Community-based Reconstruction and Simulation of a Full-scale Model of Region CA1 of Rat Hippocampus

Armando Romani, Alberto Antonietti, Davide Bella, Julian M. L. Budd, E. Giacalone, Kerem Kurban, Sára Sáray, Marwan Abdellah, Alexis Arnaudon, Elvis Boci, Cristina Colangelo, Jean-Denis Courcol, Thomas Delemontex, András Ecker, Joanne Falck, Cyrille Favreau, Michael Gevaert, Juan Hernando, Joni Herttuainen, Genrich Ivaska, Lida Kanari, Anna-Kristin Kaufmann, James King, Pramod Kumbhar, Sigrun Lange, Huanxiang Lu, Carmen Alina Lupaşcu, Rosanna Migliore, Fabien Petitjean, Judit Planas, Pranav Rai, Srikanth Ramaswamy, Michael Reimann, Juan Luis Riquelme, Nadir Román Guerrero, Ying Shi, Vishal Sood, Mohameth François Sy, Werner Van Geit, Liesbeth Vanherpe, Tamás F. Freund, Audrey Mercer, Eilif Müller, Felix Schürmann, Alex M. Thomson, Michele Migliore, Szabolcs Káli, Henry MarkrambioRxiv 2023 · DOI

TL;DR

This paper presents a comprehensive, community-driven, full-scale computational model of the rat hippocampal CA1 region, integrating vast amounts of experimental data from the synaptic to network level. This model aims to reconcile diverse experimental findings and serve as a flexible platform for investigating hippocampal function, particularly theta oscillations.

What Problem Does This Paper Address?

The CA1 region of the hippocampus is crucial for memory and spatial navigation, yet understanding its complex function has been hampered by the sheer volume and diversity of experimental data, often derived from different approaches and difficult to reconcile. Researchers face a challenge in integrating this information into a unified, quantitative framework that can explain the region's dynamics. Existing models often focus on specific phenomena or use simplified circuit structures, making it hard to generalize findings or compare results. This paper addresses the need for a comprehensive, data-driven, and scalable model that can integrate diverse experimental data and serve as a foundation for future research.

Methods

The authors employed a bottom-up approach, building the model from individual components to the full network. Key methods include:

  • Atlas-based Reconstruction: A 3D atlas of the rat CA1 region was used as a spatial framework, with parametric coordinates defining longitudinal, transverse, and radial axes.
  • Morphological and Electrical Modeling: A library of over 26,000 morpho-electrical (me-type) cell models was created by combining 2,592 morphologies with 36 electrical models. These models were systematically varied to capture neuronal diversity.
  • Connectome Algorithm: A previously developed algorithm was adapted to generate the intrinsic connectome, identifying potential synapses based on co-localization and then pruning them to match experimental bouton densities and synapse numbers.
  • Phenomenological Modeling: A model for acetylcholine's effects on neuronal excitability and synaptic transmission was developed, integrating experimental dose-response data.
  • Simulation Experiments: The model was used to simulate various experimental conditions, including changing extracellular ion concentrations, applying different stimuli, and mimicking in vitro slice preparations.
  • Data Integration and Validation: A core principle was the systematic validation of each model component against available experimental data, with assumptions made explicit and transparent.

Datasets

The paper utilized a combination of existing datasets and experimental data from various labs. Specific datasets mentioned include:

  • Morphological Reconstructions: Datasets from Sprague Dawley and Wistar rats were used for single neuron morphological reconstructions.
  • Electrophysiological Recordings: Data from Sprague Dawley and Wistar rats were used for electrophysiological recordings.
  • Hippocampal Atlas: A publicly available atlas of the CA1 region (Ropireddy et al., 2012) was used for the spatial framework.
  • Literature Data: Extensive use was made of published experimental data for constraining model parameters, validating components, and guiding simulations.

Key Insights and Ideas

  • Community-Driven Approach: The emphasis on a community-driven model, with openly available data, code, and analyses, is a significant step towards collaborative neuroscience research and reproducibility.
  • Reconciliation of Diverse Data: The model's ability to integrate data from various experimental approaches (in vitro, in vivo, different labs) is crucial for building a more complete understanding of CA1 function.
  • Theta Rhythm Generation: The study highlights that while CA3 and medial septum inputs can robustly induce theta oscillations, intrinsic CA1 mechanisms alone appear insufficient for generating regular theta activity.
  • Importance of Spatial Framework: Using a 3D atlas-based approach for cell placement and morphology orientation is critical for accurately representing the complex geometry of the hippocampus.
  • Phenomenological Modeling of Neuromodulation: Developing a phenomenological model for acetylcholine's effects allows for the investigation of neuromodulatory influences on network dynamics.

Key Results

  • Full-Scale CA1 Model: The paper successfully constructed a full-scale, atlas-based model of the rat CA1 region, comprising approximately 456,000 neurons and 821 million intrinsic synapses.
  • Theta Oscillation Reproduction: The model could reproduce theta oscillations when driven by CA3 input (particularly under in vitro-like calcium levels) and medial septum input, demonstrating its capacity to capture key network dynamics.
  • Validation Against Experimental Data: Each component of the model, from ion channels to network connectivity and synaptic physiology, was extensively validated against experimental data, showing good agreement in many cases.
  • Dose-Dependent ACh Effects: The model captured the dose-dependent effects of acetylcholine on neuronal excitability and synaptic transmission, leading to distinct activity regimes at different concentrations.
  • Frequency Propagation: The model demonstrated that intermediate frequencies (delta to low gamma band) propagate more efficiently through the Schaffer collateral pathway than very low or very high frequencies.

Limitations and Future Work

  • Missing Cell Types and Receptors: The model currently excludes neurogliaform cells and GABAB receptors, which were not present in the available datasets.
  • Simplifications in Connectivity: While the connectome algorithm is advanced, some aspects like the precise topography of afferent projections were not fully replicated.
  • Experimental Data Sparsity: For some pathways, experimental data was scarce, necessitating assumptions and leading to potential discrepancies.
  • In Vivo vs. In Vitro Conditions: While the model can simulate both, achieving robust theta oscillations under in vivo-like calcium levels proved challenging for some input types.
  • Future Work: The authors emphasize the model's potential for extension to include glial and vascular systems, as well as its role as a platform for further community-driven research and validation. The availability of the model and associated tools (hippocampushub.eu) is intended to facilitate this.

Summary updated

Abstract

Abstract The CA1 region of the hippocampus is one of the most studied regions of the rodent brain, thought to play an important role in cognitive functions such as memory and spatial navigation. Despite a wealth of experimental data on its structure and function, it has been challenging to reconcile information obtained from diverse experimental approaches. To address this challenge, we present a community-driven, full-scale in silico model of the rat CA1 that integrates a broad range of experimental data, from synapse to network, including the reconstruction of its principal afferents, the Schaffer collaterals, and a model of the effects that acetylcholine has on the system. We tested and validated each model component and the final network model, and made input data, assumptions, and strategies explicit and transparent. The unique flexibility of the model allows scientists to address a range of scientific questions. In this article, we describe the methods used to set up simulations that reproduce and extend in vitro and in vivo experiments. Among several applications in the article, we focus on theta rhythm, a prominent hippocampal oscillation associated with various behavioral correlates and use our computer model to reproduce and reconcile experimental findings. Finally, we make data, code and model available through the hippocampushub.eu portal, which also provides an extensive set of analyses of the model and a user-friendly interface to facilitate adoption and usage. This neuroscience community-driven model represents a valuable tool for integrating diverse experimental data and provides a foundation for further research into the complex workings of the hippocampal CA1 region.
BibTeX
@misc{romani2023community,
  author = {Armando Romani and Alberto Antonietti and Davide Bella and Julian M. L. Budd and E. Giacalone and Kerem Kurban and Sára Sáray and Marwan Abdellah and Alexis Arnaudon and Elvis Boci and Cristina Colangelo and Jean-Denis Courcol and Thomas Delemontex and András Ecker and Joanne Falck and Cyrille Favreau and Michael Gevaert and Juan Hernando and Joni Herttuainen and Genrich Ivaska and Lida Kanari and Anna-Kristin Kaufmann and James King and Pramod Kumbhar and Sigrun Lange and Huanxiang Lu and Carmen Alina Lupaşcu and Rosanna Migliore and Fabien Petitjean and Judit Planas and Pranav Rai and Srikanth Ramaswamy and Michael Reimann and Juan Luis Riquelme and Nadir Román Guerrero and Ying Shi and Vishal Sood and Mohameth François Sy and Werner Van Geit and Liesbeth Vanherpe and Tamás F. Freund and Audrey Mercer and Eilif Müller and Felix Schürmann and Alex M. Thomson and Michele Migliore and Szabolcs Káli and Henry Markram},
  title = {{Community-based Reconstruction and Simulation of a Full-scale Model of Region CA1 of Rat Hippocampus}},
  year = {2023},
  doi = {10.1101/2023.05.17.541167},
  howpublished = {bioRxiv},
}
Figure