Analysis of brain images from young children
Jose Dolz and a team from CHU Sainte-Justine receive a grant from IVADOWednesday, August 5, 2020
IVADO has awarded a Fundamental Research Project Grant to a team of researchers from ÉTS and CHU Sainte-Justine for their work involving the development of new artificial intelligence (AI) tools for better visualizing the brains of newborns.
There is no longer any doubt that AI and data banks have contributed to the emergence of significant advances in the healthcare sector. In fact, a team of researchers from CHU Sainte-Justine, working on a study in collaboration with Jose Dolz, a Researcher in the Software and Information Technology Engineering Department at ÉTS, has developed an innovative AI-based technique to better define the different sections of the brain in newborns during magnetic resonance imaging (MRI) exams. Improving the segmentation of the brain tissue of newborns allows for a rapid analysis of the brain with increased reliability.
Within the context of its Fundamental Research Funding Program, the Institute for Data Valorization (IVAD0) has awarded a grant to these researchers that will enable them to pursue their work. The study, entitled “Next generation neonatal brain segmentation built on HyperDense-Net, a fully automated real-world tool”, was published in Frontiers in Neuroscience in March 2020.
The neonatal brain is extremely vulnerable to the biological consequences of prematurity or birth asphyxia, which can lead to cognitive, motor, language and behavioural impairment. One must wait several years before it is feasible to test certain key aspects of the functioning of newborn brains, which significantly hinders the progress of neonatal neuroprotection. As is the case with adults, researchers and clinicians require objective tools in order to immediately assess the efficacy of administered therapies.
The results of the aforementioned study reveal that imaging of the neonatal brain using MRI can close the gap and provide the required tools. Working with a team headed by Jose Dolz, the researchers at CHUSJ were able to adapt existing tools to the specificities of the neonatal setting. However, neonatal brain imaging using MRI represents a considerable challenge, because the growth and maturation of the brain during this time of life is intense. According to Jose Dolz, who co-authored the study with Dr. Gregory A. Lodygensky, a Neonatologist at CHU Sainte-Justine and Professor at Université de Montréal: “The IVADO grant will enable us to use the latest developments in AI tools to improve our existing neonatal brain segmentation tools, especially HyperDense-Net, and to develop a tool that will allow for the determination of objective brain maturation among newborns”.
An expert in artificial intelligence, the analysis of medical images and deep learning applied to visual recognition, Professor Dolz develops solutions aimed at enabling machines to analyze large volumes of data for the purpose of comparison and establishing links. He explains: “The issues facing the medical field today revolve around obtaining rapid results that are accurate, interpretable and reliable, which is exactly what we are trying to accomplish”.
About José Dolz
Jose Dolz earned his Master’s degree in Telecommunication Engineering in 2010, and holds a PhD in Applied Mathematics from Lille 2 University of Health and Law. He was awarded a Marie Curie Grant for his doctoral studies, during which he developed methods for automatizing the segmentation of organs at risk of brain cancer. He joined ÉTS in 2016 as a Post-Doctoral Fellow. His research focuses on deep learning and the optimization of medical imaging analysis methods.
To find out more:
Artificial intelligence helping very young brains (ÉTS Website)
Using Deep Convolutional Neural Networks for Neonatal Brain Image Segmentation (Frontiers in Neuroscience)
Communications ServiceAll the news