Blaise Yvert: Getting the Brain to Talk

Blaise Yvert has one goal – restore speech to people who are paralyzed and who have lost their vocal abilities. Along with his team at the Grenoble Institute of Neuroscience, he is developing a system capable of decoding the brain signals associated with speech, so that it can be produced by an external device. This is referred to as a brain-computer interface.

Will Yvert’s research restore speech to those who have lost it? This is what the Inserm Research Director leading the Neurotechnologies and Network Dynamics team at the Grenoble Institute of Neuroscience is hoping. For the past ten years, he has been working on the development of a brain-computer interface to decode the brain signals of speech and reproduce the words of people who are unable to utter them. His project was recently selected as part of the Impact Santé program, funded by the France 2030 investment plan and coordinated by Inserm. Brain Implant, the scientific consortium he has formed for this project, has received three million euros to develop a new brain implant that will improve the accuracy of speech reconstruction based on brain activity. 

From engineer to researcher

This desire dates back to his engineering studies at École centrale de Lyon and Cornell University in the US. « I was drawn to research and wanted to develop health technologies, especially for people with disabilities. I knew several people with disabilities when I was younger and it’s a cause I hold dear », explains Yvert. Once he graduated in 1993, the young engineer was hired by an Inserm human electrophysiology research unit in Lyon. « The team sought to mathematically locate the brain regions responsible for the signals recorded on the surface of the head. This was something I was particularly interested in », he recalls. During two postdocs, one in Finland and the other in Germany, the researcher used this approach to identify the auditory areas. But he realized that, even for very simple sounds, the cortex activation pattern is too complex to be finely understood with non-invasive recordings. « So I thought: let’s develop more sophisticated systems, for a more precise look at what happens in the neural networks. »

Towards a new technology

With this goal in mind, and after obtaining a research fellowship at Inserm, Yvert joined in 2003 a research unit in Bordeaux that focuses on neural networks in the developing spinal cord. There, he initiated a partnership with the French Alternative Energies and Atomic Energy Commission (CEA) in Grenoble and the ESIEE engineering school in Paris, which has academic laboratories, to develop microelectrode networks to enable detailed exploration of neural tissue activity in vitro. An initial prototype was finalized three years later. Through multiple collaborations, he continued to improve this technology, particularly with new materials to increase the performance of electrodes (platinum, diamond and, more recently, graphene).

Then, Yvert wanted to put his research to work for patients. With this project in mind, he spent a year at Brown University in the US, in a research unit that led the way in implantable brain-computer interfaces in humans. Back in France, he joined the Grenoble Institute of Neuroscience and began his project on decoding brain signals of speech. In particular, he collaborated with the Clinatec institute created by the CEA, « a unique environment for creating new rehabilitation strategies for people with paralysis », he believes.

The interface to which Yvert devoted his work is aimed particularly at people with “locked-in syndrome” (LIS). Although they cannot move or speak due to complete paralysis, their cognitive faculties are intact. “The cortical activities produced when they want to say something are always present, so if we can decode them with our implants, we can reproduce what they want to say”, hopes the researcher. An initial clinical trial is expected to start in 2025, « if the regulatory procedures go well », he warns. This trial will include people with LIS who will be equipped with an implant developed by Clinatec, positioned on the surface of the brain. “This device provides signals that are highly stable over the long term, with wireless transmission through the skin », he explains.

Pursue and accelerate development

At the same time, the scientist does not forget the fundamental aspect, which has always been a source of motivation in his work. « For example, we’re exploring the brain activity of a new animal model that is very vocal – the pig. This model allows us to test new, more efficient types of implants for potential future use in humans. It will also be possible to see whether there are similarities between the data collected in animals and humans ».

In order to finely decode brain activity, he believes that the devices will still need to be improved, by increasing the number of electrodes, and by innovating in materials and integrated electronics. This is the goal of the Brain Implant project. « We want to create a technological building block that would serve both basic research and to develop brain-computer interfaces for clinical use in different indications: to restore speech or other motor functions », he explains. These developments and their challenges for people and society are inevitably accompanied by ethical questions around which Yvert has set up processes of reflection, conducted in collaboration with philosophers and patient organizations. And as if all of this were not enough, the researcher has also led, since early 2025, the Grenoble Initiative in Medical Devices (LabEx GIMeD), a research partnership on medical devices. « The aim is to bring together multidisciplinary units that develop health technologies, including teams specialized in the humanities and social sciences, to reflect on the implications of these technologies. New projects are expected to emerge from this ecosystem », he outlines for the future.

Looking back, Yvert notes that risk-taking during his career has been successful. “Going from non-invasive brain recording in humans to the technological development of in vitro systems took me out of my comfort zone. But in the end, this leap was essential in preparing for the development of an interface that, I hope, will one day be able to provide real services to patients », he concluded.

Blaise Yvert is an Inserm Research Director and head of the Neurotechnologies and network dynamics team at Grenoble Institute of Neuroscience (unit 1216 Inserm/Grenoble-Alpes University) in Grenoble.