Differential Diagnosis of Parkinsonian syndromes using artifical intelligence and multimodal MRI



Main objectives

The main objectives of the projects are : i) to improve the accuracy of automated classification techniques for the differenciation of parkinsonian subjectts, ii) to determine the added value of MRI and automated classification in the early stages of parkinsonian syndromes, iii) to extend our classification methods to other forms of parkinsonian syndrome. To make these techniques applicable in the clinic, we will propose new deep learning approaches capable of simultaneously extracting relevant biomarkers and providing diagnostic categorization at the individual level.


Parkinsonian syndrome of neurodegenerative origin is most often caused by Parkinson's disease. More rarely, it is said to be atypical, belonging to the spectrum of tauopathies (progressive supranuclear palsy and corticobasal degeneration) or synucleinopathies (multisystematic atrophy, and dementia with Lewy bodies). Clinical differentiation of Parkinsonian syndromes is often difficult, especially in the early stages of the disease. Magnetic resonance imaging (MRI) provides specific biomarkers that can be used to train and validate artificial intelligence algorithms to accurately classify parkinsonian subjects.

Methodology/Technology used

Artificial intelligence methods (machine learning) for automatic patient classification and biomarkers extraction from multimodal brain MRI images.



AP-HP, Inria, CNRS, Inserm, Sorbonne Université, Institut du Cerveau

Other projects



US Caractérisation et prédiction de la survenue de formes graves ou létales du COVID-19 à partir des données issues de l’EDS de l’AP-HP

Names of partners involved
AP-HP, Inria & Centrale Supélec

read more