Projects During Graduate School:

	Genes Underlying Brain Development This was my first project where I served as a collaborator with the Sheng Laboratory, showing that the gene "Shank" is an important player in the development of brain connections. They had demonstrated that its expression causes the postsynaptic side of connections to grow larger structurally, and I performed functional measurements which found that the conductance across the synapses were stronger as a result. Shank's most interesting effect is in recruiting together several molecular parts that are vital for the function of synaptic connections, and activating signals to trigger the construction of more distant parts of the synapse. The work was published in the journal Neuron.
	A New Mechanism for Strengthening Excitatory Connections This work, conducted in the Liu Laboratory, provided a biophysical description of a new way to influence the strength of excitatory connections. Using electrophysiological and molecular techniques, I showed that the filling of synaptic vesicles by the glutamate transporter "VGLUT1" controls the amount of transmitter released, and the consequent electrophysiological amplitude of the synapse's "quantal size". We also demonstrated that this mechanism for controlling the strength of a synapse from the "sending side" responds dynamically to changes in activity levels, and thus could adjust synapses to accommodate activity in the network. Details of the work are reported in The Journal of Neuroscience.
	A Relationship Between the Number of Synapses and their Strengths in Hippocampal Neurons Here I worked on a "micropatterning" system that would allow us to grow cultures of brain circuits with finely controlled geometry and connectivity designed by a computer. I then applied intracellular electrophysiology and confocal imaging to measure the strength, pattern, and molecular makeup of synaptic connections as the size of networks was geometrically scaled from small to large. We found that when the size of a network is scaled artificially, the neurons in that network make connectivity tradeoffs, either spreading their connections out among many partners, or focusing them narrowly with a few strong partners. A summary of this work can be read in the Journal of Neuroscience.
	Observation and Stimulation of Population Activity in Neuronal Cultures An ongoing side-project with the Sur Laboratory throughout my graduate studies (or distraction, as some others called it) involved building an experimental system to stimulate multiple inputs converging on a single cell. I took glass chips with embedded microelectrodes, and used digital logic to route analog signals to multiple neurons while recording from a single neuron. The work contributed to a patent and produced some interesting possibilities that I am now hoping to make use of in my postdoctoral work, permitting me to examine interactions between synapses in controlled ways.

Undergraduate Projects:

	Computational Models of How People Hear Working with the Bass Lab, I developed a neuronal modeling system for biophysically simulating single compartmentalized neurons on the Cornell supercomputer. The research question was to understand how neurons might biophysically respond to and distinguish between multiple sounds colliding on one ear. The system evolved model neurons that behaved similarly to real ones we recorded electrophysiologically from a teleost fish. A summary of this work appeared in the Journal of Computational Neuroscience.
	My First Data Acquisition System Working with the Walcott Lab, this was the first (simple) electrophysiology system I built, to allow us to inexpensively record electrical impulses under the control of a computer and user-friendly software. The software was written for an A/D card I found lying around but was later compatible more generically. It was definitely harder with previous technology than it would be now! It provides a graphical interface and control modules to allow automated analysis of input, as well as customizable delivery of stimulation. The system now has some built-in analyses for single-cell recording, and can synchronize sensory stimulation with several channels of recorded neural activity.
	A Technology for Web-Based Experiment and Simulation Working with the Cornell Neuromuscular Biomechanics Lab in the early days of the web, I took the popular simulation environment Mathematica and wrote code building a general tunnel to it from HTML. As a result any Mathematica simulation could be run interactively in a web page. Similarly, I showed how web servers can be used to direct experiments, going on in many parts of the country, while keeping results and analysis centralized. These foundations culminated in a website, featuring a complete biomechanical simulation for the human index finger.
	Search for a Built-In Magnetic Compass Working with the Walcott Lab, I conducted behavioral and electrophysiological experiments to figure out how some animals are able to perceive the Earth's magnetic field and use it like a compass. Using the honeybee Apis melifera, I built an artificial environment that could divert bees coming out of their hive into a magnetic stimulating chamber. I equipped the "disorientation" chamber with photosensors to observe the bees' movements and programs to analyze on-the-fly whether any applied fields were disrupting their orientation. I also tried poking around the bee's nervous system to try to isolate the compass structurally, but never found it.
	Parallel Methods for Searching Complex Spaces This work during my Master's degree was to find time-efficient ways to search complex systems, since my interest was how to optimize parameters of neurons and networks to emulate empirical input-output relationships. I wrote programs that ran simplex minimization, simulated annealing and genetic algorithms on parallel nodes of supercomputers to provide a fast and coordinated search of non-linear spaces. It was exciting to watch the system get better at navigating nasty terrain, and the principles I learned relating to multi-dimensional search were definitely instructive for how to reduce empirical questions to critical parameters now.