Maxime Janbon

Profile

Project Title: Information and criticality in neural activity during sensing and learning.

Summary: The aim of this project is to uncover organizational principles behind the integration of sensory stimuli in animal brains and how these principles allow for flexible learning and use of the encoded information in a variety of tasks.
 
The first strand of this project will focus on different characterisations of stimuli-evoked neural activity. Indeed from a dynamical perspective, a growing body of work suggests that neural activity is poised between order and disorder: it is at or near a critical phase transition. Such a state confers some optimal information processing properties to the system. From an information-theoretic perspective, recent developments in partial information decomposition have paved the way for a precise breakdown of how sensory information is represented in a biological system. In this section I would like to compare these two descriptions and explore how they interact in biologically constrained models such as artificial neural networks and connected artificial neurons (eg Wilson Cowan networks).
 
The second strand of this project will explore the role of inter-areal neural feedback in determining how task-related information is represented in animal brains. A cross species comparison will be made to investigate how feedback and representation change as a function of brain anatomy. The link between this change and cognition and task performance will then be explored in the different animal models used.

Supervisor:

• Prof Lars Chittka

Research