Former PhD Student
I am a Ph.D student at the ANC as a member of the DTC since 2008. My background is in electronic engineering, but I have always been interested in the brain. I am ultimately interested in how the human brain can give rise to 'consciousness'. This is a tricky question, and a convenient stepping stone is understanding how the nervous systems of simpler creatures work. In particular, locomotor circuits are attractive, because we know the output is the motor system.
I am working in collaboration with Prof. Alan Robert's Xenopus tadpole experimental lab in Bristol. I am keen to understand how a relatively small number of neurons in the spinal cord of Xenopus laevis tadpole can lead to different locomotive behaviours of swimming and struggling, and what the crucial features of such a network are. In particular I am currently investigating a type of interneuron known as a dIN; which are thought to be important because they are believed to synchronise the firing of neurons during swimming cycles.
The tadpole spinal cord is composed of rows of classes of neurons. Their interconnectivity and individual firing responses to injected current are thought to be critical to network function. To investigate this further, I am currently investigating the roles of different channel densities on the firing properties of the classes of neurons. To do this I build Hodgkin-Huxley style models of the individual channels, and then run large parameter sweeps in which I vary the densities of the various channels.
My supervisors are Prof. Alan Roberts (Bristol) and Prof. David Willshaw (Edinburgh).
I am also very interested in the roles of computer science in neuroscience.
* How quickly can we convert our hypotheses about how things work into simulations.
* How do produce reproducable research - how hard is it for someone else to try out my modelling.
* How do we manage large datasets - as we develop better and better methods for acquiring data, this will become harder!
* Standards - how to we share data with our collaborators?
As researchers in computational neuroscience we are very dependant on the tools we use; I am interested in the development of well designed, open, extensible tools and standards. In modelling, a small typo can lead to completely incorrect results; and it is currently time-consuming to reimplement other peoples models to check them. As more and more experimental results becomes available, I feel data management and reproducible analysis are areas that will become increasingly important.
I work mainly in Python and NEURON.
I enjoy tutoring, and have tutored various courses while I have been a post-graduate at Edinburgh (OO Programming, Computation & Logic, Software Eng, Haskell, Neural Computation, Data & Analysis).
I am currently a teaching assistant for the first-year OO-Programming Course, in which we have developed a system, `infandango`, which allows students to upload their Java and have it marked online.
With another post-graduate student, I also run another course, "Introduction to Research Computing", in which we talk about practical computer science tools and techniques, such as version control, dependancy management, data management and software development, for students involved in computer modelling who are not from a computer science background.
With some colleagues, I also run a Python discussion group, `WhyPy`, we meet up once a week to discuss ideas, libraries and useful Python techniques we have come across.
Configurations of neurons in networks are able to create specific behaviours
Related Publications and Presentations
and Alan Roberts,
"A roles for ratios? How ion channel densities could define neuronal firing properties",
Society for Neuroscience (SfN), 2010, 46.8/H58(Sat-PM), 605.
Matthew P Down,
James PJ Withers,
and Michael Hull,
"Algorithms for automatic neuron tracing in noisy 3D image stacks",
Neuroscience Day, Edinburgh, 2010.