Is it possible to use HERBS to visualize the fiber locations and probes of multiple subjects in the same figure?

[skrigel]

We want to visualize all images from all subjects in the same figure, so we have to plot them all together in the same coordinate space. The idea is to match the position of each fiber to one atlas section, like in the images from the paper linked below (but in the atlas space rather than plotting by hand). We would also like to visualize each probe against the atlas.

Here's a link to the paper to get a better idea of what I'm talking about: https://pubmed.ncbi.nlm.nih.gov/37495689/

[JingyiGF]

Oooops, you ask something big without providing clear and detailed information. As you might know from other issues and discussions, my background is not in Neuroscience. So may I ask what do you mean by "subjects"? "fiber"? I am sorry I have no idea what this figure is about. Honestly, I do not have time to read the paper, and I have no access to read the full paper. So I do not know what I can do to help you with the current info you provided. If you could explain a bit more what you are trying to do, with text and drawings, things will go easier I hope.

[JingyiGF]

The above is for register probes in volume atlas, if you want to register to PAXINOS, for example, you have to read https://github.com/JingyiGF/HERBS/blob/main/Tutorial/Atlas_Related/upload_slice_atlas.md. Not so many people tested this functionality for it, so if you found something unusual, I will fix :-)

[matiasandina]

Hmmmm, the design was after users loaded all probes that were saved, users should link the probe object that want to compare and then click the 2D-button to visualise them in 2D atlas slice. But as you can see, it was under development ;-)

Yes, this would be it, but ideally we would do it as a script or set of function calls rather than clicking on each animal file location, loading probes, linking them, etc. I'm not exactly sure how to solve the "probe in plate problem" since the probe usually spans more than one section, but showing all sections with annotated probe would be good enough.

I'm not sure how much of a development from your end this would be. We are happy to help if given some pointers for the way you would approach this

[JingyiGF]

Ah, now I see what you mean. So you are saying all probes are not lying perfect on the same cut, but you would like to visualize them on to a perfect cut. Right? This is a totally different story than what I thought. In this case, you have to do it with code. HERBS 2D will not support this. What you probably want to do,

1. save the voxel coordinates (x, y, z) for the insertion and terminus.
2. plot the atlas cut, for example, for coronal cut, plot a line go through insertion and terminus's (x, z).
   This is just for visualisation, as lenght traveled in each brain region of a probe will definitely be different when plotting the probe that cross multiple sections into a single section.

[matiasandina]

Not exactly. What I am saying is this: The thickness of the probe is greater than the section (400 um vs 100 um). In addition, there will never be perfect alignment between the insertion angle and the cut.

That means we want to trace the probe in whatever sections it appears (multiple sections) and visualize the probe in all those sections using a programmatic approach instead of the GUI. Each animal would have a set of atlas plates with portions of probe in 2D. Because animals are targeted to the same location, we would show multiple animals for each AP as shown in the plots above.

It's somewhat clear how to use the x,z cords. But it's not clear how to find the atlas cut for the specific location of the probe in herbs. How do you find and plot that atlas level ?

Is it possible to plot the atlas reference plate, as shown in the reference atlas instead of the tissue pics, with delineated regions in schematic fashion?

[JingyiGF]

Hmmm, ok. If you go to the folder where you save your atlas, you will see there are multiple files for the processed atlas. Find the pre-xxx pickle files, load it. Atlas data are saved inside as dictionaries of numpy arrays, which in another word, a 3d matrix with numbers of color levers. The length of x direction is number of sagital cuts in the processed atlas. And length of y direction is the number of coronal cut and so on. Say A is the matrix of atlas data, you want data for coronal cut 35, then A[:, 35,:] is the data for the atlas slice. Is this the problem? Or? Coordinates related to bregma will not work here, the easiest is to use voxel coordinates. Fetching atlas slices is easy, unless you want to get a rotated cut like what you can do in HERBS. For rotated cuts, I have to make a API.

Note: rotated cuts are numerical calculated that are not perpendicular to any axis. Whereas a perfect cut is perpendicular to a given axis.