Skip to main content
[an error occurred while processing this directive]

[an error occurred while processing this directive]
HomePrograms >

2012 Neukom Scholars

Winners of the Neukom Scholars Program have been announced with awards of up to $1,000 per term for up to two terms.

The programs goal is to fund undergraduate students engaged in faculty-advised research in the development of novel computational techniques or the application of computational methods to problems in the Sciences, Social Sciences, Humanities, and the Arts.

Third and fourth year students are eligible to apply for these competitive grants. Winners for Summer, Fall and Winter terms are:

The Eyes are the Window to the Soul

Troy Dildine '13

Thalia Wheatley, Advisor (Psychological and Brain Sciences)

The proverb, "The eyes are the window to the soul" is an explicit stating of something we seem to know intuitively: looking at someone's eyes plays an important roll in how we communicate with them. However, research on the perception of eyes is still fairly new. Neuroimaging has shown that one area, the posterior superior temporal sulcus (pSTS) is particularly sensitive to eye motion and is activated when viewing shifts of eye gaze. However, pSTS is not typically implicated in the processing of mental states. Instead, the medial prefrontal cortex (mPFC) and right temporo-parietal junction have been shown to be active for tasks that involve identifying the mental states of others; yet they are not implicated in the processing of eye movements. This apparent double dissociation between the processing of eye movements and the processing of mental states is at odds with behavioral work showing that the eye region and eye movement are particularly important for understanding the mental states of others.

Our goal is to examine the neural substrates of how humans are able to determine whether or not they are viewing the eyes of a person who is also engaged. To test this, we will show 30 Dartmouth undergraduates dynamic videos of human eyes that were previously recorded. These videos will be taken from an earlier study that recorded the eye region of the face while people did two things: paid attention to an interesting external stimulus (engaged eyes) and did a boring task where they were free to mind-wander (unengaged eyes). Subjects will view engaged/unengaged eye videos as well as control videos of moving objects. This will allow for direct comparisons between brain responses for the recognition of social engagement and social disconnect. This will bring to light important information about how the brain processes social interaction and recognizes minds.

A New Scientific Programming Language and Improving the Concept of Programming Library

Michael Edgar'10

Prof. Bill McKeeman, Advisor (Computer Science)

Our summer project is the implementation of a new scientific programming language. The language objectives are usability and performance. The research objectives also include improving the concept of programming library, introducing declaration-free strong typing, simplifying objects, simplifying and extending the range of numerical values, and extending the imperative model to include functions as first-class values. A novel form of overloading allows for a natural approach to function extension, simplifying and generalizing dispatch beyond nominal types.

One specific goal is the incorporation of (1979) Hehner & Horspool algorithms for rational arithmetic, based on the p-adic number system. This notation has many attractive properties, including simple arithmetic algorithms closely mirroring the 2's-complement operations implemented in computer hardware. It also features a variable-length representation with compact representations for integers and many (but not all) rationales. However, this representation has, to our knowledge, no implementations exist in either software or hardware.

Understanding Water Quality Stressor Interactions

Emily Kyker-Snowman,'11

Celia Chen, Advisor (Biology)

Current environmental regulations for water quality generally use individual toxin concentrations as a measure of toxicity, ignoring significant interactions between toxins and other stressors. In reality, anthropogenic and natural factors interact in complex ways, and the effects of multiple stressors in combination can be greater or less than the sum of the effects of individual stressors (Folt et al. 1999). Given that some stressors have much more detrimental impacts upon aquatic communities when combined than they do alone (Heugens 2003), it is crucial that regulations take into account multiple stressor effects, and therefore that the academic community develops an understanding of the ways these stressors interact.

My work with the Neukom Institute would start by modifying previous programs and adding my own code in R to create a program that can properly handle these statistical methods as they apply to aquatic stressors. With effective and efficient software, I would be able to analyze new data and previously gathered data from my advisor's lab, test the ability of these new methods to handle data from aquatic systems, and begin building the framework for a database of information about multiple stressor interactions. My hope would be to make the program open-source and easy for other researchers to use so that the application of these statistical methods could become commonplace.

N400 Research in Dartmouth College's Reading Brains Lab

Anna Roth, '13

Dr. Donna Coch, Advisor (Department of Education)

This research project uses Event-Related Potential (ERP) methodology to investigate reading acquisition, especially as it affects late-elementary school children. The project revolves primarily around collecting high-quality data from first an adult control group, then from a representative sample of local third, fourth and fifth graders. By assessing how adult readers process visual input, and comparing this to how the developing readers process input, the project will learn when mature processing begins to develop and what this development looks like. Perhaps the findings of this project will better inform late elementary school curriculum, making it more sensitive to children's cognitive abilities.

Much control subject testing on college-aged participants in the lab has already been conducted, something I worked on much of last year. Testing on developing readers has commenced, and will continue throughout the next year with a goal N=24 students each from grades three, four, and five. Over the summer, I hope to work one-on-one with Dr. Coch to learn how to run these younger subjects and analyze their data as opposed to the data of college students. My proposed role in this project consists of gathering behavioral and ERP data from college-aged and young subjects and processing this data by analyzing behavioral tests and creating ERP from EEG input through computational analysis. Working in the lab last year, I learned how to operate ERP-related technical equipment, from setting up a subject run to performing follow-up analysis on resultant data. This year, I will be able to become more autonomous and complete the different levels of data analysis independently, as well as learning to run ANOVA statistical analysis. This project successfully integrates computational analysis as a crucial tool to inform educational policy, and it would be an honor to continue working on it.

Emotodrome Project

Luke Fowlie, '13

Hany Farid, Advisor (Computer Science)

We are building an iOS app that will allow worldwide users to each create a unique, evolving pattern-shape that reflects and reacts to their movements. A scalable view, from global to personal, will allow users to see all logged-in participants within their chosen view scale. In the graphical view, pattern-shapes will evolve in a 3D environment through which viewers can move by tapping, tilting and moving the personal access device. The project will also track some global ice, through arctic sensors that send real-time data streams. The ice will also create pattern-objects within the piece. For sound, ice data streams will be fed through an emotive classifier, which will determine the human-scale emotive equivalent of their wave pattern. After classification, the data stream will be encoded with human vocal formants appropriate to its class, which will be produced as sound, letting the ice speak.

High Energy Eigenfunctions of the Helmholtz Equation on "Chaotic" Domains

Kyle Konrad '12

Alex Barnett, Advisor (Mathematics)

My research explores statistics of nodal domains in high energy eigenfunctions of the Helmholtz equation on "chaotic" domains. Chaotic domains have non-integrable dynamics of a point particle ("billiard ball") bouncing inside the domain. These high energy eigenfunctions (Fig. 1) arise as vibrational modes of a drum head as well as solutions of the Schrodinger equation for a particle in a "box." They have applications in acoustics, engineering, data analysis, and address fundamental questions in the mathematical physics area of "quantum chaos." For nonintegrable domains we must deal with boundary conditions for which we cannot find an exact solution to the Helmholtz equation. These types of boundary conditions are interesting because they produce highly complex eigenfunctions at high energies.

The goal of my research is to validate these conjectures using high-precision numerical techniques. Vergini has devised an efficient way to compute solutions of the Helmholtz equation over non-integrable domains which has been implemented in C by Alex Barnett. Since these eigenfunctions have very fine spatial oscillations at high energies, the cost of computing these solutions rapidly increases with higher resolution. Thus we are constrained to work with solutions on a coarsely sampled grid. Using only coarse sampling can result in situations when the boundaries of nodal domains become ambiguous (fig. 2). In order to resolve these ambiguities we will adaptively interpolate the numerical solutions around such regions. This interpolation will be done using an expansion in Bessel functions, which span the solution space of the Helmholtz equation.

Wear and Stress Properties of Knee Implants

Natalie Burkhard '12

Professor Douglas Van Citters, Advisor (Engineering Sciences)

Total knee arthroplasty (TKA) is a clinically successful solution for diminishing pain and restoring mobility of the knee1 and treating advanced osteoarthritis 2,3,4. There are more than 500,000 TKA procedures in the United States annually with a five-year survivorship of 97.2%5,6. However, failed implants pose a significant economic and physical burden that is likely to intensify: the number of patients requiring TKA and revision TKA is projected to grow by 673% and 601% respectively by 20307. Mechanical failure/wear is the second most common knee implant complication, accounting for 16.1% of all TKA revision surgeries8. The articular surfaces of the knee implant deteriorate, causing severe pain, limitation or loss of function, and/or deformity of the joint9. The tibial component's artificial bearing surface is made of ultra-high molecular weight polyethylene (UHMWPE) and must withstand large forces while allowing wide-range mobility for rolling, gliding, and rotation10. The repetitive sliding of the UHMWPE against the femoral component's cobalt-chrome alloy during cyclic loading eventually results in polyethylene breakdown11. High levels of cyclic stress in tibial components frequently exceeds the fatigue limit and yield stress of UHMWPE and is the major cause of early failure12.

Development of a Particle-in-Cell Computer Model for
Turbulent Heating in Plasmas

Manaure Francisquez '11

Barrett Rogers, Advisor (Physics and Astronomy)

World population is rapidly increasing, and with it the demand for energy. By 2050 the expected increase in population to nine billion people may increase that demand threefold.1 The current trend of a world energy sector consisting of 80% fossil fuels will prove environmentally and financially devastating. No single technology will be able to meet the energy demand that we will face in the future, which will reach an astonishing 114.3 quadrillion Btu by 2035 in the U.S. alone.2 Our economies will need to strike the balance between fossil fuels, renewables, nuclear fission and hopefully nuclear fusion. Nuclear fusion, the process by which the Sun is able to generate energy occurs when atomic nuclei collide together and release energy. The current approach to achieve fusion energy is through magnetic confinement. In this scheme a hot plasma3 (~108 degrees) is confined in a toroidal vessel called a tokamak where tritium and deuterium (hydrogen isotopes) collide and recombine to yield helium and an energetic neutron4, compared to radioactive uranium and plutonium produced in nuclear fission.

This project is a colossal endeavor that will begin this spring and continue through summer 2012, the two terms we seek your support for. The first four weeks of spring will be dedicated to literature review, mastering the mathematics of gyrokinetics and understanding the structures necessary to implement a second level parallelism. Weeks five to seven will be dedicated to algorithm design, where professor Barrett Rogers and I will exhaustively discuss CPU load-balancing and how to diagnose turbulence and inhomogeneous heating in our code. Programming will begin towards the end of the spring, during week eight, and will carry on for much of the summer and fall 2012. The objective is to begin statistical validation of our code by October and have a working version of the code in the first quarter of 2013.

Biomechanics of Vertical Climbing in Humans and
Material Properties of Bone Daggers from
Indigenous Populations of Papua New Guinea

W. Kyle Heppenstall '13

Nathaniel Dominy, Advisor (Anthropology)

In July of 2011, I was awarded a grant from the Goodman Scholars program in the anthropology department to travel to Hayward, Wisconsin to collect energetics data on competitive pole climbers (who use a rope wrapped around the pole and special toe gaffes to race up a 60 or 90 foot wooden pole) at the Lumberjack World Championships. We used an Oxycon Mobile System – a type of portable cardiopulmonary stress testing system – to determine the breath‐by‐breath metabolic response of pole climbers during walking, jogging, and climbing. For the seven pole climbers that conducted trials with us, we monitored their heart rates, the volume of oxygen consumed, and the volume of carbon dioxide exhaled. This spring, I will analyze the respirometry data (taken every five seconds by the Oxycon

Mobile) to determine the relative metabolic cost of the locomotor behaviors we studied. Although the relationship between volume of oxygen consumed and energy used is well established for aerobic exercise, it is not well established for part aerobic, part anaerobic exercises like climbing. Therefore, I plan to use a quantitative model based on 02/CO2 ratios to accurately determine the energetic cost of pole climbing. Though it is not surprising that vertical pole climbing is an energetically costly activity, our quantitative results will allow comparison to other energetically costly activities like sprinting or swimming that are part of the human locomotor repertoire.

The second part of my proposed research is to continue a project determining the material properties of Sepik bone daggers. In the indigenous populations of the Sepik River region in Papua New Guinea, bone daggers were essential tools in hunting and warfare. These bone daggers were either made from human femurs or cassowary (a large flightless bird, genus Casuarius) tibias. The goal of this project is to determine how resilient to deformation each type of dagger is under a variety of different forces. Using the daggers from the Hood Museum, we took CT scans of 10 different daggers (5 of each type) at the Dartmouth‐Hitchcock Medical Center in the fall of 2011. Using the CT scans in conjunction with the software packages Mimics and Solidworks, I will perform Finite Element Analysis – a computational technique to determine how a material deforms under different forces– to reconstruct models of each type of bone dagger. Using the models, I will simulate different forces on the daggers to quantify how much compressional, tensile, or torsional force the bone dagger can withstand along any axis. From there, I can make conclusions about which type of bone dagger is more resilient to deformation, and likely more valuable in warfare.

Labov's Three Classic Principles of Language and Gender

Stephanie Pignatiello '12

James Stanford, Advisor (Linguistics & Cognitive Science Department)

Labov's (1990, 2001) three classic principles of language and gender (cited in Wolfram & Schilling-Estes 2006:237) are that (1) for stable sociolinguistic variables, men use a higher frequency of nonstandard forms than women, (2) in change from above (the level of consciousness), women favor incoming prestige forms more than men, and (3) in change from below, women are most often the innovators in language change.
Empirical sociolinguistic research consistently finds that Labov's three principles are generally accurate in a wide range of different societies around the world. Many possibilities may underlie these results, including the influence of traditional gender roles, lack of power in society, women taking the responsibility to serve as cultural and linguistic models for their children, the "boys will be boys" attitude that men should be "tough" and therefore should rebel social rules such as standard pronunciations, whereas women may be expected to be "cultural reproducers" and "symbolic border guards," etc. (Yuval-Davis 1997:23, 37).

In the real world, it is difficult to separate such sociocultural/historical factors from biological factors. As a result, it is still not known whether Labov's three principles are the simple result of basic social roles and interactions, or whether there could be more fundamental biological or physiological factors at work. My project will use agent-based modeling to isolate the social factors from any biological factors. The study will use Java to program a model of interactions between "gendered" agents with different dialect features. This type of multi-agent modeling simulates a world according to certain basic parameters as defined by the researchers, and has been used for previous sociolinguistic research. Fagyal et al. (2010) used agent-based modeling to examine the effects of social networks on language change. Stanford & Kenny (2012) and Kenny & Stanford (2011) used agent-based modeling to examine new aspects of the transmission and diffusion of dialects, a project which was supported by the Neukom Institute.

The agents are not biologically sexed, of course. Instead, we will give them particular gender roles in the virtual world. "Female" agents will be assigned the stereotypical roles of child-raising, etc. "Male" agents will "work outside the home" more than female agents do. We will also include economic and social prestige incentives for agents as they interact, and this will model the power differential in the real-world gender hierarchy. In this way, this type of world makes it possible to test Labov's gender principles more directly without the complex cultural/historical influences that complicate any real-world society.


Last Updated: 7/28/15