BrainHack 2024: NiPreps

name: title
layout: true
class: center
---
layout: false
count: false

.middle.center[

## @ BrainHack 2024

### Oscar Esteban
#### CHUV | Lausanne University Hospital

###### [www.nipreps.org/assets/BrainHack2024](https://www.nipreps.org/assets/BrainHack2024)
]

---
name: newsection
layout: true

.perma-sidebar[

<a rel="license" href="http://creativecommons.org/licenses/by/4.0/"><img alt="Creative Commons License" style="border-width:0; height: 20px; padding-top: 6px;" src="https://i.creativecommons.org/l/by/4.0/88x31.png" /></a>
 Standardizing neuroimaging workflows

]

---

# Why standardizing?

.boxed-content[
.distribute[
* Increase reliability -- classic test theory:
  * Repetition
  * Standardization

* Other
  * Pushing the *truck factor* above 1.0.
  * Engage users

*Standardized Preprocessing in Neuroimaging: Enhancing Reliability and Reproducibility* doi:[10.31219/osf.io/42bsu](https://doi.org/10.31219/osf.io/42bsu), chapter 3.1 of *Methods for analyzing large neuroimaging datasets* doi:[10.17605/OSF.IO/D9R3X](https://doi.org/10.17605/OSF.IO/D9R3X) edited by R. Whelan and H. Lemaître.
]]

---

# Facets to standardize

.boxed-content[
.distribute.large[
* **Outer interface** (inputs and outputs). Also, consider an internal data structure format for high throughput.
* **Implementation** (WDL vs. programmed): code styling, best practices, BIDS-Apps.
* **Modularize** (see *NiPreps* and *TemplateFlow*).
* Use **containers** to ensure software delivery and reproducibility.
* Implement testing and **continuous integration** to catch errors early and streamline development.
* **Version** the workflow and its components to ensure compatibility and track changes, issue **LTS**.
* Promote **community** and social standardization through collaboration, documentation, telemetry, and open-source practices.
* **Visual reporting**
]]

---

# "Analysis-grade" data

.larger[
The *NeuroImaging PREProcessing toolS* (*[NiPreps](https://nipreps.org).org*) augment scanners to produce *analysis-grade* data (= **directly consumable by analyses**)
]

.pull-left[

***Analysis-grade* data** is an analogy to the concept of "*sushi-grade (or [sashimi-grade](https://en.wikipedia.org/wiki/Sashimi)) fish*" in that both are:

.large[**minimally preprocessed**,]

and

.large[**safe to consume** directly.]
]

.pull-right[
<img align="right" style='margin-right: 50px' src="https://1.bp.blogspot.com/-Osh4H4WXka0/WlMJmVgkZTI/AAAAAAAAEMY/GynUzSomJ-EBiyqv2m-maiOyKSM7SOmNACLcBGAs/s400/yellowfin%2Btuna%2Bsteaks%2Bnutrition.jpg" />
]

---

<img src="../nipreps-chart.png" width="63%" /> 
NiPreps (<a href="https://doi.org/10.31219/osf.io/ujxp6">Esteban et al., 2020</a>)

---

# BIDS-Apps: subject-wise parallelization

<a href="https://doi.org/10.1371/journal.pcbi.1005209">
<img src="../journal.pcbi.1005209.g002.png" width="90%" /> 
(Gorgolewski et al., 2017)
</a>

---

## The individual report

<video controls="controls" width="70%"
 name="Video Name" src="../fmriprep-report.mov"></video>

---

template: title
layout: false
.middle[

<img src="https://github.com/oesteban/fmriprep/raw/f4c7a9804be26c912b24ef4dccba54bdd72fa1fd/docs/_static/fmriprep-21.0.0.svg" width="95%" />

]

---

# fMRIPrep's usage

---

# fMRIPrep's usage

---

# *TemplateFlow*

<img src="https://www.templateflow.org/assets/templateflow_fig-birdsview.png" width="60%" /> 
(<a href="https://doi.org/10.1038/s41592-022-01681-2">Ciric et al., 2022</a>)

---

# *TemplateFlow* | Client

---

# Conclusion

<img src="https://media.springernature.com/full/springer-static/image/art%3A10.1038%2Fs41596-020-0327-3/MediaObjects/41596_2020_327_Fig1_HTML.png?as=png" width="90%" /> 
(<a href="https://doi.org/10.1038/s41596-020-0327-3">Esteban et al., 2020</a>)

---

.boxed-content[
.middle.center[
# Thanks!

### Questions?

]
]