Trying to understand the markings on the glass brain plots

,
1

I’m trying to figure out the schematic markings on the glass brain plots. I dug into the code for a while, but didn’t find where they are from. They don’t seem to line up with sulci, as I would expect.

Central sulcus

The long front, mostly vertical line would seem to be the central sulcus, but does not line up with the tissue maps for pre- and post-central gyri from the Harvard-Oxford atlas. Unknown markings designated with arrows (which is basically all of them).

fig = plt.figure(figsize=(20, 20))
display = plot_glass_brain(None, figure=fig)
display.add_overlay('harvardoxford-cortical_prob_Precentral Gyrus.nii.gz', alpha=0.8, cmap='Reds', threshold=30)
display.add_overlay('harvardoxford-cortical_prob_Postcentral Gyrus.nii.gz', alpha=0.8, cmap='Blues', threshold=30)

Screen Shot 2022-04-29 at 12.23.31 PM
Screen Shot 2022-04-29 at 12.23.31 PM1920×753 83.6 KB

IFG

The two short vertical lines would seem to delineate the sulci beside pars opercularis and pars triangularis, but according to the HO atlas, they do not.

fig = plt.figure(figsize=(20, 20))
display = plot_glass_brain(None, figure=fig)
display.add_overlay('harvardoxford-cortical_prob_Inferior Frontal Gyrus, pars triangularis.nii.gz', alpha=0.8, cmap='Reds', threshold=30)
display.add_overlay('harvardoxford-cortical_prob_Inferior Frontal Gyrus, pars opercularis.nii.gz', alpha=0.8, cmap='Blues', threshold=30)

Screen Shot 2022-04-29 at 12.24.00 PM
Screen Shot 2022-04-29 at 12.24.00 PM1938×766 142 KB

Second question: are there different markings available, or can I use my own illustrations with lines for the glass brain plots?

Thanks for any enlightenment you’d care to share :slight_smile:

2

Hi @JohnAtl –

These are SVG images distributed with Nilearn and accessible here. They’re hand-drawn, so you could certainly re-use the functions with your own files ! They’d just need to be in the same format.

I think an important – and confusing – point with glass brains is that the have no depth information. So the axial rendering corresponds to all potential cuts, and therefore the denoted features are all collapsed into a single view. This makes it difficult to do precise anatomical localization !

It is not impossible: For the pre- and post-central gyri, for example, the sagittal view seems to better capture these features. But in general, I’d recommend against using glass brains for inference on exact structures.

HTH,

Elizabeth

3

Yes!

Just reviewing with my advisor, and wondering ‘What is that?’ ‘If that’s this, why is it up there?’, etc.

Thanks!

1 Like