Marja-Leena Linne aims to unravel the mysteries of the brain

I wonder what the weather will be like tomorrow.

The forecast says it will be rainy and windy, but no meteorologists have climbed to the top of a mountain to confirm the arrival of a weather front. Instead, data collected by weather stations is fed to a computer that calculates how the weather will change over time.

This approach also lends itself to the study of the human brain: scientists feed experimental data to a sophisticated computational model, which can then be used to simulate and predict brain activity under different conditions.

“My greatest dream is to understand what is happening in the human brain at the cellular and molecular levels and how these processes contribute to whole-brain dynamics,” says Docent Marja-Leena Linne. She leads the Computational Neuroscience (CNS) research group based on the Hervanta campus of Tampere University.

Here in the depths of the Sähkötalo building, scientists push the boundaries of computational neuroscience. For instance, the CNS group has conducted research demonstrating that glial cells in the brain actively contribute to learning.

Glial cells were long believed to play a passive role, providing physical support for neurons and helping to remove waste products from the brain. However, a study by the CNS group confirmed that glial cells are critical for the transmission of information from one neuron to another across synapses.

The study was conducted under the umbrella of the Human Brain Project (HBP), a major EU-funded research initiative that came to an end last autumn. The HBP took 10 years, some €600 million and more than 150 organisations from 19 countries and resulted in more than 2,500 publications.

“The HBP was a massive undertaking but extremely interesting. We now have an excellent network of contacts all over Europe that opens up new opportunities for collaboration,” Linne says.

Computational modelling helps to illuminate the brain’s inner workings

Without computational modelling, it would be next to impossible to study the cellular mechanisms of human brain function and the multitude of connections between cellular-level activity and whole-brain activity.

Brain activity can be indirectly measured without opening the skull with a non-invasive technique called functional Magnetic Resonance Imaging (fMRI). Invasive studies are only possible in special cases. For example, scientists may be able to piggyback on epilepsy surgery to study the brain.

Computational modelling helps scientists not only to shed light on fundamental brain mechanisms but also to investigate the onset of different brain disorders. Scientists use computational models to test and validate their hypotheses before moving on to real-world experiments with cell cultures, brain samples and animal models. This saves a great deal of time and money.

Photo: Jonne Renvall

“One advantage of computational modelling is its ability to mimic complex interactions between multiple factors. We can, for example, manipulate all the parameters and observe the effects on the entire system,” Linne says.

In the future, it may even be possible to create whole-brain models of individual patients. The models could be used in hospitals to support the diagnosis of brain diseases.

In fact, a clinical trial is already underway in 13 hospitals in France to test whether personalised brain models of patients with epilepsy are effective as a diagnostic tool. Computational neuroscience is already making its way into clinical practice.

EBRAINS facilitates and accelerates brain research

As computational models are not yet accurate enough to reliably predict human brain activity under all conditions, more research is needed.

The pan-European Human Brain Project was launched to develop the necessary technologies and infrastructure to advance this research. One of the outcomes of the HBP is EBRAINS, a digital infrastructure that gives scientists easy access to the latest research in neuroscience as well as high-performance computing resources, all in an effort to accelerate brain research and the discovery of new treatments for brain diseases across Europe.

The CNS group headed by Linne participates in the further development of EBRAINS to ensure the platform serves scientists as effectively as possible.

“We are striving to make the resources accessible even to scientists with limited training in computational modelling and simulation. A further goal is to make sure new data is automatically incorporated into the computational models to improve their capacity to accurately predict brain activity,” Linne points out.

EBRAINS is funded under the EU’s Horizon Europe programme and is available to scientists, clinicians, students, patients and all citizens interested in their health.

The Google Maps of the brain reveals brain regions and gives access to datasets

Perhaps the greatest scientific breakthrough resulting from the Human Brain Project is the digital brain atlases that provide comprehensive maps of neuroanatomy and brain regions.

Linne compares these brain atlases to Google Maps. First, scientists can see the different brain regions that resemble countries, seas and mountains on a map. Then they can zoom in to get a closer look at the structural details and neurons, just like Google Maps lets you explore cities, villages and buildings up close. 

“With Google Maps, you can find local restaurants and open their websites. Similarly, the brain atlases contain links to the original sources of data and to the functional brain models created from these datasets,” Linne describes.

Photo: Jonne Renvall

Another major discovery in the Human Brain Project was to show how changes in neuronal activity are reflected at the whole-brain level and in the brain signals that can be recorded from outside the skull. Expertise in glial cells was brought to this study by the CNS group headed by Linne.

“It was previously believed that neuronal activity could be directly inferred from fMRI signals. We were the first to create a precise computational model demonstrating that glial mechanisms also contribute to the signals,” she says.

Interests shifted from power grids and transformers to neural communication

Perhaps surprisingly, Marja-Leena Linne did not start out with a plan to become a neuroscientist.

She originally studied physics and electrical engineering at the then Tampere University of Technology. During the summers, she worked at a power plant, sizing distribution transformers, mapping power grids and performing grid condition assessments.

In the spring of her fourth year of studying, Linne came across an advertisement for a summer job to study electrical signalling in neurons. Immediately captivated, she decided to apply.

“That summer radically changed the course of my career. I am still on the same path.”

Linne specialised in biomedical engineering while studying for a master’s degree and after graduation ended up working in Germany, developing brain imaging technology. While writing her doctoral dissertation, she also studied neurobiology, molecular biology and genetics. Linne’s knowledge of electrical engineering came in handy when she was building amplifiers for detecting weak signals transmitted by neurons.

Linne has been leading the Computational Neuroscience group on the Hervanta campus for two decades. The group is currently studying, for example, the interactions between neurons and glial cells in different brain regions, the mechanisms of memory formation and learning, and the cellular and network mechanisms underlying the pathogenesis of neurodevelopmental disorders. 

The group is also developing advanced artificial intelligence that replicates the human brain’s ability to learn.

“It is fruitful to develop AI from a neuroscientific perspective. The human brain only needs a small amount of energy and data to learn complicated things. We develop computational models that attempt to emulate this process,” Linne says.

Read more about the EBRAINS infrastructure