Saskia.Frisby@mrc-cbu.cam.ac.uk
01223 731874
Research summary for everyone
When we see a tiger, read the word “tiger” or listen to the word “tiger”, how do we know that those different sensory inputs refer to the same thing? My research is about the code used by the brain to link those pieces of information together. To investigate this code, I analyse signals from electrodes placed right onto the surface of people’s brains (they are put there to help doctors plan surgery to cure those people’s epilepsy, but the data gathered are useful to researchers too). We would like to be able to measure those signals in healthy people as well as patients, but some research suggests that this type of brain code might be difficult to detect with scanners. I am therefore working to improve scanning techniques to make this code easier for us to study. I then hope to be able to study neural codes non-invasively in healthy people.
Research summary for specialists
My PhD project has two strands. The first is characterisation of the neural code used to represent semantic information. Initial evidence suggests that this information is represented in the anterior temporal lobe as neural patterns that are distributed across space and that change rapidly in time during processing. During my PhD I will use electrocorticography (ECoG) data to test and explore this hypothesis, and ultra-high-field functional magnetic resonance imaging (7T-fMRI) to investigate whether these codes are still visible when spatial resolution is sacrificed.
The second strand is the development of better research methods for revealing distributed, dynamic codes. For example, magnetic field inhomogeneities in the anterior temporal lobe mean that the region is plagued by “drop-out” and distortion problems when standard fMRI protocols are used; these issues must be addressed before 7T-fMRI can be used to replicate results from ECoG. With the Knowledge and Concepts Lab at the University of Wisconsin-Madison, I am also developing a new analysis technique, Representational Similarity Learning, to improve our ability to study neural codes of different kinds.