Neuronal mechanisms of language processing

Timothée Proix has received an SNF Starting Grant and will start this research at ETH Zurich in the summer. The grant aims to discover how our brains’ neuronal activity allows us to build complex sentences from simple sounds and words. This will also include a clinical study offering a unique opportunity to record single neuron activity in humans during speech processing.

Hi Timothée, you received an SNF Starting Grant for your project “From spikes to language: bundling and binding in neural manifolds”. Congratulations! What are the study’s primary goals in a few sentences?
My project aims to understand the neural mechanisms by which we combine individual sounds, first into meaningful words and then complex sentences. Specifically, we will study groups of individual neurons to discover how neurons interact together to make these combinations possible. To do this, we will use both unique recordings of neuronal activity in the human brain and mathematical modeling.

Investigating language processing is interesting in a variety of research fields. Are there any inter- or transdisciplinary cooperations planned?
One important collaboration we have planned is with neurologists in several Swiss hospitals. We will collect datasets using electrodes implanted intracranially for clinical reasons in patients with epilepsy. These datasets are unique and can also be used to address other research questions. Another collaboration that will take place during the project is with researchers at the Institute of Neuroinformatics, my host institution. They are working on algorithms based on neuronal action potentials, the almost instantaneous activity that neurons use to communicate with each other.

How did the idea for the project emerge?
The project builds on the research I have been doing for the last four years in my Ambizione project. We have been working on how the different elements of language, such as the sounds or the word meanings, can be inferred from the neuronal activity in the human brain. The next logical question is to ask how these elements are combined to form complex sentences. The key ideas lie in the mechanisms at the neuronal level that we propose. Recent discoveries in machine learning and dynamical systems theory [editor’s note: a field of mathematics to investigate complex dynamic systems] inspire these mechanisms.

What was your personal motivation to initiate the project?
Creating complex sentences is a skill uniquely developed in humans compared to other animals. An exciting research question is to understand what in the human brain makes this possible. Answering this question could help to understand one of the neuronal mechanisms that allows humans to perform complex cognitive tasks. This is an important motivation for many neuroscientists, including myself.

Are you anticipating any specific challenges during the study?
The datasets we collect will be particularly large and rich, presenting unique challenges for analysis. However, we have already developed several tools that should help in this process. On the theoretical side, the algorithms we will develop make specific hypotheses about how complex sentences can be described by neuronal activity. These may be incorrect; in which case we would test alternative hypotheses.

What’s the study’s significance for society?
One important potential application of our research is helping patients with language disorders recover from stroke. An experiment we will conduct can be used to develop personalized treatments for these patients. Another critical application is to apply the algorithms we will create to neuromorphic computers. These computers operate similarly to the brain by using action potentials, making them more energy efficient than today’s computers.

Anything else about the project or the people involved, you would like to tell?
This project is possible because many patients suffering from epilepsy agree to participate in our language experiments, knowing that the benefits will not directly affect them but may help other people suffering from language disorders. I would like to take this opportunity to thank them, their contribution is invaluable to many research projects, including this one.

Thank you and all the best for your research!

Useful links:

Images:
Joshua Hoehne on unsplash