Is it possible to simulate the human brain? Yes, and that is the goal of The Virtual Brain, an open-source neuroinformatics platform. Interview with Professor Petra Ritter, PhD, Director of the Brain Simulation Section at Charité – Universitätsmedizin Berlin, Johanna Quandt Professor for Brain Simulation at Berlin Institute of Health at Charité (BIH), co-founder of The Virtual Brain, and leader of the Testing and Experimentation Facility Health AI and Robotics TEF-Health.
Is it possible to simulate the human brain – a complex yet unexplored organ?
It is possible, yet at a degree of resolution that is much more abstract and simplified than the real brain.
How did the Virtual Brain Project start? And what obstacles did you have to overcome at the beginning?
There were many different starting points, such as building mathematical models of the different components of animal and human brains: single neurons, neuronal circuits, and whole—but simplified—brains.
The beauty is that now these different lines of research are getting connected, and each approach is considered in light of the others, leading to multi-scale computational models of the brain.
I can imagine that to create a virtual twin of a brain, a lot of data is needed. Will the recently adopted European Health Data Space help?
Absolutely. The European Health Data Space is a regulation in preparation that aims to increase the availability of human clinical data for secondary use in research.
Access to more data will significantly enhance the development of human digital twin technologies, allowing for more accurate and comprehensive models.
What fascinates you most about the Virtual Brain?
It’s the opportunity to decipher the principles of human cognition while simultaneously creating the basis for technologies that could greatly benefit patients.
What surprised you most during the simulations? Were there any “wow” moments when the model became increasingly accurate?
One of the most surprising findings was that the virtual brains of individuals with higher fluid intelligence (the capability of thinking and solving difficult problems) took slightly more time in simulations before making decisions, but their decisions were more accurate.
Initially, we did not have access to the response time data of these 650 individuals. After our surprising discovery, we requested this data, and the reaction times of the real individuals confirmed the simulations of their digital twins. It was a remarkable validation of our model.
The virtual brain is an open science platform. Don’t you worry that some results of your research could be misused, for example, to impact our political or shopping decisions?
Indeed, precautions must be taken to prevent abuse. The AI Act—published on June 12, 2024, in the Official Journal of the European Commission—serves as a legal instrument to help prevent misuse.
Will there be a moment when you will be able to say: We have it – a complete model of a virtual brain? Or will it be a never-ending project?
Any prediction would be mere speculation. Yet, progress is speeding up, and we certainly will see an enormous evolution of these technologies.
What is needed to speed up research in this area? Quantum computers?
Quantum computers are indeed an exciting new technology that might enable us in the future to simulate brains in a way that much better approximates the functioning of real brains. However, while this expectation is promising, it currently lacks concrete proof. Significant work is required by the neuroscience community to provide the first evidence supporting this potential
Science is slowly demystifying all the secrets of life. It started with the Human Genome Project, and now we are decoding the human brain. Do you believe that through science, we will know ourselves better as humans and be able to strive for a better world? What will it do to us as humans if we one day find out that our lives are just zeros and ones?
One clear distinction between biological systems and computers is that our lives are not merely zeros and ones. Biological systems are far more complex and nuanced. The fact that quantum computers may eventually account for this complexity suggests a significant improvement in the computational modeling of brains compared to traditional computers.
Engineers around the world are trying to build AI that works exactly like the human brain. Having made progress in your research, do you think it’s possible? If yes, is there a way to create a general artificial intelligence?
I believe that AI inspired by the knowledge about real biological brains will rapidly become more powerful and develop towards general artificial intelligence. As we continue to understand and replicate brain functions, I see no reason that would prevent this development.
Another project you are leading is “Testing and Experimentation Facility Health AI and Robotics” (TEF-Health). Can you please explain its purpose?
TEF-Health is a consortium of 52 partners funded with €60 million – half from the European Commission and half from member states – for five years. It aims to build a sustainable European infrastructure that provides scientific and technological support for AI providers, such as SMEs, and notified bodies to validate AI systems to ensure compliance with the AI Act.
The AI Act includes TEFs as an implementation instrument to ensure compliance and guide the deployment of AI. Four sectorial TEFs have been established under the Digital Europe program: Healthcare (TEF-Health), Smart Cities (Citcom.ai), Agriculture (Agrifood-TEF), and Manufacturing (AI-Matters). This project is crucial for fostering innovation while maintaining rigorous standards for safety and efficac
