Connectome-constrained networks predict neural activity across the fly visual system.

Autor: Lappalainen JK; Machine Learning in Science, Tübingen University, Tübingen, Germany.; Tübingen AI Center, Tübingen, Germany.; Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, USA., Tschopp FD; Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, USA., Prakhya S; Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, USA., McGill M; Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, USA.; Computation and Neural Systems, California Institute of Technology, Pasadena, CA, USA., Nern A; Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, USA., Shinomiya K; Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, USA., Takemura SY; Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, USA., Gruntman E; Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, USA.; Dept of Biological Sciences, University of Toronto Scarborough, Toronto, Ontario, Canada., Macke JH; Machine Learning in Science, Tübingen University, Tübingen, Germany.; Tübingen AI Center, Tübingen, Germany.; Max Planck Institute for Intelligent Systems, Tübingen, Germany., Turaga SC; Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, USA. turagas@janelia.hhmi.org.
Jazyk: angličtina
Zdroj: Nature [Nature] 2024 Oct; Vol. 634 (8036), pp. 1132-1140. Date of Electronic Publication: 2024 Sep 11.
DOI: 10.1038/s41586-024-07939-3
Abstrakt: We can now measure the connectivity of every neuron in a neural circuit 1-9 , but we cannot measure other biological details, including the dynamical characteristics of each neuron. The degree to which measurements of connectivity alone can inform the understanding of neural computation is an open question 10 . Here we show that with experimental measurements of only the connectivity of a biological neural network, we can predict the neural activity underlying a specified neural computation. We constructed a model neural network with the experimentally determined connectivity for 64 cell types in the motion pathways of the fruit fly optic lobe 1-5 but with unknown parameters for the single-neuron and single-synapse properties. We then optimized the values of these unknown parameters using techniques from deep learning 11 , to allow the model network to detect visual motion 12 . Our mechanistic model makes detailed, experimentally testable predictions for each neuron in the connectome. We found that model predictions agreed with experimental measurements of neural activity across 26 studies. Our work demonstrates a strategy for generating detailed hypotheses about the mechanisms of neural circuit function from connectivity measurements. We show that this strategy is more likely to be successful when neurons are sparsely connected-a universally observed feature of biological neural networks across species and brain regions.
(© 2024. The Author(s).)
Databáze: MEDLINE
Externí odkaz: