Post-Doctoral Research Visit F/M POSTDOC2024-MATHNEURO Creating Ageonostics: A New Clinical Tool to Diagnose Ageotypes across Multiscale Human Lifespan Data - Montpellier, France - INRIA

Description

Contexte et atouts du poste

The MathNeuro project-team in the Inria branch at the University of Montpellier focuses on understanding brain activity at multiple scales (i.e. from single-cell, microcircuits to large scale brain processes). MathNeuro has a core at the intersection of Mathematics and Neuroscience, but it also employs a multidisciplinary approach combining theory, computational modelling, as well as, data science. In line with this, our collaborations involve experts from diverse fields, including experimentalists, clinicians, and data scientists. The research lines of MathNeuro include questions related to plasticity (both synaptic and non-synaptic) and associated learning rules, memory, excitability, with applications in both normal and pathological brain states, as well as, ageing brain. Recent research portfolios include novel mathematical classification of neuronal bursting, explanation of asynchronous neurotransmitter release (via competition of timescales) and application areas such as, understanding migraine, epileptic seizures, Alzheimer's Disease (AD) and ageing.

MathNeuro resides within a rich scientific environment, which includes three permament members, one postdoc, two phd students, and benefits from a large network of international collaborators. The MathNeuro team is also embedded within the ecosystem of the AI Institute in Nice, co-organized by Inria. Detailed information of MathNeuro can be found at our webpage ( and

Mission confiée

The project underpinning this postdoc position is motivated by unique longitudinal multiscale and multimodal datasets on aging (from genes/multi-omics, inflammatory markers, brain data to environmental factors) to which we have obtained access. Namely, the Baltimore Longitudinal Study on Aging (BLSA, collected by NIH, USA, see Ageonostics) for diagnosing and explaining ageotypes (i.e., pathways to ageing) as observed within the BLSA and SLAS datasets. Methodologies will involve state-of-the-art combination between data-science tools such as AI, Topological and Geometrical Deep Learning, Structural Recurrence Analysis and multiscale dynamical systems. Indeed, together with our partners on this project, we have all the necessary expertise in order to pursue this endeavor [1-5].

The central hypothesis for this project is based on recent developments that posit the link between long genes and aging as well as age-related diseases. It is found that long-genes are more vulnerable to environmental factors (e.g., toxins, epigenetic factors) which then induce large-scale phenotypes that lead to different ageing pathways (i.e., ageotypes) and neurodegenerative diseases (e.g., AD). We aim to determine key biomarkers of ageing and neurodegenerative diseases under this hypothesis, which will lead to novel ways to diagnose and possibly motivate the design of new drugs.

As mentioned above, this project will benefit from a strong network of collaborators (Prof. Luigi Ferrucci, NIH, USA; Prof. Alan A. Cohen, Columbia University, USA; Prof. Tamas Fülöp, University Hospital Sherbrooke, Canada; Prof. Serafim Rodrigues, BCAM, Spain; Dr. Fernando Santos, University of Amsterdam, Netherlands), a research ecosystem at the forefront of data science, and unique datasets.

Keywords: aging studies, multiscale data science, deep learning, topological data analysis.

References:

[1] Carmantini, G. S., Beim Graben, P., Desroches, M., & Rodrigues, S. . A modular architecture for transparent computation in recurrent neural networks. Neural Networks 85 : DOI: [3] Fülöp, T., Desroches, M., Cohen, A. A., Santos, F. A. N., & Rodrigues, S. . Why we should use topological data analysis in ageing: towards defining the "topological shape of ageing". Mechanisms of Ageing and Development 192 : DOI: [5] Santos, F. A.N.,(...), Desroches, M., Rodrigues, S., Schooheim, M., Douw, L. & Quax, R. . Emergence of high-order functional hubs in the human brain. bioRxiv, DOI:

Principales activités

The main objective of this postdoc is to perform higher-order multi-layered network analysis of key subparts of the available datasets. We will study complex interactions, identifying higher-order interactions invariances across hierarchical interactome, higher-order hubs and centrality measures based on a number of methods, focusing on deep learning and topological data analysis.

Compétences

Candidates should be familiar with Machine learning, Multi-layered network analysis and Topological data analysis.

Avantages

Rémunération

Duration: 18 months
Location: Sophia Antipolis, France
Gross Salary per month: 2788€ brut per month