Computational modelling of neural mechanisms of brain plasticity

The goal is to model and understand plasticity effects in biological and artificial neural networks by studying learning dynamics in neural networks with plasticity rules and compare results with experiments.

Illustration

Fig. 1: Extracellular matrix structure (green) wrap around the neuron (red). When plasticity is required, e.g. during learning, specialized enzymes cleave the matrix structure to enable synapse remodelling. This project will address how such processes can be modelling with artificial neural network models.

The brain shows an immense ability to change and learn from experience while retaining long-term memories for decades. Still, the underlying mechanisms of brain plasticity and its relation to the complex interactions between billions of brain cells are mostly unresolved.

We have a strong experimental activity on the neural mechanisms underlying plasticity[1]. Insights from these experiments, hold great potential to develop biologically inspired artificial neural networks that will impact our understanding of artificial intelligence, improve existing algorithms, and provide a deeper understanding of biological neural processes.

Based on insight from laboratory experiments, you will in this project develop recurrent neural network model as highly simplified models for the dynamics of learning, the role of inhibitory cells, adaptation and plasticity. You will have access to large-scale data of high temporal and spatial resolution that allow direct comparison with simulations. You will work in an interdisciplinary environment between physics, computer science and bioscience that combines experiments and modeling, providing a basis for new breakthroughs in our understanding of both artificial and biological neural networks.

The project include collaborations with leading groups at Harvard University and the University of California San Diego.

[1] E. H. Thompson, K. K. Lensjø, M. B. Wigestrand, A. Malthe-Sørenssen, T. Hafting, M. Fyhn, Removal of perineuronal nets disrupts recall of a remote fear memory, Proceedings of the National Academy of Sciences (2017) 115, 607-612

Requirements

  • You must have a master degree in physics, computational neuroscience, artificial intelligence or similar.
  • Documented experience from scientific computing or neural network simulations will be considered an advantage.

Supervisors

Professor Anders Malthe-Sørenssen

Professor Gaute Einevoll

Call 2: Project start autumn 2022

This project is in call 2, starting autumn 2022. 

Published Sep. 25, 2020 12:30 PM - Last modified Oct. 14, 2020 10:22 AM