Meeting Minutes Feb 22, 2023
===

###### tags: `Templates` `Meeting`

:::info
- **Date:** Feb 22, 2023 17:00(UTC)
- **Agenda:** [link](https://hackmd.io/OhqyaRmgS4q2YkHNdZnuOg)
- **Participants:**
- **Minutes by:** Elodie <[e-mail]>
:::

### Official introduction of Elodie Savary
During the summer, the CZI open-source software grant was accepted. Several aspects of the grant are funded for two years. The proposal focused on community building. In particular, the proposal's cornerstone is hiring a Research Engineer with community management skills.
Elodie has been recently hired for this position. She will be responsible for all the action items regarding the CZI funding, particularly the organization of the Roundups.
### CZI EOSS5 award
The R01 grant focuses on individual *-Preps* (*fMRIPrep*, *dMRIPrep*, *ASLPrep*, *PETPrep*, *MRIQC*) and underlying neuroimaging developments (e.g., *SDCFlows*) while the CZI EOSS5 focuses on community building and scientific software engineering aspects.
The action items from the project proposal of the CZI grant are summarized here: https://github.com/orgs/nipreps/projects/4/views/1
### Reactivating the adoption of a governance model
A new version of the of governance documents is availiable in [the GOVERNANCE repository](https://github.com/nipreps/GOVERNANCE).
In particular, they define different roles for the people contributing to the organization and/or individual projects:

* Steering committee members: responsible for all technical oversight, project approval and oversight, policy oversight, and trademark management for the Organization.
* Maintainers : responsible for organizing activities around developing, maintaining, and updating the Project.
* Contributors : people who actively help or have previously helped the Project .

The details about the different roles can be found in the [Charter](https://github.com/nipreps/GOVERNANCE/blob/main/org-docs/CHARTER.md) and [the Project charter](https://github.com/nipreps/GOVERNANCE/blob/main/org-docs/PROJECT-CHARTER.md).

The plan is to assign MAINTAINERS to all projects listed below within the next six (6) weeks, to allow two weeks before the next *Roundup* on April 19.

* fMRIPrep
* MRIQC
* dMRIPrep
* SDCFlows
* Niworkflows
* sMRIPrep
* eddymotion
* fMRIPrep-rodents
* NiBabies & fMRIPrep-infants
* TemplateFlow
* PETPrep (To be confirmed)
* ASLPrep (Taylor Salo will discuss with Ted Satterthwaite)

We need to nominate an *acting NiPreps Steering Committee* that will ensure the MAINTAINERS are nominated, propose the first NiPreps Steering Committee with three (3) members or five (5), drawn from the MAINTAINERS of the projects before the next Roundup (April 19th, 2023) and organize the election.
The aim is to elect the Steering Committee at the next meeting.
At the moment two persons volonteered for the *acting NiPreps Steering Committee*: Chris Markiewicz and Elodie Savary. If you want to contribute please write to: [e-mail].

### Proposal about SDCFlows

SDCFlows is an attempt to have one place to handle the susceptibility distortion correction for all tools. The most recent version switched to a single representation of the field map. The attempts to make the new framework work led to a series of bugs. We want to set a MAINTAINERS list for SDCFlows. The MAINTAINERS will talk to other projects in order to outsource part of the work List of possible projects:

* AFNI: this is one of the largest gaps in AFNI (IMHO). The `3dQWarp` would need some improvement to generate B-Spline coefficient fields, not just the field. That is a hairy problem because we will fall into the compatibility issue (unless they accept NiTransforms/SDCFlows conventions, which is unlikely given their LPS coordinate system, among other factors.
* FSL: this is the most obvious package where help could come from, but they are not going to change (for many reasons, some more legitimate than others, but overall all justifiable).
* ANTs: this could be the best option for expansion, perhaps reaching out to Nick could help. On the downside list there is the lack for s-forms support, and some of the interpolation features of NiTransforms are missing in ITK. Nonetheless, there's been talks with Russ about a TOPUP implementation with ANTs which would make a lot of sense.
* An agreement on the internal representation could be placed under the NiTransforms' responsibility - this suggestion solves nothing, but indicates that we need to also understand/acknowledge searching for this idea trascends SDCFlows' domain.
*TORTOISE: C++ library, recently made open-source, GPL license. Eventual pitfall: ITK does not accept S(4) matrices and defines all transformations -> tricks must be applied to use ITK.

### Organization of hackatons
We want to organize two hackathons after ISMRM 2023 (Toronto) and OHBM 2023 (Montreal).
Members of the community based in Toronto and Montreal are very welcome to offer themselves to help organize (Please write to [e-mail]).
Last year's event led to successful projects:

* Separating fmriprep fitting method: The idea is to generate the transformation only to limit computational cost. The project is currently continued by Chris Markiewicz and Robert Smith. It is now implemented in a branch of sMRIPrep and will soon be implemented for fMRIPrep.
* Nireports: Module that compiles all visual generation tools. Now paused but will be restarted soon.
* Pydra conversion of MRIQC: Will restart next summer
* Eddymotion: implementation and benchmarking for diffusion MRI, will be generalized in the future.
### NiPreps Logo
Within the EOSS5 grant, the design of a logo was included. OE requested a first proposal with four options of logo + typography.
We will organize a poll to find the best option. This option can still be improved. Therefore the poll will allow people to enter comments.

### MIGAS technical update
MIGAS is designed for project usage and error reporting. It is currently integrated into fMRIPrep, MRIQC, and NiBabies. MIGAS can also be used to send messages to software. Mathias Goncalves is currently setting up a dashboard for the data.

### MRIQC

* SynthStrip has recently replaced traditional skull stripping. The main drawback is the RAM usage of SynthStrip.
* A [QC protocol](https://www.frontiersin.org/articles/10.3389/fnimg.2022.1073734/full) has now been published to explain how to interpret the reports generated by fMRIPrep and MRIQC. A [educational notebook](https://www.nipreps.org/qc-book/auto-qc/iqms_interpretability.html) has also been released.
* A [project](https://osf.io/7vqzr/) on predicting motion artefacts from MRIQC metrics with machine learning models has been submitted to OHBM.
* Mckenzie P. Hagen and Céline Provins are becoming MAINTAINERS of MRIQC.