Dyer, Alexander James;
(2025)
Inferring the Physical and Functional Structure of Neural Networks from their Activity.
Doctoral thesis (Ph.D), UCL (University College London).
![[thumbnail of Dyer_Thesis.pdf]](https://discovery.ucl.ac.uk/style/images/fileicons/text.png) |
Text
Dyer_Thesis.pdf - Submitted Version Access restricted to UCL open access staff until 1 February 2026. Download (10MB) |
Abstract
Given the complexity and fine-scale of biological neural networks, their topography (spatial arrangement in the brain) and topology (patterns of connectivity) are challenging to experimentally determine. For example, complete wiring diagrams (connectomes) exist only for extremely simple organisms such as C. elegans. As such methods for inferring the structural and functional properties of neural networks are required in lieu of techniques for direct observation. Here, we investigate the inverse problem of inferring network properties from the network activity. First, we investigate the problem of inferring the topography of entorhinal grid cells within a module, based on their toroidal firing behaviour. Through numerical simulations, including selforganising maps and comparison with experimental data, we predict that the grid cell topography broadly matches that observed in orientation selective neurons in V1, except for two circular dimensions instead of one, corresponding to the two phase axes of the grid cell torus. Second, we systematically investigate the problem of network inference in the setting of artificial neural networks. After proposing a research agenda for incrementally building towards full network inference, we address the first task of assigning artificial neurons to locations in a network of known architecture, specifically the LeNet image classifier. We evaluate a supervised learning approach based on features derived from the eigenvectors of the activation correlation matrix. Experiments highlighted that for an image dataset to be effective for accurate localisation, it should fully activate the network and contain minimal confounding correlations. No single image dataset was found that resulted in perfect assignment, however perfect assignment was achieved using a concatenation of features from multiple image datasets.
| Type: | Thesis (Doctoral) |
|---|---|
| Qualification: | Ph.D |
| Title: | Inferring the Physical and Functional Structure of Neural Networks from their Activity |
| Language: | English |
| Additional information: | Copyright © The Author 2025. Original content in this thesis is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licence (https://creativecommons.org/licenses/by-nc/4.0/). Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms. Access may initially be restricted at the author’s request. |
| UCL classification: | UCL UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Life Sciences UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Life Sciences > Div of Biosciences |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10204055 |
Loading...
Loading...Loading...Loading...Archive Staff Only
 |
View Item |
