#Tag · bonfire.cafe

What happens when elegant neural coding theory meets biological reality?

For our Inside Discovery series, Veronika Koren explains why adding messiness — noise, metabolism, inhibition — still leads to efficient brains.
https://elifesciences.org/inside-elife/75921c16/inside-discovery-where-theory-meets-biology?utm_source=mastodon&utm_medium=social&utm_campaign=organic

RE: https://fediscience.org/@eLife/115871742804185192

Cool example of #theory meeting #biology: shows that efficient coding principles survive when models include realistic circuitry, noise, and metabolic constraints. A reminder that good theory does not require oversimplification.

#NeuralCoding #CompNeuro #Neuroscience

Fabrizio Musacchio

@FabMusacchio@mastodon.social · 2 weeks ago

eLife

@eLife@fediscience.org · 2 weeks ago

What happens when elegant neural coding theory meets biological reality?

RE: https://fediscience.org/@eLife/115871742804185192

#NeuralCoding #CompNeuro #Neuroscience

Fabrizio Musacchio

@pixeltracker@sigmoid.social · 2 months ago

🧠 New preprint by Tilbury et al: Characterizing #NeuronalPopulation geometry with #AI equation discovery

The approach generates & evaluates 100s of candidate equations, finding "peaky" non-Gaussian tuning functions whose Fourier structure matches power-law dimensionality observed in real #V1 pops. Links shape of single- #neuron tuning to #PopulationLevel geometry using both data fits & analytical derivations.

🌍 https://doi.org/10.1101/2025.11.12.688086

#CompNeuro #Neuroscience #NeuralCoding #PopulationDynamics

2 media

Fig. 1. Oriented stimuli produce a high-dimensional population code not captured by standard tuning curve models.

Fig. 5. Effect of tuning peakiness on a simulated hyperacuity task.

Characterizing neuronal population geometry with AI equation discovery

Fabrizio Musacchio

@pixeltracker@sigmoid.social · 2 months ago

🧠 New paper by Safaai et al. (2025): parietal #cortex output populations show highly structured, task-dependent population geometry. Using multi-area recordings and circuit modeling, they show that #parietal populations display organized task-related patterns rather than uniform mixed coding, and that distinct output groups shape how decisions are routed to downstream targets:

🌍 https://doi.org/10.1038/s41593-025-02095-x

#Neuroscience #NeuralCoding #ParietalCortex #PopulationDynamics #DecisionMaking

Fig. 1: Differences in the activity of neurons projecting to distinct cortical targets.

Specialized structure of neural population codes in parietal cortex outputs