BrainIAK Tutorials

[manojneuro]

Link to the repository: https://github.com/brainiak/brainiak-tutorials

Title of notebook: There are 13 notebooks, from basics to advanced techniques.

Authors: Manoj Kumar, Cameron T. Ellis, Qihong Lu, Hejia Zhang, Mihai Capota, Theodore L. Willke, Peter J. Ramadge, Nicholas B. Turk-Browne, Kenneth A. Norman

Short layman description: Advanced fMRI analyses have the potential to answer questions that mainstream methods cannot. BrainIAK aims to integrate these cutting-edge techniques into a single, accessible Python environment. To help users get started, we have created the following set of tutorials based on courses taught at Princeton and Yale Universities.

Description: The binder platform will enable anyone to access and run these tutorials, reducing the burden of installations and availabilty of computational resources to learn advanced fMRI analysis.

Link to notebook icon: https://github.com/brainiak/brainiak.github.io/blob/master/assets/img/logo.svg

Track: _tutorials preprint: https://osf.io/j4sbc/

Why is that resource useful for the neuroscience community:

The analysis of brain activity, as measured using functional magnetic resonance imaging (fMRI), has led to significant discoveries about how the brain processes information and how this is affected by disease. However, the most powerful analyses are complex and computationally intensive, creating a high barrier for entry into this field. Furthermore, the available materials to learn these methods are limited, the software is rarely open source, and the analyses are often difficult to run on large datasets without cluster computing. We have created interactive software tutorials that make it easy to understand and execute analyses on fMRI data using the BrainIAK package — an open-source package built in Python. We have released these tutorials freely to the public and have significantly reduced the burden of computational resources needed by the learner, by making it possible to run the tutorials using a web browser and an internet connection. We hope that this access and usability of the tutorials and underlying code — a virtual compendium for how to program and optimize the latest fMRI analyses — combined with the ease of running the tutorials without large computing resources, will accelerate training and discovery in cognitive neuroscience.

How much data storage space do you expect to use: e.g. 25 GB

How much execution time do you expect users to need: e.g. 2 - 8 hours depending on how much time they spend on the exercises

[ltetrel]

https://github.com/neurolibre/brainiak-tutorials

[ltetrel]

@manojneuro did you try the binder link since the last time ? Any other thoughts on this, if not I will consider your submission as resolved.

[pbellec]

@ltetrel @manojneuro please confirm that the notebooks run as expected on neurolibre. I will invite reviewers for the submission as soon as it's the case.

[manojneuro]

I have been traveling after OHBM, so just seeing these messages. I'll check and get back in a couple of days.

Manoj

On Mon, Jun 24, 2019 at 8:53 AM Pierre Bellec notifications@github.com wrote:

@ltetrel https://github.com/ltetrel @manojneuro https://github.com/manojneuro please confirm that the notebooks run as expected on neurolibre. I will invite reviewers for the submission as soon as it's the case.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/neurolibre/submit/issues/4?email_source=notifications&email_token=AGHDOMOKBZA7WF2NMOFDMU3P4A445A5CNFSM4HWHKR62YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODYLUU3Y#issuecomment-504842863, or mute the thread https://github.com/notifications/unsubscribe-auth/AGHDOMLYFV7AV44UPC7SGXLP4A445ANCNFSM4HWHKR6Q .

[manojneuro]

@ltetrel @pbellec I have started testing out the notebooks. The install looks good for the most part. I did encounter an error in 2 notebooks. The 10-isc notebook errors out when plotting an image. This is a nilearn function error and I suspect that it might be an older version. It works fine on google Colab. Another notebook, 12-hmm errors on check for a directory. As the folders on binder are read-only, this statement is causing an error. I will create an issue on github for these errors. Otherwise things are looking very good.

[manojneuro]

@ltetrel @pbellec Update: I have logged issues here: https://github.com/neurolibre/brainiak-tutorials/issues. Here is a summary. 2 notebooks need a little more memory. One notebook most likely needs a newer version of nilearn. One notebook fails on a directory check. We'll need to come up with a workaround for this.

All these notebooks are working fine on Google Colab.

[mathieuboudreau]

All these notebooks are working fine on Google Colab.

How well are they working on the NeuroLibre binderhub? https://binder.conp.cloud/

[mathieuboudreau]

Or are the issues you're describing issues that you had on neurolibre? Sorry it wasn't clear to me. If so, screenshots might be helpful to see what some of the errors are thrown.

[manojneuro]

These errors are on neurolibre. They work fine on google Colab.

[manojneuro]

@mathieuboudreau I have added the screenshot to the issue.

[ltetrel]

A reviewer have agreed to provide feedback on a part of this submission. You can follow it here neurolibre/brainiak-tutorials#7

[ltetrel]

A reviewer have agreed to provide feedback on a part of this submission. You can follow it here neurolibre/brainiak-tutorials#8

[pbellec]

@manojneuro there are several open comments by reviewers on https://github.com/neurolibre/brainiak-tutorials

You have also been invited to join the team on that repository, to simplify interactions. Please follow up at your earliest convenience. If you anticipate important delays before being able to address these comments, please let our team know.

[manojneuro]

@mathieuboudreau @pbellec @ltetrel Thanks for installing and reviewing the tutorials and providing detailed feedback. Over summer, I have limited bandwidth and will not be able to work exclusively on this, but will address the review comments for each tutorial. It will be at a slower pace, but we will get it done.