July 25, 2019 tedana newsletter

[tedana-newsletter]

Ported message from our previous tinyletter.com system:

tedana monthly(ish) newsletter

Dear contributors and users,

tedana is continuing a monthly(ish) newsletter (subscribe here) to keep the community abreast of our activities and projects.

We would love to hear your feedback on this newsletter. You can chat with the team in our Gitter channel here or email this month's editor, Joshua Teves, at x.

Thank you, and enjoy!

The tedana development team

OHBM Poster and Hackathon

There was a solid presence of multi-echo fMRI at the Organization for Human Brain Mapping meeting in June. We counted at least 33 posters and talks that used multi-echo fMRI. These included both research that applied existing multi-echo denoising techniques and presented new approaches to denoising with multi-echo data. The tedana project had several prominent roles at the conference. We presented a poster and software demonstration specifically for tedana and led a discussion in the open science room with over 20 participants. These forums helped us better focus several key roles for tedana.

1. The tedana software can continue to develop in a direction that supports multiple approaches for denoising multi-echo fMRI data.
2. Not every developer is currently interested in bringing their denoising approaches into tedana, but can play a coordinating role to make sure the range of multi-echo methods are known to users whether or not they run within the tedana code
3. There is an unmet need for practical information on how to collect and analyze multi-echo data

 

Based on the feedback on points 2 and 3, we have started to expand the background documentation on our readthedocs site to provide general recommendations and resources for current and potential users of multi-echo fMRI. Whether or not people contribute to the tedana code, we welcome contributions to this documentation that should benefit the wider multi-echo fMRI community. If you have or know of multi-echo publications that are missing from our list please add them. We don't yet have a documentation page for other openly accessible denoising approaches, but will happily add that section when other developers start to contribute information they want to share there.

Tedana Hackathon!

We are planning a tedana-focused hackathon in Bethesda, Maryland on November 6-8, 2019. The hackathon will be the first time many of the developers of this code will be working together in the same room. If you are interested in attending in person, contributing remotely, or have suggestions of priorities for this hackathon please contact Dan Handwerker as soon as possible!

Test Coverage

We have increased our test coverage from 42% to 79%! This is very exciting. Special thanks to Taylor Salo and Monica Yao for their hard work in making more tests, and to Taylor for migrating all of our tests onto one platform.

Hot Topics

There has been a lot of discussion about how to incorporate documentation submisions without asking submitters to participate via GitHub. We're interested in feedback for how this process might be managed. There's also been quite a few discussions about how the math of multi-echo fMRI works for some model estimation steps and component decision metrics. We've put up a lecture by Javier Gonzalez-Castillo here that we hope will help clarify some details.

Good First Issues 

Just a friendly reminder that if you'd like to contribute, we maintain a list of good first issues that we think would be a great way to get started!

Merged Changes

These changes are since the previous newsletter on June 3rd, 2019.

Fixes

- Fix PCA to use normalized variance explained (Taylor Salo)

Documentation Changes

- Corrections to TEDPCA documentation (Taylor Salo)
- Add this newsletter to README (Kirstie Whitaker)
- Add references spreadsheet to multi-echo page on readthedocs (Dan Handwerker)
- Add OHBM updates and poster to readthedocs (Elizabeth DuPre)
- Fix some documentation errors (Stephan Heunis)
- Add notebooks to explain TE dependence (Taylor Salo)
- Generate boilerplate and citations for a tedana run (Taylor Salo)
- Fix broken links (Taylor Salo)
- Automatically update readthedocs copyright year (Monica Yao)

Functionality Changes

- Add low memory option (Taylor Salo)
- Use normalized variance explained in PCA decision tree (Taylor Salo)

Testing Changes

- Migrate all tests to CircleCI (Taylor Salo)
- Add smoke tests to decay.py (Monica Yao)

Refactoring

- Changes model to metrics module (Joshua Teves)

Thank you to Taylor, Kirstie, Dan, Elizabeth, Stephan, Monica, and Josh for your excellent work! Congratulations to Stephan and Monica on being first-time contributors, thank you!