Romain Brette

"The Brain, In Theory" is out today!

A short excerpt in The Transmitter:

https://www.thetransmitter.org/theoretical-neuroscience/the-brain-in-theory-an-excerpt/

I wrote a book: "The brain, in theory" (to be published by Princeton University Press).

First chapter and TOC:
http://romainbrette.fr/WordPress3/wp-content/uploads/2024/12/The-brain-in-theory-TOC-and-first-chapter.pdf

New preprint: “A dynamical perspective on biological reproduction”

https://hal.science/hal-05491732

I replace von Neumann's self-reproducing machine with a dynamical model of reproduction, in which invariant reproduction occurs as a result of convergence to a fixed point - ie, an emerging property. The genome then becomes a transmissible developmental constraint, not a representation of the organism. With many examples from Paramecium biology ;)

Just published: "Theory of axo-axonic inhibition".

https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1013047

I show that chandelier cells, which target the AIS of principal cells of cortex and hippocampus, inhibit by raising the spike threshold in an ohmic fashion (proportional to the synaptic current). This is the case also if the synaptic current is depolarizing (as long as the reversal potential is below spike threshold).

I'm very happy to announce that I’m moving to the Institute for Intelligent Systems and Robotics in Paris! An interdisciplinary institute with a natural interest in embodied cognition, among other themes that are dear to me.

https://www.isir.upmc.fr/?lang=en

I reviewed Farid Zahnoun's book, The embodiment of meaning:
https://rdcu.be/dJaTO

Great book demonstrating the incoherence of functionalism, the mainstream doctrine that claims that mind is essentially the causal organization of functional states.

Why large language models or deep nets are not conscious:

http://romainbrette.fr/notes-on-consciousness-xi-why-large-language-models-are-not-conscious/

(Where I argue that we don't know what consciousness is exactly, but we do have an idea of what consciousness is not)

Our paper on the interaction of Paramecium with obstacles is out:

https://royalsocietypublishing.org/doi/10.1098/rsos.221645

It's an empirical analysis of what Paramecium, the "swimming neuron", does when it encounters an obstacle: slide, swim backward (spike), slide then swim backward (spike).

On the existential risks of artificial intelligence
http://romainbrette.fr/on-the-existential-risks-of-artificial-intelligence/

Many have commented on whether this is plausible or a fantasy. I give a different perspective: misaligned autonomous human-made entities are plausible because they already exist, except not in the technological domain. These are the ones we should worry about first.

Our paper on the Paramecium model is out:

https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1010899

Paramecium is a swimming unicellular organism where action potentials trigger directional changes. We built a model of the action potential of Paramecium, coupled with kinematics, so the model swims and turns and navigates in structured environments.

There's a competitive call for PhD fellowships in neuroscience in Paris:
https://dim-cbrains.fr/en/phd-program/dim-cbrains

Students can apply for the subject I proposed: "Neuroscience of escape behavior in Paramecium, the “swimming neuron”"

@PhiloNeuroScie@neuromatch.social Yes that's usually part of what is considered an elementary step (eg in Church's lambda calculus or Turing stuff).

The point is that it's something that is reducible to a finite set of elementary stuff. The set you choose defines your model of computation, and the Church-Turing thesis says that the intuitive ones are equivalent.