Current projects
Your project!
Is there a topic you are particularly interested in? A tool you would like to work on? Submit your own project!
Projects can cover any type of software development, but also can be a proposal for a focus discussion or, a self-training using public resources, and even more! All project proposals will be considered, please contact us if you are unsure whether your proposal can be submitted as a project. To submit a project, please email us a project title and a short description with an image.
An fMRI meta-analysis
Little is known about the brain mechanisms of recovery after stroke. Here we will focus on two rehabilitation methods of the upper limb: constraint induced movement therapy and motor imagery. After reviewing the literature, we selected 32 articles. We want to perform a meta-analysis on fMRI data to understand the cerebral mechanisms of these rehabilitation programs, using tools such as NeuroVault, Brainspell and/or NIDM-Results.
Contact: Simon Butet
MNE BIDS
MNE BIDS is a python package to automatically convert existing files into BIDS compatible datasets.
Contact: Mainak Jas
Discuss fMRI processing methods and tools
What you always wanted to try and never find the time too, starting with for example denoising, modelling for task based fMRI, whole brain or ROI analysis, correction for multiple comparisons, DCM.
Contact: Elise Bannier
Self-training: fMRI in Python
This project aims at processing a task-based fMRI dataset with Python tools only (Nipype, nistats, nilearn, nibabel etc...). The pipeline shall include the following steps: reading a BIDS format dataset, performing the pre-processing stages (slice timing, realignment, coregistration, normalisation), specifying design matrices (GLM 1st and 2nd level) and generating a results report.
Contact: Quentin Duché
Develop ASL processing pipeline using Anima to compute CBF maps
The goal is to build a complete pipeline for quantifying cerebral blood flow (and possibly other perfusion maps) from ASL and M0 images using Anima, an open source software for medical image processing. The project consists in combining different stages of ASL preprocessing images using functions already available in Anima
Contacts: Julie Coloigner, Antoine Legouhy
Nipype looks like a great framework to define small processing bricks and automatically assemble them to form pipelines. Other well known softwares like Nifty* or ANTS already have their interfaces. The idea of this project is to study how workflows and interfaces are made in Nipype through the example of creating some from the Anima open-source software.
Contact: Olivier Commowick
Acquiring and processing neurofeedback EEG data with MENSIA Modulo
We have a new 8 channels EEG cap from MENSIA Modulo, and we would like to see how it works in a first time. Then we want to acquire several neurofeedback sessions, convert the data to a matlab formating, before processing the EEG signals and modelling the neurofeedback scores to improve the NF filtering.
Contacts: Claire Cury and Giulia Lioi
BIDS importer/exporter for Shanoir
We will develop a new feature for the brain imaging database Shanoir to allow for import of datasets organised using the Brain Imaging Data Structure (BIDS). We will use public datasets from OpenfMRI as test data.
Contact: Michael Kain
Implementing a simple Brain-Computer Interface with the Emotiv portable EEG
The goal of the project would be to investigate the capability of the Emotiv EPOC 16-channel portable EEG to design a BCI applications. We can start with a simple binary classification case and let's see where we finish !
Contact: Pierre Maurel
An overview of medInria plugins creation
medInria 3.0 was recently released and offers to developers who want to share their work a way to create plugins (either open source or not) so that they appear in medInria automatically. This micro-tutorial will concentrate on showing a few examples of plugins and what are the steps necessary to create them in medInria 3.0.
Contact: Olivier Commowick
A Deep Learning playground
A practical on how to use Google’s Tensor Flow framework (i.e. Tensor Flow in Python) to create artificial neural networks for deep learning. We will try to achieve an easy understanding of the complexities of the Tensor Flow framework, build a network from scratch with Python, and - why not - use a convolutional neural network for image processing.
Contact: Francesca Galassi
Automatic quality control on large dataset
We would like to address the relation between data quality and error on quantitative biomarkers by looking at large databases. We propose to discuss how to define quality indexes (QI), and how to infer the relation between QI and the bias or the variability of different biomarker. We are open to collaboration to work on the large database acquired at CENIR (~ 17000 T1w scans with a consequent proportion of rescan due to bad quality of the first acquisition).
Contact: Romain Valabregue
Rule-based neuroimaging workflows with SnakeMake
Trying the SnakeMake workflow engine, initially developed for bioinformatics application, in the context of neuroimaging pipelines.
Contact: Alban Gaignard
From MS excel sheets to semantic bioimaging markers : representing and querying cerebral vascular measures with the QIBO ontology
In the context of intracranial aneurysms, try the QIBO ontology to represent and share quantitative imaging biomarkers (vascular geometrical features), design some sample queries to filter and select subsets of imaging markers.
Contact: Alban Gaignard
