WIP Special & Methods

WIP: How might humans and artificial agents flexibly perform multiple cognitive tasks?

Authors: Holly Haines, Laurence Hunt, Mark Woolrich 

Presented by Holly Haines

Abstract: Humans encounter a diverse array of cognitive tasks in the environment. Our ability to flexibly perform this myriad of tasks and leverage learned skills to rapidly solve novel tasks is a remarkable, yet poorly understood, hallmark of human behaviour. Owing to its dramatic expansion during human evolution and recruitment across many different cognitive tasks, prefrontal cortex (PFC) is likely a key contributor to multitask behaviour. Specifically, PFC is thought to implement and manipulate task representations in a way that influences interactions between distributed networks of brain regions to achieve the sensory-to-motor mappings required for task performance. Existing work in humans and artificial neural networks indicates that compositional task representations (i.e., consisting of reusable task components) may be particularly useful when recombined to solve different learned, and even previously unseen, tasks. This study will investigate neural task representation in humans learning to perform a large, compositional multitask set across multiple MEG scanning sessions. Representational similarity analysis methods will be used to investigate features of task representation topography (e.g., compositional task axes, dimensionality). Analysis of brain network dynamics will investigate the role of PFC in coordinating cortical interactions to achieve flexible multitask behaviour.

WIP: Investigating the Effects of SSRI Manipulation on the Neural and Behavioural Correlates of Effort-Based Decision-Making using fMRI

 Presented by Eva-Periche Tomas

Abstract: Maintaining motivation and dedicating effort are essential for an individual’s success. Apathy, often a manifestation of pathological loss of motivation, is prevalent across various brain disorders and while pharmacological treatments have been used to address apathy and anhedonia in these conditions, our understanding of their mechanistic targets remains unclear.

In this study, our goal is to investigate how increased serotonin levels may influence the domains of behavioural and social motivation. To achieve this, 50 young healthy participants will undergo a course of 7 days of either placebo or Escitalopram, followed by fMRI scanning alongside an effort-based decision-making task exploring self-benefiting and prosocial behaviours. We hope this study will offer new insights into the mechanistic foundations of the brain mechanisms underpinning motivation. 

WIN’s new compute cluster and the retirement of Jalapeno 

Presented by Duncan Mortimer 

Abstract: WIN has a new compute cluster, with new facilities for remote access and data analysis. This talk will provide a demonstration of the system and allow questions about usage and how it differs from the legacy system.