I am currently part of the MULTI-FACT study (Multi-centric study of Fetal Abnormal Cortical Trajectory on Fetal MRI), focusing on analyzing fetal brain development using advanced MRI techniques. During this project, I have: Reconstructed and analyzed over 600 fetal T2 MRI scans. Implemented a comprehensive pipeline for fetal brain processing, from raw scan to surface mesh generation. Developed a user-friendly GUI for brain reconstruction, designed for clinical application. Conducted cortical surface analysis to compare healthy and growth-restricted fetuses.Tools: Python, Qt5, Pytorch, HPC, Docker, Singularity, FSL, ANTs. Reconstructed and analyzed over 600 fetal T2 MRI scans. Implemented a comprehensive pipeline for fetal brain processing, from raw scan to surface mesh generation. Developed a user-friendly GUI for brain reconstruction, designed for clinical application. Conducted cortical surface analysis to compare healthy and growth-restricted fetuses.Tools: Python, Qt5, Pytorch, HPC, Docker, Singularity, FSL, ANTs.

As a recipient of the MAGNIMS-ECTRIMS Research Fellowship, I worked on the Virtual Brain on MS project, focusing on generative whole-brain modeling in patients with Multiple Sclerosis (MS). Key accomplishments in this project include:- Developed two data-driven computational models of brain activity (stable and oscillatory).- Processed and analyzed structural, functional, and diffusion MRI data from 697 subjects.- Identified associations between model parameters and clinical indicators of MS.- Evaluated the conservation hypothesis in structural brain connectivity of MS patients.Tools: The Virtual Brain, Python, Mrtrix3, MATLAB, CONN, C++, Freesurfer, HPC.

In my PhD thesis, I focused on developing statistical and machine learning methods for the early detection and progression assessment of Alzheimer's disease (AD), a multifactorial neurodegenerative condition. My work covered the integration of longitudinal neuroimaging data with other biomarkers to better understand disease dynamics. Key contributions include:- A comprehensive review of machine learning applications in AD using longitudinal neuroimaging, highlighting how these methods improve disease modeling and computer-aided diagnosis.- A study on AD subtyping using novel plasma-based blood markers, identifying distinct patient subgroups and analyzing their progression using neuroimaging data.- The development of a new method based on recurrent multimodal variational autoencoders to model disease progression across different subjects and time-points.- An investigation into the impact of APOE ε4 gene dosage on hippocampal shape using a multivariate surface analysis in an ε4-enriched cohort.

Working on CAMOMILE project (https://camomile.limsi.fr) for automatic face detection and annotation on video sequences. Utilitzant Deep learning i Convolutional Neural networks per Face recognition and verification.

Suport a Relacions Internacionals i congressos de la UPC.