Inria@SiliconValley Post-Doc laureate Nina Miolane is one of the thirty promising young researchers recognised by UNESCO and the L’Oréal Foundation in 2016. She has been awarded a “Women in Science” Fellowship for her virtual patient research in the category “Exploring the brain, a new world to conquer”.
The thirty laureates were chosen this year from more than a thousand applicants by an independent jury of academicians based on academic excellence, the originality of their scientific work, and their desire to transmit their passion to the younger generations.
Nina Miolane conducts her research within the Inria-Stanford associate team “Geomstats” team (specializing in geometric statistics applied to digital anatomy), under the supervision of Inria Research Director Xavier Pennec (ASCLEPIOS Inria Team at Sophia-Antipolis) and Stanford Professor Susan Holmes, on the geometric hierarchical model of cerebral anatomy and MRI-based modelling of the human brain. She will defend her PhD thesis on “Geometric Statistics for Computational Anatomy” in December, before heading to Stanford with a post-doctoral fellowship from Inria@Siliconvalley.
Just twenty-five years old, Nina Miolane is savouring this social and financial recognition of her work. To her, this honour represents the opportunity to explain the focus and impact of her research to the general public, and, by her example and achievements, to encourage girls and young women to pursue exciting careers in science. The Fellowship will allow her to continue promoting science among young people and to further her studies with a short formation in neurology in order to better understand the medical challenges in this field, which she hopes to address in her future research
Maths and Physics
Her academic background is impeccable: she completed secondary school at the French-German Lycée de Buc near Paris, and then two years of preparatory classes at the prestigious “Ginette” campus in Versailles, before entering Polytechnique, a top French engineering school. But when asked how her passion for maths began, Nina Miolane says it was a gradual process, something that evolved as she learned more about the different branches and their applications. After completing a Masters in theoretical physics, applying Riemmanian geometry to the description of space-time, she turned her attention to digital imaging. She spoke with us about choosing this new focus. “I realised that although I loved maths, I wanted to apply them to something that could have a direct impact on society. My choice of dissertation subject reflects that ambition. In working with my doctoral advisor Xavier Pennec, I saw that Riemmanian geometry was used in medical image analysis. So I stopped theoretical physics and transferred my maths knowledge to the field of medical imaging. The same mathematics that I used to describe black holes I now use to model brain shapes .”
“A Virtual man for Medecine of the Future”
Using maths and digital technologies for advances in medical practice: Nina Miolane’s project develops digital anatomy, a key field in the future of medicine. This multidisciplinary science combines mathematics and computer technology in order to analyse medical images (scans or MRIs [Magnetic Resonance Imaging]) and thus, by calculating statistics on these digital images of the inside of the human body, create a digital model of a healthy patient, or in the case of Nina Miolane’s research, a digital model of the human brain.
“A doctor’s expertise is related to his experience: the more patients he has seen in his career, the better equipped he is to accurately diagnose a new patient. With medical imaging and databases, we are essentially giving a doctor direct access to a multitude of patients. And instead of analysing these images with his eyes, he can use the computing power of a supercomputer and 3D imaging ,” explains Nina Miolane.
In practice, using a database of MRIs of healthy brains, she has developed a mathematical theory (geometric statistics) that allows medical images to be processed in order to digitally represent the human anatomy, as well as its healthy and diseased variations. The cerebral anatomy model will be stratified to account for the diversity of brain morphologies found in healthy individuals. It will present the large-scale common anatomy, and the possible variations occurring on a smaller scale.
Although this project concerns images of the brain, the work is generic enough to be applicable to other organs down the road. In the short term, it will help improve our knowledge of human anatomy, and in the medium term, comparison with medical images of a virtual patient with a healthy anatomy will aid diagnosis thanks to the development of new digital tools.
Towards a preventive approach to medecine
The goal of this type of research, combining powerful computing with digital databases, is to develop a preventive approach to medicine, for example, making it possible to diagnose the probability of a patient developing a neurodegenerative disease such as Alzheimer’s or Parkinson’s before the symptoms appear.
“Within the near future, this type of model could pave the way for technologies to aid presymptomatic diagnosis ,” sums up Nina Miolane.