The CRM (www.crm.cat) is an independent research centre in the CERCA network since 2008, with the purpose of promoting both advanced training and research in all areas of mathematics, in collaboration with universities and other research institutions in Catalonia. CRM’s mission includes transferring knowledge based on the centre's research to society and aiming at having a societal impact.
Research at CRM covers a wide spectrum of fields, from the most fundamental areas (algebra/number theory/geometry/topology, analysis and PDEs, combinatorics and dynamical systems) to applied mathematics (in the fields of biomedicine, climate change mitigation and adaptation and natural hazards).
The Computational Neuroscience — Vidaurre group at the CRM is looking for a PhD student to join the project “A new integrative statistical framework to connect symptoms to mechanisms in brain disease” project code "Actuación ATR2024-155014 financiada por MICIU/AEI/10.13039/501100011033"
This contract is envisaged to start on September 1st, 2025, and will last until August 30th 2029.
The prediction of progression (prognosis) in Alzheimer’s disease is of immense potential importance for the personalization of treatment (especially with emerging new disease-modifying therapies) and for planning for future care needs. However, prognostic models are currently crude and unreliable. This project will capitalize on a very large longitudinal dataset from a single dementia clinic from Barcelona to build and test detailed models to predict individual progression of disease over time. ACE Alzheimer’s Centre in Barcelona is the largest dementia clinic in Europe, which sees >8000 patients per year. The ACE database has >20,000 patients from the past 10 years who have been evaluated every 6-12 months through the course of their disease. Available data include clinical features, neuropsychological test performance, MRI and PET imaging, CSF biomarkers and genetics. However, while the neuropsychological data is rich and very complete, the neuroimaging data of ACE is present only in a reduced subset of the subjects. On the other hand, large-scale public datasets (e.g., ADNI) provide extensive imaging for many subjects, but lack the depth of clinical data that ACE has. To leverage this potential, we will foundationally models on neuroimaging data using public datasets to learn meaningful feature representations, which will then be applied on ACE’s neuroimaging data and combined with ACE’s neuropsychological data to predict disease trajectories.