| title | tags | |
|---|---|---|
PracticalMEEG workshop in Aix-en-Provence |
|
PracticalMEEG aims at providing an intensive three–day training to MEG and EEG analysts. It will provide its attendees with the ability to create a full analysis pipeline with exemplar (or their own) data in one or several of four leading software dedicated to MEG and EEG analysis (Brainstorm, FieldTrip, EEGLAB, and MNE-python). On demand, we hope to provide attendees with the opportunity to bridge across toolboxes and weave their own complex multi-software pipeline. More details can be found here.
The entire meeting will take place at the building named Le Cube on the Aix-Marseille University – Schuman Campus. The address is 29 Av. Robert Schuman, 13100 Aix-en-Provence.
Dec 14-16, 2022
Anne-Sophie Dubarry and Maximilien Chaumon are the local organizers.
The FieldTrip tutorials will be presented and tutored by Robert Oostenveld (from the Donders Institute for Brain, Cognition and Behaviour, Nijmegen, NL), Marie-Constance Corsi (Inria, Paris, FR) and Laure Spieser (Laboratoire de Neurosciences Cognitives, Aix-en-Provence, FR).
There will also be other experts from MNE-Python, BrainStorm and EEGLAB with their respective tutorials, and jointly we will organize the plenary lectures.
- Preprocessing raw data and computing ERPs/ERFs
- Dealing with artifacts
- Time-frequency analysis using Hanning window, multitapers and wavelets
- Creation of headmodels and sourcemodels for source reconstruction
- Reconstructing source activity using beamformers
- Group-level statistics with parametric and non-parametric methods
Please take a quick look at the FieldTrip reference paper if you have not done so already. If you have an hour or so, you can watch this introductory lecture. Note that more recorded lectures are available as video.
Please download the most recent version of the FieldTrip toolbox from here or from GitHub. As we did some last miniute bugfixes to make all tutorials run smoothly, you should have version 20221207 or later.
Please read this FAQ on how to set your path. After adding the FieldTrip main directory to your path, you should type ft_defaults which will add the required subdirectories (which depend on your MATLAB version). We recommend to put ft_defaults in your MATLAB startup.m file.
We will use a small datainfo_subject.m function that specifies the input and output data files for each participant. You can download that from the code folder on our download server. Please organize the workshop data and code as follows:
/Volumes/SamsungT7/practicalmeeg2022/
├── README
├── code
│ ├── README.md
│ ├── atlas_subparc374_8k.mat
│ ├── datainfo_subject.m
│ └── ...
└── ds000117-pruned
│ ├── CHANGES
│ ├── README
│ ├── dataset_description.json
│ ├── derivatives
│ ├── participants.tsv
│ ├── stimuli
│ ├── sub-01
│ └── ...
└── derivatives
├── anatomy
├── groupanalysis
├── raw2erp
├── sensoranalysis
└── sourceanalysisThe /Volumes/SamsungT7/practicalmeeg2022 directory is where I have the data on my laptop (actually on an external SSD), for you that would be somewhere else. There should be a code directory with (at least) the datainfo_subject.m function, there should be the ds000117-pruned directory with the raw data in subdirectories, and there should be a derivatives directory with one subdirectory per analysis step.
The derivatives directory contains (or will contain) the results of the analysis, you can download precomputed results as .mat files from our download server. For some steps that is not needed and you will computre them yourself, but for some lengthy analysis you will continue to work with the precomputed results of the previous step.
You can get the data that we will work with from Zenodo or on OpenNeuro. The Zenodo version is pruned and only contains the raw data files needed for this workshop. The OpenNeuro version contains all data (and hence is very large).
You can alternatively download the raw data from the ds000117-pruned folder on our download server. Note that this is a WebDav server and you should use a WebDav client like CyberDuck or FileZilla. On Windows you can also map it as a network drive. By using a WebDav client you can maintain the folder structure when downloading, which is important!
Besides the raw data, we will also use some already processed data (as some computations take too long). Specifically, we have prepared and shared the headmodel and sourcemodel for the selected subject. Furthermore, we have processed all subjects up to and including to the source level, as that is what we will use for the group analysis. This processed data is available from the derivatives folder on our download server. Please note that this is not the same as the derivatives that is contained within the BIDS ds000117-pruned data: that only contains the MaxFiltered data and the FreeSurfer output, not the FieldTrip-processed data.