A low-dimensional hypothesis for cognition and the human brain

Abstract

The human brain is among the most complex biological systems known, yet a growing body of evidence suggests that its activity consistently exhibits low-dimensional organisation. Across behavioural states, recording modalities, spatial scales, and analytical approaches, large-scale neural dynamics can often be described by a surprisingly limited number of latent processes expressed through coordinated large-scale spatiotemporal patterns. This observation presents a fundamental paradox: how can a system capable of remarkably flexible cognition operate within such constrained dynamical structure? Here, we propose a low-dimensional hypothesis for cognition. We argue that cognition depends on the emergence of low-dimensional latent processes from complex neural activity, their reconfiguration according to internal and external demands, and the interactions among them. Within this framework, cognition arises not from unconstrained high-dimensional neural activity, but from the efficient organisation of neural dynamics into latent processes and their interactions. We review converging evidence for low-dimensional organisation across brain activity and cognition and outline a set of falsifiable predictions concerning cognitive demands, learning, development, aging, neurological and psychiatric disorders, pharmacological interventions, and social interaction. Finally, we consider whether this principle extends across spatial scales and whether analogous organisation emerges in artificial neural networks. Altogether, we propose low-dimensional organisation as a candidate principle for understanding the neural substrate of cognition.

Whole-brain modeling of dynamic causal circuits in human cognition using amortized variational inference
Byeongwook Lee, Louis Rouillard, Leticia Levin Diniz +8
HAL9 Aug 2026PreprintLicence not statedNeuroscience StatisticsHAL
The Platonic brain bridge hypothesis: human brain networks as an architectural prior for multimodal large language models
Pengfei Zhang, Biao Tian, Xiangang Li +1
arXiv14 Sep 2026PreprintCC BY 4.0Full textNeurosciencearXiv
Hippocampal Neuronal Overflow and Cortical Reconstruction: A Hypothesis on the Formation of Neural Networks Unique to the Human Brain
Jue Wang
OSF Preprints24 Oct 2026PreprintCC BY 4.0Full textNeuroscience General medicineOSF
High-Dimensional Multi-omic Mapping of Post-Mortem Human Brain Using Iterative Indirect Immunofluorescence Imaging on Xenium-Processed Tissues
B. Karaahmet, T. Lama, A. Wang +7
bioRxiv16 Sep 2026PreprintCC BYFull textNeurosciencebioRxiv
A Comparison of Brain Metabolic Connectivity Methods Topology, Cognition, Age, Structure, and Genetics
H. Deery, S. Zhao, E. Liang +2
bioRxiv16 Sep 2026PreprintAll rights reservedFull textNeurosciencebioRxiv

Cite

BibTeX RIS

BibTeX

@misc{bonetti2026lowdimensional,
  title = {{A low-dimensional hypothesis for cognition and the human brain}},
  author = {Leonardo Bonetti and Barbara A. P. Grosjean and Chiara Malvaso and Nikita Kudriavtsev and Gemma Fernandez-Rubio and Mattia Rosso},
  howpublished = {PsyArXiv},
  year = {2026},
  month = aug,
  doi = {10.31234/osf.io/sn26e_v2},
  url = {https://osf.io/preprints/psyarxiv/sn26e_v2/},
  note = {CC BY 4.0},
}

RIS

TY  - UNPB
TI  - A low-dimensional hypothesis for cognition and the human brain
AU  - Leonardo Bonetti
AU  - Barbara A. P. Grosjean
AU  - Chiara Malvaso
AU  - Nikita Kudriavtsev
AU  - Gemma Fernandez-Rubio
AU  - Mattia Rosso
PY  - 2026
DA  - 2026/08/31
PB  - PsyArXiv
DO  - 10.31234/osf.io/sn26e_v2
UR  - https://osf.io/preprints/psyarxiv/sn26e_v2/
AB  - The human brain is among the most complex biological systems known, yet a growing body of evidence suggests that its activity consistently exhibits low-dimensional organisation. Across behavioural states, recording modalities, spatial scales, and analytical approaches, large-scale neural dynamics can often be described by a surprisingly limited number of latent processes expressed through coordinated large-scale spatiotemporal patterns. This observation presents a fundamental paradox: how can a system capable of remarkably flexible cognition operate within such constrained dynamical structure? Here, we propose a low-dimensional hypothesis for cognition. We argue that cognition depends on the emergence of low-dimensional latent processes from complex neural activity, their reconfiguration according to internal and external demands, and the interactions among them. Within this framework, cognition arises not from unconstrained high-dimensional neural activity, but from the efficient organisation of neural dynamics into latent processes and their interactions. We review converging evidence for low-dimensional organisation across brain activity and cognition and outline a set of falsifiable predictions concerning cognitive demands, learning, development, aging, neurological and psychiatric disorders, pharmacological interventions, and social interaction. Finally, we consider whether this principle extends across spatial scales and whether analogous organisation emerges in artificial neural networks. Altogether, we propose low-dimensional organisation as a candidate principle for understanding the neural substrate of cognition.
ER  -

Abstract

Related

Cite