SYNAPSE: How can we better understand the brain using imperfect data?

Image 1 : graphe représentant les relations entre différents éléments d’un même réseau

SYNAPSE is a research project dedicated to graph analysis. A graph is a mathematical tool used to represent the relationships between different elements within a single network (see Figure 1). Graph-based representations are used in many fields, such as biology, social sciences, and medicine. In the medical field, they can be used, in particular, to represent the connections between different regions of the brain. Today, the data used to create these graphs are often noisy and uncertain. This can be due to many factors. In the medical field, graphs are created based on medical imaging such as MRIs or CT scans. The equipment used, its quality, its positioning on the body, or even the conditions in which the exam is performed are factors that can lead to irregularities in the collected data.

The SYNAPSE project specifically aims to account for these irregularities in the constructed graphs and in all resulting analyses. To achieve this, the project’s research team will developp new mathematical tools and artificial intelligence techniques to refine the results by taking these uncertainties into account. The graphs will therefore be more reliable and robust, enabling a more refined analysis of medical data.

JCJC Call for project

Mira Rizkallah was laureate of the JCJC (Jeunes Chercheuses et Jeunes Chercheurs - Young Researchers) call for project, organized by the ANR (French National Research Agency). This initiative aims to support research projects led by young scientists and to promote the development of original research.

Portrait of Mira Rizkallah, project lead for SYNAPSE

Mira Rizkallah is an associate professor in Centrale Nantes, within the Laboratory of Digital Sciences of Nantes (LS2N). Her research focuses on signal and image processing, graph representations, and machine learning on graphs, with applications in the analysis of biomedical and clinical data.

Her current work focuses on:

• Processing EEG (brain electrical activity), ECG signals (heart electrical activity) and analyze of their connectivity
• Machine learning on graphs for disease diagnosis and prognosis
• Medical applications, particularly for breast cancer screening and the study of diffuse large Bcell lymphoma (a type of blood cancer)

Her goal: to develop new graph-based methods (at both the individual and population levels) to better understand complex biomedical data and contribute to medical diagnosis and prognosis.

Mira earned her PhD in signal and image processing from the University of Rennes 1 and Inria Rennes in 2019. She received her engineering degree and a master’s degree in telecommunications from the University of the Holy Spirit in Kaslik (USEK) in Lebanon.

Watch the video profile of Mira Rizkallah