5/ The dynamics look like they might be driven by a local recurrent circuit. We report evidence of such a circuit and show that its emergence correlates with the shifts.
Excited to share that our paper is now out in Neuron @cp-neuron.bsky.social (dlvr.it/TM9zJ8).
Our perception isn't a perfect mirror of the world. It's often biased by our expectations and beliefs. How do these biases unfold over time, and what shapes their trajectory? A summary thread. (1/13)
6/ The shifts aren’t well-captured by the existing temporal models we tried. A target for future modeling work - directly fitting models on these data (which clearly have a lot of signal.
Yeah. I see a lot of "if chatGPT coded all of it, how do you know it's right?" - but for me the more important thing is "if chatGPT coded all of it, how did you learn anything???"
1/X Our new method, the Inter-Animal Transform Class (IATC), is a principled way to compare neural network models to the brain. It's the first to ensure both accurate brain activity predictions and specific identification of neural mechanisms.
Preprint: arxiv.org/abs/2510.02523
4/ The dynamics correspond to increased functionality for decoding hard images. Images that are harder for a model to classify are decoded later from the neural population, as representations become more complex.
3/ We document an area-wide phenomenon, not specific to subpopulations of neurons. Most individual electrodes are predicted by deeper DNN layers over the course of the response.
dlvr.it
People exhibit biases when perceiving features of the world, shaped by both external
stimuli and prior decisions. By tracking behavioral, neural, and mechanistic markers
of stimulus- and decision-rela...
Usually the process of writing code makes you realize you don't know the solution in enough detail. If AI fills in the blanks my making assumptions that you're not sufficiently aware of that could be dangerous. So seems like having the experience of having written code yourself would help
