Supervisor: Mark Humphries
School of Psychology

Description:

Psychology and neuroscience experiments on rule-learning typically take the form of discrete trials in which a participant is asked to make a choice between two or more alternatives. A correct choice elicits a reward — in some experiments the reward delivery is probabilistic. The goal set for the participant is to learn the rule leading to the correct choice by trial and error. To do so, learning can be guided by feedback (reward; error signals), predictive cues (lights, sounds), or both. Such experimental designs can be equally applied whether the participant is human, rodent, or bee.

When analysing the behaviour of a participant, we typically want to know when that participant has learnt the correct rule, and how that participant arrived there. To know the answers to these questions, for any trial we want to be able to ask: what choice strategy are they using now? In this project we will develop a Bayesian approach to answering that question, based on the idea of quantifying the probability that a given strategy is being implements. We will apply this approach to a range of experimental data from human and animal subjects (e.g. Maggi et al 2018). With the fine-grained analysis of behavioural strategies in hand, in the final part of the project we will look at the neural correlates of the behavioural strategies in neural population recordings from rodent cortex.

Requirements:

• Matlab and/or Python programming experience. A strong quantitative background in e.g. physics, mathematics, computer science or engineering

Reading:

• Maggi, S.; Peyrache, A. & Humphries, M. D. (2018) An ensemble code in medial prefrontal cortex links prior events to outcomes during learning. Nature Communications, 9, 2204

Please see full job listing here.