Dan Goodman<p>Is anarchist science possible?</p><p>As an experiment, we got together a large group of computational neuroscientists from around the world to work on a single project without top down direction. Read on to find out what happened.</p><p>The project started as a tutorial on a new technique at the <span class="h-card" translate="no"><a href="https://neuromatch.social/@CosyneMeeting" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>CosyneMeeting</span></a></span> 2022. We realised that the technique was easy and cheap for anyone to use with a lot of low hanging fruit.</p><p><a href="https://neural-reckoning.github.io/cosyne-tutorial-2022/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">neural-reckoning.github.io/cos</span><span class="invisible">yne-tutorial-2022/</span></a></p><p>At the tutorial we announced a 1-2 year open research project that anyone could join, starting from the materials of the tutorial, and a few starting questions we found interesting, but with no other constraints. We were inspired by the Polymath Project in mathematics.</p><p><a href="https://en.wikipedia.org/wiki/Polymath_Project" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">en.wikipedia.org/wiki/Polymath</span><span class="invisible">_Project</span></a></p><p>31 people contributed to the project, joining for monthly meetings to discuss progress. All code was publicly available throughout, and when we started writing up the work in progress was also fully public. You can see that version here: <a href="https://comob-project.github.io/snn-sound-localization/paper" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">comob-project.github.io/snn-so</span><span class="invisible">und-localization/paper</span></a></p><p>Not everyone who was involved made it to the paper (didn't respond or couldn't find contact details), and not all are on Mastodon, but authors include: <span class="h-card" translate="no"><a href="https://neuromatch.social/@marcusghosh" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>marcusghosh</span></a></span> <span class="h-card" translate="no"><a href="https://fediscience.org/@TomasFiers" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>TomasFiers</span></a></span> <br><span class="h-card" translate="no"><a href="https://neuromatch.social/@GabrielBena" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>GabrielBena</span></a></span> <span class="h-card" translate="no"><a href="https://sigmoid.social/@rory" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>rory</span></a></span> </p><p>We used GitHub and Jupyter notebooks to coordinate development, with a website showing everyone's current code and results to make collaboration easier. We used <span class="h-card" translate="no"><a href="https://fosstodon.org/@mystmarkdown" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>mystmarkdown</span></a></span> and GitHub actions to automate this.</p><p><a href="https://comob-project.github.io/snn-sound-localization/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">comob-project.github.io/snn-so</span><span class="invisible">und-localization/</span></a></p><p>So how did it work out? Well, some things went well and others not so well. We published a paper with our results and reflections on the process. If you're interested in spiking neural networks, sound localisation, or anarchist science, check it out:</p><p><a href="https://www.eneuro.org/content/12/7/ENEURO.0383-24.2025" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">eneuro.org/content/12/7/ENEURO</span><span class="invisible">.0383-24.2025</span></a></p><p>Generally, the infrastructure we built worked well, as did the monthly meetings. Starting from the tutorial was a good decision because it gave everyone a common reference and meant they could easily get started.</p><p>However, the lack of direction meant that we didn't achieve very coherent results in the end. We don't think this is a catastrophic problem, but when we try again, this is something we'd like to address. If you have thoughts or would like to be involved, get in touch!</p><p>Ultimately, we didn't achieve a scientific breakthrough in this project, but we did show that without top down direction or any specific funding, we could organise a large group of scientists to work together and publish their research in a good journal. We think that's a hopeful sign for the future!</p><p><a href="https://neuromatch.social/tags/neuroscience" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>neuroscience</span></a> <a href="https://neuromatch.social/tags/computationalneuroscience" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>computationalneuroscience</span></a> <a href="https://neuromatch.social/tags/compneuro" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>compneuro</span></a> <a href="https://neuromatch.social/tags/anarchism" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>anarchism</span></a> <a href="https://neuromatch.social/tags/science" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>science</span></a></p>
#compneuro
0 posts0 participants0 posts today
New #TeachingMaterial available: Functional Imaging Data Analysis – From Calcium Imaging to Network Dynamics. This course covers the entire workflow from raw #imaging data to functional insights, including #SpikeInference & #PopulationAnalysis. Designed for students and for self-guided learning, with a focus on open content and reproducibility. Feel free to use and share it 
https://www.fabriziomusacchio.com/blog/2025-07-13-function_image_analysis/
How can we test theories in neuroscience? Take a variable predicted to be important by the theory. It could fail to be observed because it's represented in some nonlinear, even distributed way. Or it could be observed but not be causal because the network is a reservoir. How can we deal with this?
Increasingly feel like this isn't a theoretical problem but a very practical one that comes up all the time. I'd be interested if anyone has seen anything practical that addresses this.
Almost last call to register for UK neural computation conference in London July 10-11. Registration deadline is July 1st. We have some great talks and posters as well as a session on funding with ARIA.
Look forward to seeing you all there. Now click here 
Preprint alert 
New insights into the tradeoff of effort and delay costs! A collaboration with the Wikenheiser lab #neuroscience #compneuro https://www.biorxiv.org/content/10.1101/2025.06.03.657635v1
The #Italian Network of #ComputationalNeuroscience announced its 2025 conference:
Palazzo della Salute, Padova, Italy 
September 22–24, 2025
Submission deadline: June 7, 2025 https://www.incn.it/
A 3-day deep dive into the brain — from models to data, theory to technology.
#CompNeuro #Neuroscience
Late-breaking poster call!
Submit your abstract for CNS*2025 in Florence, Italy, July 5–9
Poster-only presentations | Deadline: June 8
#compneuro #CNS2025 
How do babies and blind people learn to localise sound without labelled data? We propose that innate mechanisms can provide coarse-grained error signals to boostrap learning.
New preprint from @yang_chu.
https://arxiv.org/abs/2001.10605
Thread below
arXiv.orgLearning spatial hearing via innate mechanismsThe acoustic cues used by humans and other animals to localise sounds are subtle, and change during and after development. This means that we need to constantly relearn or recalibrate the auditory spatial map throughout our lifetimes. This is often thought of as a "supervised" learning process where a "teacher" (for example, a parent, or your visual system) tells you whether or not you guessed the location correctly, and you use this information to update your map. However, there is not always an obvious teacher (for example in babies or blind people). Using computational models, we showed that approximate feedback from a simple innate circuit, such as that can distinguish left from right (e.g. the auditory orienting response), is sufficient to learn an accurate full-range spatial auditory map. Moreover, using this mechanism in addition to supervised learning can more robustly maintain the adaptive neural representation. We find several possible neural mechanisms that could underlie this type of learning, and hypothesise that multiple mechanisms may be present and interact with each other. We conclude that when studying spatial hearing, we should not assume that the only source of learning is from the visual system or other supervisory signal. Further study of the proposed mechanisms could allow us to design better rehabilitation programmes to accelerate relearning/recalibration of spatial maps.