On your first morning you will be met by Sarah and/or one of the administration team either at the ICN, MPC or Bedford Way, depending on where your desk will be based. They will arrange for you to have a building induction, introduce you to key people, get your access card and show you to your desk and computer. There are a few things you need to do within your first few weeks for both UCL and the lab, see below for a checklist. If there is anything you seem to be missing, please ask Sarah, Steve or senior lab members and we’ll point you in the right direction.
Make sure you have been added to:
Make a start on the key readings (see next)
The work we do is strongly influenced by computational approaches to the brain and mind. As such, it’s useful to at least familiarise yourself with some of the common modeling tools and frameworks. Not all of these will be relevant to your project, but it’s likely that they will touch upon shared assumptions. In particular, Bayesian modeling of stochastic percepts, decisions and confidence estimates is a common theme, so investment here will pay dividends. A useful starting point for these topics is the Neuromatch Academy online course materials, in particular, Linear Algebra, Statistics and Bayesian Modeling book by Ma, Kording and Goldreich. The Neuromatch NeuroAI course also now has a section on computational approaches to consciousness.
Sam Gershman has collected a wonderful set of resources for deepening your background knowledge of cognitive science here.
We aim to take a thoughtful approach to statistical modeling and data analysis, rather than just using tools “out of the box”. Many of these approaches are grounded in the principles of generative modeling and (typically Bayesian) inference. Richard McElreath’s Statistical Rethinking is an excellent (open-access) starting point.
For an introduction to fMRI analysis, a good starting point is the Handbook of Functional MRI Analysis by Poldrack, Mumford and Nichols. This will provide the core concepts that are a feature of most analysis pipelines and packages such as SPM.
For most cognitive computational neuroscience, being able to program is also important. We use a range of programming languages for different projects (Matlab, Python, R, Julia), and try to use free open-source tools where we can. That said, Matlab is still a workhorse for many of our projects, particularly neuroimaging data analysis using SPM. Fluency in at least one core language that you can use for analysis, simulation and plotting is essential. We also often use probabilistic programming languages (STAN, JAGS, Turing) for model fitting - again, fluency in one language often means you can easily switch between different packages.
To build a common knowledge base for everyone in the lab, we recommend reading the following theoretical / review papers:
In addition, please check out our papers published in the past 12-18 months, particularly preprints, to get a sense of what we’re actively pursuing. These are all available on our Publications page.
For more general overviews and informal discussions, you can download podcasts and interviews that Steve and other lab members have participated in over the years on our Media page.