EUVIMEDOtkrijte EUVIMED
Open access · CC BY 4.0

Modelling discrete states and long-term dynamics in functional brain networks

SungJun Cho iD, Rukuang Huang iD, Chetan Gohil iD, Oiwi Parker Jones iD, Mark W Woolrich iD

Imaging Neuroscience · 2026-05-26 · 4

DOI: 10.1162/imag.a.1237 ↗

At a glance

Journal article in Imaging Neuroscience, published online on 2026-05-26. Written by SungJun Cho and 4 co-authors. Institution: Oxford Centre for Integrative Neuroimaging (OxCIN), University of Oxford, Oxford, United Kingdom. 57 references. The full text is freely available (open access, CC BY 4.0). Funded by: Medical Research Council, Hertford College, University of Oxford, Nuffield Department of Clinical Neurosciences, University of Oxford.

SažetakSubjectsBibliografski podaciCiteRelated studies

Abstract (original language)

Functional brain network dynamics underlie fundamental aspects of human cognition and behaviour, including memory, ageing, and a range of clinical disorders. It has been shown that ongoing brain network dynamics can be reliably inferred at fast, sub-second timescales from electrophysiological data using unsupervised machine learning. However, these methods often struggle with inherent trade-offs. For example, Hidden Markov Models (HMMs) have been used to infer categorical brain network states that provide good interpretability but do not model long-range temporal structure. Recently, deep learning approaches using recurrent neural networks (e.g., Dynamic Network Modes) have been proposed to model long-range temporal dependencies, but at the expense of interpretability. In this paper, we introduce Dynamic Network States (DyNeStE) to address this problem. This new model employs amortised Bayesian inference with recurrent neural networks to model long-range temporal structure and uses a Gumbel-Softmax distribution to enforce categorical states for greater interpretability. In both simulations and real resting-state magnetoencephalography data, DyNeStE was able to recover plausible dynamic brain network states and showed superior performance over the HMM in capturing long-range temporal dependencies in network dynamics. These dynamic networks were reproducible across independent data splits and build on established HMM-based findings. Together, these results highlight DyNeStE as an interpretable and temporally informative framework, capable of representing large-scale neural activity as discrete state transitions while capturing transient and long-range brain network dynamics.

Subjects

Author keywords

MEGdynamicselectrophysiologymachine learningresting-state networks

Authors and institutions

  • Oxford Centre for Integrative Neuroimaging (OxCIN), University of Oxford, Oxford, United Kingdom.
  • Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, United Kingdom.
  • Department of Psychiatry, University of Oxford, Oxford, United Kingdom.
  • Department of Engineering Science, University of Oxford, Oxford, United Kingdom.

ORCID profiles: SungJun Cho, Rukuang Huang, Chetan Gohil, Oiwi Parker Jones, Mark W Woolrich

Funding

  • Medical Research Council
  • Hertford College, University of Oxford
  • Nuffield Department of Clinical Neurosciences, University of Oxford
  • EPSRC Centre for Doctoral Training in Health Data Science
  • Wellcome Trust
  • Royal Society
  • National Science Foundation
  • NFRF

Cite

APA

Cho, S., Huang, R., Gohil, C., Jones, O. P., & Woolrich, M. W. (2026). Modelling discrete states and long-term dynamics in functional brain networks. Imaging Neuroscience, 4. https://doi.org/10.1162/imag.a.1237

Vancouver

Cho S, Huang R, Gohil C, Jones OP, Woolrich MW. Modelling discrete states and long-term dynamics in functional brain networks. Imaging Neuroscience. 2026;4. doi:10.1162/imag.a.1237

BibTeX
@article{Cho2026,
  author = {Cho, SungJun and Huang, Rukuang and Gohil, Chetan and Jones, Oiwi Parker and Woolrich, Mark W},
  title = {Modelling discrete states and long-term dynamics in functional brain networks},
  journal = {Imaging Neuroscience},
  year = {2026},
  volume = {4},
  doi = {10.1162/imag.a.1237}
}
Start for free →