Development of optical instrumentation and signal analysis for biomedical applications

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Request accessible transcript

Transcript (if available)

Content
Copyright 2018
Development of Optical
Instrumentation and Signal Analysis
for Biomedical Applications

By

Samantha Elizabeth McBirney

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

August 2018

ii
Dedication

To my husband, my mother, my father, my brother, my sister, and Granma.

iii
Acknowledgements

“Success consists of going from failure to failure without loss of enthusiasm.”
~ Winston Churchill

I have learned the true meaning of “It takes a village” in the past several years. To
say my work would not have been possible without the support, love, and guidance of
those I am about to address is a vast understatement. I am forever indebted to each and
every one of you – know that, if ever I have the chance to do for you what you have done for
me, I would be honored to do so.
Of course, I would like to thank my advisor, Professor Andrea Armani, who accepted
me into her research group in early 2012. Choosing a research advisor is arguably one of
the most important decisions of one’s life, and I am truly fortunate to say that, no matter
how many times I may be given the chance to choose another, I would choose her time and
time again. Her support, her generosity, and her sincerity have always been far above and
beyond what I ever could have reasonably expected from my advisor. Dr. Armani has
provided me with endless opportunities that have made an indelible impact on my time as
a graduate student. Not only has Dr. Armani impressed me with her academic prowess, but
I have also had the fortune of having a much deeper relationship with her, opening up to
her on a personal level in times of great need. For that, in particular, I will be forever

iv
grateful. I cannot thank her enough for believing in me, inspiring me, and guiding me
through the past several years.
The research contained in this dissertation took place across several labs. I was
incredibly fortunate to have had the opportunity to work with Dr. Annie Wong-Beringer at
the USC School of Pharmacy, Dr. Joseph Kinetz at UC San Diego, and Dr. Hossein Ameri at
the USC Roski Eye Institute, all of whom were gracious enough to foster collaborations by
opening up their lab spaces to me. In the last couple of months, I began a collaboration with
Dr. Mary Galinski at Emory University, and I cannot thank her enough for being willing to
start working together before our proposal is even funded, giving me the ability to take the
malaria diagnostic to a place I never expected prior to graduation.
I never would have made it to USC without the life-changing experiences had as an
undergraduate student at the California Institute of Technology and UC Berkeley. Both of
these schools changed my life, in more ways than I can count, and I have become grateful
for all experiences had, even if not particularly grateful in the moment. These experiences
molded me into the person I am today, so, in short, thank you. I am unbelievably proud to
be a Golden Bear and a Trojan for life.
I would like to thank all of the incredible students I have had the opportunity to
mentor and work with over the years – particularly, Lexie Scholtz, Bernard Chen, Kaitlyn
Olah, and Kristina Kaypaghian. Each of these students is exceptionally gifted, and the
malaria work would not have been possible without their help across so many disciplines.
You each have such promising futures. I cannot wait to see where life takes you, and I can
only hope to be a witness to your continued successes.

v
I would also like to thank all of my lab mates at USC. Dr. Jason Gamba – you were the
very first person I met in the Armani Lab, and you set the tone, showing me what a warm,
welcoming environment this lab would be. That first meeting with you was invaluable. My
fellow lab mates created, and continue to create, a truly exceptional work environment,
making it a pleasure to work here. Not everyone is so fortunate to have such a wonderful
experience during his or her Ph.D., so thank you all for contributing so immensely.
Nearing the end, I would like to thank Mom, Dad, my brother, J.D., and my sister,
Serena. Where do I begin? To say I am blessed to have you all as my family does not come
even close to reality. You are unable to choose your family, but, thankfully, I would pick
each of you out of a crowd of billions. Thank you, Mom and Dad, for encouraging me to
pursue my passions as a child, for teaching me the value of a work ethic, for believing in me,
for standing up for me before I was able to stand up for myself. You’ve taught me to find the
beauty in the rollercoaster, and I would choose that rollercoaster over the carousel any day
of the week with the two of you by my side. J.D., I love you. Your support and confidence in
me when I was young played such an important role in my development. I hope I can make
you proud. Serena, you are one of the strongest women I know. To watch you grow into the
person you are today has been an honor. You have been through so much, and, despite
whatever obstacles life throws at you, you persist. I am unbelievably proud to have you as
my sister. I would also like to thank Granma. Simply put, you were a better person than any
of us deserved, and I miss you every single day. Granpa, although I never had the chance to
meet you, I know you live on in me. You have all taught me invaluable lessons that I will
carry with me wherever I go.

vi
Most of all, I would like to thank my incredible husband, Collin. You are my best
friend, an extraordinary teammate, someone who can always make me laugh. Marrying you
is truly the best decision I have ever made and ever will make. It is your unfailing love and
support that have pulled me through some of my darkest times. You are there to celebrate
the highs, and you are there by my side to help me out of the lows. Thank you for uprooting
your life in the Bay Area to move to LA for me to pursue my Ph.D. We moved down here as
barely more than children, with nothing to our name, and we now have a beautiful life we
have built here over the past almost-six years. Thank you for being the world’s greatest
puppy and kitty dad. Thank you for always being ready to try something new. I cannot wait
to see where this next chapter of life takes us. You are my rock, you are my world, and I
love you, der.

vii
Table of Contents
DEDICATION II
ACKNOWLEDGEMENTS III
TABLE OF CONTENTS VII
LIST OF TABLES X
LIST OF FIGURES XI
ABSTRACT XVI
CHAPTER 1: INTRODUCTION AND OVERVIEW 1
1.1 MOTIVATION 1
1.2 CHAPTER OVERVIEW 3
1.3 REFERENCES 5
CHAPTER 2: BACKGROUND AND RELATED WORK 7
2.1 SENSOR THEORY 7
2.1.1 OVERVIEW 7
2.1.2 RECEIVER OPERATING CHARACTERISTIC (ROC) CURVES 8
2.1.3 RESPONSE TIME 10
2.1.4 OTHER SENSOR ATTRIBUTES 12
2.1.5 THE SPIDER CHART 14
2.2 SPECTROSCOPY 16
2.2.1 OVERVIEW 16
2.2.2 ABSORPTION SPECTROSCOPY 17
2.2.3 DIFFERENTIAL ABSORPTION SPECTROSCOPY 21
2.3 LASERS AND BTBI 22
2.3.1 TRAUMATIC BRAIN INJURY: AN OVERVIEW 22
2.3.2 LASERS: AN OVERVIEW 25
2.3.3 USING LASERS TO STUDY TRAUMATIC BRAIN INJURY 26
2.4 REFERENCES 28
CHAPTER 3: WAVELENGTH-NORMALIZED SPECTROSCOPIC ANALYSIS OF BACTERIA
GROWTH RATES 31
3.1 SIGNIFICANCE AND BACKGROUND 31
3.1.1 THE STAGES OF MICROBIAL GROWTH 31
3.1.2 CURRENT METHODS OF MEASURING CELL GROWTH 32
3.1.3 BACTERIA OF INTEREST 34
3.2 EXPERIMENTAL METHODS 36
3.2.1 PREPARATION OF BACTERIAL CULTURES 36
3.2.2 BACTERIAL CULTURE DILUTION OPTIMIZATION 36
3.2.3 COLONY COUNTING 43
3.2.4 OD600 MEASUREMENTS 44

viii
3.2.5 MULTI-WAVELENGTH DIFFERENTIAL ABSORPTION SPECTROSCOPY 44
3.3 RESULTS AND DISCUSSION 46
3.4 CONCLUSIONS 50
3.5 REFERENCES 51
CHAPTER 4: PORTABLE DIAGNOSTIC FOR MALARIA DETECTION IN LOW-RESOURCE
SETTINGS – SYSTEM OPTIMIZATION 54
4.1 SIGNIFICANCE AND BACKGROUND 55
4.1.1 CURRENT METHODS OF DIAGNOSIS 56
4.1.2 USING HEMOZOIN AS AN INDICATOR OF INFECTION 56
4.2 EXPERIMENTAL METHODS 59
4.2.1 SYNTHESIS OF FE3O4 NANOPARTICLES BY CHEMICAL CO-PRECIPITATION 59
4.2.2 SYNTHESIS OF Β-HEMATIN 59
4.2.3 DEVICE EVOLUTION 61
4.2.4 MATHEMATICAL MODELING 68
4.2.5 VERIFICATION OF DEVICE 68
4.3 RESULTS AND DISCUSSION 72
4.3.1 VERSIONS 1 AND 2: OPTIMIZATION OF MAGNETIC STRENGTH AND TIME 72
4.3.2 VERSION 3: MALARIA-INFECTED BLOOD RESULTS 77
4.3.3 VERSION 5: FE3O4 NANOPARTICLE RESULTS 79
4.3.4 VERSION 5: MALARIA-INFECTED BLOOD RESULTS 82
4.3.5 VERSION 5: Β-HEMATIN RESULTS 83
4.4 CONCLUSIONS 88
4.5 REFERENCES 88
CHAPTER 5: PORTABLE DIAGNOSTIC FOR MALARIA DETECTION IN LOW-RESOURCE
SETTINGS – TESTED IN WHOLE RABBIT BLOOD 92
5.1 MALARIA’S MILITARY IMPACT 92
5.2 EXPERIMENTAL METHODS 94
5.2.1 SYNTHESIS OF Β-HEMATIN BY ACETATE-MEDIATED PRODUCTION 94
5.2.2 DEVICE SET-UP 94
5.2.3 MATHEMATICAL MODELING 96
5.2.4 VERIFICATION OF DEVICE 96
5.3 RESULTS AND DISCUSSION 97
5.4 CONCLUSIONS 100
5.5 REFERENCES 101
CHAPTER 6: PORTABLE DIAGNOSTIC FOR MALARIA DETECTION IN LOW-RESOURCE
SETTINGS – TESTED WITH MALARIA-INFECTED NONHUMAN PRIMATE SAMPLES 102
6.1 EXPERIMENTAL METHODS 102
6.1.1 BLOOD SAMPLE PREPARATION 102
6.1.2 DEVICE SET-UP 103
6.1.3 MATHEMATICAL MODELING 104
6.1.4 VERIFICATION OF DEVICE 105
6.2 RESULTS AND DISCUSSION 105
6.3 CONCLUSIONS 107
6.4 REFERENCES 108

ix
CHAPTER 7: ELUCIDATING THE MECHANISM OF BLAST-INDUCED TRAUMATIC BRAIN INJURY
109
7.1 SIGNIFICANCE AND BACKGROUND 109
7.1.1 BLAST-INDUCED TRAUMATIC BRAIN INJURY 110
7.1.2 BLAST ENERGY TRANSFER TO THE BRAIN 111
7.2 EXPERIMENTAL METHODS 113
7.2.1 NEURAL SLICE PREPARATION 113
7.2.2 NEURAL SLICE TRANSPORTATION 114
7.2.3 BLAST-INDUCTION SET-UP 115
7.2.4 VERIFICATION OF SET-UP 117
7.2.5 EXAMINATION OF CELLULAR PROCESSES 118
7.3 RESULTS AND DISCUSSION 118
7.4 CONCLUSIONS 121
7.5 REFERENCES 121
CHAPTER 8: CONCLUSIONS AND FUTURE WORK 123
8.1 CONCLUSIONS 123
8.1.1 BACTERIA PROJECT 123
8.1.2 MALARIA PROJECT 124
8.1.3 BTBI PROJECT 125
8.2 FUTURE WORK FOR MALARIA PROJECT 126
8.2.1 MAGNETIC TRAPS 126
8.2.2 POLARIZED LIGHT SCATTERING SPECTROSCOPY 128
8.2.3 DEVICE CONFIGURATIONS 130
8.2.4 SAMPLE TESTING 131
8.3 FUTURE WORK FOR BTBI PROJECT 132
8.4 REFERENCES 134

x
List of Tables
Table 2-1: Factors known to affect sensor utility .................................................................................... 14

Table 4-1: Specifications of magnets tested. ............................................................................................. 69

xi
List of Figures
Figure 2-1: Sample receiver operating characteristic (ROC) curves. (a) Curves for different
levels of detection confidence in a single environment, with a fixed response time. (b)
Curves for different environments at a fixed detection confidence and fixed response
time. Typically, an urban environment has much greater noise than does a desert
environment. Therefore, at a fixed sensitivity, a sensor operating in an urban setting
will have a higher false positive rate than a sensor operating in a desert setting. ......... 10

Figure 2-2: The sensor response time is the time interval between when the agent
concentration reaches the sensor’s specified sensitivity (or detection threshold) and
the time that the sensor issues a detection signal. ........................................................................ 11

Figure 2-3: (a) Rise-time definition, where Tr represents the response time, and (b) fall-
time definition, where Td represents the response time. ........................................................... 11

Figure 2-4: Example of a spider chart, allowing designers to quickly determine areas that
can be improved or identify tradeoffs that should be made to optimize sensor function.
............................................................................................................................................................................. 16

Figure 2-5: Electron excitation. (a) Electron orbiting the nucleus of an atom. A photon is
absorbed by the sample, and (b) the particles are then excited to a higher-energy state.
............................................................................................................................................................................. 18

Figure 2-6: Schematic showing the attenuation of radiation passing through a sample. I0 is
the incident light from the source, and I is the light transmitted through the sample to
the detector (in this case, the eye). Due to absorbance, I < I0. ................................................. 18

Figure 2-7: Plot of transmittance (in dark blue) and absorbance (in pink) as a function of
wavelength (λ, nm) for CrCo0.6Ni0.4FeO4 annealed at 1000°C.
13
............................................. 20

Figure 2-8: Simple physics of a blast wave. ............................................................................................... 24

Figure 2-9: Electron excitation and emission. (a) Electron orbiting the nucleus of an atom.
The lasing medium is pumped, releasing energy (E = hc/λ) that is absorbed by the
electron. (b) The electron is then excited to a higher-energy state. (c) The electron
subsequently releases this energy in the form of photons, (d) thereby returning to its
original ground state. ................................................................................................................................. 26

Figure 3-1: Rendering of (a) S. aureus and (b) P. aeruginosa. Adapted with permission from
reference 25, The Optical Society. ........................................................................................................ 35

Figure 3-2: Optical images of the four S. aureus bacteria strain cultures grown on blood agar
plates at different dilutions. (a) All strains plated at 10
-6
dilution. (b) KH38 and LAC91

xii
plated at 10
-10
dilution, HH36 at a 10
-8
dilution, and HH49 at a 10
-9
dilution. (c) All
strains plated at a 10
-10
dilution. Adapted with permission from reference 25, The
Optical Society. .............................................................................................................................................. 37

Figure 3-3: Dilutions 1:1, 1:5, 1:10, 1:15, and 1:20 tested at a single time point for (a) HH36
strain, (b) HH49 strain, (c) KH38 strain, and (d) LAC91 strain. .............................................. 39

Figure 3-4: Dilutions 1:1, 1:3, 1:9, 1:27, and 1:81 tested every 30 minutes for (a) HH36
strain, (b) HH49 strain, (c) KH38 strain, and (d) LAC91 strain. .............................................. 40

Figure 3-5: Dilution 1:27 tested every 15 minutes over the course of 24 hours for (a) HH36
strain, (b) HH49 strain, (c) KH38 strain, and (d) LAC91 strain. .............................................. 41

Figure 3-6: Dilution 1:40 tested every 15 minutes over the course of 24 hours for (a) HH36
strain, (b) HH49 strain, (c) KH38 strain, (d) LAC91 strain, and (e) PA01. ......................... 42

Figure 3-7: (a) UV-Vis spectra for S. aureus strain HH49. Arrows indicate wavelengths
chosen for subsequent analysis. (b) UV-Vis measurements at wavelengths of interest
plotted at select intervals over time. (c) OD600 measurements plotted at select
intervals over time. Error bars are shown for (b) and (c); however, the error is so
small that the error bars are smaller than the symbols. Reprinted with permission
from reference 25, The Optical Society. ............................................................................................. 47

Figure 3-8: Change in absorbance over time for seven wavelengths, including 600 nm,
exhibiting significant or little change. The specific bacteria plotted are: (a) S. aureus,
HH36 strain, (b) S. aureus, HH49 strain, (c) S. aureus, KH38 strain, (d) S. aureus, LAC91
strain, and (e) P. aeruginosa, PA01 strain. Error bars are shown on each plot; however,
the error is so small that the error bars are smaller than the symbols. Reprinted with
permission from reference 25, The Optical Society. .................................................................... 48

Figure 3-9: Wavelength-normalized absorption and OD600 for each strain plotted as a
function of time. The specific bacteria plotted are: (a) S. aureus, HH36 strain, (b) S.
aureus, HH49 strain, (c) S. aureus, KH38 strain, (d) S. aureus, LAC91 strain, and (e) P.
aeruginosa, PA01 strain. Error bars are shown on each plot for both wavelength-
normalized absorption and OD600 measurements; however, the error is so small for
the wavelength-normalized absorption that most of the error bars are smaller than the
symbols, while the error is clearly visible for OD600 measurements. Reprinted with
permission from reference 25, The Optical Society. .................................................................... 50

Figure 4-1: Hemozoin production. (a-d) The formation of hemozoin and subsequent release
into the bloodstream. (a) The parasite (shown in blue) remains in the liver,
reproducing until it is released into the bloodstream. (b) Free heme is generated, as a
byproduct of hemoglobin consumption by the parasite. (c) Heme is aggregated into an
insoluble crystal known as hemozoin. (d) Hemozoin is released into circulation during
erythrocyte lysis. Hemozoin remains in circulation from several days to weeks,

xiii
without affecting phagocyte viability. (e) Scanning electron microscopy image of β-
hematin. ........................................................................................................................................................... 57

Figure 4-2: The beginning stages of the device. (a) Version 1 was all free-space optics,
designed for preliminary experiments. (b) Version 2 was the first semi-portable
version. (c) With the incorporation of a 3D-printed sample holder, we were
streamlining the system. This version was also self-contained within a box, as all
future versions were as well – however, the box was not included in photos for the
sake of being able to see the internal workings of the device. (d) The use of a laser
diode was a vast improvement over the laser pointer used, making the device more
user-friendly and allowing for continuous use for up to 36 hours. ....................................... 63

Figure 4-3: (a) The live version and (b) a rendering of the fifth version of the device, used
for the majority of further experiments. The incorporation of a linear actuator and a
rotating sample holder were hugely significant. ........................................................................... 65

Figure 4-4: Early-stage results with versions 1 and 2 of the device. (a) Two dilutions of
spherical Fe3O4 nanoparticles were tested – one visibly higher in concentration than
the other. The sample with higher concentration had a greater change in power. (b-e)
Fe3O4 nanoparticles in two different solvents tested with five magnets of varying
shape and field strength over a range of dilutions. Each plot shows that, with greater
field strength, we see a greater change in power, even at the same concentration. (b)
Fe3O4 nanoparticles in water at a 1:128 dilution. (c) Fe3O4 nanoparticles in water at a
1:256 dilution. (d) Fe3O4 nanoparticles in 75% glycerol at a 1:2 dilution. (e) Fe3O4
nanoparticles in 75% glycerol at a 1:32 dilution. All y axes have units of µW. ................ 74

Figure 4-5: Results for spherical Fe3O4 nanoparticles tested over a range of concentrations
and with magnets of varying field strengths applied after either 30 seconds or 60
seconds of baseline data collection. ..................................................................................................... 76

Figure 4-6: First experiments conducted with semi-portable system, and first time
recording change in optical power in real-time. ............................................................................ 77

Figure 4-7: Two different visualizations for results from malaria-infected blood samples.
Samples 16 and 19 are controls, while Samples 23, 26, and 27 are infected. (a) The
controls are relatively flat when compared to the infected samples, and there are
increases in transmitted power when the magnetic field is applied to infected samples,
as expected. (b) The controls remain close to the 0.0 point, indicating little change in
transmission, while the infected samples are more spread out along the x axis,
indicating more significant changes in transmission. ................................................................. 78

Figure 4-8: Spherical Fe3O4 nanoparticles tested over a range of dilutions in (a) water, (b)
ethanol, (c) 10% PEG, and (d) 15% PEG. As the dilution increases, the change in signal
decreases, as is expected. ......................................................................................................................... 80

xiv
Figure 4-9: Spherical Fe3O4 nanoparticles tested with a green laser. (a) 1:2000 dilution in
water, (b) 1:4000 dilution in water, (c) 1:2000 dilution in 10% PEG, (d) 1:4000
dilution in 10% PEG. .................................................................................................................................. 81

Figure 4-10: Experimental and mathematical results for spherical Fe3O4 nanoparticles
tested in (a) water, (b) 10% PEG, and (c) 15% PEG. Solid lines show experimental
data; shaded regions show ranges provided by mathematical modeling. .......................... 82

Figure 4-11: Results from (a) whole and (b) lysed malaria-infected blood samples. ............. 83

Figure 4-12: Experimental and mathematical results for β-hematin tested in (a) 10% PEG
and (b) 15% PEG. The working range of the device for β-hematin concentrations in (c)
10% PEG and (d) 15% PEG. Insets show the working range of specific concentrations
plotted in (a) and (b). ................................................................................................................................. 85

Figure 4-13: Experimental and mathematical results for β-hematin tested in 10% PEG and
15% PEG. Experimental results are solid lines; mathematical results are dashed lines.
(a) Highest concentrations in 10% PEG, (b) medium concentrations in 10% PEG, (c)
lowest concentrations in 10% PEG, (d) highest concentrations in 15% PEG, (e)
medium concentrations in 15% PEG, (f) lowest concentrations in 15% PEG. ................. 86

Figure 4-14: Reproducibility of results for (a) 10% PEG and (b) 15% PEG. The
measurement was performed iteratively using the same set-up. .......................................... 87

Figure 5-1: Schematic of the device. The light from the laser diode passes through the
sample and is detected on the power meter. The concentration of magnetic
nanoparticles is detected by changing the position of the magnet, which is mounted on
a computer-controlled motorized stage. The entire system is enclosed in a dark box to
minimize noise from ambient light. ..................................................................................................... 95

Figure 5-2: (a) Experimental and mathematical results for β-hematin tested in rabbit blood.
(b) The working range of the device for β-hematin concentrations in rabbit blood. (c)
Reproducibility of results. The measurement was performed iteratively using the same
set-up. ............................................................................................................................................................... 99

Figure 5-3: Experimental and mathematical results for β-hematin tested in rabbit blood.
Experimental results are solid lines; mathematical results are dashed lines. (a) Higher
concentrations in rabbit blood, (b) lower concentrations in rabbit blood. ...................... 100

Figure 6-1: Schematic of the device. ............................................................................................................ 104

Figure 6-2: (a) Experimental and mathematical results for P. cynomolgi-infected M. mulatta
blood. (b) The working range of the device. (c) Reproducibility of results. The
measurement was performed iteratively using the same set-up. Legends for (a) and (c)
show parasite count per µL. .................................................................................................................. 107

xv
Figure 7-1: Simple physics of a blast wave. ............................................................................................. 111

Figure 7-2: Rendering of blast-induction set-up, complete with six-well plate of neurons on
microscope stage. ...................................................................................................................................... 117

Figure 7-3: (a-c) Isolated hippocampal neurons, 16 days post-plating, imaged upon arrival
to our lab (after transportation). ........................................................................................................ 119

Figure 7-4: Live cell imaging using calcein-AM. Scale bars represent 100 μm. (a) Green
spots indicate the presence of live cells before application of a laser pulse to the well.
(b) A lack of green fluorescence indicates the lack of live cells after application of a
laser pulse, showing that the cells lost their membrane integrity within 30 minutes of
the simulated blast. ................................................................................................................................... 120

Figure 8-1: An illustration of the improved magnetic trap, with the sample shown in grey,
magnetic nanoparticles in black dots, the laser beam in red, and the magnets in
translucent blue. (a-b) Following the baseline measurement of light transmission, the
magnet, mounted on an actuator, moves down, (c) pulling magnetic nanoparticles into
an aggregate at the bottom of the cuvette. (d) The magnet then moves back up, (e)
pulling the aggregate back into the beam path at the peak of its movement, where the
magnet would be removed for the second measurement. ...................................................... 128

Figure 8-2: All device configurations to be tested in the future: (a) A single-magnet system
with transmission-based detection, (b) a single-magnet system with polarization-
based detection, (c) a magnetic trap with transmission-based detection, and (d) a
magnetic trap with polarization-based detection. ...................................................................... 131

xvi
Abstract
Optics first originated as a field of study centuries ago with the development of
lenses by the ancient Egyptians and Mesopotamians. They never could have imagined how
these early lenses would evolve and what opportunities they would open up with this field.
Practical applications segued into theory, and all of this has catapulted us into our current
state. Optics is presently known as the branch of physics that involves the behavior and
properties of light, including its interactions with matter and the construction of
instruments that use or detect it. This dissertation details the development and application
of new tools created to further our understanding of a variety of topics, ranging from
bacteria to malaria to blast-induced traumatic brain injury. Each project utilizes various
properties and capabilities of light to study topics that are, otherwise, seemingly
disconnected.
Here, I will share work I have completed to-date and ideas for future work to be
passed on to others. In the first part of my dissertation, I explain a novel method I
developed to determine microbial growth using multi-wavelength spectroscopic
measurements. Measuring and monitoring bacterial growth rates plays a critical role in a
wide range of settings. Having a fuller understanding of the growth cycles of bacteria
known to cause severe infections and diseases will lead to a better understanding of the
pathogenesis of these illnesses, leading to better treatment and, ultimately, the
development of cures.

xvii
The second section of my thesis details what has been my primary focus for the last
two years – the development of a malaria diagnostic. Exploiting the magnetic properties of
a byproduct of the malaria parasite present in the bloodstream, we have developed a
device capable of detecting malarial infection at clinically relevant concentrations. The
implications of this device are incredibly far-reaching – if successful, this tool could save
hundreds of thousands of people every year who die from malaria due to inaccurate or late
diagnosis. I am hoping that one of my mentees will continue this work once I am gone in an
effort to fully transition this device from the lab to reality.
The final section of this dissertation details my study of blast-induced traumatic
brain injury. Presently, little is known about this specific type of traumatic brain injury,
despite it causing a multitude of injuries and deaths, particularly among troops. I have
studied the properties of a system representative of the brain in an effort to see how the
system changes upon being exposed to a laser-induced blast. The hope is that, once the
mechanisms of neuronal death due to blast-induced neurotrauma have been clarified, we
may then use these results to design a preventive measure to avoid these injuries
altogether.

1
Chapter 1: Introduction and Overview
“I am driven by two main philosophies: know more about the world than I knew
yesterday, and lessen the suffering of others. You’d be surprised how far that gets you.”
~ Neil deGrasse Tyson

1.1 Motivation
Since the dawn of scientific experiments, one series of questions that has plagued
mathematicians, philosophers, physicists, and inventors alike is, what is light, how does it
travel, and how can we use light and its unique properties to our advantage? As early as
700 BCE, answers to these questions started appearing when ancient Egyptians and
Mesopotamians decided to polish crystals in an attempt to replicate the optical properties
of water. The Nimrud lens, also known as the Layard lens, is a nearly 3000-year-old piece of
rock crystal that is one of the first lenses ever to be made.
1
While we are unsure as to
whether it was used as a magnifying glass, a tool to start fires, or simply a decorative piece,
this was the start of something tremendous.
These first steps in optics made primarily in Asia and the Middle East fueled the
imagination of Greek and Roman physicists, mathematicians, philosophers, and inventors
whose experiments formed the basis of classical optics. Around 250 BCE, Archimedes set
up an optical defense mechanism for the King of Sicily to defend the city against the
Romans.
2
Around 30 CE, Seneca wrote about the magnifying effects of liquids in

2
transparent vessels.
2
Hero of Alexandria defined the law of refraction a few decades later,
and, just after the turn of the century, Ptolemy wrote a five-volume textbook on optics.
2,3

After the theory came more applications – the year 1284 CE brought the creation of the
first wearable eyeglass, marking an expansion of optics research that enabled scientists to
make discoveries at an unprecedented rate.
2
In the 17
th
century alone, we were gifted with
the brilliant minds of: Johannes Kepler, who provided a correct explanation of vision and
the function of the pupil, cornea, and retina; Robert Hooke, whose book Micrographia on
optical microscopy brought us not only important observations using an early microscope
but also introduced us to the wave theory of light; René Descartes, whose Discourse on the
Method is the first known publication of the Law of Refraction; Christiaan Huygens,
remembered especially for his wave theory, published in his Traité de la lumière, making it
the first mathematical theory of light; and Isaac Newton, whose book Opticks was accepted
as the greatest achievement in light research at the time.
2,3

Advances in the field of optics continued with the work of countless other
individuals, including Thomas Young, James Maxwell, Max Planck, Albert Einstein, Niels
Bohr, and Paul Dirac, to name a few. It is because of the work done by all of these scientists
and more that optics has developed into the field it has today. Optics is a field of limitless
potential, as evidenced by its many applications found in a variety of technologies and
everyday objects. Many people use eyeglasses or contact lenses, and optics is integral to the
functioning of many consumer goods, including cameras. Rainbows and mirages are
examples of optical phenomena. Optical communication provides the backbone for both the
Internet and modern telephones. But before optics can reach that level, it must be used in
fundamental research. It is in this fundamental research realm that this dissertation lays.

3
This dissertation combines optics and biomedical engineering to develop novel
devices and new measurement and analysis techniques, as well as the application of these
tools to describe and understand various biological and medical phenomena. This
dissertation covers work in three specific areas: spectroscopic analysis of bacteria growth
rates, malaria detection, and blast-induced traumatic brain injury. While seemingly
unrelated topics, the use of optics is the common thread tying all of them together.

1.2 Chapter Overview
The focus of this dissertation is to improve our understanding of various biological
phenomena through the use of optics. Before this research is detailed in depth, however, it
is prudent to take a brief look at the associated background.
We begin this review in Chapter 2 with an extensive review of sensor theory and the
use of optics in the context of biomedical engineering, specifically spectroscopy and the
study of traumatic brain injury.
Chapter 3 details a novel approach for characterizing bacteria concentrations in
growth media. The standard approach currently used in microbiology settings is optical
density (OD) measurements, based on measuring the optical absorption of a sample at a
single wavelength.
4,5
Here, we use the conventional OD technique to measure the growth
rates of two different species of bacteria while also analyzing these samples over the entire
UV-Vis wavelength spectrum, allowing for a multi-wavelength normalization process to be
implemented. Our hypothesis was that, by conducting full spectrophotometric analyses on

4
samples as opposed to only looking at a single wavelength of light, perhaps it is possible to
glean more information on the growth patterns of samples, thereby shedding more light on
the different growth stages. This chapter details this multi-wavelength normalization
process and how it is used to characterize bacterial growth rates, allowing for accurate
quantification of rates with high fidelity at low concentrations.
In Chapter 4, we cover the development and experimental verification of a novel
malaria diagnostic tool based on hemozoin detection. Hemozoin is a magnetic nanoparticle
byproduct generated by the malaria parasite and is found in the blood of a patient infected
with malaria.
6,7
Leveraging hemozoin’s unique magnetic and optical properties, we
developed a portable diagnostic system for malaria detection. Chapter 4 details the process
of designing various versions of the device, and our methodology is tested in earnest,
detecting the presence of spherical Fe3O4 nanoparticles and β-hematin, which is a
hemozoin mimic, in solvents of similar viscosity to blood. We also develop predictive
mathematical models that explain the observed signals. Chapter 5 takes the diagnostic one
step further, detecting the presence of β-hematin in rabbit blood samples, proving its
potential to be used in the field with whole blood samples obtained from patients.
In Chapter 6, the device has once again evolved, as have the experiments conducted.
By this time, our newly found collaborators at Emory University have sent us blood
samples from nonhuman primates infected with malaria. In testing our device with these
samples, we find that the device is capable of detecting a parasite count as low as 25
parasites/μL, which is considered early-stage malaria. Other diagnostics currently in use

5
struggle to detect parasitemia levels below 100 parasites/μL, so this marks a huge
milestone in the evolution of the malaria diagnostic.
Chapter 7 delves into the use of optics to study blast-induced traumatic brain injury,
or bTBI. Traumatic brain injuries can be broadly categorized as originating from blunt
impact or a blast. While the precise neuronal mechanisms of blunt impact traumatic brain
injuries have been studied in depth, particularly as they relate to sports injuries and
automobile accidents, there is not much known about the mechanisms of bTBI. bTBI occurs
specifically when a pressure wave emanating from a source comes into contact with the
skull, ultimately manifesting in symptoms of traumatic brain injury.
8,9
We developed an
optical-based method to study what happens on the individual neuron scale when a blast
occurs, simulating the blast using a laser pulse.
Chapter 8 concludes this dissertation with a brief review of the major advancements
resulting from this work and an outline of relevant future work. The chapter details future
work to study blast-induced traumatic brain injury and the phenomena of microcavitation
bubbles in greater detail. The chapter concludes with a number of ways to enhance
sensitivity of our malaria diagnostic tool, as well as next steps for bringing this device to
market.

1.3 References
1 Enoch, J. M. & Lakshminarayanan, V. Duplication of unique optical effects of ancient
Egyptian lenses from the IV/V Dynasties: lenses fabricated ca 2620-2400 BC or

6
roughly 4600 years ago. Ophthal Physl Opt 20, 126-130, doi:Doi 10.1016/S0275-
5408(99)00053-8 (2000).
2 Ben-Menahem, A. Historical encyclopedia of natural and mathematical sciences. 1st
edn, (Springer, 2009).
3 Bertolotti, M. The history of the laser. (Institute of Physics Pub., 2005).
4 Myers, J. A., Curtis, B. S. & Curtis, W. R. Improving accuracy of cell and chromophore
concentration measurements using optical density. BMC Biophys 6, 4,
doi:10.1186/2046-1682-6-4 (2013).
5 Shao, J., Xiang, J., Axner, O. & Ying, C. Wavelength-modulated tunable diode-laser
absorption spectrometry for real-time monitoring of microbial growth. Appl Opt 55,
2339-2345, doi:10.1364/AO.55.002339 (2016).
6 Gluzman, I. Y. et al. Order and specificity of the Plasmodium falciparum hemoglobin
degradation pathway. J Clin Invest 93, 1602-1608, doi:10.1172/JCI117140 (1994).
7 Goldberg, D. E., Slater, A. F., Cerami, A. & Henderson, G. B. Hemoglobin degradation
in the malaria parasite Plasmodium falciparum: an ordered process in a unique
organelle. Proc Natl Acad Sci U S A 87, 2931-2935 (1990).
8 Cernak, I. & Noble-Haeusslein, L. J. Traumatic brain injury: an overview of
pathobiology with emphasis on military populations (vol 30, pg 255, 2010). J Cerebr
Blood F Met 30, 1262-1262, doi:10.1038/jcbfm.2009.203 (2010).
9 Owen-Smith, M. Bomb blast injuries: in an explosive situation. Nurs Mirror 149, 35-
39 (1979).

7
Chapter 2: Background and Related Work
This dissertation details my efforts to contribute to and further the field of
biomedical engineering using optics as my primary tool. My research hinges on three key
disciplines: sensors, spectroscopy, and the study of traumatic brain injury. Before delving
into my specific endeavors, we begin with a brief review of these topics.

2.1 Sensor Theory
2.1.1 Overview
Broadly speaking, a sensor is a device whose purpose is to detect events or changes
in the environment and send this information to receivers. Sensor technology is a
burgeoning field as sensors are allowing us to monitor our surroundings in unparalleled
ways. Sensors take on many different forms and have a wide variety of uses, including in
industry, for safety and security, for communications, and for detection of biological and
chemical agents. For example, temperature, humidity, and glucose sensors monitor
individual containers inside of a refrigerated grocery truck, reducing spoilage in fish or
produce.
1,2
Humidity sensors may also be used in forests to monitor for fire danger.
3

Firefighters use wireless sensors and place them throughout a burning building in order to
map hot spots and flare-ups, also allowing them to create an emergency communications
network.
1
The potential uses for sensors are truly endless.

8
In recent years, the Department of Defense conducted a study known as the
Chemical and Biological Sensor Standards Study (CBS3) to address many shortcomings in
the field of sensor development and validation.
4
Standards and protocols previously used
were inadequate for the ever-expanding requirements, so the DoD developed a new set of
sensor metrics and measurement protocols for the evaluation of sensor efficacy. A panel of
chemical and biological detection experts came together and evaluated the problem at
hand, creating concrete, viable solutions. At the core of the mission lies the fact that sensors
must be highly performing, efficient, reliable, easy-to-use, and readily integrated.
Here, we will delve more deeply into this study, specifically focusing on sensor
metrics and attributes.
2.1.2 Receiver Operating Characteristic (ROC) Curves
The performance of a sensor is characterized by a number of interrelated
parameters, including sensitivity, probability of correct detection, false positive rate, and
response time. The sensor’s Receiver Operating Characteristic (ROC) is created by plotting
sensitivity, or the true positive rate (TPR), against the false positive rate (FPR) at various
threshold values.
4-6
The construction of ROC curves is an efficient method to visualize the
trade-offs for the performance of a particular technique or sensor system for a given set of
sensor conditions
The sensitivity of the sensor is generally the minimal detectable agent concentration,
given in units of particles or mass per unit air volume. In isolation, sensitivity is not a very
telling metric. Due to noise and confounding effects of interferents in the environment,

9
detecting an agent is not assured during operation outside of the lab. A false negative is a
test result indicating that a condition does not hold, when, in fact, it does. This is commonly
due to interferents or noise that cause the sensor to fail to detect the agent. A false positive,
also known as a “false alarm”, is a result that indicates a given condition exists, when, in
fact, it does not.
Sensitivity is meaningless without indicating the probability of detection, or
detection confidence, and the false positive rate, and these are only meaningful when
generated in the context of a real-world environment as opposed to a highly controlled
laboratory setting. Detection confidence is the probability of detecting an agent when there
is one present at or above the specified concentration. Sensor sensitivity, detection
confidence, false positive rate, and response time are all related and all depend on the
sensor’s operating environment.
An example of a set of ROC curves is shown in Figure 2-1. Figure 2-1(a) shows a
family of curves for different levels of detection confidence in a single environment with a
fixed response time. Figure 2-1(b) shows an example of curves for a fixed detection
confidence in various environments. As is shown in these figures, ROC curves can be useful
when testing and validating a sensor to aid in optimizing parameters.

10

Figure 2-1: Sample receiver operating characteristic (ROC) curves. (a) Curves for
different levels of detection confidence in a single environment, with a fixed
response time. (b) Curves for different environments at a fixed detection confidence
and fixed response time. Typically, an urban environment has much greater noise
than does a desert environment. Therefore, at a fixed sensitivity, a sensor operating
in an urban setting will have a higher false positive rate than a sensor operating in a
desert setting.
2.1.3 Response Time
Another important metric for continuously operating sensors is the response time,
meaning the time interval between the arrival of the target agent and the sensor detection
declaration. Sensors do not change output state immediately when an input parameter
change occurs. Instead, it will change to the new state over a period of time, and this period
is called the response time, illustrated in Figure 2-2. The response time can be defined as
the time required for a sensor output to change from its previous state to a final stable
value within a tolerance band of the correct new value. The required response time for a
given sensor is highly dependent on the intended use.

11

Figure 2-2: The sensor response time is the time interval between when the agent
concentration reaches the sensor’s specified sensitivity (or detection threshold) and
the time that the sensor issues a detection signal.
Figure 2-3 shows two types of response time. In Figure 2-3(a), the curve represents
the response time following an abrupt positive going step-function change of the input
parameter. The curve in Figure 2-3(b) is a decay time in response to a negative going step-
function change of the input parameter.

Figure 2-3: (a) Rise-time definition, where Tr represents the response time, and (b)
fall-time definition, where Td represents the response time.

12
2.1.4 Other Sensor Attributes
In addition to performance metrics, there are several other important attributes
that affect the sensor’s utility. Metrics and attributes are categorized as such based on the
relative ease with which one can both define and measure attributes as compared to the
more complex metrics previously discussed.
The initial cost is any cost incurred during the design and construction process, and
it is this cost that affects how a sensor is deployed and the number deployed. This cost can
include planning, preliminary engineering, and project design; an environmental impact
report; project-related staff training; construction costs; equipment purchases; and other
items. Disposable sensors should have low initial costs and be very inexpensive, while non-
disposable sensors could cost orders of magnitude more.
Another related aspect is the operating cost, which is comprised of any costs
incurred after the initial acquisition cost. These expenses are related to the operation of a
device, component, piece of equipment, or facility – the cost of resources used just to
maintain the sensor’s existence. This includes both logistic and maintenance costs,
consumable supplies, repair parts, and operator training. Operating costs can range from
very low for disposable sensors to lifetime costs that greatly exceed the initial cost of the
sensor for more paramount sensors. In a situation where only one sensor is maintained, a
higher operating cost may be more tolerable than in a situation where large numbers of
sensors are deployed.

13
Power consumption is another important attribute, referring simply to the electrical
energy per unit time needed to operate the sensor. Sensors could be battery powered,
require an AC line, or anything in between.
Maintenance consists of the actions taken to keep the sensor in a serviceable
condition or restore it to serviceability, including expenses incurred to keep the sensor in
good condition and/or good working order. When purchasing a sensor that requires
upkeep, consumers often consider not only the initial cost, but also the item’s ongoing
maintenance.
Even when maintenance is carefully followed, sensor failures may still occur based
on the reliability of the sensor. Reliability is the probability that an item will perform to its
intended function for a specified interval under stated conditions. The longer the sensor
performs without experiencing an unexpected failure, the better the reliability.
Ruggedness is a term used for sensors designed to operate in extremely harsh
environments and conditions. Rugged devices are designed to work in extreme
temperatures and weather conditions, to withstand shocks and vibrations, to be
impervious to being dropped, and to be dustproof and waterproof.
The form factor is an aspect of design that defines and prescribes the size, shape,
weight, and other physical specifications. Form factor is of particular concern when sensors
are frequently moved, such as in the case of a point-of-care diagnostic tool. Small, self-
contained, portable sensors are highly desirable in this role. Small form factor is typically
less critical from the facility standpoint because sensors will usually remain in place.

14
The final attribute, environmental considerations, is the set of guidelines meant to
protect the environment and living things. These include issues such as safe disposal of
reagents and used consumables to excessive noise and laser eye-safety. These can have a
serious impact on sensor acceptance.
Thus, the overall performance of a sensor can be characterized by a number of
related metrics, while acceptance and implementation of a sensor can be influenced by
these aforementioned attributes. The exact requirement for each of these metrics and
attributes is highly dependent on the use of the sensor.
Table 2-1: Factors known to affect sensor utility.
Environmental Factors Economic Factors Sensor Characteristics
Temperature range Cost Sensitivity
Humidity effects Availability Range
Corrosion Lifetime Stability
Size Repeatability
Power consumption Linearity
Self-test capability Error
Ruggedness Response time
Frequency response

2.1.5 The Spider Chart
To capture the overall performance of a sensor, it is common to use a spider chart,
or radar chart as it is also commonly known. This is a graphical method of displaying

15
multivariate data in the form of a two-dimensional chart of three or more quantitative
variables represented on spokes projecting from a central point. Each spoke represents one
of the variables, and the values of the variables are encoded into the lengths of the spokes,
making spoke length proportional to the magnitude of the variable. A line is then drawn
connecting the data values for each spoke.
4,7

The spider chart in Figure 2-4 integrates all of the sensor metrics into one visual
chart. Each metric or attribute relevant to the sensor being evaluated is assigned to a
spoke, and the tick marks on each spoke represent measured or predicted values
associated with the respective metric/attribute. Values improve as they radiate out from
the origin. There are three values on each spoke – the center value representing an average
acceptable value, the inner and outer values representing the error bar. The innermost
value is minimally acceptable, while the outermost value is the point at which one reaches
diminishing returns. Connecting the values provides a footprint of the sensor performance
– the larger the area encompassed, the better the sensor. This footprint also becomes
valuable in comparing multiple sensors with common requirements.

16

Figure 2-4: Example of a spider chart, allowing designers to quickly determine areas
that can be improved or identify tradeoffs that should be made to optimize sensor
function.

2.2 Spectroscopy
2.2.1 Overview
The history of spectroscopy began with Newton’s optics experiments in the late
1660s and early 1670s.
8
However, it wasn’t until the early 1800s that experimental
advances enabled spectroscopy to become a more precise and quantitative technique,

17
specifically with the development of the spectrometer. At the end of the 19
th
century,
spectroscopy was still limited to the absorption, emission, and scattering of visible,
ultraviolet, and infrared electromagnetic radiation. It has since expanded to include other
forms of electromagnetic radiation (X-rays, microwaves, and radio waves) and other
energetic particles (electrons and ions).
8

Although there are many differences in spectroscopic instrumentation and
techniques, all techniques leverage the properties of electromagnetic radiation, or light.
Light is a form of energy whose behavior is described by the properties of both waves and
particles. It consists of oscillating magnetic and electric fields that propagate through space
along a linear path with constant velocity. A light wave is characterized by its velocity,
amplitude, frequency, phase angle, polarization, and direction of propagation. The
wavelength is defined as the distance between successive maxima. For ultraviolet and
visible light, the wavelength is typically expressed in nm, while for infrared radiation, it is
usually given in nm.
When light and matter interact, there are many events that can take place. One type
of interaction is known as absorption, which forms the basis for absorption spectroscopy.
2.2.2 Absorption Spectroscopy
In absorption spectroscopy, a beam of electromagnetic radiation passes through a
sample. Typically, most of the radiation passes through the sample without a loss in
intensity. However, at selected wavelengths dependent upon the makeup of the sample, the
radiation’s intensity is attenuated. At these wavelengths, a photon is absorbed by the

18
sample. This photon is destroyed, its energy is acquired by the sample, and the particles
undergo a transition from a lower-energy state to a higher-energy state, or excited state, as
shown in Figure 2-5. When the sample absorbs a photon, the number of photons passing
through the sample decreases, as shown in Figure 2-6.
9-12

Figure 2-5: Electron excitation. (a) Electron orbiting the nucleus of an atom. A photon
is absorbed by the sample, and (b) the particles are then excited to a higher-energy
state.

Figure 2-6: Schematic showing the attenuation of radiation passing through a
sample. I0 is the incident light from the source, and I is the light transmitted through
the sample to the detector (in this case, the eye). Due to absorbance, I < I0.
This attenuation of radiation is described quantitatively by two separate, but
related, terms – absorbance and transmittance. Absorbance is the more common unit for

19
expressing the attenuation of radiation as it is a linear function of the sample’s
concentration. This is a measurement of the decrease in the number of photons passing
through the sample, and it is plotted as a function of the photon’s energy. Absorbance is
given by:

A=−log
I
I
0
⎛
⎝
⎜
⎞
⎠
⎟
(2.1)
Transmittance, on the other hand, is the ratio of the number of photons that pass
through the sample, I, to the number of incident photons, I0 (as shown in Figure 2-6):

T=
I
I
0
(2.2)
Both transmittance and absorbance measure photons that are absorbed or
scattered, and they are each related to the optical depth, absorptivity, and concentration of
the sample. Absorbance and transmittance are inversely related – that is, the more a
particular wavelength of light is absorbed by a substance, the less it is transmitted – and
this relationship is logarithmic in nature.

20

Figure 2-7: Plot of transmittance (in dark blue) and absorbance (in pink) as a
function of wavelength (λ, nm) for CrCo0.6Ni0.4FeO4 annealed at 1000°C.
13

Every atom and molecule in existence has unique absorbance and transmittance
spectra. For nanoparticles, size, shape, concentration, aggregation, and refractive index all
contribute to the measure UV-Vis spectra. For example, scattering by a very small particle
(approximately 2 nm in diameter) is negligible, so the optical spectrum of nanoparticles is
almost entirely due to absorption. Scattering is also highly sensitive to aggregation state –
specifically, as nanoparticles aggregate, scattering increases. When nanoparticles are in a
transparent solvent, only the spectral features of the nanoparticle need be considered.
However, for the malaria diagnostic, once we move beyond system optimization
experiments, we are studying nanoparticles in blood, which brings in other variables. When
looking at UV-Vis spectra for nanoparticles in blood, we have to take into account the
spectral features of the solvent, which can add a significant amount of noise. All of the same

21
scattering and absorption that occurs for nanoparticles is amplified because there are small
particulates in the blood with their own optical properties that inevitably contribute to the
UV-Vis spectra. Additionally, the color and opacity of the solvent medium affect UV-Vis
spectra, so the color of blood and its opaque nature both impact the spectroscopic
measurements.
Because every atom or molecule has their own unique UV-Vis spectra, these spectra
can be used to detect, identify, and quantify information about a given sample. Absorption
spectroscopy is one of the best methods for determination of impurities in organic
molecules. If an impurity is present, additional peaks can be observed and compared.
14-16

Spectroscopy is also useful in elucidating the structures of organic molecules. The presence
or absence of specific peaks and the combination therein of peaks can provide important
structural details.
17,18

2.2.3 Differential Absorption Spectroscopy
Standard absorption spectroscopy involves the use of a background, or
normalization, spectrum. This necessitates taking a measurement using a sample that only
contains the solvent medium. All subsequent spectra are normalized using this background
measurement by subtracting the absorbance of the solvent medium from the absorbance of
the sample of interest at each time point. This is the standard approach when testing
samples that have a universal control.
However, in differential absorption spectroscopy, only differential intensities are
important, which eliminates the need for a separate control measurement. By monitoring

22
the change in intensity between two different states and calculating the delta between
measurements, we normalize a sample against itself. This strategy is often used when a
universal control is not available – for example, with human specimens. When testing blood
samples, there is no universal control because every blood sample is unique. Color,
viscosity, opacity, blood cell density, and presence of particulates are but a few variables to
consider when comparing blood samples between humans, or even blood samples from the
same subject collected at different time points. So the use of differential spectroscopy not
only eliminates the need for a separate normalization step, but also reduces the effect of
sample-to-sample variation, which is inherent in human specimens.

2.3 Lasers and bTBI
2.3.1 Traumatic Brain Injury: An Overview
Traumatic brain injury, often referred to as TBI, occurs when an external force
injures the brain and there is a sudden acceleration or deceleration within the skull. In
addition to the damage caused at the moment of injury, a variety of events in the minutes to
days following may result in secondary injury.
19-21
TBI can be classified in many different
ways, and no two brain injuries are alike. TBI often results in physical, cognitive, social,
emotional, and behavioral symptoms, each ranging in severity.
22-27
Some symptoms appear
immediately, while others may not manifest for days or weeks, if not years, after the injury.
Outcomes can range from complete recovery to permanent disability or death.

23
While causes vary and include falls, vehicle collisions, and violence, causes can be
placed into one of two broad categories – blunt impact, or blast-induced. Blunt impact
forms of TBI include injuries sustained from slips, trips, or falls, sports-related injuries,
automobile accidents, and the like. Blast-induced traumatic brain injury, or bTBI, is an
injury more common among troops and is often caused by an explosion. One thing that has
been discovered is that TBI manifests itself very differently when triggered by an explosion
as opposed to blunt impact. In the case of a blunt impact, TBI is a more typical concussion
with localized inflammation, edema, and neuronal and glial cell death.
28-30
However,
explosive blasts often result in a high-pressure wave moving out from an explosion at a
high speed. Blasts begin with detonation of an explosive material, and the overpressure
wave, or blast wave, starts with a single pulse of increased air pressure that lasts a few
milliseconds, with a negative pressure immediately following.
31
The blast wave progresses
from the source as a sphere of compressed and rapidly-expanding gases, displacing an
equal volume of air at high velocity.
32
The blast wave consists of the front of high pressure
that compresses the surrounding air and falls rapidly to negative pressure and is the main
determinant of the primary blast injury. However, a person exposed to an explosion will be
subjected not only to a blast wave but also to the high-velocity wind traveling directly
behind the shock front of the wave.
32

24

Figure 2-8: Simple physics of a blast wave.
There are several potential modes for blast energy transfer to the brain, which bring
us to several hypotheses as to precisely what aspect of the blast wave results in cellular
injury and death. But the reality is that we do not yet have a firm grasp on the mechanism
of bTBI, and it is this lack of understanding that keeps us from developing predictive
models and creating improved preventive measures and treatment methods. In order to
isolate and verify which hypotheses are correct, it is necessary to perform experiments at
all different levels of biological complexity (molecular, cellular, tissue, etc.). Currently, the
primary methods which are being pursued to understand the molecular and the cellular
mechanisms rely heavily on fluorescent methods; for example, using multi-wavelength
fluorescent microscopy to image the deformation of the cell or the production of specific
proteins under different shear rates.
33,34
However, due to limitations in multi-color
fluorescent imaging, only a few proteins can be monitored simultaneously.
35
Developing an
alternative method of monitoring cell behavior when exposed to a simulated blast is crucial

25
in allowing a fuller understanding of the mechanism of bTBI, and using a laser-based
approach seems to be a viable path.
2.3.2 Lasers: An Overview
Laser is an acronym, to many people’s surprise. It stands for Light Amplification by
Stimulated Emission of Radiation (LASER), and it is a very intense, highly directional beam
of light.
36
Although there are many different types of lasers, they all have certain essential
features. As is shown in Figure 2-9, the lasing medium is pumped, at which point electrons
absorb energy and become excited, thereby moving from a lower-energy to a higher-
energy orbit. When the electron releases this energy and returns to its ground state, it
emits energy in the form of photons. The photon emitted has a very specific wavelength
that depends on the state of the electron’s energy when the photon is released.

26
Figure 2-9: Electron excitation and emission. (a) Electron orbiting the nucleus of an
atom. The lasing medium is pumped, releasing energy (E = hc/λ) that is absorbed by
the electron. (b) The electron is then excited to a higher-energy state. (c) The
electron subsequently releases this energy in the form of photons, (d) thereby
returning to its original ground state.
Lasers are unique in that they are artificially created beams of energy that possess
intense brightness, can travel long distances, and can be used in countless applications in
everyday life. Lasers differ from other light sources in that they emit light coherently, both
spatially and temporally. Spatial coherence allows a laser to be focused to a tight spot and
propagate long distances with minimal loss of intensity. Temporal coherence allows lasers
to emit light with a very narrow spectrum, meaning they can emit a single color of light.
36

There are many different types of lasers. Lasers are commonly designated by the
lasing mechanism, which is related to the gain material and the emission wavelength.
37

Laser applications are endless and include laser printers, barcode scanners, DNA
sequencing instruments, fiber-optic communications, military and law enforcement
devices, and laser lighting displays for entertainment.
38
One such laser application is using
laser light to recreate a blast, allowing us to study traumatic brain injury on a more
localized level.
2.3.3 Using Lasers To Study Traumatic Brain Injury
Laser ablation of tissue with nanosecond laser pulses has been widely studied in the
past, specifically in medical applications. When a laser pulse interacts with a material,
optical energy is absorbed by a combination of linear and nonlinear absorption, dependent

27
upon both the laser light irradiance and the material properties.
39
Lasers have also been
used to trigger microcavitation, thereby selectively destroying cells.
40,41
By bringing these
two applications together, laser pulses have been used to simulate a blast and trigger
neuronal death in a manner similar to blast-induced neurotrauma.
One of the many hypotheses as to what aspect of the blast wave results in cellular
injury and death revolves around microcavitation. Cavitation can occur in a liquid instantly
set into motion by an impulsive force.
42
It is not the formation, but rather the collapse, of
these cavitation bubbles that often results in severe damage. The brain is suspended in
cerebrospinal fluid, lending itself toward these fluid-based hypotheses, so the effects of
blasts that cause traumatic brain injury from cavitation have been investigated.
43,44
To
study this microcavitation theory, laser pulses have been used to create microcavitation
bubbles in a variety of media.
It is known that cells are loaded with microparticles by phagocytosis, and when
irradiated with short laser pulses (nano- or picosecond time range), these microparticles
absorb laser energy and release this energy as heat. For a brief moment, they can reach
temperatures significantly higher than 100
°
C. This causes localized vaporization of the fluid
surrounding the microparticles, leading to the generation of transient vapor bubbles
known as microcavitation.
40,45,46
The lifetime of a bubble is typically 0.1-1 μs. When these
bubbles expand and subsequently collapse, they breach the cell membrane and destroy the
cell.
40,46
Cell damage by microcavitation precedes cell damage by other thermal methods
when the irradiation is delivered in single pulses fewer than 3 μs in duration.
47
More on
this will be discussed in Chapter 7.

28

2.4 References
1 Mehrotra, P. Biosensors and their applications - A review. J Oral Biol Craniofac Res 6,
153-159, doi:10.1016/j.jobcr.2015.12.002 (2016).
2 Monosik, R., Stredansky, M., Tkac, J. & Sturdik, E. Application of enzyme biosensors
in analysis of food and beverages enzyme and microbial technology. Food Analytic
Methods 5, 40-53 (2012).
3 Holden, Z. A. & Jolly, W. M. Modeling; topographic influences on fuel moisture and
fire danger in complex terrain to improve wildland fire management decision
support. Forest Ecol Manag 262, 2133-2141, doi:10.1016/j.foreco.2011.08.002
(2011).
4 Carrano, J. Chemical and Biological Sensor Standards Study., (US Department of
Defense, Arlington, VA, USA, 2005).
5 Cotte-Rodriguez, I. et al. Analysis of gaseous toxic industrial compounds and
chemical warfare agent simulants by atmospheric pressure ionization mass
spectrometry. Analyst 131, 579-589, doi:10.1039/b513605a (2006).
6 Young, M., Fan, W., Raeva, A. & Almirall, J. Application of Receiver Operating
Characteristic (ROC) Curves for Explosives Detection Using Different Sampling and
Detection Techniques. Sensors 13, 16867-16881, doi:DOI 10.3390/s131216867
(2013).
7 Saary, M. J. Radar plots: a useful way for presenting multivariate health care data. J
Clin Epidemiol 61, 311-317, doi:10.1016/j.jclinepi.2007.04.021 (2008).
8 Bertolotti, M. The history of the laser. (Institute of Physics Pub., 2005).
9 Walsh, A. Atomic absorption spectroscopy: a foreword. Appl Opt 7, 1259,
doi:10.1364/AO.7.001259 (1968).
10 Sullivan, J. V. & Walsh, A. The isolation and detection of atomic resonance lines. Appl
Opt 7, 1271-1280, doi:10.1364/AO.7.001271 (1968).
11 Price, W. C. & Walsh, A. D. The absorption spectra of benzene derivatives in the
vacuum ultraviolet. Proc R Soc Med 191, 22-38 (1947).
12 Walsh, A. D. The absorption spectrum of acetaldehyde in the vacuum ultra-violet.
Proc R Soc Med 185, 176-182 (1946).
13 Singhal, S., Bhukal, S., Singh, J., Chandra, K. & Bansal, S. Optical, X-Ray Diffraction, and
Magnetic Properties of the Cobalt-Substituted Nickel Chromium Ferrites
Synthesized Using Sol-Gel Autocombustion Method. Journal of Nanotechnology, 1-6,
doi:10.1155/2011/930243 (2011).
14 Jamro, G. H. & Frei, R. W. Determination of copper impurities in nickel and cobalt
salts by atomic absorption spectroscopy. Microchimica Acta 58, 429-432 (1970).
15 Karaiskaj, D., Stotz, J. A., Meyer, T., Thewalt, M. L. & Cardona, M. Impurity absorption
spectroscopy in 28Si: the importance of inhomogeneous isotope broadening. Phys
Rev Lett 90, 186402, doi:10.1103/PhysRevLett.90.186402 (2003).

29
16 Sasaki, T., Sakamoto, T. & Otsuka, M. Detection of Impurities in Organic Crystals by
High-Accuracy Terahertz Absorption Spectroscopy. Analytical Chemistry 90, 1677-
1682 (2018).
17 Horng, M. & Quitevis, E. L. Visible Absorption Spectroscopy and Structure of Cyanine
Dimers in Aqueous Solution: An Experiment for Physical Chemistry. Journal of
Chemical Education 77, 637 (2000).
18 Maréchal, Y. The molecular structure of liquid water delivered by absorption
spectroscopy in the whole IR region completed with thermodynamics data. Journal
of Molecular Structure 1004, 146-155 (2011).
19 Bellander, B. M. et al. Secondary insults following traumatic brain injury enhance
complement activation in the human brain and release of the tissue damage marker
S100B. Acta Neurochir (Wien) 153, 90-100, doi:10.1007/s00701-010-0737-z
(2011).
20 Elward, K. & Gasque, P. "Eat me" and "don't eat me" signals govern the innate
immune response and tissue repair in the CNS: emphasis on the critical role of the
complement system. Mol Immunol 40, 85-94 (2003).
21 Stahel, P. F., Morganti-Kossmann, M. C. & Kossmann, T. The role of the complement
system in traumatic brain injury. Brain Res Brain Res Rev 27, 243-256 (1998).
22 Morton, M. V. & Wehman, P. Psychosocial and emotional sequelae of individuals
with traumatic brain injury: a literature review and recommendations. Brain Inj 9,
81-92 (1995).
23 Berger, E., Leven, F., Pirente, N., Bouillon, B. & Neugebauer, E. Quality of Life after
traumatic brain injury: A systematic review of the literature. Restor Neurol Neurosci
14, 93-102 (1999).
24 Zuckerman, S. L. et al. Functional and Structural Traumatic Brain Injury in
Equestrian Sports: A Review of the Literature. World Neurosurg 83, 1098-1113,
doi:10.1016/j.wneu.2014.12.030 (2015).
25 Beadle, E. J., Ownsworth, T., Fleming, J. & Shum, D. The Impact of Traumatic Brain
Injury on Self-Identity: A Systematic Review of the Evidence for Self-Concept
Changes. J Head Trauma Rehabil 31, E12-25, doi:10.1097/HTR.0000000000000158
(2016).
26 Oyesanya, T. O. & Ward, E. C. Mental Health in Women With Traumatic Brain Injury:
A Systematic Review on Depression and Hope. Health Care Women Int 37, 45-74,
doi:10.1080/07399332.2015.1005307 (2016).
27 Stuhmiller, J. H. Biological response to blast overpressure: a summary of modeling.
Toxicology 121, 91-103 (1997).
28 Andriessen, T. M. et al. Epidemiology, severity classification, and outcome of
moderate and severe traumatic brain injury: a prospective multicenter study. J
Neurotrauma 28, 2019-2031, doi:10.1089/neu.2011.2034 (2011).
29 Losiniecki, A. & Shutter, L. Management of traumatic brain injury. Curr Treat Options
Neurol 12, 142-154, doi:10.1007/s11940-010-0063-z (2010).
30 Timmons, S. D. Current trends in neurotrauma care. Crit Care Med 38, S431-444,
doi:10.1097/CCM.0b013e3181ec57ab (2010).
31 Cernak, I., Savic, J., Ignjatovic, D. & Jevtic, M. Blast injury from explosive munitions. J
Trauma 47, 96-103, doi:Doi 10.1097/00005373-199907000-00021 (1999).

30
32 Cernak, I. & Noble-Haeusslein, L. J. Traumatic brain injury: an overview of
pathobiology with emphasis on military populations (vol 30, pg 255, 2010). J Cerebr
Blood F Met 30, 1262-1262, doi:10.1038/jcbfm.2009.203 (2010).
33 Basch, A. & Fazekas, I. G. Increased permeability of the blood-brain barrier following
experimental thermal injury of the skin. A fluorescent and electron microscopic
study. Angiologica 7, 357-364 (1970).
34 Guan, Y., Dong, P., Wan, C., Yang, J. & He, M. The assessment of brain injury with
measurement of calcium fluorescent intensity of a vital brain slice. J Extra Corpor
Technol 34, 185-189 (2002).
35 Wiedenmann, J., Oswald, F. & Nienhaus, G. U. Fluorescent proteins for live cell
imaging: opportunities, limitations, and challenges. IUBMB Life 61, 1029-1042,
doi:10.1002/iub.256 (2009).
36 Pan, M. W. et al. Spatial and temporal coherence of broad-area lasers with grating
feedback. J Opt Soc Am B 15, 2531-2536, doi:Doi 10.1364/Josab.15.002531 (1998).
37 Patil, U. A. & Dhami, L. D. Overview of lasers. Indian J Plast Surg 41, S101-113
(2008).
38 Duley, W. W. An encyclopedia of laser applications. Anal Chem 55, 1355A-1356A,
doi:10.1021/ac00263a774 (1983).
39 Perez-Gutierrez, F. G., Camacho-Lopez, S. & Aguilar, G. Time-resolved study of the
mechanical response of tissue phantoms to nanosecond laser pulses. Journal of
Biomedical Optics 16, doi:Artn 115001
10.1117/1.3644380 (2011).
40 Leszczynski, D., Pitsillides, C. M., Pastila, R. K., Rox Anderson, R. & Lin, C. P. Laser-
beam-triggered microcavitation: a novel method for selective cell destruction.
Radiat Res 156, 399-407 (2001).
41 Mills, B. M. et al. Microcavitation and spot size dependence for damage of artificially
pigmented hTERT-RPE1 cells. P Soc Photo-Opt Ins 5319, 245-251,
doi:10.1117/12.530331 (2004).
42 Pan, Z. et al. Cavitation onset caused by acceleration. P Natl Acad Sci USA 114, 8470-
8474, doi:10.1073/pnas.1702502114 (2017).
43 Rosenfeld, J. V. et al. Blast-related traumatic brain injury. Lancet Neurol 12, 882-893,
doi:10.1016/S1474-4422(13)70161-3 (2013).
44 Hong, Y., Sarntinoranont, M., Subhash, G., Canchi, S. & King, M. A. Localized Tissue
Surrogate Deformation due to Controlled Single Bubble Cavitation. Exp Mech 56, 97-
109, doi:10.1007/s11340-015-0024-2 (2016).
45 Lin, C. P. & Kelly, M. W. Cavitation and acoustic emission around laser-heated
microparticles. Applied Physics Letters 72, 2800-2802, doi:Doi 10.1063/1.121462
(1998).
46 Lin, C. P., Kelly, N. W., Sibayan, S. A. B., Latina, M. A. & Anderson, R. R. Selective cell
killing by microparticle absorption of pulsed laser radiation. Ieee Journal of Selected
Topics in Quantum Electronics 5, 963-968, doi:Doi 10.1109/2944.796318 (1999).
47 Brinkmann, R. et al. Origin of retinal pigment epithelium cell damage by pulsed laser
irradiance in the nanosecond to microsecond time regimen. Laser Surg Med 27, 451-
464, doi:Doi 10.1002/1096-9101(2000)27:5<451::Aid-Lsm1006>3.0.Co;2-1 (2000).

31
Chapter 3: Wavelength-Normalized Spectroscopic
Analysis of Bacteria Growth Rates

3.1 Significance and Background
Optical density (OD) measurements are the standard approach used in microbiology
for characterizing bacteria concentrations in culture media. OD is based on measuring the
optical absorption of a sample at a single wavelength, and any error will propagate through
all calculations, leading to reproducibility issues. Here, we use the conventional OD
technique to measure the growth rates of two different species of bacteria, Pseudomonas
aeruginosa and Staphylococcus aureus. The same samples are also analyzed over the entire
UV-Vis wavelength spectrum, allowing a distinctly different strategy for data analysis to be
performed. Specifically, instead of only analyzing a single wavelength, a multi-wavelength
normalization process is implemented. When the OD method is used, the detected signal
does not follow the logarithmic growth curve. In contrast, the multi-wavelength
normalization process minimizes the impact of bacteria byproducts and environmental
noise on the signal, thereby accurately quantifying growth rates with high fidelity at low
concentrations.
3.1.1 The Stages of Microbial Growth

32
Understanding and accurately characterizing the different stages of microbial
growth play a key role in a wide range of fields spanning from therapeutic design and
development to diagnostics and disease prevention. Microbial growth is characterized by
four distinct phases: lag phase, exponential or logarithmic phase, stationary phase, and
death phase. The lag phase is considered to be the initial period of slow growth, during
which time cells are enlarging and synthesizing critical proteins and metabolites. This
phase is followed by the exponential growth phase, sometimes known as the logarithmic
growth phase. During growth, cell division proceeds at a constant rate with the number of
cells doubling every unit of time, known as the doubling, or generation, time. Generation
times for bacteria vary widely and can range from minutes to days. During the stationary
phase, conditions begin to become unfavorable for growth, and bacteria stop replicating.
During the death phase, cells lose viability.
1-3
In order to fully map out this process, a
measurement method must be able to take readings quickly, iteratively, and reliably over
long periods of time. Additionally, the technique should be resistant to potentially
confounding signals, such as waste byproducts from the cells and thermal drifts.
3.1.2 Current Methods of Measuring Cell Growth
As a result of its importance, a wide range of methods have been developed for
determining the growth rate, ranging from ultrasensitive flow cytometry to automated
counting of bacteria colonies in images. When accuracy is critical, flow cytometry is
typically the preferred solution. However, it requires the bacteria to be fluorescently
labeled, either through genetic engineering (e.g. GFP, RFP) or with a fluorescent tag.
4,5
In
contrast, colony counting does not have this requirement, but it is significantly less

33
sensitive and is prone to large error values.
6
The importance of these measurements in a
wide range of settings from biomanufacturing to basic research has motivated other label-
free methods based on spectroscopy to be developed.
The rise in advanced spectroscopy methods has provided a broad toolset for
biologists to choose from when performing these analytical measurements. For example,
using Raman and FTIR spectroscopy, researchers have monitored the formation of
individual bacterium and bacteria layers and the layers’ response to different antibiotics,
enabling rapid optimization of therapeutics.
7-11
However, research has also shown that the
results can be strongly dependent on the instrument configuration. Additionally, the
methods are extremely time-consuming and rely on access to advanced instrumentation.
Given the important role that these results play throughout microbiology, a quantitative
rapid technique for determining bacteria concentration in cell media can significantly
impact and improve a wide range of applications.
12

In industrial and microbiology lab settings, optical density (OD) measurements have
become the preferred approach because of their simplicity and rapid time to answer. Based
on optical spectroscopy, an OD measurement characterizes the amount of light that is lost
due to scattering and absorption at a single wavelength.
13
In bacteria analysis, 600 nm is
frequently used, resulting in the nomenclature OD600. It is presumed that the OD
correlates directly with the cell concentration.
14
However, it is known that cells and
bacteria generate byproducts as they grow. Therefore, in addition to optical loss due to the
bacteria, these byproducts also contribute to the signal. Additionally, given the diversity of

34
bacteria morphologies which will give rise to very different scattering profiles, it would
seem advantageous to use a multi-wavelength analysis approach to maximize the signal.
While the OD600 technique is attractive for its simplicity, relying on a single
wavelength can significantly increase the impact of these morphological and environmental
variations on the detected signal. Therefore, taking into account recent advances in the
field of optical spectroscopy and signal analysis, it is critical to rigorously evaluate the
efficacy and accuracy of the current OD600 approach, particularly given previous work
demonstrating discrepancies between different spectrophotometers.
12
By conducting full
spectrophotometric analyses on samples as opposed to only looking at a single wavelength
of light, perhaps it is possible to glean more information on the growth patterns of samples,
thereby shedding more light on the different growth stages.
3.1.3 Bacteria of Interest
To study this hypothesis, we chose to measure the growth kinetics of Pseudomonas
aeruginosa and Staphylococcus aureus. These two bacteria strains were chosen for the
present study for both their similarities and their differences. Their similarities lie in the
manner of infection and virulence, while their differences lie in more fundamental aspects;
namely, their shape, motility, and growth patterns. P. aeruginosa is one of the leading
nosocomial pathogens, responsible for 10-15% of nosocomial infections worldwide, and it
has joined the rank of “superbugs” due to its resistance to practically all antimicrobial
drugs available on the market.
15-17
Additionally, this bacterium is very adaptable, and it is
able to continuously develop new resistance mechanisms as new antimicrobial agents are
created.
17-19
Similarly, S. aureus is a dangerous and versatile pathogen that causes a variety

35
of severe diseases; most frequently skin and respiratory tract infections.
20
S. aureus is the
most common nosocomial pathogen, and it is associated with high morbidity and mortality,
causing clinical disease in 2% of all patient admissions.
21,22
Like P. aeruginosa, S. aureus is
also an extraordinarily adaptable pathogen with a proven ability to develop resistance.
While P. aeruginosa and S. aureus exhibit many similarities, there are also several
key differences between these two bacteria, shown in Figure 3-1. P. aeruginosa is a Gram-
negative rod-shaped bacterium, known for growing in isolation from other colonies. Almost
all strains are motile by means of a single polar flagellum.
23
S. aureus is a Gram-positive,
non-motile, small, round-shaped bacteria, typically found growing in grape-like clusters.
24

Characterizing the growth cycles of P. aeruginosa and S. aureus will lead to a better
understanding of the pathogenesis of infections caused by these bacteria. Additionally,
these two bacteria provide a rigorous and biologically relevant test of our system, allowing
us to verify the universality of the method.

Figure 3-1: Rendering of (a) S. aureus and (b) P. aeruginosa. Adapted with
permission from reference 25, The Optical Society.

36
3.2 Experimental Methods
3.2.1 Preparation of Bacterial Cultures
Four different clinical strains of S. aureus (LAC91, KH38, HH49, and HH36) and one
laboratory reference strain of P. aeruginosa (PA01) were reconstituted from frozen stock.
Clinical strains have a clinical origin, whereas laboratory reference strains have been
subcultured for decades since their first isolation. P. aeruginosa comes from a strain
originally isolated in 1955, and subcultures have been passaged on laboratory media and
shared among microbiological laboratories all over the world since. It is important to note
that, in the course of sequential in vitro passage, laboratory reference strains may have
significantly differentiated from non-passaged clinical samples. After lightly scratching the
surface of the frozen stock with a sterile inoculating loop, bacterial cultures were
suspended in 5 mL broth (TSB) using a vortex mixer and grown in a shaking incubator at
37°C for 24 hours. The culture was kept in a shaking incubator for the duration of the
experiments.
3.2.2 Bacterial Culture Dilution Optimization
At the start of the measurements, the cultures were diluted. Two different dilutions
were used in the present series of measurements.
To perform the colony counting, the cultures were plated at a 10
-10
dilution, diluted
in TSB. This dilution factor was chosen after several trial runs for the purpose of ensuring a
representative number of colonies. If there are too many colonies present when plating,

37
counting them is not possible, and the numbers are too high to be relevant, while if there
are too few colonies, there are not enough for the count to be significant. Additionally, in
order for colonies to form visibly, there must be a minimum number of bacteria. Colonies
were first plated at dilutions of 10
-1
through 10
-5
. These dilutions were too dense to even
differentiate the colonies from each other. Figure 3-2(a) shows HH49, LAC91, and HH36
plated at a 10
-6
dilution. As is clear in the figure, these dilutions were still too high to count.
Figure 3-2(b) shows KH38 and LAC91 at a 10
-10
dilution, HH36 at a 10
-8
dilution, and HH49
at a 10
-9
dilution. From this, we determined that plating at a 10
-10
dilution would ensure the
most accurate colony counts, as shown in Figure 3-2(c). This is the dilution used for
subsequent colony counting measurements.

Figure 3-2: Optical images of the four S. aureus bacteria strain cultures grown on
blood agar plates at different dilutions. (a) All strains plated at 10
-6
dilution. (b)
KH38 and LAC91 plated at 10
-10
dilution, HH36 at a 10
-8
dilution, and HH49 at a 10
-9

dilution. (c) All strains plated at a 10
-10
dilution. Adapted with permission from
reference 25, The Optical Society.

38
For the OD600 and the multi-wavelength normalization measurements, the level of
dilution was 1:40. This dilution was performed in 15-minute intervals, in time with the
spectroscopic measurements. At each interval, 50 µL from the overnight culture was added
to 1950 µL of fresh sterile TSB, resulting in a 1:40 dilution. This concentration and testing
frequency were optimized to match the working range of the spectrophotometers used for
these measurements (Biochrom WPA Spectrawave S1200 Spectrophotometer and Tecan
Magellan Sunrise Spectrophotometer). The first set of dilutions is shown in Figure 3-3 for
the four S. aureus strains where we tested dilutions 1:1, 1:5, 1:10, 1:15, and 1:20 at a single
time point. We then decided to cover a wider range of dilutions, as shown in Figure 3-4,
testing 1:1, 1:3, 1:9, 1:27, and 1:81 every 30 minutes. From here, we decided to focus on a
1:27 dilution (Figure 3-5), increasing the frequency with which each sample was tested –
testing every 15 minutes for 11 hours with three final time points at 23.75, 24, and 24.25
hours. We ultimately chose to proceed with a 1:40 dilution (Figure 3-6), testing every 15
minutes for 11 hours with two final time points at 23.75 and 24 hours, to make the dilution
amounts rounder numbers.

39

Figure 3-3: Dilutions 1:1, 1:5, 1:10, 1:15, and 1:20 tested at a single time point for (a)
HH36 strain, (b) HH49 strain, (c) KH38 strain, and (d) LAC91 strain.

40

Figure 3-4: Dilutions 1:1, 1:3, 1:9, 1:27, and 1:81 tested every 30 minutes for (a)
HH36 strain, (b) HH49 strain, (c) KH38 strain, and (d) LAC91 strain.

41

Figure 3-5: Dilution 1:27 tested every 15 minutes over the course of 24 hours for (a)
HH36 strain, (b) HH49 strain, (c) KH38 strain, and (d) LAC91 strain.

42

Figure 3-6: Dilution 1:40 tested every 15 minutes over the course of 24 hours for (a)
HH36 strain, (b) HH49 strain, (c) KH38 strain, (d) LAC91 strain, and (e) PA01.

43
3.2.3 Colony Counting
Colony-counting measurements were performed as an alternative to OD600. Five
hours into the spectroscopy measurements, the bacteria strains were plated to yield an
approximate concentration. S. aureus strains were plated on blood agar plates. 10 µL was
drawn from each dilution and dropped onto the plate. Using a sterile loop, the bacteria
were spread over a section of the plate to create a streak in one quadrant. This procedure
was repeated with a fresh sterile loop for each of the subsequent S. aureus strains, filling up
each of the four quadrants of the blood agar plate. A similar procedure was used to plate
the P. aeruginosa strain; however, a Pseudomonas Isolation Agar (PIA) plate was used.
Plates were incubated overnight at 37°C, and colonies were counted 24 hours later.
Figure 3-2 shows the optical images of the S. aureus bacteria cultures plated at
varying dilutions. The bacterial concentration for each strain is calculated in colony-
forming units per milliliter, the standard method used for obtaining microbial
concentrations. To determine this value, the number of colonies present is counted. The
number of colonies is then multiplied by the dilution factor and divided by the volume
plated, in milliliters. Given that the dilution factor chosen was 10
-10
and strains were plated
at 10 µL (or 10
-2
mL), we can calculate the concentration in colony-forming units per
milliliter by multiplying the number of colonies by 10
12
. Colony counts for each strain, as
shown in Figure 3-2(c), were as follows: 113 colonies for KH38, 15 colonies for LAC91, 21
colonies for HH36, and 35 colonies for HH49. Therefore, the bacterial concentration is
approximately determined to be on the order of 10
13
- 10
14
colony-forming units per
milliliter for each bacteria strain. However, as can be seen, the size of the different colonies

44
greatly varies, and several of the colonies appear to be merging. Both of these can result in
a large error in this value.
3.2.4 OD600 Measurements
OD600 measurements were performed using a Tecan Magellan Sunrise
Spectrophotometer. This specific instrument is the standard one used for this
measurement in a core facility. At the start of each measurement interval, a background or
normalization spectrum was taken using a sample that only contained the TSB. All spectra
were normalized using this spectrum by subtracting the absorption from the growth media
from the absorption of the cultures for each bacteria strain at each time point. All
measurements were conducted at room temperature and taken at 15-minute intervals for
11 hours, starting 24 hours after inoculation. At each interval, three separate 150 µL
samples were measured, per 1:40 dilution of each strain, yielding three values that were
then averaged to obtain a more representative OD600 measurement. No subsequent data
analysis was conducted on these measurements, per the protocol.
3.2.5 Multi-Wavelength Differential Absorption Spectroscopy
The UV-Vis transmission spectra from the bacterial suspensions were recorded
using a Biochrom WPA Spectrawave S1200 Spectrophotometer. At the start of each
measurement interval, a background or normalization spectrum was taken using a sample
that only contained the TSB. All spectra were normalized using this spectrum by
subtracting the absorption from the growth media from the absorption of the cultures for
each bacteria strain at each time point.

45
All measurements were conducted at room temperature and taken at 15-minute
intervals for 11 hours, starting 24 hours after inoculation, yielding 45 data points for each
strain. The resulting optical spectra were analyzed two ways: (1) the absorption at 600 nm
was determined, and (2) a multi-wavelength normalization analysis was performed. The
absorption at 600 nm is determined from the UV-Vis absorption spectra. These results can
be directly compared with those from the OD600 system.
The multi-wavelength normalization analysis is more complex. First, every
spectrum is converted into a three-point wavelength-averaged spectra. The wavelength
average accounts for minor error or drift in the wavelength accuracy of the spectrometer
between measurements. Then, two wavelengths and the associated absorbances are
identified: (1) a wavelength that experiences a minimum change in absorption over time
(λΔmin), and (2) a wavelength that experiences a maximum change in absorption over time
(λΔmax). In the final step, the normalized absorbance is calculated by dividing the two
absorbance values ((Δαmax)/(Δαmin)).
As mentioned previously, the motivation for performing this self-normalization
process is that the bacteria culture sample is a dynamic system, and, as the bacteria
replicates, it generates byproducts. The signal being measured is the result of both
scattering from particles with diameters larger than the wavelength as well as absorption
from particles smaller than the wavelength. In the present case, these two regimes
correspond to the bacteria (diameter > wavelength) and to the byproducts (diameter <
wavelength). Therefore, a simple background subtraction using the initial growth media
does not account for the byproduct accumulation.

46

3.3 Results and Discussion
Figure 3-7(a) shows a subset of the 45 UV-Vis spectra for the S. aureus strain HH49
after converting to the three-point wavelength-averaged spectra. As can be observed, some
wavelengths experience a greater time-dependent change than other wavelengths.
To more clearly see this behavior, the absorbance values at the six wavelengths
indicated by black arrows in Figure 3-7(a) as well as 600 nm, indicated by a red arrow, are
plotted in Figure 3-7(b) alongside the results from the OD600 measurements. As
mentioned previously, these specific wavelengths were selected to either have minimum or
maximum time-dependent change. Several key points can be concluded based on these
results. First, the change in absorbance at 600 nm actually falls in between these two
extremes and, therefore, is not ideal for detection of bacteria growth. Additionally, all of the
wavelengths with a large change are below 600 nm. When one considers the size of S.
aureus, the reason for this dependence becomes evident – S. aureus is approximately 600
nm in diameter.
26

It is important to verify that better instrumentation is not responsible for the
improved signal at the different wavelengths. To study this parameter, the reading at 600
nm from the UV-Vis is compared to the OD600 measurements. Figure 3-7(c) plots the 600
nm absorbance values from the spectrophotometer and the OD600 measurements together
to show the comparison. The general trends are nearly identical, as is the total signal
change, though there is a constant offset. This constant offset could arise from several

47
different parameters including a wavelength calibration error – for example, if one of the
systems was not precisely measuring 600 nm. However, given that it is a constant offset,
the results between the two systems are in good agreement.

Figure 3-7: (a) UV-Vis spectra for S. aureus strain HH49. Arrows indicate wavelengths
chosen for subsequent analysis. (b) UV-Vis measurements at wavelengths of interest
plotted at select intervals over time. (c) OD600 measurements plotted at select
intervals over time. Error bars are shown for (b) and (c); however, the error is so
small that the error bars are smaller than the symbols. Reprinted with permission
from reference 25, The Optical Society.
Using these results, seven wavelengths were selected for each bacterium based on
the same criteria (Figure 3-8). Based on the total change in absorbance over time (αfinal –
αo), the wavelengths that undergo the maximum change and the minimum change for each
bacterium were identified. The absorption values from these wavelengths were
subsequently used to determine the normalized differential absorption ((Δαmax)/(Δαmin)).
For all four of the S. aureus strains, the wavelength that undergoes the maximum change is
446 nm, and the wavelength that undergoes the minimum change is 787 nm. For P.

48
aeruginosa, 644 nm changes the most while 600 nm changes the least, despite the fact that
this is the wavelength used in the typical OD600 measurements. These results were
reproduced several times (N=3), and the ideal wavelengths associated with each bacterium
remained constant.

Figure 3-8: Change in absorbance over time for seven wavelengths, including 600
nm, exhibiting significant or little change. The specific bacteria plotted are: (a) S.
aureus, HH36 strain, (b) S. aureus, HH49 strain, (c) S. aureus, KH38 strain, (d) S.
aureus, LAC91 strain, and (e) P. aeruginosa, PA01 strain. Error bars are shown on
each plot; however, the error is so small that the error bars are smaller than the
symbols. Reprinted with permission from reference 25, The Optical Society.
Figure 3-9 shows the normalized differential absorption (left axis) and the OD600
measurement (right axis) for each bacterium as a function of time. Several important points
become immediately apparent upon analyzing the results. First, while both the wavelength-
normalized data and the OD600 data show a general increase in absorption with time, for

49
the S. aureus, the functional form of this increase is completely different, as seen in Figure
3-9(a) through 3-9(d). Specifically, for all four strains, the wavelength-normalized data
exhibits a well-defined log growth. In contrast, the OD600 shows a linear trend. It has been
well established that the expected growth curve is logarithmic.
27
Second, for the case of P.
aeruginosa, given the noise in the data, the OD600 is unable to detect growth reliably
because the noise in the data is larger than the signal, whereas the wavelength-normalized
approach easily detects the growth curve with high fidelity (Figure 3-9(e)). Therefore, for
P. aeruginosa, this analysis method is truly an enabling approach.
Using the results shown in Figure 3-9, the doubling time for each bacteria strain is
calculated. Discarding the lag and stationary phases of growth, we fit the linear portion of
the exponential growth curve to a line. The slope of this line is known as µ. Dividing µ into
ln(2) yields the doubling time. Doubling times were as follows: 22.42 ± 2.1 minutes for
LAC91, 33.44 ± 3.0 minutes for KH38, 37.98 ± 5.4 minutes for HH49, 24.92 ± 1.8 minutes
for HH36, and 157.17 ± 19.4 minutes for PA01. S. aureus has previously reported doubling
times of between 24 and 40 minutes,
27,28
while P. aeruginosa has reported doubling times
of between 100 and 200 minutes.
29,30
Therefore, there is excellent agreement between the
current results and previous results in the field. In contrast, the OD600 data is unable to be
fit to an exponential function. As a result, it is not possible to calculate doubling times from
the OD600 measurements for comparison.

50

Figure 3-9: Wavelength-normalized absorption and OD600 for each strain plotted as
a function of time. The specific bacteria plotted are: (a) S. aureus, HH36 strain, (b) S.
aureus, HH49 strain, (c) S. aureus, KH38 strain, (d) S. aureus, LAC91 strain, and (e) P.
aeruginosa, PA01 strain. Error bars are shown on each plot for both wavelength-
normalized absorption and OD600 measurements; however, the error is so small for
the wavelength-normalized absorption that most of the error bars are smaller than
the symbols, while the error is clearly visible for OD600 measurements. Reprinted
with permission from reference 25, The Optical Society.

3.4 Conclusions
The present work proposes and demonstrates that, by implementing a wavelength-
normalization step in the data analysis, the accuracy of characterizing the growth rate of a
bacteria culture will be significantly improved over the conventional OD600 technique. The
proposed method is verified using two distinctly different types of bacteria, P. aeruginosa
and S. aureus, and the results obtained are in good agreement with previous values. In

51
contrast, due to poor absorbance at 600 nm, the classic OD600 measurement method is
unable to detect the growth rate reliably. Our wavelength-normalization protocol to detect
bacteria growth rates can be readily and easily adopted by research labs, given that it only
requires the use of a standard spectrophotometer and implementation of a straightforward
data analysis method. Measuring and monitoring bacteria growth rates plays a critical role
in a wide range of settings, spanning from therapeutic design and development to
diagnostics and disease prevention. Having a full understanding of the growth cycles of
bacteria known to cause severe infections and diseases will lead to a better understanding
of the pathogenesis of these illnesses, leading to better treatment and, ultimately, the
development of a cure.

3.5 References
1 Ede, S. M., Hafner, L. M. & Fredericks, P. M. Structural changes in the cells of some
bacteria during population growth: a Fourier transform infrared-attenuated total
reflectance study. Appl Spectrosc 58, 317-322, doi:10.1366/000370204322886672
(2004).
2 Kim, K. S. & Anthony, B. F. Importance of bacterial growth phase in determining
minimal bactericidal concentrations of penicillin and methicillin. Antimicrob Agents
Chemother 19, 1075-1077 (1981).
3 Rolfe, M. D. et al. Lag phase is a distinct growth phase that prepares bacteria for
exponential growth and involves transient metal accumulation. J Bacteriol 194, 686-
701, doi:10.1128/JB.06112-11 (2012).
4 Black, C. B., Duensing, T. D., Trinkle, L. S. & Dunlay, R. T. Cell-based screening using
high-throughput flow cytometry. Assay Drug Dev Technol 9, 13-20,
doi:10.1089/adt.2010.0308 (2011).
5 Button, D. K. & Robertson, B. R. Determination of DNA content of aquatic bacteria by
flow cytometry. Appl Environ Microbiol 67, 1636-1645,
doi:10.1128/AEM.67.4.1636-1645.2001 (2001).
6 Hazan, R., Que, Y. A., Maura, D. & Rahme, L. G. A method for high throughput
determination of viable bacteria cell counts in 96-well plates. BMC Microbiol 12,
259, doi:10.1186/1471-2180-12-259 (2012).

52
7 Berrier, A., Schaafsma, M. C., Nonglaton, G., Bergquist, J. & Rivas, J. G. Selective
detection of bacterial layers with terahertz plasmonic antennas. Biomed Opt Express
3, 2937-2949, doi:10.1364/BOE.3.002937 (2012).
8 Hamasha, K. et al. Sensitive and specific discrimination of pathogenic and
nonpathogenic Escherichia coli using Raman spectroscopy-a comparison of two
multivariate analysis techniques. Biomed Opt Express 4, 481-489,
doi:10.1364/BOE.4.000481 (2013).
9 Jo, Y. et al. Label-free identification of individual bacteria using Fourier transform
light scattering. Opt Express 23, 15792-15805, doi:10.1364/OE.23.015792 (2015).
10 Jung, G. B., Nam, S. W., Choi, S., Lee, G. J. & Park, H. K. Evaluation of antibiotic effects
on Pseudomonas aeruginosa biofilm using Raman spectroscopy and multivariate
analysis. Biomed Opt Express 5, 3238-3251, doi:10.1364/BOE.5.003238 (2014).
11 Mazhorova, A. et al. Label-free bacteria detection using evanescent mode of a
suspended core terahertz fiber. Opt Express 20, 5344-5355,
doi:10.1364/OE.20.005344 (2012).
12 Naik, P. & D'Sa, E. J. Phytoplankton light absorption of cultures and natural samples:
comparisons using two spectrophotometers. Opt Express 20, 4871-4886,
doi:10.1364/OE.20.004871 (2012).
13 Myers, J. A., Curtis, B. S. & Curtis, W. R. Improving accuracy of cell and chromophore
concentration measurements using optical density. BMC Biophys 6, 4,
doi:10.1186/2046-1682-6-4 (2013).
14 Shao, J., Xiang, J., Axner, O. & Ying, C. Wavelength-modulated tunable diode-laser
absorption spectrometry for real-time monitoring of microbial growth. Appl Opt 55,
2339-2345, doi:10.1364/AO.55.002339 (2016).
15 Blanc, D. S., Petignat, C., Janin, B., Bille, J. & Francioli, P. Frequency and molecular
diversity of Pseudomonas aeruginosa upon admission and during hospitalization: a
prospective epidemiologic study. Clin Microbiol Infect 4, 242-247 (1998).
16 Breidenstein, E. B., de la Fuente-Nunez, C. & Hancock, R. E. Pseudomonas
aeruginosa: all roads lead to resistance. Trends Microbiol 19, 419-426,
doi:10.1016/j.tim.2011.04.005 (2011).
17 Strateva, T. & Yordanov, D. Pseudomonas aeruginosa - a phenomenon of bacterial
resistance. J Med Microbiol 58, 1133-1148, doi:10.1099/jmm.0.009142-0 (2009).
18 McGowan, J. E., Jr. Resistance in nonfermenting gram-negative bacteria: multidrug
resistance to the maximum. Am J Infect Control 34, S29-37; discussion S64-73,
doi:10.1016/j.ajic.2006.05.226 (2006).
19 Pechere, J. C. & Kohler, T. Patterns and modes of beta-lactam resistance in
Pseudomonas aeruginosa. Clin Microbiol Infect 5 Suppl 1, S15-S18 (1999).
20 Deurenberg, R. H. & Stobberingh, E. E. The evolution of Staphylococcus aureus. Infect
Genet Evol 8, 747-763, doi:10.1016/j.meegid.2008.07.007 (2008).
21 Lindsay, J. A. & Holden, M. T. Staphylococcus aureus: superbug, super genome?
Trends Microbiol 12, 378-385, doi:10.1016/j.tim.2004.06.004 (2004).
22 Otto, M. Staphylococcus aureus toxins. Curr Opin Microbiol 17, 32-37,
doi:10.1016/j.mib.2013.11.004 (2014).
23 Lister, P. D., Wolter, D. J. & Hanson, N. D. Antibacterial-resistant Pseudomonas
aeruginosa: clinical impact and complex regulation of chromosomally encoded

53
resistance mechanisms. Clin Microbiol Rev 22, 582-610, doi:10.1128/CMR.00040-09
(2009).
24 Lowy, F. D. Staphylococcus aureus infections. N Engl J Med 339, 520-532,
doi:10.1056/NEJM199808203390806 (1998).
25 McBirney, S. E., Trinh, K., Wong-Beringer, A. & Armani, A. M. Wavelength-normalized
spectroscopic analysis of Staphylococcus aureus and Pseudomonas aeruginosa
growth rates. Biomed Opt Express 7, 4034-4042, doi:10.1364/BOE.7.004034 (2016).
26 Harris, L. G., Foster, S. J. & Richards, R. G. An introduction to Staphylococcus aureus,
and techniques for identifying and quantifying S. aureus adhesins in relation to
adhesion to biomaterials: review. Eur Cell Mater 4, 39-60 (2002).
27 Qiu, J. et al. Subinhibitory concentrations of thymol reduce enterotoxins A and B and
alpha-hemolysin production in Staphylococcus aureus isolates. PLoS One 5, e9736,
doi:10.1371/journal.pone.0009736 (2010).
28 Domingue, G., Costerton, J. W. & Brown, M. R. Bacterial doubling time modulates the
effects of opsonisation and available iron upon interactions between Staphylococcus
aureus and human neutrophils. FEMS Immunol Med Microbiol 16, 223-228 (1996).
29 Goldova, J., Ulrych, A., Hercik, K. & Branny, P. A eukaryotic-type signalling system of
Pseudomonas aeruginosa contributes to oxidative stress resistance, intracellular
survival and virulence. BMC Genomics 12, 437, doi:10.1186/1471-2164-12-437
(2011).
30 Yang, L. et al. In situ growth rates and biofilm development of Pseudomonas
aeruginosa populations in chronic lung infections. J Bacteriol 190, 2767-2776,
doi:10.1128/JB.01581-07 (2008).

54
Chapter 4: Portable Diagnostic for Malaria Detection in
Low-Resource Settings – System Optimization

Despite significant success in therapeutic development, malaria remains a
widespread and deadly infectious disease in the developing world. Given the nearly 100%
efficacy of current malaria therapeutics, the primary barrier to eradication is lack of early
diagnosis of the infected population. However, this population includes not just
symptomatic patients, but also asymptomatic patients. While significant efforts and
resources have been invested in developing methods to diagnose those exhibiting
symptoms and seeking treatment, a rapid and simple broad-population screening method
capable of identifying all malaria carriers has not been realized. Yet, until the entire
malaria-infected population receives treatment, the disease will exist. Here, we report the
development of a portable, self-contained, magneto-optic technology for early-stage
diagnosis of malaria. This chapter focuses on the development of the device and early-stage
experiments. Malaria pigment, known as hemozoin, is a waste product of the malarial
parasite, and it is known to have unique magnetic and optical properties, different from any
other components of blood. Using β-hematin, a hemozoin mimic, we demonstrate detection
limits of <0.0087 µg/mL, corresponding to <26 parasites/µL. This parasitemia level
corresponds to early-stage malaria.

55
4.1 Significance and Background
In spite of tremendous efforts made for its eradication, malaria remains a major
global health problem. It is still a leading cause of morbidity and mortality; however, the
burden of the disease rests primarily in the developing world. Nearly half of the world’s
population currently lives in malaria-endemic regions.
1
Each year, more than 200 million
people are infected, and more than 500,000 individuals die from malaria and its
complications.
2-7
Seventy percent of deaths occur in children aged under five years.
7
These
rates are particularly frustrating given the efficacy of current therapeutics. Specifically, if
an infected person is identified early, treatment is nearly 100% effective when properly
prescribed and administered.
8,9
Therefore, to eradicate malaria, it is necessary to identify
and to treat the currently infected population.
Unfortunately, like typhoid fever, HIV, and many other diseases, this population
consists of two distinct groups: symptomatic and asymptomatic carriers. Current
diagnostic methods have focused on the former group, resulting in the development of a
diagnostic that is unable to completely address this challenge. One need not go far back in
history to know why undiagnosed asymptomatic carriers pose such a problem. Mary
Mallon, known as “Typhoid Mary”, was an asymptomatic carrier of typhoid fever.
10
Mary
was a cook for several families in New York City at the beginning of the twentieth century,
and she also cooked for the soldiers.
11,12
Mary appeared perfectly healthy, but she carried
the infectious agent without becoming sick herself, ultimately infecting 51 people, three of
whom died, before she was quarantined.
10
Asymptomatic chronic blood-stage carriers of
malaria pose a similar threat in that mosquito transmission and, therefore, the spread of

56
the disease remain possible. As such, the development of a broad population-wide
screening method is needed. Such a method must be inexpensive and be able to quickly
identify an infected person to allow for isolation and treatment. In contrast, current
technologies for point-of-care malaria diagnostics are time-consuming, expensive, and
costly; thus, meeting none of the key criteria for a broad-population screening diagnostic.
4.1.1 Current Methods of Diagnosis
The two most commonly used methods are light microscopy, or imaging, and
antibody-based diagnostic tests. Light microscopy of blood smears has been the gold
standard for malaria diagnosis for over a century,
8,9,13,14
and, among methods currently
used, it is the most reliable and sensitive. However, it is low-throughput, requires
significant expertise and training, and is both labor-intensive and expensive as it requires
the use of high-powered microscopes.
1,15,16
To accelerate detection and address these
issues, the rapid diagnostic test (RDT) was developed. These tests rely on the presence of
histidine-rich protein 2 (HRP-2), specific to P. falciparum, and aldolase, a panmalarial
antigen found in all Plasmodium species.
17-20
However, like any antigen-based diagnostic,
reliability is governed by stability of the reagent(s), which, in a low-resource environment,
is a fundamental concern.
21
Therefore, moving away from immunoassay-based approaches
is desirable.
4.1.2 Using Hemozoin as an Indicator of Infection
An alternative strategy presented itself when hemozoin was discovered. Initially
called the malaria pigment,
22
hemozoin is a magnetic nanoparticle byproduct of the

57
parasite formed during the intraerythrocytic growth cycle.
23-28
Notably, unlike all other
naturally occurring materials in the blood, hemozoin exhibits paramagnetic behavior,
responding to low magnetic field strengths. Therefore, if found in a patient’s blood,
hemozoin is indicative of malarial infection.
23,29-34

Figure 4-1: Hemozoin production. (a-d) The formation of hemozoin and subsequent
release into the bloodstream. (a) The parasite (shown in blue) remains in the liver,
reproducing until it is released into the bloodstream. (b) Free heme is generated, as
a byproduct of hemoglobin consumption by the parasite. (c) Heme is aggregated into
an insoluble crystal known as hemozoin. (d) Hemozoin is released into circulation
during erythrocyte lysis. Hemozoin remains in circulation from several days to
weeks, without affecting phagocyte viability. (e) Scanning electron microscopy image
of β-hematin.
As shown in Figure 4-1, the malaria parasite uses hemoglobin as its primary
nutrient source, leading to parasite growth and asexual replication while also generating
monomeric heme, which is highly toxic to the parasite. As the parasite is unable to excrete
the free heme and does not possess a heme oxygenase to recover the iron and detoxify the
heme, heme is converted by the parasite in a crystallization process to form the insoluble
hemozoin.
27,35
The morphology of hemozoin varies depending on the parasite species,
36,37

58
though the nanocrystals typically have an elongated rod-like shape with a length ranging
from 200 nm to 1 µm.
Given its specificity as an indicator for malaria and its unique optical and magnetic
properties, hemozoin is an ideal marker for a malaria diagnostic, and a wide range of
methods have been developed for its detection. Two of the more popular methods with
abilities to consistently detect clinically relevant concentrations of hemozoin are laser
desorption mass spectrometry and Raman spectroscopy.
1,38-40
Other methods include flow
cytometry and polarization microscopy.
41-44
However, similar to conventional smear
optical imaging, these methods are time- and equipment-intensive and require extensive
training on the part of the user to yield accurate results. In parallel with the development of
these detection systems, a hemozoin mimic has been created, known as β-hematin,
allowing for the study of hemozoin without the need to handle malaria-infected samples. β-
hematin and hemozoin from P. falciparum share the same unit crystal structure and the
same magnetic and optical properties, making β-hematin the standard hemozoin mimic
used in the field.
36,45

Leveraging these unique properties of hemozoin, we have developed a portable
optical diagnostic system for malaria detection. Every aspect of the system is optimized for
detection of clinically relevant concentration ranges (<1 to 5 µg/mL
46-48
) in low-resource
environments. Finally, we have developed a predictive model that explains the observed
signals.

59
4.2 Experimental Methods
4.2.1 Synthesis of Fe3O4 Nanoparticles by Chemical Co-Precipitation
To synthesize Fe3O4 nanoparticles, we followed the standard chemical co-
precipitation route.
49
Ferrous chloride (0.4 g) and ferric chloride (1.1 g) were added to
deionized water (20 mL) and heated to 80°C under argon in a three-necked, round-
bottomed flask while magnetically stirring the mixture. Ammonium hydroxide solution (5
mL, 28%) was injected into the solution. It was heated for 1 hour. Oleic acid (1.8 g) was
added, and the entire solution was neutralized with HCl. Fe3O4 nanoparticles were
magnetically decanted and washed three times with deionized water, and then dissolved in
hexane. Further size selection of nanoparticles was done using an Eppendorf 5804
ultracentrifuge. As a final step, the hexane was evaporated, leaving the dry nanoparticulate
residue in powder form.
4.2.2 Synthesis of β-Hematin
To synthesize β-hematin, we first followed a standard protocol found in the
literature.
29
15 mg hemin (Sigma) was dissolved in 3.0 mL of 0.1 M NaOH and stirred in a
glass titration cell connected to a thermostatted bath. 0.30 mL of 1 M HCI and 1.74 ml of
12.9 M acetate (pH 5) pre-incubated at 60°C were added to the solution. After 30 minutes,
the reaction mixture was removed from the cell, cooled on ice for 5 minutes, and then
filtered on an 8 pm cellulose acetate/nitrate Millipore filter type SC and extensively washed
with water. The solid was dried over silica gel and phosphorus pentoxide at 37°C for 48
hours.

60
This protocol proved very difficult to reproduce – specifically, the preparation of the
acetate buffer – which prompted me to reach out to Dr. Timothy Egan, the first author on
the paper publishing this protocol. He confirmed that preparation of the buffer was the
most difficult part, namely because 12.9 M acetate buffer is so close to saturation at
elevated temperatures and is supersaturated at room temperature that small changes in
conditions and the sodium acetate used affect it dramatically. For example, the protocol
works with trihydrate sodium acetate, but not with anhydrous sodium acetate, despite
there being no specifications listed in the protocol.
He then referred me to an updated synthesis protocol his lab developed, a standard
acetate-mediated production route.
50
This protocol used 9.7 M acetate buffer instead of
12.9 M acetate buffer. This is diluted to 4.5 M in the reaction medium, so it does not matter
that a lower concentration stock is used. This protocol was developed specifically to
address reproducibility issues with the original protocol. Hemin (90 mg, Fluka) was
dissolved in 10 mL of NaOH (0.1 M) and neutralized with 1 mL of HCl (1 M). To this, 9.25
mL of acetate buffer (9.7 M, pH 4.8) was added, and the mixture was incubated for 1 hour
at 60°C. Obtaining a pH of 4.8 in the acetate buffer proved difficult, so it is noteworthy that
the ratio of sodium acetate and acetic acid was calculated so as to give a pH of 4.8 based on
the pKa value of acetic acid.
After incubation, the reaction was quenched with water, and the mixture was cooled
over ice. The resulting precipitate was collected via filtration and extensively washed with
water. To remove any unreacted hemin, the air-dried precipitate was placed in a 15 mL
Falcon tube with 1 mL of an aqueous pyridine solution consisting of 5% (v/v) pyridine,

61
40% (v/v) acetone, and 0.02 M HEPES (pH 7.4). To get pure β-hematin requires extensive
washing of the product in this step – a further extension of the procedure, courtesy of Dr.
Timothy Egan.
This well-shaken mixture was diluted to 10 mL with water, centrifuged for 10
minutes, and the supernatant discarded. The resulting precipitate was washed with water
until the supernatant was clear. Finally, the precipitate was collected via filtration (Sigma,
SCWP09025 EMD MILLIPORE) and left to dry in the vacuum oven.
4.2.3 Device Evolution
Every disease is unique – and this uniqueness can often be leveraged to intelligently
design a diagnostic that is inherently specific and selective. In the case of malaria, as the
malaria parasite digests red blood cells, hemozoin is created. Notably, magnetic byproducts
or nanoparticles are not normally present in a healthy patient’s blood. Therefore, the
presence of hemozoin can be directly correlated with the malaria disease state, and, as a
magnetic nanoparticle, it can be collected or moved by magnetic fields.
The sensing instrument is based on differential optical spectroscopy. Specifically, by
monitoring the change in optical power before and after a magnet is applied, we are able to
determine the concentration of nanoparticles present in a solution. This strategy is elegant
in its simplicity as it eliminates the need for a separate normalization step and reduces the
effect of sample-to-sample variation, which is inherent in human specimens.
The first version of the device, developed in early 2016, was based entirely on free-
space optics, meaning, among other things, it was not portable and was expensive to

62
operate. This version, shown in Figure 4-2(a), was truly only meant for proof-of-concept
experiments – to verify that, as the magnet pulls the magnetic nanoparticles out of the
beam path, the optical transparency of the solution increases, manifesting as a change in
optical power. A 980 nm laser was connected to an optical fiber. This optical fiber emitted
980 nm light directed at the sample (a 3 mL polymethyl methacrylate cuvette), and this
light was collected by another optical fiber carefully aligned on the side opposite the
emitting fiber. All of this was connected to a computer that plotted power in real-time. The
magnet was mounted to an optical post and manually moved closer to the sample for
testing.
The second version of the device, developed in late May 2016, was the first quasi-
portable version, shown in Figure 4-2(b). In this version, we originally attempted to use a
handheld laser source emitting 635 nm wavelength (HLS635, ThorLabs). However, there
was a defect in the item, so we used a red laser pointer as the light source, and our detector
was the S120C photodetector from ThorLabs. This was a vast improvement over the first
version of the device as it did not require fine alignment of two optical fibers to obtain a
signal. The photodetector was connected to a powermeter (PM100USB, ThorLabs), which
was then connected to a laptop to plot and record data in real-time. The magnet remained
mounted to an optical post, necessitating manual movement of the magnet, and the sample
remained a 3 mL polymethyl methacrylate cuvette. All of this was screwed onto a
breadboard (MB1012, ThorLabs), making the system entirely portable.

63

Figure 4-2: The beginning stages of the device. (a) Version 1 was all free-space optics,
designed for preliminary experiments. (b) Version 2 was the first semi-portable
version. (c) With the incorporation of a 3D-printed sample holder, we were
streamlining the system. This version was also self-contained within a box, as all
future versions were as well – however, the box was not included in photos for the

64
sake of being able to see the internal workings of the device. (d) The use of a laser
diode was a vast improvement over the laser pointer used, making the device more
user-friendly and allowing for continuous use for up to 36 hours.
In early June, we evolved to version 3, shown in Figure 4-2(c), which came with two
significant improvements – a custom 3D-printed sample holder and a box to enclose the
entire system. The box not only made the system more portable, but it also decreased the
effects of ambient light on the measurements taken, making the device more sensitive.
However, we were still working with a laser pointer, which meant that I not only had to
depress the button myself throughout all measurements taken, but the laser pointer also
died after approximately six hours of use. The next version of the device, Figure 4-2(d),
came quickly with the purchase of a laser diode (CPS635, ThorLabs). Using KAD11NT and
FMP1, both from ThorLabs, to mount and adjust the axes of the laser diode, we were
steadily making the system more user-friendly. We also incorporated a battery pack (CPS1,
ThorLabs) to power the laser diode for up to 36 hours of continuous use.
Finally, we arrived at version 5 – the version used for a significant portion of
subsequent experiments. As shown in Figure 4-3, the most meaningful improvement here
was the incorporation of a linear actuator to automatically bring the magnet in close
proximity to the sample and hold it in position for a set period of time. The sample holder
was also redesigned, allowing for 360° rotation which would mix the sample and
redisperse the nanoparticles within the solution between measurements. This allowed full
automation of the entire system, so we could run tests within the enclosed box at the press
of a button, including mixing the sample and testing multiple times in a row to ensure
reproducibility.

65

Figure 4-3: (a) The live version and (b) a rendering of the fifth version of the device,
used for the majority of further experiments. The incorporation of a linear actuator
and a rotating sample holder were hugely significant.
It is important to note the magnet must be initially located sufficiently far from the
sample such that the magnetic field (B) equals 0 inside the sample cuvette. By decreasing
the gap distance between the magnet and the sample, the magnetic field increases, and this
gap must be decreased quickly enough that the magnetic field strength has two states: on
and off. This approach allows label-free detection of the nanoparticles to be performed in
real-time as well as self-normalization. Initial experiments optimized the magnet strength,

66
which, in turn, determined the system footprint and the detection time. Many different
magnets were tested throughout the process, all of which are detailed in Table 4-1. For
versions 1 through 4, the primary magnet used was ZBX084PC-PNK from K&J Magnetics.
This magnet was originally chosen because it is very strong, thereby ensuring fast
interaction with the magnetic nanoparticles in solution. However, while a stronger magnet
decreases the time-to-signal, enabling higher sample throughput, a stronger magnet also
increases the requisite initial gap distance between the sample and the magnet, thereby
increasing the system footprint. Given that this device is intended for use in low-resource
settings and the goal is for it to be lightweight and portable, we wanted to choose a magnet
that would allow for a manageable footprint and weight while also being strong enough to
interact with the nanoparticles and provide timely results. Ultimately, we chose a magnet
strength for version 5 such that sample throughput was one sample every 8-10 minutes
with an overall footprint under 10” x 12”.
The system weighs fewer than 10 pounds and is entirely self-contained within a
dark plastic box, which not only increases portability but also decreases the effects of
ambient light on the system. The system can operate strictly off of battery power for
approximately eight hours of continuous use. While the battery pack could power the laser
diode for up to 36 hours, the limiting factor is the laptop to which the data is output.
If we refer back to Sections 2.1.2 and 2.1.4, we start to understand why these
considerations are so important. This diagnostic is not being designed for use in a
laboratory or a hospital setting – the intent is to be able to use it in malaria-endemic
regions, so we are tailoring the device to the predicted environment. The most important

67
metrics being considered in the design of this device are sensitivity, power consumption,
initial and operating costs, maintenance, ruggedness, and form factor. Creating a device
that is entirely enclosed within a dark box not only increases sensitivity by decreasing the
affects of noise and other interferents in the environment, but it also contributes to a more
practical form factor by making the device more portable. Because the device is able to
operate solely off of an external battery pack, we keep power requirements low such that
no electrical outlets are needed as electricity is not guaranteed in rural areas affected by
malaria. By using 3D-printed and off-the-shelf components, we are keeping initial costs
low, and operating costs are low due to the fact that the only part of the device that needs
to be replaced for each individual test is the cuvette, which costs on the order of 10 to 15
cents per. Maintenance costs are low in that, if one part of the device malfunctions or
breaks, that individual part can be readily replaced by purchasing another to swap in – the
entire device does not need to be replaced. Ruggedness is an important consideration when
designing a device to operate under harsh conditions. We have designed this system with
the environment of malaria-endemic regions in mind – humidity being one of the more
difficult conditions for many previous diagnostics to battle. Many RDTs have shorter shelf
lives and produce less accurate results when used in hot, humid environments, though this
is not something that affects our system. Lastly, we spent a lot of effort considering the
form factor – that is, the size, weight, shape, and portability of the device. This was
considered when determing what magnet to use for testing as well as in the incorporation
of a plastic box to enclose the entire system. In future generations of the device, we
continue to work on all aspects of the device to make it as user-friendly and readily
adaptable in malaria-endemic regions as possible.

68
4.2.4 Mathematical Modeling
To understand the sensing signal, a mathematical model was created by lab member
Dongyu Chen. While the model is based on optical spectroscopy, it also incorporates the
time-dependent nature of the signal that arises from the interaction of the magnetic field
with the magnetic nanoparticles. As has been well established, the optical signal in optical
transmission spectroscopy is primarily due to scattering and absorption.
51,52
In the present
work, the signal is dominated by the scattering component. As such, the power of the
transmitted light is inversely related to the number of particles in the path of the laser
beam. Mathematical modeling results are included in subsequent figures, showing that our
experimental data is in excellent agreement with the modeling.
4.2.5 Verification of Device
Different methodologies were used with different versions of the device, so here I
will go over, in detail, the protocols used for every test conducted.
For preliminary experiments with version 1, no specific concentrations were used.
Instead, there were simply two solutions of spherical Fe3O4 nanoparticles – one that was
visibly very concentrated, and one that was mostly clear with very few nanoparticles
present. We wanted to show two cases: (1) the optical power changed significantly when
there were more nanoparticles present, and (2) the optical power changed very little when
there were few nanoparticles present. Each sample was placed in line with the system, the
initial power was measured, the magnet (ZBX084PC-PNK, K&J Magnetics) was then pulled
into place, and the change in power was recorded as an absolute number (measurements

69
were not recorded in real-time). The sample was then vortexed to redisperse the
nanoparticles in solution and tested again for a total of seven trials with each sample.
Once this behavior was confirmed, we started to look at different solvents, more
specific dilutions, and different magnet strengths. Water, 50% glycerol, and 75% glycerol
were used as the solutions. Because blood is ultimately the solvent to be used in this device,
we added glycerol to increase the viscosity of the solution in a predictable manner, more
closely mimicking the viscosity of blood. Fe3O4 nanoparticle dilutions of 1:1, 1:2, 1:32, and
1:128 were tested, with the original stock simply being very concentrated nanoparticles.
We also tested magnets of varying strengths and shapes – notably, some of these magnets
had sinkholes or rounded edges. More details for each magnet tested are included in Table
4-1, all sourced from K&J Magnetics. Pull force, as referenced in the table, is the maximum
pull force generated between a single magnet and a thick, ground, flat, steel plate. Again, we
simply wanted to show that there was a change in optical power that corresponded to the
dilution of nanoparticles – the higher the dilution, the lower the change in power – so the
testing protocol was the same as above.
Table 4-1: Specifications of magnets tested.
Low1 Low2 Med1 Med2 High1 High2
K&J
Magnetics
Part Number
ZD202 ZB66 ZD84PC-
PNK
ZBX084PC-
PNK
ZB8 BX0X04-
N52
Dimensions 1/8” dia.
x 1/16”
thick
3/8” sq.
x 1/16”
thick
½” dia. x
¼” thick
1” x ½” x ¼”
thick
1” x ½”
x 1/8”
thick
1” x 1” x
¼” thick
Weight (oz.) 0.003 0.0381 0.198 0.381 0.255 1.08

70
Pull force
(lbs)
0.48 1.67 3.45 8.26 9.67 37.17
Surface field
(Gauss)
4844 1550 3775 3723 2100 3032

We then decided to test two different baseline durations. Up until this point, the
testing set-up allowed 30 seconds for the collection of an initial reading, at which point the
magnet was pulled in closer. We were curious as to whether allowing more time for an
initial reading would affect the overall change in optical power – whether we might get a
more accurate baseline measurement or not. So we looked at 30-second and 60-second
durations for a wider range of Fe3O4 nanoparticle dilutions, this time looking at 1:2, 1:4,
1:8, 1:16, 1:32, 1:64, 1:128, 1:256, and 1:512. We continued to test the same magnets listed
above. Again, we wanted to see if we could correlate higher dilutions to lower changes in
power while also determining whether a longer baseline had an effect.
Version 2 did not see many experiments as we quickly moved on to version 3.
However, version 2 marked a significant milestone in that the system was now semi-
portable, so I ran some experiments to test that the concept behind the device still worked.
The ZD84PC-PNK magnet was used for these experiments, and I simply looked at 1:1 and
1:5 dilutions of Fe3O4 nanoparticles.
Version 3 was the first time that malaria-infected blood samples were tested, in
collaboration with Dr. Kailash Patra at the University of California, San Diego. I originally
began using the ZD84PC-PNK magnet, however, it did not appear strong enough when
interacting with nanoparticles in a more viscous (albeit diluted) blood sample, so I began

71
using the ZBX084PC-PNK magnet. For these experiments, I would begin by recording
power, turning the laser on, then placing the sample in the sample holder. After 30 seconds,
the magnetic field was applied. 30 seconds later, the magnet was removed, the sample was
mixed with a stir bar to redisperse the hemozoin (while remaining in the sample holder
and in the beam path), and the magnetic field was applied again. The magnetic field was
applied a total of three times to each sample to see if we were obtaining similar changes in
power across the tests.
Version 4 did not see any experiments, and then we ran hundreds of experiments
with version 5. We tested solvents of varying viscosities, including water, ethanol, 10%
polyethylene glycol (PEG), 15% PEG, and 25% PEG. Again, PEG was added to the solvent to
increase viscosity. We tested a red laser (CPS635, ThorLabs) and a green laser (CPS532,
ThorLabs). We tested malaria-infected blood at UCSD, spherical Fe3O4 particles, and β-
hematin. Regardless of what sample was being tested, each sample was analyzed using the
diagnostic system in the same manner. First, a background was taken using a null or blank
solution (water, ethanol, 10% PEG, or 15% PEG). Then, 3 mL of the sample was pipetted
into an empty cuvette, and, with the magnet located away from the sample, the box was
closed. A second background was taken for 30 seconds to compare with the initial
background to ensure that changing the samples had not disrupted the system, and then
the magnet was moved into place using the motorized stage. Notably, the box was not
moved during either of these measurements, enabling the continuous acquisition of data
and observation of the onset of magnetic field-induced nanoparticle motion. The total
measurement duration was 10-20 minutes with an acquisition rate of 600 points/minute,

72
for a total of 6,000-12,000 points/measurement. All measurements were performed
multiple times (N≥3) on the same solution to verify reproducibility of the measurement.
Initial experiments with version 5 of the device looked strictly at dilutions of Fe3O4
particles, so there were not specific concentrations being tested. As the device evolved, we
made the concentrations more precise. This not only allowed for more reproducible
results, but it also allowed us to determine the true limit of detection and the full working
range of the device, both of which are important metrics for any diagnostic. The Fe3O4
particles were dried out, weighed, and re-suspended in water, 10% PEG, and 15% PEG at
approximately 4.5 mg/mL. Using serial dilution, a range of concentrations of each solution
was made from 10 µg/mL to <1 µg/mL. The β-hematin particles were re-suspended in 10%
and 15% PEG at 3.93 mg/mL concentration. Due to increased β-hematin size and
aggregation effects, it was not possible to uniformly suspend β-hematin in water. Using
serial dilution, a range of concentrations of each solution was made from 393.3 µg/mL to
0.0087 µg/mL. The concentration units used are μg/mL because this is the standard unit of
measurement for malaria infections and parasitemia levels, allowing us to determine
whether the concentrations being tested were relevant or not – clinically relevant
concentrations of hemozoin in the blood range between <1 and 5 μg/mL, so the goal is for
this device to detect well below 1 μg/mL.

4.3 Results and Discussion
4.3.1 Versions 1 and 2: Optimization of Magnetic Strength and Time

73
Preliminary experiments with the free-space optics set-up showed several
promising results that allowed us to move forward with further versions of the device. A
subset of results is shown in Figure 4-4. All y axes in this figure have units of µW. We were
able to manipulate the magnetic nanoparticles by decreasing the distance between the
magnet and the sample, and this had a direct impact on the signal. Figure 4-4(a) shows the
first set of results obtained using the free-space optics set-up, showing that there was a
correlation between higher concentrations of nanoparticles and a greater change in the
power detected. Figure 4-4(b) through 4-4(e) looks specifically at how stronger magnets
cause a greater change in power, regardless of solvent or nanoparticle dilution factor.

74

Figure 4-4: Early-stage results with versions 1 and 2 of the device. (a) Two dilutions
of spherical Fe3O4 nanoparticles were tested – one visibly higher in concentration
than the other. The sample with higher concentration had a greater change in power.
(b-e) Fe3O4 nanoparticles in two different solvents tested with five magnets of

75
varying shape and field strength over a range of dilutions. Each plot shows that, with
greater field strength, we see a greater change in power, even at the same
concentration. (b) Fe3O4 nanoparticles in water at a 1:128 dilution. (c) Fe3O4
nanoparticles in water at a 1:256 dilution. (d) Fe3O4 nanoparticles in 75% glycerol at
a 1:2 dilution. (e) Fe3O4 nanoparticles in 75% glycerol at a 1:32 dilution. All y axes
have units of µW.
Figure 4-5 delves more deeply into the optimization of magnet strength, showing
results obtained by testing spherical Fe3O4 nanoparticles over a range of dilutions with
magnets of varying strengths. The final magnet selection was designed to optimize: (1) the
speed of acquiring a stable signal and (2) the overall footprint of the system. A stronger
magnet allowed data to be acquired faster, as the magnetic nanoparticles moved in the fluid
faster. However, a stronger magnet could also influence the sample from farther away. As
such, the initial starting point of the magnet or separation distance between the magnet
and the sample was determined by the strength. Therefore, stronger magnets increased the
instrument footprint. We observed that, as expected, with a uniform magnet shape and a
stronger magnetic field, we obtained a better signal.
Figure 4-5 also shows the study of background noise level and the amount of data
needed to accurately characterize this signal. This signal can be influenced by both fast and
slow timescale factors, such as detector noise and laser noise (short term) or laser drift
(slow). There was not a significant difference in overall signal change whether we applied
the magnetic field after 30 seconds or 60 seconds, as shown in Figure 4-5.

76

Figure 4-5: Results for spherical Fe3O4 nanoparticles tested over a range of
concentrations and with magnets of varying field strengths applied after either 30
seconds or 60 seconds of baseline data collection.
Lastly, while we did not conduct many experiments with version 2 of the device
before moving on to version 3, we did, for the first time, record the change in optical power
in real-time as opposed to simply obtaining single values for the two magnetic field states.

77

Figure 4-6: First experiments conducted with semi-portable system, and first time
recording change in optical power in real-time.
4.3.2 Version 3: Malaria-Infected Blood Results
Figure 4-7 shows results from testing malaria-infected blood samples at UCSD. With
this version of the device, since I was using a laser pointer and manually moving the
magnet, sharp peaks in the transmission were often the result of me bumping the testing
set-up, the laser pointer being moved by accident, and/or the stir bar getting in the way of
the light path when mixing the sample between cycles. Samples 16 and 19 represent the
controls – uninfected blood samples – while samples 23, 26, and 27 are infected. Figure 4-
7(a) shows that the controls are reasonably flat when compared to the infected samples, in
which we see increases in transmission when the magnetic field is applied. Figure 4-7(b) is
a histogram plot, allowing us to visualize the data in another format and look specifically at

78
differential power between the two magnetic field states. We can see that the controls
remain tightly spaced around the center 0.0, which is what we would expect since we do
not expect a change in power between when the magnetic field is at a distance and when it
is aligned with the sample. However, the infected blood samples show more significant
changes in transmission as they are further from the center point.

Figure 4-7: Two different visualizations for results from malaria-infected blood
samples. Samples 16 and 19 are controls, while Samples 23, 26, and 27 are infected.
(a) The controls are relatively flat when compared to the infected samples, and there

79
are increases in transmitted power when the magnetic field is applied to infected
samples, as expected. (b) The controls remain close to the 0.0 point, indicating little
change in transmission, while the infected samples are more spread out along the x
axis, indicating more significant changes in transmission.
4.3.3 Version 5: Fe3O4 Nanoparticle Results
The spherical Fe3O4 nanoparticle experiments allowed for further optimization of
system device. First, different solvents were tested of varying viscosities to see how the
nanoparticles behaved in viscous solutions more similar to blood than water. Figure 4-8
shows a subset of these results for water, ethanol, 10% PEG, and 15% PEG. While we
planned to test with 25% PEG, it proved to be too viscous. We can see that, for the most
part, regardless of the solvent, as the nanoparticle dilution increases, the change in signal
decreases, which is to be expected – higher dilutions represent lower concentrations of
nanoparticles. So the fewer nanoparticles there are in the solution (and, by association, the
beam path), the less of a differential there is when the nanoparticles are aligned to the
magnet and pulled out of the beam path.

80

Figure 4-8: Spherical Fe3O4 nanoparticles tested over a range of dilutions in (a)
water, (b) ethanol, (c) 10% PEG, and (d) 15% PEG. As the dilution increases, the
change in signal decreases, as is expected.
Next, we tested a green laser to see if this would yield better results than the
previously used red laser. These results using Fe3O4 nanoparticles are shown in Figure 4-9.
It appeared that the green laser yielded lower overall power; however, the results were
still positive – we saw expected behavior.

81

Figure 4-9: Spherical Fe3O4 nanoparticles tested with a green laser. (a) 1:2000
dilution in water, (b) 1:4000 dilution in water, (c) 1:2000 dilution in 10% PEG, (d)
1:4000 dilution in 10% PEG.
Finally, we began experiments where, instead of using dilutions of Fe3O4
nanoparticles, we precisely weighed them out to determine the limit of detection and
working range of the device and correlate our experimental results to the mathematical
model. The concentration units used are μg/mL because this is the standard unit of
measurement for malaria infections, allowing us to gauge whether the concentrations
being tested were relevant or not – clinically relevant concentrations of hemozoin in the
blood range between <1 and 5 μg/mL. The results from these rounds of testing are shown
in Figure 4-10. The upper and lower bounds are plotted based on +15% and -15% of the
modeling results. Notably, there is excellent agreement between the modeling and the
experimental results within these bounds.

82

Figure 4-10: Experimental and mathematical results for spherical Fe3O4
nanoparticles tested in (a) water, (b) 10% PEG, and (c) 15% PEG. Solid lines show
experimental data; shaded regions show ranges provided by mathematical
modeling.
4.3.4 Version 5: Malaria-Infected Blood Results
Version 5 was also brought back to UCSD for further testing with malaria-infected
blood. This time, both whole and lysed blood samples were tested to see if lysed blood
yielded better results than whole blood. The green laser was also tested at this time –
however, it was unable to penetrate the blood and reach the photodetector on the other

83
side, so work continued with the red laser. A subset of these results is shown in Figure 4-
11, showing that we are seeing noticeable increases in transmission upon application of the
magnetic field, even when compared to the uninfected control. The control does show
oscillation – much more so than solvents used in the lab – however, this is to be expected.
Hemozoin is not the only nanoparticle capable of interacting with the beam path in blood.
Blood contains many small particulates that, while they are not interacting with the
magnetic field, they are floating and suspended in the blood, thereby interacting with the
beam path and affecting transmitted power without being indicative of infection.

Figure 4-11: Results from (a) whole and (b) lysed malaria-infected blood samples.
4.3.5 Version 5: β-hematin Results
A subset of the results from the β-hematin measurements in 10% PEG and 15% PEG
are shown in Figure 4-12(a) and 4-12(b), respectively, with the entire data set shown in
Figure 4-13. The experimental results are fit using our mathematical expression with
excellent agreement. The slight fluctuation in the transmission can be attributed to random
particle-particle interactions. Because of the fundamental nature of these interactions, it

84
was not possible to predict absolute values at any point in time. As such, to capture the
overall transmission change due to the particle motion, lower and upper bounds on the
signal were calculated, giving a ±15% bound around the model fitting results. By fitting the
experimental data, the coefficients f(ρ) and δ1 are determined, and f(ρ) fits a sigmoidal
curve, as expected. Also, for the same type of particle, δ1 is larger for solutions with lower
viscosity, which agrees with its definition.
As expected, the signal is dependent on the concentration of β-hematin present in
the solution. At concentrations as low as 0.0087 µg/mL, we are still seeing a sizeable
increase in transmission upon application of the magnetic field as compared to the base
noise level (signal-to-noise ratios equal 5.03 and 12.41 for 10% PEG and 15% PEG,
respectively). Additionally, the noise does not appreciably increase as the concentration
increases. There is also excellent agreement between the experimental data and the ranges
provided by the simulations.
Figure 4-12(c) and 4-12(d) show the characterization of the working range of the
device. As expected, it has a sigmoidal response. While a sensitivity level of <1 µg/mL may
seem unremarkable when taken in the broader context of ultra-high sensitivity diagnostics,
it is important to remember that the relevant detection concentrations for malaria are <1
to 5 µg/mL. Additionally, this device has shown the ability to detect concentrations a full
order of magnitude below clinical relevance and still differentiate the signal from the noise.

85

Figure 4-12: Experimental and mathematical results for β-hematin tested in (a) 10%
PEG and (b) 15% PEG. The working range of the device for β-hematin concentrations
in (c) 10% PEG and (d) 15% PEG. Insets show the working range of specific
concentrations plotted in (a) and (b).

86

Figure 4-13: Experimental and mathematical results for β-hematin tested in 10%
PEG and 15% PEG. Experimental results are solid lines; mathematical results are
dashed lines. (a) Highest concentrations in 10% PEG, (b) medium concentrations in

87
10% PEG, (c) lowest concentrations in 10% PEG, (d) highest concentrations in 15%
PEG, (e) medium concentrations in 15% PEG, (f) lowest concentrations in 15% PEG.
Figure 4-14 investigates the reproducibility of the instrument’s signal. Each sample
was tested three times to monitor the behavior of the system over three separate cycles. As
can be seen in Figure 4-14, the signal produced by the device is fairly consistent over each
cycle for each concentration tested. This result verifies that the signal produced by the
device is reproducible, which is a key feature of a diagnostic tool.

Figure 4-14: Reproducibility of results for (a) 10% PEG and (b) 15% PEG. The
measurement was performed iteratively using the same set-up.

88

4.4 Conclusions
In conclusion, we have developed a portable, self-contained, malaria diagnostic
based on magneto-optic technology. We took the device from an idea, to a device that
worked but could only be used in a laboratory setting, to a portable device with heightened
sensitivity, low associated costs, and vastly improved form factor. The size, weight, and
power requirements of this system make it ideal for use in low-resource environments as a
screening diagnostic. By exploiting the optical and magnetic properties of the hemozoin
byproduct of the malaria parasite, we hope to detect the presence of the parasite and
thereby diagnose patients with malaria at an earlier stage than other diagnostics currently
used. This device has been shown capable of detecting clinically relevant concentrations of
magnetic nanoparticles in solvents of similar viscosity to blood, and we have verified the
device’s functionality with both mathematical modeling and experiments. Spherical Fe3O4
nanoparticles were tested, as well as a hemozoin mimic known as β-hematin. Ultimately,
this device would be detecting malaria in a patient’s blood sample; therefore, the
immediate next step is testing the device with blood samples to see if detection is still
possible.

4.5 References
1 Demirev, P. A. Mass spectrometry for malaria diagnosis. Expert Review of Molecular
Diagnostics 4, 821-829 (2004).

89
2 Breman, J. G., Alilio, M. S. & Mills, A. Conquering the intolerable burden of malaria:
what's new, what's needed: a summary. Am J Trop Med Hyg 71, 1-15 (2004).
3 Guerin, P. J. et al. Malaria: current status of control, diagnosis, treatment, and a
proposed agenda for research and development. Lancet Infect Dis 2, 564-573
(2002).
4 Kager, P. A. Malaria control: constraints and opportunities. Trop Med Int Health 7,
1042-1046 (2002).
5 Morens, D. M., Folkers, G. K. & Fauci, A. S. The challenge of emerging and re-
emerging infectious diseases. Nature 430, 242-249, doi:10.1038/nature02759
(2004).
6 Phillips, R. S. Current status of malaria and potential for control. Clin Microbiol Rev
14, 208-226, doi:10.1128/CMR.14.1.208-226.2001 (2001).
7 World Health Organization. World malaria report 2016. (World Health Organization,
2016).
8 Nabarro, D. & Mendis, K. Roll Back Malaria is unarguably both necessary and possible
[round table discussion]. (2000).
9 World Health Organization. Regional Office for Europe. Roll Back Malaria highlights
2000-2001. (WHO Regional Office for Europe, 2002).
10 Bartoletti, S. C. Terrible typhoid Mary : a true story of the deadliest cook in America.
(Houghton Mifflin Harcourt, 2015).
11 Brooks, J. The sad and tragic life of Typhoid Mary. CMAJ 154, 915-916 (1996).
12 Marineli, F., Tsoucalas, G., Karamanou, M. & Androutsos, G. Mary Mallon (1869-
1938) and the history of typhoid fever. Ann Gastroenterol 26, 132-134 (2013).
13 Greenwood, B. The molecular epidemiology of malaria. Trop Med Int Health 7, 1012-
1021 (2002).
14 Makler, M. T., Palmer, C. J. & Ager, A. L. A review of practical techniques for the
diagnosis of malaria. Ann Trop Med Parasitol 92, 419-433 (1998).
15 Butykai, A. O., A.; Kocsis, V.; Szaller, D.; Bordács, S.; Tátrai-Szekeres, E.; Kiss, L.F.;
Bóta, A.; Vértessy, B.G.; Zelles, T.; Kézsmárki, I. Malaria pigment crystals as magnetic
micro-rotors: key for high-sensitivity diagnosis. Scientific Reports (2012).
16 Reyburn, H. et al. Overdiagnosis of malaria in patients with severe febrile illness in
Tanzania: a prospective study. BMJ 329, 1212, doi:10.1136/bmj.38251.658229.55
(2004).
17 Bigaillon, C., Fontan, E., Cavallo, J. D., Hernandez, E. & Spiegel, A. Ineffectiveness of
the Binax NOW malaria test for diagnosis of Plasmodium ovale malaria. J Clin
Microbiol 43, 1011, doi:10.1128/JCM.43.2.1011.2005 (2005).
18 Dimaio, M. A., Pereira, I. T., George, T. I. & Banaei, N. Performance of BinaxNOW for
diagnosis of malaria in a U.S. hospital. J Clin Microbiol 50, 2877-2880,
doi:10.1128/JCM.01013-12 (2012).
19 Mboera, L. E. et al. Comparison of the Paracheck-Pf test with microscopy, for the
confirmation of Plasmodium falciparum malaria in Tanzania. Ann Trop Med Parasitol
100, 115-122, doi:10.1179/136485906X78571 (2006).
20 Murray, C. K., Bell, D., Gasser, R. A. & Wongsrichanalai, C. Rapid diagnostic testing for
malaria. Trop Med Int Health 8, 876-883 (2003).

90
21 Ashley, E. A. et al. Evaluation of three parasite lactate dehydrogenase-based rapid
diagnostic tests for the diagnosis of falciparum and vivax malaria. Malar J 8, 241,
doi:10.1186/1475-2875-8-241 (2009).
22 Brown, W. H. Malarial pigment (hematin) as an active factor in the production of the
blood picture of malaria. J Exp Med 18, 96-106, doi:DOI 10.1084/jem.18.1.96 (1913).
23 Bohle, D. S., Kosar, A. D. & Stephens, P. W. Phase homogeneity and crystal
morphology of the malaria pigment beta-hematin. Acta Crystallogr D Biol Crystallogr
58, 1752-1756 (2002).
24 Sherman, I. W. & Hull, R. W. The Pigment (Hemozoin) and Proteins of the Avian
Malaria Parasite Plasmodium-Lophurae. Journal of Protozoology 7, 409-416, doi:DOI
10.1111/j.1550-7408.1960.tb05989.x (1960).
25 Sherman, I. W. & Mudd, J. B. Amino Acid Composition of Hemozoin ( Malaria Pigment
). Journal of Protozoology 11, 23-& (1964).
26 Sherman, I. W., Ting, I. P. & Ruble, J. A. Characterization of the malaria pigment
(hemozoin) from the avian malaria parasite Plasmodium lophurae. J Protozool 15,
158-164 (1968).
27 Slater, A. F. et al. An iron-carboxylate bond links the heme units of malaria pigment.
Proc Natl Acad Sci U S A 88, 325-329 (1991).
28 Yamada, K. A. & Sherman, I. W. Plasmodium-Lophurae - Composition and Properties
of Hemozoin, the Malarial Pigment. Exp Parasitol 48, 61-74, doi:Doi 10.1016/0014-
4894(79)90055-9 (1979).
29 Egan, T. J., Ross, D. C. & Adams, P. A. Quinoline anti-malarial drugs inhibit
spontaneous formation of beta-haematin (malaria pigment). FEBS Lett 352, 54-57
(1994).
30 Gluzman, I. Y. et al. Order and specificity of the Plasmodium falciparum hemoglobin
degradation pathway. J Clin Invest 93, 1602-1608, doi:10.1172/JCI117140 (1994).
31 Goldberg, D. E., Slater, A. F., Cerami, A. & Henderson, G. B. Hemoglobin degradation
in the malaria parasite Plasmodium falciparum: an ordered process in a unique
organelle. Proc Natl Acad Sci U S A 87, 2931-2935 (1990).
32 Orjih, A. U. On the mechanism of hemozoin production in malaria parasites:
activated erythrocyte membranes promote beta-hematin synthesis. Exp Biol Med
(Maywood) 226, 746-752 (2001).
33 Pandey, A. V., Singh, N., Tekwani, B. L., Puri, S. K. & Chauhan, V. S. Assay of beta-
hematin formation by malaria parasite. J Pharm Biomed Anal 20, 203-207 (1999).
34 Tripathi, A. K., Garg, S. K. & Tekwani, B. L. A physiochemical mechanism of hemozoin
(beta-hematin) synthesis by malaria parasite. Biochem Biophys Res Commun 290,
595-601, doi:10.1006/bbrc.2001.6231 (2002).
35 Egan, T. J. Recent advances in understanding the mechanism of hemozoin (malaria
pigment) formation. J Inorg Biochem 102, 1288-1299,
doi:10.1016/j.jinorgbio.2007.12.004 (2008).
36 Noland, G. S., Briones, N. & Sullivan, D. J., Jr. The shape and size of hemozoin crystals
distinguishes diverse Plasmodium species. Mol Biochem Parasitol 130, 91-99
(2003).
37 Oliveira, M. F. et al. Structural and morphological characterization of hemozoin
produced by Schistosoma mansoni and Rhodnius prolixus. FEBS Lett 579, 6010-
6016, doi:10.1016/j.febslet.2005.09.035 (2005).

91
38 Demirev, P. A. et al. Detection of malaria parasites in blood by laser desorption mass
spectrometry. Anal Chem 74, 3262-3266 (2002).
39 Hobro, A. J., Konishi, A., Coban, C. & Smith, N. I. Raman spectroscopic analysis of
malaria disease progression via blood and plasma samples. Analyst 138, 3927-3933,
doi:10.1039/c3an00255a (2013).
40 Scholl, P. F. et al. Rapid detection of malaria infection in vivo by laser desorption
mass spectrometry. Am J Trop Med Hyg 71, 546-551 (2004).
41 Burnett, J. L., Carns, J. L. & Richards-Kortum, R. In vivo microscopy of hemozoin:
towards a needle free diagnostic for malaria. Biomed Opt Express 6, 3462-3474,
doi:10.1364/BOE.6.003462 (2015).
42 Delahunt, C., Horning, M. P., Wilson, B. K., Proctor, J. L. & Hegg, M. C. Limitations of
haemozoin-based diagnosis of Plasmodium falciparum using dark-field microscopy.
Malar J 13, 147, doi:10.1186/1475-2875-13-147 (2014).
43 Jun, G., Lee, J. S., Jung, Y. J. & Park, J. W. Quantitative determination of Plasmodium
parasitemia by flow cytometry and microscopy. J Korean Med Sci 27, 1137-1142,
doi:10.3346/jkms.2012.27.10.1137 (2012).
44 Rebelo, M. et al. A novel flow cytometric hemozoin detection assay for real-time
sensitivity testing of Plasmodium falciparum. PLoS One 8, e61606,
doi:10.1371/journal.pone.0061606 (2013).
45 Jaramillo, M. et al. Synthetic Plasmodium-like hemozoin activates the immune
response: a morphology - function study. PLoS One 4, e6957,
doi:10.1371/journal.pone.0006957 (2009).
46 Butykai, A. et al. Malaria pigment crystals as magnetic micro-rotors: key for high-
sensitivity diagnosis. Sci Rep 3, 1431, doi:10.1038/srep01431 (2013).
47 Chen, K., Yuen, C., Aniweh, Y., Preiser, P. & Liu, Q. Towards ultrasensitive malaria
diagnosis using surface enhanced Raman spectroscopy. Sci Rep 6, 20177,
doi:10.1038/srep20177 (2016).
48 Newman, D. M. et al. A magneto-optic route toward the in vivo diagnosis of malaria:
preliminary results and preclinical trial data. Biophys J 95, 994-1000,
doi:10.1529/biophysj.107.128140 (2008).
49 Gass, J., Poddar, P., Almand, J., Srinath, S. & Srikanth, H. Superparamagnetic Polymer
Nanocomposites with Uniform Fe3O4 Nanoparticle Dispersions. Advanced
Functional Materials 16, 71-75, doi:10.1002/adfm.200500335 (2006).
50 Kuter, D. et al. Insights into the initial stages of lipid-mediated haemozoin
nucleation. CrystEngComm 18, 5177-5187, doi:10.1039/c6ce00866f (2016).
51 Peticolas, W. L., Fanconi, B., Tomlinson, B., Nafie, L. A. & Small, W. Inelastic light
scattering from biological and synthetic polymers. Ann N Y Acad Sci 168, 564-588
(1969).
52 Smith, R. G. Optical power handling capacity of low loss optical fibers as determined
by stimulated Raman and brillouin scattering. Appl Opt 11, 2489-2494,
doi:10.1364/AO.11.002489 (1972).

92
Chapter 5: Portable Diagnostic for Malaria Detection in
Low-Resource Settings – Tested in Whole Rabbit Blood

While Chapter 4 focuses on experiments to develop and optimize our malaria
diagnostic, this chapter focuses on conducting experiments with whole rabbit blood. Here,
we report the second stage of the development of a portable, self-contained, magneto-optic
technology for early-stage diagnosis of malaria. While still using β-hematin, a hemozoin
mimic, we demonstrate detection limits of <0.0081 µg/mL, corresponding to <26
parasites/µL. This parasitemia level again corresponds to early-stage malaria. These tests
were conducted in whole rabbit blood with no sample preparation, making this a viable
device for use outside of a laboratory setting.

5.1 Malaria’s Military Impact
One common misconception is that the impact of malaria is limited to developing
countries. Given the global reach of the U.S. military, malaria has long been considered a
significant threat to the health and effectiveness of military forces. Dr. Stephen Hoffman, a
Navy doctor and former director of the malaria program at the Naval Medical Research
Center, was quoted in 1996 as saying, “In every military campaign this century, we lost
more casualties to malaria than bullets.” This observation is particularly compelling given
that none of the troops were diagnosed with malaria while abroad. However, this is not

93
surprising given the combination of malaria’s long latency period and the short standard
deployment time.
Since the end of the Cold War, U.S. military forces have increased the frequency,
location, and scope of overseas operations, often deploying troops to malaria-endemic
regions, including Afghanistan, Iraq, the Republic of Korea, and various countries in Central
America, South America, and Africa.
1
More than one-third of deployed troops are exposed
to multiple regions endemic for malaria.
1
During the Gulf War, malaria existed in the
Euphrates River Valley of Iraq,
2
and many cases of malaria were reported among troops
who crossed into southern Iraq.
3,4
Operations Enduring Freedom and Iraqi Freedom saw
troops serving either exclusively in Afghanistan or in both Afghanistan and Iraq. Malaria is
endemic in areas of both countries.
5

While standard protocols for preventing malarial infection of troops are in place,
compliance is often an issue. Further complicating diagnostic efforts, if standard malaria
chemoprophylactic medications are being taken, latent infections from relapsing species
may not present until they leave endemic regions and stop taking these medications. Thus,
infections acquired in one area may not manifest clinically until months later, making it
difficult, if even possible, to determine times and locations of various infection
acquisitions,
6
which can hinder outbreak containment. Secondly, field commanders may be
reluctant to report illnesses perceived as trivial in the context of a forward military
operation (such as vomiting and diarrhea).
7
Therefore, rapid, broad spectrum screening
methods for malaria – caused by all five species of Plasmodium – are important to ensure

94
the health of our troops and to reduce the probability of domestic occurrences upon return
from deployment.

5.2 Experimental Methods
5.2.1 Synthesis of β-Hematin by Acetate-Mediated Production
To synthesize β-hematin, we followed the standard acetate-mediated production
route from Section 4.2.2.
8
Hemin (90 mg, Fluka) was dissolved in 10 mL of NaOH (0.1 M)
and neutralized with 1 mL of HCl (1 M). To this, 9.25 mL of acetate buffer (9.7 M, pH 4.8)
was added, and the mixture was incubated for 1 hour at 60°C. After incubation, the reaction
was quenched with water, and the mixture was cooled over ice. The resulting precipitate
was collected via filtration and extensively washed with water. To remove any unreacted
hemin, the air-dried precipitate was placed in a 15 mL Falcon tube with 1 mL of an aqueous
pyridine solution consisting of 5% (v/v) pyridine, 40% (v/v) acetone, and 0.02 M HEPES
(pH 7.4). This well-shaken mixture was diluted to 10 mL with water, centrifuged for 10
minutes, and the supernatant discarded. The resulting precipitate was washed with water
until the supernatant was clear. Finally, the precipitate was collected via filtration and left
to dry over P2O5.
5.2.2 Device Set-Up
The sensing instrument used for these experiments is fundamentally the same as
shown in Figure 4-3, based on differential optical spectroscopy. Specifically, by monitoring

95
the change in optical power before and after a magnet is applied, we are able to determine
the concentration of nanoparticles present in a solution.

Figure 5-1: Schematic of the device. The light from the laser diode passes through the
sample and is detected on the power meter. The concentration of magnetic
nanoparticles is detected by changing the position of the magnet, which is mounted
on a computer-controlled motorized stage. The entire system is enclosed in a dark
box to minimize noise from ambient light.
The device operates in much the same way as before, with one small change. As
shown in Figure 5-1, light from a 635 nm laser diode (CPS635, ThorLabs) passes through a
3 mL polymethyl methacrylate micro-cuvette to a photodetector (S120C, ThorLabs). A
micro-cuvette is used in this set-up as opposed to a cuvette (used in Chapter 4) so as to
allow a smaller sample volume of blood to be tested, following animal use guidelines of
replacement, reduction, and refinement. Up until this point, we have been able to
adequately replace rabbit blood with clear solvents of similar viscosity. However, the

96
diagnostic reached a point where blood replacements were no longer sufficient and we had
to use true blood samples.
Again, it is important to note that the magnet is initially located sufficiently far from
the sample, such that the magnetic field (B) equals 0 inside the sample cuvette. By
decreasing the gap distance between the magnet and the sample, the magnetic field
strength increases. In the present work, this gap was decreased quickly enough that the
magnetic field strength had two states: on and off. This approach allows label-free
detection of the nanoparticles to be performed in real-time as well as self-normalization.
For the present work, the magnet strength was chosen such that sample throughput was
one sample every 8-10 minutes with an overall footprint under 10” x 12”.
5.2.3 Mathematical Modeling
To understand the sensing signal, a mathematical model was created. The model
designed and applied to experiments in Section 4.2.4 held with the whole rabbit blood
experiments as well, allowing us to use the same model for all experiments.
5.2.4 Verification of Device
To validate the diagnostic capability of the system and verify the model in blood
samples, we performed a series of measurements with β-hematin suspended in whole
blood, collected from live Dutch Belted rabbits and mixed with anticoagulant (Camco®
Sequester-Sol® Liquid Anticoagulant) to prevent clotting during transportation to the lab.
A series of solutions was then made, resuspending the β-hematin particles in whole blood

97
at 4.0 mg/mL concentration. Using serial dilution, a range of concentrations was made from
400 µg/mL to 0.0081 µg/mL.
Each sample was then lysed by sonication for 60 seconds
9
and analyzed using the
diagnostic system in the same manner. First, a background was taken using a null or blank
solution (uninfected rabbit blood). Then, 500 µL of the sample was pipetted into an empty
micro-cuvette, and, with the magnet located away from the sample, the box was closed.
Previously, a 3 mL volume was needed for testing, so being able to decrease this volume
six-fold is a significant improvement when considering taking blood from a patient in the
field. A second background was taken for 30 seconds to compare with the initial
background to ensure that changing the samples had not disrupted the system, and then
the magnet was moved into place using the motorized stage. Notably, the box was not
moved during either of these measurements, enabling the continuous acquisition of data
and observation of the onset of magnetic field-induced nanoparticle motion. The total
measurement duration was 10 minutes with an acquisition rate of 60 points/minute, for a
total of 600 points/measurement. All measurements are performed multiple times (N=3)
on the same solution to verify reproducibility of the measurement.

5.3 Results and Discussion
A subset of the results from the β-hematin measurements in whole rabbit blood is
shown in Figure 5-2 with the entire data set shown in Figure 5-3. As expected, the signal is
dependent on the concentration of β-hematin present in the solution, as shown in Figure 5-

98
2(a). At concentrations as low as 0.0081 µg/mL, we are still seeing a sizeable increase in
transmission upon application of the magnetic field as compared to the base noise level
(signal-to-noise ratios equal 4.03). Additionally, the noise does not appreciably increase as
the concentration increases. There is also excellent agreement between the experimental
data and the ranges provided by the simulations.
Figure 5-2(b) shows the characterization of the working range of the device. As
expected, it has a sigmoidal response. Again, while a sensitivity level of <1 µg/mL may
seem unremarkable when taken in the broader context of ultra-high sensitivity diagnostics,
it is important to remember that the relevant detection concentrations for malaria are <1
to 5 µg/mL. This device has shown the ability to detect concentrations a full order of
magnitude below clinical relevance and still differentiate the signal from the noise.
Figure 5-2(c) investigates the reproducibility of the instrument’s signal. Each
sample was tested three times to monitor the behavior of the system over three separate
cycles. As can be seen in Figure 5-2(c), the signal produced by the device is fairly consistent
over each cycle for each concentration tested. This result verifies that the signal produced
by the device is reproducible, which is a key feature of a diagnostic tool.

99

Figure 5-2: (a) Experimental and mathematical results for β-hematin tested in rabbit
blood. (b) The working range of the device for β-hematin concentrations in rabbit
blood. (c) Reproducibility of results. The measurement was performed iteratively
using the same set-up.

100

Figure 5-3: Experimental and mathematical results for β-hematin tested in rabbit
blood. Experimental results are solid lines; mathematical results are dashed lines.
(a) Higher concentrations in rabbit blood, (b) lower concentrations in rabbit blood.

5.4 Conclusions
In conclusion, we have confirmed that the malaria diagnostic developed in Chapter 4
is capable of detecting clinically relevant concentrations of synthetic hemozoin in whole
rabbit blood. This is of great significance for two reasons: (1) conducting optical tests in
blood is no small feat, given the viscosity and opacity of blood as well as the presence of
innumerable particulates, and (2) this device was designed to detect the presence of
malaria in a patient’s blood sample. Now that it has proven capable of detection in whole
rabbit blood and we have again verified the device’s functionality with both mathematical
modeling and experiments, it is becoming more realistic that this device be used in malaria-
endemic regions as a screening diagnostic. By identifying both asymptomatic and
symptomatic populations in malaria-endemic regions, early-stage diagnosis and

101
therapeutic intervention can occur, both of which are crucial to successfully eliminate this
disease.

5.5 References
1 Ciminera, P. & Brundage, J. Malaria in U.S. military forces: a description of
deployment exposures from 2003 through 2005. Am J Trop Med Hyg 76, 275-279
(2007).
2 Young, R. C., Jr., Rachal, R. E. & Huguley, J. W., 3rd. Environmental health concerns of
the Persian Gulf War. J Natl Med Assoc 84, 417-424 (1992).
3 Hyams, K. C. Gulf War Syndrome: potential role of infectious diseases. Curr Opin
Infect Dis 12, 439-443 (1999).
4 Hyams, K. C. et al. Diarrheal disease during Operation Desert Shield. N Engl J Med
325, 1423-1428, doi:10.1056/NEJM199111143252006 (1991).
5 Wallace, M. R. et al. Endemic infectious diseases of Afghanistan. Clin Infect Dis 34,
S171-207, doi:10.1086/340704 (2002).
6 Kim, W. et al. Detection and size measurement of individual hemozoin nanocrystals
in aquatic environment using a whispering gallery mode resonator. Opt Express 20,
29426-29446, doi:10.1364/OE.20.029426 (2012).
7 Matson, D. O. Norovirus gastroenteritis in US Marines in Iraq. Clin Infect Dis 40, 526-
527, doi:10.1086/427508 (2005).
8 Kuter, D. et al. Insights into the initial stages of lipid-mediated haemozoin
nucleation. CrystEngComm 18, 5177-5187, doi:10.1039/c6ce00866f (2016).
9 Brown, R. B. & Audet, J. Current techniques for single-cell lysis. J R Soc Interface 5
Suppl 2, S131-138, doi:10.1098/rsif.2008.0009.focus (2008).

102
Chapter 6: Portable Diagnostic for Malaria Detection in
Low-Resource Settings – Tested with Malaria-Infected
Nonhuman Primate Samples

In the last couple of months of my Ph.D., we started a cross-country collaboration
with Professor Mary Galinski at Emory University. She conducts malaria research using
nonhuman primates, and, despite the fact that our proposal has yet to be funded, she was
generous enough to send me blood samples from nonhuman primates infected with
malaria. Here, we report findings from our first round of experiments, demonstrating
detection limits of 25 parasites/µL. Again, this parasitemia level corresponds to early-stage
malaria. The significance of these results is that we are no longer testing blood doped with
specific concentrations of a synthetic version of hemozoin – we are detecting the presence
of malaria in blood from infected nonhuman primates.

6.1 Experimental Methods
6.1.1 Blood Sample Preparation
Macaca mulatta monkeys, commonly known as rhesus macaques, were infected
with Plasmodium cynomolgi. P. cynomolgi is a nonhuman primate parasite sharing genetic

103
and biological characteristics with P. vivax, including relapses.
1
As a result, it has been used
as a model for P. vivax because the infection in rhesus macaques displays clinical and
parasitological features pertinent to modeling the pathology of P. vivax.
M. mulatta monkeys were inoculated with 2 x 10
5
cryopreserved P. Cynomolgi
sporozoites, injected into the saphenous vein. Parasites were synchronized, meaning only
the ring stage was present in these samples. Parasitemia levels were monitored daily, and
at day 21 after inoculation, when parasitemia reached 5% rings, a blood draw was taken
from the animal. This draw represented 250,000 parasites/µL. The sample was diluted 10x
from 250,000 parasites/µL to 25 parasites/µL with uninfected M. mulatta blood. Samples
were sonicated at 60 Hz for 60 seconds to lyse the red blood cells, rendering the sample
noninfectious, allowing for shipment from Emory University to USC and effectively freezing
the parasite level.
6.1.2 Device Set-Up
The sensing instrument used for these experiments is fundamentally the same as
used in Chapter 5, with a few small changes shown in Figure 6-1. Previously, we were using
the ZBX084PC-PNK magnet from K&J Magnetics. For this version, we decided to use the
BX0X04-N52 magnet (specifications shown in Table 4-1) due to its stronger magnetic field.
The 3D printing designs were changed for this version of the device, meaning that instead
of the magnet being pulled in close proximity to the sample in a linear fashion, there is now
a rotation involved. The starting position of the magnet has the face of the magnet pointed
directly upwards, and, upon starting a test, the magnet is rotated 90° downward to then
face the sample. To allow space for the rotation, the magnet’s resting position is now

104
further away from the sample, meaning we needed a stronger magnetic field to still interact
with magnetic nanoparticles in solution from a further distance.
In addition, the breadboard base was replaced with a 3D-printed base. This allowed
us to custom design the base to make it smaller, decreasing the footprint from 10” x 12” to
4” x 8”. It also significantly decreased the weight of the system – the breadboard weighs
6.55 pounds, and the new base is <1 pound, making the new system lighter and more
portable, decreasing the overall weight from 10 pounds to <5 pounds.

Figure 6-1: Schematic of the device.
6.1.3 Mathematical Modeling
To understand the sensing signal, a mathematical model was created. The model
designed and applied to experiments in Section 4.2.4 held with the nonhuman primate
experiments as well, allowing us to use the same model for all experiments.

105
6.1.4 Verification of Device
To validate the diagnostic capability of the system, each sample was analyzed using
the system in the same manner. First, a background was taken using 500 µL of a null or
blank solution – sonicated blood from an uninfected M. mulatta. Then, 500 µL of the sample
was pipetted into an empty micro-cuvette, and, with the magnet located away from the
sample, the box was closed. A second background was taken for 30 seconds to compare
with the initial background to ensure that changing the samples had not disrupted the
system, and then the magnet was moved into place using the motorized stage. The total
measurement duration was 10 minutes with an acquisition rate of 600 points/minute, for a
total of 6,000 points/measurement. All measurements are performed multiple times (N=3)
on the same solution to verify reproducibility of the measurement.

6.2 Results and Discussion
Results are shown in Figure 6-2. As shown in Figure 6-2(a), the signal is dependent
on the number of parasites present in the blood. The signal seems to saturate at a parasite
count of 25,000 parasites/µL, which is not surprising given the fact that peak parasitemias
rarely exceed 25,000 parasites per μL of blood in humans. At parasite counts as low as 25
parasites/µL, we are still seeing a sizeable increase in transmission upon application of the
magnetic field as compared to the base noise level (signal-to-noise ratios equal 3.52). Early-
stage corresponds to <100 parasites/µL, so a sensitivity level of 25 parasites/µL is quite
remarkable. This device has shown the ability to detect parasite counts nearly a full order

106
of magnitude below clinical relevance and still differentiate the signal from the noise. There
is also excellent agreement between the experimental data and the ranges provided by the
simulations.
Figure 6-2(b) shows the characterization of the working range of the device. As
expected, it appears to have a sigmoidal response. Because we were only testing five
parasite levels between 25 and 250,000 parasites/µL, we need more data points across all
concentrations to have a more accurate fit. We are presently working with our
collaborators at Emory University on the next round of samples, ensuring that we get a
fuller range of parasite counts.
Figure 6-2(c) investigates the reproducibility of the instrument’s signal. Each
sample was tested three times to monitor the behavior of the system over three separate
cycles. As can be seen in Figure 6-2(c), while the signal produced by the device is not as
consistent as previous experiments, it is fairly consistent over each cycle for each dilution
tested. The inconsistencies may be due to hemozoin settling in solution and/or aggregating,
which would affect the system’s ability to reproduce results across multiple runs. To
amend this, we plan to sonicate samples before every measurement in the future, ensuring
that there is no aggregation of hemozoin in solution.

107

Figure 6-2: (a) Experimental and mathematical results for P. cynomolgi-infected M.
mulatta blood. (b) The working range of the device. (c) Reproducibility of results.
The measurement was performed iteratively using the same set-up. Legends for (a)
and (c) show parasite count per µL.

6.3 Conclusions
In conclusion, we have confirmed that the malaria diagnostic developed is capable
of detecting clinically relevant parasite counts in blood obtained from nonhuman primates
infected with malaria. This is a significant step in the progression of this project as, prior to

108
these experiments, we had yet to work with the malaria parasite and/or hemozoin itself,
much less blood samples from animals infected with malaria. We already have plans in
place to improve these results for future experiments, specifically focusing on
reproducibility. The size, weight, and power requirements of this version of the system
make it ideal for use in malaria-endemic regions as a screening diagnostic, which often
have limited access to resources. By exploiting the optical and magnetic properties of the
hemozoin byproduct of the malaria parasite, we have detected the presence of the parasite
at parasitemia levels categorized as early-stage, giving us the ability to diagnose patients
with malaria at an earlier stage than other diagnostics currently used.

6.4 References
1 Joyner, C. et al. Plasmodium cynomolgi infections in rhesus macaques display
clinical and parasitological features pertinent to modelling vivax malaria pathology
and relapse infections. Malar J 15, 451, doi:10.1186/s12936-016-1480-6 (2016).

109
Chapter 7: Elucidating the Mechanism of Blast-Induced
Traumatic Brain Injury

7.1 Significance and Background
Traumatic brain injury (TBI) is often the result of a violent blow or jolt to the head,
or the penetration of an object into the brain tissue. Mild TBI may affect your brain cells
only temporarily, while more severe injuries can result in loss of consciousness,
hemorrhages within the brain, memory loss, and a range of other physical, social,
behavioral, emotional, and psychological symptoms. Broadly speaking, there are two types
of traumatic brain injury – blunt impact, and blast-induced. Blunt impact TBI has been a
topic of much discussion, particularly in recent years, due to collisions observed in football
and the long-term ramifications of repeated blows to the head. However, not much
attention has been paid to blast-induced TBI. It is a common form, often appearing in
troops as the result of an explosion. TBI is known to manifest very differently when
triggered by an explosion as opposed to blunt impact, yet little is known about the precise
mechanism of neuronal death. In this project, I plan to use a sensing platform to conduct
both molecular- and cellular-level experiments in an effort to better understand the
mechanisms of neuronal injury and death resulting from exposure to explosions. This work
will have immediate impact on treatment, and the results could be used in the future to
develop improved preventive measures.

110
7.1.1 Blast-Induced Traumatic Brain Injury
Since the onset of the Global War on Terror, also known as Operation Iraqi Freedom
and Operation Enduring Freedom, traumatic brain injury has been of particular relevance
to both military and civilian health care sectors. TBI, or more specifically blast-induced
traumatic brain injury (bTBI), has been nicknamed the “signature injury” of this war
because it is so prevalent amongst troops and civilians
1
– over 73% of all U.S. military
casualties in this Global War on Terror are caused by explosive weaponry.
2
Media coverage
of casualties in Iraq and Afghanistan have raised awareness of the damaging effects of TBI
and its high incidence among troops returning from those fronts.
Explosions are typically the result of Improvised Explosive Devices (IEDs), and one
thing that has been discovered is that TBI manifests itself very different when triggered by
an explosion as opposed to blunt impact. In the case of a blunt impact, as seen in
automobile accidents, collisions between football players, and other similar incidents
involving a direct blow to the head, TBI is a more typical concussion with localized
inflammation, edema, and neuronal and glial cell death. However, explosive blasts often
result in a high-pressure wave moving out from explosion at a high speed. Blasts begin with
detonation of an explosive material, and the overpressure wave, or blast wave, starts with a
single pulse of increased air pressure that lasts a few milliseconds, and a negative pressure
immediately follows (Figure 7-1).
3
The blast wave progresses from the source as a sphere
of compressed and rapidly- expanding gases, displacing an equal volume of air at high
velocity.
2
The blast wave consists of the front of high pressure that compresses the
surrounding air and falls rapidly to negative pressure and is the main determinant of the

111
primary blast injury. However, a person exposed to an explosion will be subjected not only
to a blast wave but also to the high-velocity wind traveling directly behind the shock front
of the wave.
2

Figure 7-1: Simple physics of a blast wave.
7.1.2 Blast Energy Transfer to the Brain
There are several potential modes for blast energy transfer to the brain. Among
these various hypotheses is the speculation that blast-induced head acceleration and
effects upon the skull may lower cerebral intravascular and intracranial pressures
sufficiently to generate microcavitation bubbles that, in turn, can collapse with pressures
exceeding 1000 atm.
4,5
The proposed “fluid hammer” effect occurs by means of the blast
acting on the thorax and abdomen, forcing blood up into the closed cranial vault and

112
increasing local pressures.
6
This can result in vascular trauma and cerebral hemorrhage,
among other things.
At the moment, there are several hypotheses as to precisely what aspect of the blast
wave results in cellular injury and death. Injuries may result from stresses and strains as
the brain tissue is stretched by the blast wave. Some modeling suggests that the pressure
differentials resulting from reflections within the skull could lead to formation of
microcavitations, posing that cellular injury and death is the result of the formation and
subsequent collapse of microcavitation bubbles. When these bubbles expand and collapse,
they generate tremendous heat and often a high speed jet of fluid, thereby breaching the
cell membrane and destroying the cell.
7
While these forces operate at a very localized level,
a single cavitation event can lyse more cells if the cells are close in proximity.
8
However,
alternate views propose a more molecular-based theory, in which the heat generated by
the bubble collapse initiates a protein-signaling cascade, which ends in cell death.
Understanding the precise mechanism of bTBI is critical in order to develop
predictive models and create improved preventive measures and treatment methods. In
order to isolate and to verify which hypotheses are correct, it is necessary to perform
experiments at all different levels of biological complexity (molecular, cellular, tissue, etc.).
Currently, the primary methods which are being pursued to understand the molecular and
the cellular mechanisms rely heavily on fluorescent methods; for example, using multi-
wavelength fluorescent microscopy to image the deformation of the cell or the production
of specific proteins under different shear rates. However, due to limitations in multi-color
fluorescent imaging, only a few proteins can be monitored simultaneously. By developing

113
an alternative method of monitoring cell behavior in 3D cultures capable of detecting
multiple analytes in real-time, researchers will be able to significantly expand the scope of
their experiments, study multiple parameters simultaneously, and increase the reliability
of their results. Therefore, a comprehensive re-design of current sensing platforms is
needed to fully understand bTBI.
In this chapter, we discuss the beginning steps taken toward developing a multi-tier
approach to address this challenge. This work will have immediate impact on the treatment
of blast-induced neurotrauma, and, ultimately, it could be used to provide guidance in the
development of improved preventive devices.

7.2 Experimental Methods
7.2.1 Neural Slice Preparation
Primary cultures of rat hippocampal neurons are widely used to reveal cellular
mechanisms in neurobiology. Here, we use primary hippocampal neurons derived from
fetal rats to study the neural response to a blast. All neural slice preparation was conducted
in the Arnold Lab at USC.
Sterility is always a factor when growing primary cell cultures, so the greatest
caution is exercised to ensure the most sterile environment possible. A pregnant rat was
euthanized at approximately 16 to 17 days post-fertilization by gassing and decapitation.
While this is the standard operating procedure for the Arnold lab, this is something to

114
consider, given that anesthesia is known to cause brain cell death.
9,10
Using sterile
dissecting scissors and forceps, an opening in the mid-ventral side of the rat was made to
reveal the body cavity. Prenatal pups were then removed, decapitated, and we then
removed the entire brain by opening the cranium of the pup from the back of the neck to
the nose. The cerebellum is removed, and an incision is made down the midline of the brain
to separate it into two hemispheres. The meninges are then removed – while this is not
necessary, the presence of the meninges can make dissection of the hippocampus more
difficult due to the toughness of the membrane. The hippocampus then becomes more
visible as a curved structure that starts in the distal part of the hemisphere and bends
ventrally. The inner, concave side is facing a ventricle and already free, so, to isolate the
hippocampus, one must cut along the convex outer side. After dissection, we gently lift the
hippocampus and transfer into a small tissue culture dish with warmed HBSS (37°C) under
a cell culture hood.
After tissue dissociation and plating, the neural slices were kept in an incubator
until DIV 16 (days in vitro), at which point I picked them up and transported the plates
back to the Armani lab to test. Neurons in culture are typically only considered fully mature
at DIV 14, and DIV 21 is considered very mature, even old, as the cells stop transfecting at
that age. So DIV 16 was a reasonable middle ground for testing the neurons.
7.2.2 Neural Slice Transportation
Successfully transporting the neurons from the Arnold lab to our lab was a project
in and of itself. At the onset of this venture, the students in the Arnold lab leading the
neural preparation were hesitant about my ability to successfully transfer the plated neural

115
slices back to our lab for testing, namely because neurons are incredibly fragile. At that
time, our lab was in Vivian Hall of Engineering, and the Arnold lab was in Irani Hall. There
was construction going on for the building of Michelson Hall between the two, so I had to
take a more circuitous route to and from the Arnold lab to obtain the neurons. This
involved more walking, and walking past very noisy construction zones. The temperature
of the neurons was unstable during transport as I simply transported them in a white
Styrofoam box, and if the temperature of their environment rises or falls too much, the
neurons will die. So there were many factors that made transportation difficult.
After a few failed attempts to bring the neurons back to the Armani lab, I found
some solutions that made transportation more likely to succeed. By transporting them in a
white Styrofoam box with a heating pad inside, placed below the plate, I was able to
maintain the temperature within a reasonable range long enough for me to get the plate
back to our lab and into the incubator, set to 37°C, 5% CO2, and 88-95% humidity. I also
decided to conduct experiments on Friday evenings, when there was not only less foot
traffic from other students, but construction had also ended for the day. Lastly, once the
neurons were in the incubator in our lab, I let them stabilize for a period of time before
testing them as opposed to immediately placing them in the set-up upon arrival. With these
few modifications, I was able to keep the neurons healthy en route to our lab for testing
and confirm successful transport with microscopy.
7.2.3 Blast-Induction Set-Up
Given that simulations are to mimic the phenomena occurring in the brain during
and immediately after an explosion, experiments are to be carried out at internal body

116
temperature. Previous work has shown that there are two viable options for producing
microcavitation bubbles in vitro: near-IR laser, and ultrasound exposure.
While we first attempted to use an ultrasound transducer in the lab to induce
microcavitation, it broke early on in experiments, so we switched to pursuing the near-IR
laser approach. It has been demonstrated that nanosecond laser pulses are capable of
producing microcavitation bubbles that expand and collapse around individual cells.
7
Cells
readily absorb near-infrared radiation. They absorb pulsed laser energy and release this
energy as heat, thus vaporizing surrounding fluid. This generates temporary vapor bubbles
known as microcavitations around the cells, and when these bubbles collapse a few 100 ns
after irradiation, they breach the cell membrane and destroy the cell. Based on previous
work, cell damage by microcavitation appears to precede damage by other thermal
methods when irradiation is delivered in short pulses, so heating of the cell is not expected
to interfere.
7

A rendering of the testing set-up is shown in Figure 7-2. In short, we first place the
six-well plate of neurons on the microscope stage to examine whether the cells appear
healthy after transportation. We then have it arranged such that, when ready to study the
cells’ response to a laser pulse, we deliver a pulse using a 1050 nm laser to each individual
well. The pulse was 50 milliseconds in duration. Before and after delivering the pulse, we
are able to monitor the neurons’ appearance and behavior to watch their response in real-
time.

117

Figure 7-2: Rendering of blast-induction set-up, complete with six-well plate of
neurons on microscope stage.
7.2.4 Verification of Set-Up
Once neurons have been successfully isolated and grown in culture and transported
to the lab, the first step is confirming that the set-up is functioning properly in real-time
and confirming expected cell behavior using immuno-fluorescence. Prior to inducing
microcavitations, calcein-AM is added to the neurons. Viable cells take up the polyanionic
dye internally, at which point calcein-AM is converted to calcein via cellular esterase
enzymes. As long as the cell maintains the integrity of its cellular membrane, the calcein is
unable to diffuse out.
7
Calcein produces an intense uniform green fluorescence in live cells
(excitation/emission ~495 nm/~515 nm), allowing for real-time cell viability assessment
(based on cell membrane integrity) during exposure experiments – cells that fluoresce
green contain calcein and, therefore, have cell membranes that are intact, while cells that

118
no longer fluoresce after exposure to microcavitation bubbles have cell membranes that
have been breached, allowing the fluorescently-labeled calcein to diffuse out.
First, we made a 50 μM working solution of calcein-AM (Catalog #10188-056, VWR)
by diluting in DMSO. 2 μL of calcein-AM was added for every 1 mL of cell suspension in
culture medium, and cells were incubated for 15-20 minutes at room temperature,
protected from light. Cells were then analyzed measuring green fluorescence emission for
calcein, both before and after application of the laser pulse.
7.2.5 Examination of Cellular Processes
Once the cells’ behavior and the functionality of the testing set-up were confirmed,
microcavitation bubbles were again generated within the system, and the response was
detected 30 minutes after inducing microcavitation using a caspase-3 assay (Catalog
#89156-540, VWR). Caspase-3 plays a major role in injury-induced neuronal loss after
traumatic brain injury. Neuronal apoptosis associated with the activation of caspases has
been shown after both human TBI and in various animal models.
11
So if there was an
increase in expression and activation observed after the simulated blast, then the neurons
were responding as expected.

7.3 Results and Discussion
After several failed attempts to successfully transport neurons, once a few changes
were made to my approach, we were able to successfully transport neurons and keep them

119
intact upon arrival to our lab, as shown in Figure 7-3. As extensively discussed with
multiple graduate students in the Arnold lab who study neurons, healthy neurons have a
“look” to them, as do sick neurons. There are many visible signs that a neuron is struggling,
including (but not limited to): engorged somas, exceedingly long and spindly dendrites, too
many dendrites, and/or too few dendrites. None of the neurons below exhibit signs or
symptoms that they are unhealthy.

Figure 7-3: (a-c) Isolated hippocampal neurons, 16 days post-plating, imaged upon
arrival to our lab (after transportation).
The next step was confirming that the set-up worked properly and that the cells
responded to a laser pulse by monitoring calcein-AM expression via green fluorescence. We
first started by looking for green fluorescence within five minutes of inducing
microcavitation. However, this was too short of a timeframe – it is unlikely that the neurons
would respond this quickly to the laser pulse and die within five minutes. We lengthened
this period of time to 30 minutes, giving the neurons enough time to respond, and this
yielded better results. Figure 7-4 shows fluorescence images both immediately before and
30 minutes after exposing the wells to a laser pulse. As is seen in Figure 7-4(a), there are
many green-ish spots present, indicating the presence of live cells based on plasma

120
membrane integrity and esterase activity. However, in Figure 7-4(b), we see virtually
nothing, indicating that this portion of the well is now void of live cells.

Figure 7-4: Live cell imaging using calcein-AM. Scale bars represent 100 μm. (a)
Green spots indicate the presence of live cells before application of a laser pulse to
the well. (b) A lack of green fluorescence indicates the lack of live cells after
application of a laser pulse, showing that the cells lost their membrane integrity
within 30 minutes of the simulated blast.
After several attempts to use the caspase-3 assay, no results were obtained. After
discussions with a student in the Arnold lab who has studied caspase-3, along with other
proteins along the apoptotic pathway, she confirmed that, despite having used many
caspase-3 assays in her career, she has never obtained results from one. She recommended
working more directly with the caspase-3 antibody and studying the neural response in
that manner.
Unfortunately, after just these few rounds of data collection, the Arnold lab stopped
supplying neurons, so this project was unable to move forward due to a lack of materials.

121
7.4 Conclusions
Unfortunately, due to a lack of materials rendering us incapable of continuing with
further experiments, we have simply been able to show that the laser set-up we used was
able to induce apoptosis in the neurons. The timing was unfortunate because I had just had
a conversation with a student who studies neurons, and she had given me some very
helpful information with which to design future experiments. But between that
conversation and the next round of neuron collection, our supply was cut off. These
thought experiments are outlined in Chapter 8.

7.5 References
1 Elder, G. A., Mitsis, E. M., Ahlers, S. T. & Cristian, A. Blast-induced mild traumatic
brain injury. Psychiatr Clin North Am 33, 757-781, doi:10.1016/j.psc.2010.08.001
(2010).
2 Cernak, I. & Noble-Haeusslein, L. J. Traumatic brain injury: an overview of
pathobiology with emphasis on military populations (vol 30, pg 255, 2010). J Cerebr
Blood F Met 30, 1262-1262, doi:10.1038/jcbfm.2009.203 (2010).
3 Owen-Smith, M. Bomb blast injuries: in an explosive situation. Nurs Mirror 149, 35-
39 (1979).
4 Stuhmiller, J. H. Biological response to blast overpressure: a summary of modeling.
Toxicology 121, 91-103 (1997).
5 Stuhmiller, J. H., Santee, W. R. & Friedl, K. Blast injury : translating research into
operational medicine. (Borden Institute, 2008).
6 Cernak, I., Wang, Z., Jiang, J., Bian, X. & Savic, J. Ultrastructural and functional
characteristics of blast injury-induced neurotrauma. J Trauma 50, 695-706 (2001).
7 Mills, B. M. et al. Microcavitation and spot size dependence for damage of artificially
pigmented hTERT-RPE1 cells. P Soc Photo-Opt Ins 5319, 245-251,
doi:10.1117/12.530331 (2004).
8 Williams, J. C., Jr., Woodward, J. F., Stonehill, M. A., Evan, A. P. & McAteer, J. A. Cell
damage by lithotripter shock waves at high pressure to preclude cavitation.
Ultrasound Med Biol 25, 1445-1449 (1999).
9 Stratmann, G., Sall, J. W., May, L. D., Loepke, A. W. & Lee, M. T. Beyond anesthetic
properties: the effects of isoflurane on brain cell death, neurogenesis, and long-term

122
neurocognitive function. Anesth Analg 110, 431-437,
doi:10.1213/ANE.0b013e3181af8015 (2010).
10 Seibenhener, M. L. & Wooten, M. W. Isolation and culture of hippocampal neurons
from prenatal mice. J Vis Exp, doi:10.3791/3634 (2012).
11 Stoica, B. A. & Faden, A. I. Cell death mechanisms and modulation in traumatic brain
injury. Neurotherapeutics 7, 3-12, doi:10.1016/j.nurt.2009.10.023 (2010).

123
Chapter 8: Conclusions and Future Work

In this dissertation, we have detailed the development of signal analysis and optical
instrumentation for biomedical applications. While the projects are seemingly
disconnected, the foundation for each is built on optics – specifically, the study and
manipulation of light – while the applications are all biomedical in nature, looking at the
junction between biology, engineering, and medicine. In this chapter, we will outline
conclusions derived from each project as well as future research directions and new
possibilities resulting from this work.

8.1 Conclusions
8.1.1 Bacteria Project
Understanding and accurately characterizing the different stages of microbial
growth plays a key role in a wide range of fields. To fully map out the bacterial growth
stages, we need a method that can not only take readings quickly, iteratively, and reliably
over long periods of time, but also remain resistant to potentially confounding signals.
While optical density measurements, or OD measurements, have become the preferred
approach due to simplicity and rapid time to answer, relying on a single wavelength of 600
nm can increase the impact of variations on the signal. We developed the hypothesis that,

124
by conducting full spectrophotometric analyses on samples as opposed to only looking at a
single wavelength of light, we might be able to glean more information on the growth
patterns of samples.
In studying Staphylococcus aureus and Pseudomonas aeruginosa, two leading
nosocomial pathogens known as “superbugs” due to their incredible resistance to
antibiotics and high adaptability, we discovered that, by implementing a wavelength-
normalization step in the data analysis, the accuracy of characterizing the growth rate of a
bacteria culture can be significantly improved over currently used OD 600 measurements.
When it comes to equipment needed for this new signal analysis method, our protocol only
requires the use of a standard spectrophotometer, so this is a method that can be readily
and easily adopted.
Measuring and monitoring bacteria growth rates will allow us to have a fuller
understanding of the growth cycles of bacteria, leading to a better understanding of the
pathogenesis of illnesses, leading to better treatments and, ultimately, the development of
cures.
8.1.2 Malaria Project
Malaria is the leading cause of morbidity and mortality, with the burden resting
primarily in the developing world. Each year, neary 500 million people are infected, and
upwards of 1 million individuals die from malaria and its complications. However, when
properly diagnosed and subsequently treated, malaria is a curable disease. Early diagnosis
is the leading factor of decreased morbidity, but half of the 500 million worldwide

125
infections go undiagnosed, and half of those diagnoses are incorrect. Our goal was to
develop a malaria diagnostic that is less expensive, easier to use, more accurate, and
capable of earlier-stage diagnosis than diagnostics currently available.
Every diagnostic must be both specific and sensitive. Our approach derives its
specificity from hemozoin, a magnetic nanoparticle byproduct of the parasite formed
during the growth cycle. If found in a patient’s blood, it is indicative of malarial infection. By
leveraging both the magnetic and optical properties of hemozoin, we developed a
diagnostic tool based on differential spectroscopy to detect the presence of hemozoin in a
sample. We started by studying spherical magnetic nanoparticles and a hemozoin mimic in
clear solvents of various viscosities, eventually leading to detecting clinically relevant
parasite counts in blood drawn from nonhuman primates infected with malaria.
This device has come an exceedingly long way since the start of this project. While
results so far are incredibly promising, there are still many more experiments to be run and
more versions of the device to be tested before this device can transition from the lab into
the hands of those who so desperately need it.
8.1.3 bTBI Project
While there are many different ways in which one can suffer traumatic brain injury,
there are two main types of head injuries – blunt impact, and blast-induced. Blunt impact
TBI is very common, resulting from sports collisions, falls, motor vehicle accidents, and the
like, so it has been the focus of much research over the years. However, very little is known
about the neuronal mechanisms behind blast-induced traumatic brain injury, despite its

126
prevalence among soldiers due to its occurrence as the result of a nearby explosion. Before
a preventive measure can be developed, the neuronal mechanism of death must be
understood.
Here, I design a testing set-up to induce microcavitations and simulate a blast on a
culture of neurons. The goal was to study the cells’ response to the blast by examining the
cellular processes and looking at increased or decreased expression of proteins. While the
testing set-up was proven capable of inducing a blast, unfortunately, our neuron supply
was unable to provide us with neurons anymore, making it impossible to move forward
with this project.

8.2 Future Work for Malaria Project
In this dissertation, I focused on a particular configuration of the malaria diagnostic
in which one magnet is used to pull nanoparticles out of the beam path. This increases the
optical transparency of the solvent, indicating the presence of nanoparticles by an increase
in transmission. Future work will explore the incorporation of two further additions in an
attempt to decrease the limit of detection of the device – one being a dual magnet system,
thereby creating a magnetic trap, and the other being a set of polarizers to change the
detection from transmission-based to polarization-based.
8.2.1 Magnetic Traps

127
Magnetic traps have been successfully used in biosensing applications, including the
rapid separation and detection of foodborne pathogens,
1
as well as the concentration of
particles to maximize target-probe hybridization and detection efficiency.
2-6
The goal of
incorporating a magnetic trap into the PODS is to increase the sensitivity of our device.
Initially, the idea was to create a ring of magnets around the cuvette that were set
up to interact with the magnetic nanoparticles simultaneously, thereby forcing all of them
to the exact center of the magnetic trap system which would align with the beam path.
However, we quickly learned that hemozoin is superparamagnetic, meaning that
magnetization can flip direction.
7
Regardless of how the magnets were aligned, the
hemozoin would always be attracted to the magnets and never repelled, which posed a
unique challenge.
The dual magnet system, or the improvised magnetic trap, is how we plan to
overcome this challenge, depicted in Figure 8-1. To create a similar effect, we use a micro-
cuvette, which is tapered at the bottom. The magnets (SB6C4-IN, K&J Magnetics), mounted
on an actuator on opposite sides of the micro-cuvette, move from the top of the micro-
cuvette to the bottom, thereby pulling the magnetic nanoparticles into an aggregate at the
bottom of the micro-cuvette. Dragging the nanoparticles down increases trapping
efficiency, collecting all nanoparticles present in the solution. Once this aggregate is
formed, the magnet moves back up, concentrating the nanoparticles in the sensor active
area in order to be detected, effectively increasing the nanoparticle concentration. These
trapped nanoparticles are then identified by the beam path.

128

Figure 8-1: An illustration of the improved magnetic trap, with the sample shown in
grey, magnetic nanoparticles in black dots, the laser beam in red, and the magnets in
translucent blue. (a-b) Following the baseline measurement of light transmission,
the magnet, mounted on an actuator, moves down, (c) pulling magnetic
nanoparticles into an aggregate at the bottom of the cuvette. (d) The magnet then
moves back up, (e) pulling the aggregate back into the beam path at the peak of its
movement, where the magnet would be removed for the second measurement.
The hope is that, with the incorporation of this pseudo-magnetic trap, we have more
sensitive and selective detection and increase the sensitivity of the PODS, once again
decreasing the limit of detection and the rate of false test results.
8.2.2 Polarized Light Scattering Spectroscopy
Optical imaging through tissue is a subject of much interest. However, there are
some significant obstacles that limit the efficacy of imaging through tissue – namely, the
scattering properties and the thickness of the tissue. Two things can happen when light
travels through a turbid material: (1) it can be weakly scattered and travel a direct path
through the material, or (2) it can be highly scattered and travel a more circuitous route

129
through the material.
8,9
These two categories are also known as: light due to single
scattering events, and the background due to multiply scattered light (diffusive
background). To study the coherent light, or the single scattering component, we must first
distinguish it from the diffusive background. Extraction of weakly scattered light from the
highly scattered light can decrease noise and amplify the signal.
8-11
Polarized incident light
can be used to accomplish this.
8,11-15
Polarization in this sense has been used in many
applications, including differentiating underlying structures in the imaging of human
skin,
16,17
imaging the spatial distribution of polarized light multiply scattered from cell
suspensions,
10,18
and extracting morphological information about living cells in the
presence of large background from underlying tissue.
9
It is with this knowledge that we
decided to add a modification to the PODS.
While the use of polarized light provides a direct method for background removal,
we must be conscious of where we place the polarizers. It is well known that initially
polarized light loses its polarization when traversing a turbid medium,
19,20
such as blood.
With this in mind, to study the polarized back-scattered light, we placed two linear
polarizers – the first (Polarizer Rotation Mount, 12.7 mm, 5° Grad., 0.5° Sens., M4,
Newport), between the light source and the sample; the second (LPVIS050-MP2,
ThorLabs), between the sample and the detector. The first polarizer polarizes the incident
beam, delivering collimated polarized light to the blood sample. The second polarizer
collects light backscattered from the sample and acts as an analyzer. The analyzer is
rotated to select the polarization state of the backscattered light. The output from this
analyzer is delivered to the detector, which then relays the polarization-based detection in
real-time to a laptop.

130
The hope is that, with the incorporation of this set of polarizers and the switch from
transmission-based to polarization-based detection, we are able to increase the signal-to-
noise ratio of the system, thereby decreasing the limit of detection and increasing the
overall accuracy of the PODS by keeping false-positive and false-negative rates low.
8.2.3 Device Configurations
There are four primary configurations of the PODS that have been designed and are
awaiting validation. The first is that configuration tested in Chapters 4 through 6 – the
single-magnet system with transmission-based detection, shown in Figure 8-2(a). In the
hopes of improving the limit of detection of the device, we have added two modifications to
the original PODS: (1) a magnetic trap, and (2) polarization-based detection. As a result,
there are three additional configurations to be tested, shown in Figure 8-2(b) through 8-
2(d), respectively: (1) a single-magnet system with polarization-based detection, (2) a
magnetic trap with transmission-based detection, and (3) a magnetic trap with
polarization-based detection. Each variation of the PODS still relies on a 635 nm laser diode
as the light source, a photodetector to measure the transmitted light, and magnets supplied
from K&J Magnetics held in place by custom 3D-printed parts. The goal is to further
decrease the limit of detection and keep the rates of false positive and negative test results
low with these additions.

131

Figure 8-2: All device configurations to be tested in the future: (a) A single-magnet
system with transmission-based detection, (b) a single-magnet system with
polarization-based detection, (c) a magnetic trap with transmission-based detection,
and (d) a magnetic trap with polarization-based detection.

8.2.4 Sample Testing

132
For future tests conducted with malaria-infected M. mulatta blood samples obtained
from our collaborators at Emory University, we plan to sonicate the sample before every
measurement to ensure that hemozoin does not settle or aggregate.
When it comes to sample parasitemia levels, we have two approaches, depending on
what works better for our collaborators. One is, we hope to test a fuller range of dilutions.
We tested a range of 250,000 parasites/μL down to 25 parasites/μL, but the dilutions were
10x, meaning we only had five dilutions total. These were also dilutions as opposed to
blood draws taken from the same animal at different time points, or different infection
levels. Ideally, we would like to obtain blood draws from the same animal at different
stages of the infection as opposed to diluting down a high parasitemia level.

8.3 Future Work for bTBI Project
There are many ways in which this project could be continued and innumerable
other variables to study.
For one, I consistently worked with neurons that were DIV 16. Neurons are
considered fully developed at DIV 14, so testing neurons that are both older and younger
would be interesting to see how age affects behavior and response.
Secondly, I studied the neurons while they were still in cell media. Due to the
opacity of cell media, this may not have allowed for full penetration of the laser. Another

133
experiment could be studying the neurons when in PBS, or a clearer medium allowing
fuller penetration.
One factor that may have contributed to the lack of results with the caspase-3 assay
is the fact that I waited only 30 minutes between when the microcavitations were induced
and when I looked for increased caspase-3 expression. Upon speaking with this other
student, it was determined that it is unlikely that we see an increase in caspase-3
expression in such a short period of time. It would be interesting to monitor caspase-3
levels over the course of 24 hours, choosing a few select time points along the way (such as
2 hours and 12 hours) to look for a change.
It is also possible that caspase-3 assays were not working because the cells were not
apoptotic, but simply sick. Caspase-3 is a marker that plays a significant role at the end of
the apoptotic pathway, so it would be interesting to look for markers that show up earlier
in the pathway and see if those are affected by microcavitation.
Blast-induced traumatic brain injury is so pervasive, particularly among our troops,
so severe, and there is so little known about it that it would be incredible to continue this
work. The first step towards developing a preventive measure is to have a deeper
understanding of how the neurons are responding to the pressure wave. Until we
understand the neuronal mechanisms on a more profound level, we will be unable to
decrease bTBI incidents in the field.

134
8.4 References
1 Guo, P. L. et al. Combination of dynamic magnetophoretic separation and stationary
magnetic trap for highly sensitive and selective detection of Salmonella
typhimurium in complex matrix. Biosens Bioelectron 74, 628-636,
doi:10.1016/j.bios.2015.07.019 (2015).
2 Martins, V. C. et al. Femtomolar limit of detection with a magnetoresistive biochip.
Biosens Bioelectron 24, 2690-2695, doi:10.1016/j.bios.2009.01.040 (2009).
3 Graham, D. L. et al. Magnetic field-assisted DNA hybridisation and simultaneous
detection using micron-sized spin-valve sensors and magnetic nanoparticles.
Sensors and Actuators B: Chemical 107, 936-944 (2005).
4 Ferreira, H. A., Cardoso, F. A., Ferreira, R., Cardoso, S. & Freitas, P. P.
Magnetoresistive DNA chips based on ac field focusing of magnetic labels. Journal of
Applied Physics 99 (2006).
5 Li, F., Kodzius, R., Gooneratne, C. P., Foulds, I. G. & Kosel, J. Magneto-mechanical
trapping systems for biological target detection. Microchimica Acta 181, 1743-1748
(2014).
6 Devkota, J. et al. A novel approach for detection and quantification of magnetic
nanomarkers using a spin valve GMR-integrated microfluidic sensor. RSC Advances
5, 51169-51175 (2015).
7 Inyushin, M. et al. Superparamagnetic Properties of Hemozoin. Sci Rep 6, 26212,
doi:10.1038/srep26212 (2016).
8 Gurjar, R. S. et al. Imaging human epithelial properties with polarized light-
scattering spectroscopy. Nat Med 7, 1245-1248, doi:10.1038/nm1101-1245 (2001).
9 Backman, V. et al. Polarized light scattering spectroscopy for quantitative
measurement of epithelial cellular structures in situ. IEEE Journal of Selected Topics
in Quantum Electronics 5, 1019-1026 (1999).
10 Bartel, S. & Hielscher, A. H. Monte Carlo simulations of the diffuse backscattering
mueller matrix for highly scattering media. Appl Opt 39, 1580-1588 (2000).
11 Sankaran, V., Schonenberger, K., Walsh, J. T., Jr. & Maitland, D. J. Polarization
discrimination of coherently propagating light in turbid media. Appl Opt 38, 4252-
4261 (1999).
12 Bicout, D., Brosseau, C., Martinez, A. S. & Schmitt, J. M. Depolarization of multiply
scattered waves by spherical diffusers: Influence of the size parameter. Phys Rev E
Stat Phys Plasmas Fluids Relat Interdiscip Topics 49, 1767-1770 (1994).
13 MacKintosh, F. C., Zhu, J. X., Pine, D. J. & Weitz, D. A. Polarization memory of multiply
scattered light. Phys Rev B Condens Matter 40, 9342-9345 (1989).
14 Morgan, S. P., Khong, M. P. & Somekh, M. G. Effects of polarization state and scatterer
concentration on optical imaging through scattering media. Appl Opt 36, 1560-1565
(1997).
15 Schmitt, J. M., Gandbakhche, A. H. & Bonner, R. F. Use of polarized light to
discriminate short-path photons in a multi- ply scattering medium. Applied Optics
31, 6535-6546 (1992).
16 Anderson, R. R. Polarized light examination and photography of the skin. Arch
Dermatol 127, 1000-1005 (1991).

135
17 Jacques, S. L., Roman, J. R. & Lee, K. Imaging superficial tissues with polarized light.
Lasers Surg Med 26, 119-129 (2000).
18 Demos, S. G. & Alfano, R. R. Optical polarization imaging. Appl Opt 36, 150-155
(1997).
19 Demos, S. G. & Alfano, R. R. Temporal gating in highly scattering media by the degree
of optical polarization. Opt Lett 21, 161-163 (1996).
20 Yoo, K. M. & Alfano, R. R. Time resolved depolarization of multiple backscattered
light from random media. Physics Letters 142, 531-536 (1989).

Asset Metadata

Creator McBirney, Samantha Elizabeth (author)

Core Title Development of optical instrumentation and signal analysis for biomedical applications

Contributor Electronically uploaded by the author (provenance)

School Andrew and Erna Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Biomedical Engineering

Publication Date 06/11/2018

Defense Date 04/18/2018

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

Tag absorption spectroscopy,bacteria,beta-hematin,blast-induced neurotrauma,diagnostics,hemozoin,Magnets,malaria,microbial growth,neurons,oai:digitallibrary.usc.edu:usctheses,OAI-PMH Harvest,OD600,optical instrumentation,optics,sensors,spectroscopy,traumatic brain injury

Format application/pdf (imt)

Language English

Advisor Armani, Andrea (committee chair), Marmarelis, Vasilis (committee member), Wu, Wei (committee member), Yamashiro, Stanley (committee member)

Creator Email samantha.mcbirney@gmail.com,samantha.mcbirney@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-505145

Unique identifier UC11266797

Identifier etd-McBirneySa-6334.pdf (filename),usctheses-c40-505145 (legacy record id)

Legacy Identifier etd-McBirneySa-6334.pdf

Dmrecord 505145

Document Type Dissertation

Format application/pdf (imt)

Rights McBirney, Samantha Elizabeth

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

Abstract (if available)

Abstract Optics first originated as a field of study centuries ago with the development of lenses by the ancient Egyptians and Mesopotamians. They never could have imagined how these early lenses would evolve and what opportunities they would open up with this field. Practical applications segued into theory, and all of this has catapulted us into our current state. Optics is presently known as the branch of physics that involves the behavior and properties of light, including its interactions with matter and the construction of instruments that use or detect it. This dissertation details the development and application of new tools created to further our understanding of a variety of topics, ranging from bacteria to malaria to blast-induced traumatic brain injury. Each project utilizes various properties and capabilities of light to study topics that are, otherwise, seemingly disconnected. ❧ Here, I will share work I have completed to-date and ideas for future work to be passed on to others. In the first part of my dissertation, I explain a novel method I developed to determine microbial growth using multi-wavelength spectroscopic measurements. Measuring and monitoring bacterial growth rates plays a critical role in a wide range of settings. Having a fuller understanding of the growth cycles of bacteria known to cause severe infections and diseases will lead to a better understanding of the pathogenesis of these illnesses, leading to better treatment and, ultimately, the development of cures. ❧ The second section of my thesis details what has been my primary focus for the last two years—the development of a malaria diagnostic. Exploiting the magnetic properties of a byproduct of the malaria parasite present in the bloodstream, we have developed a device capable of detecting malarial infection at clinically relevant concentrations. The implications of this device are incredibly far-reaching —if successful, this tool could save hundreds of thousands of people every year who die from malaria due to inaccurate or late diagnosis. I am hoping that one of my mentees will continue this work once I am gone in an effort to fully transition this device from the lab to reality. ❧ The final section of this dissertation details my study of blast-induced traumatic brain injury. Presently, little is known about this specific type of traumatic brain injury, despite it causing a multitude of injuries and deaths, particularly among troops. I have studied the properties of a system representative of the brain in an effort to see how the system changes upon being exposed to a laser-induced blast. The hope is that, once the mechanisms of neuronal death due to blast-induced neurotrauma have been clarified, we may then use these results to design a preventive measure to avoid these injuries altogether.