Background
The use of naturalistic stimuli, such as films, during functional MRI acquisitions is opening a new way to explore brain activity. One approach is “dynamic mode decomposition” (DMD), a latent space discovery method built upon auto-regressive models that can identify different dynamical components within the multivariate time series provided by fMRI. First analyses have shown promising results, but the aim of this project is to revisit and deepen these initial findings with the purpose of further dissemination.
Project description
This is a semester project/lab immersion. It is predominantly computational with a strong link with cognitive neuroscience.
Requirements
- The applicants are expected to have knowledge of signal processing.
- Good programming skills is a plus. While python is preferred, at least an elementary knowledge of Matlab (to read, understand, and translate code) is required.
- Desired: Travel to Geneva 1 day per week.
Please contact Dimitri Van De Ville (dimitri.vandeville@epfl.ch) with your CV and we’ll get back to you with more details about this project.