Arbor-TVB: A Novel Multi-Scale Co-Simulation Framework with a Case Study on Neural-Level Seizure Generation and Whole-Brain Propagation

Keywords: Multi-scale neural models, seizures, The Virtual Brain, Arbor

Computational neuroscience has traditionally focused on isolated scales, limiting understanding of brain function across multiple levels. While microscopic models capture biophysical details of neurons, macroscopic models describe large-scale network dynamics. Integrating these scales, remains a significant challenge. In this study, we present a novel co-simulation framework that bridges these levels by integrating the neural simulator Arbor with The Virtual Brain (TVB) platform. Arbor enables detailed simulations from single-compartment neurons to populations of such cells, while TVB models whole-brain dynamics based on anatomical features and the mean neural activity of a brain region. By linking these simulators for the first time, we provide an example of how to model and investigate the onset of seizures and their propagation to the whole brain. This framework employs an MPI intercommunicator for real-time bidirectional interaction, translating between discrete spikes and continuous activity. Its fully modular design enables independent model selection for each scale, requiring minimal effort to translate activity across simulators. The Arbor-TVB co-simulator allows replacement of nodes with biologically realistic neuron populations, offering insights into seizure propagation and potential intervention strategies. It marks a significant advancement in multi-scale modeling, providing a comprehensive computational framework for studying neural disorders and optimizing treatments.

Arbor-TVB co-simulation schematic and communication pattern. (A) In a TVB simulation of regions I, K, and P, one region P will be replaced by a proxy containing a network of detailed cells simulated in Arbor. Regions are connected via the weights of the connectome and produce an activity values based on the chosen region model. When crossing the boundary between TVB and Arbor models, care needs to be taken to convert between discrete action potentials in Arbor to continuous, region-model-specific variables in TVB. (B) Spikes generated by Arbor and TVB—converted from activity values interpreted as mean spiking rates—are exchanged using an MPI intercommunicator and the All-gather primitive. This is equivalent to concatenating all contributions from all Arbor MPI ranks and sending the result to all TVB ranks and vice-versa. (C) Activity traces for healthy and diseased neurons in the Arbor network, before and after turning on network connectivity. (D)  Activity trace for a seizure-recruited TVB node. (E) Time-distance from seizure onset in the Arbor network (star marker) in the whole brain network.

hal-05489179

Contact:
Juliette Courson juliette.courson@cyu.fr, Thanos Manos thanos.manos@cyu.fr

Affiliated/Collaborating Laboratories: Thorsten Hater¹, Juliette Courson^2,3 , Han Lu¹ , Sandra Diaz-Pier¹ and Thanos Manos²

¹JSC - Jülich Supercomputing Centre

²ETIS Lab, ENSEA, CNRS, UMR8051, CY Cergy-Paris University

³Department of Computer Science, University of Warwick

Link to the Laboratory’s Page : https://www.etis-lab.fr/

Funding: This work was supported by the NIH 1R21AG087888-01 grant. JC was supported by the LABEX MME-DII (ANR-16-IDEX-0008) PhD grant. HL was supported by EBRAINS2.0. EBRAINS 2.0 has received funding from the European Union’s Research and Innovation Program Horizon Europe under Grant Agreement No. 101147319. Open access publication funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation)—491111487.