaderi, Sushmita L. Allam, Brian Robinson, Gene Yu and et al. for their help and supports.
Finally, I would like to thank my family, especially my husband for their love and constant
supports.
viii
Abstract
To understand how the formation of long-term memory arises from the interaction between
hippocampal neurons, it is necessary to have access to individual sub-regions of the hippocampal
circuit at the single-neuron level. The success of a hippocampal prosthetic device also relies highly
upon its ability to attain resolvable unitary activities from populations of neurons in multiple
regions of the hippocampal networks over long-term. Multi-electrode arrays are the primary
interfacing device used to collect unitary activities from the hippocampus and are also critical for
closed-loop hippocampal prosthesis. To achieve multi-regional recording from the hippocampus,
multi-electrode arrays must be specifically designed to accommodate the anatomy of the
hippocampus. In this work, conformal multi-electrode arrays that matched the anatomical
curvature of the rat hippocampus were designed, developed and chronically implanted.
Optimizations of the standard microwire electrode array were made to develop a triple-region,
microwire-based multi-electrode array with varied wire lengths to target at all three sub-regions of
the rat hippocampus. To minimize the foreign body response to a rigid, penetrating neural implant,
a novel, Parylene-based multi-electrode array with a reduced elastic modulus was also developed.
This Parylene-based multi-electrode array was designed and fabricated to have 64 electrodes
positioned to match the anatomy of the rat hippocampus and allowed for simultaneous recordings
from two cell body layers of the tri-synaptic pathway. The performance of the Parylene array was
evaluated both acutely and chronically. On average, over 70 units (n=7) were simultaneously
recorded from both the CA1 and CA3 sub-regions under acute preparations. Chronically, unitary
activities with good signal-to-noise ratio were stably recorded from multiple sub-regions for at
least 70 days (n=4). The location-specific firing properties of hippocampal neurons and non-
stationary responds of place cells to manipulations of visual stimuli to the surrounding
environment were simultaneously recorded from multiple sub-regions of the hippocampus with
these conformal, hippocampal multi-electrode arrays while the animal was behaving freely in an
open field. Such multi-regional recordings are the foundation to develop mathematical models that
can fully characterize the connectivity between hippocampal sub-regions and also provided
valuable experimental data for the computational analysis of synaptic plasticity within the
hippocampal circuit, which will expand the understanding of memory encoding and decoding in
the hippocampus to a system level.
1
Chapter 1. Introduction and Background
Introduction and Background
1.1 Motivation
Advanced cognitive functions of the brain, including that of memory, rely on effective
communications between individual neurons within neural networks. The hippocampus, a deep
brain structure, has been known as the center for the formation of new, long-term, declarative
memories for decades [1][2]. The communication between hippocampal neurons via action
potentials propagating through axons and dendrites expanding to different sub-regions of the
hippocampus accomplishes the generation of new long-term memories and retrieval of existing
memories [3]. A small size electrode, i.e. an implantable extracellular microelectrode, placed close
enough to single hippocampal neurons can sense the rapid change of electrical field around the
neuron and read these electrical signals out. This extracellular recording technique provides us a
great opportunity to study neural activities of the hippocampus with the highest special resolution
[4]. With extracellular recording electrodes implanted to multiple hippocampal sub-regions, the
study of functional connections and synaptic plasticity between hippocampal sub-regions which is
the key to understanding how the brain processes memory information is also made possible [5].
Not just that the understanding of brain functions such as learning and memory requires
direct interacting with multiple cortical and subcortical regions with appropriate techniques, the
brain itself is not just the most important but also one of the most fragile organs in our body. The
hippocampus is particularly vulnerable to brain trauma and stroke [6], and it is the brain region
2
that relates to many irreversible neurodegenerative diseases such as Alzheimer’s disease [7][8].
The development of a computational model based, implantable prosthetic devices, hippocampal
prosthesis, that can restore the function of the intact hippocampus would have a profound impact
on the quality of life of patients with diseases related to the hippocampus [9][10]. The first
technical challenge that we face is a perfect neural interface which can get recordings from
populations of neurons in multiple sub-regions of the hippocampal circuit.
In addition, biological factors include the foreign body response to the implant and
inflammatory responses which cause the formation of a glial scar around the implant over time
will isolate recording electrodes from the neural tissue and prevent the electrodes from collecting
neural signals over long-term [11]. Therefore, the development of intracortical electrodes capable
of collecting unitary activities from populations of neurons with high signal-to-noise (SNR) ratio
over long periods of time is necessary. In summary, the development of conformal neural
recording devices which can record unitary activities from multiple sub-regions of the
hippocampal circuit over long-term will not only benefit the development of memory prostheses
but can also add insights to the understanding of underlying mechanism of memory encoding and
decoding and will in turn help the understanding of how high level cognitive functions are
accomplished by the brain.
1.2 The hippocampus
The clear link between the hippocampus and memory functions has been known for several
decades. After the obvious memory impairment shown following the removal of the medial
temporal lobe of patient H. M., more and more studies have focused on the function of one major
structure within the medial temporal lobe, the hippocampus [12]. From rodents to human beings,
there is considerable data which supports the involvement of the hippocampus in spatial
information processing [13][14], novelty detection [15] and the formation of new, long-term,
declarative memories [16].
3
If a slice of tissue is taken along the longitudinal axis of the curved tube-like brain structure,
it can be seen that hippocampal neurons and interneurons lie compactly within the thin cell body
layers that curve into a double-C shaped structure. Anatomically, the hippocampus is composed
of three cytoarchitectonically distinct subdivisions: the cornu ammonis (CA) 1, CA3, and dentate
gyrus (DG) [17]. In between cell body layers, lies the axon bundle that provide inputs to each sub-
region and dendrites that receives inputs. The major, unidirectional axon bundle that connects all
three sub-regions of the hippocampal formation is called the tri-synaptic pathway. In short, DG
cells receive inputs from the entorhinal cortex (EC) through the perforant pathway and gives rise
to the mossy fiber pathway which project to the CA3. The CA3 then projects to the CA1 sub-
region via Schaffer collaterals [18]. This intrinsic, cascade, and excitatory connection has long
been hypothesized to have considerable importance in information processing and the modulation
of synaptic efficiency. In addition, the clear anatomical difference between hippocampal sub-
regions CA3 and CA1 as well as the DG [19] and diverse polymodal inputs that each hippocampal
sub-region receive from cortical regions have engendered proposals for distinct functions of these
different sub-regions. Lesion studies in rat models, which show that damages to individual
hippocampal sub-regions have caused different impairment in the ability of the rat to detect spatial
novelty [15], add more evidences to the theory that unique functions are associated with each
hippocampal sub-region. Simultaneous recordings of neural activities from multiple hippocampal
Figure 1. Schematic representation of the rat brain and the relative location of the left hippocampus.
Pictured on the top right shows a transverse slice of the hippocampus with the diagrammatic
representation of the tri-synaptic pathway. The bottom diagram shows the major neural circuits between
hippocampal sub-regions.
DG CA3 CA1 EC
Perforant Pathwawy
Schaffer
collaterals
4
sub-regions are necessary to the investigation of roles that each sub-region plays in long-term
memory formation and the study of synaptic plasticity in the hippocampal system during memory
formation.
1.3 Wireless Implantable Bioelectronic System for Neuroscience
Research (WIBSNR)
In the field of biomedical engineering, animal study is the one essential step before any
technique can be applied to clinical use. Rodents are the most commonly used animal model since
they are not only easy to handle, and cost efficient, their genetic, biological and behavior
characteristics closely resemble those of humans. In addition, many diseases and symptoms of
human can be replicated in rodents such as rats. Especially in neuroscience studies, researches on
behaving rats have accumulated considerable knowledge about brain functions and has built up
the foundation of the understanding of underlying mechanisms of memory formation.
Traditionally, the study of neural activities in behaving rats involves the implantation of
recording components to desired brain regions, connecting the recording electrodes to data
acquisition systems with bundles of wires and collecting neural data while the animal performs
certain tasks. To what extend a neuro-scientific question can be answered is highly limited by these
research tools. For example, the number of electrodes implanted directly decides the amount of
Figure 2. The concept of the wireless implantable bioelectronic system (WIBSNR) for neuroscience
research. The illustrative diagram shows components that necessary for the WIBSNR.
5
brain regions and the number of neurons that can be reached. The property of these electrodes also
determines whether bi-directional interactions with neural tissue is possible [20]. High-quality
signal amplifications and efficient noise rejections all rely on exceptional design of the acquisition
system. How neural data is transferred also affects the complicity of behavioral experiments
conducted. With the development of material science, electrode fabrication techniques, integrated
circuits design and minimizing of electronic components, it is the right time to develop wireless,
fully implantable bioelectronic systems specifically for rodent neuroscience research (WIBSNR).
Such a system will be capable of interfacing with neural tissues in a bi-directional manner
(recording and stimulation) continuously over extended periods while the fully un-tethered animal
is behaving normally in a close-to-nature environment. Enabled by WIBSNR, it will be possible
to identify the neural correlations of natural behaviors. Such experiments can also provide valuable
insights to the understanding of functional neural connectivity within neural networks and provide
experimental data for computational studies of brain functions. Electrode arrays that can
simultaneously interface with multiple brain regions within a neural network over long periods of
time is a crucial element for such a system.
1.4 Hippocampal Prosthesis
A prosthesis usually refers to an artificial device that is used to replace a missing body part.
In the central and peripheral nervous system, neural prostheses, which are designed to restore or
replace certain functions of the nervous system, have been developed in the past several decades.
The cochlear implant is one of the most famous and commonly used neural prosthetic device. By
converting sounds with different frequencies and amplitudes to electrical stimulations in
corresponding regions of the auditory nerve with varied stimulation duration, frequency and
strength, the hearing ability of patients with hair cell impairments can be partially or even fully
restored [21]. Retinal prosthesis is another sensory prosthesis recently been approved by the FDA
for clinical use to treat retinitis pigmentosa [22]. Strength and location of light are represented by
electrical stimulation to small populations of neural cells in particular region of the retina which
then send signals to downstream to intact visual centers of the brain. This way degenerated neural
cells in the retina can be bypassed and partial sight restoration is made possible.
The success of neural prosthetic devices in sensory function restorations has greatly
encouraged the development of cortical prosthesis which focuses on restoring motor or sensory
6
functions through direct interactions with particular regions of the central nervous system, and the
brain. Similar as those sensory prostheses mentioned above, a cortical prosthesis that provides
restoration of motor functions follows the same strategy. This cortical, motor prosthesis utilizes
the well-established correlation of motor cortical neural ensemble firing with arm movements,
using the brain-computer interfaces (BCIs) technique, and achieved control of arm movements
through scalp recorded neural events. Such cortical prosthesis can provide relief and improvement
of motor functions in patients who suffer from pathological conditions [23].
Although those sensory and motor prostheses mentioned above target at various parts of
the nervous system, there is one common feature shared by them all: these prostheses rely on the
direct mapping between sensory or motor functions and neural activities. This direct mapping
becomes inapplicable when a prosthesis aims to restore high-level, cognitive functions such as the
memory function. One of the most advanced cognitive prostheses to date is the hippocampal
prosthesis developed by Dr. Theodore Berger’s team at the University of Southern California and
Wake Forest University. Activities from ensembles of CA3 and CA1 neurons in the hippocampus
Figure 3. Diagrams of sensory prostheses, with a cochlear implant shows at top left and a retinal
prosthesis shows at top right. The diagram of the hippocampal prosthesis is shown at the bottom.
7
of rats were recorded and provided data for the development of a multi-input/multi-output (MIMO)
computational model which characterizes nonlinear information transformations performed by the
healthy hippocampus. When electrical stimulates were sent through CA1 electrodes following the
stimulation pattern predicted by the MIMO model, the memory performance of animals whose
hippocampal synaptic transmission which had been blocked pharmacologically were recovered
[10]. For such a hippocampal prosthesis, a set of well-designed electrodes capable of recording
from multiple hippocampal sub-regions of behaving animals over long-term is not only necessary
for the identification of neural coding of memories but is also an essential part of the hippocampal
prosthesis itself.
1.5 Recording from the Brain
Microwire electrodes arrays have the longest history of being used for recordings from
individual neurons in the brain. It is constructed with thin metal wires insulated with biocompatible
polymers with limited exposition at the wire tip. Since microwires can provide sufficient
mechanical rigidity for penetrating brain tissue and has a relatively low impedance for signals in
the spike frequency range [24], it has been widely used to interface with neurons in deep brain
Figure 4. Illustrations of rigid multi-electrode arrays. Picture on the left shows the sketch of a
conventional microwire electrode array with a zoom-in view of the exposed microwire tip. Picture on
the right shows the layout of a liner silicon probe arrays with a zoom in view of multiple recording sites
evenly placed along a silicon shank.
Microwire Electrode Array
Silicon Probe
Array
8
structures such as the hippocampus over the past several decades. Commercially available
microwire electrodes arrays with uniformed wire length remains one of the most standard
recording tool this day. Under the best circumstances, unitary activities from single neuron can be
recorded for more than a year [25]. However, the fact that each microwire has only a single
exposed tip to interface with surrounding neural tissue greatly limits the number of recording sites
implanted. This limitation has been overcome, in part, by microfabricated silicon-based neural
probes, produced using standard micromachining techniques common to the semiconductor
industry. Silicon-based probes can include dozens to hundreds of recording sites whose location
along the shank and electrode geometries can be precisely controlled, allowing for recordings from
multiple neurons along the length of the neural track [26].
Despite the success of both silicon and microwire devices in animal and human subjects, a
growing body of scientific research suggests the material composition of these devices is ill-suited
for reliable, long-term recording applications. The high elastic moduli of silicon and metal
(hundreds of GPa) compared to the brain (~10
-6
GPa) [27] is frequently cited as a cause of tissue
irritation and subsequent immune response. It has been posited that damage to the brain incurred
by the micromotion of the rigid implants instigates tissue trauma in the form of local neuronal
death and glial encapsulation around the implantation site [28][29], reducing recording quality or
eliminating resolvable signals in as little as a few weeks [30]. Although microdrives can maximize
device lifetime by periodically advancing the probe to a new area of tissue to temporarily bypass
fibrotic encapsulation, this approach precludes the ability to reliably target the same region
chronically [31].
A potential design solution is to reduce the rigidity of the probe by replacing
micromachined silicon with thin-film flexible polymers, such as polyimide, SU-8, or poly-(para-
chloro-xylylene) (Parylene C) [32]–[34]. Recent work, including modeling, in vitro, and in vivo
studies have found that the use of compliant substrates mitigates inflammation and preserves
neuronal density when compared to rigid materials [28]. By adapting micropatterning and
micromachining techniques to polymer substrates, several iterations of polymer-based probes and
9
arrays have been successfully deployed for acute and chronic recording applications in animal
models [33]–[35].
1.6 Objectives
The advancement of recording technique greatly promoted the understanding of brain
functions. Implantable microwires allows for the investigation of activities of single neurons and
small groups of neurons without impalement. With extracellular recording electrodes, neural
activities from single unit can be isolated and studied for hours, or even days, in behaving animals.
It is those advanced neural interfacing techniques makes the comprehensively characterization of
physiology and connectivity of single neurons possible. To further answer scientific questions such
as how neurons inside our brain process perceptions, thoughts, and even memories, techniques
capable of identifying populations of neurons within the neural circuits of interests at single-neuron
resolution during behavior are required.
This work focuses on the development of multi-electrodes arrays that are capable of
recording unitary activities simultaneously from multiple sub-regions of the hippocampal network
from behaving animals. Optimization to the standard uni-length microwire electrodes array will
be made to achieve simultaneously recording from all three sub-regions of the tri-synaptic pathway.
Microwire electrode arrays that conformal to the anatomical structure of the rat hippocampus will
Figure 5. A flexible, multi-electrode probe fabricated with Parylene C. Diagrams on the top row show
the design of the probe and three designs of pocket for microwires that can be used to assist the insertion
of the probe. Photos on the bottom show fabricated Parylene probes.
10
be designed, implanted chronically and evaluated in vivo. With a growing body of research
indicating the clear benefit of using flexible material to reduce tissue responds to the invasive
implants [36], the possibility of using a flexible, polymer, the Parylene C, as the structural material
of a multi-electrode array will also be explored. A conformal, Parylene-based, multi-electrode
array will be designed and fabricated. Corresponding package that allows the connection of such
an interface to conventional data acquisition systems and insertion techniques for the flexible
Parylene array will also be developed. In addition, the performance of the Parylene array will be
evaluated both acutely and chronically. At last, behavioral experiments will be conducted to show
how such conformal multi-electrode arrays can help in the understanding of memory formation
and identification of synaptic plasticity in the hippocampal circuit.
1.7 References
[1] H. Eichenbaum, “Hippocampus : Cognitive Processes and Neural Representations that Underlie
Declarative Memory The hippocampus serves a critical role in declarative,” Neuron, vol. 44, pp.
109–120, 2004.
[2] H. Eichenbaum, “The hippocampus and mechanisms of declarative memory.,” Behav. Brain Res.,
vol. 103, no. 2, pp. 123–33, Sep. 1999.
[3] H. Eichenbaum, T. Otto, and N. J. Cohen, “The hippocampus—what does it do?,” Behav. Neural
Biol., vol. 57, no. 1, pp. 2–36, Jan. 1992.
[4] M. M. Heinricher, “Principles of Extracellular Single-Unit Recording,” Microelectrode Rec. Mov.
Disord. Surg., vol. c, pp. 8–13, 2004.
[5] D. Song, B. Robinson, R. Hampson, V. Marmarelis, S. Deadwyler, and T. Berger, Sparse large-
scale nonlinear dynamical modeling of human hippocampus for memory prostheses, vol. PP, no. 99.
2016.
[6] M. J. Kotapka, D. I. Graham, J. H. Adams, and T. A. Gennarelli, “Hippocampal pathology in fatal
non-missile human head injury,” Acta Neuropathol., 1992.
[7] G. Allen et al., “Reduced hippocampal functional connectivity in Alzheimer disease,” Arch. Neurol.,
2007.
[8] D. P. Devanand et al., “Hippocampal and entorhinal atrophy in mild cognitive impairment:
Prediction of Alzheimer disease,” Neurology, 2007.
[9] T. W. Berger et al., “Restoring lost cognitive function.,” IEEE Eng. Med. Biol. Mag., vol. 24, no. 5,
pp. 30–44, 2005.
11
[10] T. W. Berger, R. E. Hampson, D. Song, A. Goonawardena, V. Z. Marmarelis, and S. a Deadwyler,
“A cortical neural prosthesis for restoring and enhancing memory.,” J. Neural Eng., vol. 8, no. 4, p.
046017, Aug. 2011.
[11] S. A. Deadwyler, E. Hampson, W. Forest, and N. Carolina, “Hippocampal Ensemble Activity during
Spatial Performance in Rats,” vol. 16, no. 1, pp. 354–372, 1996.
[12] W. B. Scoville and B. Milner, “LOSS OF RECENT MEMORY AFTER BILATERAL
HIPPOCAMPAL LESIONS,” J. Neurol. Neurosurg. Psychiatry, vol. 20, no. 1, pp. 11–21, 1957.
[13] J. O’Keefe, “Place units in the hippocampus of the freely moving rat,” Exp. Neurol., vol. 51, no. 1,
pp. 78–109, 1976.
[14] E. Park, D. Dvorak, and A. a Fenton, “Ensemble place codes in hippocampus: CA1, CA3, and
dentate gyrus place cells have multiple place fields in large environments.,” PLoS One, vol. 6, no.
7, p. e22349, Jan. 2011.
[15] I. Lee, M. R. Hunsaker, and R. P. Kesner, “The role of hippocampal subregions in detecting spatial
novelty.,” Behav. Neurosci., vol. 119, no. 1, pp. 145–53, Feb. 2005.
[16] N. Burgess, E. A. Maguire, and J. O’Keefe, “The human hippocampus and spatial and episodic
memory,” Neuron. 2002.
[17] R. Lorente De Nó, “Studies on the structure of the cerebral cortex. II. Continuation of the study of
the ammonic system.,” J. für Psychol. und Neurol., vol. 46, pp. 113–117, 1934.
[18] R. S. G. Jones, “Entorhinal-hippocampal connections : a speculative view of their function,” vol. 16,
no. 2, 1993.
[19] D. G. Amaral and M. P. Witter, “The three-dimensional organization of the hippocampal formation:
A review of anatomical data,” Neuroscience, vol. 31, no. 3, pp. 571–591, 1989.
[20] S. F. Cogan, “Neural Stimulation and Recording Electrodes,” Annu. Rev. Biomed. Eng., vol. 10, no.
1, pp. 275–309, 2008.
[21] O. Macherey and R. P. Carlyon, “Cochlear implants,” Current Biology. 2014.
[22] James D. Weiland and Mark S. Humayun, “Retinal Prosthesis,” vol. 25, no. 3, pp. 289–313, 2016.
[23] D. J. Guggenmos et al., “Restoration of function after brain damage using a neural prosthesis,” Proc.
Natl. Acad. Sci. U. S. A., 2013.
[24] D. R. Humphrey and E. M. Schmidt, “Extracellular Single-Unit Recording Methods *,” 1977.
[25] J. C. Williams, R. L. Rennaker, and D. R. Kipke, “Long-term neural recording characteristics of
wire microelectrode arrays implanted in cerebral cortex,” pp. 303–313, 1999.
[26] Z. Fekete, “Recent advances in silicon-based neural microelectrodes and microsystems: A review,”
Sensors Actuators, B Chem., vol. 215, pp. 300–315, 2015.
12
[27] K. Scholten and E. Meng, “Materials for microfabricated implantable devices: a review,” Lab Chip,
vol. 15, no. 22, pp. 4256–4272, 2015.
[28] T. D. Y. Kozai, A. S. Jaquins-Gerstl, A. L. Vazquez, A. C. Michael, and X. T. Cui, “Brain tissue
responses to neural implants impact signal sensitivity and intervention strategies,” ACS Chem.
Neurosci., vol. 6, no. 1, pp. 48–67, 2015.
[29] R. Biran, D. C. Martin, and P. A. Tresco, “Neuronal cell loss accompanies the brain tissue response
to chronically implanted silicon microelectrode arrays,” Exp. Neurol., vol. 195, no. 1, pp. 115–126,
2005.
[30] R. J. Vetter, J. C. Williams, J. F. Hetke, E. a. Nunamaker, and D. R. Kipke, “Chronic neural
recording using silicon-substrate microelectrode arrays implanted in cerebral cortex,” IEEE Trans.
Biomed. Eng., vol. 51, no. 6, pp. 896–904, 2004.
[31] A. Jackson and E. E. Fetz, “Compact movable microwire array for long-term chronic unit recording
in cerebral cortex of primates.,” J. Neurophysiol., vol. 98, no. 5, pp. 3109–3118, 2007.
[32] C. Hassler, T. Boretius, and T. Stieglitz, “Polymers for neural implants,” Journal of Polymer Science,
Part B: Polymer Physics, vol. 49, no. 1. pp. 18–33, 2011.
[33] B. a Wester, R. H. Lee, and M. C. LaPlaca, “Development and characterization of in vivo flexible
electrodes compatible with large tissue displacements,” J. Neural Eng., vol. 6, no. 2, p. 24002, 2009.
[34] P. J. Rousche, D. S. Pellinen, D. P. Pivin, J. C. Williams, R. J. Vetter, and D. R. Kipke, “Flexible
polyimide-based intracortical electrode arrays with bioactive capability,” IEEE Trans. Biomed. Eng.,
vol. 48, no. 3, pp. 361–370, 2001.
[35] L. Luan et al., “Ultraflexible nanoelectronic probes form reliable , glial scar – free neural integration,”
no. February, pp. 1–10, 2017.
[36] S. P. Lacour, G. Courtine, and J. Guck, “Materials and technologies for soft implantable
neuroprostheses,” Nat. Rev. Mater., vol. 1, no. 10, 2016.
13
Chapter 2. Development of the Triple-Region Microwire Electrodes Array
Development of the Triple-Region Microwire
Electrodes Array
Microwires have a long history of being used as implantable electrodes to record neural
activities chronically from the brain [1]. Neuronal activity causes a transmembrane current and the
summation of potentials from multiple neurons spreading through the extracellular medium can
be measured as voltage, refers to as local field potential (LFP). The largest-amplitude currents that
occur across the somatic membrane is called action potentials or spikes can also be measured as
voltage. The magnitude of LFP and spikes greatly depends on the size and shape of the neuron and
its distance between recording electrodes [2]. Insolated, fine microwires with exposed tips inserted
to brain regions of interests and placed close enough to individual neurons can sense these
membrane currents and read electrical signals out from surrounding neural tissues. Over the past
several decades, the method of using microwires for record neural activities has not fundamentally
changed. It remains one of the most standard electrodes used in today, giving long-lasting, unitary
recordings from neurons for more than a year in particular cases [3][4]. To obtain access to
populations of individual neurons in both fixed and behaving animals, microwire electrodes arrays
formed by groups of fine microwires, organized in different geometrical configurations were
developed [5]. With growing demands of simultaneous recordings from large samples of neurons,
14
distributed across multiple cortical and sub-cortical brain regions in fully awake and behaving
animals, microwire electrodes arrays with both 2-D and 3-D configurations were fabricated. This
chapter will focus on the design of a 3-D, microwire electrodes array which targets all three sub-
regions of the rat hippocampus simultaneously. Multi-regional recordings achieved with the triple-
region microwire electrodes array will also be shown in this chapter.
2.1 Design of the Triple-Region Microwire Electrodes Array
Since the extracellular spike amplitude decrease rapidly as the distance between the neuron
and the recording electrode increase [6], the key to obtain high-amplitude, high-SNR, unitary
signals from a neuron is to precisely place the recording electrode as close to the target as possible.
To acquire simultaneously recording from the curved cell body layers of multiple sub-regions of
the rat hippocampus, the layout of recording electrodes needs to capture the curvature of the
hippocampus and the variation of spacing between cell body layers at different locations of the
hippocampus. Based on the anatomical distribution of hippocampal neurons, groups of microwires
with different length were arranged together. A rat brain atlas, histological results together with
neural activities recorded in vivo were referred to for the design of the triple-region microwire
electrodes array.
Figure 6. Measurement of the spacing between the CA1 and the CA3 sub-regions at different anterior
to posterior locations of the rat hippocampus from brain atlas.
AP= -2.85
AP= -3.25
AP= -3.70
AP= -3.90
AP= -4.20
AP= -2.45
Superimposed CA1 and CA3
5 1 2 5 2 0 1 5 1 0 3 0 3 5 4 0 4 5 5 0 5 5 6 0 6 5 7 0 7 3
ec
fx
opt
mtt
cpd fi
al v
cc
df
al v
alv
alv
amc
cing
LHAp
VM
PF
MDl
PVp
V 3 h
V 3h
BLAp
LA
E Pd EPv
sm
MH
LH CL
PVT
V3t
R T
V L
V L m
CA3
C A2
CA1
IG
4
3
2
1
5
R S P v
R S P d
SS p
rf
PIR
1
2
3
6a
b
DGmb
DGlb
C O Apl
C O Apm
VIP
FC
chpl
CM
BMAp
VPM
VPMpc
VPL
VPLpc
hf
IMD
LP
PO
PH
S TN
PA
E NT l
LGd
DGlb
so
sr
mo
sg
po
sg
mo
slm
slu
PMv
PMd
T U te
L H A vm
vlt
L H A v l
P S T N
S P F m
NL AL
LHAsfpm
IL
CA3so
sp
s t
S UB v
1
2
3
1 3 4
5
6
P E R I
im
f r
b sc
T R
P T Lp
E C T
T E a
AUDv
AUDp
AUDd
sps
spd
chf3
TMd
em
ZI
LGvl
sr
2
b a
sp
sg
3
P AA
3
I
I
30 Degree
1.30 mm
15
2.1.1 Brain atlas measurements
Bregma, the anatomical point on the rat skull at which the coronal suture is intersected
perpendicularly by the sagittal suture was used as the reference point. Depth differences between
the CA1 and the CA3 cell body layers of rat hippocampus at locations between 2.45 - 4.00 mm
posterior to the Bregma were measured from a commonly used rat brain atlas [7]. As illustrated in
Figure 3 and listed in Table 1. The average depth difference between the CA1 and the CA3 cell
body layer is 1.16 ± 0.05 mm. In addition, with a 30 degree implantation angle between the mid-
line and the electrodes array, the DG can also be reached.
Table I. Depth differences measured from a brain atlas
A-P (mm) 2.45 2.85 3.25 3.75 3.90
M-L (mm) 2.40 2.65 3.00 3.25 3.40
Depth Difference 1.08 1.15 1.20 1.15 1.20
2.1.2 Histological measurement of the depth of cell body layers
a) Methods
Both the Institutional Animal Care and Use Committee (IACUC) and the Department of
Animal Resources of the University of Southern California (DAR, USC) reviewed and approved
all animal experiments and surgical procedures. Three male Sprague-Dawley (SD) rats greater
than 3 month and weighted between 300 to 450 g were used for the measurement of depth
difference between the CA1 and CA3 cell body layers from histological slides.
16
The animal was pre-anesthetized in an induction chamber with Isoflurane mixed with
oxygen gas. Then the animal was deeply anesthetized with an intraperitoneal (IP) injection of a
mixture of ketamine and xylazine. Once the animal reached a surgical plane of anesthesia, an
incision was made to the abdominal wall right beneath the rib cage. After the diaphragm was
exposed, an incision was made onto the diaphragm and a cut through the rib up to the collarbone
was made on both side of the rib cage. The rib cage was then placed over the head of the animal
with hemostat to provide a clear view of the heart and major blood vessels. Small incisions were
made to both the left ventricle and right atrium. A blunt-tipped needle was then inserted into the
ascending aorta through the cut on the left ventricle and the animal was perfused with 4 °C saline
followed by 4 °C, 10% formalin. After perfusion, the brain of the animal was carefully dissected
and further fixed with 10% formalin at 4 °C overnight. Fixed brain tissue was then dehydrated
with 18% sucrose at 4 °C overnight. After the dehydrated brain tissue been embedded into 3%
agarose block, 50 µm coronal slices was made with a vibratome (Lecia, VT 1200) and stained with
crystal violent to make all cells visible. The distance between the CA1 and the CA3 cell body
layers was measured from brain slices between 2.45 to 4.20 mm posterior to the Bregma.
Figure 7. Graphic illustration of operation steps for whole body perfusion on rat.
17
b) Measurement of brain slices
The distances between the CA1 and the CA3 cell body layers in different regions along the
anterior-posterior (AP) axis from brain slices were showed in Figure 5. The depth difference
between the CA1 and the CA3 ranged from 0.90 mm to 1.30 mm. On average, the CA3 layer is
1.16 ± 0.18 mm deeper than the CA1 layer.
Figure 8. Representative crystal violet stained coronal brain sections obtained from three different
anterior to posterior locations of one rat hippocampus together with the plot of average depth differences
measured from three animals at various coordinate.
18
2.1.3 Relative depths of multiple sub-regions of the hippocampus calculated
from neural recordings
Complex spikes, a group of action potentials fired with less than 5 ms inter-spike intervals,
usually with decreasing amplitude, are characteristic firing patterns of hippocampal pyramidal
cells [8]. In the CA1 region, the generation of complex spikes is driven by regenerative dendritic
plateau potentials, produced by correlated entorhinal cortical and CA3 inputs that simultaneously
depolarize distal and proximal dendritic domains [9]. These high-frequency trains of action
potentials are believed to represent an information-rich communication signal enabling reliable
neurotransmission and increasing coding capacity [10][11]. Microwire electrodes arrays with
uniform wire length were implanted to either the CA1 or the CA3 sub-region of the rat
hippocampus to determine the spacing between these two cell body layers. Neural activities were
monitored during the insertion of all electrode arrays to provide feedback of locations of microwire
arrays. While the recording electrode located in the cortical region, commonly, single spikes were
detected. As the electrode advanced into the hippocampal region, complex spikes were observed.
Figure 9. Spike activities recorded during the implantation of microwire electrode arrays. a) shows single
spikes recorded from cortical regions while b) shows an example of complex spikes recorded from the
CA3 region. Compare to single spikes, complex spikes show much shorter inter-spike-intervals (less
than 5 ms).
0.52 0.56 0.61
−600
0
400
Time sec
Amplitude uV
0.8 0.85 0.89
−600
0
400
Time sec
Amplitude uV
a) b)
19
The detection of complex spikes was used as an indicator that the recording electrode was
advanced into the hippocampal region.
a) Methods
Male SD rats greater than three months and weighted between 300 g to 450 g were used
for implantations of uni-length microwire electrodes arrays. Each animal was pre-anesthetized in
an induction chamber with mixed Isoflurane and oxygen. Then the animal was deeply anesthetized
with an IP injection of a mixture of ketamine and xylazine. The appropriate level of anesthesia was
maintained with an inhalation of isoflurane mixed with oxygen (0.5-1.5 L/min) and constantly
checked with toe-pinch reflex test during the surgery. The animal’s body temperature was
measured with a rectal probe thermometer and maintained at appropriate level during the surgery
Figure 10. Illustration of surgical setups and diagrams of layouts of uni-length microwire electrodes
arrays. a) shows the stereotactic, surgical frame used in the experiment. Together with a schematic
representation of brain operation procedure. Data acquisition setup during the implantation is showed
on top right. b) a representative sketch of a 2x8, uni-length microwire array and a pic of the 2x8
microwire electrodes array.
Electrode
Rat brain and
Hippocampus
Microwire Electrode Array
Uni-length Microwire array
a)
b)
A
Input
Ref.
10000x
300-3KHz
20
with a heated warm pad. First, anesthetized animal was fixed onto a stereotactic, surgical frame
with two ear bars. An incision on the skin along the mid-line of the rat skull was made. After the
skull of the rat was made flat between the bregma and lambda, three small holes were drilled for
anchor screws.
One of the anchor screws was contacting cerebrospinal fluid to provide global ground for
the recording electrode. In addition to the holes for ground screws, an extra small hole was drilled
above the cerebellum for the insertion of the ground wire. A 2.5 × 4.0 mm cranium window was
made with a dental drill and the underneath dura and pia (blood vessel layer) layer above the right
hippocampus was carefully removed with tweezers and cotton tips to allow the insertion of the
microwire electrodes array.
Two types of uni-length, microwire electrodes arrays were used in this experiment. All
microwire arrays (MicroProbes for Life Science, Gaithersburg, MD) were consisted of 16,
stainless steel microwires which was insulated with Teflon and was 25 µm in diameter. In addition
to recording channels, each microwire array also had a reference channel which was 250 µm away
from one of the recording microwire. The reference was made of the same material as the recording
electrodes only with a bigger exposed electrode tip so that the reference electrode can only pick
up local field potentials. The local field potential can be used as reference signals for differential
amplifiers to help a better identification of unitary activities. Each microwire array also had a
ground wire which was a thin, stainless steel microwire with no insulation layer to provide global
ground for the recording. One type of microwire array was composed with two groups of eight
microwires with the same length. The separation between groups was 250 µm and the spacing
between wires was also 250 µm. The other type of microwire array had three microwire groups
with six wires in the middle group and five wires in other two groups. The spacing between each
group was 250 µm and the separation between each wire was 200 µm.
21
After the brain surface was exposed, a microwire array was advanced into the brain tissue
at a speed of 10 µm/sec to the cortex. During the insertion, neural activities recorded with the
microwire array were bandpass filtered from 300 Hz to 10, 000 Hz and differentially amplified for
10000 times with a single channel differential amplifier (DAM50 Extracellular Amplifier, WPI,
Sarasota, FL). Due to the limitation of the amplifier, only one recording channel was monitored at
a time. A customized switchboard was used to allow the selection of any one of those sixteen
recording channels. All sixteen recording channels were frequently checked throughout the
insertion. As the microwire arrays approached the hippocampal region, the advancement of the
microwire was paused every 5 to 10 µm to check neural activities on all sixteen channels at each
depth. The observation of complex spikes at the depth between 2.2 - 3.2 mm ventral to the bregma
was considered as recordings from the CA1 cell body layer while the detection of complex spikes
in deeper regions (3.6 - 4.4 mm from the bregma) was considered as recordings from the CA3 cell
body layer. The depth at which complex spikes were detected were recorded. After the microwire
array reached desired regions, a thin layer of dental cement (Stoelting Inc., Wood Dale, IL) was
Figure 11. Illustration of different spike types encountered at different brain region. Single spikes from
cortical regions is color coded with green, complex spike from the CA1 region is colored as red and
complex spikes from the CA3 region is represented in dark blue. While the implantation of a microwire
array, single spikes was first recorded. As the array approached the CA1 region, complex spikes were
recorded.
200 uV
20 msec
Single spike from cortical region
Complex spikes from the CA1
Complex spikes from the CA3
22
applied around the microwire array and over the cranium windows to hold the electrode in place.
The ground wire was twisted onto the ground screw and the rest of the ground wire was inserted
into the small anchor hole drilled above the cerebellum. The remaining part of the microwire array
were then lowered down to reduce the overall exposed connector height above the animal’s head.
More dental cement was then applied to cover the microwire and lower half of the connector with
only the top part of the connector been exposed for future connection with the data acquisition
system. After recording, the animal was deeply anesthetized with an IP injection of a ketamine and
xylazine mixture. Direct current (DC) of 300 µA was applied to half of the recording channels for
5 secs to make electrical lesions at the tip of the microwire. Then the animal was perfused with
4 °C saline, 4 °C 10% formalin and 4 °C 3% KFe (potassium ferrocyanide in 10% formalin)
followed perfusion procedures described in section 2.1.2. 50 µm thick brain slices were prepared
and stained with crystal violent to verify tip locations.
23
b) Measurement results
Twenty-three animals were implanted with the uni-length microwire array for the
measurement of depth difference between the CA1 and the CA3 region. Sixteen animals were
implanted with uni-length microwire arrays (both 2x8 arrays and arrays with three microwire
groups) to the CA3 region while seven animals were implanted with uni-length microwire arrays
to the CA1 region. Tip locations were verified with histological results. Seven implantations were
either too shallow or too deep for the CA3 region and three implantations were not located in the
CA1 cell body layer. Those animals were excluded from the analyze. The implant coordinates of
the rest nine implantations to the CA3 region were listed in Table II. During the insertion, the depth
at which complex spikes from the CA1 region was observed was also recorded and listed in Table
II. Calculated from the depth at which complex spikes were recorded, the spacing between the CA1
and the CA3 region ranged from 1.00 mm to 1.30 mm. On average the CA3 cell body layer was
1.13 ± 0.12 mm deeper than the CA1 cell body layer.
Figure 12. Depth difference between the CA1 and CA3 sub-region estimated from neural activities. a)
shows representative complex spikes recorded from the CA1 and CA3 region. b) shows depths at which
complex spikes from the CA1 and the CA3 were recorded while the implantation of nine uni-length
microwire arrays.
1 2 3 4 5 6 7 8 9
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Animal Number
Depth Difference (mm)
CA3
CA1
200 µV
20 ms
CA1
CA3
a)
b)
24
Table II. Measure of depth difference based on neural activities
A-P (mm) 2.45 2.85 2.60 2.60 2.45 2.45 2.45 2.45 2.47
M-L (mm) 2.60 2.60 2.75 2.75 2.44 2.40 2.40 2.40 2.40
CA1 Depth 3.10 3.10 3.05 3.20 3.05 3.30 2.95 3.00 2.90
CA3 Depth 4.15 4.35 4.20 4.20 4.35 4.35 4.05 4.25 4.08
Depth
Difference
1.05 1.25 1.15 1.00 1.30 1.05 1.10 1.25 1.18
Implant coordinates of four, CA1 implantations were listed in Table III. The final
microwire array depth of those nine, CA3 implantations and four, CA1 implantations were
compared. As listed in Table II, the average depth of the CA3 region is 4.23 ± 0.12 mm and the
average depth of the CA1 region is 2.56 ± 0.18 mm.
Table III. Final implantation depth for CA1
A-P (mm) 2.72 2.72 2.45 2.70
M-L (mm) 2.85 2.85 2.50 2.85
CA1 Depth 2.43 2.51 2.80 2.60
2.1.4 The triple-region microwire electrodes array
According to the measurement of depth difference from brain atlas and brain slices together
with depth difference calculated from implantations of uni-length microwire electrodes arrays, the
layout of the triple-region microwire electrodes array was designed. The triple-region microwire
array was composed by sixteen, 25 µm Teflon coated stainless steel wires arranged into three
groups differed in length. With two longer groups of wires targeted at the CA3 and the DG
respectively while the shorter group targeted at the CA1 cell body layer. Both recording groups
for the CA3 and the DG region were 1300 µm longer than the group that targeted at the CA1 region.
25
Two longer groups each had five recording channels and was in the same sagittal plane with 1000
µm spacing in between. The shorter group had six channels and located between two longer groups
on another sagittal plane 250 µm lateral to the longer ones. The separation between each wire is
200 µm. A reference electrode was also placed to the CA1 recording group. Another un-insulated
microwire coming directly off the connector was used as ground wire.
2.2 in vivo Evaluation of the Triple-Region Microwire Electrodes
Array
To verify whether the 3D, triple-region microwire electrodes arrays can capture the
curvature of the rat hippocampus and recording from the cell body layer of multiple hippocampal
sub-regions, the triple-region electrodes arrays were implanted to rats following the same
procedure described in section 2.1.3, methods. Again, neural activities were recorded both during
the implantation surgery and after the animal fully recovered from the surgery.
Figure 13. Layout of the triple-region microwire electrodes array. The top left shows a photo of the
triple-region microwire array and the relative size of the array. Diagrams on the left and bottom show
the dimension of the microwire array.
CA3
CA1
DG
Hippocampus CA1
DG
CA3
B
D
D
C
A= 200 µm
B= 1300 µm
C= 1000 µm
D= 250 µm
End view
C
D
A
B
26
2.2.1 Recording with the triple-region microwire electrodes array
Seven triple-region microwire electrodes arrays were implanted to six animals. For five of
those animals, one triple-region electrodes array was implanted to the right hippocampus. For one
animal, two triple-region electrodes arrays were implanted to both hippocampi. During the
implantation, all seven implanted triple-region electrodes arrays recorded from at least two sub-
regions of the hippocampus. However, only one electrodes array was able to record unitary
activities from all three sub-regions of the hippocampus. Since in-surgery data was collected with
a single-channel amplifier and an oscilloscope, spike sorting was not conducted to neural activities
recorded and the number of units was not counted. Instead, the number of wires that recorded
spikes during the insertion was counted. On average, 5.43±1.72 out of 16 channels recorded spikes
during the implantation. For each sub-region, 2.29±1.80 wires reached the CA1 layer, 1.71±1.25
wires recorded signals from the CA3 layer and 1.43±1.40 wires recorded spikes from the DG
region.
After the animal fully recovered from the surgery, chronic data was recorded from all seven
implants while the animal running freely in an open field. Two electrodes arrays among those
Figure 14. Shows an example of in vivo recordings collected with a triple-region microwire array.
Unitary activities were recorded from all three sub-regions and tip locations of the triple-region array
were verified with histological slices.
c)
600 µV
10 ms
300 µV
10 ms
300 µV
10 ms
100 µV
200 µs
250 µm
27
seven triple-region microwire electrodes arrays recorded from multiple sub-regions of the rat
hippocampus. Figure 14. Shows spike activities recorded from all three sub-regions of the
hippocampus simultaneously and the region from where those signals were recorded verified
through histological stain. However post-implantation histology results showed that only limited
number of microwires (1 - 2) were located close enough to desired cell body layers.
2.2.2 Limitation of the triple-region microwire electrodes array
Simultaneous recordings from the CA1, the CA3 and the DG verified that the concept of
using microwire arrays with multiple lengths to target brain regions at different depth is valid.
However, in vivo recording results mentioned above implies more of limitation than feasibility.
The first limitation is caused by the nature of microwires, with only one recording site per wire,
relatively less number of recording channels can be implanted with the same amount of tissue
displacement. Compare to silicon probes [12] and polymer-based neural probes [13] orders of
magnitude more recording sites can be placed along probe shanks without greatly increase the
amount of brain tissue been damaged. Secondly, the fabrication of multi-length microwire
electrodes arrays still rely on the manufact of fine assembly jigs to hold and align individual
microwires [14]. The accuracy of wire arrangement and repeatability might not fulfill the demands
of precisely targeting at multiple thin (less a 100 µm), cell body layer of brain structures such as
the hippocampus. In addition, the curvature of the hippocampal cell body layers posts more
difficulties on the fabrication of conformal microwire electrodes arrays.
2.3 Summary
In summary, driven by the growing and primary demand of simultaneous recording from
populations of single neurons distributed in the hippocampal circuit, we modified the layout of
conventional 2-D microwire electrodes array which is commonly composited with microwires
with uniformed length and designed a 3-D, triple-region microwire electrodes array according to
the main anatomical characteristics and contours of the rat hippocampus. With groups of
microwires with specifically adjusted wire length, hippocampal cell body layers located at
different vertical depth were simultaneously targeted. Acute and chronic recordings together with
histological results verified the possibility of using this triple-region microwire electrodes array to
collect neural activities from multiple sub-regions of the hippocampus. However, the limited
28
number of microwires that reached individual sub-region and relatively low units yield shown by
those in vivo results imply that although the concept is durable but there are considerable
difficulties to practically design, fabricate and apply conformal, multi-regional, microwire
electrode arrays. The potential of using other advanced and progressing nano-fabrication technique
in neural probe development should be explored and evaluated.
2.4 References
[1] F. Strumwasser, “Long-term recording from single neurons in brain of unrestrained mammals,”
Science (80-. )., 1958.
[2] G. Buzsáki, C. A. Anastassiou, and C. Koch, “The origin of extracellular fields and currents-EEG,
ECoG, LFP and spikes,” Nature Reviews Neuroscience. 2012.
[3] M. A. L. Nicolelis et al., “Chronic, multisite, multielectrode recordings in macaque monkeys,” 2003.
[4] S. Suner, M. R. Fellows, C. Vargas-Irwin, K. Nakata, and J. P. Donoghue, “Reliability of signals
from a chronically implanted, silicon-gased electrode array in non-human primate primary motor
cortex,” IEEE Trans Neural Syst Rehabil Eng, vol. 13, no. 4, pp. 524–541, 2005.
[5] J. C. Williams, R. L. Rennaker, and D. R. Kipke, “Long-term neural recording characteristics of
wire microelectrode arrays implanted in cerebral cortex,” pp. 303–313, 1999.
[6] D. a Henze, Z. Borhegyi, J. Csicsvari, a Mamiya, K. D. Harris, and G. Buzsáki, “Intracellular
features predicted by extracellular recordings in the hippocampus in vivo.,” J. Neurophysiol., vol.
84, no. 1, pp. 390–400, 2000.
[7] L. W. Swanson, “Brain maps 4.0—Structure of the rat brain: An open access atlas with global
nervous system nomenclature ontology and flatmaps,” J. Comp. Neurol., 2018.
[8] J. O’Keefe, “Place units in the hippocampus of the freely moving rat,” Exp. Neurol., vol. 51, no. 1,
pp. 78–109, 1976.
[9] S. R. Balind et al., “Diverse synaptic and dendritic mechanisms of complex spike burst generation
in hippocampal CA3 pyramidal cells,” no. 2019.
[10] J. E. Lisman, “Bursts as a unit of neural information: Making unreliable synapses reliable,” Trends
in Neurosciences. 1997.
[11] F. Zeldenrust, W. J. Wadman, and B. Englitz, “Neural coding with bursts—Current state and future
perspectives,” Frontiers in Computational Neuroscience. 2018.
[12] A. Berényi et al., “Large-scale, high-density (up to 512 channels) recording of local circuits in
behaving animals,” J. Neurophysiol., vol. 111, no. 5, pp. 1132–1149, Mar. 2014.
29
[13] J. E. Chung et al., “High-Density, Long-Lasting, and Multi-region Electrophysiological Recordings
Using Polymer Electrode Arrays,” Neuron, vol. 101, no. 1, p. 21–31.e5, 2019.
[14] G. Lehew and M. A. L. Nicolelis, “State-of-the-art microwire array design for chronic neural
recordings in behaving animals,” in Methods for Neural Ensemble Recordings, Second Edition,
2007.
30
Chapter 3. Development of the Parylene-Based Multi-Electrode Array
Development of the Parylene-Based Multi-
Electrode Array
Traditionally, lifetimes of neural recordings implants are greatly limited by the body’s
nature immune response against any foreign implant which causes neuronal death and glial
scarring. Acute response to any implanted electrodes including the local activation of microglia
and the migration of microglia toward implant sites [1][2]. Within approximately six weeks, the
implant is encapsulated with a thin but dense layer of astrocytes [3]. These inflammatory responses
and the formation of gliosis around neural implants are suggested to be responsible for poor signal
quality and long-term recording failure [4]. There are many factors that might affect the degree of
immune responses including the size, shape and the material of the neural implant. More and more
evidences show that the physical and mechanical mismatch between the neural implant and the
brain tissue is one critical cause for long-term recording failure [5]. The immune reaction is posited
to be exacerbated by micromotion between the rigid implant and soft surrounding brain tissue and
attenuates the quality of recordings over time [6].
To narrow the gap between neural tissues which are several orders of magnitude softer than
metal or silicon neural implants, the application of ‘soft’ material with low Young’s moduli and
low hardness in the neural implant fabrication developed rapidly in the past few years. Parylene C
is a micro-machinable and USP class VI biocompatible polymer commonly used in the biomedical
31
field for chronic implants. The Parylene was first used as coating for implantable electronics [7][8]
since its chemically inert property and high dielectric strength. In addition, the Parylene can be
deposited uniformly and conformal to the substrate and form an optimal encapsulation barrier.
These characteristics made the Parylene one of the most popular material in the Mciro-Electro-
Mechanical System (MEMS) field. Fabrication of thin films of Parylene is relatively easy compare
to other flexible polymers such as polyimide and polydimethylsiloxane (PDMS). The Parylene can
be deposited under room temperature and the deposition process of Parylene is compatible with
conventional micromachining and photolithographic process [9]. In addition, the Parylene thin
film is high in flexibility mean while strong in mechanical strength. Because these advantages, the
application of Parylene rapidly expended to the field of MEMS neural probes. Thin film electrode
with high-density recording sites was fabricated with the Parylene as both substrate and insulation
layer for both recording from and deliver stimulation to the retina [10]. Recently, the application
of the Parylene is further expanded. Penetrating neural probes using the Parylene as structural
material were fabricated and implanted to obtain high resolution electromyogram [11] and unitary
activities from cortex [12]. Results from these advanced applications of the Parylene indicates a
promising prospect of the application of using Parylene as the structural material for neural probes
targeting at multiple deep-brain regions simultaneously.
In this work, we further explored the possibility of using the flexible, Parylene C as the
structural material for long-term, sub-cortical recordings from multiple sub-regions of the
hippocampus of behaving arts. A Parylene multi-electrode array which targets at the CA1 and the
CA3 cell body layers of the rat hippocampus simultaneously was designed and fabricated. The
corresponding insertion technique for the flexible multi-electrode array which limited insertion
damage to surrounding tissue to the cross-section of the array itself was also developed.
3.1 Design of the Parylene Multi-Electrode Array
With experiences and information accumulated from the development of the triple-region,
microwire electrodes array, the layout of the Parylene multi-electrode array was designed. We
decided to have eight shanks in each Parylene array with 250 µm spacing between shanks to span
2,000 µm region of the dorsal hippocampus along the septal-temporal axis. The array was
constructed from layered thin films of Parylene C (10-18 µm thick) and Pt (2000 Å thick). Parylene
provides both structural support and electrical insulation. To targeting at both the CA1 and the
32
CA3 cell body layers, the sandwiched metal layer in each probe contained two groups of Pt
electrodes and their traces. Pt was chosen to serve as the electrode material due to its
biocompatibility, high electrical conductivity, and inertness in the body. Since the deepest
hippocampal layer of interest sits at a depth of 4.0 - 4.5 mm from the bregma, a probe length of
5.5 mm was used to reach all hippocampal regions of interest.
As shown in previous chapter, principal neurons in the hippocampus pack into cell body
layers that is tens of microns thin and curves into a double-C structure and is functionally divided
into the CA1, CA3 and DG sub-regions. In order for penetrating neural probes to record from each
thin cell body layer of the hippocampus simultaneously, electrode sites must be accurately
patterned along the length of each probe with some level of redundancy to overcome anatomical
variations between animals. Therefore, in the design, each sub-region was targeted by a group of
four linear electrodes spaced by 40 µm. This ensured that at least one electrode site resides in each
targeted cell body layer. The appropriate positions of electrode recording sites along the probe
shanks were determined through a combination of rat brain atlas and histological measurements
alongside in vivo recordings collected with both uni-length microwire electrodes arrays and the
triple-region electrodes arrays described in the previous chapter. The exposed electrode area was
30 µm in diameter (electrode sites were patterned 50 µm in diameter and the edges were
surrounded by insulation). Electrode diameters were chosen to improve their selectivity to
individual neurons and limit their impact on the overall width of the shank, while balancing the
need for sufficient surface area for reducing electrode impedance (<1 MΩ at a frequency of 1 kHz)
and noise.
The electrodes were connected to contact pads by 5 µm wide traces, with 5 µm spacing.
This small trace width was chosen to decrease the overall width of each probe and thereby reduce
the foreign body response to the implant. These design criteria enabled the realization of a probe
whose width spans from 110 µm at its most tapered part, to 150 µm at its widest part. The shape
and critical dimensions of the array and individual probes are presented in Figure 15.
33
3.2 Fabrication and Electrical Packaging
Typically, a polymer-based probe is composed by a thin conductive metal layer
sandwiched between two insulating and supporting polymer layers. The thickness of the metal
layer is usually on the order of hundreds of nm while the thickness of individual insulation layer
is around 10 µm. Each layer of material can possess intrinsic stress, resulted from the deposition
process, and thermal stress, a result of mismatch in thermal expansion of different materials. Both
types of stress can result in undesired curling which can cause extra difficulties in straight insertion
and precise placement of the flexible probe. Since both intrinsic stress and thermal stress closely
relate to the temperature metal and polymer materials exposed to during the fabrication [13],
especially metal deposit procedure, different metal deposit methods were tested. In an attempt to
straighten the Parylene-based multi-electrode array, the thickness of the top insulation layer was
also modified to balance stresses apposed on the probe. In addition, all Parylene arrays were
annealed post-fabrication with heat and pressure. Thermal annealing is proved to increase Parylene
chain entanglement which can enhance the bounding strength between Parylene-Parylene interface
Figure 15. Diagrammatic representation of the detailed layout of the Parylene-based mlti-electrode
array. Picture on the left shows the overall view of the Parylene multi-electrode array. The arrangement
of recording sites on each Parylene shank was adjusted to match the curvature of the hippocampal cell
body layers. Picture on the right shows layout and dimension of an individual Parylene shank.
1200 µm
anterior
around
5.5 mm
CA3
CA1
DG
0.07 mm 0.05 mm
250 µm
150 µm
110 µm
30 µm
40 µm
side view
10 µm
10 µm
34
[14]. This step can also reduce stress mismatch between different layers so that assist in
straightening the devices.
3.2.1 Optimization of fabrication procedure
The fabrication of a Parylene-based multi-electrode array is accomplished following
several steps. Figure 16. briefly illustrated the fabrication procedure of a single-metal-layer
Parylene array. First, a base layer of Parylene is vapor-deposited on the silicon wafer followed by
the deposition of a uniform layer of photoresist. The outline of electrode recording sites, traces and
contact pads are lithographically patterned on the photoresist. In general, the edge of photoresist
is processed to have a negative profile so that allowing the sidewall of the photoresist to escape
metal encapsulation. This feature can ease the future lift-off of undesired metal parts. After the
surface that would be coated with Pt is exposed and prepared with short O2 treatment, a thin layer
of Pt is deposited. The metal layer is then encapsulated by another layer of Parylene. Electrodes
and contact pad are subsequently exposed by deep reactive iron etching (DRIE) with an O2 plasma
Figure 16. Illustration of fabrication procedures of the Parylene multi-electrode array. a) shows steps
that involved in the fabrication of the Parylene array. b) shows the negative profile of photoresist. This
negative profile can prevent the encapsulation of side walls of photoresist by thin metal layers and ease
the lift-off of undesired metal.
1. Deposit parylene 2. Spin photoresist (PR) 3. Expose PR
4. Evaporate metal 5. Strip PR 6. Deposit parylene
7. Spin PR 8. Expose PR
9. Etch parylene
10. Strip PR
11. Peel release in
water bath
Silicon wafer
Parylene
Metal
Photoresist
a) b)
Negative profile of the PR
Metal layer deposited on PR
with negative profile
35
and the probes is cut out from the substrate. The device is released from the silicon carrier wafer
in a water bath.
a) Metal deposition methods
Electron beam (E-beam) evaporation and sputtering deposition are two commonly used
vacuum techniques in microfabrication, nanofabrication and semiconductor industries to deposit
thin films of various conductive materials to the surface of objects. Both techniques were used to
deposit the thin Pt layer containing electrodes and connective traces in this work. E-beam
deposition method utilizes an intense, electron beam which is generated from a filament and
steered via electric and magnetic fields to strike source metal and vaporize it within a vacuum
environment. As the metal is heated, its surface atoms will have sufficient energy to escape and
traverse the vacuum chamber to the substrate [15]. The sputtering method of thin film deposition
also requires a vacuum environment. A controlled gas is introduced into a vacuum chamber, and
electrically energizing a cathode to establish a self-sustaining plasma. The exposed surface of the
cathode is a slab of the metal material to be deposited onto the substrates. The gas atoms lose
electrons inside the plasma to become positively charged ions, which are then accelerated into the
target and strike with sufficient kinetic energy to dislodge atoms or molecules of the target metal.
This sputtered metal now constitutes a vapor stream, which attracted by the anode, the surface
need to be coated, and hits the substrate [16].
Compare to e-beam deposition method, all sputtering procedures are conducted under a
relatively low-temperature environment. The deposition of the 2000 Å thick layer of Pt onto the
Figure 17. Schematic representations of different metal deposition method. a) shows the illustration of
typical e-beam deposition method while b) shows the diagram of sputtering deposition method.
a) b)
36
substrate with the sputtering method was tested first. However, since sputtered Pt move diffusively
in the vacuum chamber, reach all surfaces within the vacuum chamber and condense after
undergoing a random walk. The thin Pt layer still encapsulated sidewalls of the photoresist in some
cases, even the sidewall was with negative profile. This phenomenon caused difficulty in the
following lift-off procedure. Although most undesired metal parts were lifted-off after 6 hours of
soaking the whole wafer in 50
o
C acetone followed by vigorous swabbing of the metal features.
Those steps resulted in unwanted damages to the probe’s structure. The mechanical swabbing step
was especially harmful to the fine metal features.
Because the potential lift-off difficulty faced with sputtering deposition method, e-beam
deposition technique was then experimented. No lift-off problems were experienced for all e-
beamed devices. However, cracks or fractures in the metal features were noticed in devices whose
Pt layer was e-beamed using the e-beam machine at USC. An abnormal high temperature of the
wafers was noticed during the e-beam procedure. This overheating was hypothesized to be the
cause of metal cracks and fractures. Concurrent to the overheating problem encountered with the
e-beam machine at USC, metal deposition was conducted at Caltech. Since the e-beam machine at
Caltech has a much longer throw distance (28 inches at Caltech, 14 inches at USC) which provide
longer cool down period for the metal, no overheating of the substrate was observed during the
metal deposition procedure and the 2000 A thick Pt layer was deposited without metal cracking.
Since the lift-off of metal deposited via e-beam method was significantly easier than sputtering
method, e-beam metal deposition method was used for all future devices.
b) Straighten the device
Curvature of the fully fabricated, Parylene-based multi-electrode array was noticed on both
devices fabricated with e-beam metal deposition method and sputtering metal deposition method.
Stress mismatch, particularly stress in the thin film Pt layer, was the major cause of the curvature
of the Parylene arrays [13]. In an attempt to balance the stress in each layer, the thickness of the
insulation Parylene layer was adjusted. Parylene arrays with 10, 14 or 18 µm thick insulation layers
were fabricated and the radius of array curvature were measured and listed in Table IV. Annealing
is the process aimed to enhance Parylene chain entanglement and reduce stresses mismatch
between different layers [17]. Released devices were sandwiched between two 0.33 mm thick
Teflon sheets which were then temporarily fixed between two glass slides with mini clips. Those
37
devices were then placed in an oven, vacuum purged three times with nitrogen gas and then
annealed, under vacuum for 48 hours at 200
o
C. Radius of curvatures of devices post-thermal
forming were list in Table V. Comparison of the radius of curvature between devices fabricated
with different parameter (insulation layer thickness, metal deposition method) and radius of
curvature before and post thermal forming, devices whose metal layer been e-beamed between two
10 µm thick Parylene layers had the smallest curvature.
Table IV. Effects of the thickness of insulation layer and metal deposition method on
Parylene array curvature
Insulation layer thickness Metal deposition method Radius of curvature
(mm)
Sham N/A -165.43
10 µm/10 µm Sputtering 11.04 ± 0.24
10 µm/14 µm Sputtering 102.47 ± 16.53
10 µm/18 µm Sputtering -42.96 ± 6.25
10 µm/10 µm E-beam 75.74 ± 60.46
Table V. Effects of thermal forming on Parylene array curvature
Insulation layer thickness Metal deposition method Radius of curvature
(mm)
Sham N/A -633.07 ± 281.67
10 µm/10 µm Sputtering 72.51 ± 10.06
10 µm/14 µm Sputtering -145.84 ± 36.97
10 µm/18 µm Sputtering -39.78 ± 7.83
10 µm/10 µm E-beam 568.81 ± 857.65
38
3.2.2 Final fabrication scheme
A base layer of 10 µm thick Parylene C (Specialty Coating Systems, Indianapolis, IN) was
deposited via chemical vapor deposition (CVD) onto a dehydrated, prime 4” silicon wafer.
AZ5214 (Integrated Micro Materials, Argyle, TX) image-reversal photoresist (step 1: 8 s, 500 rpm,
step 2: 45 s, 2,000 rpm) was patterned via photolithography to define the electrodes and leads. The
wafer was cleaned in O2 plasma, and 2000 Å Pt was deposited by either sputter coating (LGA Thin
Films, Santa Clara, CA) or electron beam evaporation (Caltech Kavli Nanoscience Institute,
Pasadena, CA). The Pt features were defined by lift-off in acetone (50 °C) accompanied by gentle
brushing. An insulation layer of Parylene, 10, 14, or 18 µm thick was subsequently deposited via
chemical vapor deposition. A 15 µm thick layer of AZ4620 photoresist (Integrated Micro Materials,
Argyle, TX) was spun (step 1: 5 s, 500 rpm, step 2: 45 s, 1,200 rpm) and patterned to produce an
etch mask, and the array shape was defined using a switched chemistry process in a DRIE tool that
alternated between fluoropolymer deposition (C4F8) and oxygen plasma etching [32]. This first
etch mask was removed in acetone, and a second mask of AZ4620 subsequently deposited 30 µm
thick (two spins separated by softbake; step 1: 8 s, 500 rpm and step 2: 45 s, 2,000 rpm). In a
second etch, the electrodes were exposed and the device fully defined by etching through the
remaining layer of Parylene. Finally, the resist mask was stripped, the wafer rinsed sequentially
with acetone, isopropanol and deionized water, and devices were released by gently peeling the
Figure 18. a) Photograph of fabricated array highlighting the neural interface (consisting of electrodes
patterned onto probe shanks), Parylene ribbon cable, and contact pads which provide contact to the
electrical packaging system. b) Close-up of electrodes and traces.
electrodes
ribbon
cable
probe
shanks
contact pads
5 mm 500 µm
(a) (b)
39
device away from the native oxide layer of the silicon wafer while immersed in water. Any
remaining photoresist was removed by soaking released devices for 5 minutes in baths of acetone,
isopropanol, and water. Cleaned devices were sandwiched between Teflon and thermally annealed
for 48 hours at 200 °C in a nitrogen purged oven under vacuum.
During the early stage of this project, all Parylene-based multi-electrode arrays with
smooth metal surface and minor shorted traces were considered as functional devices and were
used for testing and implantation. Both devices fabricated with e-beam metal deposition method
and sputtering metal deposition method were tested and implanted. The thickness of insulation
Parylene layers varied between 10, 14 and 18 µm. However, future Parylene arrays will all be
fabricated with the e-beam metal deposition method and the e-beam machine with long throw
distance will be used. To minimize the cross-section size of the Parylene array, the insulation layer
will be 10 µm and the curvature of the Parylene array will be treated with an annealing process
post-fabrication.
Figure 19. Shows the two PCBs used to connect the Parylene multi-electrode array with the data
acquisition system. The bottom picture on the left shows the first PCB. The ZIF connector on the bottom
connected with one Parylene array. The top picture on the left shows the second PCB which convert the
SSB6 connector to two Omnetic connectors. The photo on the right side shows one animal been
implanted with the Parylene electrode array. The first PCB was half-covered with dental cement.
32 Channel Omnetic Connector
PBC 1
PBC 2
ZIF
SSB6
1 cm
40
3.2.3 Electrical packaging
Individual shanks of the hippocampal array connect into a flat, flexible cable which
terminate in Pt contact pads mated to a 71 pin zero-insertion-force (ZIF) connector (Hirose Electric
Co., Japan). Parylene contact pads were supported with 0.002” thick polyether ether ketone (PEEK)
tape with a 2.3 mil thick acrylic adhesive (CS Hyde Co., Lake Villa, IL) which enabled insertion
of the Parylene array into the ZIF connector. The ZIF used in this work has a space-saving double
row design with a 200 µm pitch between contact pads and a total length of 1.58 cm. The electrical
packaging used during in vitro electrochemical testing consisted of a single PCB that served as an
adapter between the ZIF and two 0.1” spacing header boards, whose pins were easily connected to
the working electrode. Both cyclic voltammetry (CV) cleaning and electrode impedance
spectroscopy (EIS) were run on each of the 64 individual electrodes using this setup.
Two types of connection methods were used for in vivo testing. Due to the concern of the
size of part that been permanently mounted on an animal’s head, the first connection system used
a two-printed circuit board (PCB) setup. It was consisted of two, mated printed circuit boards
(PCB), as depicted in Figure 19. The first PCB was permanently mounted to the rat’s head with
dental cement and supported the ZIF connector and an SSB6 PCB-to-PCB receptacle connector
(Molex Incorporated, Lisle, IL). The dimension of the first PCB was minimized to prevent
disturbing of the animal’s normal behaviors. The horizontal length of the first PCB was 1.8 cm
and was only 2 mm wider than the ZIF connecter. The vertical height of the PCB was 1.7 cm.
Besides the room for the SSB6 and ZIF connectors, an additional 5 mm spacing was added for the
application of dental cement for chronical fixation. The second PCB was detachable and supported
a male SSB6 connector and two 32 position dual row nano-miniature Omnetics connectors
(Omnetics Connector Corporation, Minneapolis, MN). The Omnetics connectors were connected
during the surgery and during recording sessions to establish connection from the electrical
packaging to recording systems.
During test phases, we notice that the SSB6 connector would wear and tear with repeated
attachment and detachment. This leaded to unstable connections between the first and second
41
PCBs and introduced extra motion noises to the recording. In some cases, the SSB6 connector was
fully broken which made it impossible to get continuous recording from moving animals. To
address this issue, a single PCB was designed to replace the two-PCBs setup. Through adjust the
orientation of two Omnetic connectors, the single PCB setup was only less than 5 mm taller after
been chronically fixed onto an animal’s head.
3.3 Development of Array Insertion Techniques
Long, thin, flexible polymer shanks will mechanically buckle under very small loads,
deforming the array during insertion. Because of the flexibility of the Parylene probes, direct
insertion of a bare Parylene array will cause buckling of the array shanks. To avoid buckling and
to enable a successful implantation of a Parylene array to desired deep-brain region, the
hippocampus, a special insertion technique is needed to facilitate surgical insertion of the flexible
Parylene muli-electrode array.
3.3.1 Insertion shuttles
Insertion shuttles are stiff structures commonly used in the insertion and placement of soft,
flexible, polymer based neural probes [18]. Typically, a shuttle device is made of ridged materials
which can provide enough mechanical support for the soft probe to be inserted into brain tissue
[19]. Considering the risk that overhanging parts of a flexible probes might get caught in superficial
Figure 20. Schematic representations of the two connection methods used for interface the Parylene
array with data acquisition systems. The figure on the left shows a two-PCB design and the figure on
the right shows a single PCB design. The dashed red line represents the final height of PCBs that will
stand out of the dental cement cap which will been chronically mounted on the animal’s head.
Two-PCB Setup One-PCB Setup
Detachable
Headstages
Detachable
Headstages Omnetic
Connector
PCB1
PCB2
Dental Cement
42
tissue during insertions, the cross-sectional dimension of a shuttle device is usually the same or
bigger than the width of the soft probe [20]. This increasing in acute insertion trauma directly
conflicts with the requirement of minimizing the surgical damage made during implantation.
Minimize the surgical damage is considered as one key aspects for establishing a reliable and scar-
free neural-probe interface [21]. Another challenge in successfully implementing a shuttle device
lie in the integration between the shuttle and the flexible probe. Ideally, the attachment between
the shuttle and the probe need to be strong during the insertion so that decoupling of the two will
not happen before the probe been delivered to desired locations. Meanwhile, after the insertion is
completed, the probe needs to be released from the shuttle easily and the retraction of the shuttle
device should not displace the already in-place probe. However, displacements of flexible probes
post removing of shuttle devices are common and a displacement of 3% of the implantation
distance is considered as negligible [19]. To reach the CA3 region of the rat hippocampus which
is around 4.00 mm deep from the brain surface, this displacement is expected to be on the order of
one hundred µm which is almost comparable to the thickness of cell body layers of the
hippocampus. Based on limitations described above, the application of shuttle device is not an
ideal approach for the insertion of the Parylene array to desired, thin cell body layers of the rat
hippocampus with high accuracy.
3.3.2 Dissolvable brace for insertion support
The straight insertion of a single flexible probe into brain tissue requires the force that the
probe can handle without buckling exceeds the force necessary to penetrate brain tissue. The force
that induces buckling of a flexible probe can be estimated from eq. 1 by modeling an individual
probe as a mechanical beam of width w, thickness t, length L, and Young’s modulus E. This model
assumes one end to be clamped to the insertion tool and the other pinned in the x-y plane as soon
as it contacts brain tissue. For a 28 µm thick, 150 µm wide, and 5.5 mm in length single Parylene
shank (Young’s modulus of 3.2 GPa), we would expect it to certainly buckled at only 0.6 mN.
However, the insertion force required to penetrate brain tissue is commonly accepted to be 1 mN
[29].
𝐹𝐹 𝑏𝑏 𝑏𝑏 𝑏𝑏𝑏𝑏 𝑏𝑏 𝑏𝑏 𝑏𝑏 𝑏𝑏 =
𝜋𝜋 2
𝐸𝐸 𝐸𝐸 𝑡𝑡 3
5. 88 𝐿𝐿 2
(eq. 1)
43
This obstacle has traditionally restricted the application of flexible probes to superficial, cortical
targets that are < 3 mm deep, as buckling force threshold inversely relates to the effective probes
length and shorter probes may be inserted without buckling.
Based on Euler’s critical load equation (eq. 1), by halving the length L, Fbuckling is increased
four-fold to 2.3 mN (estimated) which is beyond the minimal forced required to penetrate brain
tissue. To insert the Parylene array, without resorting to rigid materials or increasing the cross-
sectional footprint, we designed a temporary brace to reduce the effective length of the Parylene
array during insertion, and by using a dissolvable material, the brace can be removed before
inserting the remaining length of the array into the brain.
This dissolvable brace was fabricated from polyethylene glycol (PEG) and validated in
mechanical tests before it was used in surgical implantations. In preparation for implantation, the
Parylene arrays were partially encapsulated in the bio-degradable PEG brace. The brace was
created by casting molten PEG around the completed Parylene array using a mold prepared from
three layers of 0.5 mm thick polydimethylsiloxane (PDMS) and assembled as depicted in Figure
16. The first PDMS layer serves as the base of the mold. The second and third layers sandwich the
array, supported by an acrylic backing, and outlined a cavity which defines the thickness and shape
Figure 21. Schematic representations of criteria for the insertion of flexible probes. a) Diagram shows
different insertion results. If the force that a flexible probe can handle exceeds the minimal insertion
force for brain tissue penetration, the flexible array can be inserted straightly. On the other hand, if the
buckling force of an array is less than the force for brain tissue penetration, the flexible array will be
bend. b) shows the negative correlation between the length of the Parylene array and buckling threshold
of the array. Insertion force required for a straight insertion into brain tissue is highlighted with light
green and the length of the bare Parylene array is indicated with a red arrow.
1 2 3 4 5 6 7
0
2
4
6
8
10
12
14
16
x 10
−4
Effective Length (mm)
Force (N)
F_brain < F_buckling
Straight
Insertion
F_brain > F_buckling
Bended
Probe
a) b)
Force Required for
Straigth Insertion
Total Array
Length
44
of the PEG brace. The Parylene array contained within the PDMS assembly was heated in a 50 °C
oven for 30 seconds. Molten PEG of molecular weight 3,350 (Sigma-Aldrich, Darmstadt,
Germany) was injected into the cavity and a PDMS coverslip was placed on top to wipe away
excess PEG. The entire assembly was then cooled at room temperature for 5 minutes until the PEG
solidified. The encapsulated device was then removed from the mold carefully.
3.4 Summary
Parylene C has been used as an insulation coating on many electronics in biomedical
implants. We further expanded the application of Parylene C in the biomedical field through the
development of a 64-channel, Parylene-based, multi-electrode array. This novel, penetrating
polymer array used the Parylene C as both the insulation and structural materials. The array
featured eight electrodes on eight individual Parylene shanks. Instead of simply having linear
Figure 22. PDMS mold used to create the PEG brace that supports probe shanks during implantation.
The mold was comprised of three PDMS sheets: a base layer, a second layer which accommodates a
clear acrylic insertion backing that supports the array as its probes span the mold cavity, and a third
layer which is gently aligned to and laid over the probe tips to complete the mold. Molten PEG was
injected into the mold cavity to brace the top length of the probes temporarily, during implantation, to
allow for probe insertion into brain without buckling. a) Photograph of PDMS assembly surrounding
array with the mold cavity outlined by dotted red line. b) Exploded view of PDMS layers and array on
top of clear acrylic backing.
45
electrodes placed evenly along individual Parylene shanks, the layout of electrodes on each
Parylene shank was specifically arranged to conformal to the anatomical distributions of
hippocampal neurons. With recording groups placed at different vertical depths, two cell body
layers, the CA1 and the CA3, were targeted simultaneously. With advanced nano-fabrication
techniques provided by our collaborators, straight and fully functional Parylene multi-electrode
arrays were fabricated. The corresponding insertion method involving the application of a
dissolvable brace to reduce probe buckling without the use of extra insertion shuttle was also
developed. Customized PCBs were designed and fabricated to interface the Parylene multi-
electrode array with a conventional neural data acquisition system. The testing and evaluation of
the performance of this Parylene multi-electrode array under both in vitro and in vivo conditions
will be detailed in the following chapter.
3.5 References
[1] M. L. Block, L. Zecca, and J. S. Hong, “Microglia-mediated neurotoxicity: Uncovering the
molecular mechanisms,” Nature Reviews Neuroscience. 2007.
[2] M. L. Block and J. S. Hong, “Microglia and inflammation-mediated neurodegeneration: Multiple
triggers with a common mechanism,” Progress in Neurobiology. 2005.
[3] D. H. Szarowski et al., “Brain responses to micro-machined silicon devices,” Brain Res., 2003.
[4] J. N. Turner et al., “Cerebral astrocyte response to micromachined silicon implants.,” Exp. Neurol.,
vol. 156, no. 1, pp. 33–49, 1999.
[5] P. J. Rousche, D. S. Pellinen, D. P. Pivin, J. C. Williams, R. J. Vetter, and D. R. Kipke, “Flexible
polyimide-based intracortical electrode arrays with bioactive capability,” IEEE Trans. Biomed. Eng.,
vol. 48, no. 3, pp. 361–370, 2001.
[6] S. P. Lacour, G. Courtine, and J. Guck, “Materials and technologies for soft implantable
neuroprostheses,” Nat. Rev. Mater., vol. 1, no. 10, 2016.
[7] H. L. Brydon, G. Keir, E. J. Thompson, R. Bayston, R. Hayward, and W. Harkness, “Protein
adsorption to hydrocephalus shunt catheters: CSF protein adsorption.,” J. Neurol. Neurosurg.
Psychiatry, vol. 64, no. 5, pp. 643–7, 1998.
[8] R. D. Dickerman et al., “Failure rate of frontal versus parietal approaches for proximal catheter
placement in ventriculoperitoneal shunts: Revisited,” J. Clin. Neurosci., vol. 12, no. 7, pp. 781–783,
2005.
[9] W. F. Gorham, “A New, General Synthetic Method for the Preparation of Linear Poly-p-xylylenes,”
J. Polym. Sci. Part A-1 Polym. Chem., vol. 4, no. 12, pp. 3027–3039, 1966.
46
[10] D. C. Rodger et al., “Flexible parylene-based multielectrode array technology for high-density
neural stimulation and recording,” Sensors Actuators, B Chem., vol. 132, no. 2, pp. 449–460, 2008.
[11] C. Metallo, R. D. White, and B. a. Trimmer, “Flexible parylene-based microelectrode arrays for
high resolution EMG recordings in freely moving small animals,” J. Neurosci. Methods, vol. 195,
no. 2, pp. 176–184, 2011.
[12] J. T. W. Kuo et al., “Novel flexible Parylene neural probe with 3D sheath structure for enhancing
tissue integration.,” Lab Chip, vol. 13, no. 4, pp. 554–561, 2013.
[13] J. A. Thornton and D. W. Hoffman, “Stress-related effects in thin films,” Thin Solid Films, vol. 171,
no. 1, pp. 5–31, 1989.
[14] J. Charmet, J. Bitterli, O. Sereda, M. Liley, P. Renaud, and H. Keppner, “Optimizing parylene C
adhesion for MEMS processes: Potassium hydroxide wet etching,” J. Microelectromechanical Syst.,
vol. 22, no. 4, pp. 855–864, 2013.
[15] P. K. S. S. Harsha, Principles of Vapor Deposition of Thin Films. 2006.
[16] E. Alfonso, J. Olaya, and G. Cubillos, “Thin Film Growth Through Sputtering Technique and Its
Applications,” in Crystallization - Science and Technology, 2012.
[17] S. Dabral et al., “Stress in thermally annealed parylene films,” J. Electron. Mater., vol. 21, no. 10,
pp. 989–994, 1992.
[18] S. H. Felix et al., “Insertion of Flexible Neural Probes Using Rigid Stiffeners Attached with
Biodissolvable Adhesive,” J. Vis. Exp., no. 79, pp. 1–12, 2013.
[19] T. D. Y. Kozai and D. R. Kipke, “Insertion shuttle with carboxyl terminated self-assembled
monolayer coatings for implanting flexible polymer neural probes in the brain,” J. Neurosci.
Methods, vol. 184, no. 2, pp. 199–205, 2009.
[20] S. Felix et al., “Removable silicon insertion stiffeners for neural probes using polyethylene glycol
as a biodissolvable adhesive,” Proc. Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. EMBS, pp. 871–
874, 2012.
[21] L. Luan et al., “Ultraflexible nanoelectronic probes form reliable , glial scar – free neural integration,”
no. February, pp. 1–10, 2017.
47
Chapter 4. Evaluation of the Parylene Multi-Electrode Array
Evaluation of the Parylene Multi-Electrode
Array
Neural interfacing devices intend for successful long-term implant in the brain must meet
a series of strict criteria in the electrical, mechanical and biological arenas. Electrically, appropriate
insulation and conductivities must be maintained over long periods in a warm, saline environment.
Mechanically, the neural implant needs to tolerant any stress induced by all handling, cleaning and
implantation procedures. In addition, it also needs to be capable of withstanding possible
micromotions between the brain tissue and the implant post-implantation. Biocompatibility is a
more critical concern for any neural implants intends for chronic use. Biologically, the minimal
requirement for any implantable devices is that it should not cause any excessive immune response.
The characterization of properties of the Parylene multi-electrode array both in vitro within an
artificial environment that mimics the biological situation and in vivo will be described in this
chapter.
4.1 Benchtop Testing of the Parylene Multi-Electrode Array
Before the implantation of the fully functional Parylene multi-electrode array to animals,
together with our collaborator, comprehensive benchtop testing and evaluation about the
mechanical and electrical properties of the Parylene array were conducted.
4.1.1 Electrode impedance
The electrical property of an electrode can be evaluated electrochemically through
electrode impedance spectroscopy (EIS). Benchtop EIS usually involves the application of a small
alternating voltage between a recording electrode fabricated on the Parylene multi-electrode array
48
and a counter electrode immersed in a conductive solution. By measuring the current, the complex
impedance of the electrode-solution interface, both in magnitude and phase, can be calculated.
Through repeated testing at different voltage frequencies (commonly from 10 Hz to 100 kHz), the
electrical property of an electrode can be characterized. The frequency of a neural spike is around
1 kHz. Impedance magnitude measured at 1 kHz can help to predict the ability of an electrode to
record low amplitude voltages. An electrode with high signal-to-noise (SNR) recordings usually
associates with low electrode impedance.
Prior to PEG encapsulation and insertion into the ZIF, hippocampal arrays were CV
cleaned and the electrical property of each electrode was analyzed using EIS. CV (-0.2 V to 1.2 V,
scan rate of 250 mVs
-1
) was run on each individual electrode trace for 30 cycles in 0.5 M H2SO4
purged with N2 for five minutes prior to scanning. A 1 cm
2
Pt plate was used as a counter electrode
and an Ag/AgCl electrode was used as a reference electrode. EIS was performed in 1x phosphate
buffered saline (PBS) (OmniPur 10x PBS, EMD Chemicals, Darmstadt, Germany) with an
excitation voltage of 25 mV (AC) over frequencies from 1 Hz to 0.1 MHz. Electrodes with
impedance measuring greater than 2 MΩ at 1kHz, or with a phase spectrum that did not transition
from resistive (close to 0°) to capacitive (close to -90°) with decreasing frequency, were considered
open circuits and were not used for average impedance calculations.
Figure 23 .Representative CV of electrode from a sputtered hippocampal array. The first CV curve is
relatively narrow, indicating a small electroactive area. As the cycles progress the current response
broadens at different voltages indicating increased cleanliness of the electrode surface area which
allows for more surface reactions and current flow. At cycle 30 this CV now mimics the traditional CV
of Pt in H 2SO 4.
49
CV cleaning of electrodes successfully removed residual scum left over from processing
from the surface of the electrodes, as evidenced by higher magnitude currents (lower impedances)
in each of the 30 consecutive CV cycles. Electrochemical impedance, measured at 1 kHz,
decreased 37% (n=8) from a mean of 801 ± 29 kΩ to 507 ± 60 kΩ following CV cleaning. In a
larger survey of electrodes (n=256 electrodes across four arrays) the mean electrochemical
impedance of clean electrodes at 1 kHz was calculated as 677 ± 297 kΩ.
4.1.2 in vitro testing of the insertion strategy
The feasibility of the insertion of the bare Parylene multi-electrode array with temporarily
shorted effective length was first tested with brain phantoms. Agarose blocks (0.5% in
concentration) were used to simulate the consistency and density of brain tissue [1]. Sham arrays,
which were identical to fully functional Parylene arrays in dimension but lacked the thin metal
layer (both the base Parylene layer and the insulation layer was 10 µm), was attached to an acrylic
Figure 24. Illustrations of procedures involved during the insertion of the Parylene array with the help
of a dissolvable PEG brace. a) shows schematic representations of the insertion steps of a Parylene array
into agarose. b) shows front view photos collected during the insertion of a sham Parylene array into
agarose with the implantation strategy described in this chapter. c) shows photos of both front and side
view collected after the sham array been fully inserted.
Acrylic backing
Parylene
array
PEG
Agarose
Agarose Agarose
Fully
inserted
arrray
a)
b)
Front view Sideview
c)
2 mm
PEG
Agarose
50
backing with the PEG brace. The PEG brace also encapsulated the top part of the Parylene array
with only less than 2 mm of the array tip been exposed. The acrylic backing was then fixed onto a
micromanipulator. The exposed tip of the embedded Parylene array was slowly inserted into the
brain phantom by fine advancement of the micromanipulator. As the PEG brace reached the
surface of the agarose block, drops of saline were applied to dissolve the brace. At the point that
remining part of the Parylene array was exposed, the full Parylene array was inserted into the brain
phantom straightly.
4.2 Acute, in vivo Evaluation of the Parylene Multi-Electrode Array
in vitro, Benchtop testing of the Parylene multi-electrode array shown that the insertion of
bare Parylene arrays with supportive PEG brace is a potential method that can be applied to in vivo
Figure 25. in vivo insertion of the bare Parylene array to the rat hippocampus with supportive PEG
brace. The photo on top left shows the exposed tip of the bare Parylene array been inserted. The top
right photo shows the PEG brace dissolving with added saline. The photo on the bottom left shows the
newly exposed Parylene array after the PEG brace washed away with saline, and the bottom right photo
shows the bare Parylene array been fully inserted.
PEG
Brace
Tip of the
Parylene Array
PEG Dissolved
with Saline
Newly Exposed
Parylene Array Fully Inserted
Parylene Array
51
preparation. We first tested the feasibility of this method with sham devices. A sham Parylene
array was identical to fully functional arrays in geometry and dimension only with no metal
electrodes and traces been deposited. Then the performance of fully functional Parylene arrays on
recording from the rat hippocampus was evaluated with acute preparations.
4.2.1 in vivo insertion of the Parylene multi-electrode array
Two sham Parylene arrays encapsulated in PEG braces and supported with acrylic backings
were implanted into both hippocampi of an animal. All animal preparation and animal care were
in accordance with the IACUC guidelines and approved by the DAR at USC. A SD rat older than
3 months of age and weighing between 300 - 450 g was used. The animal was pre-anesthetized
with an intraperitoneal (IP) injection of a ketamine and xylazine mixture. Anesthesia was
maintained intra-operatively through the delivery of an inhaled mixture of isoflurane and oxygen
delivered to the animal through a nose cap. Negative toe pinch withdrawal reflexes tested
Figure26. Histological results that verified the straight insertion of a sham Parylene array with the
insertion method described in chapter 2.3.2. Locations of individual Parylene shanks were pointed out
with red arrows. A zoomed in view of locations of five shanks is also shown.
52
throughout surgery confirmed appropriate anesthetic level. A stereotactic frame was used to hold
the animal in place and to keep the animal’s head flat. To expose the brain surface above the
implantation site, 2 x 4 mm cranial windows were drilled away above both the right and left dorsal
hippocampi and the dura and blood vessels were carefully removed with forceps. The exposed,
bare tips of sham Parylene arrays were poised above the implantation site located at ~2.5 mm
posterior to bregma and ~2.45 mm lateral of the midline (bilaterally). A micromanipulator that
attached to the stereotactic frame was used to support and advance the acrylic backing during the
implantation. Parylene arrays were slowly inserted into the brain until the bottom edge of the PEG
brace reached the surface of the brain. The brace was then gradually dissolved away with saline
solution and the newly exposed array length was advanced in increments of 0.05 mm at a speed of
10 µm/s. After the full array (5 mm) was inserted, the subject was perfused followed procedures
described in Chapter 2.1.2. After implanted arrays were removed, 50 µm thick brain slices were
prepared and stained with crystal violent. Successful insertion of the Parylene array in deep brain
region was confirmed in histological slices (Figure 26).
Figure 27. Data acquisition setup during the implantation of the Parylene array. a) shows the picture of
an animal with the Parylene array implanted. The array was fixed onto the skull with dental cement
covered half of the first PCB. The detachable second PCB was connected to a customized 64 channel
switch board. b) shows the connection between the second PCB to the switch board with zoom in of the
second PCB shows on the right and the 64-channel switch board on the left.
64 channel
switch board
PCB 2 with SSB6 and
Omnetic connecters
From animal To amplifier
a) b)
53
4.2.2 Acute, in vivo recording with the Parylene multi-electrode array
Before the insertion of functional Parylene multi-electrode arrays, all Parylene arrays were
inspected under microscope to check for short circuits and cracks on the connective trace which
indicate open circuits. The electrical property of each recording electrode was also evaluated
through EIS. Only electrodes that shown no sign of open or short circuits and whose impedance
measured under 1 kHz fall in normal range were considered functional electrodes.
Eleven functional Parylene multi-electrode arrays that passed electrochemical testing were
implanted into eleven animals. The implantation procedure was similar to steps for the insertion
of sham Parylene arrays but with some modifications. Following same insertion procedures
detailed in Chapter 4.2.1, functional Parylene arrays were inserted into the rat brain. Neural signals
were monitored throughout the implantation procedure for the presence of complex spikes which
is considered as the characteristic firing pattern of hippocampal pyramidal cells. The presence of
complex spikes helped to confirm proper placements of the Parylene array in both the CA1 and
CA3 layers of the hippocampus. In order to collect neural activities during the implantation, beside
the cranial window four additional small holes, for anchor screws and the placement of ground
wire were drilled on the rat skull. One anchor screw also doubled as a ground by making contact
with the cerebrospinal fluid. During the implantation, the Parylene array were connected to an
oscilloscope and a speaker for signal display. Neural activities were bandpass filtered from 300
Hz to 10,000 Hz and amplified 1,000 times with a single channel, differential amplifier (DAM50
Extracellular Amplifier, WPI, Sarasota, FL). The Parylene array was first advanced in small
increments at a speed of 10 µm/s, once the first group of electrodes began recording spikes from
the more superficial, CA1 layer of the hippocampus (located at ~ 3 mm deep from the surface of
the skull), advancement of the Parylene array was frequently paused to monitor and record neural
activities at each depth. Then the array was further advanced until the second group of electrodes
reached the CA1, at which point the first group of electrodes recorded from the deeper, CA3 layer.
During four implantations, neural activities from individual channels were recorded from the
oscilloscope. After the array had reached desired target location, a thin layer of dental cement was
54
applied over the insertion site to hold the array in place. The ground wire was twisted onto the
screw that acted as the ground electrode for added stability, and the tip of the ground wire was
inserted into the hole above the cerebellum. For implantations that the two-PCB setup were used,
the electrical packaging was lowered to the cranium and additional dental cement was then applied
to create a ‘cap’ of dental cement that encapsulated the array and the first half of the permanent
PCB, with care taken to leave the SSB6 receptacle at the top of the PCB exposed for future PCB
to PCB connections during experiments. For Parylene arrays that implanted with the single PCB
setup, dental cement was applied to cover the lower half of the PCB and left the two Omnetic
connectors exposed for future connection.
Representative data describing neural recordings obtained from the implantation surgery is
shown in Figure 28. Implantation coordinate, number of functioning shanks and number of Parylen
shanks which simultaneously recorded from both desired regions are listed in Table VI. We
considered each cluster of four closely spaced electrodes to be a ‘recording group’. Recording
groups containing at least two electrodes that passed electrochemical inspection prior to
Figure 28. Complex spikes (burst of 2‐ 6 single spikes of decreasing amplitude with ≤ 5 ms interspike
intervals) recorded from the CA1 and CA3 during the second implantation surgery. a) 1 s long recording
traces from electrodes on the same probe recording from both the CA1 and CA3 once the array has
reached its target location. b) Zoom‐ in view showing complex spikes recorded from the same electrodes
that, initially during implantation lay within the CA1 sub‐ region, and, as the probe was advanced,
eventually reached the CA3. Color coded anterior‐ posterior sections of the hippocampus illustrate the
change in hippocampal depth along the anterior‐ posterior direction.
AP= -2.45
AP= -4.20
AP= -3.90
AP= -3.70
AP= -3.25
AP= -2.85
CA1
100 µV
20 ms
CA3
CA1
shank 2 shank 5
S8 - CA1
S2 - CA1
S4 - CA1
S6 - CA1
S7 - CA1 S7 - CA3
S2 - CA3
S4 - CA3
S6 - CA3
S8 - CA3
200 µV
200 ms
(a) (b)
55
implantation were considered functional electrode groups. A total of 57 (out of 64) recording
groups were considered functional and were monitored during four implantation surgeries. Unitary
activities and complex spikes were successfully recorded from 35 recording groups (61.4% of
functional recording groups). In each of the four implantation surgeries, signals were obtained
from electrodes in both the CA1 and CA3 sub-regions. Probes with two or more working electrodes
in both the CA1 and CA3 electrode groups were considered functional probes. A total of 28
functional probes (out of 32) were implanted, and at the final implant location, 13 of these shanks
recorded unitary activities from both the CA1 and CA3. As the insertion procedure progressed, the
depth at which complex spikes from these two sub-regions were detectable was recorded.
Table VI. Implantation coordinates, final target placement, and number of functional probes across
four array insertions.
Implantation X, Y
implantation
coordinates
Depth of CA1
complex
spikes
Depth of
final
placement
Number of
functional probes
that reached both
CA1 and CA3
1 2.64, 2.45 - 2.95 - 4.15 2/5
2 2.49, 2.45 - 2.91 - 4.10 5/6
3 2.43, 2.45 - 2.98 - 4.15 2/6
4 2.49, 2.45 - 2.50 - 4.10 4/8
56
For other seven implantations, neural activities were also monitored but not recorded
during the insertion of the Parylene array. Monitoring neural activities during insertion helped the
insertion of Parylene arrays to appropriate depth. After the Parylene array was implanted to desired
depth, it was chronically fixed onto the animal’s head with dental cement following steps described
above. Immediately after the surgery, signals from all 64 channels were recorded for 5 to 10 mins
from animals still under light anesthesia with a 64 channel, data acquisition system (OmniPlex,
Plexon Inc., Dallas, TX). Spike sorting based on principle component analysis (PCA) were
Figure 29. Simultaneous recordings from all 64 channels of a Parylene multi-electrode array implanted
to a rat recorded immediately after the implantation surgery. The animal was still lightly anesthetized.
Color coded traces on the top show a half-second recording from all 8 implanted Parylene shanks with
signals from the CA1 show on the top half and signals from the CA3 show on the bottom. Purple traces
at the bottom row show a zoomed-in view of complex spikes and single spikes recorded with the sixth
shank.
200 uV
100 mSec
Zoom in View
Complex Spikes
CA1
CA3
57
conducted to recorded neural signals with build-in software (Offline Sorting, Plexon Inc., Dallas,
TX).
Unitary activities from both the CA1 and the CA3 sub-regions were recorded with all seven
implanted Parylene multi-electrode arrays. Details of these seven implantations, including the
number of functional shanks on each array, the number of shanks that recorded spike activities
from the animal, the number of shanks recorded from both the CA1 and the CA3 sub-regions of
the hippocampus and the number of units recorded immediately after the implantation were listed
in Table VII. A total of 52 (out of 56) functional shanks were implanted. Spike activities were
recorded with 48 (92.31% of implanted functional shanks) Parylene shanks. Among these shanks,
25 Parylene shanks recorded unitary activities from both the CA1 and the CA3 regions. After spike
sorting, a total of 496 units, with 187 from the CA1 sub-region and 309 from the CA3 sub-region,
were recorded from seven, lightly anesthetized animals. On average, over 70 (70.8) units were
recorded from each animal. Simultaneous recordings from all 64 channels from one implanted
animal were shown in Figure 29.
Table VII. Details of seven implantations from which neural signals were recorded with data
acquisition system immediately after implantation surgery.
Figure 30. Spike sorting with PCA method.
58
Implantation Number
of
functional
shanks
Number of
shanks
recorded
spike
activities
Number of
shanks
reached
both sub-
regions
Number
of CA1
units
Number
of CA3
units
Total
number of
units
1 8/8 8/8 5 36 49 85
2 8/8 6/8 Bended 7 35 42
3 8/8 6/8 5 37 50 87
4 7/8 7/7 4 32 27 59
5 6/8 6/6 4 9 70 79
6 7/8 7/7 4 48 23 71
7 8/8 8/8 3 18 55 73
4.3 Chronic, in vivo Evaluation of the Parylene Multi-Electrode Array
Acute, in vivo neural data recorded simultaneously from multiple sub-regions of the rat
hippocampus with the Parylene multi-electrode array indicated promising prospect of the
application of the Parylene array chronically. Eight out of those eleven animals implanted with
Parylene arrays were given 7-16 days post implantation to fully recover, after which neural
activities were recorded with the 64 channel data acquisition system while the animal freely
roamed the open field.
59
4.3.1 Chronic fixation of the Parylene multi-electrode array and the package
The procedure of chronic implantation of Parylene multi-electrode arrays was similar to
the process detailed in session 4.2.2. In short, after desired brain surface of the animal to be
implanted were exposed, the tip of the Parylene array which was pre-embedded and supported
with PEG brace and acrylic backing was inserted slowly. As the bottom edge of the PEG brace
approached brain surface, drops of saline were applied to dissolve the bottom part of the PEG
brace. The brace was then gradually dissolved away and the newly exposed array length was
advanced. If the PEG brace was directly submerged in saline, it would take < 20 min to fully
dissolve. During the insertion advancement of the array was frequently paused to monitor neural
activities at each depth. Therefore, the dissolution of the PEG brace was carefully controlled to
avoid unnecessary array exposure during the surgical pauses. To accomplish this, a small amount
of saline was applied directly to the part of the PEG brace to be dissolved and any excess saline
was blotted away. After the Parylene array reached desired location, dental cement was applied to
form a cap which fixed the array to anchor screws chronically.
Figure 31. Chronic fixation of the Parylene array and the two-PCB package with three anchor screws.
a) shows three anchor screw been drilled into the skull. b) shows fixation of the first PCB with dental
cement and the picture of an animal after the Parylene array and head-mount package been chronically
mounted onto the skull. c) shows a side view and bottom view of a dental cement cap been pulled out
of the animal’s head after the animal been euthanized.
Three
Anchor
Screws
a). b).
PCB One
Dental
Cement
Cap
c).
Dental cement cap removed
from the animal’s head
PCB One
Attached Skull
60
The chronic fixation of the Parylene array was directly adapted from the fixation of
microwire electrodes arrays. At first four implantations, the two-PCB setup detailed in Chapter
3.2.3 were used. Three anchor screws were drilled into the animal’s skull to help secure the
location of the Parylene array. Dental cements that coved all three anchor screws, the full array
and the lower half of the first PCB was applied to chronically fix the Parylene array and its package.
This chronic fixation method was firm and stable for the two-PCB setup. With repeated connects
and disconnects performed daily for the first two weeks after the animal fully recovered and at
least once a week after one-month post-implantation, the dental cement cap stayed stable on the
animal’s head. At the end of the experiment, the dental cement cap was removed from animal’s
Figure 32. Schematic representations of the single PCD head-mount package and the chronic fixation
method which used 5 anchor screws. a) shows the dimension of the single PCB. Photos on the right
show a side view and bottom view of one dental cement cap and the single PCB package fall off an
animal’s head during one recording session. The bottom view shows no sign of bone attachment to any
of the anchor screws. b) shows the location of five anchor screws and the bone window for the Parylene
array. The photo on the right shows an animal after the implantation of a Parylene array and the single
PCB package been chronically mounted on the animal’s head.
Holes for Anchor Screws
Bone Win-
dows for
Electrode
Ground
a).
b).
61
skull. Small bone fractions that attached to anchor screws were noticed which indicated a firm
bonding.
For the following seven implantations, the one-PCB package (detailed in Chapter 3.2.3)
were used to eliminate extra noises introduced by the unstable connection between SSB6
connectors. At first, the Parylene array and the single PCB was chronically anchored onto the
animal’s head with three anchor screws. However, the increased PCB size, two Omnetic
connectors and the extra amounts of dental cement needed to mount the single PCB posed a
challenge to the long-term fixation of the new head-mount package to the animal’s skull. Four
chronically mounted dental cement caps together with the implanted Parylene arrays fell off the
animal’s head two to eight weeks post implantation. On the detached dental cement cap, no bone
fraction was observed.
To add more supports to the head-mount piece and to improve the stability of the chronic
fixation method, a smaller drill bit was used to create tighter holes for self-tapping anchor screws
for subsequent implantations. In addition, two more anchor screws were used to strengthen the
bonding between dental cement cap and the animal’s skull. Seven more animals were implanted
with the modified mounting methods, only one dental cement cap fall off the animal’s head again
during chronical recording sessions one-month post implantation. All other dental cement cap
stayed stable for up to twelve months with this optimized method.
4.3.2 Long-term recording with the Parylene multi-electrode array
The Parylene multi-electrode array was also chronically implanted into animals. After one
to two weeks recovery, in vivo signals were recorded while the animal explored freely in an open
field. Unitary activities, including both complex spikes and single spikes, together with local field
potentials (LFP) were recorded from both the CA1 and the CA3 region with the Parylene array.
a) Animal preparation
Before the implantation of the Parylene multi-electrode array, the amount of food provided
to male Sprague-Dawley (SD) rats greater than three months were controlled to maintain the
animal’s weight at 85% to 90% of free-feeding body weight. All animals have ad libitum access
to water. Then the animal was introduced to a circular recording chamber daily. The animal was
encouraged to run around and explore the entire chamber with small food pellets been randomly
62
scattered on the floor of the recording chamber. After the animal got familiar with the recording
environment, a Parylene array was implanted following steps described in Chapter 4.2.2 and 4.3.1.
In summary, the Parylene array was implanted through stereotaxic brain surgery. After
implantation, the Parylene array and the package was chronically fixed onto the animal’s head
with dental cement.
After implantation, the animal was given one to two weeks to fully recover from the
operation. Buprenorphine was admitted through subcutaneous injection to the implanted animal
24 to 72 hours post-implantation. In some animals, Buprenorphine Slow Release (BUP SR) which
was designed to slowly release BUP over 72-hour period was injected on time at the time of the
implantation surgery. The animal was examined daily to prevent severe pain and infection. The
animal was euthanized if the animal developed any infection that cannot be controlled.
b) Data acquisition
Behavioral training and chronic recording were conducted in a round open field. The
chamber is solid black with 76 cm in diameter and 50 cm in height. The open field located in a
customer-made copper mesh (Georgia Copper LLC) covered faraday cage which helps in reducing
of power line noise from other equipment in the recording environment. The recording chamber
Figure 33. A view of the recording chamber captured with the overhead camera. Location of the animal
was represented with a (x, y) coordinate at each time stamp. The shape of the animal was well identified
by the “counter tracking method” and highlighted with yellow color. The moving trace of the animal
was illustrated with the green line.
70 cm
50 cm
(50, 70)
x
y
63
was also surrounded by a black curtain to isolate the chamber from lab environment. The chamber
was illuminated by a white LED strip which shaped into a circle placed on top of the faraday cage.
Behaviors of the animal was captured by an over-head camera.
Moving traces of the implanted animal were recorded and processed with a behavioral
research system (CinePlex, Plexon Inc., Dallas, TX). Movements of the animal were recorded with
a camera oversees the whole recording chamber. The distance that each pixel in the camera view
represented was calibrated pre-recording. The animal’s physical position in the recording chamber
was represented by a (x, y) coordinate which is the measurement of distances of the animal from
the reference axes, the bottom and left edge of the camera view, at each time stamp. Two methods
were used to track the position of the animal in the recording chamber. The object contour tracking
method compares the color of the animal with the background color and find the whole-body shape
of the desired object. The center of gravity of the shape is calculated and tracked. Since the white
color of SD rats and the black background color used in the experiment formed a sharp contrast,
the counter of the animal was able to be well defined with this tracking method. The color makers
tracking method tracks particular colors that can be defined pre-recording. Two color taps (red and
green) were attached to each side of the headstage during recording. Tracking moving traces of
those color marker provided extra information about the head movement and head direction of the
animal beside locations of the subject. Which makes the future evaluation of the correlation
between head directions and neural activities of the hippocampus possible.
Correlated neural activities were collected with a 64 channel, data acquisition system
(OmniPlex, Plexon Inc., Dallas, TX). Neural signals were amplified 20 times with the headstage
and further amplified with a differential amplifier with adjustable gain for each channel. Amplified
signals were digitized at 40 KHz and stored for off-line analysis. Spike activities were detected,
and spike sorting was conducted using an off-line software (Offline Sorter, Plexon Inc., Dallas,
TX). Low frequency component of the raw, continuous data was filtered out by applying a
Butterworth four-pole, low-cut filter with 250 Hz cutoff frequency. Neural activities across the
spike detecting threshold (set at higher than three times of the standard derivation of filtered signal)
were considered as spikes. Cross-channel artifacts and high amplitude artifacts were manually
removed. Single units were manually sorted with principle component analysis (PCA) method.
64
c) Long-term recording result
After recovery, simultaneous recording traces of all 64 electrodes from eight animals were
collected and spike sorting was conducted to the recorded data offline. All eight arrays recorded
unitary activities from behaving animals, including complex spikes from hippocampal pyramidal
cells. Chronic neural recordings were monitored over the period of one week to almost a year post-
implantation. Units with SNR above 3 were constantly recorded from all eight implantations. A
Figure 35. Chronic recordings obtained from one implantation of the Parylene multi-electrode array
one-month post implantation. Spike waveforms were superimposed for each unit. Five units recorded
from the CA1 region were shown with light blue while thirteen units recorded from the CA3 were shown
with purple.
Figure 34. Example of multiple units recorded from the CA3 sub-region with one recording site on the
Parylene muli-electrode array. Different units were color coded and corresponding SNRs for each unit
were listed in the figure. A zoom in view of part of the continuous data which contains two group of
complex spikes from two units is shown on top right.
Complex Spikes
Yellow Unit
Complex Spikes
Green Unit
SNR = 15 SNR = 5.7 SNR = 3.6 SNR = 2.7
200 µV
20 ms
100 µV
200 µs
65
qualitive evaluation of numbers of units recorded from multiple sub-regions of the rat
hippocampus showed that five implanted Parylene array simultaneously recorded multiple units
from both the CA1 and the CA3 sub-regions. At least 91 units, with 67 from the CA3 and 24 from
the CA1, were recorded from eight implanted animals. Up to four neural units were recorded from
a single electrode. The one Parylene array which obtained the most chronic unitary activities from
both the CA1 and CA3 regions recorded a total of 18 units seven weeks post-implantation. With
five units from the CA1 and thirteen units from the CA3.
Average spike amplitudes and baseline noise levels for eight implanted Parylene multi-
electrode array cross the recording period of twelve weeks post-implant were calculated and
plotted in Figure 35. The baseline noise level ranged from 16.48 µV to 49.39 µV. On average, the
noise level at the first week post-implantation was 24.77 ± 5.82 µV and the noise level at the tenth
week post-implantation was 28.80 ± 6.82 µV. The noise level of all eight implanted Parylene array
stayed stable over the period of 10 weeks post-implantation. Chronic experiment for one animal
(animal 4) was terminated at two weeks and at four weeks for another animal (animal 6) due to the
failure of the head-mount dental cement cap. For another two implantations (animal 1 and animal
Figure 36. The plot of average noise level and average amplitude of spikes over the duration of twelve
weeks of eight chronically implanted Parylene arrays.
Average
Noise Level
Average Spike
Amplitude
66
8), the average spike amplitude continuously decreased over time. For animal 1, this decreasing
was likely to be caused by the unstable connection between SSB6 connectors. This made the
number of functional channels reduced greatly over time. For animal 7, the average signal
amplitude varied for the first 6 weeks then stayed relatively stable after that. This variation agreed
with the fact that it takes around six weeks for the immune respond to stabilize [2]. For other two
implanted Parylene arrays, the average spike amplitude was stable over the period of 10 weeks.
4.3.3 Life-time testing
For one particular implantation, neural activities were continuously checked every week
for an extended period of time to test the life-time performance of the Parylene multi-electrode
array. Neural activities were recorded for two to four times a week for the first two months after
implantation then recorded weekly until almost one-year post-implantation. The animal was given
ad libitum access to both food and water after two months. The animal’s body weight and health
condition were also monitored weekly to ensure the animal was not suffered from the chronic
Figure 37. Recording of neural activities over extended time period with the Parylene multi-electrode
array.
67
implantation and any age related disease. Redness that been occasionally noticed on the skin
around the dental cement cap was treated with bacitracin zinc ointment. The head-mount, dental
cement cap stayed stable and bonded to the skull firmly for 362 days. No motion of the dental
cement cap was noticed over this extended period of period. The average noise level of channels
that recorded unitary activities remained stable for one year, only changed from 23.70 µV at two
weeks post-implantation to 24.16 µV at 362 days post-implantation. Multiple units with SNR
greater than three were recorded during every recording session over this extended period. The
average spike amplitude was 102.28 µV at thirteen days after implantation and this number was
101.90 µV at around 330 days post-implantation. At 362 days after the implantation of the Parylene
array, a unit with SNR greater than 8.9 (spike amplitude was 211.73 µV and the noise level was
23.73 µV) was still been recorded.
4.3.4 Immune responds to the Parylene multi-electrode array
In response to foreign implants, glial cells such as astrocytes proliferate to form a thin,
dense scar to isolate the implant from normal, intact brain tissue [3]. The activation of glial cells
induced by the implantation of the Parylene multi-electrode array and the effect of the Parylene
array to surrounding neurons was evaluated through immunohistochemical (IHC) analysis at both
one-month and six-months post-implantation.
A sham array, which lacked metal electrodes and traces, but was otherwise identical to
fully fabricated arrays, was implanted into the right hippocampus of a single rat. This animal was
sacrificed at one-month post-implantation, at which time the animal was deeply anesthetized with
an IP injection of ketamine and xylazine and intracardially perfused with 10% paraformaldehyde.
The rat brain was dissected from the cranium and the hippocampal array was removed from the
brain. After the brain was dehydrated in 18% sucrose overnight, the tissue was then embedded in
3% agarose for slicing. 50 µm thick transverse slices cut with a vibratome (Lecia, VT1000 S).
Slices taken at a depth of 2.7 mm and 2.75 mm were stained with antibodies for glial fibrillary
acidic protein (GFAP) and neuronal nuclei (NeuN) respectively for the identification of astrocytes
and neurons and counterstained with hematoxylin. Sections of tissue on the same histological slice,
on the same hemisphere, but > 500 µm away from the implant site were chosen as control regions
for comparison. Radial rings of 25 µm were drawn around probe cross-sections and corresponding
control regions. Color thresholding in ImageJ software (National Institutes of Health, Bethesda,
68
MD) was used to identify and count the concentration of astrocytes in each annulus. Rectangular
bins of 50 µm width, whose length matched the thickness of the cell body layer were used to count
the concentration of neurons in implanted and control regions.
The concentration of astrocytes and neurons around the implantation site of a sham array,
measured one month post implantation, are presented in Figure 37, 38, Table VIII and Figure 39.
Tissue slices were taken at a depth of 2.7 mm and 2.75 mm and stained for GFAP and NeuN
respectively with hematoxylin counterstain. At 2.7 mm deep, the probes deviated slightly from the
cell body layer, whereas at 2.75 mm the probes cleanly hit the cell body layer. Control regions
were chosen to mimic the offset of each probe from the cell layer in order to control for the
Figure 38. Transverse, 50 µm thick hippocampal slice implanted with sham array that was stained with
GFAP to highlight astrocytes in brown, taken at a depth of 2.7 mm. Purple corresponds to hematoxylin
counter staining; dense purple strip is the DG of hippocampus. Array was removed prior to tissue
slicing. (a) Black arrows indicate locations of five probes of the array visible in the DG of the
hippocampus and (b) is the corresponding control area from the same tissue slice. (c & d) Color
thresholding of (a) and (b), respectively, to measure astrocytic density in 25 µm rings around the central
three probes and corresponding control regions, labelled A, B, and C. Rings reach 150 µm from probe
cross-sections. Scale bars are 100 µm.
69
heterogeneity in cellular distribution. Radial rings of 25 µm were drawn around probe cross-
sections and the corresponding control regions. Color thresholding was used to create a color mask
which allowed for the calculation of the concentration of astrocytes. Rectangular annuli fit to the
thickness of the cell body layer were used to calculate the concentration of neurons since neurons
largely lie within the cell layer. The concentration of the cell of interest was defined to be the
fraction of area in each ring occupied by the stain of interest.
Figure 38 shows histology images of the implanted and control regions with astrocytes
stained in brown with GFAP (first two panels). Panels (c) and (d) of Figure 38 have a superimposed
color mask in red over regions populated by astrocytes. Astrocytes, a glial support cell in the brain,
Figure 39. Transverse, 50 µm thick slice of hippocampus implanted with sham array that was stained
with NeuN to highlight neurons in brown, taken at 2.75 mm deep. Purple corresponds to hematoxylin
counter staining; dense purple strip is the DG of the hippocampus. Array was removed prior to tissue
slicing. (a) Black arrows indicate locations of five probes of the array visible in the DG of the
hippocampus and (b) is the corresponding control area from the same tissue slice. (c & d) Color
thresholding of (a) and (b), respectively, to measure density of neurons in 50 µm thick rectangles around
the central three probes and corresponding control regions. Scale bars are 100 µm.
70
are known to increase in number around areas of injury and attempt to wall-off foreign implants
through scar formation. An increase in astrocytic concentration around the implant site serves as
an immune marker for the severity of the immune response against the individual probes in a neural
array. An examination of these images reveals a small increase in astrocyte concentration at rings
close to the implant site compared to control, non-implanted regions. The concentration of
astrocytes in each radial ring surrounding three probes tracks, and the corresponding values from
a control region, are presented in
Table . The concentration of astrocytes within 100 µm of the implantation location was 8.5% ±
3.3% higher, on average, than their control counterparts. However, no significant increase of
astrocytes beyond those in the control region was noticed at distances greater than 100 µm from
the probe track (t test value of p=0.67).
A comparison of neuronal concentration (neurons were stained in brown with NeuN) is
presented in Figure 39 and presented graphically in Figure 40, in order to evaluate whether or not
neural death occurred near the implant site. Rectangular bins that were 50 µm wide whose lengths
matched the varying thickness of the cell body layer were used to compare the concentration of
Figure 40. Concentration of neurons in implanted and control rectangular bins. Green data points
correspond to neuronal concentration in the bins surrounding the implant site. Non-solid circles
correspond to the center of three probes in the array and naturally have the lowest concentration of
neurons. Solid blue horizontal line represents the average neuronal concentration across all control bins;
dashed lines and blue shaded region represent ± 1 standard deviation of the average concentration.
Neuronal concentration returns to control levels in between probes in the array.
Distance Along Hippocampal Layer (µm)
Concentration of Neurons (%)
71
neurons across the site of array implantation to a control region. A comparison between implanted
and control sites reveals that neuronal concentrations return to normal in the ~ 100 µm space
between adjacent probes in the array.
Table VIII. Density of astrocytes (area populated by astrocytes/area of ring) in rings
After all behavioral and neural data were recorded one animal which was implanted with
a functional Parylene array for six months was euthanized following steps descripted in Chapter
2.1.2. After the brain was removed from the animal, 50 µm thick brain slices were prepared with
a cryostat microtome. The rat brain was first kept in optimal cutting temperature (OCT) jell and
was kept frozen at -80 ℃ for two to four hours. Then the OCT block was fixed onto a cut stage
with more OCT at -20 ℃. Brain slices were then cut inside the -20 ℃ chamber and picked up with
positively charged glass slides. Slices taken at a depth around 3.15 mm were stained with GFAP.
The hippocampus at the intact side was used as control for the evaluation of activation of
astrocytes. The concentration of astrocytes in 75×200 µm rectangle bines on both left and right
side of the implanted location were measured and compared with similar regions of contralateral
brain tissue. The concentration of astrocytes increased at the implanted site and remained slightly
Dis. P_A C_ A P_B C_B P_C C_C Probe
average
Control
average
25 20.7 3.0 14.3 2.1 4.7 0.4 13.2± 8.1 1.8± 1.3
50 14.8 6.4 9.9 2.0 7.7 0.1 10.8± 3.6 2.8± 3.2
75 15.7 9.4 14.5 2.0 14.3 1.5 14.8± 0.7 4.3± 4.4
100 9.8 14.6 9.3 2.2 9.8 0.4 9.7± 0.3 5.7± 7.8
125 4.9 15.9 12.0 9.0 12.5 4.0 9.8± 4.2 9.6± 6.0
150 10.7 11.7 9.2 11.9 7.7 7.3 9.2± 1.5 10.3± 2.6
72
higher compared to control region at 75 µm away from the shank center, but the concentration of
astrocytes went back to control level at around 100 µm away from the Parylene shank.
Figure 41. GFAP stain at 6 months post-implantation.
Figure 42. The comparison of astrocytes concentration between the implanted side and control side.
73
The respond of neurons to the Parylene after the implantation of the Parylene array for six
months was evaluated through NeuN stain. Neurons were stained in brown and the concentration
of neuron along the cell body layer was measured and compared with control regions that were on
the same hemisphere but at least 500 µm away from the implant site. The decrease of neuron
concentration was observed around the implanted site but went back to control level at around 200
µm away from the implant site.
4.4 Comparison between Microwire Electrodes Arrays and the
Parylene Multi-Electrode Array
To further evaluated the performance of the Parylene multi-electrode array comparatively,
the quality of neural recordings obtained with the Parylene array was compared with that recorded
with microwire electrode arrays which are still the most commonly used conventional neural
recording tools. The quality of recordings collected under both acute preparations and chronic
Figure 43. NeuN stain of brain tissue collected from an animal implanted with the Parylene array for
six months. The comparison of neuron concentration between the implanted site with control region
along the cell body layer was shown with the plot on the right.
74
preparations were compared. In addition, the performance of the Parylene array was also compared
to neural recording probes developed by other groups.
4.4.1 Comparison of signals recorded under acute preparations
Recordings of 1 sec continuous data collected from the insertion surgery of both the
Parylene multi-electrode array and conventional microwire electrodes arrays were compared.
Recording quality include average peak-to-valley spike amplitudes, background noise, and SNR
were examined. SNR was calculated using the following formula:
SNR =
A
2 ∗ SD
n o is e
Figure 44. Spike activities recorded with the Parylene array during implantation. a) shows an example
of complex spikes recorded with the Parylene array (left) and one example of that recorded with
microwire arrays (left). Baseline noise levels were represented with red dash lines while spikes crossed
spike detecting threshold were highlighted with red dots. b) shows the comparison of average spike
amplitude, maximal spike amplitude and noise level between the Parylene array recordings to
microwire array recordings. * p=0.69, ** p=0.22, *** p=0.32. c) shows the comparison of average SNR
and maximal SNR between signals recorded with the Parylene array and microwire array.
Ave_SNR Max_SNR
0
2
4
6
8
10
12
14
SNR
Microwire
Parylene
Ave_Amp Max_Amp Noise
0
200
400
600
800
Amplitude (µV)
Microwire
Parylene
*
**
***
Parylene recordings Microwire recordings
Noise level
a)
b) c)
75
where A, the mean amplitude of spikes, is the average of peak-to-valley voltage of waveforms in
each 1 s trace. SDnoise is the standard deviation of the background noise. Each 1 s long recording
trace was low pass filtered to remove baseline drift. Spikes were defined as peaks with negative
amplitudes greater in magnitude than a threshold of two standard deviations of the filtered
recording trace. After removing spikes from the trace (the removed segments include both 400 µs
before and 1,200 µs after threshold crossing) the standard deviation of the baseline noise was
calculated. The maximum SNR achieved by one of electrodes on a Parylene array was 26.1. One-
second recordings collected from 96 electrode sites on those four implanted Parylene arrays were
compared to one-second recordings recorded with 70 individual microwires implanted to seven
animals. All recordings were collected from cell body layers of the rat hippocampus (including
both the CA1 and CA3). Figure 41. b) shows the mean peak-to-valley spike amplitudes and
background noise levels across all working electrodes recorded during surgery, while Figure 41.
c) displays the average SNR values achieved by each Parylene array. Each of these two figures
contains a comparison to data recorded with conventional microwire arrays.
Spike sorting was not conducted to recordings obtained during the implantation surgery. A
direct comparison of units yields of microwire arrays and Parylene arrays was not performed. The
Parylene multi-electrode array was designed to have four recording electrodes with close spacing
and targeted at a single cell body layer. Each four recording electrodes was considered as a
recording group. In total, there are sixteen recording groups on each Parylene array which matches
the number of microwires within the microwire electrodes array used in our experiment. A
comparison was made between the number of recording groups that recorded unitary activities
during the insertion surgery and immediately after chronic fixation to the number of microwires
recorded spike activities during implantation surgeries. From all eleven implanted Parylene multi-
electrode array, 107 recording groups (given an average of 9.73 ± 2.20 groups per animal) recorded
unitary activities during and immediately after the implantation surgery. For seven animals been
implanted with a sixteen channel uni-length microwire electrodes array, 9.00 ± 2.65 microwires
recorded spikes during the implantation surgery.
76
4.4.2 Comparison of signal quality under chronic preparations
The average noise level and average spike amplitude of an implanted Parylene multi-
electrode array was also compared to data recorded with a microwire electrode array over a period
of 45 days post-implantation. The representative example of the signal quality of this microwire
electrode arrays was illustrated in Figure 42. This particular microwire array represented one of
the most successful implantations of microwire arrays since up to 29 well-isolated units with a
highest SNR over 16 cross the period of one-month post-implantation were recorded with this
microwire array. Both spike amplitudes and baseline noise level of signals recorded with this
microwire array were stable over time. A comparison of the quality of chronic recordings obtained
from on animals implanted with the Parylene array representative example showed that the average
noise level and average spike amplitude was also stable and the quality of chronic recordings of
the Parylene array were comparable to the best performance of microwire electrode arrays (Figure
42).
4.4.3 Literature comparison
In this work, the study was not focused on the long-term evaluation of the performance of
microwire electrodes arrays. However, there are literatures available which the comprehensive
evaluation of the signal quality of recordings obtained with microwire arrays over extended time
periods was detailed. In one study [4] the failure modes of tungsten microwire electrodes arrays
were characterized over time periods ranged from 1 day to 260 days post-implantation of 16-
Figure 45. Preliminary comparison of average spike amplitude and baseline noise level between
microwire recordings and Parylene array recordings over one-month post implantation.
15 20 25 30 35 40 45
0
50
100
150
200
250
Days Post Implantation
Amplitude (uV)
Parylene Array
Average Amplitude
Noise Level
15 20 25 30 35 40 45
0
50
100
150
200
250
Days Post Implantation
Amplitude (uV)
Microwire Array
Average Amplitude
Noise Level
77
channel, 50 µm diameter, polyimide insulated tungsten microwire arrays to the rat’s cortex. At 260
days post-implantation, the array yield, defined as the percentage of microwires in an array that
were able to isolate single action potentials, dropped dramatically from around 65.00 ± 16.30 %
to 6.25 %. Compare to the one Parylene multi-electrode array which recorded high SNR signals at
362 days post-implantation, the array yield (percentage of recording groups that unitary activities
were recorded) at 273 days post-implantation was 37.50 % and the number was 25.00 % at 362
days post-implantation.
In another study, the tissue preservation and recording quality for chronic multi-electrode
implants was analyzed comprehensively [5]. Thirty-two channel, tungsten microwire arrays were
implanted to rat cortex and after six months brain slices from the implanted animal were stained
with GFAP. A comparison of GFAP stain between the implanted side and the contralateral, intact
side shown an activation of astrocytes at least 200 µm around the microwire.
Figure 46. GFAP stains of brain slices collected 1 month, 3 months and 6 months after the implantation
of tungsten microwire electrodes arrays. Scale bar was 100 µm. The astrocytes displayed a non-activated
morphology with cells presenting non-hypertrophic astrocytic activation, as reflected by the presence
of cells displaying hypertophic cell bodies and shorter and thicker processes (arrowheads).
78
4.5 Summary
In this chapter, the possibility of using Parylene-based multi-electrodes array for acute and
chronic neural recordings from multiple sub-regions of the rat hippocampus was explored. Straight
insertions of the Parylene array to both brain phantoms and the hippocampus of a rat verified the
feasibility of using a dissolvable brace to assist the insertion of the bare, flexible Parylene array.
Unitary activities were successfully recorded from the CA1 and CA3 cell body layers of the rat
hippocampus under both acute and chronic preparations. The recording of over 70 units
simultaneously from both sub-regions of the hippocampus from lightly anesthetized animals right
after implantation verified that the layout of electrodes placed on the Parylene array was conformal
to the curvature of the rat hippocampus. Chronic recordings result shown that both the noise level
and average spike amplitude was stable for a period of ten weeks post-implantation. Units with
high SNR recorded from at 362 days post-implantation implies the great potential of the
application of the Parylene array for long-term recordings from behaving animals. A comparison
of average spike amplitudes, noise levels and SNRs between signals recorded with microwire
electrodes arrays and Parylenn arrays shown that the signal quality of the Parylene array is
comparable to the conventional microwire electrodes array. Histological staining of tissue
surrounding the implant at one-month post-implantation reveals that astrocytic density increases
near the insertion site (as compared to control) but then returns to normal within 100 µm from the
implant site. Neuronal density was observed to decrease in the immediate vicinity of implant sites
but returned to control levels in the ~100 µm space between probes of the same Parylene array.
Minimal damage to the neuronal tissue was observed in tissue slices at the one-month mark. GFAP
stain of brain tissues collected after six months’ implantation of the Parylene array shown a slightly
increase of astrocytes at 90 µm away from the center of the implantation site. This result added
more support to previous reports that flexible implants may mitigates tissue response to foreign
implants.
4.6 References
[1] a Weltman, H. Xu, K. Scholten, T. W. Berger, D. Song, and E. Meng, “Deep Brain Targeting
Strategy for Bare Parylene Neural Probe Arrays,” Hilt. Head Conf., pp. 3–6, 2016.
[2] D. H. Szarowski et al., “Brain responses to micro-machined silicon devices,” Brain Res., 2003.
[3] K. C. Cheung, “Implantable microscale neural interfaces,” no. January, pp. 923–938, 2007.
79
[4] A. Prasad et al., “Comprehensive characterization and failure modes of tungsten microwire,” 2012.
[5] M. A. M. Freire et al., “Comprehensive Analysis of Tissue Preservation and Recording Quality from
Chronic Multielectrode Implants,” vol. 6, no. 11, pp. 23–27, 2011.
80
Chapter 5. Neuroscience Applications of Conformal , Multi-Electrode Arrays
Neuroscience Applications of Conformal,
Multi-Electrode Arrays
In the later 20th century, the discovery of the usage of electrodes to measure nerve impulses
from brain tissue led to new techniques to study brain functions. Over several decades, although
lesion studies and advanced imaging techniques contributed considerably to the understanding of
the brain’s structure and function, recordings of electrical signals from single neuron, small group
of neurons and large populations of neurons still contributed the most in the understanding of how
the brain process variety of sensory information and what roles it plays in dictating complex
behaviors. With access to neurons within different brain regions, the studies of what areas of the
brain are responsible for certain behaviors and actions have been made possible. However, to
investigate and understand how neural networks encode and decode sensory, behavioral and high-
level cognitive information, simultaneously interacting with groups of neurons in multiple brain
regions is necessary. With unitary activities recorded from multiple sub-regions of the
hippocampus, features of individual hippocampal regions can be compared and characterized.
More importantly, with simultaneous recordings from multiple hippocampal sub-regions, data
driven input-output models can be built to investigate the connectivity within the hippocampal
circuit and its contribution to the formation of new long-term memory. In addition, with stable
chronic recordings from behaving animals the evaluation of short-term and long-term plasticity
between hippocampal sub-regions can also be made possible. As one major application of
conformal, hippocampal, multi-electrode arrays, obtain long-term, stable recordings of unitary
activities simultaneously from multiple sub-regions of the hippocampal circuity will advance the
81
perception of functional connectivity within the hippocampal formation. Such experimental data
is also the foundation to develop mathematic models that can help to characterize the encode and
decode of memories and will provide valuable insights about the mechanisms underlying the
formation and retrieval of long-term memory.
Investigation of the correlation between neural activities and behaviors is the most
straightforward approach to study brain functions. Recordings of neural activities from behavioral
animals and human has built up the fundamental knowledge base of how the brain processes multi-
dimensional sensory inputs and how the brain control individual muscle to accomplish fine
movements [1]. Early studies from Foerster, Penfield and their collaborators revealed the
correlation between sensations from different regions of the body and localized activation of
cortical areas. These findings led to the generation of a receptive map of the somatosensory cortex
in the parietal lobe [2] which still been referred to frequently during early diagnose of brain regions
involved in traumatic brain injuries and stokes. Besides somatosensory input, vision is one of the
most important sensations that human and other mammalian rely on the most to percept rich
information from surrounding environments. Recordings from single cells in the cat visual cortex
revealed the orientation tuning, one of the major transformation in the organization of receptive
fields in the early visual pathway [3]. By mapping receptive fields of neurons in visual cortex,
orientation tuning in single neurons, columnar organization of ocular dominance and the effect of
visual deprivation to cortical development were studies [4][5]. These studies advanced the
understanding of how brain process large-scale sensory inputs and extract valuable information
from noisy environment. These studies also promoted the development of mathematical models
which can capture computational features existing in real visual cortex [6] which in turn brought
insights into the creation of prosthetic devices [7][8].
Different from sensory-perceptual of the brain and nervous system, it is relatively difficult
to reveal the direct correlation between certain brain region to high-level cognitive functions.
However, the hippocampus together with associated medial temporal-lobe structures, plays a
critical role in the formation of new long-term memory has been known with certainty ever since
Scoville and Milner’s report of profound amnesia in the patient H.M. following bilateral surgical
resection of hippocampal system in an attempt to control epileptic seizures [9]. This phenomenon
motivated intensively studies of the hippocampus over the past several decades. The clear
contribution of the hippocampus in the ability of human to accumulate information from
82
experiences and to express memory explicitly proved the critical role of the human hippocampus
plays in declarative memory system. Although it is still under debate whether other mammalians
have the capacity for conscious recollection, and the ability to distinction between ‘explicit’ and
‘implicit’ memory expression [10] like human, considerable supporting data is available to verify
the participation of the hippocampus in spatial information processing [11]–[14] and the
construction and modification of a relational representation [15][16] both in human and other
species.
5.1 Recording of Place Cells from Multiple, Hippocampal Sub-
Regions
Neural correlates of natural behaviors of free moving animals in an open field was first
studied with both uni-length, microwire electrode arrays and conformal, multi-electrode arrays.
Microwire electrode arrays which have been widely used to interface with neurons located in deep
brain regions, uni-length microwire electrode arrays, triple-region microwire arrays and the
conformal, Parylene multi-electrode array developed in this work were chronically implanted to
obtain unitary, neural activities from single hippocampal sub-region and simultaneously from
multiple sub-regions of the hippocampus of behaving rats.
Figure 47. Illustration of location-specified firing property of hippocampal neurons. a) shows a sketch
of the location of an animal in an environment and the correlation of a particular location to the
activation of a single unit in the hippocampus. b) shows one place cell recorded from the CA3 region
of the rat hippocampus. Moving traces of the animal were shown with black lines with every location
at which a spike was recorded represented with a red dot. The rate map of that particular unit was shown
on the right.
a) b)
Location-specified
receptive field
83
5.1.1 Hippocampal place cells
The most compelling evidence of the involvement of the hippocampus in spatial
information processing is the discovery of that some units in the CA1 region of the rat
hippocampus fired maximally when the animal was in a particular location. These cells were
named “place cells” by John O’Keefe and his colleagues, who first reported this phenomenon, and
that particular location the cell fired for was called the “receptive or place field” of the neuron [11].
Overtime, the discharge of hippocampal neurons driven by the location of the animal in space was
observed in all three sub-regions of the hippocampus [17] and cross all species including human
[18][19][20][21]. This added more supports to a direct participation of the hippocampus in the
processing of spatial information. The conspicuous behavioral correlates of hippocampal neurons
can help the understanding and derivation of other general principles of cortical networks. Place
cells of the rat hippocampus can be considered as a potentially model system to understand the
memory function of the brain. As descripted in Chapter 1.2, there are clear distinctions in neural
excitability, and anatomical connection patterns between hippocampal sub-regions which indicate
a unique biological and computational role performed by each sub-region. Recordings of activities
from place cells within each sub-region may offer considerable insight into how the brain creates
high-order cognitive representations of the world and how the hippocampus integrate multimodal
sensory input from the entorhinal cortex through the tri-synaptic pathway. Simultaneously
monitoring the activity of place cells in multiple sub-regions of the hippocampus can help to keep
in track the correlation changes between different regions which can help to identify the
transformation rule of the hippocampal circuit. This will also add more insights in how spatial
information is processed by the hierarchical, hippocampal neural network and provide an
opportunity for the evaluation of dynamic changes of functional connectivity between different
sub-regions.
The location-specific firing property of hippocampal neuros were recorded with both uni-
length microwire electrode arrays been implanted to different sub-regions of the rat hippocampus
and conformal, multi-electrode arrays that targeting at multiple hippocampal sub-regions. Neural
activities from multiple sub-regions of the hippocampus of free moving rats was recorded
simultaneously with these conformal, hippocampal multi-electrode arrays and behavioral data that
synchronized with the neural recordings was also recorded with cameras.
84
5.1.2 Recording from free moving animals
Animals used in behavioral experiments were introduced into the recording environment
daily to get familiar with the recording environment before the implantation of electrode arrays
(see Chapter 4.3.2 for details). The implantation of microwire electrodes arrays was detailed in
Chapter 2.1.3 and the procedure of Parylene array implantation was described in Chapter 4.2.2. In
short, stereotaxic surgeries were performed to open small cranial windows above the hippocampal
region. Both the dura layer and blood vessel layer were carefully removed to allow the insertion
of microwire electrodes arrays and the Parylene array. Both types of arrays were advanced into
brain tissue with small increments and neural activities were monitored throughout the insertion
procedure. After arrays were inserted to desired region, the window on the cranium was sealed
with dental cement. The dental cement also covered those anchor screws to chronically fix the
location of the implanted array and to mount the electrical package onto the animal’s skull.
All animals were given enough time (one to two weeks) to fully recover from the surgery.
After recovery, behavioral experiments were conducted in a controlled, highly simplified round
open field. The recording chamber is a solid black, round container with 76 cm in diameter and 56
cm in height. A rotatable, black platform with the same diameter as the container was placed on
the floor of the chamber. A moveable white cardboard with the same height as the recording
Figure 48. Diagram of the behavioral chamber. Sketch on the left shows the relative position and size
of the faraday cage, LED strip, black curtain and the solid black recording chamber. Sketch on the right
shows a zoom in view of the recording chamber together with the white cue card location at 6 o’clock
when viewed from overhead camera.
Commutator
LED Strip
Faraday Cage
Camera
Cue Card
76 cm
56 cm
85
chamber and covered 100
o
of the chamber’s wall was bent to fit the round inner surface and served
as a visual cue to the animal. The whole recording apparatus was isolated from external lab
environment with a black curtain which hang over the recording chamber. A customized faraday
cage (82 cm x 164 cm) was also built to protect the whole recording environment from power line
noise generated by other equipment in the lab. Illumination of the apparatus was provided by a
white LED strip which was shaped into a circle and placed on top of the faraday cage (Figure 48).
The animal’s location in the open field and neural activities were recorded following procedures
described in Chapter 4.3.2. In short, during each recording session, the animal’s moving traces
were recorded with an overhead camera while synchronized neural activities were recorded with
the 64-channel data acquisition system. Single units were manually sorted offline with principle
component analysis (PCA) method. The spatial distribution of firing rates of each single unit was
calculated with a data analysis software (Neuro Explorer, Nex Technologies, Madison, AL). Total
Figure 49. Example of complex spikes recorded from the CA1 and the CA3 region with the uni-length
microwire electrode array from behaving animals. a) shows waveforms of the unit and a continuous
view of complex spikes. Recording location was verified with histological slices. b) shows an example
of units fired complex spikes recorded from the CA3 region.
1000 µV
100 µs
250 µm
500 µV
10 ms
a)
b)
150 µV
10 ms
300 µV
100 µs
86
area of the open field was divided into 2 x 2 cm bins and the firing rate of the unit in each bin was
computed as the total number of spikes happened in the bin divided by the total time the animal
spent in that bin.
5.1.3 Place cells in individual sub-regions of the hippocampus
Neural activities together with the movement trace of the animal in the open field were
recorded after animals recovered from the implantation of uni-length microwire electrode arrays
to either the CA1 sub-region or the CA3 sub-region. Animals from which complex spikes were
observed during the implantation of the microwire electrode array and the recording location was
verified to be within the cell body layer based on histological results were included in this Chapter.
Four sixteen channel uni-length microwire electrode arrays were chronically implanted to the CA1
sub-region of four animals and six uni-length microwire electrode arrays were chronically
implanted to the CA3 sub-region of six animals. Unitary activities were recorded while the animal
was running a series of 15 min trials in the open field. Small food pellets were randomly through
onto the floor of the open field to encourage the animal exploring the entire environment. Fifty-
Figure 50. One example of multiple place fields recorded from the DG region. Picture on the left
shows the wave form of the unit (green). Top right photo shows the histological slices with the
location of the microwire circled out with red. Bottom right shows the place field of the unit.
100 µV
200 µs
DG
87
four well isolated single units with signal to noise ratio (SNR) greater than 3 were recorded from
the CA1 sub-region of four animals. Among those 54 units recorded from the CA1, 36 units (66.67
%) showed location-dependent receptive fields. 87 units with SNR greater than 3 were recorded
from the CA3 sub-region of six animals. Among all units recorded from the CA3, place fields were
observed from 54 % of them. Figure 49 shows a representative example of one unit recorded from
the CA1 region and one example of units recorded from the CA3 region while those animals were
exploring the open field.
Evaluation of histological brain slices collected from implanted animals showed that some
microwires were implanted to the DG region. Units with extremely low firing rates, one feature of
the granule cell in the DG, were recorded with some of those electrodes. Place fields of those
putative granule cells had multiple, grid-like receptive fields which were similar to those of
principal cells in the medial entorhinal cortex [22]. The medial entorhinal cortex is the direct upper
stream of the hippocampal circuit. Several other groups also reported [23][24] similar results that
some DG cells fired in multiple locations in an environment.
Figure 51. Example of place fields recorded from both the CA1 and the CA3 sub-regions of an animal
implanted with the Parylene multi-electrode array. Units shown location specified firing property were
high-lighted with light blue color. A view of the recording environment and the trace of the animal’s
location captured from the over-head camera was shown at bottom right.
88
5.1.4 Place cells simultaneously recorded from multiple sub-regions of the
hippocampus
After recovery, spike activities were also simultaneously recorded with the Parylene multi-
electrode array from multiple sub-regions of the rat hippocampus. The animal was running freely
for 5 mins sessions in the open field described in Chapter 5.1.2. Small food pellets were randomly
scattered onto the floor of the open field to encourage the animal to explore the entire chamber.
Units with location specified firing property were recorded from five animals implanted with a
Parylene multi-electrode array. From three of them, place cells were recorded simultaneously from
both the CA1 and the CA3 sub-regions of the hippocampus. A total of 44 units were recorded from
these three animals with 16 units from the CA1 sub-region and 28 units from the CA3 sub-region.
The ratio of place cells was 56.25 % and 42.86 % for the CA1 and the CA3 sub-regions
respectively.
5.2 Neural Response to Manipulations of Visual Stimulus
In addition to the clear association of the hippocampus to spatial information processing,
growing body of evidences imply that the hippocampal system is also essential for learning
complex relationships [15]. Successfully encoding of novel information and retrieval of old
information demand a continuous comparison of the current environment to the internal
representations of the environment to determine an appropriate mode of network operation. It has
been strongly suggested that the hippocampus performs as a key neural substrate for match-
mismatch process between the internal representation and the sensory information from the
environment [25][26]. However, most studies on the hippocampus was conducted in environments
that the animal already been familiar with. As a consequence, relatively little is known about how
hippocampal neural activity changes as animals learn about a novel environment [27]. In this work,
unitary activities were simultaneously recorded from the CA1 and the CA3 sub-regions of the
hippocampus with chronically implanted Parylene multi-electrode arrays. In addition, non-
stationary responds of place cells to the manipulation to the surrounding environment including
the introduction of novel visual cues and novel objects to the open field and the rotation of visual
cues were recorded.
89
5.2.1 Novel cue card
For two animals from which spike activities were recorded from both the CA1 and the CA3
sub-regions of the hippocampus with Parylene multi-electrode arrays, the responds of those
hippocampal neurons to a novel visual cue card been introduced into the recording chamber were
recorded. Recordings were first collected while the animal was running in the open field with no
visual cue card been presented. Both animals were allowed to run more than ten, 5 mins sessions
in the open field to ensure the animal was fully familiar with this cue card-free environment. Then
neural activities were recorded for three continuous 5 mins session after the white visual cue card
been added to the environment. Firing rate of each individual neurons at different places within
the open field was calculated for each 5 min session. Gradually formation of place fields was
observed from both the CA1 and the CA3 sub-region. Two 5 mins sessions were need for the
development of place fields in response to the novel visual cue been introduced into a familiar
environment. This observation agrees with the result shown in a study of the hippocampal
plasticity to novel environment exposure that hippocampal neurons required 5 to 6 mins of
experience to form stable spatial representations [27].
Figure 52. Example of the change of place fields to the added novel visual cue card. The gradually
formation of place fields was high-lighted with red rectangles.
90
5.2.2 Novel objects
For one particular animal, stable unitary activities were recorded over an extended time
period. With that animal, a test of neural responses to novel objects was conducted. After the
animal got fully familiar with the recording environment, two identical cylinders were introduced
into the recording chamber. For the first 5 min session immediately after the object been added,
no food pellet was provided. Overtime, the animal stopped exploring so food pellets were
randomly scattered onto the floor of the recording chamber to encourage the animal running around
inside the recording chamber for following sessions. The place field of hippocampal units was
calculated for each 5 min session. Units that rapidly re-mapped to new locations were observed
from both the CA1 and the CA3 sub-regions. The place field of one particular CA3 unit also
associated with the location of one object. Robust conjunctive item-place coding was observed
from a study that investigated activities of hippocampal neurons when the animal performed a
conditional discrimination task [16]. In short, activities of hippocampal neurons were recorded
Figure 53. Example of the responds of place fields to novel objects been introduced into an already
familiar recording environment. The firing of one CA3 neuron was associated with the location of one
object.
91
while the animal was learning which of two items is reworded depending on the environmental
context been presented. The result shown that a large percentage of hippocampal neurons, both in
the CA3 and in the CA1 sub-regions, developed representations of task-relevant, item-place
associations. The unit recorded in this work that shown item-associated firing property added more
support for the hypothesis that the hippocampus integrates “what” and “where” information in the
service of episodic memory [28].
5.2.3 Cue card rotation
The respond of place cells to the rotation of cue card were recorded with both uni-length
microwire electrode arrays implanted to individual hippocampal regions (either the CA1 or the
CA3) and with Parylene multi-electrode arrays simultaneous from both the CA1 and the CA3 sub-
regions. For animals been implanted with uni-length microwire electrode arrays, unitary activities
were recorded while the animal was running a series of 15 min trials in the open field with the cue
Figure 54. Examples of place fields recorded from the CA1 sub-region of the hippocampus with uni-
length microwire electrode arrays. a) shows place cells whose place fields followed the rotation of the
cue card while b) shows place cells whose place fields changed unpredictable by the cue card.
a) b)
Cue Card
Place fields followed the
rotation of the cue card
Remapping of place fields
92
card rotated either at 0
o
, 90
o
(either clockwise or counterclockwise) or 180
o
in the subsequent trial.
Post cue-card rotation, 70.0 % of those 36 units showed location-dependent receptive fields
recorded from the CA1 sub-region of four animals followed the rotation of the visual cue.
Remining cells with location-specified receptive fields either became silent after cue card rotation
or fired at different locations that was not clearly correlated to the position of the visual cue.
Examples of cells whose place fields was driven by the location of the cue card and cells whose
activities did not correlated to the cue card were shown in Figure 55.
Six animals were implanted with microwire electrode arrays targeted at the CA3 sub-
region. The cue card rotation experiment was conducted to four animals. Among those four
animals, 63 units were recorded. 32 units showed location-dependent receptive field. Post cue-card
rotation, place fields of 75 % of place cells followed the rotation of the visual cue. Similar as the
CA1 result, remining cells with location-specified receptive fields either became silent after cue
card rotation or fired at unpredictable locations post cue card rotation.
Cue card rotation experiments were also conduced to three animals implanted with the
Parylene mulit-electrode array. Animals were allowed to run more than five, 5 mins sessions in
Figure 55. Examples of place fields recorded from the CA3 sub-region of the hippocampus with uni-
length microwire electrode arrays. Place fields shows on the left shows place cells whose place fields
followed the rotation of the cue card while most right column of shows the place fields of place cells
that remapped after cue card rotation.
Cue Card
Place fields followed the
rotation of the cue card
Remapping of place fields
93
the open field with the cue card located at 3 o’clock to make sure the animal got fully familiar with
this setup. Then neural activities were recorded during following 5 mins sessions with the white
visual cue card been rotated to either 6 o’clock or 9 o’clock. A total of 35 units were recorded with
those three Parylene arrays from three animals with 23 units from the CA3 sub-region and 12 from
the CA1 sub-region. Six CA1 units and eight CA3 units were identified as place cell and 66.7 %
of CA1 place cells and 50.0 % CA3 place cells rotated their place fields with the rotation of the
cue card.
With place cells recorded from multiple sub-regions of the hippocampus, features of the
location specified firing properties of individual hippocampal regions can be compared and
characterized. Units recorded with uni-length microwire electrode arrays and Parylene multi-
electrode arrays were combined together. A total of 70 units from the CA1 and 110 units from the
CA3 sub-region were recorded. 64.3 % CA1 units shown place fields while animals were exploring
freely in the open field. 50.0 % CA3 units had place fields under the same behavioral experiment
setup. The ratio of place cells that rotated their place field with the white cue card was 69.0 % for
the CA1 sub-region and 70.0 % for the CA3 sub-region.
Figure 56. Responds of place cells to cue card rotation recorded simultaneously from both the CA1
and the CA3 sub-regions with the Parylene multi-electrode array.
94
More importantly, with simultaneous place cell recordings from multiple hippocampal
sub-regions, data driven input-output models can be built to investigate the connectivity within the
hippocampal circuit and its contribution to the processing of spatial information. In addition, with
stable chronic recordings from behaving animals the evaluation of short-term and long-term
plasticity between hippocampal sub-regions can also be made possible. A calculation of the cross-
correlation between one CA1 place cells and three CA3 units shown that the connectivity between
one neuron pair changed after the rotation of the cue card. A computational input-output model is
needed to comprehensively understand and characterize the connectivity change within the
hippocampal circuit after modifications to the surrounding environment is made. These multi-
regional recordings provided experimental data for the development of such mathematical models
and also added valuable insight in the understanding of the mechanism underly spatial and memory
information processing performed by the hippocampal neural network.
Figure 57. The ratio of place cells that rotated their place fields according to the location of a white
visual cue card in the CA1 sub-region and the CA3 sub-region respectively.
25 units
13 units
29 units
3 units
Place field rotated
Place field not rotated
No rotation experi-
ment conducted
No place field
CA1
CA3
55 units
28 units
12 units
15 units
Place field rotated
Place field not rotated
No rotation experi-
ment conducted
No place field
95
5.3 Summary
Multi-regional recordings were obtained with multi-electrode arrays while the animal was
exploring freely in the open field. Free foraging experiment involves minimal training and allows
the investigation of neural activities of the hippocampus under a close-to-nature condition. Unitary
activities recorded from multiple sub-regions of the hippocampus under such experimental setup
can help the understanding of how the hippocampal circuit process natural spatial information.
Place cells were recorded from multiple hippocampal sub-regions with both uni-length microwire
electrode arrays and the Parylene multi-electrode array. These recordings allow the investigation
of properties of place fields at different sub-regions of the hippocampus. Responds of unitary
activities recorded from multiple hippocampal sub-regions to novel visual stimuli introduced into
the familiar environment were also presented in this chapter. The gradually formation of place
fields with visual cue card added to the environment was observed from both the CA1 and CA3
sub-regions. These multi-regional recordings made the computational study of synaptic plasticity
within the tri-synaptic pathway possible and also provided experimental data for the development
non-stational mathematic models.
Figure 58. Cross-correlation between one CA1 place cell and three CA3 units simultaneously recorded
with a Parylene multi-electrode array during the cue card rotation experiment.
96
5.4 References
[1] H. Super and P. R. Roelfsema, “Chronic multiunit recordings in behaving animals: Advantages and
limitations,” Prog. Brain Res., vol. 147, no. SPEC. ISS., pp. 263–282, 2004.
[2] W. Penfield and T. Rasmussen, “The Cerebral Cortex of Man. A Clinical Study of Localization of
Function.pdf,” Academic Medicine, vol. 25. p. 375, 1950.
[3] D. H. Hubel and T. N. Wiesel, “Early exploration of the visual cortex,” Neuron, vol. 20, no. 3, pp.
401–412, 1998.
[4] D. H. Hubel and T. N. Wiesel, “Functional architecture of macaque monkey visual cortex,” R. Soc.
London, vol. 198, no. 1130, pp. 1–59, 1977.
[5] G. C. Deangelis et al., “Spatiotemporal organization of simple-cell receptive fields in the cat ’ s
striate cortex . II . Linearity of temporal and spatial summation Spatiotemporal Organization of
Simple-Cell Receptive Fields in the Cat ’ s Striate Cortex . II . Linearity of Temp,” vol. 69, no. 4,
pp. 1118–1135, 1993.
[6] D. J. Tolhurst and D. J. Heeger, “Comparison of contrast-normalization and threshold models of the
responses of simple cells in cat striate cortex,” Vis. Neurosci., vol. 14, no. 2, pp. 293–309, 1997.
[7] M. G. M. and J. P. G. W. H. Dobelle, “Artificial vision for the bline: electrical stimulation of visual
cortex offers hope for a functional prosthesis,” Science (80-. )., vol. 183, no. 4123, pp. 440–444,
1974.
[8] E. M. Schmidt, M. J. Bak, F. T. Hambrecht, C. V Kufta, D. K. O’Rourke, and P. Vallabhanath,
“Feasibility of a visual prosthesis for the blind based on intracortical microstimulation of the visual
cortex,” Brain, vol. 119, no. 5, pp. 507–22, 1996.
[9] W. B. Scoville and B. Milner, “LOSS OF RECENT MEMORY AFTER BILATERAL
HIPPOCAMPAL LESIONS,” J. Neurol. Neurosurg. Psychiatry, vol. 20, no. 1, pp. 11–21, 1957.
[10] H. Eichenbaum, “The hippocampus and mechanisms of declarative memory.,” Behav. Brain Res.,
vol. 103, no. 2, pp. 123–33, Sep. 1999.
[11] J. O’Keefe, “Place units in the hippocampus of the freely moving rat,” Exp. Neurol., vol. 51, no. 1,
pp. 78–109, 1976.
[12] F. P. Battaglia, G. R. Sutherland, and B. L. McNaughton, “Local sensory cues and place cell
directionality: additional evidence of prospective coding in the hippocampus.,” J. Neurosci., vol. 24,
no. 19, pp. 4541–50, May 2004.
[13] M. B. Moser and E. I. Moser, “Distributed encoding and retrieval of spatial memory in the
hippocampus.,” J. Neurosci., vol. 18, no. 18, pp. 7535–42, Sep. 1998.
[14] I. Lee, M. R. Hunsaker, and R. P. Kesner, “The role of hippocampal subregions in detecting spatial
novelty.,” Behav. Neurosci., vol. 119, no. 1, pp. 145–53, Feb. 2005.
97
[15] N. J. Fortin, K. L. Agster, and H. B. Eichenbaum, “Critical role of the hippocampus in memory for
sequences of events.,” Nat. Neurosci., vol. 5, no. 5, pp. 458–62, May 2002.
[16] R. W. Komorowski, J. R. Manns, and H. Eichenbaum, “Robust conjunctive item-place coding by
hippocampal neurons parallels learning what happens where.,” J. Neurosci., vol. 29, no. 31, pp.
9918–29, Aug. 2009.
[17] I. Lee, D. Yoganarasimha, G. Rao, and J. J. Knierim, “Comparison of population coherence of place
cells in hippocampal subfields CA1 and CA3.,” Nature, vol. 430, no. 6998, pp. 456–459, 2004.
[18] A. D. Ekstrom et al., “Cellular networks underlying human spatial navigation,” Nature, vol. 425,
no. 6954, pp. 184–188, 2003.
[19] J. O’Keefe, N. Burgess, J. G. Donnett, K. J. Jeffery, and E. a Maguire, “Place cells, navigational
accuracy, and the human hippocampus.,” Philos. Trans. R. Soc. Lond. B. Biol. Sci., vol. 353, no.
1373, pp. 1333–40, Aug. 1998.
[20] N. Burgess, E. A. Maguire, and J. O’Keefe, “The human hippocampus and spatial and episodic
memory,” Neuron. 2002.
[21] A. D. Ekstrom et al., “Cellular networks underlying human spatial navigation,” Nature, 2003.
[22] M. B. Moser, M. Fyhn, S. Molden, M. P. Witter, and E. I. Moser, “Spatial representation in the
entorhinal cortex.,” Science (80-. )., vol. 305, no. 5688, pp. 1258–64, 2004.
[23] J. K. Leutgeb, S. Leutgeb, M.-B. Moser, and E. I. Moser, “Pattern separation in the dentate gyrus
and CA3 of the hippocampus.,” Science, vol. 315, no. 5814, pp. 961–6, Feb. 2007.
[24] J. P. Neunuebel and J. J. Knierim, “Spatial firing correlates of physiologically distinct cell types of
the rat dentate gyrus.,” J. Neurosci., vol. 32, no. 11, pp. 3848–58, Mar. 2012.
[25] R. E. Hampson, J. D. Simeral, and S. A. Deadwyler, “Distribution of spatial and nonspatial
information in dorsal hippocampus,” Nature, 1999.
[26] J. J. Knierim, H. S. Kudrimoti, and B. L. Mcnaughton, “Interactions between idiothetic cues and
external landmarks in the control of place cells and head direction cells,” J. Neurophysiol., 1998.
[27] L. M. Frank, G. B. Stanley, and E. N. Brown, “Hippocampal Plasticity across Multiple Days of
Exposure to Novel Environments,” vol. 24, no. 35, pp. 7681–7689, 2004.
[28] H. Eichenbaum, A. P. Yonelinas, and C. Ranganath, “The Medial Temporal Lobe and Recognition
Memory,” Annu. Rev. Neurosci., 2007.
98
Chapter 6. Conclusion and Future Directions
Conclusion and Future Directions
Ever since the first recording of electrical signals from the central nervous system in the
early 1950s, billions of little action potentials generated everywhere in the brain and every moment
as long as the object is alive have fascinated generations of neuroscientists. A wealth of data about
the function and property of a certain brain region was obtained with extracellular, single-unit
recording methods [1]. The advancement in neural recording techniques together with the
development in surgical techniques, behavioral experiments and computational mathematics, keep
pushing the understanding of mechanisms underlying complex brain functions forward. With
knowledges of the nervous system accumulated over the past several decades, it is clear that
simultaneous monitoring of the activity of populations of individual neurons, distributed across
multiple, interconnected cortical and subcortical structures that define particular neural circuits for
long periods of time is the foundation to the understanding of high-level cognitive functions such
as the memory.
6.1 Development of Conformal, Multi-regional Neural Interfaces
To understand how the hippocampus accomplishes the formation of long-term memories,
simultaneous recording of unitary activities from multiple hippocampal sub-regions over long-
term from behaving animals is crucial. Towards accomplishing these goals, conformal, multi-
electrode arrays that targeting at multiple sub-regions of the rat hippocampus was developed. First,
a 3-D microwire electrodes array that having microwires targeting at cell body layers located at
different vertical depth was designed. Unitary activities simultaneously recorded from all three
sub-regions of the rat hippocampus with chronically implanted triple-region microwire electrodes
99
arrays verified the possibility of using conformal, microwire electrodes arrays for multi-regional
recordings from the hippocampus. To further increase channel counts and reduce long-term tissue
response, the possibility of using Parylene C as structural materials for penetrating, multi-electrode
arrays was also explored in this work. A novel, Parylene multi-electrode array with electrodes
placement conformal to the curvature of the rat hippocampus was designed, fabricated and
evaluated both acutely and chronically. The recording of unitary activities from multiple
hippocampal sub-regions with the Parylene array under both acute and chronic preparations
verified that the conformal design matched well with the anatomical structure of the rat
hippocampus. Units with high SNRs recorded with a Parylene array at 362 days post-implantation
implies the great potential of the application of Parylene multi-electrode arrays for ultra-long-term
recordings from behaving animals.
The methodology of develop multi-electrode arrays according to the main anatomical
contours of a specific brain region and configure the layout of electrodes to targeting at different
Figure 59. An illustrative diagram of the design of a Parylene multi-electrode array that can be
potentially used for simultaneous recording from the somatosensory cortex, motor cortex and the
hippocampus. a) shows the design of the Parylene array with different shank length. b) shows the six-
layers structure of the cortex.
Layer I
Layer II
Layer III
Layer IV
Layer V
Layer VI
500 µm
2 mm
5 mm
20 µm
110 µm
Motor
Somatosensory
Hippocampus
Cortex
Rat Brain
Layers of Cortex
Somatosensory
Cortex
Whisker Evoked Response
Somatosensory Cortex Mapping
Somatosensory
Cortex
a) b)
100
brain regions associate to certain neural circuit is not just limited to the development of
hippocampal arrays. The strategy used for the development of conformal, hippocampal arrays can
be expanded to other cortical and sub-cortical regions. For example, by placing recording
electrodes along the shank of polymer probe arrays, neural activities from multiple layers of the
cortex can be recorded. With polymer probe shanks with different lengths been integrated into
polymer shank arrays, simultaneously recording from somatosensory cortex, motor cortex and the
hippocampus can be made possible.
6.2 Multi-regional Recordings from Behaving Animals
Through the investigation of the correlation between neural activities and behaviors of an
animal the mechanism underlying complex brain functions can be revealed. A set of fundamental
behavioral experiments were conducted. Free foraging experiments provides a paradigm for the
investigation of neural activities while an animal is behaving normally. Unitary activities from
multiple hippocampal sub-regions were recorded while the animal was exploring freely in a close-
to-nature environment. In addition, the responds of units from both the CA1 and CA3 to
manipulations to visual stimuli and novel objects were recorded. These multi-regional,
experimental data can provide valuable insight into the study of synaptic plasticity within the
hippocampal circuit. Simultaneous recordings of unitary activities from populations of neurons
within the tri-synaptic pathway are also the foundation to develop computation, non-stationary
model to investigate the successive transformations of memory information that are performed by
the hippocampal neural network [2]. With such a mathematic model that can mimic the biological
function of the hippocampus, together with properly developed neural interfacing devices, a fully
101
implantable hippocampal prosthesis that can restore the memory function of a patient can be
achieved.
Figure 60. MIMO nonlinear dynamical model consists of a series of multiple- input, single-output
(MISO) models of spiking neurons that are equivalent to generalized Laguerre-Volterra models.
102
6.3 References
[1] D. R. Humphrey and E. M. Schmidt, “Extracellular Single-Unit Recording Methods *,” 1977.
[2] D. Song, B. Robinson, R. Hampson, V. Marmarelis, S. Deadwyler, and T. Berger, Sparse large-
scale nonlinear dynamical modeling of human hippocampus for memory prostheses, vol. PP, no. 99.
2016.
Abstract (if available)
Abstract
To understand how the formation of long-term memory arises from the interaction between hippocampal neurons, it is necessary to have access to individual sub-regions of the hippocampal circuit at the single-neuron level. The success of a hippocampal prosthetic device also relies highly upon its ability to attain resolvable unitary activities from populations of neurons in multiple regions of the hippocampal networks over long-term. Multi-electrode arrays are the primary interfacing device used to collect unitary activities from the hippocampus and are also critical for closed-loop hippocampal prosthesis. To achieve multi-regional recording from the hippocampus, multi-electrode arrays must be specifically designed to accommodate the anatomy of the hippocampus. In this work, conformal multi-electrode arrays that matched the anatomical curvature of the rat hippocampus were designed, developed and chronically implanted. Optimizations of the standard microwire electrode array were made to develop a triple-region, microwire-based multi-electrode array with varied wire lengths to target at all three sub-regions of the rat hippocampus. To minimize the foreign body response to a rigid, penetrating neural implant, a novel, Parylene-based multi-electrode array with a reduced elastic modulus was also developed. This Parylene-based multi-electrode array was designed and fabricated to have 64 electrodes positioned to match the anatomy of the rat hippocampus and allowed for simultaneous recordings from two cell body layers of the tri-synaptic pathway. The performance of the Parylene array was evaluated both acutely and chronically. On average, over 70 units (n=7) were simultaneously recorded from both the CA1 and CA3 sub-regions under acute preparations. Chronically, unitary activities with good signal-to-noise ratio were stably recorded from multiple sub-regions for at least 70 days (n=4). The location-specific firing properties of hippocampal neurons and non-stationary responds of place cells to manipulations of visual stimuli to the surrounding environment were simultaneously recorded from multiple sub-regions of the hippocampus with these conformal, hippocampal multi-electrode arrays while the animal was behaving freely in an open field. Such multi-regional recordings are the foundation to develop mathematical models that can fully characterize the connectivity between hippocampal sub-regions and also provided valuable experimental data for the computational analysis of synaptic plasticity within the hippocampal circuit, which will expand the understanding of memory encoding and decoding in the hippocampus to a system level.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Penetrating parylene neural probe array for dense, in vivo, chronic recordings
PDF
Next generation neural interfaces for neural modulation and prosthesis
PDF
Experimental and computational models for seizure prediction
PDF
Functional consequences of network architecture in rat hippocampus: a computational study
PDF
Simulating electrical stimulation and recording in a multi-scale model of the hippocampus
PDF
Decoding memory from spatio-temporal patterns of neuronal spikes
PDF
Intensity discrimination in single and multi-electrode patterns in cochlear implants
PDF
The electrochemical evaluation of Parylene-based electrodes for neural applications
PDF
Nonlinear modeling of causal interrelationships in neuronal ensembles: an application to the rat hippocampus
PDF
Design and application of a C-shaped miniaturized coil for transcranial magnetic stimulation in rodents
PDF
Nonlinear models for hippocampal synapses for multi-scale modeling
PDF
Strategies for improving mechanical and biochemical interfaces between medical implants and tissue
PDF
Development of implantable Parylene-based MEMS technologies for cortical applications
PDF
Investigation of the electrode-tissue interface of retinal prostheses
PDF
Fabrication and packaging of three-dimensional Parylene C neural interfaces
PDF
On the electrophysiology of multielectrode recordings of the basal ganglia and thalamus to improve DBS therapy for children with secondary dystonia
PDF
Parylene C bioMEMS for implantable devices with electrochemical interfaces
PDF
A double-layer multi-resolution classification model for decoding time-frequency features of hippocampal local field potential
PDF
Data-driven modeling of the hippocampus & the design of neurostimulation patterns to abate seizures
PDF
A million-plus neuron model of the hippocampal dentate gyrus: role of topography, inhibitory interneurons, and excitatory associational circuitry in determining spatio-temporal dynamics of granul...
Asset Metadata
Creator
Xu, Huijing
(author)
Core Title
Multi-region recordings from the hippocampus with conformal multi-electrode arrays
School
Viterbi School of Engineering
Degree
Doctor of Philosophy
Degree Program
Biomedical Engineering
Publication Date
12/05/2019
Defense Date
12/05/2019
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
conformal multi-electrode array,hippocampus,multi-region recording,OAI-PMH Harvest,Parylene
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Song, Dong (
committee chair
), Berger, Theodore (
committee member
), Hashemi, Hossein (
committee member
), Meng, Ellis (
committee member
)
Creator Email
hdbmcat2009@gmail.com,huijingx@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c89-245120
Unique identifier
UC11674974
Identifier
etd-XuHuijing-8004.pdf (filename),usctheses-c89-245120 (legacy record id)
Legacy Identifier
etd-XuHuijing-8004.pdf
Dmrecord
245120
Document Type
Dissertation
Rights
Xu, Huijing
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
conformal multi-electrode array
hippocampus
multi-region recording
Parylene