Citation
Optical coherence tomography for brain tissue analysis

Material Information

Title:
Optical coherence tomography for brain tissue analysis
Creator:
Challinor, Andrew James
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English

Thesis/Dissertation Information

Degree:
Master's ( Master of science)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Bioengineering, CU Denver
Degree Disciplines:
Bioengineering
Committee Chair:
Gibson, Emily
Committee Members:
Restrepo, Diego
Welle, Cristin

Notes

Abstract:
Ever since the emergence of Optical Coherence Tomography (OCT) as a tool for ophthalmic imaging, the additional applications of the technology for medical imaging have been expanding rapidly. The two basic techniques, time domain (TD) and spectral domain (SD) have both been applied in different situations. SD OCT has become the more commonly used method for medical applications, mainly because of its ability to produce high speed images, including real time, with less moving parts than TD OCT. In the field of brain imaging, TD OCT holds considerable promise in the ability to probe deep within a live brain without significant damage, and to produce high quality images of brain tissue, which can be of considerable benefit to neurosurgeons. OCT has the ability to not only produce high resolution (~1 micron) images, but also to provide topographical data of the area under observation, allowing 3-D images of the tissue to be constructed. The method images the tissue directly, and so does not require labelling with radioactive materials or other contrast agents, making it much safer for the patient also. This thesis demonstrates the fundamental principles required for the construction of an SD OCT system which is capable of producing 3-D scans of brain tissue slides with sufficient resolution to identify cellular damage and inclusions such as Lewy bodies.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
Copyright Andrew James Challinor. Permission granted to University of Colorado Denver to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

Downloads

This item has the following downloads:


Full Text
OPTICAL COHERENCE TOMOGRAPHY FOR BRAIN TISSUE ANALYSIS
by
Andrew James Challinor B.Sc.(Hons) Teesside University, 1983
A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Science Bioengineering Program
2017


©2017
ANDREW JAMES CHALLINOR
ALL RIGHTS RESERVED


This thesis for the Master of Science degree by Andrew James Challinor has been approved for the Bioengineering Program
by
Emily Gibson, Chair Diego Restrepo Cristin Welle
Date: July 29 2017


Challinor, Andrew James (MS Bioengineering)
Optical Coherence Tomography for Brain Tissue Analysis Thesis directed by Associate Professor Emily Gibson
ABSTRACT
Ever since the emergence of Optical Coherence Tomography (OCT) as a tool for ophthalmic imaging, the additional applications of the technology for medical imaging have been expanding rapidly. The two basic techniques, time domain (TD) and spectral domain (SD) have both been applied in different situations. SD OCT has become the more commonly used method for medical applications, mainly because of its ability to produce high speed images, including real time, with less moving parts than TD OCT.
In the field of brain imaging, TD OCT holds considerable promise in the ability to probe deep within a live brain without significant damage, and to produce high quality images of brain tissue, which can be of considerable benefit to neurosurgeons.
OCT has the ability to not only produce high resolution (~1 micron) images, but also to provide topographical data of the area under observation, allowing 3-D images of the tissue to be constructed. The method images the tissue directly, and so does not require labelling with radioactive materials or other contrast agents, making it much safer for the patient also.
This thesis demonstrates the fundamental principles required for the construction of an SD OCT system which is capable of producing 3-D scans of brain tissue slides with sufficient resolution to identify cellular damage and inclusions such as Lewy bodies.
The form and content of this abstract are approved. I recommend its publication.
Approved: Emily Gibson
IV


TABLE OF CONTENTS
I. INTRODUCTION........................................1
BACKGROUND...........................................1
OPTICAL COHERENCE TOMOGRAPHY OVERVIEW................6
APPLICATIONS OF OCT.................................12
II. SYSTEM DESIGN.....................................15
HARDWARE............................................15
SOFTWARE............................................20
STAGE CONTROL SOFTWARE..............................20
DATA ANALYSIS SOFTWARE..............................20
III. RESULTS..........................................22
IV DISCUSSION.........................................25
V FUTURE WORK........................................28
HARDWARE............................................28
SOFTWARE............................................29
REFERENCES............................................30
APPENDIX..............................................36
v


LIST OF FIGURES
FIGURE 1 - Pathways in the basal ganglia.......................................1
FIGURE 2 - Absorbance of Hemoglobin and Water.................................11
FIGURE 3 - Sample Commercial OCT System.......................................13
FIGURE 4 - Schematic of Michelson Interferometer..............................15
FIGURE 5 - Initial OCT System Layout..........................................18
FIGURE 6 - Current OCT System Layout..........................................19
FIGURE 7 - Stacked Cover Slip Arrangement.....................................22
FIGURE 8 - Stacked Cover Slips................................................23
FIGURE 9 - Fingerprint Photographs (standard size and enlarged)...............24
FIGURE 10 - Fingerprint.......................................................24
FIGURE 11 - Fourier Data of Slide Only......................................26
FIGURE 12 - Fourier Data of Slide and Single Coverslip......................26
FIGURE 13 - Fourier Data of Slide and Two Coverslips........................26
VI


CHAPTER I
INTRODUCTION
BACKGROUND
Parkinson’s disease affects the substantia nigra section of the brain, which is located in the midbrain region. It forms one of the components of the basal ganglia, and is the main area for dopamine production. Parkinson’s causes the dopanergic neurons in this area to atrophy, which disrupts the signals to other parts of the brain which control skeletal muscular actions1. This results in the typical tremors that are seen in Parkinson’s patients. A block diagram of the basal ganglia components and pathways is shown in Figure 1. The effects of the damage to the substantial nigra are indicated.
Cerebral cortex Frontal cortex
Degenerated
Jn P
Substantia * nigra 1 pars ,
o
tacreased—1► O O Diminished
Q I More tonic I inhibition
Globus pallidu>, Globus pallidus,
external segment internal segment
o
Subthalamic Q
nucleus
Increased
NEUROSCIENCE, Fourth Edition, Hgur* 18.11 (Pan 1)
Purves, D. (2012). Neuroscience. Sunderland, Mass: Sinauer Associates.
FIGURE 1 - Pathways in the basal ganglia
1


For the most part, current treatment for Parkinson’s is palliative care, as there is no current cure. Medications can help the symptoms of the disease, and these include drugs such as Levodopa and Carbidopa, which is an amino acid which increases dopamine concentrations in the brain. Other drugs such as apomorphine act as dopamine agonists. This drug can relieve Parkinson’s symptoms quickly, but is also short lived.
A more effective and longer term treatment for Parkinson’s is deep brain stimulation (DBS), which involves the implantation of electrodes into the sub thalamic nucleus which are connected to a subcutaneous device which provides electrical pulses to these areas2. The strength and frequency of the pulses are controlled. The controller, which can be set to deliver a wide range of stimulus signals.
Regrettably, this is still not a cure for Parkinson’s, and this method is not always suitable for the patient, but it is the most effective treatment currently available.
One of the main drawbacks of the DBS method lies in the method of implanting the electrodes in the patient. At the present time, there is no method of producing real time images of deep brain structures, although OCT systems have been developed to produce high frame rates using fixed position scanning (3 4). Currently an electrical method is used to aide in electrode placement. This technique consists of the surgeon inserting an electrode that reads out the local field potential activity, and this is converted into a visible and audible signal for the neurosurgeon. This is done in combination with MRI and CT scans to help guide the placement of the probe to the correct area, using stereotactic methods. Once the probe is in the correct area of the
2


brain, the surgeon then listens for changes in the audible signal to indicate areas of the brain which have been damaged, and use this to guide the placement of the electrodes.
There is no direct visualization of the affected area during the procedure, and so the electrode placement is somewhat inaccurate. In order to achieve the correct placement, several attempts may be needed, so prolonging the surgery. At times, a repeat surgery is needed to reposition the electrodes to a better location.
Parkinson’s typically develop in older adults, usually between the ages of 55 and 655, and around 0.3% of the population may be affected to some degree. The incidence among men is generally higher than women9.
The disease’s main symptoms are muscle tremors, stiffness, slow movement and balance problems7. Additionally, softening of the voice and small handwriting are common, as are mood swings including depression and anxiety. The risk of death for sufferers of any level is almost double that of non-sufferers. Because of the range of symptoms, of which most patients will have a combination of, the majority of Parkinson’s patients will not be able to maintain employment of any kind8. The cost of standard drug treatment is between $1,000 and $6,000 per year, with additional healthcare costs of between $2,000 to $20,000 per year9'10.
Treatment by a neurologist has been statistically shown to decrease the rate of deaths, and of the removal of patients from their homes into nursing facilities11.
The disease itself is a neurodegenerative syndrome, which involves several motor and non-motor neural bundles in the basal ganglia7'12'13 The disease typically disrupts the brain signaling system13'14, causing erratic firing patterns in the neurons, which in turn cause the tremors and rigidity commonly seen in patients.
3


At the beginning of treatment, patients generally have a good response to one of the drug therapies available, however after around five years or so, a range of drug related complications tend to start occurring. Many drug therapies see a gradual loss of effectiveness over time, however in Parkinson’s disease the patients frequently develop a problem where the drugs become largely ineffective. The loss of effectiveness of drug therapy can occur suddenly and unpredictably, which is an obvious concern for the patient and their safety, as the symptoms can include suddenly increased tremors, rigidity, loss of balance, speech, and swallowing ability6'15. Many patients become increasingly resistant to drugs such as the often used carbidopa-levidopa combination, which may leave them with few alternatives to alleviate their symptoms.
Patients that are currently responsive to levodopa and other factors are considered the best potential candidates for deep brain stimulation (DBS) treatment16. The area of the brain that is damaged by the disease is the substantial nigra which is responsible for dopamine production. In this region there is generally cellular damage, and inclusions called Lewy bodies appear in the cells. DBS stimulation of this area is not effective, since the cell damage has already occurred. The electrodes are typically implanted into the sub thalamic nucleus of the brain. The electrodes are connected to a device which can give an electrical stimulus to the brain, which can be varied in intensity and frequency. The effect of the electrical stimulation is to inhibit cell activity, and increasing fiber activity17 2. In particular, the firing rate and pattern of individual neurons is affected2, along with the localized release of dopamine and glutamate19'20 21, which helps control the patient’s tremors. An added bonus is that the stimulation also increases blood flow and stimulates neurogenesis22.
4


Although the individual effects of DBS are well know, it is still unclear how these actually affect Parkinson’s disease as a whole. However, the documented effect results of the treatment are well established over a range of patients.
DBS was approved by the Food and Drug Administration (FDA) in 2002 for use in Parkinson’s patients whose symptoms are not adequately controlled by medication.
The centers where the treatment is available generally select patients based a number of factors, but which generally include an inadequate response to levodopa, on-off fluctuations and tremors.
The procedure itself starts with taking a set of images of the patient’s brain using CT and MRI scans. Once the surgery starts, the surgeon drills a burr hole in the patient’s skull for the insertion of the electrical probe, and the quadrature deep brain stimulation electrodes. General anesthesia is not normally used, as the patient’s reactions during the procedure need to be monitored. Once the surgeon considers that the stimulation electrode is suitably placed, the patient’s reactions to different levels of stimulus can be monitored, particularly for reduction in tremors and rigidity, and also any adverse effects which may occur.
The costs for this type of surgery can vary anywhere from $28,000 to $50,000. In addition, there may be additional costs for controller setting adjustments24. In some case, the patient may have a poor response to the treatment, which is often caused by suboptimal positioning of the electrodes25. These can often be rectified, but will generally require additional surgery, either to reposition or to remove and replace the electrodes. This obviously increase the overall cost, and stress to the patient.
5


Although there can be difficulties, there is generally a positive outcome for the patient in the form or reduced tremors and rigidity, and also reduced levels of medication required. Some patients have had symptom improvement lasting more than 10 years23. The addition of Optical Coherence Tomography imaging to the DBS surgical process would give the surgeon real time imaging capability. This could help reduce the misplacement of the electrodes, reduce the the overall time taken for the surgery, and allow the patient to be anesthetized.
OPTICAL COHERENCE TOMOGRAPHY OVERVIEW
Optical Coherence Tomography (OCT) is an imaging method which can produce both 2 dimensional and 3 dimensional images with high resolution. It was pioneered by James Fujimoto, Eric Swanson and their team at MIT26. They originally developed the technology as a method of imaging live structures within the eye.
Since the technology first appeared in 199126, it has developed rapidly and has begun to enter a range of fields where previously imaging was difficult or impossible. The original intent was for use in ophthalmology, it is now spreading to many other areas of medical imaging.
The first commercial OCT system was produced by Carl Zeiss Meditec in 1996, and at present there are over 20 other manufacturers producing commercial systems27. The first systems based on fiber optics for intravascular imaging have only recently appeared, and are now approved not only in Europe and Asia, but the US also28'29. Other commercial applications are being developed all the time. One of the most recent is for the use in cardiovascular surgeries, which uses fibre optic technology as the coupled with gradient index (GRIN) lens technology, which allows for 3-D imaging within
6


arteries. This development has been helped by the increases in speed of imaging processors, and the reduction of component sizes in the devices29.
The advantages of OCT imaging are that it can produce 3-dimensional images with resolutions of ~1 pm30'31 â– 32 but with a tissue penetration of several millimeters, dependent on the tissue scattering parameters of the sample being imaged, it has better imaging depth than conventional confocal microscopy33. Typically the light source used is in the near infrared, which is non-ionizing, and so safe for repeated examinations of the same sample. With the appropriate electronic controller and image processor, OCT is also capable of real time imaging.
The main principles behind OCT is the use of an interferometer that detects reflected light from the imaged object and compares this with a reference beam to allow the depth profile of the sample to be calculated. This technique in conjunction with the long center wavelength and incoherent light give OCT imaging it’s unique properties over other imaging methods, such as wide field and confocal microscopy. Tissue scattering is reduced at longer wavelengths34.
There are several methods of implementing OCT. Each method contains some form of interferometer, and the main detection methods are either CCD camera or photo diode for time domain (TD) OCT, or by spectrometer for spectral domain (SD) OCT. TD OCT systems were the first to be implemented26, and perform their depth scan based on low coherence interferometry, which was originally used in optical fibers35. The wide bandwidth of the light signal used results in a low coherence length, and so only reflections from the sample which are near to the reference length show interference. In
7


order to scan the sample in depth, the reference length is changed to build up a model of the sample’s tomography.
Spectral domain OCT systems generally used a fixed reference path, and image the object depth by measuring the spectral response of the interference pattern36. The data from the spectrometer, encoded in frequency space, can be converted using a Fourier transform algorithm into an image of the scanned object37'38.
Another technique, swept source OCT, uses a narrow frequency source whose wavelength is quickly adjusted across the range used. The resulting power from the interference can then be used to create a profile of the sample. For this method only a single photodetector is required, the complication being in producing a range of frequencies for scanning the sample. This would normally be done by using an external cavity semiconductor laser with a fast tunable filter38'40'41.
There are a number of advantages and disadvantages for each method. SD OCT allows for high scan rates since there is no alteration in the reference arm length during scans, but there is more processing of the data following the scan. TD OCT does not require Fourier calculations on the data, but the scan rates are reduced because of the extra mechanical movements needed in the reference arm.
SD OCT has proved the most popular for medical imaging, mostly as a result of scanning speed, and with appropriate processing hardware42, the ability to produce real time images of the sample.
Most optical systems use a high numerical aperture (NA) to achieve a high resolution. OCT generally uses a low NA, since in applications like ophthalmology, a high NA cannot be achieved due to the distance between the outside of the eye and the
8


retina. The low NA increases the depth of field, and the axial resolution is governed by measuring the level of interference of the light source used rather than the NA.
In both types of OCT system, optical coherence microscopy (OCM) can be achieved where an image is acquired in a transverse plane to the light beam used43. This combines the usefulness of regular 2-D microscopy using confocal or some other method, with the larger imaging capability of OCT.
Although OCT has the capability of high depth penetration, this is usually limited by the light scattering effect of the tissue being imaged, and often results in a maximum usable depth of a few millimeters. To give a better image deeper into tissue or organs, OCT can be used with endoscope probes44 connected to fibre optics, which can then be used to probe directly into tissues. One of the features of OCT is it’s ability to produce depth related images of tissue structures without the need for staining or fluorophores, the method can be enhanced by the use of these techniques. The addition of contrast agents to samples can affect reflectivity and absorption, and so can be used to enhance specific features in the sample45.
OCT sensitivity is largely dependent on the detection of returning signal light over noise from the sample. This is mostly dependent on the tissue being imaged, but most OCT systems can produce sensitivities up to 100dB46.
The imaging speed varies based on the type of OCT system used. The limiting factor for TD OCT is the speed of changing the reference arm length. For SD OCT it is the acquisition and transfer rate from the spectrometer, and for SS OCT it is the sweep rate of the light source used.
9


For medical applications the image sample rate is one of the most important factors, particularly for imaging live samples, as some movement often cannot be avoided47. This is the reason why SD OCT has become the most popular method for medical imaging. Some systems are capable of scan rates in hundreds of kilohertz48, which can easily produce real time images.
OCT performance is also affected by the properties of the sample being imaged. Two main factors in this are the absorption and scattering of light at the frequency or frequencies being used, which affects the overall penetration depth of the signal48. For samples which scattering or absorption of the signal are significant will require either a signal reducing element in the reference arm49 to compensate for the reduced signal, or this may be achieved in the signal processing system50.
Biological tissues tend to scatter light, and so this becomes the main factor in imaging the sample51. This is a result of biological tissue usually being made of a large number of different components, all with their own light absorption and scattering properties.
For biological samples, absorption tends to be the least for wavelengths between 600 nm to 1300 nm, as this is the region where absorption of water and other common tissue constituents like hemoglobin and melanin are low52. The optimum wavelength to use largely depends on the constituent of the specific tissue being imaged. Longer wavelengths allow for deeper penetration into tissues, so for samples where water content is not significant, longer wavelengths can be used53. For applications where water is significant, like ophthalmic imaging, 800 nm is the most common frequency used54.
10


Wavelength (on)
Photodynamic Therapy: Current Status and Future Directions - Scientific Figure on ResearchGate. Available from: https://www.researchgate.net/262265731_fig2_Figure-6-The-optical-phototherapeutic-window-where-the-absorption-and-scattering-of [accessed 26 Nov, 2017]
FIGURE 2 - Absorbance of Hemoglobin and Water
Figure 2 above shows the relative absorbance of hemoglobin and water against wavelength.
For highly scattering samples, speckle can also be an issue in imaging. This is a phenomena which manifests itself as as fine graininess in the image produced. It is generally caused by a random distribution of scatterers in the tissue55, and can be used to determine features of the underlying tissue56'57. The downside is that if the amount of speckling is excessive, it can mask other features in the image. A number of methods are available to reduce speckle, some of which depend on varying the light signal features55, and others using data processing techniques to remove the unwanted noise58'62. Despite speckle being generally viewed as a problem, it can also be used in a positive manner. One example is in the measurement of vascular changes which can
11


occur in response to accidental trauma, but also in response to events such as the implantation of electrodes into the motor cortex for artificial limb control63, or in the case of electrode implantation for Parkinson’s disease64. Monitoring the vascular changes around the electrodes can give indications of early failure of the implants. In speckle variance analysis this is done by analyzing the amplitude changes in the OCT signal, which cause changes in the speckle pattern. These changes can then be used to visualize the changes occurring in the vasculature. This is similar to the amplitude decorrelation method, which also uses the OCT signal amplitude changes to detect vasculature changes.
Similar to imaging vasculature, OCT can also be useful in examining neural pathways65'66 as it is capable of identifying nerve fibers from surrounding tissue by differences in their scattering properties.
APPLICATIONS OF OCT
The original OCT systems were developed for use in ophthalmology, and have become the standard tool for directly examining ocular structures67'68'69. OCT can provide a 3-D image of a living retina comparable with a light microscopy image from a histological section70. OCT is a non-invasive technique, which requires no direct contact with the eye or labeling. It has become the standard diagnostic tool for a number of diseases, including macular degeneration71, glaucoma72 and diabetic retinopathy73 74.
12


Angiovueâ„¢ Imaging System. Optovue Inc.
-IIM
FIGURE 3 - Sample Commercial OCT System
In cardiology OCT has been used to identify vulnerable plaques76, and to aid in the diagnosis and characterization of atherosclerotic plaques77 that were not able to be accurately imaged with conventional techniques, such as ultrasound. OCT can be used to measure the thickness of plaque caps79, the presence of lipids below the cap75, and macrophage infiltration of the cap78. This has made OCT a valuable tool in cardiovascular diagnosis. Also in the cardiovascular area, OCT has also been applied to monitoring blood flow in capillary vessels80. This has been used to monitor the blood flow in the cortical capillary network of mice. Other applications have been found in studying the hemodynamic changes in mammalian embryos. This involved the use of a Doppler swept source OCT system for monitoring both changes in the blood vessels
13


and blood flow in mice embryos81. This could lead to advances in early diagnosis and treatment of congenital cardiovascular defects.
OCT is also being used for follow up monitoring of coronary stents82. The catheter type OCT systems can be used to examine the stent in vivo and make sure that it is correctly anchored, and there is no other issues with the placement. Catheter based OCT systems have also been developed for imaging other hollow organs with frame rates high enough to allow real time imaging83. Solid organs can also be imaged3, the main difference being the wavelength of light used for imaging, as different tissue types have different scattering and absorption properties.
Considering the current applications of OCT for imaging in vivo, extending this into brain imaging seems a reasonable step to take. Towards this end, work has been done to combine OCT with other forms of imaging currently used in neural monitoring. Optical intrinsic signal imaging (OISI) has been combined with OCT to exam activity in the rat somatosensory cortex84. This was done through a thinned skull rather than removing a skull section completely as would be required with optical imaging. At present, the current methods of identifying the affected areas of the brain in a disease like Parkinson’s are only accurate to within several millimeters at best, being able to produce 3-D images with micrometer resolution should be a significant advantage to both the patients and surgeons.
As has been previously mentioned, SD OCT has become the primary method for use in medical imaging. This is because of its higher speed imaging, and the reduced number of moving parts within the system. For these reasons, SD OCT was chosen as the method for this project.
14


CHAPTER II
SYSTEM DESIGN
HARDWARE
The hardware design of the system was based on a Michelson interferometer (Figure 4). The system was constructed using standard components supplied by ThorLabs Inc.
FIGURE 4 - Schematic of Michelson Interferometer
In Figure 4, U, Ls> Lr, and Ld are the optical path lengths for the LED output to the beam splitter, the sample arm to the beam splitter, the reference arm to the beam splitter, and the detector to the beam splitter respectively.
15


Eo is the electric field at the LED output. The electric field at the detector can then
be expressed as:
1.0
E=-E e~jk]'M+2LR+LD^ 4. -E e~jkfM+2Ls+liD^ 2 ® 1 2 ®
This can be rearranged and given in intensity: 1.1
where c is the speed of light, and n is the refractive index. Simplifying gives the following equation:
1.2
X
l+coB(2feA?)
2
Where lo is the intensity of the source diode, and lopt is the intensity at the detector.
Az is the difference in the path lengths of the reference and sample arms.
The following formula gives the axial resolution for the system. This is a relation of the central frequency of the light source being used (lambda 0) to the full bandwidth half maximum power (Alambda) of the source.
1-3 &= —
n AA
16


This formula gives the lateral resolution for the system, with f0bj being the focal length of the objective lens, and d representing the beam diameter incident on the objective lens.
1.4 Ax =
n d
Working from the principle of interferometry, the following drawing shows the initial design of the system. This was an open arrangement with the mirrors and lenses in adjustable holders. (Figure 5)
17


Reference Mirror
directed to the focusing lens for the spectrometer.
FIGURE 5 - Initial OCT System Layout
This design was used to determine the best arrangement for the system, and for performing alignment tests to produce interference between the two arms of the system The stage control and data analysis software was written based on this configuration.
After initial testing, it was found that the system was too unstable in it’s present form, and tended to move out of alignment too easily. To attempt to correct for this, the following changes were made to the design (Figure 6)
18


Splitter
Sampling
Mirror
and
Staae
Reference Mirror and Lens in Tube Housing
The LED output is collimated and the split 50:50 by the beam splitter and directed to the reference mirror and lens, and the sample. The reflected light is recombined in the splitter and
directed through the focusing lens
FIGURE 6 - Current OCT System Layout
to the spectrometer.
This new configuration proved to be considerably more stable, and could be more easily realigned if needed. Another advantage of the tube system was that there was much less interference in the system, since the light paths were almost entirely enclosed.
19


SOFTWARE
The software for the system was divided into two major sections, stage control and data analysis. The stage control section also performed data capture functions.
STAGE CONTROL SOFTWARE
This was written in LabView, and performed the main control function for the scanning stage. This involved driving the x and y-axis stepper motors, and controlling the data acquisition from the spectrometer. In addition, the LabView software also provided on screen displays of the current data being acquired, and the Fourier transform of the data, to give the user a representation of the data being acquired from the sample under observation.
The software was written as a modular event driven package, which could be extended at any time to incorporate other controls or data acquisition requirements without disrupting the existing software.
Details on the current software configuration and operations are given in Appendix 1.
DATA ANALYSIS SOFTWARE
Data analysis and image production was done using Matlab software. For the initial stage of the project the analysis was performed offline, after the data capture was completed. This was done to facilitate the processing of large datasets without the use of hardware other than regular desktop computers.
20


The image processing software was divided into two sections, the pre-processing code and the image generation code. The division was done to help speed up final image generation, since the image generation code contains segments which can be modified to improve the image, whereas the pre-processing code does not need modifications for different images.
The software uses two data files, the primary file contains the raw data from the scan of the sample, and the other file contains the meta data describing the main dat file. The meta file data consists of the following data:
The base name for the file set,
Threshold scan values for a slide with a coverslip,
Threshold scan values for slide a slide without a coverslip,
Number of motor steps per datapoint,
Number of rows per data scan,
Number of data points per row,
Scan rate value,
Number of axes divisions (for plotting),
The Picomotor step size in nanometers,
The Spectrometer resolution in pixels.
This data is then used to construct the name of the primary data file to be loaded, how to process the raw data into the final image, and how to plot the image data.
21


CHAPTER III
RESULTS
The following images are a representative sample of what the system can produce. The first plot (Figure 8) shows an image of two microscope cover slips attached to form a stair like arrangement as shown in Figure 7.
Scan parameters:
300 points per line @ 30ji/m/point 16 lines, 9.8 Mbytes (approximately)
Front View Side View
Slides
FIGURE 7 - Stacked Cover Slip Arrangement
22


HeflW;
024
021 v
mm
FIGURE 8 - Stacked Cover Slips
The next image (Figure 10) shows the scan of a fingerprint placed on a microscope coverslip. A photograph of the fingerprint is shown for comparison (Figure 9). Obviously the regular photograph cannot show the depth of the fingerprint to any significance, but it does serve to highlight the lack of smoothness in the ridges and valleys of the print.
Scan parameters:
200 points per line @ 30 ji/m/point
62 lines
25.5 Mbytes (approximately)
23


FIGURE 9 - Fingerprint Photographs (standard size and enlarged)
Dwm» Nw
• 4* • Cl Q O
FIGURE 10 - Fingerprint
24


CHAPTER IV
DISCUSSION
OCT microscopy has expanded into many area of medical imaging since its inception in ophthalmology24. Spectral domain OCT has become the most popular for medical applications, and has been the method chosen for this project.
The images shown in the results section were chosen to demonstrate the capabilities of spectral domain OCT, and they highlighted a number of features of the technique.
The method has the capability to scan a sample to a very small resolution, and coupled with this is the generation of large quantities of data. This was the main reason that offline processing of image scans was chosen. The raw data contains intensity information and differences in the path lengths of the two arms. Since different material compositions can cause variations in the absorption and scattering of the source light, interpretation of the data gathered is significant to producing an accurate representation of the sample85'86. In addition, the type of sample being scanned is also significant in its interpretation. A microscope slide with a cover slip needs the software to identify the slide and coverslip positions and interpret the data between them. The intensity of the returned signal can be considerably reduced by the coverslip. For this reason the software needs to be aware of the type of sample being imaged, and adjust its parameters accordingly to provide an accurate image. To demonstrate this, the following three images show the data plots from the analyzer as it scans across a slide with two cover slips in a stepped arrangement as in Figure7.
25


FIGURE 11 - Fourier Data of Slide Only
FIGURE 12 - Fourier Data of Slide and Single Coverslip
FIGURE 13 - Fourier Data of Slide and Two Coverslips
26


The left screen shows the amplitude of the combined reference and sample signals at the spectrometer against wavelength. The right screen shows the Fourier transform of the amplitude/wavelength data converted to give a signal intensity and distance from the baseline, in this case a microscope slide.
Figure 11 shows the beam focused on the slide only, and so the reference and signal paths are of equal length. The Fourier plot shows a signal amplitude at the baseline only.
Figure 12 shows the signal and Fourier data when the beam has moved over the first coverslip. The Fourier plot distance is representative of the thickness of the coverslip at the point where the beam is focused. The left screen shows a representation of the interference patterns from the different reference and sample lengths.
Figure 13 gives a representation of the sample beam directed at both coverslips. As can be seen, the amplitude return from the first coverslip is considerably reduced, since the signal is now partly reflected from the surface of the second coverslip, and partly attenuated by passing through the coverslip.
For slides with cover slips, the intensity of the signal returns on the sample below it will be much less than for a sample on a slide without a cover slip. The analysis software needs to be a ware of this, and be able to process the data accordingly.
27


CHAPTER V
FUTURE WORK
The current state of the OCT system has indicated the potential for further development and improvements. The potential upgrades considered at this time fall into two main areas, hardware and software.
HARDWARE
The experience gained from working on the system was that alignment of the components was critical to correct operation of the device. The majority of this work could be automated with the addition of motor driven optical components, along with a micro controller system such as a Beagle Bone. This is a low cost commercially available device, and could be used as a main controller for the OCT. The Beagle Bone itself does not incorporate analog sensors, but this could be overcome by using PIC Micro-controllers interfaced to the Beagle Bone. These are also commercially available at very low costs, in the order of $1-10 per unit. This could then be programmed to take over most of the alignment tasks for the system. In addition, changing the light source to a wide bandwidth source centered around 1300nm would give better depth resolution.
The volume of data generated is large even for a small area scan, and so the Beagle Bone could also be used as a data processor, and is capable of driving high definition monitors directly, and so could be used as a primary display device.
28


SOFTWARE
The current mixture of LabView and Matlab software could be improved by moving to a language such as Python, which would be capable of doing both the data capture and image processing itself. The Beagle Bone is capable of running Python code, so this would mean that only one language would be need throughout the project.
29


REFERENCES
1. Lee A, Gilbert RM. Epidemiology of Parkinson Disease. Neurol Clin NA. 2017;34(4): 955-965. doi:10.1016/j.ncl.2016.06.012.
2. Vitek JL. Mechanisms of deep brain stimulation: excitation or inhibition. Mov Disord 2002;17:Suppl 3:S69-S72.
3. Ahsen 00, Tao YK, Potsaid BM, et al. Swept source optical coherence microscopy using a 1310 nm VCSEL light source. 2013;21(15):18021-18033. doi:10.1364/OE. 21.018021.
4. Mariampillai A, Standish BA, Munce NR, et al. Doppler optical cardiogram gated 2D color flow imaging at 1000 fps and 4D in vivo visualization of embryonic heart at 45 fps on a swept source OCT system. 2007;15(4):1627-1638.
5. Tanner CM, Aston DA. Epidemiology of Parkinson’s disease and akinetic syndromes. Curr Opin Neurol 2000;13:427-30.
6. Fahn S. Description of Parkinson’s disease as a clinical syndrome. Ann N Y Acad Sci 2003;991:1-14.
7. Lang AE, Lozano AM. Parkinson’s disease (first of two parts). N Engl J Med 1998;339:1044-53.
8. Martikainen KK, LuukkaalaTH, Marttila RJ. Parkinson’s disease and working capacity. Mov Disord 2006;21:2187-91.
9. Dodel RC, Eggert KM, Singer MS, Eichhorn TE, Pogarell O, Oertel WH. Costs of drug treatment in Parkinson’s disease. Mov Disord 1998;13:249-54.
10. Whetten-Goldstein K, Sloan F, Kulas E, Cutson T, Schenkman M. The burden of Parkinson’s disease on society, family, and the individual. J Am Geriatr Soc 1997; 45:844-9.
11. Willis AW, Schootman M, Evanoff BA, Perlmutter JS, Racette BA. Neurologist care in Parkinson disease: a utilization, outcomes, and survival study. Neurology 2011;77:851-7.
12. Alexander GE, DeLong MR, Strick PL. Parallel organization of functionally egregated circuits linking basal ganglia and cortex. Annu Rev Neurosci 1986;9:357-81.
13. Delong MR, Georgopoulos AP, Crutcher MD, Mitchell SJ, Richardson RT, Alexander GE. Functional organization of the basal ganglia: contributions of single-cell recording studies. Ciba Found Symp 1984;107:64-82.
14. Wagle Shukla A, Okun MS. Personalized medicine in deep brain stimulation through utilization of neural oscillations. Neurology 2012;78:1900-1.
15. Fahn S. The spectrum of levodopa-induced dyskinesias. Ann Neurol 2000;47: Suppl 1 :S2-S9.
16. Okun MS, Fernandez HH, RodriguezRL, Foote KD. Identifying candidates for deep brain stimulation in Parkinson’s disease: the role of the primary care physician.
Geriatrics 2007;62:18-24.
17. McIntyre CC, Hahn PJ. Network perspectives on the mechanisms of deep brain stimulation. Neurobiol Dis 2010;38: 329-37.
18. Wichmann T, DeLong MR, Guridi J, Obeso JA. Milestones in research on the pathophysiology of Parkinson’s disease. Mov Disord 2011 ;26:1032-41.
30


19. Tawfik VL, Chang SY, Hitti FL, et al. Deep brain stimulation results in local glutamate and adenosine release: investigation into the role of astrocytes. Neurosurgery 2010;67:367-75.
20. Lee KH, Hitti FL, Chang SY, et al. High frequency stimulation abolishes thalamic network oscillations: an electrophysiological and computational analysis. J Neural Eng 2011 ;8(4):046001.
21. Lee KH, Chang SY, Roberts DW, Kim U. Neurotransmitter release from high-frequency stimulation of the subthalamic nucleus. J Neurosurg 2004;101:511-7.
22. Vedam-Mai V, van Battum EY, Kamphuis W, et al. Deep brain stimulation and the role of astrocytes. Mol Psychiatry 2012;17:124-31.
23. Castrioto A, Lozano AM, Poon YY, Lang AE, Fallis M, Moro E. Ten-year outcome of subthalamic stimulation in Parkinson disease: a blinded evaluation. Arch Neurol
011 ;68:1550-6.
24. Tomaszewski KJ, Holloway RG. Deep brain stimulation in the treatment of Parkinson’s disease: a cost-effectiveness analysis. Neurology 2001 ;57:663-71.
25. Okun MS, Tagliati M, Pourfar M, et al. Management of referred deep brain stimulation failures: a retrospective analysis from 2 movement disorders centers. Arch Neurol 2005;62:1250-5.
26. Huang D, Swanson EA, Lin CP, Schuman JS, Stinson WG, Chang W, Hee MR, Flotte, Gregory K, Puliafito CA, Fujimoto JG (1991) Optical coherence tomography. Science 254 (5035): 1178-1181
27. Smolka G (2010) Optical coherence tomography 2010: technology, applications, and markets (Market Report OM-51). Strategies Unlimited (PennWell Corporation), MountainView
28. Swanson E (2017) LightLab Imaging announces FDA clearance of C7-XRâ„¢ coronary OCT products in the United States. Opt Coher Tomogr News 5/5/2010. http:// www.octnews.ora/articles/2008632/liahtlab-imaaing-announces-fda-clearance-of-c7-xrt/. accessed 10 Apr 2017
29. Volcano Corp. (2017) Volcano Corporation announces receipt of CE mark for its OCT imaging system and catheter. Volcano Corp., San Diego ReleaselD=434187, accessed 10 Apr 2017
30. Hartl I, Li XD, Chudoba C, Ghanta RK, Ko TH, Fujimoto JG, Ranka JK, Windeler RS
(2001) Ultrahigh-resolution optical coherence tomography using continuum generation in an air-silica microstructure optical fiber. Opt Lett 26 (9):608-610
31. Povazay B, Bizheva K, UnterhuberA, Hermann B, Sattmann H, FercherAF, Drexler W, Apolonski A, Wadsworth WJ, Knight JC, Russell PSJ, Vetterlein M, Scherzer E
(2002) Submicrometer axial resolution optical coherence tomography. Opt Lett 27(20): 1800-1802
32. Drexler W (2004) Ultrahigh-resolution optical coherence tomography. J Biomed Opt 9(1):47-748.
33. Inoue S (2006) Chap 1. In: Handbook of biological confocal microscopy, 3rd edn. Springer, Berlin, ISBN: 978-0-387-25921-5
34. S. J. Matcher, M. Cope, D. T. Delpy, (1997) In vivo measurements of the wavelength dependence of tissue-scattering coefficients between 760 and 900 nm measured with time-resolved spectroscopy. Appl. Opt. 36, 386-396
31


35. Takada K, Yokohama I, Chida K, Noda J (1987) New measurement system for fault location in optical waveguide devices based on an interferometric technique. Appl Opt 26(9): 1603-1606
36. Fercher AF, Hitzenberger CK, Kamp G, El-Zaiat SY (1995) Measurement of intraocular distances by backscattering spectral interferometry. Opt Commun 117(1-2): 43—48
37. Bail MA, Hausler G, Herrmann JM, Lindner MW, Ringler R (1996) Optical coherence tomography with the “spectral radar:” fast optical analysis in volume scatterers by short-coherence interferometry. Proc SPIE 2925:298-303
38. Hausler G, Lindner MW (1998) Coherence radar and spectral radar—new tools for dermatological diagnosis. J Biomed Opt 3(1 ):21—31
39. Yun S-H, Boudoux C, Tearney GJ, Bouma BE (2003) Highspeed wavelength-swept semiconductor laser with a polygonscanner-based wavelength filter. Opt Lett 28(20): 1981-1983
40. Choma MA, Hsu K, Izatt JA (2005) Swept source optical coherence tomography using an all-fiber 1300-nm ring laser source. J Biomed Opt 10(4):044009
41. Huber R, Wojtkowski M, Taira K, Fujimoto JG, Hsu K (2005) Amplified, frequency swept lasers for frequency domain reflectometry and OCT imaging: design and scaling principles. Opt Express 13(9):3513-3528
42. Huber R, Wojtkowski M, Fujimoto JG (2006) Fourier domain mode locking (FDML): a new laser operating regime and applications for optical coherence tomography. Opt Express 14 (8):3225-3237
43. Izatt JA, Hee MR, Owen GM, Swanson EA, Fujimoto JG (1994) Optical coherence microscopy in scattering media. Opt Lett 19 (8):590-592
44. Tearney GJ, Brezinski ME, Bouma BE, Boppart SA, Pitris C, Southern JF, Fujimoto JG (1997) In vivo endoscopic optical biopsy with optical coherence tomography.
Science 276 (5321):2037-2039
45. Akkin T, Landowne D, Sivaprakasam A. Detection of neural action potentials using optical coherence tomography : intensity and phase measurements with and without dyes. 2010;2(August):1-10. doi:10.3389/fnene.2010.00022.
46. Marschall S, Sander B, Mogensen M, Jorgensen TM, Andersen PE. Optical coherence tomography-current technology and applications in clinical and biomedical research. Anal Bioanal Chem. 2011 ;400(9):2699-2720. doi:10.1007/s00216-011-5008-1.
47. Yun S-H, Tearney GJ, de Boer JF, Bouma EB (2004) Motion artifacts in optical coherence tomography with frequency-domain ranging. Opt Express 12(13):2977-2998
48. Schmitt JM, Knuttel A, Yadlowsky M, Eckhaus MA (1994) Optical-coherence tomography of a dense tissue: statistics of attenuation and backscattering. Phys Med Biol 39(10): 1705-1720
49. Hitzenberger CK, Baumgartner A, Drexler W, Fercher AF (1999) Dispersion effects in partial coherence interferometry: implications for intraocular ranging. J Biomed Opt 4(1); 144—151
50. de Boer JF, Saxer CE, Nelson JS (2001) Stable carrier generation and phase-resolved digital data processing in optical coherence tomography. Appl Opt 40(31): 5787-5790
32


51. Brezinski ME, Tearney GJ, Bouma BE, Izatt JA, Hee MR, Swanson EA, Southern JF, Fujimoto JG (1996) Optical coherence tomography for optical biopsy: properties and demonstration of vascular pathology. Circulation 93(6):1206-1213
52. Parrish JA (1981) New concepts in therapeutic photomedicine; photochemistry, optical targeting and the therapeutic window. J Invest Dermatol 77(1):45-50
53. Sharma U, Chang EW, Yun SH (2008) Long-wavelength optical coherence tomography at 1.7 pm for enhanced imaging depth. Opt Express 16(24):19712-19723
54. Povazay B, Bizheva K, Hermann B, Unterhuber A, Sattmann H, FercherAF, Drexler W, Schubert C, Ahnelt PK, Mei M, Holzwarth R, Wadsworth WJ, Knight JC, Russel PSJ
(2003) Enhanced visualization of choroidal vessels using ultrahigh resolution ophthalmic OCT at 1050 nm. Opt Express 11 (17):1980-1986
55. Schmitt JM, Xiang SH, Yung KM (1999) Speckle in optical coherence tomography. J Biomed Opt 4(1):95-105
56. Gossage KW, Tkaczyk TS, Rodriguez JJ, Barton JK (2003) Texture analysis of optical coherence tomography images: feasibility for tissue classification. J Biomed Opt 8(3):570-575
57. Hillman TR, Adie SG, Seemann V, Armstrong JJ, Jacques SL, Sampson DD (2006) Correlation of static speckle with sample properties in optical coherence tomography. Opt Lett 31(2): 190-192
58. Kulkarni MD, Thomas CW, Izatt JA (1997) Image enhancement in optical coherence tomography using deconvolution. Electron Lett 33(16):1365-1367
59. Schmitt JM (1998) Restoration of optical coherence images of living tissue using the CLEAN algorithm. J Biomed Opt 3 (1):66-75
60. Xiang SH, Zhou L, Schmitt JM (1998) Speckle noise reduction for optical coherence tomography. Proc SPIE 3196(1):79-88
61. Rogowska J, Brezinski ME (2000) Evaluation of the adaptive speckle suppression filter for coronary optical coherence tomography imaging. IEEE T Med Imaging 19(12): 1261-1266
62. Wang RK (2005) Reduction of speckle noise for optical coherence tomography by the use of nonlinear anisotropic diffusion. Proc SPIE 5690(1 ):380-385
63. Hammer DX, Lozzi A, Abliz E, et al. Longitudinal vascular dynamics following cranial window and electrode implantation measured with speckle variance optical coherence angiography. 2014;5(8):2823-2836. doi:10.1364/BOE.5.002823.
64. Hammer DX, Welle CG, Hammer DX, Lozzi A, Boretsky A, Welle CG. Acute insertion effects of penetrating cortical microelectrodes imaged with quantitative optical coherence angiography microelectrodes imaged with quantitative. 2016;3(2). doi: 10.1117/1. NPh.3.2.025002.
65. Wang H, Black AJ, Zhu J, et al. Neuroimage Reconstructing micrometer-scale fi ber pathways in the brain : Multi-contrast optical coherence tomography based tractography. Neuroimage. 2011 ;58(4):984-992. doi:10.1016/j.neuroimage.2011.07.005.
66. Wang H, Zhu J, Reuter M, et al. Neuroimage Cross-validation of serial optical coherence scanning and diffusion tensor imaging : A study on neural fiber maps in human medulla oblongata. Neuroimage. 2014;100:395-404. doi:10.1016/j.neuroimage. 2014.06.032.
67. FercherAF, Hitzenberger CK, Drexler W, Kamp G, Sattmann H
(1993) In vivo optical coherence tomography. Am J Ophthalmol 116(1 ):113—114
33


68. Swanson EA, Izatt JA, Hee MR, Huang D, Lin CP, Schuman JS, Puliafito CA, Fujimoto JG (1993) In vivo retinal imaging by optical coherence tomography. Opt Lett 18(21 ):1864-1866
69. Izatt JA, Hee MR, Huang D, Fujimoto JG, Swanson EA, Lin CP, Shuman JS,
Puliafito CA (1993) Ophthalmic diagnostics using optical coherence tomography. Proc SPIE 1877:136-144
70. Tanno N, Kishi S (1999) Optical coherence tomographic imaging and clinical diagnosis. Med Imaging Technol 17(1):3—10
71. Kaiser PK, Blodi BA, Shapiro H, Acharya NR (2007) Angiographic and optical coherence tomographic results of the marina study of ranibizumab in neovascular age-related macular degeneration. Ophthalmology 114(10): 1868-1875
72. Tan O, Li G, Ake Tzu-Hui L, Varma R, Huang D (2008) Mapping of macular substructures with optical coherence tomography for glaucoma diagnosis. Ophthalmology 115(6):949-956
73. Soliman W, Sander B, Soliman KAE-N, Yehya S, Rahamn MSA, Larsen M (2008) The predictive value of optical coherence tomography after grid laser photocoagulation for diffuse diabetic macular oedema. Acta Ophthalmol 86(3):284-291
74. Gaucher D, Sebah C, ErginayA, Haouchine B, Tadayoni R, Gaudric A, Massin P (2008) Optical coherence tomography features during the evolution of serous retinal detachment in 2716 S. Marschall et al. patients with diabetic macular edema. Am J Ophthalmol 145(2):289-296
75. van der Meer FJ, Faber DJ, Sassoon DMB, Aalders MC, Pasterkamp G, van Leeuwen TG (2005) Localized measurement of optical attenuation coefficients of atherosclerotic plaque constituents by quantitative optical coherence tomography. IEEE T Med Imaging 24(10): 1369-1376
76. Jang l-K, Bouma BE, Kang D-H, Park S-J, Park S-W, Seung K-B, Choi K-B, Shishkov M, Schlendorf K, Pomerantsev E, Houser SL, Aretz HT, Tearney GJ (2002) Visualization of coronary atherosclerotic plaques in patients using optical coherence tomography: comparison with intravascular ultrasound. J Am Coll Cardiol 39(4):604-609
77. Low AF, Tearney GJ, Bouma BE, Jang l-K (2006) Technology insight: optical coherence tomography: current status and future development. Nat Clin Pract Card 3:154-162
78. Tearney GJ, Yabushita H, Houser SL, Thomas Aretz H, Jang l-K, Schlendorf KH, Kauffman OR, Shishkov M, Halpern EF, Bouma BE (2003) Quantification of macrophage content in atherosclerotic plaques by optical coherence tomography. Circulation 107(1) :113—119
79. Kume T, Akasaka T, Kawamoto T, Okura H, Watanabe N, Toyota E, Neishi Y, Sukmawan R, Sadahira Y, Yoshida K (2006) Measurement of the thickness of the fibrous cap by optical coherence tomography. Am Heart J 152(4):755.e1-755.e4
80. Lozzi A, Agrawal A, Boretsky A, Welle CG, Hammer DX. Image quality metrics for optical coherence angiography. 2015;6(7):2435-2447. doi:10.1364/BOE.6.002435.
81. Larina I V, Sudheendran N, Ghosn M, et al. Live imaging of blood flow in mammalian embryos using Doppler swept-source optical coherence tomography.
2017;13(December 2008):10-12. doi:10.1117/1.3046716.
34


82. Gonzalo N, Serruys PW, OkamuraT, Shen ZJ, Onuma Y, Garcia-Garcia HM, Sarno G, Schultz C, van Geuns RJ, Ligthart J, Regar E (2009) Optical coherence tomography assessment of the acute effects of stent implantation on the vessel wall: a systematic quantitative approach. Heart 95(23):1913—1919
83. Tsai T, Potsaid B, Tao YK, et al. Ultrahigh speed endoscopic optical coherence tomography using micromotor imaging catheter and VCSEL technology. 2013;4(7): 898-905. doi: 10.1364/BOE.4.001119.
84. Chen Y, Aguirre AD, Ruvinskaya L, Devor A, Boas DA, Fujimoto JG. Optical coherence tomography ( OCT ) reveals depth-resolved dynamics during functional brain activation. 2009;178:162-173. doi:10.1016/j.jneumeth.2008.11.026.
85. Drexler, W. & Fujimoto, J. G. Optical coherence tomography : technology and applications. (Springer, 2008).
86. Al-Mujaini, A., Wali, U. K., & Azeem, S. (2013). Optical Coherence Tomography: Clinical Applications in Medical Practice. Oman Medical Journal, 28(2), 86-91. http:// doi.ora/10.5001/omi.2013.24
35


APPENDIX
Main Screen
Overall Structure
36


Scan Row/Column
Scan Slide
37


Move Stage
Move to Zero
38


Overall Hierarchy
39


Matlab Software
Data Pre-processing:
% preprocDataFile takes the raw data from the image scan, and pre-processes % the data to a form used by the dataAnalyzer program.
%
clc;
clear all; close all;
A = importdata('sampleMeta.txt');
fBase = cell2mat(A.textdata(1)); slideThreshold =A.data(1); printThreshold =A.data(2); steps =A.data(3); rows =A.data(4); points =A.data(5); scanR =A.data(6); axesPts = A.data(7); picoNm = A.data(8); specPts = A.data(9);
% File basename % Threshold for slide w/coverslip % Threshold for slide wo/coverslip % Motor steps per datapoint % Rows in scan % Datapoints per row % Scan rate
% Axes divisions (for printing) % Picomotor step size (nm)
% Spectrometer resolution
fNameRoot = sprintf('%s_%d_%ds_%dp_%dr', fBase, rows/2, steps, points, scanR);
samples = specPts; % spectrometer values per point
fName = cat(2, fNameRoot, '.txt'); fid = fopen(fName, 'r'); fp1 = fscanf(fid, '%f');
j = 1;
maxLen = Iength(fp1)/samples; fp2 = zeros(maxl_en, samples); fi = zeros(maxl_en, samples);
for i = 1 :maxl_en
fp2(i, :) = fp1(j:j+(samples-1)); j = j + samples; fi(i, :) = fft(fp2(i, :)); end
rowLen = points;
40


rowLenStart = 1;
rowLenEnd = rowLenStart + rowLen - 1;
% Divide points into rows, and create a cell vector of rows % Every other (2, 4, 6, ...) row gets flipped, since the scan % tracks along a row, moves up, and tracks back to the X origin % (sort of: if the motors were ok)
%
fp3 = cell(1, rows); for i = 1 :rows
fp3{i} = fp2(rowLenStart:rowLenEnd, :); if(rem(i, 2) == 0) fp3{i} = flipud(fp3{i}); end;
rowLenStart = rowLenStart + rowLen; rowLenEnd = rowLenEnd + rowLen;
end;
fName = cat(2, fNameRoot, '.mat'); save(fl\lame, 'fp3');
41


Image Analysis:
% dataAnalyzer takes the preprocessed data file and produces a 3-D plot of % the scanned image.
%
clc;
clear all; close all;
A = importdata('sampleMeta.txt');
fBase = cell2mat(A.textdata(1)); slideThreshold =A.data(1); printThreshold =A.data(2); steps =A.data(3); rows =A.data(4); points =A.data(5); scanR =A.data(6); axesPts = A.data(7); picoNm = A.data(8); specPts = A.data(9);
% File basename % Threshold for slide w/coverslip % Threshold for slide wo/coverslip % Motor steps per datapoint % Rows in scan % Datapoints per row % Scan rate
% Axes divisions (for printing) % Picomotor step size (nm)
% Spectrometer resolution
fName = sprintf('%s_%d_%ds_%dp_%dr.mat', fBase, rows /2, steps, points, scanR);
load(fl\lame, 'fp3');
ft_2 = zeros(rows, points);
for j = 1 :rows for i = 1 :points
ft = sqrt(real(fft(fp3{1,j}(i, :)))A2); if(contains(fBase, 'slide', 'IgnoreCase', true)) ft_1 =ft(100:300); ft_1(ft_1 < slideThreshold) = 0; else
ft_1 = ft(50:specPts/2); ft_1(ft_1 < printThreshold) = 0; end
ft_1 (ft_1 > 0) = 1;
ind = find(ft_1, 1); % find first index where change starts
if(isempty(ind)) ind = 0; end
ft_2(j, i) = ind; % make up vector of index points end end
42


ft_2s = zeros(rows, points); for i = 1 :rows
ft_2s(i,:) = smooth(ft_2(i, :), 10); end
surf(ft_2s); shading interp; axis tight;
xlabelCWidth: mm'); zlabel('Height: mm'); ylabel('Length: mm');
mm = (picoNm * points * steps) / 1e6; xt = linspace(0, points, 7); xticks(xt);
xlabs = linspace(0, mm, axesPts); xlabelsc = cell(axesPts+1, 1); for i=1 :length(xlabs)
xlabelsc{i}=sprintf('%.2f', xlabs(i)); end
xticklabels(xlabelsc);
yp = rows - mod(rows, 10); mm = (picoNm * yp * steps) / 1e6; ylabs = linspace(0, mm, axesPts); ylabelsc = cell(axesPts+1, 1); for i = 1 length(ylabs)
ylabelsc{i} = sprintf('%.2f', ylabs(i)); end
yticklabels(ylabelsc);
[M, I] = max(ft_2s); zp = round(max(M)/10) * 10; mm = (picoNm * zp * steps) / 1e6; zlabs = linspace(0, mm, axesPts); zlabelsc = cell(axesPts+1, 1); for i = 1 length(zlabs)
zlabelsc{i} = sprintf('%.2f', zlabs(i)); end
zticklabels(zlabelsc);
43


Full Text

PAGE 1

OPTICAL COHERENCE TOMOGRAPHY FOR BRAIN TISSUE ANALYSIS by Andrew James Challinor B.Sc.(Hons) Teesside University, 1983 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulÞllment of the requirements for the degree of Master of Science Bioengineering Program 2017

PAGE 2

© 2017 ANDREW JAMES CHALLINOR ALL RIGHTS RESERVED! " ii

PAGE 3

This thesis for the Master of Science degree by Andrew James Challinor has been approved for the Bioengineering Program by Emily Gibson, Chair Diego Restrepo Cristin Welle Date: July 29 2017 " iii

PAGE 4

Challinor, Andrew James (MS Bioengineering) Optical Coherence Tomography for Brain Tissue Analysis Thesis directed by Associate Professor Emily Gibson ABSTRACT Ever since the emergence of Optical Coherence Tomography (OCT) as a tool for ophthalmic imaging, the additional applications of the technology for medical imaging have been expanding rapidly. The two basic techniques, time domain (TD) and spectral domain (SD) have both been applied in different situations. SD OCT has become the more commonly used method for medical applications, mainly because of its ability to produce high speed images, including real time, with less moving parts than TD OCT. In the Þeld of brain imaging, TD OCT holds considerable promise in the ability to probe deep within a live brain without signiÞcant damage, and to produce high quality images of brain tissue, which can be of considerable beneÞt to neurosurgeons. OCT has the ability to not only produce high resolution (~1 micron) images, but also to provide topographical data of the area under observation, allowing 3-D images of the tissue to be constructed. The method images the tissue directly, and so does not require labelling with radioactive materials or other contrast agents, making it much safer for the patient also. This thesis demonstrates the fundamental principles required for the construction of an SD OCT system which is capable of producing 3-D scans of brain tissue slides with sufÞcient resolution to identify cellular damage and inclusions such as Lewy bodies. The form and content of this abstract are approved. I recommend its publication. Approved: Emily Gibson " iv

PAGE 5

TABLE OF CONTENTS I. INTRODUCTION1 ..................................................................................................... BACKGROUND1 ...................................................................................................... OPTICAL COHERENCE TOMOGRAPHY OVERVIEW6 ......................................... APPLICATIONS OF OCT12 ..................................................................................... II. SYSTEM DESIGN15 ................................................................................................ HARDWARE15 ......................................................................................................... SOFTWARE20 .......................................................................................................... STAGE CONTROL SOFTWARE20 .......................................................................... DATA ANALYSIS SOFTWARE20 .............................................................................. III. RESULTS22 ............................................................................................................ IV. DISCUSSION25 ...................................................................................................... V. FUTURE WORK28 ................................................................................................... HARDWARE28 ......................................................................................................... SOFTWARE29 .......................................................................................................... REFERENCES30 ......................................................................................................... APPENDIX36 ............................................................................................................... " v

PAGE 6

LIST OF FIGURES FIGURE 1 Pathways in the basal ganglia1 ................................................................ FIGURE 2 Absorbance of Hemoglobin and Water11 ................................................. FIGURE 3 Sample Commercial OCT System13 ........................................................ FIGURE 4 Schematic of Michelson Interferometer15 ................................................ FIGURE 5 Initial OCT System Layout18 .................................................................... FIGURE 6 Current OCT System Layout19 ................................................................ FIGURE 7 Stacked Cover Slip Arrangement22 ......................................................... FIGURE 8 Stacked Cover Slips23 ............................................................................. FIGURE 9 Fingerprint Photographs (standard size and enlarged)24 ........................ FIGURE 10 Fingerprint24 .......................................................................................... FIGURE 11 Fourier Data of Slide Only26 .................................................................. FIGURE 12 Fourier Data of Slide and Single Coverslip26 ........................................ FIGURE 13 Fourier Data of Slide and Two Coverslips26 .......................................... " vi

PAGE 7

CHAPTER I INTRODUCTION BACKGROUND Parkinson's disease affects the substantia nigra section of the brain, which is located in the midbrain region. It forms one of the components of the basal ganglia, and is the main area for dopamine production. Parkinson's causes the dopanergic neurons in this area to atrophy, which disrupts the signals to other parts of the brain which control skeletal muscular actions 1 . This results in the typical tremors that are seen in Parkinson's patients. A block diagram of the basal ganglia components and pathways is shown in Figure 1. The effects of the damage to the substantial nigra are indicated. FIGURE 1 Pathways in the basal ganglia " 1

PAGE 8

For the most part, current treatment for Parkinson's is palliative care, as there is no current cure. Medications can help the symptoms of the disease, and these include drugs such as Levodopa and Carbidopa, which is an amino acid which increases dopamine concentrations in the brain. Other drugs such as apomorphine act as dopamine agonists. This drug can relieve Parkinson' s symptoms quickly, but is also short lived. A more effective and longer term treatment for Parkinson's is deep brain stimulation (DBS), which involves the implantation of electrodes into the sub thalamic nucleus which are connected to a subcutaneous device which provides electrical pulses to these areas 2 . The strength and frequency of the pulses are controlled. The controller, which can be set to deliver a wide range of stimulus signals. Regrettably, this is still not a cure for Parkinson' s, and this method is not always suitable for the patient, but it is the most effective treatment currently available. One of the main drawbacks of the DBS method lies in the method of implanting the electrodes in the patient. At the present time, there is no method of producing real time images of deep brain structures, although OCT systems have been developed to produce high frame rates using Þxed position scanning (3, 4) . Currently an electrical method is used to aide in electrode placement. This technique consists of the surgeon inserting an electrode that reads out the local Þeld potential activity, and this is converted into a visible and audible signal for the neurosurgeon. This is done in combination with MRI and CT scans to help guide the placement of the probe to the correct area, using stereotactic methods . Once the probe is in the correct area of the " 2

PAGE 9

brain, the surgeon then listens for changes in the audible signal to indicate areas of the brain which have been damaged, and use this to guide the placement of the electrodes. There is no direct visualization of the affected area during the procedure, and so the electrode placement is somewhat inaccurate. In order to achieve the correct placement, several attempts may be needed, so prolonging the surgery. At times, a repeat surgery is needed to reposition the electrodes to a better location. Parkinson's typically develop in older adults, usually between the ages of 55 and 65 5 , and around 0.3% of the population may be affected to some degree. The incidence among men is generally higher than women 9 . The disease's main symptoms are muscle tremors, stiffness, slow movement and balance problems 7 . Additionally, softening of the voice and small handwriting are common, as are mood swings including depression and anxiety. The risk of death for sufferers of any level is almost double that of non-sufferers. Because of the range of symptoms, of which most patients will have a combination of, the majority of Parkinson's patients will not be able to maintain employment of any kind 8 . The cost of standard drug treatment is between $1,000 and $6,000 per year, with additional healthcare costs of between $2,000 to $20,000 per year 9, 10 . Treatment by a neurologist has been statistically shown to decrease the rate of deaths, and of the removal of patients from their homes into nursing facilities 11 . The disease itself is a neurodegenerative syndrome, which involves several motor and non-motor neural bundles in the basal ganglia 7, 12, 13 The disease typically disrupts the brain signaling system 13, 14 , causing erratic Þring patterns in the neurons, which in turn cause the tremors and rigidity commonly seen in patients. " 3

PAGE 10

At the beginning of treatment, patients generally have a good response to one of the drug therapies available, however after around Þve years or so, a range of drug related complications tend to start occurring. Many drug therapies see a gradual loss of effectiveness over time, however in Parkinson's disease the patients frequently develop a problem where the drugs become largely ineffective. The loss of effectiveness of drug therapy can occur suddenly and unpredictably, which is an obvious concern for the patient and their safety, as the symptoms can include suddenly increased tremors, rigidity, loss of balance, speech, and swallowing ability 6, 15 . Many patients become increasingly resistant to drugs such as the often used carbidopa-levidopa combination, which may leave them with few alternatives to alleviate their symptoms. Patients that are currently responsive to levodopa and other factors are considered the best potential candidates for deep brain stimulation (DBS) treatment 16 . The area of the brain that is damaged by the disease is the substantial nigra which is responsible for dopamine production. In this region there is generally cellular damage, and inclusions called Lewy bodies appear in the cells. DBS stimulation of this area is not effective, since the cell damage has already occurred. The electrodes are typically implanted into the sub thalamic nucleus of the brain. The electrodes are connected to a device which can give an electrical stimulus to the brain, which can be varied in intensity and frequency. The effect of the electrical stimulation is to inhibit cell activity, and increasing Þber activity 17, 2 . In particular, the Þring rate and pattern of individual neurons is affected 2 , along with the localized release of dopamine and glutamate 19, 20, 21 , which helps control the patient's tremors. An added bonus is that the stimulation also increases blood ßow and stimulates neurogenesis 22 . " 4

PAGE 11

Although the individual effects of DBS are well know, it is still unclear how these actually affect Parkinson's disease as a whole. However, the documented effect results of the treatment are well established over a range of patients. DBS was approved by the Food and Drug Administration (FDA) in 2002 for use in Parkinson's patients whose symptoms are not adequately controlled by medication. The centers where the treatment is available generally select patients based a number of factors, but which generally include an inadequate response to levodopa, onoff ßuctuations and tremors. The procedure itself starts with taking a set of images of the patient's brain using CT and MRI scans. Once the surgery starts, the surgeon drills a burr hole in the patient's skull for the insertion of the electrical probe, and the quadrature deep brain stimulation electrodes. General anesthesia is not normally used, as the patient's reactions during the procedure need to be monitored. Once the surgeon considers that the stimulation electrode is suitably placed, the patient's reactions to different levels of stimulus can be monitored, particularly for reduction in tremors and rigidity, and also any adverse effects which may occur. The costs for this type of surgery can vary anywhere from $28,000 to $50,000. In addition, there may be additional costs for controller setting adjustments 24 . In some case, the patient may have a poor response to the treatment, which is often caused by suboptimal positioning of the electrodes 25 . These can often be rectiÞed, but will generally require additional surgery, either to reposition or to remove and replace the electrodes. This obviously increase the overall cost, and stress to the patient. " 5

PAGE 12

Although there can be difÞculties, there is generally a positive outcome for the patient in the form or reduced tremors and rigidity, and also reduced levels of medication required. Some patients have had symptom improvement lasting more than 10 years 23 . The addition of Optical Coherence Tomography imaging to the DBS surgical process would give the surgeon real time imaging capability. This could help reduce the misplacement of the electrodes, reduce the the overall time taken for the surgery, and allow the patient to be anesthetized. OPTICAL COHERENCE TOMOGRAPHY OVERVIEW Optical Coherence Tomography (OCT) is an imaging method which can produce both 2 dimensional and 3 dimensional images with high resolution. It was pioneered by James Fujimoto, Eric Swanson and their team at MIT 26 . They originally developed the technology as a method of imaging live structures within the eye. Since the technology Þrst appeared in 1991 26 , it has developed rapidly and has begun to enter a range of Þelds where previously imaging was difÞcult or impossible. The original intent was for use in ophthalmology, it is now spreading to many other areas of medical imaging. The Þrst commercial OCT system was produced by Carl Zeiss Meditec in 1996, and at present there are over 20 other manufacturers producing commercial systems 27 . The Þrst systems based on Þber optics for intravascular imaging have only recently appeared, and are now approved not only in Europe and Asia, but the US also 28, 29 . Other commercial applications are being developed all the time. One of the most recent is for the use in cardiovascular surgeries, which uses Þbre optic technology as the coupled with gradient index (GRIN) lens technology, which allows for 3-D imaging within " 6

PAGE 13

arteries. This development has been helped by the increases in speed of imaging processors, and the reduction of component sizes in the devices 29 . The advantages of OCT imaging are that it can produce 3-dimensional images with resolutions of ~1 #m 30, 31, 32 , but with a tissue penetration of several millimeters, dependent on the tissue scattering parameters of the sample being imaged, it has better imaging depth than conventional confocal microscopy 33 . Typically the light source used is in the near infrared, which is non-ionizing, and so safe for repeated examinations of the same sample. With the appropriate electronic controller and image processor, OCT is also capable of real time imaging. The main principles behind OCT is the use of an interferometer that detects reßected light from the imaged object and compares this with a reference beam to allow the depth proÞle of the sample to be calculated. This technique in conjunction with the long center wavelength and incoherent light give OCT imaging it's unique properties over other imaging methods, such as wide Þeld and confocal microscopy. Tissue scattering is reduced at longer wavelengths 34 . There are several methods of implementing OCT. Each method contains some form of interferometer, and the main detection methods are either CCD camera or photo diode for time domain (TD) OCT, or by spectrometer for spectral domain (SD) OCT. TD OCT systems were the Þrst to be implemented 26 , and perform their depth scan based on low coherence interferometry, which was originally used in optical Þbers 35 . The wide bandwidth of the light signal used results in a low coherence length, and so only reßections from the sample which are near to the reference length show interference. In " 7

PAGE 14

order to scan the sample in depth, the reference length is changed to build up a model of the sample's tomography. Spectral domain OCT systems generally used a Þxed reference path, and image the object depth by measuring the spectral response of the interference pattern 36 . The data from the spectrometer, encoded in frequency space, can be converted using a Fourier transform algorithm into an image of the scanned object 37, 38 . Another technique, swept source OCT, uses a narrow frequency source whose wavelength is quickly adjusted across the range used. The resulting power from the interference can then be used to create a proÞle of the sample. For this method only a single photodetector is required, the complication being in producing a range of frequencies for scanning the sample. This would normally be done by using an external cavity semiconductor laser with a fast tunable Þlter 38, 40, 41 . There are a number of advantages and disadvantages for each method. SD OCT allows for high scan rates since there is no alteration in the reference arm length during scans, but there is more processing of the data following the scan. TD OCT does not require Fourier calculations on the data, but the scan rates are reduced because of the extra mechanical movements needed in the reference arm. SD OCT has proved the most popular for medical imaging, mostly as a result of scanning speed, and with appropriate processing hardware 42 , the ability to produce real time images of the sample. Most optical systems use a high numerical aperture (NA) to achieve a high resolution. OCT generally uses a low NA, since in applications like ophthalmology, a high NA cannot be achieved due to the distance between the outside of the eye and the " 8

PAGE 15

retina. The low NA increases the depth of Þeld, and the axial resolution is governed by measuring the level of interference of the light source used rather than the NA. In both types of OCT system, optical coherence microscopy (OCM) can be achieved where an image is acquired in a transverse plane to the light beam used 43 . This combines the usefulness of regular 2-D microscopy using confocal or some other method, with the larger imaging capability of OCT. Although OCT has the capability of high depth penetration, this is usually limited by the light scattering effect of the tissue being imaged, and often results in a maximum usable depth of a few millimeters . To give a better image deeper into tissue or organs, OCT can be used with endoscope probes 44 connected to Þbre optics, which can then be used to probe directly into tissues. O ne of the features of OCT is it's ability to produce depth related images of tissue structures without the need for staining or ßuorophores, the method can be enhanced by the use of these techniques. The addition of contrast agents to samples can affect reßectivity and absorption, and so can be used to enhance speciÞc features in the sample 45 . OCT sensitivity is largely dependent on the detection of returning signal light over noise from the sample. This is mostly dependent on the tissue being imaged, but most OCT systems can produce sensitivities up to 100dB 46 . The imaging speed varies based on the type of OCT system used. The limiting factor for TD OCT is the speed of changing the reference arm length. For SD OCT it is the acquisition and transfer rate from the spectrometer, and for SS OCT it is the sweep rate of the light source used. " 9

PAGE 16

For medical applications the image sample rate is one of the most important factors, particularly for imaging live samples, as some movement often cannot be avoided 47 . This is the reason why SD OCT has become the most popular method for medical imaging. Some systems are capable of scan rates in hundreds of kilohertz 48 , which can easily produce real time images. OCT performance is also affected by the properties of the sample being imaged. Two main factors in this are the absorption and scattering of light at the frequency or frequencies being used, which affects the overall penetration depth of the signal 48 . For samples which scattering or absorption of the signal are signiÞcant will require either a signal reducing element in the reference arm 49 to compensate for the reduced signal, or this may be achieved in the signal processing system 50 . Biological tissues tend to scatter light, and so this becomes the main factor in imaging the sample 51 . This is a result of biological tissue usually being made of a large number of different components, all with their own light absorption and scattering properties. For biological samples, absorption tends to be the least for wavelengths between 600 nm to 1300 nm, as this is the region where absorption of water and other common tissue constituents like hemoglobin and melanin are low 52 . The optimum wavelength to use largely depends on the constituent of the speciÞc tissue being imaged. Longer wavelengths allow for deeper penetration into tissues, so for samples where water content is not signiÞcant, longer wavelengths can be used 53 . For applications where water is signiÞcant, like ophthalmic imaging, 800 nm is the most common frequency used 54 . " 10

PAGE 17

Photodynamic Therapy: Current Status and Future Directions ScientiÞc Figure on ResearchGate. Available from: https://www.researchgate.net/262265731_Þg2_Figure-6The-optical-phototherapeutic-window-where-the-absorption-and-scattering-of [accessed 26 Nov, 2017] FIGURE 2 Absorbance of Hemoglobin and Water Figure 2 above shows the relative absorbance of hemoglobin and water against wavelength. For highly scattering samples, speckle can also be an issue in imaging. This is a phenomena which manifests itself as as Þne graininess in the image produced. It is generally caused by a random distribution of scatterers in the tissue 55 , and can be used to determine features of the underlying tissue 56, 57 . The downside is that if the amount of speckling is excessive, it can mask other features in the image. A number of methods are available to reduce speckle, some of which depend on varying the light signal features 55 , and others using data processing techniques to remove the unwanted noise 58-62 . Despite speckle being generally viewed as a problem, it can also be used in a positive manner. One example is in the measurement of vascular changes which can " 11

PAGE 18

occur in response to accidental trauma, but also in response to events such as the implantation of electrodes into the motor cortex for artiÞcial limb control 63 , or in the case of electrode implantation for Parkinson's disease 64 . Monitoring the vascular changes around the electrodes can give indications of early failure of the implants. In speckle variance analysis this is done by analyzing the amplitude changes in the OCT signal, which cause changes in the speckle pattern. These changes can then be used to visualize the changes occurring in the vasculature. This is similar to the amplitude decorrelation method, which also uses the OCT signal amplitude changes to detect vasculature changes. Similar to imaging vasculature, OCT can also be useful in examining neural pathways 65, 66 as it is capable of identifying nerve Þbers from surrounding tissue by differences in their scattering properties. APPLICATIONS OF OCT The original OCT systems were developed for use in ophthalmology, and have become the standard tool for directly examining ocular structures 67, 68, 69 . OCT can provide a 3-D image of a living retina comparable with a light microscopy image from a histological section 70 . OCT is a non-invasive technique, which requires no direct contact with the eye or labeling. It has become the standard diagnostic tool for a number of diseases, including macular degeneration 71 , glaucoma 72 and diabetic retinopathy 73, 74 . " 12

PAGE 19

FIGURE 3 Sample Commercial OCT System In cardiology OCT has been used to identify vulnerable plaques 76 , and to aid in the diagnosis and characterization of atherosclerotic plaques 77 that were not able to be accurately imaged with conventional techniques, such as ultrasound. OCT can be used to measure the thickness of plaque caps 79 , the presence of lipids below the cap 75 , and macrophage inÞltration of the cap 78 . This has made OCT a valuable tool in cardiovascular diagnosis. Also in the cardiovascular area, OCT has also been applied to monitoring blood ßow in capillary vessels 80 . This has been used to monitor the blood ßow in the cortical capillary network of mice. Other applications have been found in studying the hemodynamic changes in mammalian embryos. This involved the use of a Doppler swept source OCT system for monitoring both changes in the blood vessels " 13 Angiovue TM Imaging System. Optovue Inc.

PAGE 20

and blood ßow in mice embryos 81 . This could lead to advances in early diagnosis and treatment of congenital cardiovascular defects. OCT is also being used for follow up monitoring of coronary stents 82 . The catheter type OCT systems can be used to examine the stent in vivo and make sure that it is correctly anchored, and there is no other issues with the placement. Catheter based OCT systems have also been developed for imaging other hollow organs with frame rates high enough to allow real time imaging 83 . Solid organs can also be imaged 3 , the main difference being the wavelength of light used for imaging, as different tissue types have different scattering and absorption properties. Considering the current applications of OCT for imaging in vivo , extending this into brain imaging seems a reasonable step to take. Towards this end, work has been done to combine OCT with other forms of imaging currently used in neural monitoring. Optical intrinsic signal imaging (OISI) has been combined with OCT to exam activity in the rat somatosensory cortex 84 . This was done through a thinned skull rather than removing a skull section completely as would be required with optical imaging. At present, the current methods of identifying the affected areas of the brain in a disease like Parkinson's are only accurate to within several millimeters at best, being able to produce 3-D images with micrometer resolution should be a signiÞcant advantage to both the patients and surgeons. As has been previously mentioned, SD OCT has become the primary method for use in medical imaging. This is because of its higher speed imaging, and the reduced number of moving parts within the system. For these reasons, SD OCT was chosen as the method for this project. " 14

PAGE 21

CHAPTER II SYSTEM DESIGN HARDWARE The hardware design of the system was based on a Michelson interferometer (Figure 4). The system was constructed using standard components supplied by ThorLabs Inc. FIGURE 4 Schematic of Michelson Interferometer In Figure 4, L i, L s , L r , and L d are the optical path lengths for the LED output to the beam splitter, the sample arm to the beam splitter, the reference arm to the beam splitter, and the detector to the beam splitter respectively. " 15 Li distance laser to beamsplitter L R distance reference mirror to beamsplitter L S distance sample to beamsplitter L D distance detector to beamsplitter Beamsplitter (50/50) L R L i L S L D Reference Mirror Sample Mirror Laser Light Source Detector

PAGE 22

E 0 is the electric Þeld at the LED output. The electric Þeld at the detector can then be expressed as: 1.0 This can be rearranged and given in intensity: 1.1 where c is the speed of light, and n is the refractive index. Simplifying gives the following equation: 1.2 Where I 0 is the intensity of the source diode, and I opt is the intensity at the detector. $z is the difference in the path lengths of the reference and sample arms. The following formula gives the axial resolution for the system. This is a relation of the central frequency of the light source being used (lambda 0) to the full bandwidth half maximum power ($lambda) of the source. 1.3 " 16

PAGE 23

This formula gives the lateral resolution for the system, with f obj being the focal length of the objective lens, and d representing the beam diameter incident on the objective lens. 1.4 Working from the principle of interferometry, the following drawing shows the initial design of the system. This was an open arrangement with the mirrors and lenses in adjustable holders. (Figure 5) " 17

PAGE 24

FIGURE 5 Initial OCT System Layout This design was used to determine the best arrangement for the system, and for performing alignment tests to produce interference between the two arms of the system. The stage control and data analysis software was written based on this conÞguration. After initial testing, it was found that the system was too unstable in it's present form, and tended to move out of alignment too easily. To attempt to correct for this, the following changes were made to the design (Figure 6) " 18 Reference Mirror Sampling Mirror and Stage Light Source Mirror Mirror Spectrometer Beam Splitter Collimator Data to PC Focusing lens Focusing lens The output of the light source is collimated and then directed by two mirrors into the beam splitter, when it is split 50:50 to the sample and reference. The returned light is recombined by the splitter, and directed to the focusing lens for the spectrometer.

PAGE 25

FIGURE 6 Current OCT System Layout This new conÞguration proved to be considerably more stable, and could be more easily realigned if needed. Another advantage of the tube system was that there was much less interference in the system, since the light paths were almost entirely enclosed. " 19 Collimator Lens in Microadjustable Housing Sampling Mirror and Stage Beam Splitter Reference Mirror and Lens in Tube Housing Spectrometer Data to PC Tube with Focusing Lens LED Light Source Fibre optic cable The LED output is collimated and the split 50:50 by the beam splitter and directed to the reference mirror and lens, and the sample. The reßected light is recombined in the splitter and directed through the focusing lens to the spectrometer.

PAGE 26

SOFTWARE The software for the system was divided into two major sections, stage control and data analysis. The stage control section also performed data capture functions. STAGE CONTROL SOFTWARE This was written in LabView, and performed the main control function for the scanning stage. This involved driving the x and y-axis stepper motors, and controlling the data acquisition from the spectrometer. In addition, the LabView software also provided on screen displays of the current data being acquired, and the Fourier transform of the data, to give the user a representation of the data being acquired from the sample under observation. The software was written as a modular event driven package, which could be extended at any time to incorporate other controls or data acquisition requirements without disrupting the existing software. Details on the current software conÞguration and operations are given in Appendix 1. DATA ANALYSIS SOFTWARE Data analysis and image production was done using Matlab software. For the initial stage of the project the analysis was performed ofßine, after the data capture was completed. This was done to facilitate the processing of large datasets without the use of hardware other than regular desktop computers. " 20

PAGE 27

The image processing software was divided into two sections, the pre-processing code and the image generation code. The division was done to help speed up Þnal image generation, since the image generation code contains segments which can be modiÞed to improve the image, whereas the pre-processing code does not need modiÞcations for different images. The software uses two data Þles, the primary Þle contains the raw data from the scan of the sample, and the other Þle contains the meta data describing the main dat Þle. The meta Þle data consists of the following data: The base name for the Þle set, Threshold scan values for a slide with a coverslip, Threshold scan values for slide a slide without a coverslip, Number of motor steps per datapoint, Number of rows per data scan, Number of data points per row, Scan rate value, Number of axes divisions (for plotting), The Picomotor step size in nanometers, The Spectrometer resolution in pixels. This data is then used to construct the name of the primary data Þle to be loaded, how to process the raw data into the Þnal image, and how to plot the image data. " 21

PAGE 28

CHAPTER III RESULTS The following images are a representative sample of what the system can produce. The Þrst plot (Figure 8) shows an image of two microscope cover slips attached to form a stair like arrangement as shown in Figure 7. Scan parameters: 300 points per line @ 30µm/point 16 lines, 9.8 Mbytes (approximately) FIGURE 7 Stacked Cover Slip Arrangement " 22 Movable Stage Mirror Stacked microscope slides Mirror Slides Movable Stage Front View Side View

PAGE 29

FIGURE 8 Stacked Cover Slips The next image (Figure 10) shows the scan of a Þngerprint placed on a microscope coverslip. A photograph of the Þngerprint is shown for comparison (Figure 9). Obviously the regular photograph cannot show the depth of the Þngerprint to any signiÞcance, but it does serve to highlight the lack of smoothness in the ridges and valleys of the print. Scan parameters: 200 points per line @ 30 µm/point 62 lines 25.5 Mbytes (approximately) " 23

PAGE 30

FIGURE 9 Fingerprint Photographs (standard size and enlarged) FIGURE 10 Fingerprint " 24

PAGE 31

CHAPTER IV DISCUSSION OCT microscopy has expanded into many area of medical imaging since its inception in ophthalmology 24 . Spectral domain OCT has become the most popular for medical applications, and has been the method chosen for this project. The images shown in the results section were chosen to demonstrate the capabilities of spectral domain OCT, and they highlighted a number of features of the technique. The method has the capability to scan a sample to a very small resolution, and coupled with this is the generation of large quantities of data. This was the main reason that ofßine processing of image scans was chosen. The raw data contains intensity information and differences in the path lengths of the two arms. Since different material compositions can cause variations in the absorption and scattering of the source light, interpretation of the data gathered is signiÞcant to producing an accurate representation of the sample 85, 86 . In addition, the type of sample being scanned is also signiÞcant in its interpretation. A microscope slide with a cover slip needs the software to identify the slide and coverslip positions and interpret the data between them. The intensity of the returned signal can be considerably reduced by the coverslip. For this reason the software needs to be aware of the type of sample being imaged, and adjust its parameters accordingly to provide an accurate image. To demonstrate this, the following three images show the data plots from the analyzer as it scans across a slide with two cover slips in a stepped arrangement as in Figure7. " 25

PAGE 32

FIGURE 11 Fourier Data of Slide Only FIGURE 12 Fourier Data of Slide and Single Coverslip FIGURE 13 Fourier Data of Slide and Two Coverslips " 26 First Coverslip First Coverslip Second Coverslip

PAGE 33

The left screen shows the amplitude of the combined reference and sample signals at the spectrometer against wavelength. The right screen shows the Fourier transform of the amplitude/wavelength data converted to give a signal intensity and distance from the baseline, in this case a microscope slide. Figure 11 shows the beam focused on the slide only, and so the reference and signal paths are of equal length. The Fourier plot shows a signal amplitude at the baseline only. Figure 12 shows the signal and Fourier data when the beam has moved over the Þrst coverslip. The Fourier plot distance is representative of the thickness of the coverslip at the point where the beam is focused. The left screen shows a representation of the interference patterns from the different reference and sample lengths. Figure 13 gives a representation of the sample beam directed at both coverslips. As can be seen, the amplitude return from the Þrst coverslip is considerably reduced, since the signal is now partly reßected from the surface of the second coverslip, and partly attenuated by passing through the coverslip. For slides with cover slips, the intensity of the signal returns on the sample below it will be much less than for a sample on a slide without a cover slip. The analysis software needs to be a ware of this, and be able to process the data accordingly. " 27

PAGE 34

CHAPTER V FUTURE WORK The current state of the OCT system has indicated the potential for further development and improvements. The potential upgrades considered at this time fall into two main areas, hardware and software. HARDWARE The experience gained from working on the system was that alignment of the components was critical to correct operation of the device. The majority of this work could be automated with the addition of motor driven optical components, along with a micro controller system such as a Beagle Bone. This is a low cost commercially available device, and could be used as a main controller for the OCT. The Beagle Bone itself does not incorporate analog sensors, but this could be overcome by using PIC Micro-controllers interfaced to the Beagle Bone. These are also commercially available at very low costs, in the order of $1-10 per unit. This could then be programmed to take over most of the alignment tasks for the system. In addition, changing the light source to a wide bandwidth source centered around 1300nm would give better depth resolution. The volume of data generated is large even for a small area scan, and so the Beagle Bone could also be used as a data processor, and is capable of driving high deÞnition monitors directly, and so could be used as a primary display device. " 28

PAGE 35

SOFTWARE The current mixture of LabView and Matlab software could be improved by moving to a language such as Python, which would be capable of doing both the data capture and image processing itself. The Beagle Bone is capable of running Python code, so this would mean that only one language would be need throughout the project. " 29

PAGE 36

REFERENCES 1. Lee A, Gilbert RM. Epidemiology of Parkinson Disease. Neurol Clin NA. 2017;34(4): 955-965. doi:10.1016/j.ncl.2016.06.012 . 2. Vitek JL. Mechanisms of deep brain stimulation: excitation or inhibition. Mov Disord 2002;17:Suppl 3:S69-S72. 3. Ahsen OO, Tao YK, Potsaid BM, et al. Swept source optical coherence microscopy using a 1310 nm VCSEL light source. 2013;21(15):18021-18033. doi:10.1364/OE. 21.018021. 4. Mariampillai A, Standish BA, Munce NR, et al. Doppler optical cardiogram gated 2D color ßow imaging at 1000 fps and 4D in vivo visualization of embryonic heart at 45 fps on a swept source OCT system. 2007;15(4):1627-1638. 5. Tanner CM, Aston DA. Epidemiology of Parkinson's disease and akinetic syndromes. Curr Opin Neurol 2000;13:427-30. 6. Fahn S. Description of Parkinson's disease as a clinical syndrome. Ann N Y Acad Sci 2003;991:1-14. 7. Lang AE, Lozano AM. Parkinson's disease (Þrst of two parts). N Engl J Med 1998;339:1044-53. 8. Martikainen KK, Luukkaala TH, Marttila RJ. Parkinson's disease and working capacity. Mov Disord 2006;21:2187-91. 9. Dodel RC, Eggert KM, Singer MS, Eichhorn TE, Pogarell O, Oertel WH. Costs of drug treatment in Parkinson's disease. Mov Disord 1998;13:249-54. 10. Whetten-Goldstein K, Sloan F, Kulas E, Cutson T, Schenkman M. The burden of Parkinson's disease on society, family, and the individual. J Am Geriatr Soc 1997; 45:844-9. 11. Willis AW, Schootman M, Evanoff BA, Perlmutter JS, Racette BA. Neurologist care in Parkinson disease: a utilization, outcomes, and survival study. Neurology 2011;77:851-7. 12. Alexander GE, DeLong MR, Strick PL. Parallel organization of functionally egregated circuits linking basal ganglia and cortex. Annu Rev Neurosci 1986;9:357-81. 13. Delong MR, Georgopoulos AP, Crutcher MD, Mitchell SJ, Richardson RT, Alexander GE. Functional organization of the basal ganglia: contributions of single-cell recording studies. Ciba Found Symp 1984;107:64-82. 14. Wagle Shukla A, Okun MS. Personalized medicine in deep brain stimulation through utilization of neural oscillations. Neurology 2012;78:1900-1. 15. Fahn S. The spectrum of levodopa-induced dyskinesias. Ann Neurol 2000;47: Suppl 1:S2-S9. 16. Okun MS, Fernandez HH, RodriguezRL, Foote KD. Identifying candidates for deep brain stimulation in Parkinson's disease: the role of the primary care physician. Geriatrics 2007;62:18-24. 17. McIntyre CC, Hahn PJ. Network perspectives on the mechanisms of deep brain stimulation. Neurobiol Dis 2010;38: 329-37. 18. Wichmann T, DeLong MR, Guridi J, Obeso JA. Milestones in research on the pathophysiology of Parkinson's disease. Mov Disord 2011;26:1032-41. " 30

PAGE 37

19. TawÞk VL, Chang SY, Hitti FL, et al. Deep brain stimulation results in local glutamate and adenosine release: investigation into the role of astrocytes. Neurosurgery 2010;67:367-75. 20. Lee KH, Hitti FL, Chang SY, et al. High frequency stimulation abolishes thalamic network oscillations: an electrophysiological and computational analysis. J Neural Eng 2011;8(4):046001. 21. Lee KH, Chang SY, Roberts DW, Kim U. Neurotransmitter release from highfrequency stimulation of the subthalamic nucleus. J Neurosurg 2004;101:511-7. 22. Vedam-Mai V, van Battum EY, Kamphuis W, et al. Deep brain stimulation and the role of astrocytes. Mol Psychiatry 2012;17:124-31. 23. Castrioto A, Lozano AM, Poon YY, Lang AE, Fallis M, Moro E. Ten-year outcome of subthalamic stimulation in Parkinson disease: a blinded evaluation. Arch Neurol 011;68:1550-6. 24. Tomaszewski KJ, Holloway RG. Deep brain stimulation in the treatment of Parkinson's disease: a cost-effectiveness analysis. Neurology 2001;57:663-71. 25. Okun MS, Tagliati M, Pourfar M, et al. Management of referred deep brain stimulation failures: a retrospective analysis from 2 movement disorders centers. Arch Neurol 2005;62:1250-5. 26 . Huang D, Swanson EA, Lin CP, Schuman JS, Stinson WG, Chang W, Hee MR, Flotte, Gregory K, PuliaÞto CA, Fujimoto JG (1991) Optical coherence tomography. Science 254 (5035):1178Ð1181 27. Smolka G (2010) Optical coherence tomography 2010: technology, applications, and markets (Market Report OM-51). Strategies Unlimited (PennWell Corporation), MountainView 28. Swanson E (2017) LightLab Imaging announces FDA clearance of C7-XRª coronary OCT products in the United States. Opt Coher Tomogr News 5/5/2010. http:// www.octnews.org/articles/ 2008632/lightlab-imaging-announces-fda-clearance-of-c7-xrt/, accessed 10 Apr 2017 29. Volcano Corp. (2017) Volcano Corporation announces receipt of CE mark for its OCT imaging system and catheter. Volcano Corp., San Diego ReleaseID=434187 , accessed 10 Apr 2017 30. Hartl I, Li XD, Chudoba C, Ghanta RK, Ko TH, Fujimoto JG, Ranka JK, Windeler RS (2001) Ultrahigh-resolution optical coherence tomography using continuum generation in an airÐsilica microstructure optical Þber. Opt Lett 26 (9):608Ð610 31. Pova% ay B, Bizheva K, Unterhuber A, Hermann B, Sattmann H, Fercher AF, Drexler W, Apolonski A, Wadsworth WJ, Knight JC, Russell PSJ, Vetterlein M, Scherzer E (2002) Submicrometer axial resolution optical coherence tomography. Opt Lett 27(20): 1800 Ð1802 32. Drexler W (2004) Ultrahigh-resolution optical coherence tomography. J Biomed Opt 9(1):47Ð748. 33. InouŽ S (2006) Chap 1. In: Handbook of biological confocal microscopy, 3rd edn. Springer, Berlin, ISBN: 978-0-387-25921-5 34. S. J. Matcher, M. Cope, D. T. Delpy, (1997) In vivo measurements of the wavelength dependence of tissue-scattering coefÞcients between 760 and 900 nm measured with time-resolved spectroscopy. Appl. Opt. 36, 386-396 " 31

PAGE 38

35. Takada K, Yokohama I, Chida K, Noda J (1987) New measurement system for fault location in optical waveguide devices based on an interferometric technique. Appl Opt 26(9):1603 Ð1606 36. Fercher AF, Hitzenberger CK, Kamp G, El-Zaiat SY (1995) Measurement of intraocular distances by backscattering spectral interferometry. Opt Commun 117(1Ð2): 43Ð48 37. Bail MA, HŠusler G, Herrmann JM, Lindner MW, Ringler R (1996) Optical coherence tomography with the "spectral radar:" fast optical analysis in volume scatterers by shortcoherence interferometry. Proc SPIE 2925:298Ð303 38. HŠusler G, Lindner MW (1998) Coherence radar and spectral radarÑnew tools for dermatological diagnosis. J Biomed Opt 3(1):21 Ð31 39. Yun S-H, Boudoux C, Tearney GJ, Bouma BE (2003) Highspeed wavelength-swept semiconductor laser with a polygonscanner-based wavelength Þlter. Opt Lett 28(20): 1981 Ð1983 40. Choma MA, Hsu K, Izatt JA (2005) Swept source optical coherence tomography using an all-Þber 1300-nm ring laser source. J Biomed Opt 10(4):044009 41. Huber R, Wojtkowski M, Taira K, Fujimoto JG, Hsu K (2005) AmpliÞed, frequency swept lasers for frequency domain reßectometry and OCT imaging: design and scaling principles. Opt Express 13(9):3513 Ð3528 42. Huber R, Wojtkowski M, Fujimoto JG (2006) Fourier domain mode locking (FDML): a new laser operating regime and applications for optical coherence tomography. Opt Express 14 (8):3225Ð3237 43. Izatt JA, Hee MR, Owen GM, Swanson EA, Fujimoto JG (1994) Optical coherence microscopy in scattering media. Opt Lett 19 (8):590Ð592 44. Tearney GJ, Brezinski ME, Bouma BE, Boppart SA, Pitris C, Southern JF, Fujimoto JG (1997) In vivo endoscopic optical biopsy with optical coherence tomography. Science 276 (5321):2037Ð2039 45. Akkin T, Landowne D, Sivaprakasam A. Detection of neural action potentials using optical coherence tomography : intensity and phase measurements with and without dyes. 2010;2(August):1-10. doi:10.3389/fnene.2010.00022. 46. Marschall S, Sander B, Mogensen M, J¿rgensen TM, Andersen PE. Optical coherence tomography-current technology and applications in clinical and biomedical research. Anal Bioanal Chem. 2011;400(9):2699-2720. doi:10.1007/s00216-011-5008-1. 47. Yun S-H, Tearney GJ, de Boer JF, Bouma EB (2004) Motion artifacts in optical coherence tomography with frequency-domain ranging. Opt Express 12(13):2977Ð2998 48. Schmitt JM, KnŸttel A, Yadlowsky M, Eckhaus MA (1994) Optical-coherence tomography of a dense tissue: statistics of attenuation and backscattering. Phys Med Biol 39(10):1705 Ð1720 49. Hitzenberger CK, Baumgartner A, Drexler W, Fercher AF (1999) Dispersion effects in partial coherence interferometry: implications for intraocular ranging. J Biomed Opt 4(1):144Ð151 50. de Boer JF, Saxer CE, Nelson JS (2001) Stable carrier generation and phaseresolved digital data processing in optical coherence tomography. Appl Opt 40(31): 5787 Ð5790 " 32

PAGE 39

51. Brezinski ME, Tearney GJ, Bouma BE, Izatt JA, Hee MR, Swanson EA, Southern JF, Fujimoto JG (1996) Optical coherence tomography for optical biopsy: properties and demonstration of vascular pathology. Circulation 93(6):1206 Ð1213 52. Parrish JA (1981) New concepts in therapeutic photomedicine; photochemistry, optical targeting and the therapeutic window. J Invest Dermatol 77(1):45 Ð50 53. Sharma U, Chang EW, Yun SH (2008) Long-wavelength optical coherence tomography at 1.7 #m for enhanced imaging depth. Opt Express 16(24):19712 Ð19723 54. Pova% ay B, Bizheva K, Hermann B, Unterhuber A, Sattmann H, Fercher AF, Drexler W, Schubert C, Ahnelt PK, Mei M, Holzwarth R, Wadsworth WJ, Knight JC, Russel PSJ (2003) Enhanced visualization of choroidal vessels using ultrahigh resolution ophthalmic OCT at 1050 nm. Opt Express 11 (17):1980Ð1986 55. Schmitt JM, Xiang SH, Yung KM (1999) Speckle in optical coherence tomography. J Biomed Opt 4(1):95 Ð105 56. Gossage KW, Tkaczyk TS, Rodriguez JJ, Barton JK (2003) Texture analysis of optical coherence tomography images: feasibility for tissue classiÞcation. J Biomed Opt 8(3):570 Ð575 57. Hillman TR, Adie SG, Seemann V, Armstrong JJ, Jacques SL, Sampson DD (2006) Correlation of static speckle with sample properties in optical coherence tomography. Opt Lett 31(2):190 Ð192 58. Kulkarni MD, Thomas CW, Izatt JA (1997) Image enhancement in optical coherence tomography using deconvolution. Electron Lett 33(16):1365 Ð1367 59. Schmitt JM (1998) Restoration of optical coherence images of living tissue using the CLEAN algorithm. J Biomed Opt 3 (1):66Ð75 60. Xiang SH, Zhou L, Schmitt JM (1998) Speckle noise reduction for optical coherence tomography. Proc SPIE 3196(1):79 Ð88 61. Rogowska J, Brezinski ME (2000) Evaluation of the adaptive speckle suppression Þlter for coronary optical coherence tomography imaging. IEEE T Med Imaging 19(12): 1261 Ð1266 62. Wang RK (2005) Reduction of speckle noise for optical coherence tomography by the use of nonlinear anisotropic diffusion. Proc SPIE 5690(1):380Ð385 63. Hammer DX, Lozzi A, Abliz E, et al. Longitudinal vascular dynamics following cranial window and electrode implantation measured with speckle variance optical coherence angiography. 2014;5(8):2823-2836. doi:10.1364/BOE.5.002823. 64. Hammer DX, Welle CG, Hammer DX, Lozzi A, Boretsky A, Welle CG. Acute insertion effects of penetrating cortical microelectrodes imaged with quantitative optical coherence angiography microelectrodes imaged with quantitative. 2016;3(2). doi: 10.1117/1.NPh.3.2.025002. 65. Wang H, Black AJ, Zhu J, et al. NeuroImage Reconstructing micrometer-scale Þ ber pathways in the brain : Multi-contrast optical coherence tomography based tractography. Neuroimage. 2011;58(4):984-992. doi:10.1016/j.neuroimage.2011.07.005. 66. Wang H, Zhu J, Reuter M, et al. NeuroImage Cross-validation of serial optical coherence scanning and diffusion tensor imaging : A study on neural Þber maps in human medulla oblongata. Neuroimage. 2014;100:395-404. doi:10.1016/j.neuroimage. 2014.06.032 . 67. Fercher AF, Hitzenberger CK, Drexler W, Kamp G, Sattmann H (1993) In vivo optical coherence tomography. Am J Ophthalmol 116(1):113Ð114 " 33

PAGE 40

68. Swanson EA, Izatt JA, Hee MR, Huang D, Lin CP, Schuman JS, PuliaÞto CA, Fujimoto JG (1993) In vivo retinal imaging by optical coherence tomography. Opt Lett 18(21):1864 Ð1866 69. Izatt JA, Hee MR, Huang D, Fujimoto JG, Swanson EA, Lin CP, Shuman JS, PuliaÞto CA (1993) Ophthalmic diagnostics using optical coherence tomography. Proc SPIE 1877:136Ð144 70. Tanno N, Kishi S (1999) Optical coherence tomographic imaging and clinical diagnosis. Med Imaging Technol 17(1):3 Ð10 71. Kaiser PK, Blodi BA, Shapiro H, Acharya NR (2007) Angiographic and optical coherence tomographic results of the marina study of ranibizumab in neovascular agerelated macular degeneration. Ophthalmology 114(10):1868 Ð1875 72. Tan O, Li G, Ake Tzu-Hui L, Varma R, Huang D (2008) Mapping of macular substructures with optical coherence tomography for glaucoma diagnosis. Ophthalmology 115(6):949 Ð956 73. Soliman W, Sander B, Soliman KAE-N, Yehya S, Rahamn MSA, Larsen M (2008) The predictive value of optical coherence tomography after grid laser photocoagulation for diffuse diabetic macular oedema. Acta Ophthalmol 86(3):284 Ð291 74. Gaucher D, Sebah C, Erginay A, Haouchine B, Tadayoni R, Gaudric A, Massin P (2008) Optical coherence tomography features during the evolution of serous retinal detachment in 2716 S. Marschall et al. patients with diabetic macular edema. Am J Ophthalmol 145(2):289 Ð296 75. van der Meer FJ, Faber DJ, Sassoon DMB, Aalders MC, Pasterkamp G, van Leeuwen TG (2005) Localized measurement of optical attenuation coefÞcients of atherosclerotic plaque constituents by quantitative optical coherence tomography. IEEE T Med Imaging 24(10):1369 Ð1376 76. Jang I-K, Bouma BE, Kang D-H, Park S-J, Park S-W, Seung K-B, Choi K-B, Shishkov M, Schlendorf K, Pomerantsev E, Houser SL, Aretz HT, Tearney GJ (2002) Visualization of coronary atherosclerotic plaques in patients using optical coherence tomography: comparison with intravascular ultrasound. J Am Coll Cardiol 39(4):604 Ð609 77. Low AF, Tearney GJ, Bouma BE, Jang I-K (2006) Technology insight: optical coherence tomography: current status and future development. Nat Clin Pract Card 3:154 Ð162 78. Tearney GJ, Yabushita H, Houser SL, Thomas Aretz H, Jang I-K, Schlendorf KH, Kauffman CR, Shishkov M, Halpern EF, Bouma BE (2003) QuantiÞcation of macrophage content in atherosclerotic plaques by optical coherence tomography. Circulation 107(1):113Ð119 79. Kume T, Akasaka T, Kawamoto T, Okura H, Watanabe N, Toyota E, Neishi Y, Sukmawan R, Sadahira Y, Yoshida K (2006) Measurement of the thickness of the Þbrous cap by optical coherence tomography. Am Heart J 152(4):755.e1Ð755.e4 80. Lozzi A, Agrawal A, Boretsky A, Welle CG, Hammer DX. Image quality metrics for optical coherence angiography. 2015;6(7):2435-2447. doi:10.1364/BOE.6.002435. 81. Larina I V, Sudheendran N, Ghosn M, et al. Live imaging of blood ßow in mammalian embryos using Doppler swept-source optical coherence tomography. 2017;13(December 2008):10-12. doi:10.1117/1.3046716. " 34

PAGE 41

82. Gonzalo N, Serruys PW, Okamura T, Shen ZJ, Onuma Y, Garcia-Garcia HM, Sarno G, Schultz C, van Geuns RJ, Ligthart J, Regar E (2009) Optical coherence tomography assessment of the acute effects of stent implantation on the vessel wall: a systematic quantitative approach. Heart 95(23):1913Ð1919 83. Tsai T, Potsaid B, Tao YK, et al. Ultrahigh speed endoscopic optical coherence tomography using micromotor imaging catheter and VCSEL technology. 2013;4(7): 898-905. doi:10.1364/BOE.4.001119. 84. Chen Y, Aguirre AD, Ruvinskaya L, Devor A, Boas DA, Fujimoto JG. Optical coherence tomography ( OCT ) reveals depth-resolved dynamics during functional brain activation. 2009;178:162-173. doi:10.1016/j.jneumeth.2008.11.026. 85. Drexler, W. & Fujimoto, J. G. Optical coherence tomography : technology and applications. (Springer, 2008). 86. Al-Mujaini, A., Wali, U. K., & Azeem, S. (2013). Optical Coherence Tomography: Clinical Applications in Medical Practice. Oman Medical Journal , 28 (2), 86Ð91. http:// doi.org/10.5001/omj.2013.24 " 35

PAGE 42

APPENDIX Main Screen Overall Structure " 36

PAGE 43

Scan Row/Column Scan Slide ! " 37

PAGE 44

Move Stage Move to Zero " 38

PAGE 45

Overall Hierarchy " 39

PAGE 46

Matlab Software Data Pre-processing: % preprocDataFile takes the raw data from the image scan, and pre-processes % the data to a form used by the dataAnalyzer program. % clc; clear all; close all; A = importdata('sampleMeta.txt'); fBase = cell2mat(A.textdata(1));% File basename slideThreshold = A.data(1);% Threshold for slide w/coverslip printThreshold = A.data(2);% Threshold for slide wo/coverslip steps = A.data(3);% Motor steps per datapoint rows = A.data(4);% Rows in scan points = A.data(5);% Datapoints per row scanR = A.data(6);% Scan rate axesPts = A.data(7);% Axes divisions (for printing) picoNm = A.data(8);% Picomotor step size (nm) specPts = A.data(9);% Spectrometer resolution fNameRoot = sprintf('%s_%d_%ds_%dp_%dr', fBase, rows / 2, steps, points, scanR); samples = specPts; % spectrometer values per point fName = cat(2, fNameRoot, '.txt'); Þd = fopen(fName, 'r'); fp1 = fscanf(Þd, '%f'); j = 1; maxLen = length(fp1)/samples; fp2 = zeros(maxLen, samples); Þ = zeros(maxLen, samples); for i = 1:maxLen fp2(i, :) = fp1(j:j+(samples-1)); j = j + samples; Þ(i, :) = fft(fp2(i, :)); end rowLen = points; " 40

PAGE 47

rowLenStart = 1; rowLenEnd = rowLenStart + rowLen 1; % Divide points into rows, and create a cell vector of rows % Every other (2, 4, 6, ...) row gets ßipped, since the scan % tracks along a row, moves up, and tracks back to the X origin % (sort of: if the motors were ok) % fp3 = cell(1, rows); for i = 1:rows fp3{i} = fp2(rowLenStart:rowLenEnd, :); if(rem(i, 2) == 0) fp3{i} = ßipud(fp3{i}); end; rowLenStart = rowLenStart + rowLen; rowLenEnd = rowLenEnd + rowLen; end; fName = cat(2, fNameRoot, '.mat'); save(fName, 'fp3'); " 41

PAGE 48

Image Analysis: % dataAnalyzer takes the preprocessed data Þle and produces a 3-D plot of % the scanned image. % clc; clear all; close all; A = importdata('sampleMeta.txt'); fBase = cell2mat(A.textdata(1));% File basename slideThreshold = A.data(1);% Threshold for slide w/coverslip printThreshold = A.data(2);% Threshold for slide wo/coverslip steps = A.data(3);% Motor steps per datapoint rows = A.data(4);% Rows in scan points = A.data(5);% Datapoints per row scanR = A.data(6);% Scan rate axesPts = A.data(7);% Axes divisions (for printing) picoNm = A.data(8);% Picomotor step size (nm) specPts = A.data(9);% Spectrometer resolution fName = sprintf('%s_%d_%ds_%dp_%dr.mat', fBase, rows / 2, steps, points, scanR); load(fName, 'fp3'); ft_2 = zeros(rows, points); for j = 1:rows for i = 1:points ft = sqrt(real(fft(fp3{1,j}(i, :))).^2); if(contains(fBase, 'slide', 'IgnoreCase', true)) ft_1 = ft(100:300); ft_1(ft_1 < slideThreshold) = 0; else ft_1 = ft(50:specPts/2); ft_1(ft_1 < printThreshold) = 0; end ft_1(ft_1 > 0) = 1; ind = Þnd(ft_1, 1); % Þnd Þrst index where change starts if(isempty(ind)) ind = 0; end ft_2(j, i) = ind; % make up vector of index points end end " 42

PAGE 49

ft_2s = zeros(rows, points); for i = 1:rows ft_2s(i,:) = smooth(ft_2(i, :), 10); end surf(ft_2s); shading interp; axis tight; xlabel('Width: mm'); zlabel('Height: mm'); ylabel('Length: mm'); mm = (picoNm * points * steps) / 1e6; xt = linspace(0, points, 7); xticks(xt); xlabs = linspace(0, mm, axesPts); xlabelsc = cell(axesPts+1, 1); for i=1:length(xlabs) xlabelsc{i}=sprintf('%.2f', xlabs(i)); end xticklabels(xlabelsc); yp = rows mod(rows, 10); mm = (picoNm * yp * steps) / 1e6; ylabs = linspace(0, mm, axesPts); ylabelsc = cell(axesPts+1, 1); for i = 1:length(ylabs) ylabelsc{i} = sprintf('%.2f', ylabs(i)); end yticklabels(ylabelsc); [M, I] = max(ft_2s); zp = round(max(M)/10) * 10; mm = (picoNm * zp * steps) / 1e6; zlabs = linspace(0, mm, axesPts); zlabelsc = cell(axesPts+1, 1); for i = 1:length(zlabs) zlabelsc{i} = sprintf('%.2f', zlabs(i)); end zticklabels(zlabelsc); " 43