BCS 513: Introduction to fMRI: Imaging, computational analysis and neural representations

The core focus of the course will be on how fMRI can be used to ask questions about neural representations and cognitive and perceptual information processing.

Some of the questions that the course will address include:
- The basic fMRI signal just shows activation in different parts of the brain. How can we get from that to addressing questions about neural representations and neural information processing?
- Ways of relating neural activation to behavioural performance. Can fMRI provide information over and above what can be obtained from behaviour alone?
- Standard fMRI analysis using the General Linear Model, including preprocessing steps
- Multivariate fMRI analysis using machine learning approaches

There will also be a component, about 20% of the class, on the big-picture aspects of MRI physics and physiology which make fMRI possible:
- What is an MRI scan actually showing? What is being measured?
- When neurons become active and use more oxygen, how does that end up making a signal that can be measured with a magnet?
- What are the factors which give rise to a good fMRI signal? What sources of noise are there? To what extent can they be overcome?

Practical components:
The course will include some practical hands-on aspects: computer sessions involving doing data analysis.

Evaluation:
This will be through a mix of written assignments, computational problem-sets, and a class project. The class project will be on a topic of your choice, decided in consultation with the instructor. Example possible projects include conducting an analysis of a publicly available fMRI data-set, or writing a "grant proposal"-esque plan for an fMRI study and analysis to address a specific neuroscientific question.

Prerequisites:
There are no absolute prerequisites. This course is open to any graduate student or advanced undergraduate who is interested. The computational aspects of the class will involve using Matlab, but if you have no prior programming experience then these should still be accessible, although they will be a bit more work.