Skip to content
DraconiticResearch
Search Sources API
194,161 studies31 sourcesUpdated 38 min ago
Peer-reviewedCC BY 4.0Journal of Computational Neuroscience

A spatially discretized convolutional neural mass model for studying meso-scale spatio-temporal transformations in the rat hippocampus

  • Duy-Tan J. Pham1,2,3
  • Gene J. Yu4
  • Gianluca Lazzi3,5,6
  • Jean-Marie C. Bouteiller1,2,3,5
6 affiliations
  1. 1Neural Systems Computational Modeling Lab (NESCOM), University of Southern California, Los Angeles, CA USA
  2. 2Center for Neural Engineering (CNE), University of Southern California, Los Angeles, CA USA
  3. 3Alfred E. Mann Department of Biomedical Engineering, Viterbi School of Engineering, University of Southern California, Los Angeles, CA USA
  4. 4Brain Stimulation Engineering Lab, Department of Psychiatry and Behavioral Sciences, Duke University, Durham, NC USA
  5. 5Institute for Technology and Medical Systems (ITEMS), Keck School of Medicine, University of Southern California, Los Angeles, CA USA
  6. 6Department of Electrical Engineering, Viterbi School of Engineering, University of Southern California, Los Angeles, CA USA
Read full textDOICite

Abstract

The brain operates across multiple spatial and temporal scales, necessitating computationally efficient models that link micro-scale mechanisms to meso- and macro-scale dynamics. Here, we introduce a novel convolutional neural mass model (CNMM) that computes the meso-scale activity of spatially discretized neural populations (“neural masses") in the rat hippocampal CA3 subregion. The CNMM employs a kernel-based architecture, leveraging first-order Volterra expansions with Laguerre (temporal) and Chebyshev (spatial) basis functions to transform input spike densities from entorhinal cortex (EC), dentate gyrus (DG), and neighboring CA3 masses into output CA3 spike density. The model was trained and validated using data from a biophysically detailed large-scale mechanistic model (LSM) simulating exploratory behavior. The CNMM achieved high predictive accuracy for spike density across 32 neural masses spanning the entire extent of CA3 (mean correlation coefficient R = 0.951 ± 0.027) and replicated theta and beta oscillations consistent with experimental findings. When extended for forward modeling, the CNMM accurately predicted local field potentials (LFPs) at a single neural mass (R = 0.952), demonstrating feasibility of this approach. Kernel analysis revealed topographic gradients in afferent integration, with DG inputs dominating proximally (CA3c) and associational connections distally (CA3a), aligning with anatomical gradients. Compared to the LSM, the CNMM provided a 658-fold speedup in simulation time, 322-fold reduction in memory usage, and 183-fold less disk space for LFP predictions. This framework offers a scalable, efficient approach for meso-scale modeling of neural tissue, bridging detailed simulations with empirical data and laying groundwork for future investigations into both normal and pathological brain function.

Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1007/s10827-026-00950-9.

Related

  1. Bravais Lattice Sampling: Geometry-Guided Sparse Probing for Connected-Component Detection in 3D Discretized Spaces
    F. Carrascoza
    bioRxivPreprintCC BYFull textBioinformatics & computational biologybioRxiv
  2. Abstract representation in the hippocampus predicts spontaneously adopted structure-based behavioral strategy
    Y. Jiang, Y. Liu, X. Zhao
    bioRxivPreprintAll rights reservedFull textNeurosciencebioRxiv
  3. Unitary synaptic currents between lacunosum‐moleculare interneurones and pyramidal cells in rat hippocampus
    Sandrine Bertrand, Jean‐claude Lacaille
    The Journal of PhysiologyPeer-reviewedLicence not statedFull text34 citedNeuroscienceHAL
  4. High-dimensional hippocampus representation facilitates learning of low-dimensional abstract representation and supports generalization
    B. Wang, W. J. Johnston, S. Fusi
    bioRxivPreprintAll rights reservedFull textNeurosciencebioRxiv
  5. Hippocampus-evoked polysynaptic responses in the medial prefrontal cortex are attenuated in aged rats.
    S. Srivathsa, A. Vishwanath, V. M. Garza +2
    bioRxivPreprintAll rights reservedFull textNeurosciencebioRxiv

Cite

BibTeXRIS
BibTeX
@article{pham2026spatially,
  title = {{A spatially discretized convolutional neural mass model for studying meso-scale spatio-temporal transformations in the rat hippocampus}},
  author = {Duy-Tan J. Pham and Gene J. Yu and Gianluca Lazzi and Jean-Marie C. Bouteiller},
  journal = {Journal of Computational Neuroscience},
  year = {2026},
  month = aug,
  doi = {10.1007/s10827-026-00950-9},
  url = {https://doi.org/10.1007/s10827-026-00950-9},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - A spatially discretized convolutional neural mass model for studying meso-scale spatio-temporal transformations in the rat hippocampus
AU  - Duy-Tan J. Pham
AU  - Gene J. Yu
AU  - Gianluca Lazzi
AU  - Jean-Marie C. Bouteiller
PY  - 2026
DA  - 2026/08/10
JO  - Journal of Computational Neuroscience
DO  - 10.1007/s10827-026-00950-9
UR  - https://doi.org/10.1007/s10827-026-00950-9
AB  - The brain operates across multiple spatial and temporal scales, necessitating computationally efficient models that link micro-scale mechanisms to meso- and macro-scale dynamics. Here, we introduce a novel convolutional neural mass model (CNMM) that computes the meso-scale activity of spatially discretized neural populations (“neural masses") in the rat hippocampal CA3 subregion. The CNMM employs a kernel-based architecture, leveraging first-order Volterra expansions with Laguerre (temporal) and Chebyshev (spatial) basis functions to transform input spike densities from entorhinal cortex (EC), dentate gyrus (DG), and neighboring CA3 masses into output CA3 spike density. The model was trained and validated using data from a biophysically detailed large-scale mechanistic model (LSM) simulating exploratory behavior. The CNMM achieved high predictive accuracy for spike density across 32 neural masses spanning the entire extent of CA3 (mean correlation coefficient R = 0.951 ± 0.027) and replicated theta and beta oscillations consistent with experimental findings. When extended for forward modeling, the CNMM accurately predicted local field potentials (LFPs) at a single neural mass (R = 0.952), demonstrating feasibility of this approach. Kernel analysis revealed topographic gradients in afferent integration, with DG inputs dominating proximally (CA3c) and associational connections distally (CA3a), aligning with anatomical gradients. Compared to the LSM, the CNMM provided a 658-fold speedup in simulation time, 322-fold reduction in memory usage, and 183-fold less disk space for LFP predictions. This framework offers a scalable, efficient approach for meso-scale modeling of neural tissue, bridging detailed simulations with empirical data and laying groundwork for future investigations into both normal and pathological brain function.  Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1007/s10827-026-00950-9.
SN  - 0929-5313
SN  - 1573-6873
ER  -