About us
The lab includes researchers with backgrounds in physics, computer science and engineering. Out research focus is therefore more technical and we tackle questions of this kind:
- What is the optimal MRI sequence to capture some brain features, such as anatomy, tissue property, physiology, activity, etc.?
- How can we represent that “brain signal” with a mathematical model?
- How to perform valid statistical tests to infer the significance of an effect of interest between or within groups of subjects?
Magnetic Resonance Imaging (MRI) is quite central to our activity because of our 2 onsite research dedicated MRI scanners, a 3T Prisma & 7T Terra (Siemens). MRI acquisition research involves the development and optimization of advanced acquisition sequences. We work on techniques not commonly available in clinical routine, such as quantitative MRI, multi-shell diffusion-weighted MRI, Chemical Exchange Saturation Transfer (CEST), or Quantitative Susceptibility Mapping (QSM).
All these particular types of images require specific tools to process them. This therefore constitutes a second axis of research. Transforming the raw data into interpretable (quantitative) maps, then proceeding with their statistical analysis requires dedicated tools, which must be optimized and validated.
With Positron Emission Tomography (PET), we seek to reduce the invasiveness of their acquisition and facilitate the data processing steps for the neuroscientists. For example, to obtain quantitative maps, we aim to use “image-derived input function” (IDIF), instead of the standard “arterial plasma time-activity” curve (AIF), which is rather invasive and uncomfortable for the subjects.
The lab has a long standing experience with signal “time series” analysis. With the Sleep Lab, we work on the processing of whole-night polysomnographic EEG data processing: from multichannel data visualization to manual scoring and feature extraction.
For actimetric data, i.e. the recoding of a subject’s physical activity over the course of days or weeks, we have developed a software toolbox. This tool allows the systematic analysis and modelling of large scale, i.e. month long and 100+ subjects, actigraphy data.
Recently we worked on accurate electro-magnetic head modelling for application such as “trans-cranial direct current stimulation” (tDCS). The aim is to better understand the mechanism of this ‘brain stimulation’ technique and potentially rendering it more efficient with subject specific optimized experimental setup.
This vast wealth of data modalities requires clear and unambiguous data description and organization. We are thus pushing IT development for data curation and “data-basing”, which is particularly complex in a research setting, where each experimental protocol is different from another. Moreover we must account for the GDPR regulation and ensure the safety of human data, while striving for an “open data” approach.
