Some mental disorders and diseases of the nervous system represent major challenges in their approach. A European consortium, which includes the University of Rey Juan Carlos (URJC) and the Polytechnic University of Madrid, is looking Create virtual brain twins to develop personalized treatments and more effective.
On January 1, the “Virtual Brain Twin for Personalized Treatment of Psychiatric Disorders” project was launched, based on simulation of neural microcircuits, mathematical analysis, artificial intelligence tools and insights from psychiatric care and clinical studies.
One of the Spanish teams led by the URJC professor Oscar David Roblesdesigns, builds, and develops a unified framework for visual exploration of data at multiple levels.
At Digital Brain Twin we want to create a digital twin of the brain so that doctors can recreate the patient’s clinical situation with schizophrenia, stroke and epilepsy.
“With Digital Brain Twin, we plan to create a digital twin of the brain so that doctors can recreate the patient’s clinical situation with schizophrenia, stroke and epilepsy. With it, Simulate brain activity and decide on medication adjustmentLifestyle habits and so on,” he explains to SINC.
The goal is to generate visualization structures so that “neuroscientists can observe what they are doing.” In addition, another main task of this team is to create a visual language for all actors involved, which includes “the people who collect data, those who study and work with the physical structure, those who perform simulations, those who create models” . etc,” he adds.
A mathematical brain for simulations
The manner in which Virtual Brain Twin comes about remains to be defined. Robles acknowledges that they haven’t yet projected this common visual language framework, but believes it doesn’t necessarily have to be 3D. “When you make something three-dimensional, what you have in the foreground hides what is behind you,” he clarifies. Therefore, it makes more sense for him to wear something like capes.
In addition, he believes that it is also complicated in real time because “it is necessary to attach electrodes to people and it is a little uncomfortable.” The idea is to use and simulate precise mathematical models of how the brain works, when certain neurons are activated and when not.
“The concept is to create a mathematical organ based on a patient’s own brain and then simulate how the activity occurs,” says the researcher. Therefore, it is expected that it will be possible to vary the simulation parameters if there is an error or they do not correspond to the plan, allowing “changes and modified fiction in real time”.
The idea is that a mathematical organ can be created based on a patient’s own brain and then simulated as the activity occurs.
In this way, the effect of a variation in medication or treatment can be shown. “By entering the simulation you can define parameters, such as the drug used, its effect or the signals it stimulates. All of this can be studied using mathematical models. With the digital twin we can see how the brain works and whether you are really helping the patient,” he adds.
“With epilepsy, there are people who are very unwell and have to undergo ‘blind’ surgery in the hope that it will work. This is very invasive, so it is ideal if the medication is as individual as possible,” explains Robles. Another advantage of this system is that it avoids certain risks arising from other tests.
“In functional resonance measurements, sometimes you cannot see all the activity and in other cases there are contraindications to using radioactive isotopes. It’s not that they need to be isolated, but they can’t approach pregnant women, for example. That is why this is proposed,” says the URJC professor.
Visualization of a simulation of neuronal activity in the hippocampus. / Michele Migliori
Challenges for the future
The researcher believes that Virtual Brain Twin will be a tool based on abstract representations, that is, it will use boxes, circles and simple symbology for two reasons.
The first reason is that “a complete brain has millions and millions of structures.” The volume is so large that if you wanted to describe it in detail in 3D, you wouldn’t find enough machines to be able to recreate it. “Right now, I don’t think there’s anything in the world that can do all the math,” Robles says.
Additionally, working with symbolic representations would make it easier to work with different resolutions or different scale levels. “For example, the cerebral cortex has information structured into six layers. You can just plot the layers and visualize for each layer what you’re doing underneath,” he elaborates.
The professor admits that there are no plans to use augmented or virtual reality in this first phase, although “it could be possible”. However, what they have used so far for these visualizations comes from the world of video games.
This gives each particle a certain degree of transparency. You can’t see exactly what’s behind it, but you can see a collection of colors that gives an idea of how much actual activity there is in a particular area of the brain. “You have a lot of information at a glance. This is one of the goals: to use technology that is not strictly scientific for scientific purposes,” says the researcher.
Three years until this digital twin
The project was funded with 10 million euros as part of the Horizon Health Europe program and will run for four years. The University of Marseille, led by Professor Viktor Jirsa, leads the consortium.
The roadmap drawn up by the consortium assumes that a digital model of the brain will be available within three years and medical cases will be taken into account. “There will be a real impact. But it would be crazy to think that this will be the case in any hospital on a large scale by then,” concludes Robles.