Abstract
Neuron models built from experimental data have successfully predicted observed voltage oscillations within and beyond training range. A tantalising prospect is the possibility of estimating the unobserved dynamics of ion channels, which is largely inaccessible to experiment, from membrane voltage recordings. The main roadblock here is our lack of knowledge of the equations governing biological neurons which forces us to rely on surrogate models and parameter estimates biassed by model error. Error correction algorithms are therefore needed to infer both observed and unobserved dynamics, and ultimately the actual parameters of a biological neuron. Here we use a recurrent neural network to correct the outputs of a surrogate Hodgkin-Huxley (HH) model. The reservoir-surrogate HH model hybrid was trained on the voltage oscillations of a reference HH model and its driving current waveform. Out of the six reservoir-surrogate model architectures investigated, we identify one that most accurately recovers the reference membrane voltage and ion channel dynamics. The reservoir was thus effective in correcting model error in an externally driven nonlinear oscillator and in reconstructing the dynamics of both observed and unobserved state variables from the reference model mimicking an actual neuron.
Data availability
The data and code that support the findings of this study are available within the article and from the authors.
Code availability
References
Brookings, T., Goeritz, M. L. & Marder, E. Automatic parameter estimation of multicompartmental neuron models via minimization of trace error with control adjustment. J. Neurophysiol. 112, 2332–2348 (2014).
Mirallave-Pescador, A. et al. Bayesian networks for risk assessment and postoperative deficit prediction in intraoperative neurophysiology for brain surgery. J. Clin. Monit. Comput. 38, 1043 (2024).
Nogaret, A. Approaches to parameter estimation from model neurons and biological neurons. Algor. 15, 168 (2022).
Kalnay, E. Atmospheric Modelling, Data Assimilation and Predictability (Cambridge University Press, 2003).
Takens, F. Strange attractors in turbulence in dynamic systems, 366 (Springer Warwick), (1981).
Sprott, J. C. Chaos and Time-Series Analysis (Oxford University Press, 2003).
Lillacci, G. & Mustafa, K. Parameter estimation and model selection in computational biology. PLoS Comp. Biol. 6, e1000696 (2010).
Lynch, E. P. & Houghton, C. J. Parameter estimation of neuron models using in-vitro and in-vivo electrophysiological data. Front. Neuroinform. 9, 10 (2015).
Toth, B. A., Kostuk, M., Meliza, C. D., Margoliash, D. & Abarbanel, H. D. I. Dynamical estimation of neuron and network properties I: Variational methods. Biol. Cybern. 105, 217 (2011).
Wächter, A. & Biegler, L. T. On the implementation of an interior-point filter line-search algorithm for large-scale nonlinear programming. Math. Prog. Ser. A 106, 25 (2005).
Nogaret, A., Meliza, C. D., Margoliash, D. & Abarbanel, H. D. I. Automatic construction of predictive neuron models through large scale assimilation of electrophysiological data. Sci. Rep. 6, 32749 (2016).
Mirams, G. R., Clerx, M., Whittaker, D. G. & Lei, C. L. Optimal experimental designs for characterising ion channel gating by filling the phase-voltage space of model dynamics. Math. Med. Life Sci. 1, 2375494 (2024).
Morris, P. G., Taylor, J. D., Paton, J. F. R. & Nogaret, A. Single-shot detection of alterations across multiple ionic currents from assimilation of cell membrane dynamics. Sci. Rep. 14, 6031 (2024).
Kravaris, C., Hahn, J. & Chu, Y. Advances and selected recent developments in state and parameter estimation. Comput. Chem. Eng. 51, 111–123 (2013).
Wells, S. A., Taylor, J. D., Morris, P. G. & Nogaret, A. Inferring the dynamics of ionic currents from recursive piecewise assimilation of approximate neuron models. PRX Life 2, 023007 (2024).
O’Leary, T., Sutton, A. C. & Marder, E. Computational models in the age of large data sets. Curr. Opin. Neurobiol. 32, 87 (2015).
Gutenkunst, R. N. et al. Sloppy parameter sensitivites in systems biology models. PLoS Comput. Biol. 3, 1871 (2007).
Lukosevicius, M. & Jaeger, H. Reservoir computing approaches to recurrent neural network training. Comput. Sci. Rev. 3, 127–149 (2009).
Pathak, J., Hunt, B., Girvan, M., Lu, Z. & Ott, E. Model-free prediction of large spatiotemporal chaotic systems from data: A reservoir computing approach. Phys. Rev. Lett. 120, 024102 (2018).
Sussillo, D. & Abbott, L. F. Generating coherent patterns of activity from chaotic neural networks. Neuron 63, 544 (2009).
Pathak, J. et al. Hybrid forecasting of chaotic processes: Using machine learning in conjunction with a knowledge-based model. Chaos https://doi.org/10.1063/1.5028373 (2018).
Duncan, D. & Räth, C. Optimizing the combination of data-driven and model-based elements in hybrid reservoir computing. Chaos https://doi.org/10.1063/5.0164013 (2023).
Hodgkin, A. L. & Huxley, A. F. A quantitative description of membrane current and its application to conduction and excitation in nerve.. J. Physiol. 117, 500 (1952).
Parlitz, U., Schumann-Bischoff, J. & Luther, S. Local observability of state variables and parameters in nonlinear modeling quantified by delay reconstruction. Chaos 24, 024411 (2014).
Letellier, C., Aguirre, L. A. & Maquet, J. Relation between observability and differential embedding for nonlinear dynamics.. Phys. Rev. E. 71, 066213 (2005).
Rey, D. et al. Using waveform information in nonlinear data assimilation. Phys. Rev. E. 90, 062916 (2014).
Parlitz, U., Schumann-Bischoff, J. & Luther, S. Quantifying uncertainty in state and parameter estimation. Phys. Rev. E. 89, 050902 (2014).
Schumann-Bischoff, J., Luther, S. & Parlitz, U. Estimability and dependency analysis of model parameters based on delay estimates. Phys. Rev. E. 94, 032221 (2016).
Aeyels, D. Generic observability of differentiable systems. SIAM J. Control Optim. 19, 595 (1981).
Sauer, T., Yorke, J. A. & Casdagli, M. Embedology. J. Stat. Phys. 65, 579 (1991).
Wells, S. A., Morris, P. G., Taylor, J. D. & Nogaret, A. Estimation of ionic currents and compensation mechanisms from recursive piecewise assimilation of electrophysiological data.. Front. Comput. Neurosci. 19, 1458878 (2025).
Burghi, T. B. et al. Quantitative prediction of intracellular dynamics and synaptic currents in a small neural circuit. Front. Comput. Neurosci. 19, 1 (2025).
Tikhonov, A. N. On the stability of inverse problems. Proc. USSR Acad. Sci. 39, 195–198 (1943).
Meliza, C. D. et al. Estimating parameters and predicting membrane voltages with conductance-based neuron models. Biol. Cybern. 108, 495–516 (2014).
Kennel, M. B., Brown, R. & Abarbanel, H. D. I. Determining embedding dimension for phase-space reconstruction using a geometrical construction. Phys. Rev. 45, 3403 (1992).
Ebato, Y. et al. Impact of time-history terms on reservoir dynamics and prediction accuracy in echo state networks. Sci. Rep. 14, 8631 (2024).
Author information
Authors and Affiliations
Contributions
JDT demonstrated the initial the proof of concept of predicting neuronal oscillations with a reservoir driven by a time dependent current. AN conceived the approach of correcting model error in neuron-based conductance models. IW performed the simulations of hybrid architectures and generated all proof of concept data. IW and AN wrote the manuscript. All authors reviewed the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix
Appendix
Parameters of the reference Hodgkin-Huxley model:
Sodium channel: \(g_{Na}\)=69 mS.cm\(^{-2}\), \(E_{Na}\)=41mV, \(V_m\)=-39.92 mV, \(dV_m\)=10 mV, \(\tau _{0,m}\)=0143 ms, \(\epsilon _m\)= 0.1 ms, \(dV_{t,m}\)=23 mV, \(V_h\)=-65.37 mV, \(dV_h\)=-17.65 mV, \(\tau _{0,h}\)=0.701 ms, \(\epsilon _h\)= 12.9 ms, \(dV_{t,h}\)=27.22 mV.
Potassium channel: \(g_{K}\)=6.9 mS.cm\(^{-2}\), \(E_{K}\)=-100mV, \(V_n\)=-34.58 mV, \(dV_n\)=22.17 mV, \(\tau _{0,n}\)=1.291 ms, \(\epsilon _n\)= 4.314 ms, \(dV_{t,n}\)=23.58 mV.
Leak channel: \(g_{L}\)=0.165 mS.cm\(^{-2}\), \(E_{L}\)=42mV, \(C=\)1.0\(\mu\)F.cm\(^{-2}\).
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Williams, I., Taylor, J.D. & Nogaret, A. Correcting model error bias in estimations of neuronal dynamics from time series observations. Sci Rep (2026). https://doi.org/10.1038/s41598-026-43346-6
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-026-43346-6
