Sign in

Ali Haydaroğlu

@haydari.bsky.social
175 followers 247 following 26 posts

neuroscientist {calcium imager, principal component analyzer} @ harvard med school. previously ucl, uoft

PostsRepliesMedia
Ali Haydaroğlu @haydari.bsky.social · 28/04/2026
There's a central registry of agents, so if one agent has a question or a task, it can check the registry to find an appropriate agent to communicate with. You can set your agents in /loops to keep checking their memos for autonomous coordinated operation.
000
Ali Haydaroğlu @haydari.bsky.social · 28/04/2026
It's flexible: Individual agents can have generic roles like "backend" or "design", or specific, like "ResNet-benchmarking". They plan and send each other memos through shared "streams" and keep their own memories in their personal scratchboards. Tested with claude VSCode.
100
Ali Haydaroğlu @haydari.bsky.social · 28/04/2026
I've been working on some software dev with agents, and I made a light plug-in that makes it easy to coordinate up to 10 agents on a project. It's called tayfa, Turkish for crew. Works well for me, I thought it might be useful for others too! github.com/alihaydarogl...
160
Ali Haydaroğlu @haydari.bsky.social · 18/12/2025
congrats Jure, this is incredibly cool data!
010
Ali Haydaroğlu @haydari.bsky.social · 16/12/2025
Thanks for the interest!! I’m curious what you mean about the UMAP, was it on the image responses?
120
Ali Haydaroğlu @haydari.bsky.social · 15/12/2025
Agreed with the point that manifolds are probably different between contexts… I guess the question is whether the manifolds are constrained by physical circuits (a la Langdon & Engel) or simply defined by tasks. I also don’t really have a horse in the race!
020
Ali Haydaroğlu @haydari.bsky.social · 15/12/2025
Yes, exactly. I agree with you that most claims of low-D neural activity are probably high-D by our definition. But, the underlying dynamics can still be low-D.
020
Ali Haydaroğlu @haydari.bsky.social · 15/12/2025
Not necessarily! See our conjecture with Valentin: given a neural nonlinearity (p-times diff’able), you can get a spectrum with slope (alpha) close to 1 from low-dim “pre-activations” (inputs of dimension d)
110
Ali Haydaroğlu @haydari.bsky.social · 14/12/2025
congrats!!
010
Ali Haydaroğlu @haydari.bsky.social · 13/12/2025
"Low-d neural activity" would probably look high-D if you plot the spectrum on a log-log scale. @engeltatiana.bsky.social made a similar point recently on @braininspired.bsky.social. Also agrees with our claim with @bio-emergent.bsky.social that low-D and high-D can be two sides of the same coin!
192
Ali Haydaroğlu @haydari.bsky.social · 13/12/2025
The power laws don't necessarily pose a contradiction to low-D dynamics!
150
Ali Haydaroğlu @haydari.bsky.social · 13/12/2025
see: openreview.net/forum?id=cGk...
openreview.net
High-dimensional neuronal activity from low-dimensional latent...
Computation in recurrent networks of neurons has been hypothesized to occur at the level of low-dimensional latent dynamics, both in artificial systems and in the brain. This hypothesis seems at...
020
Ali Haydaroğlu @haydari.bsky.social · 13/12/2025
See our paper with @bio-emergent.bsky.social et al openreview.net/forum?id=cGk...
openreview.net
High-dimensional neuronal activity from low-dimensional latent...
Computation in recurrent networks of neurons has been hypothesized to occur at the level of low-dimensional latent dynamics, both in artificial systems and in the brain. This hypothesis seems at...
010
Ali Haydaroğlu @haydari.bsky.social · 13/12/2025
My (speculative) interpretation is that spont. activity in mouse cortex is a nonlinear readout of low-D global variables. I think activity lives in a low-D manifold where the axes correspond to arousal/internal state/neuromodulation levels, and neural nonlinearities project it into high-D in cortex.
110
Ali Haydaroğlu @haydari.bsky.social · 13/12/2025
I don't think it's crazy to think there are truly low-D manifolds! We see high-D power-law eigenspectrum in mouse cortex, both in image responses and spontaneous activity (no low-D task!). But, we find that spontaneous activity has MUCH lower "pre-activation" dimensionality than image responses.
100
Ali Haydaroğlu @haydari.bsky.social · 13/12/2025
Example from our NeurIPS talk with @bio-emergent.bsky.social : In an RNN with low-D *pre-activation* dynamics, an activation function produces high-D *post-activations*. A downstream neuron could only read out boring sinusoids from low-D state, but can read any function from high-D activity!
150
Ali Haydaroğlu @haydari.bsky.social · 13/12/2025
Low-d dynamics are useful - they are noise robust and easier to control. BUT you also get a major benefit from curling up a low-D state manifold in a high-D embedding space: you can then linearly read out arbitrary functions of state!
230
Ali Haydaroğlu @haydari.bsky.social · 09/12/2025
thanks!! hmm maybe a lab retreat 👀
020
Ali Haydaroğlu @haydari.bsky.social · 08/12/2025
thank you Shuqi!! was so nice to meet you through this work :)
110
Ali Haydaroğlu @haydari.bsky.social · 08/12/2025
the paper: www.biorxiv.org/content/10.1...
010
Ali Haydaroğlu @haydari.bsky.social · 08/12/2025
Thanks to authors @shuqiw.bsky.social, Yixiao, @carandinilab.net and @kenneth-harris.bsky.social! And to our photographer @flavioh.bsky.social :)
150
Ali Haydaroğlu @haydari.bsky.social · 08/12/2025
Best part about being a scientist is the people I get to work with. Valentin (@bio-emergent.bsky.social) and I got to give a talk at NeurIPS, bridging a gap between low- and high-dim perspectives of the brain. Thankfully, the audience was (somewhat) more awake than the San Diego desert 🏜️
4324
Reposted by Ali Haydaroğlu
Valentin Schmutz @bio-emergent.bsky.social · 05/12/2025
Tomorrow at #NeurIPS2025! Oral at 10 am in UL Ballroom 20D and poster #2016 at 11 am. @haydari.bsky.social and I are looking forward to hearing your thoughts.
0184
Ali Haydaroğlu @haydari.bsky.social · 04/11/2025
congrats!!
010
Ali Haydaroğlu @haydari.bsky.social · 29/10/2025
congrats Tom, Montreal is lovely I’m sure you’ll enjoy it!
020
Reposted by Ali Haydaroğlu
Valentin Schmutz @bio-emergent.bsky.social · 19/09/2025
🎉 "High-dimensional neuronal activity from low-dimensional latent dynamics: a solvable model" will be presented as an oral at #NeurIPS2025 🎉 Feeling very grateful that reviewers and chairs appreciated concise mathematical explanations, in this age of big models. www.biorxiv.org/content/10.1... 1/2
411023
Reposted by Ali Haydaroğlu
The Transmitter @thetransmitter.bsky.social · 31/07/2025
Some neurons respond more strongly to visual stimuli than others. But the rest of the brain doesn’t prioritize signals from those neurons, according to a new study. Read more in this month’s Null and Noteworthy. By @ldattaro.bsky.social #neuroskyence www.thetransmitter.org/null-and-not...
thetransmitter.org
Null and Noteworthy: Downstream brain areas read visual cortex signals en masse in mice
The finding contradicts a theory that the regions prioritize neurons that are adept at identifying specific stimuli. Plus, a response to a study that questioned immune memory in astrocytes.
0386
Reposted by Ali Haydaroğlu
Matteo Carandini @carandinilab.net · 24/07/2025
Introducing our new favorite stimulus. A few minutes are enough to map the visual preferences of thousands of neurons. Mapping the visual cortex with Zebra noise and wavelets www.biorxiv.org/content/10.1... By Sophie Skriabine and Max Shinn, with Samuel Picard and @kenneth-harris.bsky.social
411829
Reposted by Ali Haydaroğlu
Matteo Carandini @carandinilab.net · 07/07/2025
New by Agnès Landemard (@agnesland.bsky.social‬) & co Brainwide blood volume reflects opposing neural populations Brainwide fluctuations in blood volume arise from two populations with opposite relation to brain state and distinct relationships to blood supply www.biorxiv.org/content/10.1...
16114
Reposted by Ali Haydaroğlu
Valentin Schmutz @bio-emergent.bsky.social · 09/06/2025
Our new preprint 👀
0316
Reposted by Ali Haydaroğlu
Matteo Carandini @carandinilab.net · 09/06/2025
The firing of neural populations is high-dim even if their subthreshold activity is low-dim! This work by @bio-emergent.bsky.social and @haydari.bsky.social shows how, with a solvable model, a data analysis technique, and data from mouse visual cortex: www.biorxiv.org/content/10.1...
biorxiv.org
High-dimensional neuronal activity from low-dimensional latent dynamics: a solvable model
Computation in recurrent networks of neurons has been hypothesized to occur at the level of low-dimensional latent dynamics, both in artificial systems and in the brain. This hypothesis seems at odds ...
09028
Reposted by Ali Haydaroğlu
Hayley Bounds @hayleybounds.bsky.social · 15/05/2025
Excited to share my PhD paper! In it, we use targeted 2-photon optogenetic stimulation to determine how V1 activity is read-out in a detection task. We found that network influence, not visual coding properties, predicted the impact of ensembles on behavior - contradicting our expectations (1/5).
511336
Ali Haydaroğlu @haydari.bsky.social · 17/05/2025
So cool! Congrats :)
020
Reposted by Ali Haydaroğlu
Enny van Beest @ennyvb.bsky.social · 14/04/2025
Our paper describing UnitMatch software "tracking neurons across days with high-density probes", appeared in this month's issue of #NatureMethods. Missed it? Read it here: www.nature.com/articles/s41...
Two neurons in front of a Neuropixels probe. One neuron says to the other: "Don't look, but I think we're being followed!".
0122
Ali Haydaroğlu @haydari.bsky.social · 01/04/2025
If you do multi-plane 2p imaging, you may see the same cells across planes. We built Suite3D exactly for this purpose! It removes duplicates, corrects axial movement, finds more cells with less noise, much faster than alternatives. github.com/alihaydarogl... Try on your data, and reach out!
0214