EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Toward reliable machine learning models for neural circuit inference: A diagnostic study of CNNs on spike trains

Xiaoqian Sun iD, Hui Lu, Chen Zeng iD, Rahul Simha

PLOS Computational Biology · 10.08.2026 · 22 (8) · e1014615

DOI: 10.1371/journal.pcbi.1014615 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS Computational Biology, online erschienen am 10.08.2026. Verfasst von Xiaoqian Sun und 3 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Science, School of Engineering and Applied Science, The George Washington University, Washington, District of Columbia, United…. Hauptthemen laut MEDLINE: Convolutional Neural Networks, Machine Learning, Models, Neurological, Action Potentials. Die Arbeit nennt 51 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NINDS NIH HHS, National Institutes of Health, George Washington University.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Understanding neuronal topology—how neurons are connected—is essential for uncovering neural computation principles and functional organization. However, accurately reconstructing such connectivity remains challenging due to the indirect nature of neural recordings and the complexity of network dynamics. As a first step towards this problem, a growing body of work has explored inferring monosynaptic connectivity directly from spike data. Among these, convolutional neural networks have shown promise when applied to spike-train cross-correlograms. Nevertheless, their ability to generalize across realistic experimental variability and the internal features that drive their predictions remain poorly understood. In this paper, we present a systematic benchmarking and diagnostic study of neural-network-based synaptic inference using simulations across a broad range of biophysical regimes. We show that connectivity classification and synaptic weight estimation, though often combined, rely on distinct internal representations and exhibit markedly different generalization behavior: robust connectivity models emphasize global structure in spike-train correlations, whereas weight estimation models are more sensitive to local signal amplitude and generalize less predictably. Importantly, we find that training on pooled, biologically grounded simulation data substantially improves robustness across parameter perturbations, outperforming models trained under narrow conditions. We further validate these findings in both simulated network data and an in vitro dataset from high‑density microelectrode array recordings with patch‑clamp‑verified ground‑truth connections. Models trained on diverse simulated circuits generalize effectively to novel network architectures and the experimental dataset. Together, these results demonstrate that incorporating biologically realistic diversity during training is critical for developing reliable machine-learning tools for large-scale synaptic inference from neural recordings.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Convolutional Neural NetworksMachine LearningModels, NeurologicalAction PotentialsNerve NetNeuronsAnimalsComputer SimulationComputational BiologySynapsesHumans

Autor:innen und Institutionen

  • Department of Computer Science, School of Engineering and Applied Science, The George Washington University, Washington, District of Columbia, United States of America.
  • The GW Institute for Neuroscience, The George Washington University, Washington, District of Columbia, United States of America.
  • Department of Pharmacology and Physiology, School of Medicine and Health Sciences, The George Washington University, Washington, District of Columbia, United States of America.
  • Department of Physics, Columbian College of Arts and Sciences, The George Washington University, Washington, District of Columbia, United States of America.

ORCID-Profile: Xiaoqian Sun, Chen Zeng

Förderung

  • NINDS NIH HHS
  • National Institutes of Health
  • George Washington University

Zitieren

APA

Sun, X., Lu, H., Zeng, C., & Simha, R. (2026). Toward reliable machine learning models for neural circuit inference: A diagnostic study of CNNs on spike trains. PLOS Computational Biology, 22(8), e1014615. https://doi.org/10.1371/journal.pcbi.1014615

Vancouver

Sun X, Lu H, Zeng C, Simha R. Toward reliable machine learning models for neural circuit inference: A diagnostic study of CNNs on spike trains. PLOS Computational Biology. 2026;22(8):e1014615. doi:10.1371/journal.pcbi.1014615

BibTeX
@article{Sun2026,
  author = {Sun, Xiaoqian and Lu, Hui and Zeng, Chen and Simha, Rahul},
  title = {Toward reliable machine learning models for neural circuit inference: A diagnostic study of CNNs on spike trains},
  journal = {PLOS Computational Biology},
  year = {2026},
  volume = {22},
  number = {8},
  pages = {e1014615},
  doi = {10.1371/journal.pcbi.1014615}
}
Kostenlos starten →