A sinusoidal activity profile in a circulant ring network doing path or heading integration has some nice properties - it is noise resistant and self-stabilising with a Hebbian learning rule.
This code performs the analysis and figure generation for the paper "Theoretical principles explain the structure of the insect head direction circuit" or see preprint here
If building on this analysis or these ideas, please cite our work:
@article {10.7554/eLife.53985,
article_type = {journal},
title = {Theoretical principles explain the structure of the insect head direction circuit},
author = {Vilimelis Aceituno, Pau and Dall'Osto, Dominic and Pisokas, Ioannis},
volume = 13,
year = 2024,
pages = {e91533},
citation = {eLife2024;13:e91533},
doi = {10.7554/eLife.91533},
url = {https://doi.org/10.7554/eLife.91533},
journal = {eLife},
publisher = {eLife Sciences Publications, Ltd},
}The code is available under the BSD-3-Clause License.
- What does it look like to encode with multiple harmonics at once - Figure 1
- How is an encoding with multiple harmonics affected by noise
- Comparing the 4 population networks from Ioannis's paper to the abstract network we predict - Figures 3, 4, M1, and M6
- Deriving the 4 population connectivity network from Janelia's connectome data - Figures M2-S5
- Updating both the activity and weights simultaneously - Figures 5 and M7