Research & Scientific Interests

A few words about my research.


As part of the Aramis Lab, located at the Paris Brain Institute at the Pitié Salpétrière, in Paris, my research explores new ways to understand the brain evolution, in particular the dynamics of neurodegenerative diseases, thanks to Machine Learning tools. The goal of this research is to help designing clinical trials suited for neurodegenerative diseases.


The goals of my research are three-fold:
  • Get a better understanding of the spatio-temporal patterns of brain alterations, especially by developping numerical models of the disease progression, based on longitudinal data,
  • Characterize individual patterns of progression by analyzing the impact of cofactors (gender, genes, socio-demographical information), and, by predicting individual diagnosis and disease stage, up to 5 years in advance,
  • Detecting the moment at which the patients are the most likely to respond to drugs in clinical trials. This implies to detect the biomarkers that are most likely to detect the drug effect.


To this end, I have proposed multiple methological enhancements to existing methods (see Publications). Most of them rely on a variety of tools and software I have developped :
  • Leasp: Estimation of the group-average and individual spatio-temporal trajectories. Well suited for the estimation of spatially distributed data. C++ 14. Website.
  • Leaspy: LEArning Spatiotemporal Patterns in Python. Enables to estimate the diseae progression for various biomarkers at the average and individual level. Python 3.7.
  • Digital Brain: Web application to vizualise the group-average and invidiaul evolution of the cortical thickness, the PET-FDG, the hippocampus mesh and the cognitive scores, during the course of Alzheimer's Disease. Digital-Brain

Publications

Click on the image (or title for phone-users) to read the abstract

Simulation of Virtual Cohorts increases Predictive Accuracy of Cognitive Decline in MCI Subjects
Simulation of Virtual Cohorts increases Predictive Accuracy of Cognitive Decline in MCI Subjects
In Under revision. 2019. PDF
Koval I., Allassonnière S., Durrleman S.
Riemannian Geometry Learning for Disease Progression Modelling
Riemannian Geometry Learning for Disease Progression Modelling
In IPMI. 2019. PDF
Louis M., Couronné R., Koval I., Charlier B., Durrleman S.
Simulating Alzheimer’s Disease Progression with Personalised Digital Brain Models
Simulating Alzheimer’s Disease Progression with Personalised Digital Brain Models
In Under revision. 2018. PDF
Koval I., Bône A., Louis M., Bottani S., Marcoux A., Samper-Gonzalez J., Burgos N., Charlier B., Bertrand A., Epelbaum S., Colliot O., Allassonnière S., Durrleman S.
Design of a Decision Support System for Predicting the Progression of Alzheimer's Disease
Design of a Decision Support System for Predicting the Progression of Alzheimer's Disease
In Alzheimer's & Dementia. 2018. PDF
Ansart M., Koval I., Bertrand A., Dormont S., Durrleman S.
Spatiotemporal Propagation of the Cortical Atrophy: Population and Individual Patterns
Spatiotemporal Propagation of the Cortical Atrophy: Population and Individual Patterns
In Frontiers in Neurology. 2018. PDF
Koval I., Schiratti J.-B., Routier A., Bacci M., Colliot O., Allassonniere S., Durrleman S.
Statistical Learning of Spatiotemporal Patterns from Longitudinal Manifold-Valued Networks
Statistical Learning of Spatiotemporal Patterns from Longitudinal Manifold-Valued Networks
In MICCAI. 2017. PDF
Koval I., Schiratti J.-B., Routier A., Bacci M., Colliot O., Allassonnière S., Durrleman S.

Teaching

Machine Learning in Healthcare @ Data Science Summer School
Summer 2017, 2018 & 2020
Lviv, Ukraine
Courses of Machine Learning, applied to Healthcare, at an international Summer School. They include 5 lectures of 3 hours each, and, a 4 days project, with students (Bachelor, Master and PhD) and workers.
Programming in C
2017 / 2018
University Pierre et Marie Curie
Introductory C classes for licence (Bachelor) students, including practical sessions on computer, blackboard sessions, homeworks and exams
Advanced Machine Learning
2015
Master Data Science, Telecom Paris
Supervision of practical sessions of the Advanced Machine Learniing class of the Data Science master degree
Teachers in charge of the class : Florence d'Alché-Buc & Erwan Le Pennec

Talks, Presentations & Conferences