Taking brain simulation to the next level – the multi-scale approach

HBP researchers are simulating the brain with virtual models, an approach which is yielding novel insights into the organ’s function. These advanced technologies have enabled powerful new approaches in clinical neuroscience. Now, HBP scientists are modelling multiple scales at the same time.

Simulation is a powerful tool to better understand and predict how complex systems behave and evolve. This is particularly true when looking at the human brain. Within the HBP, researchers integrate neuroscience data into models and simulate the brain on different scales. By linking experiment and theory in this way they have shed new light on brain function and have opened the door for novel clinical applications. Now, HBP researchers have taken simulation to the next level – or rather to multiple levels at once.

Different layers of a personalised brain network model. Copyright: Charité/Petra Ritter

All scientific phenomena are observed at defined scales, and brain activity is no different. However, if you focus on one scale, the trade-off is that you either miss the whole picture of what’s happening in the brain, or you lose key details at smaller scales. Many questions can only be understood by looking at several layers of brain activity at the same time – with a so-called “multi-scale” approach, which had long been considered out of reach.

Over the years, the HBP has empowered researchers to push forward and scale up different approaches of computational brain modelling and simulation, representing brain mechanisms from the smallest level to the whole brain. In the final phase of the project, different simulation engines on the EBRAINS research infrastructure are being interlinked to enable multi-scale simulation with connected platforms. The collaborative environment, computing power and digital tools of the HBP are now advanced enough to enable researchers to construct the first multi-scale models, allowing them to study the brain in unprecedented ways.

The range and predictive power of the HBP’s simulation approaches have increased in recent years, and personalised brain modelling has become a new approach for understanding and treating debilitating neurological diseases. This includes advances in the treatment of epilepsy: seizures and the effects of surgical intervention can now be predicted by simulating virtual brain models that are based on measurements of the individual patient’s brain.

The roots of this progress reach all the way back to the early stages of the HBP, when a team from Aix-Marseille University began to adapt “The Virtual Brain” simulation engine for epilepsy – the first clinical application of the technology. The team wanted to develop a tool to improve the success rates of epilepsy surgeries, which are the main therapy option for millions of people with the drug resistant form of the disease.

Since then, much has happened: the proof-of-concept studies in clinics in collaboration with epilepsy doctors were a success; a long-term clinical trial has been funded by the French state and is currently running in thirteen hospitals; industry partner Dassault Systemes came on board; “The Virtual Epileptic Patient” has become an open platform on the HBP’s EBRAINS infrastructure (Schirner et al. 2022), and the team from Marseille has founded the VB-Tech spin-off for commercialisation (see p. 55). In 2023, the team presented the detailed novel methodology that is applied in the clinical trial (Wang et al. 2023). 

Patient-specific brain imaging data are combined with computational models to locate the epileptic zone. Copyright: INS UMR 1106

While The Virtual Brain is currently being adapted to address additional diseases, the team in Marseille is using the extensive neuroscience resources on EBRAINS to push the model’s predictive power to new limits: The team is making major steps towards bringing high-resolution anatomical data from the HBP’s human brain atlas into the simulation framework. Their latest release of The Virtual Brain is now fully integrated with the digital atlas tool siibra, which facilitates the incorporation of brain region features from different sources into the computational models. This was only made possible through the tight linkage of services on EBRAINS and the close collaboration of teams in the HBP community.

HBP researchers have already applied multi-scale simulations to target Parkinson’s disease (Meier et al. 2022). A team from the Charité in Berlin has generated the first multi-scale model of how a Parkinson’s brain responds to deep brain stimulation, a common treatment, whose outcomes have, thus far, been hard to predict. Simulating electric stimulation across multiple levels of brain networks can help clinicians preview their likely effects and plan therapies accordingly.

In the case of Parkinson’s, it is not enough to focus on the activity of single neurons or the small subcortical nuclei, because it neglects what happens at the whole-brain scale. The new approach enables researchers to both monitor some structures in high detail, spatially and also at a temporal scale, and to observe the whole-brain effect of the simulation. The study marks the first published case of a multi-scale co-simulation of the human brain applied for a clinical use case, and the methodology has been made openly available on EBRAINS. This could be translated into future medical applications that improve prediction and personalisation when performing deep brain stimulation.

The multi-scale approach is also applied to HBP research on consciousness. Theoretical neuroscientists at the University of Paris-Saclay model brain states from the microscopic scale up to the whole-brain level (Goldman et al. 2023). They work in close collaboration with experimental and clinical colleagues in the HBP who study consciousness and its disorders at the University of Milan and the University of Liege. Other multi-scale approaches have yielded new insights into plasticity, a brain mechanism important for learning (van Keulen et al. 2022) (see p. 51), and into the cerebellum, a part of the brain central for motor control (De Schepper et al. 2022).

While neural modelling and simulation have made their first major steps into large-scale clinical translation, research on basic mechanisms of brain diseases using these approaches has been broad and dynamic. For example, HBP researchers have applied modelling approaches to better understand patient outcomes in ALS (Polverino et al. 2022) and effects of damage to brain connectivity in Multiple Sclerosis (Sorrentino et al. 2022), neural stimulation effects on depression (An et al. 2022 ) and brain aging (Escrichs et al. 2022), and they have shown how AI-based brain simulations can be used to improve the classification of Alzheimer’s disease, accurately classifying patients at different stages of the disease (Triebkorn et al. 2022).

By necessity, neuroscience has traditionally been separated into top-down and bottom-up approaches. Top-down refers to looking at the whole brain and extrapolating what’s happening at a smaller scale and bottom-up means analysing a smaller scale phenomenon and drawing conclusions on what happens on the whole-brain level. Combining these two approaches is now not only possible, as recent HBP breakthroughs have shown, but indeed necessary (d’Angelo & Jirsa 2022). Multi-scale modelling and simulation have not only added value to research but have become a guiding principle for modern neuroscience.

This text was first published in the booklet ‘Human Brain Project – A closer look at scientific advances’, which includes feature articles, interviews with leading researchers and spotlights on latest research and innovation. Read the full booklet here.


An S, Fousek J, Kiss ZHT, Cortese F, van der Wijk G, McAusland LB, Ramasubbu R, Jirsa VK, Protzner AB (2022). High-resolution virtual brain modeling personalizes deep brain stimulation for treatment- resistant depression: Spatiotemporal response characteristics following stimulation of neural fiber pathways. NeuroImage 249:118848. doi: 10.1016/j.neuroimage.2021.118848

Angelo E, Jirsa V (2022). The quest for multi-scale brain modeling. Trends Neurosci. 45(10):777-790. doi: 10.1016/j.tins.2022.06.007

De Schepper R, Geminiani A, Masoli S, Rizza MF, Antonietti A, Casellato C, D’Angelo E (2022). Model simulations unveil the structure- function-dynamics relationship of the cerebellar cortical microcircuit. Commun. Biol. 5(1):1240. doi: 10.1038/s42003-022-04213-y

Escrichs A, Sanz Perl Y, Martínez-Molina N, Biarnes C, Garre-Olmo J, Fernández- Real JM, Ramos R, Martí R, Pamplona R, … Deco G (2022). The effect of external stimulation on functional networks in the aging healthy human brain. Cereb. Cortex 33(1):235-245. doi: 10.1093/cercor/bhac064

Goldman JS, Kusch L, Aquilue D, Yalçınkaya BH, Depannemaecker D, Ancourt K, Nghiem TE, Jirsa V, Destexhe A (2023). A comprehensive neural simulation of slow-wave sleep and highly responsive wakefulness dynamics. Front. Comput. Neurosci. 16 doi: 10.3389/fncom.2022.1058957

Meier JM, Perdikis D, Blickensdörfer A, Stefanovski L, Liu Q, Maith O, Dinkelbach HÜ, Baladron J, Hamker FH, Ritter P (2022). Virtual deep brain stimulation: Multi-scale co-simulation of a spiking basal ganglia model and a whole-brain mean-field model with The Virtual Brain. Exp. Neurol. 354:114111. doi: 10.1016/j.expneurol.2022.114111

Polverino A, Troisi Lopez E, Minino R, Liparoti M, Romano A, Trojsi F, Lucidi F, Gollo L, Jirsa V, Sorrentino G, Sorrentino P (2022). Flexibility of Fast Brain Dynamics and Disease Severity in Amyotrophic Lateral Sclerosis. Neurology 99(21):e2395-e2405. doi: 10.1212/WNL.0000000000201200

Schirner M, Domide L, Perdikis D, Triebkorn P, Stefanovski L, Pai R, Prodan P, Valean B, Palmer J, Langford C, Blickensdörfer A, … Ritter P (2022). Brain simulation as a cloud service: The Virtual Brain on EBRAINS. NeuroImage 251:118973. doi: 10.1016/j.neuroimage.2022.118973

Sorrentino P, Petkoski S, Sparaco M, Troisi Lopez E, Signoriello E, Baselice F, Bonavita S, Pirozzi MA, Quarantelli M, Sorrentino G, Jirsa V (2022). Whole- Brain Propagation Delays in Multiple Sclerosis, a Combined Tractography-Magnetoencephalography Study. J. Neurosci. 42(47):8807-8816. doi: 10.1523/JNEUROSCI.0938-22.2022

Triebkorn P, Stefanovski L, Dhindsa K, Diaz-Cortes MA, Bey P, Bülau K, Pai R, Spiegler A, Solodkin A, Jirsa V, McIntosh AR, Ritter P, Alzheimer’s Disease Neuroimaging Initiative (2022). Brain simulation augments machine-learning- based classification of dementia. Alzheimers Dement (N Y) 8(1):e12303. doi: 10.1002/trc2.12303

van Keulen SC, Martin J, Colizzi F, Frezza E, Trpevski D, Cirauqui Diaz N, Vidossich P, Rothlisberger U, Hellgren Kotaleski J, Wade RC, Carloni P (2022). Multi-scale molecular simulations to investigate adenylyl cyclase-based signaling in the brain. Wires e1623. doi: 10.1002/wcms.1623

Wang HE, Woodman M, Triebkorn P, Lemarechal JD, Jha J, Dollomaja B, Vattikonda AN, Sip V, Medina Villalon S, Hashemi M, Guye M, Makhalova J, Bartolomei F, Jirsa V (2023). Delineating epileptogenic networks using brain imaging data and personalized modeling in drug-resistant epilepsy. Sci. Transl. Med. 15(680):eabp8982. doi: 10.1126/scitranslmed.abp8982

2023-05-24 07:47:36