Health - Personalised Healthcare
Monitoring the Effects of Concussion in Young Kids
Date:
Changed on 05/03/2026
Resulting from a blow at a young age, pediatric concussions may lead to developmental delays. However, to this day, there is no tool for predicting whether or not this will be the case. In particular, using Magnetic Resonance Imaging (MRI), there is currently no way to compare the brain images of a kid who once had a concussion to pictures of what would be a normal brain development over time. A scientist in neuroimaging at the Inria Center at the Rennes University, France, Fanny Dégeilh is devising a method to build a ‘normative’ reference standard of such brain development that would take into account the typical variability from one child to another.
Every year in the world, 2.7 million children suffer from a socalled ‘mild’ injury. What sequelae are to be expected? Sometime none. Sometime development alterations of varying severity. Could neuroimaging help predict such evolution and thus anticipate the care that might be required? At present time, no. For there exists no reference standard of typical brain development at a young age to which new images could be compared. Nothing alike the weight-for-age charts in patients’ health records. And it is precisely such kind of tool that Fanny Dégeilh will build during the next four years. The scientist is a member of Empenn[1] ), a research team at the Inria Center at Rennes University. Her project is called CoYoKi : Concussions in Young Kids. It is funded by the French Research Agency through the Young Researchers program.
Verbatim
We want to know if a concussion will affect a child’s brain development. Is everything fine and the kid evolving along what we could call a normative trajectory? Conversely, is the child diverging from this trajectory at one given time? Does it improve or worsen over time? We want to be able to detect which are the children whom brain development will be affected and to understand what mechanisms could help them.
Auteur
Poste
Inserm researcher within the Empenn project team
For a time, ‘mild’ concussions were considered rather benign. “It was posited that, as the brain has a certain plasticity, small wounds would heal well. Furthermore, young kids having very plastic brains, they would recover all the better. However, surveys have shown that such is not always the case. Yes, most children recover very well indeed. Yet, about 30% will experience difficulties.” But it is hard to predict...
Double Variability
An alternative to sedation
How to acquire MRI data of a child under five without resorting to sedation? How to keep them still for ten long minutes inside an intimidating and noisy scanning machine? Fanny Dégeilh authored an alternative non-sedated behavioral-play familiarization protocol whose details were published in Pediatric Radiology[2]in 2023.
“Prior to the session, we prepare the children to make them as relaxed as possible. We make them listen to the machine’s noise so that they will get used to it. We read them a story that I wrote and that has been illustrated by a graphic artist. The tale narrates the experience of a little squirrel who undergoes a MRI. The little squirrel changes clothes, sets in the machine and so on. This is meant to explain to the child everything that is going to happen. We build their confidence. Then, we give them a little rabbit who will enter the machine with them. And during the session, they will watch a cartoon.” Icing on the cake: “it works! We acquire high quality sharped images, and the kids are happy with the experience!”
Depending on how the child falls or gets hit, brain injuries will differ. In addition, “brain regions do not all develop at the same time. So they are not sensitive in the same way at any given time. A region under development is more vulnerable than a region already developed. And since each child has their own development dynamics, two children of the same age who would fall exactly the same way would probably have different injuries.”
Beside this double variability, research faces another problem: MRI data is sparse at such a young age. Why? Because acquiring non blurry images requires the child to stay still in the tunnel-like machine for several minutes. Quite a challenge as most children happen to be restless at this early age. In practical terms, doctors resolve to sedation. But the method is ruled out in research-only context.
During her post-doctorate at Montreal University, Fanny Dégeilh came up with an innovative protocol that enables to acquire MRI data without child sedation [ READ BOX ]. This protocol is now being used at Rennes University Hospital, a partner of this research. It applies to children between three and five arriving at the emergency department following a blow at the head.
“The doctor offers to take part in our study. If parents agree, their child will come twice for MRI acquisition. First, in the days following their visit to emergency. Then again three months later.” Thus, progressively, will be set up the first-ever cohort of young children whose concussion will be monitored through neuroimaging on a time scale. The immediate goal, here, is to “characterize the injuries for each kid. Can MRI detect them? If yes, which MRI sequences [3] enable it? Are these wounds associated with specific symptoms?” Later, in order to spot potential trajectory deviation in the brain development, it will be possible to compare these pictures to MRI images of children who had no concussion and follow a typical development.
This second set of MRI data needed to build the normative tool will come from the United States. The collection contains over 3,000 images covering 343 children under five at different stage of their brain development.
Before leveraging this dataset, the scientist will have to build what is referred to as an atlas. “Even if brains look alike, they all have slightly different shapes. Yet, for our work, we must put all of them in the same geographical space. We must make sure that the brain region of one participant is well aligned with the same region of another.”
Computing an Average Trajectory
Once the atlas done, the next challenge will consist in modeling the development variability that typically exists from one child to another. “We will take different images of a child’s brain at different moments of its development, at different ages. Then, we will mathematically link these images through what we call a longitudinal trajectory. From there, we will be able to say: this child has evolved in such way, that other one has evolved in such other fashion. And from all these individual trajectories, we will compute an average one. We will thus be able to specify that a child of a given age has, on average, a brain of such particular size. Another kid might have a brain a bit bigger or a bit smaller, but this remains in the ballpark of normality.” These standard deviations from the normative variability.
Using this reference standard, the clinician one day will be able to monitor the evolution of a child after a concussion and compare their development to the reference standard. And beside concussion, “the very same tool could be used in other pathological contexts, by the way.”
The current project is due to last 4 years. “By that time, we will have handled the methodological side and we will be getting close to a tool.” The research findings will be integrated into Empenn’s Anima open access toolkit. “But this will only be a first step. A lot of work will still be needed before contemplating a real-world utilization in a clinical context.”
[1]Fanny Dégeilh is an Inserm scientist, member of the Empenn a pluridisciplinary project team of Inria, Inserm, CNRS, Rennes University, common to UMR Irisa. The group is specialized in biomedical image analysis for disease understanding.
[2] Read : Behavioral-play familiarization for non-sedated magnetic resonance imaging in young children with mild traumatic brain injury, by Fanny Dégeilh, Jessica Lacombe-Barrios, Carola Tuerk, Catherine Lebel, Véronique Daneault, Ramy El-Jalbout, Jocelyn Gravel, Sylvain Deschênes, Josée Dubois, Chantale Lapierre, Isabelle Gagnon, Mathieu Dehaes, Thuy Mai Luu, Miriam Beauchamp, dans Pediatric Radiology, 2023.
[3] An MRI sequence is a particular setting of radio-frequencypulses and gradients, resulting in a particular image appearance.