Saturday 22 August 2015

Bridging the Gap between Behavioural and Neural Data Streams

Last week, PhD candidate Pete Cassey from the Newcastle Cognition Lab submitted his thesis (yay!). A focus of his thesis was linking behavioural and neural data, which is an effective way to advance cognitive-neuroscientific theory. Pete’s thesis contains multiple projects using two different linking approaches. Qualitative linking approaches involve fitting a cognitive model to behavioural data and then, based on the results of this model fitting, make predictions about the nature of the neural data. Pete used such linking approaches to explore the neural analogues of specific mechanisms of a cognitive decision making model. Also, in a clinical application, this approach was used to uncover latent level mechanisms involved in individuals’ suffering from Major Depressive disorder inability to disengage from negative emotional stimuli.

While qualitative linking approaches are currently standard, across the field, quantitative linking approaches are relatively new in cognitive neruoscience. With his supervisor, Scott Brown, and collaborators, Garren Gaut and Mark Steyvers (UC Irvine), Pete developed a novel joint modelling framework which tightly (quantitatively) links behavioural and neural data streams. This framework allows both data streams to be simultaneously addressed within the one modelling framework. This forces much tighter constraints on the type of relationship(s) that can exist between the two streams, allowing for more explicit tests of linking assumptions.

Pete is moving to Nashville to take up a postdoc position with Gordon Logan and Geoff Woodman at Vanderbilt University – as well as pursuing his dream of becoming a country music star. The postdoc will extend on the work of Pete’s thesis, exploring new ways of linking behavioural and neural data streams. Good luck, Pete!