Imagine unlocking the secrets of the human brain, not through dissection, but through incredibly detailed computer simulations. Sounds like science fiction? It's closer than you think! Researchers have developed groundbreaking new software that promises to revolutionize how we understand the most complex organ in our bodies. This isn't just about building a digital brain; it's about creating a tool that can mimic brain processes with unprecedented accuracy and solve challenging cognitive tasks, offering a new window into the workings of the mind.
For decades, scientists have been striving to create computer models of the brain, hoping to decipher its mysteries. These models aim to replicate the behavior and interaction of nerve cells, the fundamental building blocks of our brain. The ultimate goal? To gain a deeper understanding of how our brains function and, perhaps, even how to treat neurological disorders. But here's where it gets controversial... Previous brain models often faced a critical trade-off: either they simplified the neuron models to the point of sacrificing biological accuracy, or they focused on detailed biophysical processes but couldn't perform complex cognitive tasks. It was a frustrating "either/or" situation.
Michael Deistler, the lead author of the groundbreaking study and a researcher in Professor Jakob Macke's work group at the University of Tübingen, perfectly captures this dilemma: "Either the path is similar to that in the brain, but the result is not, or the result is correct but the process that leads there does not compare with the processes in the brain." In essence, prior models either looked like a brain but didn't act like one, or vice-versa.
Enter Jaxley, the new software developed at the Cluster of Excellence 'Machine Learning: New Perspectives for Science' at the University of Tübingen. Jaxley changes the game because it allows researchers to train brain models to both look and act like real brains. This is a crucial step forward, allowing scientists to draw meaningful conclusions about the actual brain processes from the model's behavior.
And this is the part most people miss... How does Jaxley achieve this seemingly impossible feat? The secret lies in a technique borrowed from the world of artificial intelligence: 'backpropagation of error'. This is the same method used to train artificial neural networks. Think of it like teaching a child: you provide feedback (error), and the child adjusts their behavior (parameters) until they achieve the desired outcome. Jaxley uses backpropagation to fine-tune the parameters of the brain model, such as neuron size, connection strength, and the number of ion channels, until it reliably performs the task it's been assigned. This allows the network to "learn" which features and connections are important for a specific process, enabling it to deliver correct results even with new, similar examples. The Tübingen researchers cleverly adapted this training principle to brain simulations, resulting in a significant breakthrough.
Consider this: when your brain performs a task, hundreds of parameters within the neurons are at play. The size of the neurons, the strength of their connections, the number of ion channels – all these factors contribute to the final outcome. "Many of these parameters cannot be measured. Until now this has made it impossible to develop exact simulations that produce good results," adds Deistler. Jaxley cleverly solves this by training these non-measurable parameters within the brain models. The software iteratively adjusts these values until the simulation achieves the desired result.
The result? After training, these brain models are capable of performing complex tasks like classifying images or storing and retrieving memories – tasks that require sophisticated cognitive processing. That is, the model is actually doing something useful and provides a base to study the underlying mechanisms in the brain.
"Thanks to Jaxley we can now study how neuronal mechanisms contribute to solving tasks," explains Jakob Macke, Professor of Machine Learning in Science at the University of Tübingen. "The software will allow neuroscientists to investigate the complexity of the brain and depict it in computer simulations."
But the implications extend far beyond basic research. Long-term, these simulations could revolutionize medicine. Imagine being able to understand neurological diseases better or virtually test the effects of drugs before they are administered to patients. This is a real possibility opened up by this new software.
Professor Dr. Dr. h.c. (Dōshisha) Karla Pollmann, the president of the University of Tübingen, aptly summarizes the significance of this work: "This work is a striking demonstration of how machine learning can enrich other areas of science: Artificial intelligence is a key technology which opens up new horizons for basic research." This is a powerful testament to the potential of interdisciplinary collaboration and the transformative power of AI.
But here's a question for you: Could this technology one day lead to truly conscious AI? And what ethical considerations should we be mindful of as we develop increasingly sophisticated brain simulations? Share your thoughts in the comments below! What are the limits? Are there any?
Publication:
Michael Deistler, Kyra L. Kadhim, Matthijs Pals, Jonas Beck, Ziwei Huang, Manuel Gloeckler, Janne K. Lappalainen, Cornelius Schröder, Philipp Berens, Pedro J. Gonçalves, Jakob H. Macke: Jaxley: Differentiable simulation enables large-scale training of detailed biophysical models of neural dynamics, Nature Methods (2025). https://doi.org/10.1038/s41592-025-02895-w
