# NiPreps Hackathon 2024

NiPreps Project page: https://github.com/orgs/nipreps/projects/8/views/1

The project page can be used to add issues. At some point we will be converting this brainstorming into notes there.

## Topics

### Diffusion MRI (plus PET for eddymotion)

**Team** (this is proposed by OE, but anyone can move themselves or write themselves down):
- Teresa
- McKenzie
- Jon
- Matt (remote)
- Erin (remote)
- Elodie (remote)

**MRIQC**:
- Implement new metrics / revise existing
- Increase reliability by means of atlas-based segmentation
- Finalize details in visual reports:
- Summary
- Decide on visualizations, perhaps color FA?
- Add masks on visualizations
- Train a classifier (mriqc-learn)
- Finalize protocol paper
- **Deliverable**: protocol paper & MRIQC release
- **Preparations**:

**eddymotion**:
- Finalize PET models
- Implement Gaussian Process
- Integration with MRIQC and dMRIPrep
- Comparison vs. FSL eddy
- **Deliverable**: paper draft
- **Preparations**:
- Martin - test data for PET
- Oscar - test data for DWI
-

**dMRIPrep**:
- Unclear at this point if worth planning.

**GPU**:
- SynthStrip / docker / conda / TF and PyTorch

### Functional MRI

**Team** (as above, this is just a proposal):
- Ursula
- Celine
- Elizabeth
- Melanie
- Yibei (?)
- Steven G.
- Lea
- Taylor (remote)
- Mathias
- Mary (?)

**MRIQC**:
- Revise BOLD metrics and solidify multi-echo BOLD
- [Compute only one set of IQMs and files for multi-echo datasets](https://github.com/nipreps/mriqc/issues/1184)
- Interested contributors: Celine Provins
- Include new metrics (see https://github.com/nipreps/mriqc/milestone/13)
- NiReports (see below)
- **Deliverable**: new MRIQC release
- Interested contributors: Melanie Garcia, Steven Giavasis, McKenzie Hagen

**fMRIPrep**:
- *fMRIPrep-rodents* - first "official" release
- Multiecho projects
- Improved reporting
- Make it easy to apply fMRIPrep surface transforms after running MEICA
- Taylor Salo: FWIW, I think the recommended workflow should involve denoising fMRIPrep-generated surfaces with the MEICA components, rather than transforming the MEICA-denoised data to surfaces.
- Fit-apply
* Make work good
- Denoising
- [fmripost-aroma](https://github.com/nipreps/fmripost-aroma)
- Interested contributors: Taylor Salo
- [fmripost-tedana](https://github.com/ME-ICA/fmripost-tedana)? (see above)
- Interested contributors: Taylor Salo
- Integration with physiological signal workflows using [Physiopy](https://physiopy.github.io/)?
- Interested contributors: Mary Miedema
- Build left and right flip detection tool in fMRIPrep (integrate AFNI pipeline)
- Warn flip in summary
- Interested contributors: Celine Provins
- 3dVolreg / 7T

**fMRIPrep-Infants / NiBabies**

- Finalize fit/apply port
- Sprint on documentation

### Middleware and reproducibility fundamentals

**Team** (as above, this is just a proposal):
- Basile
- Hao-Ting
- Yohan
- Feilong
- Taylor (remote)
- Ross
- Mathias
- Dylan
- Dorota
- Steven M. (remote)
- Jeff
- Julien (remote)
- Mary (?)
- Yibei
- Lea (?)

**sMRIPrep**
- finalize gradunwarp integration: [smriprep](https://github.com/nipreps/smriprep/pull/355) [fmriprep]([https://github.com/nipreps/fmriprep/pull/3064)
- Interested contributors: Basile
- Longitudinal processing + output structure ([nipreps/smriprep#429](https://github.com/nipreps/smriprep/issues/429))

**Reproducibility**
- Rounding errors in MRIQC / reliability of metrics (OE: I believe this relates to findings in Yohan's paper for fMRIPrep)
- Denoising and reproducibility:
- Rerun fMRIPrep on the benchmark datasets of the denoising paper (see https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0270895)
- Potential regression test?
- compare runtime of the workflow in fMRIPrep original derivative and down sampled derivative.

**SDCFlows**:
- MEDIC dynamic distortion correction pipeline ([nipreps/sdcflows#36](https://github.com/nipreps/sdcflows/issues/36); [nipreps/sdcflows#435](https://github.com/nipreps/sdcflows/pull/435))
- Interested contributors: Taylor Salo
- SDCflows paper
- New estimation method
- New/improved figures

**Templateflow**

- Deep revision of submission pipeline
- Template "profile page":
- Build/use visualizations in NiPreps
- Generate Markdown
- Integrate in templateflow.org
- Integrate in submission pipeline

**Nipreps/Nitransforms resampler**

- Surface resampling
- BIDS' X5 format
- Pydra workflow

**NiReports**
- PETPrep: Integration of reports
- NiReports: Add more example cases and improve documentation (https://nireports.readthedocs.io/en/latest/quickstart.html)
- From MRIQC:
- Plot physiological recordings and task events
- Possible integration with Physiopy (?)
- [Make MRIQC rating widget scrollable](https://github.com/nipreps/mriqc/issues/1153)
- Migrate group reports and/or revise them, e.g., for brushing, scrolling, etc.