CNS*2026 Halifax

Introduction
Neural population activity is often described as evolving on low-dimensional manifolds that structure neural computation and behaviour. These manifolds are typically identified empirically using dimensionality reduction techniques applied to large-scale neural recordings (Gallego et al., 2017). While such approaches successfully reveal structured population dynamics, they are largely agnostic to the underlying biophysical mechanisms that generate neural activity. Consequently, the relationship between neural manifolds and the building blocks of neural tissue – morphology, ion channel, connectivity, synapse physiology, etc. – remains undefined.

Methods
Here, we propose a framework that constructs neural population dynamics from biophysical elements. Combing cable theory with Hodgkin-Huxley membrane dynamics, axial current along neuronal processes can be balanced by total membrane current across each membrane segment. Linearizing over axial and membrane kinetic elements governed by experimentally measurable conductances and gating variables, the total membrane current is expressed by contributions from voltage-gated ion channels, ligand-gated synaptic channels, electrical synapses, and electrogenic transport mechanisms, including ion pumps and exchangers. Thus longer-timescale membrane dynamics may be explicitly taken into account.

Results
We validate this framework in the compact, stereotypic, and connectomically-defined nervous system of C. elegans (White et al., 1986). Our analysis focuses on the locomotor rich club neurons, which coordinate transitions between forward and reverse states (Fieseler et al., 2025). Biophysical parameters are estimated from current-clamp electrophysiological recordings that constrain passive membrane properties and voltage-gated conductances. Population activity in this circuit is further constrained using whole-brain Ca²⁺ imaging measurements of the same neurons during behaviour (Fieseler et al., 2025). Together, these data provide experimentally grounded estimates of the membrane and coupling parameters governing the dynamical system.


Discussion
By deriving network dynamics directly from experimentally measurable cellular and circuit properties, this framework provides a mechanistic bridge between membrane biophysics and population-level neural dynamics. In this view, neural manifolds are not only empirical low-dimensional decompositions of neural activity, but also naturally emerging physical organizations of neural tissue: a graphical interaction between membrane dynamics and neuronal morphology. This approach therefore links cellular physiology, circuit topology, and population dynamics within a unified dynamical system to provide a principled foundation for interpreting neural manifolds as geometric consequences of biophysical mechanisms.


References

1. Gallego, J. A., Perich, M. G., Miller, L. E., & Solla, S. A. (2017). Neural Manifolds for the Control of Movement. Neuron, 94(5), 978-984. https://doi.org/10.1016/j.neuron.2017.05.025
2. White, J. G., Southgate, E., Thomson, J. E., & Brenner, S. (1986). The structure of the nervous system of the nematode Caenorhabditis elegans. Philosophical Transactions of the Royal Society of London. B, Biological Sciences, 314(1165), 1-340. https://doi.org/10.1098/rstb.1986.0056
3. Fieseler, C., Lev, I., Rey, U., Hille, L., Brenner, H., & Zimmer, M. (2025). An intrinsic neuronal manifold underlies brain-wide hierarchical organization of behavior in C. elegans. bioRxiv, 2025.2003.2009.642241. https://doi.org/10.1101/2025.03.09.642241

Acknowledgement
We would like to thank the European Research Commission (Advanced Grant) and Austrian Science Fund Cluster of Excellence (Neuronal Circuits in Health and Disease) for funding.