Machine Learning for Functional Connectomics in C. elegans

Join Andrew Warrington from Stanford University as discusses C. elegans, one of the most widely studied organisms in the world.

In vivo, experimental techniques are labour intensive and are limited by what can be practically observed or perturbed. Whole-organism simulation-based methods are coming to the fore, allowing high-fidelity examination and experimentation on digital specimens. However, these methods are still largely limited to forward-simulation only and are reliant on extensive manual tuning.

There is therefore an opportunity to leverage machine learning techniques to automate and expand quantitative and functional in silico analysis. We propose leveraging whole-brain-body simulations as a strong “prior” over electrophysiology and behaviours and using experimental data to refine, develop, and individualise these models to, for instance, particular phenotypes, or even individual specimens. This approach also allows more wide-reaching and variable hypotheses to be tested more rapidly and transparently, and without the need to develop new wet-lab procedures. More broadly, Andrew will discuss the opportunities presented by the intersection of machine learning, organism-scale computational simulation, and traditional experimental techniques, to generate new neuroscientific insights and experimental practices.

The talk will take place online on Zoom, click here to join: Join Zoom Meeting here.

Meeting ID: 865 0881 0827
Passcode: +u34wJ

Organised by Karim Djemame