Computational modeling for the diagnosis of brain disorders
Computational modeling for the diagnosis of brain disorders
Computational models (e.g., joint models of longitudinal and survival data) are useful when repeated measures and survival times are available at the same time and possibly combined. Different types of data are collected for each subject at multiple time intervals before an event occurs (e.g., conversion from mild cognitive impairment (MCI) to Alzheimer's disease (AD)). These data, obtained from the ADNI (Alzheimer's Disease Neuroimaging Initiative) dataset, including biomarkers linked to amyloidosis and neurodegeneration such as CSF ABeta 1-42, CSF tau, FDG-PET, hippocampal volumes (HV) measured via MRI, as well as scores from neuropsychological tests such as the ADAS-Cog and MMSE. The objective of this thesis is to study the association between longitudinal biomarkers (e.g., HV) and the event of interest (e.g., conversion to AD), and to use this association to predict the conversion time for new topics. Several articles have demonstrated this association, but have not explicitly stated the likelihood of risk of progression. Separate analysis of longitudinal biomarkers and conversion time may lead to ineffective or biased results. Joint models process longitudinal data and event times simultaneously and provide valid and efficient inferences (L. W et al., 2011).
Computational models, MRI, biomarkers, machine learning, Alzheimer, dementias, neurodegeneration, Mild Cognitive Impairment, ADAS-Cog, MMSE.
Candidates with a Master's/engineering degree (or equivalent) in mathematics, artificial intelligence, statistics, biostatistics, computational biology or computer science (big data analysis) with mastery of R and/or Python software are eligible. and having experience in statistical modeling in the health field (considered an asset). Additionally, the candidate must be good in English.
Please send the application before September 10, 2024 to:
Cedoc.admission@ueuromed.org & a.mouiha@ueuromed.org