The transformer as a candidate canonical cortical microcircuit
- 1Stanford University
Abstract
Neuroscience hypothesizes a canonical microcircuit subserving cortical computation across brain systems. In artificial intelligence, neural networks have become a universal computational substrate. Here, we leverage newly available large-scale data and mapping techniques to compare neural network motifs to the brain. We embedded a spectrum of candidate microcircuit architectures within a common spatial macrocircuit, measuring alignment with an extensive battery of cortical signatures. We find that one specific circuit type--the transformer--uniquely captures fine-grained unit-level circuit details, population-level functional maps and neuronal activity patterns, and systems-level wiring efficiency and behavioral demands. The key ingredient enabling the transformer's success is a learned multiplicative component making relational information available without extensive long-range wiring. These results suggest striking convergence between computational motifs in natural and artificial intelligence.
Related
- A canonical microcircuit for estimating Excitation/Inhibition (E/I) balance
- CREST: A Cortical Resting-State EEG Spatial Transformer for Chronic Pain Inference
- Natural prey capture reveals flexible reconfiguration of canonical motor-cortical dynamics for online control
- ASOFormer: a Transformer-based model for predicting antisense oligonucleotide efficacy to support therapeutic candidate prioritization
- Identification of alcohol use disorder-associated and ethanol-responsive proteins and candidate biomarkers using human cortical organoids
BibTeX
@misc{shah2026transformer,
title = {{The transformer as a candidate canonical cortical microcircuit}},
author = {Y. Shah and D. L. K. Yamins},
howpublished = {bioRxiv},
year = {2026},
month = nov,
doi = {10.64898/2026.11.01.763457},
url = {https://www.biorxiv.org/content/10.64898/2026.11.01.763457v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - The transformer as a candidate canonical cortical microcircuit
AU - Y. Shah
AU - D. L. K. Yamins
PY - 2026
DA - 2026/11/01
PB - bioRxiv
DO - 10.64898/2026.11.01.763457
UR - https://www.biorxiv.org/content/10.64898/2026.11.01.763457v1
AB - Neuroscience hypothesizes a canonical microcircuit subserving cortical computation across brain systems. In artificial intelligence, neural networks have become a universal computational substrate. Here, we leverage newly available large-scale data and mapping techniques to compare neural network motifs to the brain. We embedded a spectrum of candidate microcircuit architectures within a common spatial macrocircuit, measuring alignment with an extensive battery of cortical signatures. We find that one specific circuit type--the transformer--uniquely captures fine-grained unit-level circuit details, population-level functional maps and neuronal activity patterns, and systems-level wiring efficiency and behavioral demands. The key ingredient enabling the transformer's success is a learned multiplicative component making relational information available without extensive long-range wiring. These results suggest striking convergence between computational motifs in natural and artificial intelligence.
ER -