The python for brain imaging (pybrain) workshop series is for practicing neuroimagers who want to learn more about how to work with data in Python. We assume no previous Python background, and provide a comprehensive hands-on introduction to state of the art libraries for experiment building, statistical modelling, machine learning, and multimodal neuroimaging analysis.
The workshops are hosted by the MRC Cognition and Brain Sciences Unit at the University of Cambridge, and will run online with Zoom.
Workshops will run Mondays and Tuesdays, and generally take place 9:30am-12:00pm and 13:00pm-15:30pm GMT. There is separate registration for each workshop, but please do sign up for Edwin’s introductory session if you’ve never worked with Python before.
2-3 November 2020: Introductory python, basic data analysis, experiment coding
Edwin Dalmaijer, MRC CBU, University of Cambridge, UK
This workshop runs 9:30am - 15:30pm both days. We have breaks scheduled 10:30am-11:00am, 12:00-13:00pm, and 14:00-14:30pm, during which time you may be expected to work independently on exercises. Please consult the instructions above before the workshop.
Registered attendees will receive an email with the Zoom meeting details. Please check your bulk/spam email folders if you are registered and haven’t heard from us by now, as Eventbrite emails often get marked as spam
- Introduction to Jupyter Notebooks
- Variables and variable types
- Writing your own functions
- What are classes, and how do I write my own?
- Basic data handling, from loading to plotting
- Implementing statistical tests
- Model fitting through minimisation
- A quick look at experiments*
- Using Python at home and/or in the lab
* While we will be going over how to code experiments, we won’t be able to run them in our online environment. You’ll thus have to apply what you learned at a later time.
For this workshop you will need to install Docker Desktop on your system. Please follow these instructions to setup your system. Please be aware that the download of the workshop content might take some time, as it includes the download of a docker image of ~10GB. If you run into problems, please let us know by creating a new issue here.
You are ready for the workshop and good to go, if you can open the Jupyter Notebook
program.ipynb and see something like this.
This workshop runs from 09:30 to 15:30 both days. Lunch breaks are scheduled from 12:00 to 13:00. Individual short breaks can be taken during the multiple independent exercise sessions throughout the workshop. Please consult the instructions above before the workshop. Be aware, that the setting up of your system might take some time, as it includes the download of a docker image of about 10GB.
Registered attendees will receive an email with the Zoom meeting details by November 4. Please check your bulk/spam email folders if you are registered and don’t hear from us by this date, as Eventbrite emails often get marked as spam
09:30-10:00- Workshop overview
10:00-10:45- Quick recap on Jupyter, Python and more*
10:45-12:00- Explore MRI data with Nibabel and Nilearn
13:00-13:30- How to set up your system, using Conda and Docker
13:30-15:00- Functional connectivity and machine learning
15:00-15:30- Innovations in neuroimaging tools (Part I)
15:30-...- Open ended for questions
09:30-10:00- Introduction to Nipype
10:00-11:00- Exploration of Nipype’s building blocks
11:00-12:00- Creating a Nipype Pipeline from A-Z
13:00-14:00- PyBIDS and statistical analysis of fMRI data
14:00-15:00- Multivariate pattern analysis using Searchlight and Deep Learning
15:00-15:30- Innovations in neuroimaging tools (Part 2)
15:30-...- Open ended for questions
* While we include a short recap of Jupyter and Python, we won’t be able to cover this section in details. So if you’re new to them, please make sure to take a look at this and this notebook before the session to familiarize yourself with these tools. Both notebooks can also be explored interactively within the Docker environment.
16-17 November 2020: M/EEG analysis with MNE Python
Richard Höchenberger, INRIA Saclay, France
The workshop will cover a broad range of topics to help you get to know all essential parts of MNE-Python for conducting MEG and EEG data analysis:
- loading, filtering, and inspecting raw data
- working with BIDS data
- epoching and artifact correction
- creating and visualizing evoked responses (ERP / ERF)
- contrasting evoked responses of different experimental conditions
- decoding neural responses (machine learning)
- performing time-frequency analysis
- estimating and visualizing cortical sources (source localization)
- conducting a group analysis
We will start with a 30 min warm-up at 9:00 a.m. UTC on both days. During this warm-up, we can resolve technical difficulties you might be experiencing. The actual workshop will start at 9:30.
Lunch break is from 12 to 1, and we will continue until 3:30 or longer if need be (to hopefully address all open questions).
23-24 November 2020: Diffusion MRI with DIPY
Rafael Henriques, Champalimaud Centre for the Unknown, Lisbon, Portugal
In this two-days workshop, you will learn how to use different diffusion MRI (dMRI) techniques in an open-source environment. During the workshop, I will explain the theory behind different diffusion MRI techniques. This will be followed by practical tutorials based on the Diffusion in Python (Dipy) package (dipy.org).
To participate in the workshop, you will need to install Dipy and Fury. To install the dependencies of these two packages, I suggested that you install anaconda (in the anaconda website you will find the installers for different operative systems).
After installing anaconda, open an anaconda prompt terminal and type:
conda install -c conda-forge dipy
For some Dipy visualization tools you also need to have Fury installed. For this, in the anaconda prompt terminal, type:
pip install fury
More information on Dipy and Fury installation can be found in the following page).
The workshop will start with a 30 min warm-up at 9:00 am UTC on both days. During this warm-up, we can resolve technical difficulties you might be experiencing. The actual workshop will start at 9:30. Lunch break is from 12 to 1, and we will continue until 3:30.
Below you can find the topics discussed on the two days:
- Workshop introduction
- Diffusion MRI data as 4D images and simple Apparent Diffusion Coefficient calculation
- Diffusion Tensor Imaging (DTI)
- Introduction to DTI tractography
- Exploration of different reconstruction techniques for tractography
- Probabilistic vs Deterministic Tractography
- High angular resolution diffusion imaging (DSI, Q-ball, CSA)
- Open questions and discussion
- Diffusion MRI pre-processing
- Diffusion MRI signal representation techniques (DTI, DKI, MAP-MRI)
- Diffusion MRI microstructural models (WMTI, SMT, fw-DTI)
- Combining advanced dMRI metrics and tractography
- Overview of other tools available in Dipy
- Discussion of the future directions on Dipy development
- Open questions and discussion
Johan Carlin (MRC CBU) for more details.