Romain Quentin: Understanding Learning in Order to Improve It

Specializing in cognitive computational neuroscience, a field which uses mathematical tools to develop and test theories about brain function, Romain Quentin characterizes the neural mechanisms that come into play when learning. As a recipient of ATIP-Avenir funding which has enabled him to form his own team, he wishes to identify how to improve the way our brain memorizes new skills.

While errors have often been a source of innovation or unexpected discovery when it comes to research, they are also a cornerstone of cognitive learning from a very young age. Inserm researcher Romain Quentin is focused on exploring the neural processes involved in how we learn from our mistakes – for example when readjusting the direction in which we throw a basketball after missing the hoop on the previous attempt. To understand how this happens, the researcher uses the potential of a particularly effective technique known as magnetoencephalography (MEG). « This neuroimaging method records the brain’s magnetic signals with very good temporal resolution – we are talking milliseconds, and good spatial resolution – down to the centimeters and sometimes even millimeters, » he explains. When paired with machine learning algorithms, this method allows him to identify the neural mechanisms involved in maintaining error-related information from one attempt to the next.

Portrait de Romain Quentin
Romain Quentin

To conduct this research, Quentin was awarded ATIP-Avenir funding in 2021 which has given him the opportunity to hire coworkers and set up his own team. Together they conduct behavioral experiments, brain electrophysiology tests, and non-invasive brain stimulation tests aimed at solving the neural mysteries of learning. « Based on our discoveries, we would like to be able to develop interventions to improve learning processes, whether by stimulating the brain or by developing particular mental training protocols. » But for this to be possible, it is also necessary to understand the other learning processes.

Quentin has therefore expanded the scope of his laboratory’s research to encompass « statistical learning », otherwise known as the « learning of regularities ». « This occurs when short sequences of events or stimuli repeat themselves to us in the same way. Our brain eventually predicts the rest of the sequence as soon as it begins. This is thought to be an important process for learning language: the brain makes statistics and identifies the sequences of frequent syllables and those that are more random. This enables the infant to distinguish words from each other when someone is speaking to them. » And again, the mechanisms at play have yet to be identified: « We are currently exploring several hypotheses: do two associated events have neural interaction patterns that become similar? Does the brain engage in an accelerated replay of what it has learned right after the event, as seen with mice? This kind of neural replay could help us to memorize information. »

From Physics to Neurobiology

Quentin’s attraction to neuroimaging is nothing new. Early on in his academic career, the subject of his final-year undergraduate project was physics in magnetic resonance imaging (MRI) – a medical imaging technique close to MEG. This is because Quentin is not a biologist or a doctor, but a physicist. « In high school, I was passionate about the basic and cognitive sciences. To begin with, physics won out. But when a Master’s in neuroscience and neuroimaging became available at my university, I took the opportunity to combine my two centers of interest, » he said. Following this, he set his sights on research and prepared his PhD at the Brain Institute in Paris. He became an expert in tractography, an MRI technique used to identify the connections between neural networks within the white matter – an area of the brain comprised of axons, the cellular extensions of neurons. « My dissertation consisted of characterizing conscious visual perception and improving it by subjecting participants to magnetic stimulation of certain brain regions. But when we specialize in an analytical technique, we see both its potential and its limitations. This is kind of what happened to me with tractography. » While his initial intention was to use this same method to do a post-doctoral fellowship at the U.S. National Institute of Neurological Disorders and Stroke (NIH, NINDS), in the end he turned to magnetoencephalography, which is capable of finely elucidating dynamic processes. « My research trajectory has often changed direction. Which just goes to show that a scientist’s career path is rarely linear. » The latest shift in his career was when his wife, also a neuroscience researcher, obtained a position at the CNRS. « We were both in the U.S. and expecting a baby, so logically we wanted to return to France together and continue our careers there. Two years later, he obtained his ATIP-Avenir funding and, after a few months, an Inserm Research Officer role at the Lyon Neuroscience Research Center.

« The ATIP-Avenir program is a real springboard for launching your own research theme under excellent conditions, » acknowledges the researcher. The attendant autonomy has enabled him to adapt his research to his unwavering curiosity: « I am very drawn to innovations, new projects... And I need flexibility to thrive as a researcher. With ATIP-Avenir, Inserm offers great scientific freedom. While my subject will no doubt undergo fewer changes in the future, the development of more clinical aspects that I want to conduct is very exciting. » For this, he would like to harness the high potential of MEG – which remains marginal in the clinical setting – in exploring diseases of the central nervous system. For example, the researcher is exploring specific signals recorded in people with epilepsy during seizures. Using MEG combined with machine learning, he hopes to identify the epileptogenic areas of the brain more precisely than with conventional approaches. He also wishes to identify ways to improve cognitive learning after a stroke. « The Holy Grail is to find clinical applications for the discoveries we make in basic science!  »

Romain Quentin leads the Memory, Error and Learning Group (MEL) at the Lyon Neuroscience Research Center (CRNL, Inserm unit 1028/CNRS/Université Claude Bernard Lyon 1).