Postdoc in computational neuroscience at Imperial College London
Colleen Gillon
Loading...
8/ This project was completed with my postdoc adviser, Claudia Clopath. 🙏 Thank you to the Clopath lab for support and feedback, and especially to Tom George and Justin O’Hare for invaluable comments on the manuscript. 🙏
@clopathlab.bsky.social @tomnotgeorge.bsky.social @justinkohare.bsky.social
Check out our new perspective paper in eNeuro! We explore recent advancements and future prospects in open neuroscience, and discuss how open science practices, data sharing, and structural changes can create a more collaborative and transparent scientific ecosystem
www.eneuro.org/content/12/1...
Having the most amazing time at Imbizo! What a fantastic group of students we've assembled here. Thank you to the organizers, Team House, faculty and funders! ❤️
6/ Our work also generated several experimentally testable predictions which we hope will help inspire further experimental research into the proposed role of the hippocampus in predictive learning.
2/ Predictive learning has been studied extensively in sensory cortex, where it helps explain anticipatory activity and prediction-error responses. But how are more abstract, long-range, behaviourally-relevant predictions learned?
1/ Check out our new modeling paper on predictive learning in the hippocampus! 🧠 www.biorxiv.org/content/10.6...
A central idea in neuroscience is that the brain makes sense of the world by learning to predict its environment 🧠➡️🌍.
5/ Our circuit naturally learned to anticipate features of the environment, and recapitulated known features of hippocampal activity in both 1D and 2D environments.
7/ Our simulation code, built using RatInABox 🐀🎁, is fully available online for anyone interested in exploring these questions further! github.com/colleenjg/pr...
3/ The hippocampus 🐴🌊, which is critical for memory and cognitive mapping, has been proposed to play a key role. How might it do this at the circuitry-level? In our paper, we show how key properties of hippocampal circuitry make it ideally suited for forming predictive maps of the environment.
4/ We built a circuit around two-compartment pyramidal neurons receiving distinct input streams, and trained it to balance behavioural timescale synaptic plasticity (BTSP) with distal apical dendrite inhibition.
