Neurocognitive networks for memory and perception
We use novel neuroimaging techniques to investigate how different brain networks relate to different aspects of memory and perception. By using diffusion MRI, for example, we can measure the properties of white matter tracts that connect different brain regions (the brain's 'highways'). Our work has shown that specific white matter networks seem to support different forms of information, rather than specific cognitive processes (such as memory, perception, and so on). For instance, the white matter tract connecting the hippocampus and prefrontal cortex, called the fornix, seems to support spatial representation during visual perception, autobiographical memory and navigation. We are now beginning to look at these networks in much larger groups of participants, and see how they develop across the lifespan.
To learn more, check out our recent blog in The Conversation.
'Zooming in' on the hippocampus
In recent years, studies have shown that the human hippocampus may support functions beyond long-term memory, particularly when there is a requirement to construct internal representations of spatial environments. The hippocampus, however, is not a single functional 'unit' but a complex patchwork of different subregions. In collaboration with Kim Graham, Andrew Lawrence and Slawomir Kusmia (Cardiff University), we are now using ultra-high-field MRI to explore how these smaller subregions contribute to different aspects of memory and perception.
What makes things similar?
Similarity allows us to form categories and generalise behaviours to novel situations and contexts. Classically, the similarity between objects and stimuli in the world has been conceptualised by the number of shared features, or via distance within a multidimensional representational space. While these theories have been adequate in many contexts, they are inherently limited in representing structural information - i.e., features and their interrelations. With Ulrike Hahn (Birkbeck, University of London), I have been applying and developing novel models of similarity based on transformation distance that can better capture structural information in human observers, but also in nonhuman species.