title | tags | |
---|---|---|
Advanced MEG/EEG toolkit at the Donders |
|
From 20-24 April 2020 we will again host the “Advanced MEG/EEG toolkit”. Although initially scheduled to take place at the Donders Institute in Nijmegen, the COVID-19 situation made that impossible. We are instead hosting it online.
Organizers: Robert Oostenveld and Jan-Mathijs Schoffelen, with the assistance of Sophie Arana and Joey Zhou.
This toolkit course will teach you advanced MEG and EEG data analysis skills. Preprocessing, frequency analysis, source reconstruction and various statistical methods will be covered. Furthermore, there will be a lot of attention to best practices for reproducible analysis and to Open Science. The toolkit will consist of a number of lectures, followed by hands-on sessions in which you will be tutored through the complete analysis of a MEG and EEG data set using the FieldTrip toolbox.
Even though we host the toolkit course online, we can only host a limited number of participants. To achieve good interaction between lecturers and participants - and especially tutors and participants in the hands-on sessions - we have to limit the online attendance. Where possible we will record online events and share those afterwards.
We're pleased to have the following colleagues from the Donders and from abroad as our lecturers and tutors:
- Ashley Lewis - Donders, Nijmegen
- Christoph Huber-Huber - Donders, Nijmegen
- Eleanor Huizeling
- Elie Rassi - Donders, Nijmegen
- Eric Maris - Donders, Nijmegen
- Frederik Weber - Donders, Nijmegen
- Hesham ElShafeh - Donders, Nijmegen
- Jan-Mathijs Schoffelen - Donders, Nijmegen
- Joey Zhou - Donders, Nijmegen
- Johannes Algermissen - Donders, Nijmegen
- Kristijan Armeni - Donders, Nijmegen
- Kristijan Armeni - Donders, Nijmegen
- Lau Moller Andersen - CFIN, Aarhus
- Martin Dresler - Donders, Nijmegen
- Mats van Es - Donders, Nijmegen
- Mikkel Vinding - Karolinska, Stockholm
- Robert Oostenveld - Donders, Nijmegen
- Sophie Arana - MPI, Nijmegen
- Stephen Whitmarch - ICM, Paris
- Vitoria Piai - Donders, Nijmegen
- Xiaochen Zheng - Donders, Nijmegen
The program was shared with the participants as a Google calendar, including links to the Zoom meetings and relevant online documents. This allowed them to refresh whenever updates were made. Here is a screenshot:
{% include image src="/assets/img/workshop/toolkit2020/program.png" width=500 %}
The toolkit has already been organized for many years in real-life at the Donders. To plan all changes that were needed for running it online, we used the following list to keep track of the preparations.
Since all participants are working at home on their own computer, rather that at the Donders in the training room on the computers that we normally provide (and therefore control), we asked the participants to test their MATLAB and FieldTrip installation in advance. This not only served to test their installation, but also to check that all participants knew how to share their notes in a Google doc with the tutors.
Throughout the week we wrote down some things that we noticed in a document so that we could later reflect on the lessons that we learned.
The following are the practical notes that were distributed to participants and tutors in advance (as a Google doc) and that were updated as we were going along. In italic (as here) I have added some updates that apply to the situation afterwards.
We would be very happy if you could spend a couple of minutes of your time to have a look at our Code of Conduct to make sure we all are taking responsibility to look after each other and make sure we are contributing towards an inclusive and supportive community. Please do have a look at and let us know if you have any questions regarding it. All toolkit participants are responsible to follow the rules listed here, as well as making sure that everyone in the toolkit follows it.
To access the shared calendar using Google Calendar in your browser, you go to “other calendars”, then “add from URL” and use this shortened link.
To access the shared calendar from another application (e.g., Apple Calendar), you can use the full subscription link.
If you want read/write access (to help maintain/update the schedule), please give your Google account name to Robert, Joey or Jan-Mathijs.
All tutors received a link to the shared folder on the Google drive, all participants only received a link to the documents relevant to them. The details available to the participants are not shared here, but included the list of participants with some details, the schedule with the assignments of lecturers/tutors, etc.
Although the shared Q&A documents were meant to collect questions, tutors and participants also started using them to answer questions as the lectures were still ongoing; this was a great addition to the Q&A format. The hands-on documents shared between participants and all tutors were also used to motivate participants, to ask/answer questions, and to quickly share something on screen or in the MATLAB command window (as alternative to Zoom screen sharing, which was not so smooth).
We have a shared Google doc for each Q&A session following the lectures and hands-on sessions. Please type your questions in that document, we will discuss them in the subsequent Q&A session.
We also have a Google doc for each individual participant for the hands-on sessions. Please use it to keep your notes during the practicals, to allow tutors to look along with what you have managed to complete and monitor your progress.
The tutors have access to the complete shared folder with all documents, the participants only to the subset relevant to them.
Please note that the invite link is not valid any more.
Inviting people goes via this link. We will disable this link once everyone is on board.
Participants should sign up as “Firstname Lastname” and add a profile picture to make it more personal. Tutors should also sign up as “Firstname Lastname” like everyone else, and set their status to "🤓 tutor" so that they are easily recognized. Please note that you should set it such that the status does not automatically clear itself.
Robert has a paid pro-account and is able to add/edit scheduled meetings. We have prepared 5 online meetings, one for every day, from 8:00-18:00. Each evening (around 18:30) we will start a new Zoom session; we need some time in between to download and convert the video recordings of the day.
Participants can join all these Zoom meetings without the host being present. It is not always required that a host/lecturer/tutor is present, participants can also hang out together. Please join in Zoom using your full “Firstname Lastname” details.
All tutors and participants are made co-host, and at the start of a hands-on session we create enough breakout rooms for everyone. That allows everyone to switch between their assigned breakout room and the main room when they want, but also to switch to other breakout rooms.
Note that these links to the Zoom sessions are not valid any more.
We will stream the pre-recorded lectures live (using CamTwist), but you can also choose to watch them (on YouTube) at a moment that suits you, preferably prior to the Q&A session. There are two lectures that are not pre-recorded (sleep, stats), these are only streamed live.
If you watch the lecture on YouTube, you have the advantage of stop, rewind, etc. Please write questions in the shared Q&A document.
If you watch the YouTube lecture live, you have the advantage of feeling more connected to the event, but note that we will not take live questions while streaming the YouTube video. Please write them in the shared Q&A document.
You can find the links to the online lectures in the Google calendar. We will also post all important links for each on-going session in Slack and at the top of this document.
- Intro
- Lab demo EEG-video, MEG-video
- Sleep will be live and recorded for later online sharing (see below)
- Spectral analysis
- Forward and inverse
- Beamforming
- Statistics will be live and recorded for later online sharing (see below)
- Connectivity
During the Q&A following the lectures we will go through the questions collected in the shared Google doc. This will be an interactive session and you can also directly ask your questions here or contribute to the discussion.
During the Q&A sessions following each hands-on we will do a run-through of the tutorial, explaining what you should have seen (in case it did not work for you) and answer questions that arose during the hands-on. This will be an interactive session and we expect you to contribute.
You can find the links to the shared Q&A documents in the Google calendar. We will also post all important links for each on-going session in Slack and at the top of this document (i.e., the shared "practical notes" Google doc).
- Intro lecture
- Lab demo
- ERF hands-on
- Sleep lecture
- Spectral analysis lecture
- TFR hands-on
- Forward and inverse and Beamforming lecture
- Beamforming hands-on
- Statistics lecture
- Statistics hands-on
- Connectivity lecture
- Connectivity hands-on
This section was edited and extended throughout the week. The links to the videos were also shared via Slack once available on YouTube.
- There are no recordings on this day, only the playground.
We discussed the video recordings with the participants prior to the actual recordings.
By attending the online Zoom sessions, you implicitly give us permission to record these sessions on video.
We will record the online lectures and the Q&A sessions via Zoom’s local recording function (see here), to enable archiving and sharing (e.g., via the FieldTrip website and YouTube). We will share at least with all toolkit participants (especially those that cannot attend live), and if the quality is good, also with the wider public FieldTrip community.
The layout “Shared screen with active speaker” will be used by default during the recording. By default, all attendees to the Zoom session will be assigned as co-hosts, and therefore technically would have the right to start recording on their local machines. However, as a toolkit participant you are not allowed to make any recordings yourself!
If you prefer not to be recognizable in the recordings, please keep your video disabled. You can also refrain from asking direct questions by audio, and instead use the session specific Q&A document to ask your questions in advance.
If you have any concerns with regards to the video recordings, please don’t hesitate to reach out to Robert Oostenveld.
For the hands-on sessions you should make sure you have taken the following preparations (for people that are running the hands-on sessions on the DCCN cluster, or use a Remote Desktop Connection to the workstation at the DCCN a different set of preparations apply, see below):
For people that do not use a computer at the DCCN:
-
Install a working MATLAB version
-
Download the latest release version of the FieldTrip toolbox from here or from here. For those of you familiar with git, you can also clone directly from GitHub with
git clone https://github.com/fieldtrip/fieldtrip.git
Please note that the release version corresponds to the release branch, and the development version to the master branch (which might be a few steps ahead).
-
Add FieldTrip to the MATLAB path in your MATLAB session
restoredefaultpath addpath <path_to_fieldtrip> ft_defaults
-
Download the tutorial data required for the hands-on from the ftp server by going to https://download.fieldtriptoolbox.org/tutorial/ and download Subject01.zip
You should also download some pre-computed intermediate results for each tutorial:
- https://download.fieldtriptoolbox.org/tutorial/eventrelatedaveraging/
- https://download.fieldtriptoolbox.org/tutorial/timefrequencyanalysis/
- https://download.fieldtriptoolbox.org/tutorial/beamformer/
- https://download.fieldtriptoolbox.org/tutorial/cluster_permutation_timelock/
- https://download.fieldtriptoolbox.org/tutorial/connectivity/
Downloading will take some time, especially if everyone tries to do it at the same moment. So please download in advance! If downloading from the FTP server in the web browser gives problems, please try with CyberDuck or FileZilla.
For people that use a computer at the DCCN (either a PC via remote desktop or the DCCN cluster):
-
Consider using a VNC session with MATLAB on the cluster, rather than using MATLAB through a remote desktop connection.
-
Use a MATLAB version > 2016b.
-
FieldTrip is installed on the cluster, on /home/common/matlab/fieldtrip. (on Windows PCs, this is usually mounted on H:\common\matlab\fieldrip. So you don’t need to download the latest release version of the FieldTrip toolbox from here or from here. Add FieldTrip to the MATLAB path in your MATLAB session
restoredefaultpath addpath <path_to_fieldtrip> ft_defaults
-
The data needed for the tutorials are available on
/home/common
as well, so you don’t need to download this in advance. If you insist on downloading the data in advance, you can follow the instructions of point 4 provided in the section above.The raw MEG data set of ‘Subject01’, and associated data are located in
/home/common/matlab/fieldtrip/data
. Additional hands-on session specific data are located in/home/common/matlab/fieldtrip/data/ftp/tutorial
. These data are needed once we get to the respective hands-on session:/home/common/matlab/fieldtrip/data/ftp/tutorial/eventrelatedaveraging
/home/common/matlab/fieldtrip/data/ftp/tutorial/timefrequencyanalysis
/home/common/matlab/fieldtrip/data/ftp/tutorial/beamformer
/home/common/matlab/fieldtrip/data/ftp/tutorial/cluster_permutation_timelock
/home/common/matlab/fieldtrip/data/ftp/tutorial/connectivity
You can find the links to the tutorials in the Google calendar. We will also post all important links for each on-going session in Slack and at the top of this document.
To kick off each hands-on session we will give a short introduction to the tutorial in the main Zoom meeting room. Then you will have time to work on the respective tutorial at your own pace, while staying connected to others via Zoom and Slack. As you work on the tutorial you should record your progress in your dedicated Google doc which will be shared with you in advance. Please copy and paste your terminal output and screenshots of your figures as you run the tutorial code. This way you can record any warnings that might appear in the output, some of which are expected, others potentially indicating that something is going wrong. We will be monitoring your Google doc (i.e. “look over your shoulder”) to keep track of your progress and make sure you are on the right track.
You should keep the main Zoom session open in the background during the entire hands-on. From time to time we will make announcements via the general Zoom session or invite you to join a short breakout session discussing a recurring question.
Besides the main Zoom session there will be multiple breakout rooms, one room dedicated to each participant. If you require one-on-one help from one of the tutors or you would like to have a discussion with other participants you can contact them via Slack and invite them to join you in your breakout room. During the hands-on sessions you will be granted Co-host rights which will allow you to move freely between rooms. You don’t have to stay in your own room, you can also move to another breakout room to join a fellow participant (e.g., a colleague) that is working on similar data. We will practice the breakout rooms in advance.
Next to Zoom you should also monitor the session-specific channel in Slack. This is the place where you can communicate with fellow toolkit participants, the FieldTrip organizers and tutors. You can post questions in the channel or ask for one-on-one help in Zoom. You can identify tutors on Slack through their status which will be set to "🤓 tutor".
When the hands-on session comes to an end, we will close all breakout rooms and gather in the main Zoom room for a Q&A and wrap-up session.
After 3 1/2 days of getting familiar with the FieldTrip functions and possibilities in the analyses and in designing and implementing analysis pipelines that suit specific research questions, everything you have learned will come together in the Playground session. The Playground will largely follow the format of the hands-on sessions, but you will be working on your own data and your very own analysis pipeline. These sessions will provide you with ample time to create your analysis script from scratch with the support of the Toolkit team's expertise. Tutors will be there to help you overcome the occasional noisy trial, misaligned topography or even just find that damn missing parenthesis.
The playground is a great opportunity to do a first exploration of your data or play around with different analysis options. Even if you only have preliminary data, do not hesitate to bring it along! The Playground will give you a head-start on any future analysis and get you FieldTrippin' in no time.
In case you don’t have your own MEG or EEG data yet, please contact Jan-Mathijs or Robert (preferably a bit in advance). We have a lot of different example and tutorial datasets and we can share one with you that resembles the data from a system similar to the one you will work with and/or the experiment that you plan to do in the future.
Joey organized an online pub-quiz on Wednesday evening.