An interactive introduction to computational neuroscience

For a few years I’ve given a workshop, a two hour introduction to computational neuroscience, for masters students. This has covered the Hodgkin-Huxley model of action potential generation, reduced models, and simple small-networks. Most of the content was previously matlab code, provided by other researchers, most notably Hugh Wilson and Eugene Izhikevich.

Over the last few months, we (mostly Derek Fulton and Paraskevi Mylona) have converted this matlab code to the Julia programming language. This was partly to evaluate Julia as a replacement for matlab (and the signs so far are very encouraging) and also to evaluate recent cloud-based infrastructure for running notebooks. At the eleventh hour before this year’s workshop I decided to switch to Julia for running the workshop. Thankfully, on the day all the technology worked perfectly.

To try the workshop for yourself, simply visit the following link:

This should launch an interactive Jupyter notebook, where all the graphs can be dynamically regenerated. To try this, first go to the Cell menu -> All Output -> Clear to clear al the items and then Cell -> Run All to re-generate them. (The first time you regenrate, it may take a while as it needs to load various packages.)

I think this technology is a great way of providing resources. Students do not need to load any software onto their local machine; all you need is a modern web browser and an internet connection. No more matlab licenses required! It also introduces students to the notion of reproducible documents: live figures that can be regenerated, rather than static figures.

The tutorial is a bit rough around the edges, and not really appropriate for independent study, but I intend to develop more resources in this direction, and feedback welcome!