The Tower Undergraduate Research Journal Volume V

gtto

wer

.org

Cambridge Trust &NIH/Cambridge Fellow

VOLU

ME5

ChenMichaelSPOTLIGHT

UndergraduateResearcher

MerrittMacRESEARCHFor ResearchExecutive Vice PresidentCrossDr. StephenINTERVIEW

at the Georgia Institute of Technology Undergraduate Research JournalTOWERTHE

Want to break from the crowd?there’s a lab for that.

WWW.GTRI.GATECH.EDU/CAREERS/STUDENTS

JOBS FOR STUDENTS

The Georgia Tech Research Institute

(GTRI) is a highly regarded applied research

and development organization. Founded

in 1934, GTRI is an integral part of the

Georgia Institute of Technology, one of the

world’s top-ranked research universities.

With more than 1,500 scientists, engineers

and other professionals, GTRI helps

solve the most difficult problems facing

government and industry across the nation

and around the globe.

www.gtri.gatech.edu

THETOWERUndergraduate Research Journal

at the Georgia Institute of Technology

Fall 2012, Volume 5 The Tower is an interdisciplinary research journal forundergraduate students at the Georgia Institute of

Technology. The goals of our publication are to:showcase undergraduate achievements in research;

inspire academic inquiry; and promote Georgia Tech’s commitment to under-graduate research endeavors.

THE EDITORIAL BOARD

Editor-in-ChiefTyler J. Kaplan [email protected]

Managing Editor of Submission & ReviewMohamad Ali Najia [email protected]

Managing Editor of ProductionAllison Dowell [email protected]

Business Manager Michael Bonifacio [email protected]

Associate Editor of ReviewHarsimran Mann

Associate Editor of SubmissionChristopher Harper

Associate Editors of ProductionJaeyeon LeeKatie Staples

Webmaster/Web DesignerBradley Williford

STAFF

SPECIAL THANKS

The Tower would not have been possible without the assistance of the following people:

Dr. Christopher Reaves Faculty Advisor,Undergraduate Research Opportunities Program (UROP)

Dr. Carole Moore Assistant Vice Provost for Academic AffairsMs. Marlit Hayslett Georgia Tech Research Institute (GTRI)

Mr. Mac Pitts Student PublicationsDr. Kathryn Meehan Fellowships Office

Dr. Karen Adams Fellowships OfficeMs. Beth Bryant Office of Development

Mr. Charlie Bennett LibraryMr. Michael Chen Former Tower Editor

Dr. Lisa Yaszak Literature, Communication & CultureDr. Kenneth Knoespel Literature, Communication & Culture

Dr. Steven Girardot Assistant Vice Provost for Undergraduate EducationMs. Susan Parham Head,

Scholarly Communication and Digital CurationMs. Catherine Murray-Rust Vice Provost for Learning Excellence and Dean of Libraries

Dr. Rebecca Burnett Literature, Communication & CultureDr. Thomas Orlando Chemistry & BiochemistryMr. John Toon Enterprise Innovation Institute

Dr. Milena Mihail Computing Science & SystemsDr. Pete Ludovice Chemical & Biomolecular Engineering

Dr. Han Zhang College of BusinessMr. Michael Nees Psychology

Mr. Jon Bodnar Library

Faculty Reviewers

Dr. Tibor Beseds EconomicsDr. Wayne Book Mechanical Engineering

Dr. Amy D’Unger History, Technology & SocietyDr. Monica Halka Honors Program

Dr. Melissa Kemp Biomedical Engineering, GT/EmoryDr. Narayanan Komerath Aerospace Engineering

Dr. Don Lim Aerospace EngineeringDr. Lakshmi Sankar Aerospace Engineering

Faculty Advisory Board

Dr. Rosa I. Arriaga Interactive ComputingDr. Jeffrey A Davis Electrical & Computer Engineering

Dr. Amy D’Unger History, Technology & SocietyDr. Monica Halka Honors Program

Ms. Marlit Hayslett Georgia Tech Research InstituteDr. Pete J. Ludovice Chemical & Biomolecular Engineering

Dr. Milena Mihail College of ComputingDr. Lakshmi Sankar Aerospace Engineering

Dr. Han Zhang College of Business

Production TeamJung Min Lee

Shawna HagenNabila NazaraliFitrah Hamid

Victor LeeChristopher Cassidy

Graham Rhodes

Undergraduate Reviewers

Ang LiMatias Leguizamon

Shamus MoranHifza Sakhi

Laura LanierCasey Aultman

Simisola OludareJaheda Khanam

Christopher PaceKevin Daffon

Hoki TseKunal Shah

Michael MerrittSeungHo Shin

Graduate ReviewersChris QuintoJenna Wilson

Shriradha SenguptaAmulya ChiralaShereka Banton

Pritha Bagchi

Research ScientistDaniel C. Jones

Cover design by Katie Staples

The Tower would also like to thank the Georgia Tech Board ofStudent Publications, the Undergraduate Research

Opportunities Program, the Georgia Tech Library Information Center, Auxiliary Services, and the Student Government

Association for their support and contributions.

3 | The TOWER | SPRING 2013

Dear Reader,

This year The Tower Undergraduate Research Journal embarks on another journey; one that I believe will take us to new heights and that will allow us to become an even brighter beacon for undergradu-ate research on Georgia Tech’s campus. Over the past few years Michael Chen and I worked tirelessly to build an organization with a dedicated staff and a strong reputation across campus. When Michael departed Georgia Tech for Cambridge University last year, I believe that we finally reached that goal and were ready to begin a focus on improving and refining The Tower.

The fifth volume of The Tower ushers in a number of changes related to both the design and to the content of the journal. When we began the production of this edition, I directed our staff to move towards a more simplified design while continuing to utilize magazine-like elements. The result was what I believe to be our most sophisticated design yet. Our Editor of Production, Allison Dowell, suc-cessfully incorporated professional photography with the tasteful use of white space and a simplified color palate. I must take a moment to thank Van Jensen, the Editor in Chief of the Alumni Magazine for his sage advice on how to design a brilliant journal with limited resources. Also, I must recognize Michael James, my friend and professional photographer who created many of the brilliant images you see throughout the journal.

In addition to an improved design, we have begun what I hope will become a tradition of publishing guest authors. This edition features an article written by an undergraduate researcher at the Univer-sity of Oregon, Jake McGrew. One of my goals as Editor is to improve our relations with other under-graduate research journals across the country, and believe that this is beginning of an article exchange program that will serve to foster such relationships.

Finally, I would like to take a moment to thank the many individuals who have been so generous with their time and resources in supporting me as Editor, and the journal itself.

We hope you will enjoy reading Volume Five of the Tower Undergraduate Research Journal. Please feel free to contact me at [email protected] if you have any questions, comments or suggestions!

Sincerely,

Tyler J. KaplanEditor-in-Chief

L E T T E R F R O M T H E E D I T O R


Dear Reader,

In my position as the Director of Undergraduate Research and Student Innovation, I have the pleasure of serving as the faculty advisor to the Tower Undergraduate Research Journal. During the time that I have served as their advisor, I have come to learn that the organization truly lives out their mission to showcase, inspire and promote research on Tech’s campus.

This mission coincides in a number of ways with the mission of the Undergraduate Research Opportunities Pro-gram – as we too are continuously seeking to promote undergraduate research on Tech’s campus and get more students involved. One of our greatest goals is to get students to communicate their work with the world via pub-lications or presentations. The Tower offers a great opportunity for students toengage in this, as it allows them the real world experience of going through the peer review process and having their work published in a journal that is recognized both across campus and the nation. In the area of research presentations, we encourage our students to present at the Undergraduate Research Spring Symposium, which will occur on April 11th of this year. Additionally, the Tower hosts the Undergraduate Research Kaleidoscope each semester, which is an opportunity for students to present their work in the unique Pecha Kucha format.

We believe it is important for our students who are involved in research to explore any practical applications that exist for their work. If there is an opportunity for them to develop a prototype or commercialize their work, we want them to find it. Whenever possible, we believe it is important to move research and innovations that the students are capable of into mainstream society and put them to practical use to solve the world’s problems. The InVenture Prize is an excellent way for students to do this, and the competition will be held this year on March 13 th.

I hope that you enjoy reading Volume 5 of the Tower, and that you are inspired to get involved in research at Geor-gia Tech.

Sincerely,

Dr. Christopher Reaves Director, Undergraduate Research and Student Innovation L E T T E R

F R O M T H E D I R E C T O R



TA B L E O F CO N T E N T S

0 7

N E W SThe Environmental Cost of Capitalby Amulya Ch i ra l a

0 8 Ocean Circulation, Biogeochemistry, and Climateby Dan Jones , PhD

0 9 A Look at Robotics in Healthcareby J aheda Khanam

1 0 Advances in fMRI and Electrophysiologyby Mac Merr i t t

1 1 Developing Experimental Modelsfor Measuring Neuromuscular Improvement Following Rehabilitationby S imiso l a O. Oludare

1 2 A Water Soluble Copper(I) Selective Fluorescent Probe: New Application in Direct In-Gel Detection of a Copper Chaperone by Pr i tha Bagch i

1 3

I N T E R V I E W S

S P OT L I G H T : M i c h a e l C h e nCambridge Trust and NIH/Cam-bridge Fellow, Former Tower Editor

1 5 F E AT U R E : D r. S t e v e C r o s sExecutive Vice Presidentfor Research, Georgia Institute of Technology

1 7 S P OT L I G H T: D r. J e n n i f e r Le a v e y Integrated Sciences Curriculum Coordinator, Georgia Institute of Technology

1 9

A R T I C L E SENVIRONMENTAL ENGINEERINGEnvironmental Fate and Transport of Veterinary Antibioticsby J ona than Char l e s Ca l luraAdvised by Dr. Ching-Hua Huang, School of Civil and Environmental Engineering, Georgia Insti tute of Technology

2 5 CIVIL ENGINEERINGRisk Informed Design of Offshore Wind Turbine Structures in the U.S. Outer Continental Shelfby Timothy Wade CookAdvised by Bruce R. Ell ingwood, Ph.D. , P.E. , N.A.E. , Dist . M. ASCE, The Raymond Allen Jones Chair in Civil Engineering, Georgia Insti tute of Technology

3 3 PUBLIC POLICYThe REDD+ Programme:Affects on Governance Theory, Market Theory, and a Post-Kyoto Worldby Tejas KotakAdvised by Dr. Janelle Knox-Hayes, School of Pub-l ic Policy, Georgia Insti tute of Technology

3 9 ECONOMICS: UNIVERSIT Y OF OREGONUnited Way of Lane County’s Promise Neighborhoods and the Benefits of Read-ing Readinessby Jacob McGrew and El izabeth Lohrke , University of OregonAdvised by Dr. Joe A. Stone, W. E. Miner, Professor of Economics, University of Oregon

4 9 BIOMEDICAL ENGINEERINGBOLD Signal Changes in Resting State Networks are Related to Performance on a Vigilance Taskby Mac MerrittAdvised by Dr. Shella Keilholz, School of Biomedi-cal Engineering, Georgia Insti tute of Technology and Emory University


T h e E n v i r o n m e n t a l C o s t o f C a p i t a l

Environmental friendliness seems to be a factor that influences how easily and at what cost firms can borrow money. This is shown be research conducted by Dr. Sudheer Chava from the Scheller College of Business which reveals that firms with ‘bad’ environmental profiles have 18% fewer lenders in loan syndicates and 15% fewer institutional lenders

In his working paper titled “Environmen-tal Externalities and Cost of Capital,” Dr. Chava studies how the cost of capital that firms raise depends upon their environ-mental profile.Investors and lenders prac-tice what is termed as Socially Responsible Investing (SRI) to screen out undesirable characteristics such as nature of business, amount of pollution and climate change concerns.

Cost of capital can be defined as the cost at which a firm can raise money. It is a com-bination of the cost of debt (loans, bonds issue etc.) and cost of equity (return ex-pected by shareholders). This cost depends upon the financial stability of the compa-ny. Large, stable firms can borrow or raise money at a lower cost than smaller firms lacking proven track records. Now, lend-ers are also considering another factor in

the pricing of their loan – its environmen-tal impact.

The amount of money devoted to socially responsible investing has increased 50 times over in the last 20 years. As of 2007, 12.5% of all investments made and loans lent were under SRI guidelines. A majority of banks, representing about 80% of global lending volume have agreed to consider social and environmental factors during project finance.

In addition to SRI, Equator Principles and the United Nations Environment Pro-gramme’s Statement by Banks, lenders have their own business interests which motivate them to be cautious about lending to environmentally harmful businesses. Firms indulging in environmentally harm-ful practices are subject to regulatory and compliance risks as well as lawsuits from injured parties. It is possible that these fac-tors increase their credit risk, thus making capital more expensive for them.

To address the possibility that environ-mental factors are merely a proxy for risk of default, Dr. Chava has done an ex post analysis to show that environmentally un-friendly firms aren’t prone to go bankrupt

any more that their greener counterparts are.

In addition to bank loans, he studies insti-tutional ownership of stocks and here too, the observations indicate that institutional investors are wary of investing in firms that generate hazardous waste or contrib-ute to adverse climate change. The expect-ed returns on stocks with environmental concerns are higher.

Dr. Chava concludes that there is some ev-idence that the observed positive relation between expected stock returns (spread on the bank loans) and a firm’s environmen-tal concerns is partly driven by socially responsible investors (environmentally sensitive lenders) screening out stocks with environmental concerns. The results suggest that exclusionary SRI and envi-ronmentally sensitive lending, through the higher cost of capital channel, have the po-tential to prompt firms to internalize their environmental externalities.

A m u l y a C h i r a l a

NEWS

“Now, lenders are also considering another factor in the pricing of their loan – its environmental impact.”


O c e a n C i r c u l a t i o n , B i o g e o c h e m i s t r y, a n d

C l i m a t e

The ocean is a dynamic and important part of Earth’s climate system. It exchanges a tremendous amount of heat, moisture, and carbon with the atmosphere and trans-ports them around the globe. It is also home to a copious amount of biological activity that can affect and be affected by the global carbon cycle. Understanding these and other mechanisms are major challenges for modern climate science. Dr. Taka Ito’s research group in the School of Earth and Atmospheric Sciences strives to better understand the interplay of circula-tion, biogeochemistry, and ecology in the global ocean and how those factors impact the regional and global climate on times-cales from a few months to millennia.

The Ito group uses both observational da-tasets and modeling to better understand physical and biogeochemical processes and trends. The observational datasets are typically derived from satellite, ship, and float data products, among others. The models span a wide range of complex-ity, from simple pen-and-paper analyti-

cal treatments to sophisticated computer models that run on high-performance computing clusters at Georgia Tech and at NASA. Along these lines, the Ito group works closely with Georgia Tech’s Partner-ship for an Advanced Computing Environ-ment (PACE).

Currently, the group is been focused on two important regions; namely, the Pacific Ocean and the Southern Ocean. Group members Yohei Takano and Taka Ito are working to pin down the factors that de-termine the abundance and variability of dissolved oxygen in the Pacific. Spe-cifically, they apply a variety of statistical techniques to float data in order to deter-mine the relative importance of transport and biological processes for setting oxygen concentration from monthly to decadal timescales. Significant variations in the concentration of oxygen affects marine life and may have implications for ocean eco-systems, fisheries, and the economies that depend on them.

The ocean that surrounds Antarctica (i.e. the Southern Ocean) currently absorbs a significant fraction of the carbon generated by fossil fuel consumption and transports it into the deep ocean. However, obser-vations suggest that the capacity of the Southern Ocean to absorb carbon has de-clined in recent years. In order to improve confidence in projections of the Southern Ocean carbon absorption rate over the next few centuries, we need a detailed under-standing of how Southern Ocean circula-tion, biogeochemistry, and ecology may be affected by a changing climate. Taka Ito, Dan Jones, and Wei-Ching Hsu work on various aspects of this problem, which is currently a topic of intense interdisciplin-ary discussion in the global oceanographic community.

D a n J o n e s , P h D

“Dr. Taka Ito’s research group ... strives to better understand the interplay of circulation, biogeochemistry, and ecology in the global ocean and how those factors impact the regional and global climate on timescales from a few months to millennia.”


A Lo o k a t R o b o t i c s i n H e a l t h c a r e

At the Health Systems Institute, Dr. Char-lie Kemp’s Lab is working towards en-abling robots to interact with individuals. Robots are useful to humans, especially to those requiring personal assistance, as they are time-efficient, and more cost effective than trained animals or nurs-ing assistants. Traditionally, robots were designed to avoid contact and perform preset lower order tasks in a controlled environment, such as a factory setting. It is impractical, however, when work-ing with humans to avoid contact. For humans, if the mobile manipulator is to work at home, or even at the hospital, it has to be able to adapt to its environment.

By integrating custom user interfaces, and parts into commercially available mobile manipulators, the Kemp Lab has come up with novel ways for robots to be more viable in working with humans. For ex-ample, using whole tactile sensing arms along with the principal that low contact forces are benign the robot is able to make contact with objects as long as the force is below a certain threshold.

The robot does not require any prior knowledge of the environment until it makes contact, a very useful quality in manipulating through clutter, such as retrieving a bowl from a cupboard. Cur-

rently, the Kemp Lab is working on whole body contact to allow robots to work in environments such as hospitals, complet-ing nursing tasks such as carrying pa-tients from one room to another.

In conjunction with these adaptive robots, the Kemp Lab has designed custom user interfaces allowing the user to control the robot with head trackers, lasers, and joy-sticks. For physically impaired individu-als that have the mental capabilities to control the robot, but cannot do the task themselves, these control methods are ad-vantageous. Dusty, a small compact mo-bile manipulator, for example, can pick up objects from the floor and deliver it back to the user. The user controls the ro-bot using a joystick interface.

In the Robots for Humanity project, the Kemp Lab, and Willow Garage work with Hennery Evans, a mute quadriple-gic. Hennery Evans uses the PRII robot, a commercial available robot customized to fit Mr. Evans needs. He controls the robot with a head tracker in order to do daily tasks such as shaving or even scratching an itch. The particular tasks may seem trivial, but for a disabled individual, do-ing these tasks gives them back their sense of independence. The Kemp Lab is working to making these interfaces open

sourced so anyone with a robot can down-load these behaviors and utilize them.

Through exploring methods to make ro-bots more adaptable, easy to control, and make contact with humans, the Kemp Lab is exploring new innovative frontiers in the field of robotics. Thus, they are open-ing new avenues for future assisted living strategies in busy dynamic settings such as in hospitals. Someday these robots could be able to work alongside doctors and nurses as nursing assistants.

J a h e d a K h a n a m

NEWS

“In the Robots for Humanity project, the Kemp Lab, and Willow Garage work with Hennery Evans, a mute quadriplegic.”


NEWS IN REVIEW

A d v a n c e s i n f M R I a n d E l e c t r o p h y s i o l o g y

The human brain is one of the most com-plicated collections of matter in the known universe consisting of over 100 million neurons that use electrical and chemical signals to communicate in such a way that yields consciousness. For over 100 years, researchers have attempted to understand the role of specific brain regions and the re-lationship between each brain region. In the past scientists have relied on case studies of injured patients or electrical stimulation to segregate the brain based on function. Recently new techniques have been devel-oped that can accurately test the influence of multiple brain regions. Two important techniques in this field are electrophysiol-ogy and Functional Magnetic Resonance Imaging (fMRI). Electrophysiology uses implanted electrodes to directly measure electrical activity in a specific region. This is valuable because electrical signaling is the primary means by which information travels through the brain, but electrophysi-ology lacks spatial resolution since activity is measured only at one specific implanta-tion sight. fMRI uses repeated MRI scans to measure changes in a Blood Oxygen-ation Level Dependent (BOLD) signal over

time. This type of imaging is noninvasive and provides incredible spatial resolution but specific information about the relation-ship between the BOLD signal and the un-derlying electrical activity is still unknown. Recording these two measures simultane-ously is essential for understanding the relationship between fMRI and neuronal activity; however, each technique is highly sensitive, making them very difficult to combine. The Magnetic Resonance Imag-ing and Neural Dynamics (MIND) lab at Georgia Tech and Emory under Principle investigator Shella Keilholz has developed a method for simultaneously recording these two techniques. Using an advanced recording technique combined with novel analysis methods, this group is acquir-ing valuable insight about what the fMRI BOLD signal actually means which will lead to advances the scientific understand-ing of resting state brain activity.

A methods video has been produced and published in the Journal of Visual Experi-ments on this simultaneous technique. Many complications had to be overcome in order to combine these techniques. For

instance, toothpaste had to cover the ex-posed head so as to prevent distortion in the MRI images. Parts of the electrophysi-ology time course had to be cropped out because of the interference from the radio frequency pulse produced by the scanner. Additionally, special glass electrodes were used to acquire low frequencies which pre-vious studies have ignored.

After one publication, the Keilholz group is beginning to explore the depth of this data by investigating low frequency elec-trophysiology for the first time, and ana-lyzing the dynamic behavior of electrical and BOLD signal. Using the results of this study, Georgia Tech is making great strides towards understanding basic brain function. Despite the progress of decades of neuroscience and biomedical research, very little is known about the fundamental relationship between the brain and con-sciousness. This group is using advanced acquisition and analysis techniques so that the scientific community may begin to understand the complex processes of the brain.

M a c M e r r i t t

“Recently new techniques have been developed that can accurately test the influence of multiple brain regions.”


D e v e l o p i n g E x p e r i m e n t a l M o d e l s f o r M e a s u r i n g N e u r o m u s c u l a r I m p r o v e m e n t Fo l l o w i n g R e h a b i l i t a t i o n

Following severe neurological diseases such as a stroke and Parkinson’s disease, human beings often suffer impairments in their balance and motor control systems. Many rehabilitation strategies have been developed to reduce these deficits, how-ever the clinical measures used to evaluate the changes after rehabilitation are very broad and fail to identify how the changes are coming about. This also makes it im-possible for clinical practitioners to explain why certain strategies work for some indi-viduals and not for others. To tackle this problem, the Ting Lab at the Georgia In-stitute of Technology has developed sev-eral experimental models which measure the neuromechanical responses of healthy young adults to postural perturbations such as movement of the floor while they are standing. Although Dr. Ting and her team have several models, the underly-ing theme is to study the nervous system’s anticipatory and reactive responses to dif-

ferent biomechanical conditions. The key biomechanical conditions that Dr. Ting has chosen to study are standing balance and situations of perturbed balance. In addi-tion to balance being impaired in patients with neurological diseases, she has chosen to study this condition because it is a situa-tion in which the biomechanics is well un-derstood.

To collect movement data for their experi-ments the Ting lab uses a lab equipped with a moving platform (with force mea-surement plates underneath), motion capture cameras, reflective markers and an electromyography system. With this equipment Dr. Ting and her team can si-multaneously collect movement, force and muscle activation data which is then relat-ed through their various models. Although the Ting team is currently working on pre-dictive models by testing the responses of healthy young adults, the team has also

received several grants to study the ef-fect of rehabilitation strategies on patients who have suffered spinal cord injuries and Parkinson’s disease; the team has also published a paper on the prediction of mo-tor control following a stroke. So far there have been promising results in identifying disease areas and tracking re-development of the neuromuscular system when strate-gies such as the adapted tango therapy for patient with Parkinson’s disease are used. These advances in characterizing healthy and disease state neuromechanical re-sponses is very promising for the creating many possibilities for rehabilitation and therapy. One of those great possibilities is the development of autonomous robots which provide rehabilitation strategies based on a healthy model. Although the team is a long way from completing their work, this is one of the many great possi-bilities that the work done in the Ting lab offers.

S i m i s o l a O. O l u d a r e

NEWS

“... the underlying theme is to study the nervous system’s anticipatory and reactive responses to different biomechanical condi-tions.”


A Wa t e r S o l u b l e C o p p e r ( I ) S e l e c t i v e F l u o r e s c e n t Pr o b e :

N e w A p p l i c a t i o n i n D i r e c t I n - G e l D e t e c t i o n o f a C o p p e r C h a p e r o n e

Copper is an essential trace element and required for various biological processes such as cellular respiration, connective tis-sue cross-linking, pigment formation, and antioxidant defense, but free copper is toxic because it can participate in generating free radicals [1]. Hence, copper is tightly regu-lated in living cells and disruptions in cop-per homeostasis are implicated in numer-ous diseases, including Menkes’, Wilson’s, and Alzheimer’s disease [2]. Several copper binding proteins are integral parts of cop-per homeostatic machinery and work in conjunction to buffer the cellular copper pool [3]. Copper chaperone proteins deliver copper to subcellular locations to metallate

various copper proteins and thus these pro-teins are an important component of copper homeostasis [4]. While numerous research efforts have been made to understand the mechanisms of copper uptake, distribution, and regulation in the body, many questions still remain unanswered regarding molecu-lar mechanisms of copper trafficking and the subcellular locations or components in-volved in transient storage of copper.

Cation selective fluorescent probes have become important analytical tools for the detection of biologically important metal cations. Designing effective fluorescent probes for transition metals with variable oxidation states, such as Cu(I), is challeng-

ing due to the presence of metal-induced fluorescence quenching pathways. More-over, many fluorescent probes used in bio-logical research are lipophilic in order to facilitate lipid bilayer permeation, but high lipophilicity often leads to aggregation in aqueous solution. This aggregation can dra-matically alter the photophysical properties of the fluorophores and likely to produce artifacts when applied to biological sys-tems. Copper(I) selective fluorescent probes are prone to be especially lipophilic due to their thioether based cation binding sites. This design strategy is employed to achieve selective binding of a soft ligand like sulfur towards the soft Cu(I) cation.

To address these problems, Christoph J. Fah-rni and his lab developed a water-soluble fluorescent probe CTAP-2 that gives 65-fold fluorescence turn-on response selectively to Cu(I) in aqueous solution which was pub-lished in the Journal of the American Chem-ical Society [5]. This turn-on response of the probe in presence of Cu(I) was achieved by taking advantage of photoinduced electron transfer (PET) as a fluorescence quenching pathway. PET based fluorescent probes consist of a fluorophore and an electron do-nor unit which contains the cation binding site electronically decoupled from the fluo-rophore. In the absence of copper(I), fluo-rescence is quenched by an electron transfer from the donor unit to the excited fluoro-

phore followed by charge recombination to return to the ground state by a non-radia-tive pathway. Upon copper(I) binding, PET becomes thermodynamically less favorable and cannot compete with fluorescence emission, and therefore, the fluorescence intensity increases. Although CTAP-2 was cell permeable despite being water soluble, the featured biological application of this probe involved development of a novel pro-teomics method. Metal ions bound to metal-loproteins stay intact during separation of the proteins by non-denaturing polyacryl-amide gel electrophoresis (PAGE). When applied directly to a non-denaturing PAGE gel, CTAP-2 selectively gave fluorescence response to copper-bound Atox1, a human copper chaperone, but not to apo-Atox1, or proteins with other metal ions as cofactors such as zinc binding protein carbonic anhy-drase. Moreover, this method was selective towards solvent accessible metal ion as the probe did not respond to Cu/Zn superox-ide dismutase, a copper protein with steri-cally inaccessible copper site. These studies demonstrate the potential of cation selective fluorescent probes as convenient and inex-pensive tools for the in-gel detection of met-al-binding proteins with accessible metal binding sites.

Pr i t h a B a g c h i

“... disruptions in copper homeostasis are implicated in numerous diseases”


M i c h a e l C h e nCambridge Trust and NIH/Cambridge Fellow, Former Tower Editor

INTERVIEWS

SPOTLIGHT:

D r. K a t h r y n M e e h a n

Michael Chen first visited the Fellowships Office during his freshman year at Georgia Tech. He discussed future fellowship ap-plications and began considering the ones for which he might apply. By the begin-ning of his sophomore year he knew he wanted to apply for a Goldwater Scholar-ship, which is open to sophomores and ju-niors in STEM areas, and began working that fall with Dr. Karen Adams in the Fel-lowships Office.

The first hurdle for a Goldwater applicant is obtaining Georgia Tech nomination since the university can nominate only four stu-dents each year for national consideration. Michael was ready for the early January deadline because he worked on his essays during fall term. The campus Goldwater Committee nominated him, and in late March of his sophomore year he learned that he was a national Goldwater Scholar. Because he won the scholarship his sopho-more


year, he held the $7500 award both his ju-nior and senior years.

Michael’s undergraduate research with Dr. Nicholas Hud in the Georgia Tech Center for Chemical Evolution focused on the structure, function, synthesis and evo-lution of nucleic acids and resulted in an article they published in The Proceedings of the National Academy of Sciences, USA.

During his junior and senior years Mi-chael was Editor of The Tower. During his tenure he increased the frequency of publication, diversified content to reflect the cutting-edge research by undergradu-ates at Georgia Tech, and promoted the importance of undergraduate research to the administration and to students. He in-creased the number of students involved in peer review and increased appropriate advertising for the journal. The high qual-ity of The Tower owes much to Michael’s dedication and attention to detail.

In the spring of Michael’s junior year he visited the Fellowships Office several times to discuss which fellowships to pur-sue his senior year. He knew he wanted to attend Cambridge University and work with Prof. Shankar Balasubramanian to continue his work on nucleic acids. By changing the structures within these acids he wanted to work to create treatments for illnesses such as Huntington’s disease, and the lab at Cambridge was where he needed to be.

Michael applied for several fellowships and was interviewed for a Gates award. He was disappointed when he did not receive this fellowship to attend Cambridge but found success when he was awarded both a Cambridge Trust Fellowship and an NIH/Cambridge Fellowship to do his PhD. He is now in his first year at Cambridge and lives at Churchill College on the outskirts of Cambridge. Churchill houses about 30% Americans and is a modern structure

compared with many of the colleges in Cambridge. There is a great gym, and Mi-chael cycles to his lab and says a bike is the only way to travel there. Churchill has for-mal halls, and sometimes a Nobel Laureate shows up for dinner.

The topic of Michael’s PhD is the chemi-cal biology of nucleic acid structures. Most people are familiar with the double helix, but nucleic acids such as DNA and RNA can take up many different shapes and forms. Scientists showed that nucleic ac-ids can also form triple helices and qua-druple helices known as a G-quadruplex. He is looking at how G-quadruplexes act as switches and recognition motifs in the cell by studying a new type of enzyme that recognizes and binds quadruplexes.

Professor Balasubramanian expects gradu-ate students to be first author on at least two publications and to publish in the sci-entific world during PhD work. Michael went from his undergraduate degree to be-ing a CPGS and is seeking a Certificate of Postgraduate Study. When he passes his exams at the end of the first year, he will formally be a PhD student. He is now in a lab with about 30 other students and finds there is great mentoring at Cambridge. Fel-low graduate students help newbies navi-gate administrative requirements and tend to required training classes.

Doing research at Cambridge is different from doing research in the U.S. He finds the British system encourages people to balance work and life. At Cambridge peo-ple do not work in labs over the weekend, and students are encouraged to travel and take advantage of what Cambridge and its environs offer. It is possible to travel by both train and plane cheaply, and Michael is already taking advantage of these possi-bilities by seeing Stonehenge, London mu-seums, the National Gallery, and the Brit-ish Museum (including the Rosetta Stone and mummies such as Ramses). He took a

trip from Ipswich to the North Sea with the Cambridge Yacht Club and enjoyed sailing on the open sea for the first time.

A difference between the US and British PhD systems is that in the UK the degree is to be completed in a maximum of four years, and students may be asked to leave if they do not complete their research in that time. The big weeding out process occurs during the first year and through the exam at the end of that time. If students make it through the first year, they will likely com-plete the PhD in the required time.

Looking back at the applications he pur-sued, Michael said that one of the big things he learned was to apply early and not wait until the last minute. He talked with Fellowship advisors well in advance and began writing essays early. A success-ful fellowship application takes planning.

One of the things Michael has already learned at Cambridge is that a PhD is what a person makes it. The research discipline he learned at Georgia Tech will be used to make the most of all that Cambridge offers.


INTERVIEWS


Dr. Steve Cross, who initially began his ca-reer at Georgia Tech in September of 2003as the Director of GTRI and now serves as the Executive Vice President for Research,was originally inspired to study engineer-ing when John F. Kennedy announced his“grand challenge” to send a man to the moon by the end of the decade. Shortly af-ter Cross graduated from high school, Neil Armstrong landed on the moon, which further cemented Cross’ passion for engi-neering and achieving the impossible. Ac-cording to Cross, “with the right inspira-tion and leadership, groups of people can come together and innovate and make the impossible a reality.”

After discovering his passion for engineer-ing, Dr. Cross began his studies at the Uni-versity of Cincinnati in the school of Aero-space Engineering. However, because of a slump in the field of Aerospace the late six-ties, he decided to make the switch to Elec-trical and Computer Engineering. While at the University of Cincinnati, he participat-ed in undergraduate research beginning

in his third year of studies. When faced with the decision between space applica-tions and the field of medicine, he chose to research with a professor who worked on early applications of biomedical engineer-ing. This research experience allowed him to work on ways to cure nearsightedness, which involved automated control model-ing for ocular applications. Although he did not go on to become a medical doctor, Cross cites his experience in undergradu-ate research as one of the many reasons he became passionate about the research in-dustry and is in the position he is in today.

After completing his undergraduate de-gree, Dr. Cross’ career was interrupted by the Vietnam War, which allowed him to join the Air Force and work as an engi-neer before he was offered an opportunity pursue a Ph.D in Electrical Engineering at the University of Illinois at Urbana-Cham-paign. Because of his interest in the combi-nation of computer systems and psycholo-gy, Dr. Cross focused his Ph.D on artificial intelligence, which was at the time called a

“knowledge based system.” He specifical-ly applied these computer programs that could mimic the knowledge and expertise of a human expert to air traffic control sys-tems.

In his current position at Georgia Tech, Dr. Cross believes that undergraduate re-search plays a critical role in the education-al value that Georgia Tech can contribute to the nation and the world. Undergraduate research is valuable for the corporations that partner with Georgia Tech because it allows them to outsource to students the exploration of disruptive ideas that would otherwise pose a financial risk to the com-pany. When asked to give a piece of advice to those looking to participate in under-graduate research, Dr. Cross says simply, “jump in and go for it!”

For more details on the important role stu-dent play in the evolving research strategyfor Georgia Tech, please visit: http://ti-nyurl.com/ak9d2v6

Executive Vice Presidentfor Research

Dr. Stephen E. Cross

Ty l e r J . K a p l a n

FEATURE:


INTERVIEWS

A native Georgian from Decatur, Dr. Jen-nifer Leavey graduated from Georgia Tech with a degree in Chemistry and a certifi-cate in Biochemistry before moving on to Emory, where she gained her Ph.D. in Im-munology and Molecular Pathogenesis. She has done a research fellowship at the Center for Tropical and Emerging Global Diseases at UGA and Post-Doc. in the di-vision of endocrinology at Emory. Dr. Leavey worked heavily in the School of Biology before moving to her current posi-tion as the Integrated Sciences Curriculum Coordinator for the College of Sciences.

Dr. Leavey’s current position was created as part of the Clough Commons. Because every introductory science lab is located in the same building, the faculty wanted to capitalize on this space to start an interdis-ciplinary science community, beginning at the freshman level. Dr. Leavey works closely with the introductory lab science coordinators to move toward incorporat-ing more interdisciplinary content into the labs. She also works with the teaching faculty in the college of sciences, using a

newsletter and website (ise.gatech.edu) to try and share new teaching methods among the faculty.

Dr. Leavey’s interests extend past her work at Tech. She currently plays guitar in an Atlanta-based band, Catfight! Before join-ing the band, she didn’t know any of the other members She first heard of the band by answering an ad. Catfight! still plays occasional reunion shows, but primarily played throughout the musicians’ times in graduate school . They started in 1995, and the band regularly played though 2005. These years were filed with touring and they even had CDs released. Her favorite guitar is a Gibson L6S. Dr. Leavey told a story of how she used to have a Fender Telecaster Custom, but it got stolen years ago. The band even has a few songs in iTunes and her favorite one from the al-bum, Stomp! Shout! Scream!, is “Back off My Baby.”

One of Dr. Leavey’s most recent contribu-tions to Georgia Tech is Charged Maga-zine. It began just over a year ago, with

the goal of bringing the creativity from outside of her work life into her work life: building websites, playing music, and do-ing other artistic endeavors! She wanted to create a space that was lighthearted, fun, and exciting--but also maintained a focus on science. Dr. Leavey stated, “It is the kind of website that I hope will appeal to young people who haven’t decided what they want to do with their lives yet. I also want to engage students who are inter-ested the field of science, who are not sci-ence majors and won’t have to take many science classes. We even want to engage adults who like science and are interested in it, but don’t necessarily want to go pick up a research magazine. They want to have what they read be engaging.”

There are a wide variety of articles pub-lished by Charged Magazine. Some writ-ers pick a certain difficult topic from a class they enjoy in school, and write about it in a way that is fun and quirky, but engaging. For example, one author wrote an article about her vertical caving trip. This student was taking physics at the time, and she

Dr. Jennifer LeaveyIntegrated Sciences Curriculum Coordinator

M i c h a e l B o n i f a c i o

SPOTLIGHT:


researched which physics principles ap-plied to vertical caving. Charged Maga-zine features interviews, and they hope to continue getting more as time goes on. A recent, published article features an inter-view with a Berkeley graduate who started a company to use recycled coffee grounds to grow mushrooms. They have now made an aquaponic system with a miniature fish tank to fertilize an herb garden.

The primary writers for Charged Maga-

zine are undergraduate and graduate students from Georgia Tech. Charged Magazine recently released an article from a curator at Zoo Atlanta, and Dr. Leavey stated that she would like to get more ar-ticles like this from off campus. Charged Magazine currently uses local editors, but they are poised to move to digital editing to make their reach far beyond Tech. For now, Charged Magazine plans to remain only a website with a rolling publication,

but at some point, they may consider do-ing a print version.

To see more about Charged Magazine, vis-it http://chargedmagazine.org/.


ARTICLES


E n v i r o n m e n t a l Fa t e a n d Tr a n s p o r t o f Ve t e r i n a r y A n t i b i o t i c s

J o n a t h a n C h a r l e s C a l l u r a

1. INTRODUCTIONThe implications of trace contaminants such as pharmaceutical products in environmental systems has been a cause for increas-ing concern in recent decades, stemming from detection of select compounds in natural systems. Veterinary antibiotics, tradition-ally applied as feed additives, are utilized to prevent disease and promote growth among livestock and poultry. One study sug-gests that this application accounts for 24.6 million pounds, or approximately 70% of total antimicrobial use in the United States (Mellon, Benbrook, & Benbrook, 2001). As the leading producer in the nation, the state of Georgia generated over 1.3 million broil-ers, the industry term used for chickens grown for the purpose of meat production, in 2010 (National Agricultural Statistics, 2011).

Bedding materials such as wood shavings from broiler operations are routinely mixed with feathers, wasted feed, and broiler excre-

ment to produce poultry litter that is then used as a grassland fertilizer due to its nitrogen, phosphorus, and potassium content (Nichols, Daniel, & Edwards, 1994). Since up to 90% of the dosage passes directly through the animals’ digestive systems and be-comes incorporated into the litter, the antibiotics may potentially be leached into sediments and soil, as well as nearby surface and groundwater sources. A recent study of several large-scale swine and poultry feeding facilities noted that antimicrobial compounds were detected in 67% of surface and groundwater sources within close proximity to the farms (Campagnolo, et al., 2002). An il-lustration of the mobilization pathways may be seen in figure 1. One primary concern of the presence of such microcontami-nants is the development of antibiotic-resistant bacteria which would compromise the efficacy of current treatment methods. Furthermore, much remains unknown about the toxicity associ-ated with chronic low-level exposure to anthropogenic stressors.

ABSTRACTVeterinary antibiotics such as ionophores and tetracyclines are commonly used in farming operations to prevent disease and promote growth rates in poultry, cattle, and swine. Since waste products from treated animals are used as fertilizers, there is a growing concern that these compounds may leach into the soil and water supply causing chronic low-level exposure to humans and leading to the development of an-tibiotic-resistant bacteria. Due to close structural similarities and physical properties, nigericin was determined to be an effective surrogate standard for the detection of the monensin, salinomycin, and narasin. Aluminum sulfate addition was considered as a potential treatment method to reduce antibiotic mobility, which resulted in an approximately 80% reduction of recovery rates for tetracyclines in poultry litter samples. Strong anion exchange cartridges were used in tandem with standard hydrophilic-lipophilic balanced cartridges which resulted in more rapid extractions and faster sample processing. Demeclocycline exhibited the potential to be used as a surrogate standard for tetracy-clines due to close similarities in detection levels. This study has developed a method for detecting antibiotics in several types of environmen-tal media and reinforced potential treatment methods as a means to reduce the risk of exposure to microcontaminants in the water supply.

M i c h a e l J a m e s21 | The TOWER | SPRING 2013

Ionophores (IP) and tetracyclines (TC) are particularly relevant due to the frequency of their usage as anticoccidials in the poul-try and livestock industries. The IP under review are monensin (MON), salinomycin (SAL), narasin (NAR), and nigericin (NIG), while observed TC include oxytetracycline (OTC), chlortetra-cycline (CTC), tetracycline (TTC), and demeclocycline (DMC).

Surrogate standards are compounds that are not expected to be present in environmental samples but have similar structure and properties to the target analytes and may indicate any unusual matrix effects. The first goal of this study is to further develop experimental methods and to validate the selection of NIG as a surrogate standard for IP. Aluminum sulfate (alum) may be applied to poultry litter prior to soil fertilization to decrease phosphorus runoff and ammonia volatilization (Moore, Daniel, & Edwards, 2000). Antibiotics are expected to exhibit reduced mobility when subjected to alum treatment due to decreased sample pH and potential Al-antibiotic complexation reactions. While recent studies have primarily employed the use of hy-drophilic-lipophilic balanced (HLB) solid phase extraction (SPE) cartridges in their analyses, some literature suggests a strong anion exchange (SAX) and HLB tandem cartridge setup (Han-sen, Bjorklund, Krogh, & Halling-Sorensen, 2009); (Blackwell, Lutzhoft, Ma, Halling-Sorensen, Boxall, & Kay, 2004). Another purpose of this project is to attempt to improve TC detection through the use of the SAX-HLB configuration and pH alteration.

2. METHODS

2-A. Rainfall Runoff ExtractionThe first of the two primary analysis methods involves anti-biotic detection from a liquid matrix. An experimental field at the University of Georgia (Athens, GA) was isolated and treated with poultry litter fertilizer. The area was covered in order to prevent mobilization from unexpected rains and sub-jected to a rainfall simulation from a spray tower and the sur-

face runoff was collected in amber bottles for storage. The samples were then sent to the Georgia Institute of Technol-ogy (Atlanta, GA) for analysis and kept under refrigeration.

Liquid samples were measured to 200 mL volumes and spiked with antibiotics prior to being passed through a vacuum-powered 0.45 um glass fiber filter to remove excess suspended solids which may interfere with analysis. The samples were then adjusted to a pH of 6-7 spiked with IP and TC to a concentration of 1000 PPB. Disodium ethylenediamine tetraacetate (Na2EDTA) was added to a concentration of 1 mM as a buffer and the samples were passed through HLB extraction cartridges under vacuum pres-sure. Once fully passed, the cartridges were washed with 5 mL of deionized water and dried to prepare them for elution. Two 3 mL quantities of methanol (MeOH) were slowly passed through the cartridges after a soaking period of 5 minutes and the con-tents were collected in glass vials beneath the vacuum manifold. 2-B. Soil and Poultry Litter ExtractionFor solid phase experiments, samples were weighed out to 1 g and spiked directly with antibiotics before being spread on plastic wrap to dry for 2-3 hours. Samples from the alum addition experi-ments were also spiked with 0.2 g of finely ground aluminum sul-fate powder. Once dry, the soil or poultry litter was transferred to centrifuge tubes and mixed with an extraction solution whose con-tents are shown in tables 1, 2, or 3 depending on the method used.

Mcllvain’s Buffer at pH 7 with 10 mM Na2EDTA

(mL)

MeOH (mL) Total Vol-ume (mL)

10 10 20

0.1M Citric Acid (ml)

0.2 M Na2H-

PO4 (ml)

10mmol Na2EDTA solution

(ml)

DI water (ml)

MeOH (ml)

Total Volume (ml)

7.95 2.05 2 3 5 20

0.1M Citric Acid (ml)

0.2 M Na2H-

PO4 (ml)

10mmol Na2EDTA solution

(ml)

DI water (ml)

MeOH (ml)

Total Volume (ml)

3.68 6.32 2 3 5 20

The tubes were shaken for 30 minutes at 400 RPM on an agita-

Figure 1. Depiction of antibiotic mobilization pathways.

Table 1. Standard Extraction Solution

Table 2. Method 1 Extraction Solution (pH = 3)

Table 3. Method 2 Extraction Solution (pH = 6)


tion table then placed on a centrifuge machine for 10 minutes at 5000 RPM. The 20 mL of supernatant fluid from the centri-fuge process was separated and diluted with 460 mL of water before another batch of extraction solution was added to the centrifuge tube and shaken once more. The supernatant flu-id was separated again, bringing the final diluted volume to 500 mL. Next, the SPE process was conducted as outlined in the water sample section, with the aforementioned SAX-HLB tandem cartridges used in the alum addition experiments.

2-C. Clean-Up StepsFluids from the SPE process were blown to dryness under vacu-um then mixed with 5 mL ethyl acetate and 0.5 mL 0.1 M NaCl to separate analytes from impurities. The resulting dirty water layer was pipetted out of the vial and the samples underwent vacuum volume reduction once more. The vials were then injected with equal parts MeOH and 0.1 M NaCl as well as 25 uL of simatone, an internal standard, and taken to be analyzed on liquid chroma-tography-mass spectrometry (LC-MS) equipment which utilizes electrospray ionization (ESI) and selected ion monitoring (SIM) to effectively determine sample composition. Figures 2 and 3 illus-trate overviews for the extraction procedures for liquid and solid matrices, while figure 4 shows the clean-up procedures in detail.

3. RESULTS

3-A. Surrogate Standard ValidationTo validate the applicability of NIG as a surrogate standard for the selected compounds, a series of spiked samples were pro-cessed from all relevant matrices and the peak areas were plotted with their corresponding concentrations.

Figure 5 shows peak area readings from the LC-MS software. The unitless value is related to the solution concentration. The spiked concentrations are based off a calibration curve obtained from samples of known concentrations. This data was then normalized using NIG as the denominator for both values, producing the in-formation shown in figure 6. Figure 7 shows normalized data for SAL soil samples and NAR spiked rainfall runoff samples.

3-B. Aluminum Sulfate Addition

Figure 2. Water sample extraction process.

Figure 3. Soil and poultry litter sample extraction process.

Figure 4. Sequential clean-up process.

Figure 5. MON peak area vs. concentration for various media.


Calibration curves for each compound were created by prepar-ing standard solutions of 20, 40, 200, 500, 1000, and 2000 PPB as shown in figure 8. A linear regression was performed for the curves and the resulting second-order polynomial equations were used calculate concentrations for the poultry litter samples. Three different methods were used in preparing the samples: method 1, method 2, and alum addition. Method 1 followed the standard solid sample extraction steps using the ‘Method 1’ ex-traction solution, while method 2 utilized the other correspond-ing extraction solution. The supernatant fluid from method 2 was also injected with hydrochloric acid until a pH of 3 was reached. The ‘Alum PL’ sample followed standard solid ex-traction methods except 0.2 g of aluminum sulfate was added to the litter prior to spiking with antibiotics. Based on the cal-culated concentrations, recovery rates were determined accord-ing to the following equation with the results shown in figure 9:

Recovery (%) = Cs- Cx

S

4. DISCUSSION

4-A. Surrogate Standard ValidationThe relationships developed by the data in figures 5-7 show a correlation between the amount of NIG and the amount of other IP detected in a sample. The differences between figures 5 and 6 illustrate the benefits of NIG, as evident by the convergence of all samples around the standard line which accounts for variance based on the matrix type. Based on this connection, we may use NIG as an indicator of data quality since extreme deviations from the observed results may suggest errors or interference. Nigeri-cin was determined to be a suitable surrogate standard for the

Where CS = concentration in spiked samples, CX = concentration in unspiked samples, & S = spiking concentration

Figure 9. TC recovery rates for alum poultry litter and regular spiked poultry litters.

Figure 6. MON normalized area vs. normalized concentration.

Figure 7. SAL and NAR normalized area vs. normalized concentration.

Figure 8. TC calibration curve.


selected ionophores which supports the proposed hypothesis.

4-B. Aluminum Sulfate AdditionThe observed recovery rates for tetracyclines in this study (fig-ure 9) followed hypothesized behavior, resulting in a significant mobility reduction of samples with aluminum sulfate. DMC ex-hibited the potential to be used as a surrogate standard for TC detection based on the similarities in recovery rates with the other TC species. Methods 1 and 2, differing only in timing of pH reduction, each reported recovery rates of approximately 25%, compared to less than 5% for alum samples. These values account for the negligible existing levels of TC in blank samples and are likely below 50% due to the particularly prevalent matrix interference and signal suppression in soil and poultry litter. It is important to note that without any sort of pH alteration, TC were not detected in solid matrices so all samples were acidified. Method 1 was expected to yield slightly higher recovery rates due to acidification of the matrix which creates a stable environ-ment for tetracyclines, but only marginal improvements were noted when compared to method 2. Method 2 samples were extracted with a neutral pH extraction solution and later the supernatant fluids from the centrifuge process were acidified. Although the SAX-HLB cartridge setup marginally increased detection when compared to earlier experiments, the great-est benefit appeared to be improvements in extraction speed. Since the SAX cartridges capture humic acid and natural organic matter which is present in both soil and poultry litter, they pre-vented clogging and shortened the time required for processing.

5. CONCLUSIONSThis study develops a method for determining the fate and mo-bility of antibiotics in several compartments of the environment. By validating the selection of nigericin as a surrogate standard for ionophores, more effective analyses may be conducted in the future. Analysis of tetracyclines in various environments high-lighted the need for pH reduction to aid in detection. Perhaps most importantly, the restriction of antibiotic mobility, which oc-curred as a result of aluminum sulfate addition to soil and litter, provides a potential means for pollution prevention as it is al-ready a common practice in preventing nutrient runoff. By mix-ing alum treatment with litter fertilizers, farming operations may deter the transport of microcontaminants into surface runoff and groundwater sources, effectively limiting exposure levels to the public water supply. This work provides groundwork for further experimental development by future researchers in the rapidly developing field of environmental antibiotic contaminant study.

Future WorkIn order to fully understand the transport of ionophores and tetracyclines, further studies must be conducted. The effects of each matrix on analyte detection and the corresponding sig-nal suppression must continue to be evaluated. While NIG was thoroughly validated as a surrogate for MON, similar ex-periments must be conducted for the remaining matrices asso-ciated with SAL and NAR. Finally, the effects of stacking and storage methods on compound detection must be understood.

6. REFERENCESBlackwell, P. A., Lutzhoft, H.-C. H., Ma, H.-P., Halling-Sorensen, B., Boxall, A. B., & Kay, P. (2004). Ultrasonic extraction of veterinary antibiotics from soils and pig slurry with SPE clean-up and LC-UV and fluorescence detection. Talanta, 1058-1064.

Campagnolo, E. R., Johnson, K. R., Karpathi, A., Rubim, C. S., Kolpin, D. W., Meyer, M. T., et al. (2002, November). Antimicrobial residues in animal waste and water resources proximal to large-scale swine and poultry feeding operations. Science of The Total Environment, 299(1-3), 89-95.

Hansen, M., Bjorklund, E., Krogh, K. A., & Halling-Sorensen, B. (2009). Analytical strategies for assessing ionophores in the environment. Trends in Analytical Chemistry, 28(5), 521-533.

Mellon, M., Benbrook, C., & Benbrook, K. L. (2001). Hogging It: Esti-mates of Antimicrobial Abuse in Livestock. Cambridge, MA: Union of Concerned Scientists.

Moore, P. A., Daniel, T. C., & Edwards, D. R. (2000). Reducing Phos-phorus Runoff and Inhibiting Ammonia Loss from Poultry Manure with Aluminum Sulfate. Journal of Environmental Quality, 29(1), 37-49.

National Agricultural Statistics. (2011). Poultry - Production and Value : 2010 Summary. Washington, DC: United States Department of Agricul-ture.

Nichols, D. J., Daniel, T. C., & Edwards, D. R. (1994). Nutrient Runoff from Pasture after Incorporation of Poultry Litter or Inorganic Fertil-izer. Soil Science Society of America.


ARTICLES


R i s k I n f o r m e d D e s i g n o f O f f s h o r e Wi n d Tu r b i n e S t r u c t u r e s i n t h e U. S . O u t e r C o n t i n e n t a l S h e l f

T i m o t h y Wa d e C o o k

INTRODUCTIONIn order to meet future energy demands, the United States (U.S.) will have to not only test its ability to tap current known exhaust-ible energy sources but also expand and transition its energy portfolio to renewable energies.

Although the U.S. does not have any installed wind farms off-shore to date, there are a number of proposed projects in vari-ous stages of approval, the most notable being the Cape Wind Project, a proposed 130-turbine offshore wind farm planned for Nantucket Sound off the coast of Massachusetts (Transportation Research Board (TRB), 2011). Developers of the project have been seeking approval for 10 years; in 2011, Cape Wind became the first offshore wind farm project approved in the U.S. by both state and federal governments. This landmark approval reflects

both the determination of the project developers as well as re-cent legislative action favoring offshore wind development. To invigorate the development of renewable energy portfolios, the U.S. Department of Energy (DOE) recently published an ambi-tious initiative which aims for 20% of U.S. energy to be supplied by wind power (54 gigawatts from offshore) by 2030, with an in-terim goal of 10 GW offshore by 2020. In order to ensure success-ful deployment of the offshore wind industry the DOE identified two critical objectives: reducing the cost of energy and reducing the time to deployment (DOE, 2010).

There are many reasons why offshore wind farms are attractive when compared with onshore farms despite their higher initial capital cost, including favorable wind climatology, ability to upscale (increase utilization) and proximity to power demand.

ABSTRACTThe United States has enormous potential offshore wind energy resources in the Atlantic, Pacific, Great Lakes and the Gulf of Mexico. However, progress in developing these resources has lagged behind that in Western Europe and no offshore wind farms have been built to date in the U.S. Con-tinental Shelf. Uncertainties in U.S. siting and design criteria, specific regulations and standards, along with a lack of experience have challenged development by increasing both cost and the time to deployment. The reliability of offshore wind turbine farms is critical to industry success and should be secured efficiently with respect to cost. The ability to employ probabilistic risk management and decision theory in the design process of support structures would afford more transparent system reliabilities and more flexibility in design compared with prescriptive design standards. A general framework for risk informed design of offshore wind turbine structures is demonstrated on a typical monopole support structure. The struc-tural parameters are manipulated to adjust the risk and to achieve the desired wind turbine performance at acceptable cost. In order to implement such a design procedure in practice, regulations must stipulate clear performance requirements in terms of system reliability for project approval.

T i m o t h y Wa d e C o o k27 | The TOWER | SPRING 2013

Offshore, the wind is more consistent and the wind velocity is greater at lower elevations. Construction on the water allows larger structures that are not challenged by some onshore con-straints (i.e. highway transportation). Moreover, the 28 Coastal and Great Lakes states account for 78% of the national energy demand. Also, the major cities located in these states generally pay higher costs of energy, further conducing the utility and cost competitiveness of strategically sited offshore wind farms (DOE, 2010).

The outer continental shelf (OCS) of the U.S. has the potential to provide 2,957 GW of gross (neglecting siting constraints) avail-able offshore wind energy within 50 miles of shore, which trans-lates to approximately 3 times the current capacity of the national grid (TRB, 2011). The U.S. is well positioned with the resources and the means currently available to become a leader in offshore wind technology and utilization.

The offshore wind industry is challenged by the uncertainties in-herent in any nascent industry/technology. In order to compete against other energy sources (which may be subsidized and al-ready benefiting from the economies of scale), offshore wind en-ergy must be deliverable at a relatively attractive cost. To achieve cost goals, offshore wind facilities need to be highly reliable and durable (DOE, 2011). Offshore wind farm reliability is key in se-curing financing, insurance, social acceptance, market contracts and safety for both long-term and short-term perspectives.

Quantitative reliability assessments of an engineered system involve considering the probability that a system will success-fully meet defined performance criteria for a defined period of time. Adopting a consistent probabilistic design framework, in which the uncertainties and the reliability of the design are transparent to the designer, allows the engineer to incorporate and optimize with respect to project-specific risks and produce comparable designs. While risk-informed design may require a higher level of competency by the engineer than traditional pre-scriptive methods, offshore wind turbine design is particularly well suited to benefit from a project-specific risk-informed design approach. For example, the design phase of an offshore wind turbine involves less than 4% of its lifecycle cost (DOE, 2010), and the presence of current uncertainties in the design process chal-lenging the predictability of cost and reliability estimations may significantly influence total costs. Additionally, offshore wind turbines are designed in groups for a particular wind farm and have relatively little variability in structural configuration, thus facilitating even larger capacity for refining and updating reli-ability design and analysis methods compared with most civil engineering projects, which are typically unique.

STUDY OBJECTIVESThe objectives of this study project are two fold:

1) To introduce the concept of a risk-informed approach to performance assurance of offshore wind turbines located in the U.S. OCS, and2) To demonstrate a general framework for risk-informed design of offshore wind turbine support structures with respect to a target level of reliability.

Basic reliability principles are introduced and a reliability anal-ysis and design of a monopole support structure is performed based on an assumed acceptable level of risk to illustrate the con-cepts.

METHODS

Elements and Application of Risk-Informed Approach to Performance AssuranceThe offshore wind industry is challenged by a lack of empiri-cal observations or historical benchmarks from which to derive experience-based design criteria. Many agencies and countries, primarily European, have developed comprehensive yet largely prescriptive standards, and none are applicable, without sig-nificant modification, in the U.S. (TRB, 2011). In general, devel-opment of a risk informed basis for structural design requires:

• Definition of structural components and groups• Identification of important failure modes or limit states for components and systems• Stochastic models for uncertain parameters• Quantified performance goals (i.e. acceptable reliability level)• Standardized method and assumptions for reliability calculation• Risk-consistent design criteria to achieve the performance objec-tives

Reliability-based Formulation of Design CriteriaIn order to design for adequate performance, system require-ments can be translated into so-called limit state conditions from which equations can be developed that separate the ac-ceptable region of performance from what is considered the failure region. Defining the demand, S, and the capacity, R in a structural system, a safety margin, M, can be defined as: M = R - S (1)

A positive M in Eq. 1 represents adequate performance and a negative value represents failure. In the presence of un-certainty, R and S are random variables, and their uncertain-ties are modeled by their probability distributions. The fail-ure condition, or limit state, is defined by the inequality of M


being less than zero because the demand exceeds the capac-ity (R) of the system. Given probabilistic descriptions of R and S, the probability of failure, Pf, can be determined as: PF = P[M < 0] = P[R - S < 0] (2)

The probability of failure can be viewed as a risk metric, the com-plement of which is known as reliability. Thus, design for a stip-ulated value of Pf provides the basis of risk-based design. For example, the design parameters of the support structure or tower could be adjusted to achieve a target reliability allowing for eco-nomical and optimal design solutions by balancing decisions based on material consumption, performance requirements, fail-ure consequences and probability of failure for each group/com-ponent (Sorensen & Toft, 2010). Note that Pf is a subjective mea-sure, in the sense that it is dependent on the information available and the engineering models and assumptions used in perform-ing the calculation. Thus consistent reliability analysis methods and assumptions are critical, and the reduction of uncertainties in engineering models allows for better reliability estimates.

Example Development: Reliability-Based Design of Monopole Turbine Support To demonstrate the general concepts, we consider a typical monopole structure modeled to support a 5 MW turbine off the east coast of the U.S. (see Figure 1). The structural elements for the monopole include the pile and the tower. For simplicity, the wind turbine is assumed to be parked and the only limit state considered is the onset of yielding in the pile due to an overturn-ing moment (OTM) from a combination of actions due to wind, wave and current. The wind turbine is modeled to represent the NREL 5 MW baseline turbine defined by Jonkman, et al (2009) with a yaw misalignment of about 8 degrees. The site and en-vironmental conditions were extracted from MMI Engineering (2009), and the support structure was developed to be compa-rable to the monopole defined in that report. The environmen-tal conditions reflect data consistent with siting south of Mas-sachusetts and Rhode Island between Martha’s Vineyard and Block Island, and the water depth is assumed to be 25 meters.

Thus, given:1. Structural elements and configuration2. Proposed site and stochastic description of hazards and en-vironment3. Definition of limit state4. Assumed target level of reliability: Pf-target = 10-4 per year (DNV, 2007; TRB, 2011)

We seek the following goals:

1. More transparent structural system reliability2. Ability to employ probabilistic risk assessment and manage-men procedures when adjusting structural parameters

PROCEDURE AND MODELS

Structural Reliability Analysis and Design ProcedureThe primary engineering analysis tools used for this project in-cluded MATLAB, GTSTRUDL and GTSELOS. MATLAB’s ran-dom number generation capabilities were utilized to simulate 100 random independent environmental conditions (i.e. wind-wave-current parameters) and to perform the reliability analysis using Monte-Carlo simulation methods. GTSTRUDL was used for structural analysis and evaluating structural response, while GTSELOS was used to calculate the structural demand from combinations of wind-wave-current. Figure 2 demonstrates the general procedure followed in a simple flow chart.

Figure 1. Typical Mono-pole Offshore Wind Turbine Structure.

Figure 2. Flow chart of reliability-based design procedure.


Turbine ModelOnly parked turbine conditions are considered. For simplicity, the NREL 5 MW baseline turbine is modeled as a point mass at the top of the support tower. An equivalent flat plate area of 292 m2 was estimated from loads data calculated for this turbine published by the American Bureau of Shipping (ABS, 2011) to model the viscous drag forces on the parked rotor nacelle assem-bly (RNA), assuming an overall drag coefficient Cd=1.28. This is assumed to represent roughly an 8 degree yaw misalignment. It is noted that the magnitude of the drag loads on the RNA are sensitive to the degree of yaw alignment (ABS, 2011).

Support Structure ModelThe support structure defined in this report consists of two com-ponents: the pile and the tower. The pile is a single diameter extending from the penetration depth (60 m below mud line) to 10 m above the mean water level. To model a tapered tower, the tower defined in this report consists of 20 equal-length pipe seg-ments with incrementally decreasing diameters and thicknesses from bottom to top. Consistent with common practice, the pre-liminary design of the support structure conforms to a target natural frequency range of 0.20 to 0.34 Hz to avoid resonance with the rotor and blade passing frequencies of the turbine in op-erating conditions (MMI, 2009). The initial structural model has a natural frequency of 0.241 Hz. Structural damping of all modes was assumed to be 1%. The tower and initial pile properties are defined in Table 1.

Property Pile Tower

Base Diameter (m) 6. 5 6

Base Thickness (m) 0. 065 0. 03

Top Diameter (m) 6. 5 3. 78

Top Thickness (m) 0. 065 0. 019

Total Length (m) 95 77. 6

Density (kg/m3) 8500 8500

Damping Ratio 1% 1%

Young's Modulus, E (GPa)

210 210

Shear Modulus (GPa) 80. 8 80. 8 The specified steel density supplied by NREL (8500 kg/m3) is larger than typical steel values to account for paint, welds, and flanges (Jonkman et al., 2009). The pile is assumed to be an open tube driven to the target penetration depth with the interior filled

to the mud line with soil. Pile-soil interaction is modeled using horizontal and vertical linear soil springs, estimated from soil spring data assumed by MMI (2009). The mass of the soil inside the pile is modeled as a uniform added inertia mass in the hori-zontal direction based on an assumed soil density of 18 kg/m^3. The structural model can be seen in Figure 3.

Environmental Load DataThe wind turbine structure is assumed to be set on a flat sea bed. The only environmental load conditions considered in this study are effects of wind, wave and current. Data supplied by MMI (2009) including scattergrams of:• Wind speed (10m, 1hr mean), U10m,1hr vs. Significant wave height, HS• Significant wave height, HS vs. Average zero-crossing wave period, TZ• Maximum wave height, HMAX vs. HS

Along with 4 extreme storm conditions was used to estimate wind-wave-current condition parameters and statistical de-scriptions. A Gumbel (Type 1) distribution was used to model

Table 1. Initial Pile and Tower Properties

Figure 3. Support Structure Model


U10m,1hr as shown in Figure 4 with location and scale param-eters determined to be 15.28 and 7.95 respectively.

The 50- and 100-year markers in Figure 4 indicate the annual maximum wind speeds with a 50-year and 100-year return pe-riod, respectively. The relation between U10m,1hr and the mean significant wave height, HS-mean, is shown by the best-fit power relation in Figure 5, which is based on selected average points from referenced scattergrams and extreme storm states.

From HS-mean, a random value for HS was generated with the assumption of a normal distribution about the mean and a modi-fied standard deviation of 0.434 to account for an apparent 85% correlation between U10m,1hr, and HS observed in the refer-

enced data. The 100 environmental events were simulated by generating random values of U10m,1hr and corresponding val-ues of HS. The deterministic relationships for the remaining pa-rameters are shown below.

HMAX *= 2.144HS0.8719 (3)

T2 *= 4.29HS0.3512 (4)

T *= 1.2T2 (5) CS = 0.0091U10m, 1hr (6)

Note that the designation * indicates a relationship defined by MMI (2009).

Wind DemandAn empirical power law description of the wind speed profile shown in Eq. 7 was assumed with an exponent of 0.11 for ex-treme wind conditions, as suggested by IEC 61400-1 (2005).

U(z) = U10m, 1hr ( z )0.11 (7) 10m

in which z is the height above mean sea level. Wind vis-cous drag forces were calculated by Eq. 8:

FD = 1/2ρV2 CD Aproj (8)

where V is the wind velocity (m/s) CD = 0.5 (cylinder) Aproj is the projected area of the surface (m) ρ = 1.225 (kg/m3)

Wave and Current DemandStructural demands due to waves were calculated by GTSELOS using 5th Order Stokes Wave Theory. For simplicity, maximum breaking wave height and wave slam are not considered in this study. The values for the drag and inertia coefficients are con-sistent with MMI (2009). The current velocity at depth z, C(z), is defined as:

C(z) = Cs ( h0-z ) (9) h0

in which h0=50m is the reference depth for wind-generated cur-rent (DNV, 2007).

Figure 4. Defined Gumbel probability density function of annual maximum wind speed

Figure 5. Relationship for HS-mean and U10m,1hr


RELIABILITY ANALYSIS RESULTS AND DESIGN

Structural Demand, SThe 100 random environmental conditions were input as param-eters for load calculation in GTSELOS and the maximum over-turning moment (OTM) in the pile due to the applied loads was determined in GTSTRUDL. A Gumbel distribution was chosen to model the max OTM demand (see Fig. 6) because it is a com-mon distribution type for modeling intensities due to maximum extreme environmental events (Ang & Tang, 2006). The distri-bution parameters (location and scale factor) were estimated by determining a best-fit line of the data assembled on a Gumbel probability plot.

Structural Capacity, RThe capacity, in terms of OTM, of the structural model was de-termined by incrementally increasing four extreme event storms and identifying the OTM values when the defined limit state was reached. The average was taken to be the nominal capac-ity, OTMnom, which was 587.7 MN-m. The mean value of yield strength of typical construction grades of steel is approximately 10% higher than the specified nominal yield strength, while the coefficient of variation of a fabricated shape would be approx-imately 12% (Ellingwood, 2000). Thus, the mean OTM capac-ity, OTMcap, was assumed to be 10% larger than OTMnom, or 646.5 MN-m. A lognormal distribution and a coefficient of varia-tion equal to 12% was assumed to describe the OTM capacity as shown in Figure 7.

Reliability Analysis and CheckA Monte Carlo simulation method was used for reliability calcu-lations. With the estimated S and R distributions of the wind tur-bine support structure defined, 25 million random samples were taken from S and R, and the probability of failure was quantified by the ratio of the number of samples for which M was less than zero divided by the total number of samples. The probability of failure of the initial structure, Pf-1, was calculated to be 2.1x10-6 per year. Although Pf-1 is less than Pf-target, implying the risk is less than the level stipulated for design, an efficient design would have a Pf less than but near Pf-target¬ so the structure is not un-necessarily expensive in terms of cost and materials etc.

Design and Risk ModificationsGiven the environmental conditions, structure configuration, and turbine, the design engineer can adjust the risk associated with failure induced by OTM, for example, by adjusting the OTM capacity (i.e. altering pile thickness, diameter, or selecting an al-ternate strength steel). To demonstrate the concept, it is assumed that decreasing the pile thickness is the most beneficial approach to increasing the risk to be nearer the target level. Thus, the pile thickness was reduced from 6.5 cm to 5 cm below the mud line. The natural frequency of the modified structure was 0.234 Hz, which remains within the target range. The analysis procedure was repeated with the same assumptions and the probability of failure of the modified structure, Pf-2, was calculated to be 3.1x10-4 per year, which is close to Pf-target and completes the risk-based design procedure.

Figure 6. Gumbel probability density function of maximum annual OTM demand, S, for initial structure

Figure 7. Lognormal probability density function of OTM resistance, R, for initial structure


CONCLUSIONS AND FUTURE DEVELOPMENTOffshore wind turbine development, particularly in the U.S., would benefit from standardized risk informed design proce-dures. A general framework for risk informed design has been demonstrated on a typical monopole support structure sited on the U.S. OCS. Given consistent reliability analysis methods and assumptions, the Pf can be used as a risk metric for adjustment of design parameters with respect to cost to achieve an acceptable level of reliability. Examples of helpful methods and assump-tions to be developed include region-specific statistical distribu-tions for environmental parameters, recommended methods for determining distribution parameters, and consistent structural modeling assumptions. Additionally, in order to implement such a design procedure, regulations would need to define clear performance requirements (i.e. a P¬f-target) for project approval. As the industry is just being deployed in the U.S., regulators and standards organizations have the opportunity to endorse a risk informed design basis from the very beginning.

ACKNOWLEDGEMENTSThis report was developed as part of an undergraduate indepen-dent study project. The author is deeply honored to have had the opportunity to study and grow under the advisement of Dr. Bruce R. Ellingwood. His unsurpassed advice and expertise in structural reliability, risk assessment, and standards develop-ment proved invaluable and enlightened the author. Special thanks are also due to Dr. Emkin for providing software critical to the project, GTSTRUDL and GTSELOS.

REFERENCESAmerican Bureau of Shipping (ABS). (2011). Design Standards for Off-shore Wind Farms. 2, 79-90.

Ang, A.H.-S & Tang, W.H. (2006). Probability Concepts in Engineering: Emphasis on Applications to Civil and Environmental Engineering. John Wiley & Sons. ISBN 047172064X

Det Norske Veritas (DNV). (2007). Offshore Standard DNV-OS-J101 De-sign of Offshore Wind Turbine Structures. Oslo, Norway.

Ellingwood, B. (2000). LRFD: implementing structural reliability in professional practice. Engrg. Struct. 22(2):106-115.

International Electrotechnical Commission (IEC). (2005). IEC 61400-1 Ed. 3, Wind Turbines – Part 1: Design requirements.

Jonkman, J. M., Butterfield, S., Musial, W. & Scott, G. (2009). Definition of a 5-MW Reference Wind Turbine for Offshore System Development. National Renewable Energy Laboratory, Golden, Colorado, Report No. NREL/TP-500-38060

MMI Engineering Inc. (2009). Comparison of API & IEC Standards for Offshore Wind Turbine Applications in the U.S. Atlantic Ocean: Phase II. MMI Project No. MMW554.

U.S. Department of Energy (DOE). (2010). Creating an Offshore Wind Industry in the United States: A Strategic Work Plan for the Unites States Department of Energy, Fiscal Years 2011-2015.

U.S. Department of Energy (DOE). (2011). Offshore Resource Assess-ment and Design Conditions Public Meeting – Summary Report. Wind and Water Power Program Energy Efficiency and Renewable Energy, U.S. Department of Energy.

Sorensen, John D., & Toft, Henrik S. (2010). Probabilistic Design of Wind Turines. Energies 2010, 3, 241-257; doi:10. 3390/en302024

Transportation Research Board (TRB). (2011). Structural Integrity of Offshore Wind Turbines; Oversight of Design, Fabrication, and Instal-lation. Transporation Research Board of the National Academies. Washington, D.C. 20001. Special Report 305.


ARTICLES


T h e R E D D + Pr o g r a m m e :A f f e c t s o n G o v e r n a n c e T h e o r y, M a r k e t T h e o r y, a n d a Po s t - Ky o Wo r l d

Te j a s K o t a k

INTRODUCTIONThe majority of the scientific community agrees that anthropo-genic climate change is being caused by the release of greenhouse gas (GHG) emissions into the atmosphere, the main one being carbon dioxide. As climate change is an international issue, in-ternational measures organized under the United Nations have been implemented. The Kyoto Protocol, which is considered the foremost international treaty to reduce anthropogenic climate change, entered into force in 2005 and is due to expire at the end of 2012. The future efforts to mitigate climate change in a post-Kyoto world have been discussed at the international level, and the Reducing Emissions from Deforestation and Forest Degra-dation Programme (REDD+) is an international agreement that aims to address the issue of climate change due to deforestation and forest degradation while also encouraging the enhancement of forest stocks in developing countries with tropical forests.

Along with GHG emissions from anthropogenic sources, such as automobiles and industrial processes, the deforestation and degradation of forest land adds to net GHG emissions by re-leasing the CO2¬¬ stored in trees and by removing a carbon sink. Land-use activities and deforestation are estimated to contribute between 20-25% of global GHG emissions (Madei-ra 2008), and current worldwide forest stocks are estimated to contribute to roughly one third of carbon abatement (Sohgen & Mendelsohn 2003). The governance system of the REDD+

Programme includes mechanisms which involve both govern-ments and markets, and it potentially commodifies (forests) a resource based on its potential to sequester pollutants rather than pure commercial value. It therefore represents an interest-ing development in both governance theory and market theory.

THE REDD+ PROGRAMMEThe REDD+ Programme, as described in the Bali Action Plan, is an international agreement on creating “Policy approaches and positive incentives on issues relating to reducing emissions from deforestation and forest degradation in developing coun-tries; and the role of conservation, sustainable management of forests and enhancement of forest carbon stocks in developing countries” (UNFCCC 2008a); the “+” indicating the phrase fol-lowing the semicolon. In 2008, the UN-REDD Programme was launched to prepare developing nations for “REDD readiness” stages which will help them in creating the capacity and infra-structure for full-scale REDD+ initiatives (Johns, Johnson, & Greenglass 2009). REDD+ Programme was approved under the UNFCCC during the 2010 Conference of Parties (COP) in Can-cun, and it aims to create a governance system which addresses the GHG emissions resulting from deforestation and forest deg-radation while supporting forest conservation, sustainable for-estry, and the enhancement of forest carbon stocks in developing nations with large forest stocks. As it is set up under the UN-FCCC, the primary players in implementing the Programme will

ABSTRACTThe Reducing Emissions from Deforestation and Forest Degradation Programme (REDD+) is an international agreement that aims to address the issue of climate change due to deforestation and forest degradation while also encouraging the enhancement of for-est stocks in developing countries with tropical forests. This paper aims to address the ways in which REDD+ would fundamen-tally affect governance and market theories if it is applied, and it discusses REDD+’s potential as a successor of the Kyoto Protocol.


be the participating governments. The official Programme sets up the framework, overall structure, and goals, but each indi-vidual nation can negotiate the terms under which it is imple-mented. The forest industry has traditionally been handled either by local, sub-national governments, or by private companies.

FINANCIAL AND MARKET STRUCTUREThe Carbon Fund is one market-mechanism the REDD+ Pro-gramme will implement. The structure of the Fund is detailed in the Forest Carbon Partnership Facility’s (FCPF’s) 2008 Infor-mation Memorandum. Developing countries which have rec-ognized REDD+ Programmes and have gained certified emis-sions reductions (CERs) will be remunerated by the Fund based on the amount of reductions they have achieved. The monetary value of emission reductions for each nation will differ based on an assessment by the World Bank which will consider the na-tion’s individual REDD+ Programme and the method in which the reductions were gained. This function of the Fund effectively makes a commodity out of REDD-based CERs by placing a solid, monetary value on them. The Fund acts as a market mechanism known as a “payment for environmental services” (PES) to act as an incentive for developing nations to participate in REDD+ Pro-grammes (Anglesen et al. 2008). It may also provide incentives to sub-national governments, indigenous peoples, and private sec-tor entities based on the structure of the CER programme and the agreement between the national governments and the FCPF on who will receive the PES. The Carbon Fund faces many risks as PES system including the CERs staying constant after they have been purchased and the Carbon Fund being large enough. The Fund currently has a set target budget of $200 million, and is only planned to be active during the REDD readiness phases of national implementation. Once nations are considered “REDD ready,” different market mechanisms will have to be in place in order to secure long-term effectiveness of the REDD+ Programme.

One such mechanism is the creation of a carbon market for CERs created through REDD+ activities. Similar to the Carbon Fund’s structure, a REDD+ market calls for the creation of CER credits, but the credits are not purchased by a single buyer as in the Fund. Rather, they are bought and sold in a marketplace among par-ticipating players, which may include national and sub-national governments along with private sector entities. One main issue with making CERs is making appropriate measurements of emis-sion reductions due to the complexity of measuring how much carbon has been sequestered. Current carbon markets create a set amount of CERs within the system before the trading begins. REDD+ credits, on the other hand, would be created after the emission reductions occur, which means that there is an uncer-

tainty to how many could enter the market. There is a worry of flooding the current international CER market (which is conduct-ed primarily in the European Union) with cheap REDD+ credits (Fry 2008; Hamilton 2008). The goal of carbon markets is to pro-vide a cost-effective method for reducing GHG emissions, but if it becomes too cheap for participating nations and companies to stay within the cap then it becomes difficult to progressively lower the cap amount – which is part of the planned evolution of cap-and-trade systems. The market structure of REDD will even-tually need to change from using the Carbon Fund to something else, but the viability of using a carbon market is of much concern.

OVERLAP OF GOVERNANCE STYLESIn the current structure of the REDD+ Programme, most of au-thority rests either in the national governments who are imple-menting their own programmes, or with the international in-stitutions of the UN and the World Bank. Due to this current structure, there is no direct involvement of non-public or non-government governance within the hierarchy. Despite this fact, market governance plays a role in REDD+. The Carbon Fund and the possible carbon market future both provide market-based incentives for developing governments, local and indig-enous peoples, and private sector player to participate in the Programme (Angelsen et al. 2008). These players are given the incentive of a PES to entice them to be involved with REDD+ ac-tivities whether they be setting up the infrastructure for a full Programme implementation or sustainable forestry techniques.Market governance is still relatively unused at this stage in REDD+ implementation. Current goals are to set up the infrastructure in developing nations so that REDD+ activities can be properly initiated in a few years time. The time at which the REDD+ Pro-gramme is set up both internationally and in developing nations is when market governance will become the driving force of the Programme’s goals. Market governance within the Programme is reliant on the creation of some form of carbon market to manage the carbon reduction resulting from REDD+ activities. If a carbon market is formed then market-based incentives will move the Programme forward by funding the participants in developing nations and providing the desired results. There is an extent to which the market incentives of acting in compliance with the goals of the Programme will outweigh those of the commodity markets of logging and forestry. Once this limit is reached, the market system may help keep a balance of afforestation versus defor-estation rates through competition with the commodity market.

IMPLICATIONS OF REDD+Political and economic evolution is stimulated not just by the debates and discourses of academia, but by actual implementa-


tions of these ideas. The REDD+ Programme is an application of theories that are debated in political, economic, and environ-mental circles, and it provides a case study that can be exam-ined in the context of those theories and the additions it offers to them. In its current form, it can be looked at from the stand-point of its potential to add to the theories and discourses as it is too soon to see if it will succeed or not. When taking into ac-count its goals and framework and potential future evolution, the REDD+ Programme has many implications for governance theory, market theory, and the future of climate change policy.

GOVERNANCE THEORYGovernance, while separate from government, does tend to draw its authority and power from a government. The REDD+ Programme draws its authority in part from two main sources. The international governing body of the United Nations and the UN-REDD Programme provide the primary authority as the source of the REDD+ Programme and the international overseer respectively. The national governments which enact their own programmes give the Programme authority as they are taking it off print and paper and making it a reality. Traditional hierar-chies of both government and governance take the form of a cen-tral national government on top and subnational governments under it. The addition of an international system that is both above the national governments and overlaying the entire hier-archy brings a new dynamic into the implications of governance.

MARKET THEORYWithin the market there are costs, benefits, and externalities. The negative externalities of these activities include the effects on lo-cal populations, the effect on the biodiversity of the area, and the release the CO¬¬2 into the environment. The REDD+ Programme aims to address these externalities in forest-based industries by turning them into commodities. The potential of providing a PES to developing nations and local participants is a step towards in-creasing the importance of forests and the social equity of local and indigenous peoples as they are receiving payment for what they do. The PES method compensates for the externalities, but it does so outside of the market.

In order to see compensation within the market, a REDD+ car-bon market would need to be created. If a carbon market does become the future of the REDD+ Programme the international carbon market will be expanded and be seen in greater signifi-cance than it is now. The carbon market turns the environmen-tal externality of emissions into a commodity that has tangible value both in the cap-and-trade system of the carbon market as

well as monetary value that can be used in the traditional mar-ket. The emission externalities are accounted for by commodif-ing them and placing them into a free trade market, which al-lows for both cost-effective reduction of emissions and the use of capitalistic mechanisms to do so. Carbon markets have existed since the early 1990s, so they are not new, but they do provide cases in which market mechanisms are employed by the public sector rather than staying with private sector players.

The REDD+ Programme, with both the current PES system and the possible carbon market future, is using market mechanisms to incite a change in activities, and the major difference here is that the Programme is based in the public sector. If the public sec-tor takes on a more active role as a participant within the market then the future market will have a very different dynamic com-pared to todays where the private sector has the most influence.

SOCIAL EQUITYThe official structure of the Programme is built to include the needs and opinions of local and indigenous peoples as they will be the ones who are most effected by REDD+ activi-ties. The inclusion of the local and indigenous peoples within the governing and decision-making processes expands the Programme’s influence on social equity. If REDD+ activi-ties succeed in achieving their goals then the lifestyles and lo-cal economies of these peoples will be greatly enhanced as well as become far more sustainable than they currently are.The argument of social equity is one of great idealism rather than rationalism. It is a hope that individuals, nations, and corporations will work together for the common good of all of humanity. Envi-ronmental agreements and actions should focus more on increas-ing and stabilizing the social equity of groups which are affected by the actions and of humanity at large, but the argument is not one that is regarded highly due to its lack of tangible results. Market-based arguments have taken precedence in environmental debates because changes in the market are able to produce easily measur-able economic results as well as the desired changes in action.

The REDD+ Programme makes use of the market argument as a means to achieve social equity for the local and indigenous peoples of developing countries. This use of market mechanisms may prove to be an effective method of gaining social equity, but it invalidates the core of social equity argument by placing the economy higher than the quality of life for human beings. A POST-KYOTO WORLDThe Kyoto Protocol officially went into force in 2005, and is due to


expire at the end of 2012. The Protocol is largely considered to have been ineffective in achieving its goals of climate change abatement, but it still represents the largest step towards international climate governance. A new framework is needed to fill the gap that the Protocol is leaving, and the REDD+ Programme may be able to fill that gap. While the Programme has much narrower focus than the Protocol, it can potentially have a much greater long-term effect.

The goal of the Kyoto Protocol was to reduce the emissions of developed nations whereas the goal of the REDD+ Programme is centred on the carbon sequestration in developing nations. While reducing the emissions of developed nations is indeed important, allowing developing nations to grow within a framework that re-duces their emissions while also creating sustainable economies is more important for significant progress to be made in the twen-ty-first century. If the developing nations that are participating in the Programme grow within a REDD+ framework, they will be able to become global players without the worry of reducing their effects on the environment down the line as the current de-veloped nations are. The Protocol essentially kept the status quo and the global mentality of the twentieth century by wanting the developed nations to alter their behaviour while allowing the de-veloping nations grow in an unsustainable fashion. As mentioned earlier, the CDM of the Protocol did allow for projects in devel-oping nations, but infrequent projects are not the same as setting up a framework for sustainable development. The REDD+ Pro-gramme can act as the first step into a post-Kyoto world by pro-viding a new climate change deal and by offering the potential for changing the status quo that the Protocol aimed to keep in place. CONCLUSIONCreating significant reductions in the amounts of GHGs that are emitted due to human activities has been a difficult en-deavour due to developed countries (the United States in particular) having economic-based objections to reduction programmes. The REDD+ Programme is an international agree-ment that aims to reduce the rates of deforestation and for-est degradation in developing countries while also increasing and enhancing their total forest stocks. The economic objec-tions of developed countries are made into moot points by fo-cusing the Programme on reductions in developing countries.The Kyoto Protocol is due to expire at the end of 2012, and the REDD+ Programme can be considered a replacement for it as a major international climate change agreement. Unlike the Pro-tocol, though, the Programme alters the focus of climate change programmes from developed nations to developing nations. The shifts of responsibility of GHG reductions from developed countries to the developing countries may be viewed as an un-fair burden for them, but the Programme sets up a framework

which allows developing nations to grow and develop in such a way that will result in economies that are both environmen-tally and economically sustainable. The REDD+ Programme represents a modern climate change deal that changes the sta-tus quo for international climate change programmes through the shift of responsibility while creating the potential to alter the status quo of international politics and the international mar-ket by promoting sustainable growth in developing nations.

REFERENCESAngelsen, A., Streck, C., Peskett, L., Brown, J., & Luttrel, C. (2008). What is the right scale for REDD? The implications of national, subnational, and nested ap-proaches. CIFOR, 15.

Bell, S., Hindmoor, A., & Mols, F. (2010). Persuation as Governance: A State-Centric Relational Perspective. Public Administration, 88(3), 851-870.

Evans, P., & Rauch, J. E. (1999). Bureaucracy and Growth: A Cross-National Analysis of the Effects of “Weberian” State Structures on Economic Growth. American Sociological Review, 64(5), 748-765.

Fry, I. (2008) Reducing Emissions from Deforestation and Forest Degradation: Opportunities andPitfalls in Developing a New Legal Regime. Review of European Community & International Environmental Law, 17, 166-182.

Hamilton, I. (2008). REDD carbon markets: Proposals compared. carbonposi-tive. Retrieved March 31, 2011, from http://www.carbonpositive.net/viewarticle.aspx?articleID=1209

Lederer, M. (2010). “Establishing an Effective and Legitimate REDD System - Is There Anything We Can Learn from the CDM?” Paper presented at the annual meeting of the Theory vs. Policy? Connecting Scholars and Practitioners, New Orleans Hilton Riverside Hotel, The Loews New Orleans Hotel, New Orleans, LA Online Retrieved 2011-03-07 from http://www.allacademic.com/meta/p416044_index.html

Madeira, Erin C.M. (2008). “Policies to Reduce Emissions from Deforestation and Degradation (REDD) in Tropical Forests: An Examination of the Issues Fac-ing the Incorporation of REDD into Market-Based Climate Policies.” Resources for the Future Discussion Paper.

Sohgen, B. & Mendelsohn, R. (2003) An Optimal Control Model of Forest Car-bon Sequestration. American Journal of Agricultural Economics, 85(2), 448-457.

Streck, C. (2004). New Partnerships in Global Environmental Policy: The Clean Development Mechanism. The Journal of Environment & Development, 13(3), 295-322.

UNFCCC. (2008a). Report of the Conference of the Parties on its Thirteenth Ses-sion, Held in Bali from 3 to 15 December 2007.

UNFCCC. (2008b). UN Collaborative Programme on Reducing Emissions from Deforestation and Degradation in Developing Countries (UN-REDD): Frame-work Document.

UNFCCC. (2009). Clean Development Mechanism (CDM). Clean Development Mechanism. Retrieved February 28, 2011, from http://unfccc.int/kyoto_protocol/mechanisms/clean_development_mechanism/items/2718.php

World Bank. (2010). “Africa’s First Large-Scale Forestry Project Under the Kyoto Protocol.” Press release. Retrieved 30 March 2011 from: http://go.worldbank.org/8B352TV090


GTTOWER.ORG | 39

ARTICLES

ABSTRACTIn this paper, we measure statistical relationships between defining characteristics of incoming kindergartners and their initial literacy scores. Our analysis focuses on four elementary schools in Oregon’s Springfield School District: two Promise Neighborhood schools and two com-parable non-Promise Neighborhood schools. Using scores from the literacy benchmark tests each incoming student takes upon entering kin-dergarten—controlling for variables such as family income, English language learners, gender, special education, and ethnicity—we find the defining characteristics with the most significant relationships that influence literacy scores. In the absence of a fully randomized experi-mental design, we give policy suggestions to United Way of Lane County to more effectively increase early literacy in the Lane County, as well as offer advice on the kinds of additional information that would permit a more definitive future study of the Promise Neighborhoods.

J a c o b M c G r e w & E l i z a b e t h Lo h r k e

A r t i c l e E xc h a n g ew i t h t h e U n i v e r s i t y O f O r e g o n :

U n i t e d Wa y o f L a n e C o u n t y ’s Pr o m i s e N e i g h b o r h o o d s a n d t h e B e n e f i t s o f R e a d i n g R e a d i n e s s

I. INTRODUCTIONThe national United Way Promise Neighborhoods movement was created in 2010 to develop a continuum of “cradle through college and career” (Promise 2010) solutions to improve the educational and developmental outcomes of children living in the United States’ most distressed neighborhoods. Based on the work of Geoffrey Canada in the Harlem Children’s Zone, Promise Neigh-borhoods could be an efficient solution to releasing thousands of children from the lifelong effects of poverty by developing a full continuum of supports for children, prenatally through emerg-ing adulthood, in families, schools and neighborhoods with the support of a broad range of community partners found in the

sectors of education, business, social service, health, government, faith, and many more (Promise 2010). Children who enter school unprepared to learn tend to face more obstacles throughout their schooling and have a lower degree of long-term success in their adult lives. United Way of Lane County is focused on building a foundation for a successful life for every child by increasing the number of children who enter school ready to learn.

In Lane County, Oregon, thirteen of sixteen school districts use either Dynamic Indicators of Basic Early Literacy Skills (DIBELS) or EasyCBM to measure incoming kindergartners’ early literacy


skills and assess how prepared they are to learn to read, an im-portant sign of school readiness (Promise 2010). Assessed skills include letter recognition, sound fluency and print familiarity. While standardized testing may be an imperfect gauge of student potential, it is currently the best available measure. United Way of Lane County (UWLC) began collecting and aggregating liter-acy score data from all participating school districts in 2010, only to discover disturbing results. More than half of all children en-tering kindergarten in Lane County do not have the early literacy skills they are predicted to need for success in school based on the early literacy benchmarks set up the creators of each assessment tool. Two Promise Neighborhoods have been established in the county’s lowest scoring communities: the first in the Springfield School District and the other in Eugene’s Bethel School District. In these two combined neighborhoods, 82 percent of children en-tering kindergarten do not meet the early literacy benchmark, as compared to 56 percent across the rest of Lane County. The intent of the Promise Neighborhoods is to concentrate resources on pi-loting innovative programs to improve incoming kindergartners’ school readiness and so as to close the school achievement gap between students in all neighborhoods across Lane County.

2. BACKGROUNDIn 2010, United Way of Lane County aligned its community in-vestment process with its established 2020 goals in education, in-come and health. UWLC’s primary education goal is for all chil-dren to enter school ready to learn. This goal is broken down into three specific outcomes:

• Children enter school with age-appropriate early language and literacy skills.• Children enter school with age-appropriate social and emo-tional development.• Parents have the knowledge and tools to be actively involved in their child’s development and education.

UWLC’s strategic education investments include parenting edu-cation programs, childcare improvement efforts, and early learn-ing programs (Promise 2010).

During preliminary discussions with United Way’s Associate Di-rector of Education, Holly Mar Conte, our research team received proposed project goals that would give UWLC a compelling case for strategic investment in the Promise Neighborhoods. This project consisted of three distinct projects:• Prepare a literature review with a strong focus on the short- and long- run indirect costs of children entering school who are unprepared to learn to read, and gather background information on literacy testing.

3044

26

At Benchmark

Some Risk

High Risk

18

31

51At Benchmark

Some Risk

High Risk

27

27

46

At Benchmark

Some Risk

High Risk

45

17

38

At Benchmark

Some Risk

High Risk

2010 Promise NeighborhoodEarly Literacy Data

2011 Promise NeighborhoodEarly Literacy Data

2010 Lane CountyEarly Literacy Data

2011 Lane CountyEarly Literacy Data

Figure 1: Literacy Data for Lane County

Figure 2: Map of Springfield School District


• Identify the impact of UWLC’s Strategic Investments in the Promise Neighborhoods after controlling for factors such as fam-ily income, English language learners, gender, special education, and ethnicity.• Make recommendations for linking data from UWLC- fund-ed programs to school records for data tracking and future as-sessment. This would include recommending questions for fu-ture surveys given out at kindergarten registration.

This analysis focused on two different elementary schools in the Springfield Promise Neighborhoods: Two Rivers-Dos Rios (formerly Brattain) and Maple elementary schools. At 14- and 47-percent, respectively, these elementary schools have some of the highest percentages of students falling short of early literacy benchmarks throughout Lane County (Promise 2010).

3. LITERATURE REVIEW This paper’s purpose is to give statistical evidence to the impor-tance of early childhood development (ECD) programs so that policy makers have a bigger incentive to fund similar programs. The Brookings Institution’s William Dickens, Isabel Sawhill, and Jeffrey Tebbs (2006) noted that it is difficult for politicians to al-locate money towards long-term investments such as ECD be-cause they often face immediate pressures to fund ongoing or immediate aid programs. Though often under funded by both state and local governments, ECD programs have consistently been shown to have a notably higher return for each dollar spent compared to most other programs, both in the short-run and the long run (Rolnick & Grunewald 2003). The economic benefits in-clude a higher likelihood of high school graduation, which leads to a decreased chance of participants committing future crimes or having to rely on welfare benefits (Belfield et al 2005) as well as increased civic involvement and a lower chance of unplanned pregnancies (der Gaag & Tan 1998).

Though arguments are made that extreme poverty and low pa-rental education are the causes of under-performance in school rather than the lack of ECD, 20-year longitudinal data suggests that preschool cognitive and behavioral functioning is highly predictive of literacy in young adulthood, even when the ef-fects of family environmental characteristics, including living ar-rangements, the quality of the home environment, maternal edu-cation, and income are held constant. But it does not stop at just preschool or kindergarten; grade failure in elementary school is also associated with literacy, but this effect disappears after con-trolling for the measure of preschool abilities (Baydar et al 1994). This suggests that grade failure throughout elementary school and beyond is not precisely correlated with literacy at the time of the test, but instead dependent on literacy abilities learned at the

preschool level.

It is crucial that early childhood development programs are im-plemented as soon as possible. A key finding of University of Minnesota’s Judy Temple and Arthur J. Reynolds (2007) is that the economic returns from high-quality preschool programs are much higher than educational interventions implemented after a child enters school. The University of Cincinnati’s Victoria Pur-cell-Gates and Karin Dahl (1991) found that early literacy plays a crucial role in raising academic achievement, “The children who were the most successful at reading and writing at the end of first grade began kindergarten with more highly and broadly developed schemata about written language as compared to the children who were the least successful.”

The Promise Neighborhood program was only recently imple-mented in Lane County, so long-term effects will need to be interpreted from similar studies. We make some assumptions using a cost-benefit analysis of the High/Scope Perry preschool Program, which collected data on 40-year old individuals who at-tended the program as children (Belfield et al 2005). In the Perry study, program costs were compared against treatment impacts on educational resources, earnings, criminal activity, and welfare receipts. The treatment group obtained significantly higher earn-ings than the control group who did not receive the program. For the general public, higher tax revenues, lower criminal jus-tice system expenditures, and lower welfare payments easily out-weigh program costs; they re-paid $12.90 for every $1 invested. Even though the individual returns through this program were only around 6 percent, the returns to society were more than 12 percent (Heckman & Masterov 2007). The largest program gains came primarily from reduced crime by males. While Lane County jails are being forced to close numerous beds and lay off multiple workers due to budget cuts, the amount of crime in the county is not decreasing quickly enough to deal with these jail space shortages. In the long run, enriched ECD programs are the least-cost, most effective way to reduce crime, far more effective per dollar than increased expenditures on police or incarceration (Heckman & Masterov 2007).

The Harlem Children’s Zone, a model for the Promise Neighbor-hoods, has also been widely cited recently in comparable research studies. Results from the Harlem Children’s Zone (HCZ) suggest that high-quality schools are crucial in increasing achievement among low-income students. Harvard University’s Will Dobbie and Roland G. Fryer, Jr. (2010) provided the first empirical test of the causal impact of attending Promise Academy charter schools in the HCZ on educational outcomes, with a focus toward see-ing whether schools alone can eliminate the achievement gap


or whether the issues that poor children bring to school are too much for educators alone to overcome. The Promise Academy was found to be successful at boosting achievement in both math and English language arts in elementary school. “High-quality schools or community investments coupled with high-quality schools drive these results, but community investments alone cannot” (Dobbie & Fryer 2010). Focusing more funding on ECD programs in the Promise Neighborhoods is an extremely cost ef-fective solution for the long run sustainability of Lane County.The benefits of cognitive readiness for entering kindergartners do not stop with higher test scores and early literacy skills; re-search increasingly shows the importance of social-emotional development in a child’s readiness to learn. In a study utilizing a sample of 356 four-year-old children attending Head Start, the behavioral aspects of school readiness, including classroom participation, pro-social behavior, and aggression control were related to cognitive readiness assessments given at the start of the prekindergarten year (Bierman et al 2009). It was found that classroom participation and pro-social behavior each accounted for unique variance in cognitive readiness, while aggressive be-havior was associated with low levels of executive function skills. It was concluded that the promotion of competencies associated with classroom participation and pro-social behavior may be particularly critical to cognitive readiness in prekindergarten, which supports the holistic approach used in the Promise Neigh-borhoods. Social-emotional data was not tracked in this study, but we recommend tracking for this data in the future.

4. METHODOLOGY

4.1 Basic Structure To perform a meaningful analysis, we had two main goals:• Analyze the effect of a variety of incoming kindergarten stu-dents’ characteristics, gathered by each school district, on stu-dent fall literacy assessment scores.• Analyze the direct effects of the Promise Neighborhood on fall literacy assessment scores.

A multiple linear regression using Ordinary Least Squares (OLS) would have been the easiest and the most precise way for us to estimate these effects, but after we received the data from the school districts and ran our regressions, we realized this analysis would not be so simple. Besides the data not being uniform in col-lection or organization, the sample sizes were not large enough and the Promise Neighborhood data was strongly affected by se-lection bias—which will be discussed later in this paper. This led to very large standard deviations in all regressions, findings that were not statistically significant (having a high p-value), and a

very small coefficient of determination, or R2. The p-value is the probability of obtaining a test statistic at least as extreme as the one that was actually observed, R2 is most often seen as a number between 0 and 1.0, used to describe how well a regression line (created by estimated data) fits a given set of data. An R2 near 1.0 indicates that a regression line fits the data well, while an R2 clos-er to 0 indicates a regression line does not fit the data very well. R2 provides a measure of how well future outcomes are likely to be predicted by the model. A low R2 was expected since we were dealing with a non-randomized pool of data on a small sample size of children. Because of these problems, we were unable to establish any causal links since we could not accurately estimate coefficients for most of the characteristics. In order to perform a meaningful analysis, we modified our goals to:

• Measure the statistical relationships between the defining characteristics we received from the school districts and fall lit-eracy assessment scores.• Measure the statistical relationship between a kindergarten student being in a Promise Neighborhood and that student’s fall literacy assessment score, then analyze the difference in relation-ship after controlling for defining characteristics.

These adjustments would allow us to produce a more meaningful analysis for United Way since our findings would be based off of statistically significant finding. Statistical significance measures whether observations reflect a pattern rather than just chance. Based on previous studies on similar programs such as Harlem Children’s Zone and Perry Preschool, we expected to find a nega-tive correlation between literacy scores and characteristics such as English Language Learners (i.e. language spoken at home may not be English, while literacy tests are in English) and low in-come household (i.e. low parental education level, or simply a lack of disposable income to purchase at-home reading materi-als or enroll the child in early childhood development programs) and a positive correlation between literacy scores and living in a Promise Neighborhood (Belfield 2005, Dobbie 2011). In our con-clusion we will discuss the importance of a solid experimental design in analyzing educational programs such as the Promise Neighborhoods.

4.2 Data AcquisitionOur data was collected from Promise Neighborhood schools and comparable schools in the Springfield School District during each school’s kindergarten registration. The Promise Neighborhoods program was piloted in January of 2010, allowing us to obtain data from the 2010-2011 school year (denoted as 2010) and the 2011-2012 school year (denoted as 2011). Working with United Way’s Associate Director of Education, Holly Mar Conte, and


Springfield School District’s Director of Elementary Education, Sara Ticer, we received data on each kindergarten student in the two districts. The comparison schools were also chosen by Sara Ticer.

4.3 Scoring Characteristics for Each NeighborhoodDistricts throughout Lane County implement different systems to measure reading readiness in kindergarten students. Bethel School District uses DIBELS, while Springfield School District uses EasyCBM. Each assessment implements different tests and grading scales to measure literacy. DIBELS tests on initial sound fluency and letter naming fluency while EasyCBM tests on let-ter sounds and letter naming. A child is considered “low-risk” in the DIBELS assessment if he or she receives a score of 8 or above on each test (DIBELS n.d.), while EasyCBM is scored on norms so the score corresponding to a student’s risk level changes each year depending on the class (Richards n.d.). In the Bethel School District, the minimum score was 0 while the maximum was 97 with a mean of 23.74 and a standard deviation of 20.31. In the Springfield School District, which is the district we observed, the minimum score was 0 and the maximum was 97 with a mean of 18.32 and a standard deviation of 17.53.

4.4 VariablesOur variables are listed below with explanations of what was measured and how each was measured. Our reference group was white male kindergarten students who were not on free or reduced lunch, were not registered for special education classes, and were native English speakers. Reference groups, also known as comparison groups, are used in order to evaluate and deter-mine the nature of a given individual or other group’s character-istics.

4.4.1 Dependent VariableSCOREi = The literacy benchmark score of the ith student, as test-ed in the fall of kindergarten year. This score is the sum of Let-ter Names (LN) and Letter Sounds (LS) using EasyCBM for the Springfield School District.

4.4.2 Independent VariablesFEMi = A dummy variable that is 1 if the ith student is female and 0 if the student is male.

LUNCHi = A dummy variable that is 1 if the ith student qualified for Free or Reduced Lunch and 0 if the student did not. This was our proxy to identify low-income households. Households with incomes at or below 130% of the poverty level qualify for free lunches. Households at 130-185% of the poverty level qualify for

reduced lunches (Income n.d.).

SPEDi = A dummy variable that is 1 if the ith student is enrolled in Special Education classes and 0 if the student is not. In Oregon, students are placed in Special Education classes if they are evalu-ated as having one of the following: intellectual disability; hear-ing impairment, including difficulty in hearing and deafness; speech or language impairment; visual impairment, including blindness; deaf-blindness; emotional disturbance; orthopedic or other health impairment; autism; traumatic brain injury; or spe-cific learning disabilities (Oregon 2011).

ESLi = A dummy variable that is 1 if the ith student is enrolled in an English as a Second Language class and 0 if the student is not. This was our proxy to identify non-native English speakers.

ETHHISPi = A dummy variable that is 1 if the ith student is His-panic or Latino and 0 if the student is not.

ETHBLACKi = A dummy variable that is 1 if the ith student is Black or African American and 0 if the student is not.

ETHASIANi = A dummy variable that is 1 if the ith student is Asian or Pacific Islander and 0 if the student is not.

ETHAMERINDi = A dummy variable that is 1 if the ith student is American Indian or Native Alaskan and 0 if the student is not.

ETHMIXEDi = A dummy variable that is 1 if the ith student is mixed ethnicity and 0 if the student is not.

PNi = A dummy variable that is 1 if the ith student lived in a Prom-ise Neighborhood and 0 if the student did not.

The Springfield Promise Neighborhood schools were Maple El-ementary and Two Rivers-Dos Rios (Brattain) Elementary, while the non-Promise Neighborhood schools were Moffitt Elementary and Riverbend Elementary.

4.5 Empirical ModelingA fitted linear regression model can be used to identify the re-lationship between a characteristic variable Zi and the response variable Scorei when all the other characteristic variables in the model are “held fixed”. Specifically, the interpretation of βj is the expected change in Scorei for a one-unit change in Zi when the other characteristic variables are held fixed.

Using the variables listed above, we were able to create the fol-lowing multiple linear regression to measure the statistical rela-tionship between the literacy benchmark score of the ith student and the unique characteristics of that student in the Springfield School District. We pooled the 2010 and 2011 classes of kinder-


gartners together in order to increase our sample size since both years were tested on the same assessment and would not be much different from each other in terms of either score or char-acteristics.

4.5.1 Springfield School District (sample size=388)

SCOREi = β0 + β1FEMi + β2LUNCHi + β3SPEDi + β4ESLi + β5ETHHISPi + β6ETHBLACKi + β7ETHASIANi +

β8ETHAMERINDi + β9ETHMIXEDi + ui

After correcting for White Standard Error—which assumed that the errors or disturbances in the regression have the same vari-ance (and therefore the same standard deviations) across all ob-servation points—this model allowed us to measure the separate statistical relationship for each characteristic variable we received from the schools for both years (Huber-White n.d.). A statistical relationship is not necessarily causal, so in order to determine causality, a larger, randomized experimental design would be necessary for future analyses.

After analyzing the relationships between each characteristic variable and the pretest score, we wanted to gauge whether the score discrepancies between the Promise Neighborhood schools and non-Promise Neighborhood schools still remained the same when these variables are held constant. We did this by finding the coefficient of PNi in the regression:

SCOREi = β0+ β1PNi + ui

Then comparing this relationship to the coefficient of PNi in the same regression with the characteristic variables controlled for. If the coefficient of PNi was significantly lower after controlling for the other variables, then it can be said that the composition of each neighborhood had more influence on reading readiness than the Promise Neighborhoods—whether it was really the Promise Neighborhoods that were creating a positive change to reading readiness or some other factor such as a higher percent-age of native English speakers in the area.

5. EMPIRICAL ANALYSIS

5.1 Variable Relationships

5.1.1 Springfield School District

Characteristics:From the standard Springfield characteristic regression, four out

of the nine characteristics were statistically significant. The coeffi-cients of LUNCHi, SPEDi, and ETHASIANi were -8.64, -8.28, and +5.21 points on average, respectively, which were all considered significant at the 1 percent statistical significance level. These three coefficients were highly correlated with literacy cores. The coefficient of ESLi was -5.97 points on average, which was sig-nificant at the 5 percent significance level. Though this coefficient was less correlated to scores than the past three, it was still highly correlated with score outcomes. It should be noted that there was only one Asian kindergartner in the Springfield School District in both years, so the coefficient for ETHASIANi does not necessar-ily speak for all Asians or Pacific Islanders. The reference group received 28.17 points on average.

Table 1: Springfield Characteristic Variables


Promise Neighborhood:

In the regression using only PNi, the coefficient for PNi was not statistically significant, but the 95 percent confidence interval was

between -5.34 and 1.66. Confidence intervals consist of a range of values that act as good estimates of the unknown population pa-rameter. The level of confidence of the confidence interval would indicate the probability that the confidence range captures this true population parameter given a distribution of samples. This value is represented by a percentage, so when we say, “we are 99% confident that the true value of the parameter is in our confi-dence interval”, we express that 99% of the observed confidence intervals will hold the true value of the parameter. Keep in mind that in infrequent cases, none of these values may cover the value of the parameter. Using the Promise Neighborhood data, we say that there is a 95 percent chance that the Promise Neighborhoods have an effect from -5.34 to 1.66 (Confidence 1999). Unfortunate-ly, not much can be said about PNi since the coefficient was not

statistically significant and the 95 percent confidence interval was so wide with a range of 7 points.

When we added the characteristic variables back in, PNi was still

not close to being statistically significant, but the 95 percent con-

fidence interval became slightly smaller—between -5.06 and 1.80 with a range of 6.86 points. This led us to believe there were ef-fects that PNi was picking up from characteristics not controlled

Table 2.2: Springfield PNi Relationship Controlling for Variables

Figure 4: Change in Springfield Promise Neighborhood Confidence Interval, as Variables are Held Constant

Figure 3: Statistically Significant Characteristic Relationships in Springfield School District (Reference Group Avg. Score: 28.17)

Table 2.1: Springfield PNi Effect Not Controlling for Variables


for in our regressions, which will be discussed in the next section of this paper.

5.2 Analysis of EstimatesThe most significant and perhaps most interesting finding was that in every regression we performed, LUNCHi and SPEDi were always statistically significant at the 1 percent significance level with a negative coefficient ranging from -7 points to -10 points on average. Considering the reference groups in the Springfield School District, a 10-point decrease is a 37 percent decrease in Springfield schools’ reading readiness scores. This means that it was highly likely that a student in either a special education class or on free or reduced lunch would receive an early literacy score that was 37 percent lower than his or her peers. Since LUNCHi was our proxy for family income, it is a reasonable inference that a child coming from a lower income family income is at higher risk of entering school being unprepared to begin learning. This may be due to a range of factors such as an absence of necessary parental training or intervention, low parental education level, or simply a household lack of disposable income to purchase at-home reading materials or enroll the child in early childhood development programs—all problems that United Way seeks to address through the Promise Neighborhoods.

ESLi, our proxy for English language learners, was statistical-ly significant at the 5 percent level in almost every regression, each time leading to a negative relationship ranging from -5 to -8 points on average. In the Springfield regression, none of the ethnicity variables were negative and statistically significant; therefore there may be more non-English speaking ethnicities in Springfield that we did not control for in our model, such as Vietnamese or Russian. If this is the case, a stronger focus on Eng-lish language learners of all backgrounds throughout the city of Springfield could prove to be a huge impact on literacy rates.

5.3 Data LimitationsWhile the Promise Neighborhood regressions showed negative coefficient possibilities in the 95 percent confidence intervals, as more characteristic variables were added, the range of the confi-dence interval shrank even as it became more positive. If more characteristic variables were added to the regression, the 95 per-cent confidence interval would likely reach a range between two positive numbers in both districts. This would indicate the Prom-ise Neighborhoods likely have a positive, if minor, relationship to fall literacy scores.Another important finding from the analysis of each regression was that the variables we used explained very little, meaning

there were many more characteristic variables we did not control for. Our highest R2 was 0.125 and our lowest was 0.093. With the variety of characteristic variables we had avail-able, we were only able to describe between 9.3 percent and 12.5 percent of the characteristics statistically associated with fall lit-eracy scores. We likely observed this outcome due to our small pool of data and the fact that each school district only tracks a few characteristics for incoming students. We address this prob-lem later in the paper by giving United Way of Lane County sug-gestions on variables to track in the future.

In order for UWLC to conduct more meaningful analyses in the future, either stronger experimental design or better data is re-quired. When dealing with real-world analyses, collecting better data—such as more variables, larger sample sizes, and better-organized lists—is likely the best solution. As more variables are tracked and sample sizes continue to grow each school year, causal links between scores and Promise Neighborhood pro-grams may appear. As stated above, we began to see this slight positive trend in the Promise Neighborhood confidence inter-vals as we controlled for available variables. More years of data combined with more variables being tracked should allow future analysis of the Promise Neighborhoods to better determine their statistical significance with early childhood literacy.

5.4 Policy ImplicationsOne policy we recommend implementing in all Lane County schools is uniform organization and compiling of student data, possibly monitored by the Oregon Department of Education. This will be discussed in further detail shortly.

Using the data available, income and special education had the largest statistical relationship to low fall literacy scores, with non-native English speakers being the next most significant re-lationship. These characteristics should be taken into account as programs in the Promise Neighborhoods seek to decrease the number of high-risk students in all affected schools.

Observing that income is such a significant factor in a child’s abil-ity to be ready to learn by the time they enter school, UWLC could consider focusing a higher percentage of its funding on “Income” projects. Positively affecting income—“moving the needle” as it is often referenced at United Way—would likely affect education by reducing the effects of low-income households on children throughout the county.

UWLC already has efforts underway to help non-native English speakers. They fund parenting education programs and KITS in


Spanish, reaching out to families with young children by providing materials in English and Spanish. As long as these programs are well advertised and provided in English and Spanish, the score discrep-ancies between native and non-native English speakers who speak Spanish should decline. The next step should be also offering other languages found in Lane County, such as German or French.

6.SUGGESTIONS FOR FUTURE TRACKINGAs we compiled data from Springfield School District, the most important suggestion we could offer would be to work toward a uniform data collection program. This would make outcomes from different districts in Lane County easier to analyze. The simplest solution would be for all schools in Lane County—and ideally all schools in the state of Oregon—to decide on EasyCBM, DIBELS, or another program for student assessments. Implementing a uniform testing program would allow for easier school comparisons and cre-ation of a student database that could be easily accessed, allowing for much simpler analysis.

When first obtained, the Springfield data contained holes such as missing gender, missing fall literacy scores, and missing special education data. These holes were inconsistent across schools, likely due to the fact that each school organizes student data differently. Not only is a uniform data collection program important, a uniform data storage system is equally necessary. Whether this is a mutually agreed upon template or a master database to which all schools in Lane County contribute, some safeguard for data uniformity is es-sential for meaningful future analysis at city, county, and state levels.

Another useful characteristic to track would be a universal literacy pre-test given at age 3 or 4. If a majority of Lane County children took a pre-test before enrolling in Promise Neighborhood programs, the effects of those programs and experiences leading up to kinder-garten entrance would be easier to analyze (Walstad 1988). Future analysis using pre-test scores would focus on the positive (or nega-tive) change in reading readiness that Promise Neighborhoods of-fer rather than solely restating the already known fact that Promise Neighborhoods are places in communities with children coming into school unprepared to learn.

In the Promise Neighborhood schools, kindergarten teachers also completed a social-emotional scorecard for each student. The teach-ers gave each student an “emotional difficulties” rating as well as a “pro-social” score. We subtracted the total difficulties score from the pro-social score to create a basic “emotional score” for each student. If this same evaluation was done throughout Lane County schools, further research could look for a correlation between certain vari-ables and social-emotional scores, including between social-emo-tional scores and literacy test scores. This analysis would hopefully emphasize the effectiveness of the holistic approach to school readi-ness that UWLC strives to provide. It could also show the effect of Promise Neighborhoods on a child’s early social abilities.

Another potentially useful characteristic to track would be the num-ber of siblings in each incoming student’s household and whether these siblings are older or younger than the child under study. Nega-tive effects from having too many younger siblings that might dis-tract parental literacy teaching as well as positive effects from having older siblings who can help the kindergartner learn could appear in a future analysis. Asking how many children are in the family along with their ages would provide these useful variables.

7. CONCLUSIONS AND FUTURE RESEARCHThough our analysis led to results we did not predict, our outcome is still of great use to the policy makers of United Way of Lane County. We will continue to work with UWLC to look for ways that help Lane County school districts collect and compile data, allowing for easier economic analysis in the future. If each school district in Lane County were given a uniform way to collect and compile data and advised about which variables to track, programs could be analyzed for efficiency much sooner. This would allow the most affordable and efficient programs to be implemented more quickly.

One of the more significant challenges we had with the data was dealing with selection bias in the Promise Neighborhoods. Selection bias is systematic error due to a non-random sample of a popula-tion, causing some members of the population to be less likely to be included than others, resulting in a biased sample. UWLC placed Promise Neighborhoods in the two lowest scoring schools in each district; therefore these schools already had a predisposition for low literacy rates. Conducting a regression on the effects of the Promise Neighborhoods on early reading readiness simply verified that the Promise Neighborhoods were placed in lower performing neighbor-hoods, which we already knew.

This regression should be run again in the future, when the pool of kindergarten students is much larger and more characteristics are being followed. Following the suggestions for additional charac-teristics to follow, UWLC will have data on pre-test scores, size of household, and student social and emotional scores. Adding these variables to a regression—along with participation in Early child-hood development programs such as KITS and Headstart—should begin to show the true effects of the Promise Neighborhoods and of all other ECD programs on fall literacy scores.

As seen in the literature review, the significant case studies of pro-grams that affect reading readiness were all long-run case studies. Most robust early childhood development analyses contain at mini-mum twenty- and forty-year follow-ups. The benefits of programs funded by United Way, including those in the Promise Neighbor-hoods, should be more robustly observed upon following up after the affected children become adults and begins contributing to soci-ety. Ideally, a randomly selected group of children from the Promise Neighborhoods would be tracked longitudinally throughout their


lives alongside a randomly selected group of students from the con-trol schools. This would allow policy makers and United Way do-nors to observe the long-term effects of strategic investment in early child development programs such as those offered in the Promise Neighborhoods.

One last interesting finding was that fall literacy scores were not as highly correlated to future reading scores as anticipated. There were some students in Bethel who received 0 points on their fall assess-ment but then received 27 points three months later on their winter assessment while a different student who received 30 points on their fall assessment then received 5 points on their winter assessment. This result supports the possibility that improvement and learning may be more important to literacy scores than strict reading readi-ness at the start of kindergarten, or it could also be a data integrity issue. If future regressions could include winter or spring test scores to look for a correlation between reading readiness in the fall and how each student performs throughout the year, this could open up new conversations about educational policies.

REFERENCESBaydar, N., Brooks-Gunn, J., & Furstenberg, F. F. (1994). Early Warning Signs of Functional Illiteracy: Predictors in Childhood and Adolescence. Philadelphia: National Center on Adult Literacy.

Belfield, C. R., Nores, M., Barnett, S., & Schweinhart, L. (2005). The High/Scope Perry Preschool Program: Cost–Benefit Analysis Using Data from the Age-40 Followup. The Journal of Human Resources, 41, 162-190.

Bierman, K. L., Torres, M. M., Domitrovich, C. E., Welsh, J. A., & Gest, S. D. (2009). Behavioral and Cognitive Readiness for School: Cross-domain Associations for Children Attending Head Start. Social Development, 18(2), 305-323.

Confidence Intervals - Statistics Teaching Tools - New York State Depart-ment of Health. (1999). New York State Department of Health. Re-trieved August 14, 2012, from http://www.health.ny.gov/diseases/chron-ic/confint.htm

DIBELS 6th Edition Benchmark Goals. (n.d.). DIBELS Data System. Re-trieved August 12, 2012, from https://dibels.uoregon.edu/benchmark.php

Dickens, W. T., Sawhill, I., & Tebbs, J. (2006). The Effects of Investing in Early Education on Economic Growth. Washington, D.C.: The Brookings Institution.

Dobbie, W., & Fryer, R. G. (2011). Are High-Quality Schools Enough to Increase Achievement Among the Poor? Evidence from the Harlem Children’s Zone. American Economic Journal, 3(3), 158-187.

Heckman, J. J., & Masterov, D. V. (2007). The Productivity Argument for Investing in Young Children. Review of Agricultural Economics, 29(3),

446-493.

Huber-White Standard Errors - Glossary - Dictionary Definition of Hu-ber-White Standard Errors. (n.d.). Economics at About.com. Retrieved August 12, 2012, from http://economics.about.com/library/glossary/bldef-huber-white-standard-errors.htm

Income Eligibility Guidelines. (n.d.). Food & Nutrition Service Home Page. Retrieved August 10, 2012, from http://www.fns.usda.gov/cnd/Governance/notices/iegs/IEGs.htm

Oregon Revised Statutes - 2011 Edition. (2011). Chapter 343 — Special Education and Other Specialized Education Services. Retrieved August 10, 2012, from www.leg.state.or.us/ors/343.html

Promise Neighborhoods. (2010). United Way of Lane County, Oregon | Live United | Non-Profit Organization | Eugene, Springfield, Oregon. Retrieved May 1, 2012, from http://www.unitedwaylane.org/what-we-do/community-impact/education/success-by-6/promise-neighbor-hoods/

Purcell-Gates, V., & Dahl, K. L. (1991). Low-SES Children’s Success and Failure at Early Literacy Learning in Skills-Based Classrooms. Journal of Reading Behavior, 23(1), 1-35.

Richards, D. (n.d.). Effective Teaming: The Hardest Part of an RTI System . EasyCBM. Retrieved August 12, 2012, from http://webcache.google-usercontent.com/search?q=cache:xnxP96jK1nIJ:www.oregonrti.org/files/u2/Parkrose%2520%2520-%25204-16- 12%2520-%2520Teaming.pptx+&cd=1&hl=en&ct=clnk&gl=us&client=firefox-a

Rolnick, A., & Grunewald, R. (2003, December). Early Childhood De-velopment: Economic Development with a High Public Return. The Re-gion, 163, 6-12.

Temple, J. A., & Reynolds, A. J. (2007). Benefits and costs of investments in preschool education: Evidence from the Child–Parent Centers and re-lated programs. Economics of Education Review, 26, 126-144.

Walstad, W. B., & Soper, J. C. (1988). A Report Card on the Economic Literacy of U.S. High School Students. The American Economic Review, 78(2), 251-256 .

der Gaag, J. v., & Tan, J. (1998). The Benefits of Early Child Development Programs: An Economic Analysis. Washington, D.C.: The World Bank

ACKNOWLEDGEMENTSSpecial thanks to Dr. Joe A. Stone and Dr. Bruce Blonigen of the Uni-versity of Oregon Department of Economics; Holly Mar Conte of United Way of Lane County; and Sara Ticer of the Springfield School District. In addition, we thank the staff of The Tower undergraduate research journal for giving us the chance to be published in a great journal.


ARTICLES


B O L D S i g n a l C h a n g e s i n R e s t i n g S t a t e N e t w o r k s a r e R e l a t e d t o Pe r f o r m a n c e o n a V i g i l a n c e Ta s k

M a c M e r r i t t

INTRODUCTIONOne of the principle goals of biomedical research in the past de-cade has been to reverse engineer the human brain. Traditional cognitive research has focused on localizing functions to specific brain regions, and as a result several techniques have been devel-oped to Determine where specific functions are mapped in vari-ous regions of the brain. These techniques include the observation of brain injury patients, functional neuro-imaging, and neuro-stimulation. Functional magnetic resonance imaging (fMRI) is one of the tools that has been crucial for this type of analysis. By taking repeated MRI scans, fMRI produces a time array of images

where the relative contrast in each image is related to the pro-portion of oxygenated blood. Neural activity requires energy, and energy production requires oxygen, so when neural activity rises, oxygenated blood flows to that region in order to meet the need of the cell. These changes in blood oxygenation level can be deceted with spatial resolution as small as a few millimeters.

Although functional localization has provided valuable informa-tion about basic neurological processes, it is now clear that higher level processing relies on the interactions across brain regions (Friston 2011). Biswal, et al. (1995) were the first to characterize the functional relationship between spatially separated regions.

ABSTRACTPurpose and Background: Recent research has shown that spatially separated brain regions often display functional synchrony that relates to brain state and human performance. Two important anti-correlated functional networks that are seen with resting state functional mag-netic resonance imaging (fMRI) are the default mode network (DMN) and the task positive network (TPN). Analytically defining these two networks to better understand their behavior may have a critical impact on understanding higher level function and human performance.

Methods: 17 participants were scanned using fMRI in two different states: while performing the psychomotor vigilance task (PVT) and while in resting state. Using seed based correlation defined networks, the behavior of the TPN and DMN were tested for dynamic behavior after the onset of the task and the shifts in the magnitude of the signal in each network was compared to reaction time on the PVT using a linear regression.

Results and conclusion: The signal in each network changed significantly in response to the task (TPN increased with a peak at 6 sec-onds, DMN decreased with a peak at 12 seconds). The magnitude of the increase in the signal within the TPN was significantly related to response time on the PVT. This study validates a network generation technique that can be used in future studies to further inves-tigate the behavior of functional networks, and it shows a relationship between shifts within the TPN and behavior on this vigilance task.


By selecting a specific seed region and determining the voxels with the greatest correlation to that seed, Biswal discovered that at very low frequencies (0.08-0.1 Hz) spatially separated brain re-gions activate in a temporally synchronous manner. These rest-ing networks were related to behavior through comparison with traditional task based fMRI analysis to show that the regions associated with performance of a specific task were also corre-lated in time during resting state scans. Since the initial study by Biswal, et al., many studies have used seed based correlation as well as other analysis methods to explore functional interactions within the brain. In 2001, Raichle discovered a “default mode network” (DMN) in the brain using functional neuroimaging positron emission tomography (Raichle et al. 2001). The default mode network is one the most prominent functional networks in the brain comprised of the precuneus, angular gyri, and medial prefrontal cortex. The DMN was found to be anti-correlated with the “task positive network” (TPN) which becomes active when people perform cognitively demanding tasks (ie. attention or working memory tasks). Anatomical nodes of the TPN include the dorsal lateral prefrontal cortex, the promoter cortex, and the inferior parietal cortex (Fransson 2005).

The psychomotor vigilance task (PVT) is a reaction time task that is used to assess sustained attention. It is a basic vigilance test that involves fixating on a dot that undergoes a slight shift in color at random times. Participants are instructed to signal as quickly as possible when the dot shifts colors. Within well rest-ed subjects, previous studies have related performance on the PVT to specific brain regions using fMRI and the general linear model. Activation in the regions of the brain associated with the TPN has been shown to relate to faster performance, and activa-tion within the regions of the brain associated with the DMN has been associated with slower performance (Drummond et al. 2005). In this work we expand on previous work by using seed based correlation to define the default mode network and task positive network. We show that both resting-state networks change in response to the onset of a task, and that the extent of this change is related to performance as defined by reaction time on the PVT.

METHODSSeveral healthy individuals were recruited (9 males and 8 fe-males) with ages ranging from 18 to 26. Data was acquired at Georgia Institute of Technology/ Georgia State University Center for Advanced Brain Imaging, and all studies were performed in compliance with the Georgia Institute of Technology Institutional Review Board. fMRI was performed on all 17 individuals using a Siemens Trio 3T whole body MRI scanner (Echo-planar imag-

ing, number of slices= 4, slice thickness = 2mm, repetition time = 300ms, echo time = 30ms) while simultaneously performing a PVT. The simultaneous image and task paradigm was repeated for four fMRI runs. During each run, the subject was instructed to look at a centrally located black dot and respond immediately by pressing a button when the dot changed color to navy blue. Delay time between task onsets were randomly assigned be-tween 10 and 480 seconds. In addition to the four task fMRI runs, each subject participated in two resting state scans where they were instructed to lie still and fixate on a black dot. Participants were informed that the dot would not change color during rest-ing state scans. The overall order of the scans was counter bal-anced between order options to control for time effects (Option 1: Resting State-PVT-PVT-Resting State-PVT or Option 2: PVT-PVT-Resting State-PVT-PVT-Resting State).

Preprocessing is the initial step in fMRI data analysis. Preprocess-ing was conducted using Matlab and the Matlab programs SPM8 and AFNI. During this process, image sequences were aligned with relevant anatomical regions and adjusted to compensate for variance due to time and motion. T1 weighted anatomical images for each subject were segmented into white matter, grey matter, and cerebrospinal fluid maps. Motion parameters and mean signal in white matter were regressed from EPI data. Each voxel’s time course was normalized to mean zero and unit vari-ance. The left precuneus was reverse normalized from the MNI brain to be registered to the functional EPI scans and then later be used as a seed region. EPI sequences were slice time corrected and motion corrected using AFNI. The initial 100 TRs (30 s) were removed to account for stabilization time, and the EPI images were spatially blurred with a Gaussian sigma 2x2x1 voxels and size 3x3x1 voxels. The EPI time sequence was then band pass filtered using a finite impulse response (FIR) filter to 0.01 to 0.08 Hz. Finally, the EPI sequences were normalized and detrended according to Thompson et al. (2012). Scans were excluded from analysis for excessive movement (greater than 3.4 mm) or failure to respond to the task.

In order to generate and analyze the functional networks, the Pearson product-moment correlation coefficient (r) was calcu-lated between the mean normalized time course for a seed region (the left precuneus in the most dorsal image) and each voxel’s normalized time course for the entire fMRI run. The 1,639 vox-els in gray matter most correlated with the seed (1,630 voxels / 10% of all voxels) in gray matter were used to define the DMN, and the 1,639 voxels in gray matter least correlated with the seed were considered the TPN (see fig.2). This network generation technique was chosen because the precuneus is an easily identifi-able region in the default mode, and the task positive mode is


assumed to be anti-correlated with the default mode (Fransson 2005). The mean activity in the DMN and TPN was evaluated in 2 s windows beginning at the onset of the stimulus through 24 s post stimulus. The resulting Blood oxygenation level dependent (BOLD) signal from each time window after the task onset was compared to the mean BOLD signal between 16 and 0 seconds prior to the onset of the task within each network. A linear re-gression comparing BOLD signal change within each network to the response time for that trial was then performed for each time window.

RESULTSA summary of the results can be seen in figure 2. Both networks changed significantly in response to the task onset (alpha value of 0.05 adjusted for multiple comparisons with Sequential Good-ness of Fit, multiple T tests). The mean BOLD signal in the de-fault mode network decreased in magnitude after the onset of the

task. This reduction in signal intensity was significantly non-ze-ro at 8, 10, 12, and 14 seconds after the task peaking at 12 seconds. Figure 1 shows the dynamics of both networks after the onset of the task. Conversely, the mean BOLD signal in the task positive network increased in magnitude after the onset of the task. This increase was significantly positive at 6, 8 and 10 seconds after the task peaking at 6 seconds. Using a linear regression, the change in signal within the task positive network was significantly re-lated to performance in terms of reaction time at time shifts 6 and 8 (p-value<.05).

DISCUSSIONThe present study shows that performance on the PVT is relat-ed to changes within resting state defined networks. This sup-ports previous findings regarding the neural correlates of the PVT, and validates the use of seed based correlation to define the default mode network and task positive network during the PVT. Thompson et al. (2012) used the network defining method presented in this work to relate the correlation between the two networks to trial by trial performance both inter- and intra- indi-vidually. The present analysis provides validity to the network generation technique used in Thompson’s work by showing that the behavior of both functional networks are similar to the re-sponse characteristics seen in networks defined using the Gen-eral Linear Model (Drummond, 2005). Furthermore, the present study contributes to a larger body of research suggesting that resting state functional network analysis is critical for under-standing and predicting human behavior.

An interesting and unexplored result of this study is the differ-ence in the response time for each network. The DMN peak re-sponse was 12 seconds after the onset of the task whereas the

Figure 1: Images of the default mode (left) and the task positve mode (right)for one representative subject. The red highlighted region represents the seed

region.

Figure 2: The BOLD signal shift within each network after the onset of the task.


TPN peak response was only 6 seconds after the onset of the task. While this delay does not appear to be related to performance on the task, it may suggest a different hemodynamic response func-tion of negative DMN and positive TPN, or it may relate to a re-lationship between the spatiotemporal structures of the two net-works. The delay also may be related to the pathway by which the task triggers each network. Further work with greater spatial resolution may be necessary to better understand this phenom-enon.

FUTURE WORKThis study opens the door for two types of follow up work: 1. Work to better understand functional neural networks. 2. Work to modulate functional neural networks in a beneficial way.To better understand the networks, future studies may use more complicated performance metrics and full brain scanning rather than four slice scans to better understand the extent of each net-work spatially, and to what extent each of the networks relates to human behavior. Within this framework, more advanced analy-sis methods may reveal interesting characteristics of resting state networks. For example, several techniques including wavelet analysis and sliding window correlation are being developed in order to characterize the dynamic behavior of functional net-works in time.

Finally future studies may attempt to modulate the behavior of these functional networks with stimulation techniques such as transcranial direct current stimulation (tDCS), transcranial mag-netic stimulation (TMS), or transcranial alternating current stim-ulation (tACS). TDCS and tACS use small amounts of electrical current (<2 mA) in order to alter cortical excitability. Specifically the tDCS animal set up that is currently being developed in the Keilholz lab may be used in future work to understand how brain stimulation alters resting state network behavior.

REFERENCESBiswal B, Yetkin FZ, Haughton VM, Hyde JS. (1995): Functional connec-tivity in the motor cortex of resting human brain using echo-planar MRI. Magn Reson Med 34(4):537-41.

Drummond SP, Bischoff-Grethe A, Dinges DF, Ayalon L, Mednick SC, Meloy MJ. (2005): The neural basis of the psychomotor vigilance task. Sleep 28(9):1059-68.Fransson P. Spontaneous low-frequency BOLD signal fluctuations: An fMRI investigation of the resting-state default mode of brain function hypothesis. Hum Brain Mapp 2005;26:15–29.

Friston K. (2011): Functional and effective connectivity: A Review. Brain Connectivity 1(1):13-36Nitsche M, Cohen L, Wassermann E, Priori A, Lang N, Antal A, Paulus

W, Hummel F, Boggio P, Fregni F, Pascual-Leone A. (2008): Transcranial direct current stimulation: state of the art 2008. Brain Stimulation 1 206-223

Raichle ME, MacLeod AM, Snyder AZ, Powers WJ, Gunsnard DA, Shul-man GL. (2001): A default mode of brain function. Proc Natl Acad Sci U S A 98(2):676-82.

Rossini P, Rossi S, (2007): Transcranial magnetic stimulation diagnostic, therapeutic and research potential. Neurology 68(7) 484-488

Thompson G, Magnuson M, Merritt M, Schwarb H, Pan W, McKinley A, Tripp L, Schumacher E, Keilholz S, (2012) Short time windows of cor-relation between large scale functional brain networks predict vigilance intra-indivdually and inter-individually. Human Brain Mapping (ac-cepted, doi: 10.1002/hbm.22140)

Zaehle T, Rach S, Herrmann C, (2010): Transcranial Alternating Cur-rent Stimulation Enhances Individual Alpha Activity in Human EEG. PlosOne 5(11) ACKNOWLEDGEMENTSThe author would like to thank Garth Thompson and Matthew Magnuson for performing the analysis that motivated this study, Hillary Schwarb for collecting the MRI data, Wen-Ju Pan for his helpful suggestions, as well as Shella Keilholz, Andy Mckinley, and Eric Schumacher for guidance through the research process.


GTTOWER.ORG | 55


GTTOWER.ORG | 57


GTTOWER.ORG | 59


GTTOWER.ORG | 61

Documents

The Tower Undergraduate Research Journal Volume V