WHAT IS COGNITIVE AND COMPUTATIONAL NEUROSCIENCE?

Cognition is broadly defined to include a wide spectrum of faculties that we all take for granted as part of our mental toolkit: learning, decision-making, attention, reasoning, memory, language, and motor control... to name a few. Appropriately, cognitive neuroscience is the subfield of neurobiology charged with elucidating the neurobiological underpinnings of these faculties. 

Computational Neuroscience, on the other hand, is a subfield of neurobiology in which mathematical tools are used to develop and test theories of brain function. This commonly incorporates diverse approaches from computer science, physics, psychology, engineering, and mathematics, in order to understand how the nervous system performs its computations, and why. 

Cognitive and computational neurosciences constitute a major research theme here at DRCMR. Our long-term vision is to pioneer new methodologies for bridging between computational modelling of cognition and neuroimaging and to use this to understand brain function in health and disease. The key among these methodological efforts is to develop advanced methods for fitting computational models directly to neural data, allowing us to map how different computational and cognitive variables are encoded and patterned over the anatomy of the brain. 

The cognitive and computational research theme comprises a cluster of five different research groups each asking different types of question. We meet as a group once every six weeks to exchange ideas and figure out ways of working together. We run a week-long PhD school “Cognitive and Computational Neuroimaging” approximately every three years.

DECISION-MAKING AND REWARD

There is an active community of scientists working on decision making and reward within this research theme. Each are asking different questions. How do we make choices in risky settings? Can we change our willingness to take risks and how? How do we make decisions in sequential settings, where we have to make many decisions that may interact over time? How do we find rewards in a complex environment? do we forage in ways similar to other animals? How does our decision making change when we up the stakes? What if we could lose lots of money? How do rewards depend on our homeostatic state? These questions are formalised through economic, cognitive, and computational models, the predictions of which are incorporated into the analysis of EEG and fMRI data. With the encoding of computational variables mapped in healthy brains, these studies are being extended to clinical populations on and off medication, to infer the impact of pathologies on these processes. 

The Computational Neuroscience of Reward group seeks to build fundamental theories of reward value that are grounded in our physiology, our evolutionary history, and even abstract mathematical principles. 

The group collects a diversity behavioral data (choices, physical effort, ratings…), physiological data (metabolic states, endocrine state) and neural data (fMRI, structural MRI, diffusion) and relates these to theoretically inspired questions using computational and cognitive models. Through collaboration these models can be tested in animal behavior also.    

SENSORY SYSTEMS

Also with the research theme, we have a number of sensory neuroscientists engaged in research on the neural basis of audition, using detailed anatomical functional mapping procedures and multivariate computational methods.