Top-Down Feedback Is Filtering, Not Prediction: A Falsifiable Dissociation in a Lattice Memory Substrate

Read full text Summarise with VIRAT
Authors: Gabriel Aubin-Moreau
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

We present a minimal experiment in a lattice memory substrate (VCML) that produces a clean falsification of the predictive coding claim that top-down feedback carries generated predictions. Using an interference T-maze with random noise distraction, we show that the top-down signal (fieldM to mid_mem attractor feedback) does not predict, generate, or reconstruct — it simply filters noise by pulling fast memory traces toward stored patterns. Direct+Attractor = chance at all settling times (0.502). Mid-only+Attractor = perfect at all settling times (1.000). No generative model, no prediction error, no Bayesian inference. One line of code: mid_mem += beta * (fieldM - mid_mem).

Fields of Study & Subjects

Memory and Neural Mechanisms Embodied and Extended Cognition Neural dynamics and brain function

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
25 March 2026
Language
EN
Document type
Preprint