Add context in the notebook

[jcohenadad]

Currently the notebook only has one fig for spinal cord results, without context.

We need to:

• [x] add an intro
• [ ] add rationale/methods for brain analyses
• [ ] add rationale/methods for SC analyses
• [ ] discussion/conclusion

[zelenkastiot]

Hello @jcohenadad, I can add chapters in the book in no time I only need the markdown text for each chapter, since I don't know what exactly to put inside as context.

If you send me the text I can manage to add the chapters, latex formulas, please let me know how you envisioned it.

[jcohenadad]

Introduction

Magnetic resonance imaging (MRI) comes with different flavours-- or as physicists would call it: pulse sequences. A pulse sequence is a bit like a musical orchestra. There are different instruments and each play at a particular time and with a particular pitch (frequency) and amplitude. In an MRI pulse sequence, the instruments are the hardware (gradients, RF controller, antenna, etc.). By driving the orchestra in a smart way, we can obtain more than a pretty picture of the inside body. In fact, we can obtain numbers that reflect the state of the tissue microstructure. For example, if there is a demyelinating lesion in multiple sclerosis, some sequences would enable to quantify the amount of demyelination. These particular sequences fall under the umbrella term Quantitative MRI, or qMRI.

MRI data acquired with a special pulse sequence are then analyzed. The analysis includes complex processing such as image non-linear registration, filtering and noise removal. Then, a biophysical model is usually fitted to the data, in order to extract physically meaningful parameters such as the density of neuronal cells, the orientation of white matter fibers or the relative concentration of brain metabolites. Given its important applications in the diagnosis and prognosis of traumas, neurological diseases and cancers, qMRI offers lots of promises and has become extremely popular in the neuroscience and pharmaceutical research community.

However, many of these qMRI techniques are currently not used routinely in clinics. One of the main reason is that data processing is long and tedious, as illustrated in Figure XX.

_Figure XX. Illustration of a typical qMRI processing pipeline. The "start" and "finish" images are modified from the comics Piled Higher and Deeper by Jorge Gabriel Cham._

Also, quantitative MRI data require complex analysis pipelines that are often executed manually and hence suffer from poor reproducibility. By being popular and requiring multiple high-level expertise at the same time, the field of qMRI is also cursed with an increasing amount of studies that other researchers have hard time reproducing [https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.1002506, https://onlinelibrary.wiley.com/doi/full/10.1002/mrm.27939]. A good resource illustrating this reproducibility crisis is the recent special issue about "Reproducibility in Neuroimaging" in the highly-respected NeuroImage journal (https://www.sciencedirect.com/journal/neuroimage/special-issue/102ML28LZ8W). It has become clear that in order to properly use of qMRI, the neuroimaging community has to educate and promote transparent acquisition and analysis tools.

[jcohenadad]

@zelenkastiot thank you, I'll take it from there and directly add content in the notebook, it will be easier