Cortical columns and active inference

Research and Theory
1

Is anybody here interested in Active Inference? I learned about predictive processing from Jeff Hawkins’ book “On Intelligence”. Last year I got to know Daniel Friedman, President of the Active Inference Institute https://www.activeinference.institute and I have been getting to know this theory, especially the mathematical foundations, which I have my own thoughts on.

I attend Karl Friston’s Theoretical Neurobiology meetings where speakers present their work (40 minutes) and we all can discuss it (30 minutes) and then Karl Friston relates it to Active Inference and the Free Energy Principle. The meeting tomorrow is on cortical memory. Here is the zoom link:

and I include more information below.

Dear all,

We will have one Theoretical Neurobiology (TNB) meeting next week, on Tuesday, December 9, 2025 at 2:30pm UK time. Don Tucker will be giving a presentation titled ‘Excitatory-Inhibitory Resonance Establishes Cortical Memory’, and the meeting will be chaired by Will. Don M. Tucker is a neuropsychologist and electrophysiologist (bel.company).

Abstract for the presentation: We propose that experience and memory emerge from structured resonance between excitatory (E) and inhibitory (I) population waves in the cerebral cortex, constrained by cortical geometry and organized by subcortical controls. This mechanism of E–I resonance links deep (layer 5-6) excitatory waves, generating limbifugal predictions, with superficial (2-3) inhibitory networks, mediating error-correction on the basis of sensory input in predictive coding formulations of active inference. When the low-frequency (theta) excitatory waves resonate with in-phase (n:m coupled) higher frequency (gamma) inhibitory waves, the resonance reflects shared information and creates transient standing‑wave “write windows” that persist long enough to induce early LTP in the local synaptic network.

The meeting will be conducted online, at this link (it’s the same meeting link every week). Please see this document with screenshots to guide you on how to join the meetings via your web browser. Additionally, here is the link to view the TNB calendar, and here is the iCal link to add the TNB calendar to your own calendar.

Looking forward to seeing you there!
with best wishes,

Riddhi

on behalf of the TNB Chairs (Riddhi, Peter, Robert, Will, and Miguel)

TNB website | TNB YouTube Channel

_______________________________________________
Fil.neurobiology mailing list –fil.neurobiology@ucl.ac.uk
To unsubscribe send an email tofil.neurobiology-leave@ucl.ac.uk

3 Likes
2

Hi @AndriusKulikauskas, I am currently trying to get a grasp of Active Inference and have heard about the active inference institute before. As I loosely understand it, Active Inference models are primarily concerned with minimizing free energy / Bayesian surprise (or whatever it is called– again– loose understanding) and I am not sure how Monty fits into this framework.

I understand some of the connections between Monty and predictive coding which is said to be a close relative of Active Inference: there is hierarchy, top-down connections (on the future development roadmap). Do model hypotheses and the voting mechanisms count as error computations similar to predictive coding? Is that something entirely different or is there something within Monty that is more aligned with the predictive coding point of view?

Could you please share your thoughts about what you consider to be the similarities and differences between Active Inference and Monty? I am currently trying to figure this out for myself as well.

1 Like
3

I feel like active inference is more of a mathematically-abstracted “cousin” of Hierarchical Temporal Memory, and is a little less relatable to the Thousand Brains theory. Active inference could perhaps be leveraged in a larger-scale Thousand Brains system down the road, maybe as an alternative to HTM in learning modules for example, but it’s less applicable to Monty in its current state.

1 Like
4

Hi @sebastianruizsebas , great question, I’m in the process of updating our existing FAQ page with some more details on active inference and predictive coding - you can check out the current draft at this PR link here. See section, “What is the Relationship of the Thousand Brains Theory to Active Inference and the Free Energy Principle? What About Predictive Coding?”

I think that should answer your questions about the difference, but it’s definitely a confusing topic, particularly if it’s something you’re only starting to get into, so feel free to ask clarifying questions!

3 Likes