Page 1 of 1

Need help to interpret Cortical fold morphometric attributes

Posted: Tue Apr 05, 2016 4:37 pm
by ruslana
Hi all,

I am looking for any documentation regarding cortical fold attributes (as well as edge attributes).
I read a cortical graph by using aims library in python, and get a list of folds - each of them has a list of attributes attached. Some attributes are intuitive and easy to interpret, but others are not (at least for me)...
For example,
- what exactly 'size' measures (there are other attributes, such as surface area and various volumes, so how the size attribute differs?)
- how exactly the depth and thickness attributes relate?
- what are the moments?
- what is the difference between 'hull_normal' and 'normal' directions?
- what does a direction weight mean (i.e. 'hull_normal_weight')?
- what 'direction' is (there are already hull_direction, normal, and normal_hull, so what does this 4th direction attribute mean?)
After desperately searching the documentation online, I could not find answers to these questions.
I would highly appreciate if anyone could please point to any available documentation, or even properly commented source file where I could read about folds and edges attributes.

Thanks ahead,

Re: Need help to interpret Cortical fold morphometric attributes

Posted: Thu Apr 07, 2016 9:22 am
by riviere

There are 2 sets of sucus measurements: those directly in sulcal graphs (.arg files), which measure each fold element (not whole sulci). From your question I think I understand that it is what you are looking at, and I don't think it is what you want, or at least they cannot be used directly alone because they are measured on stuctures which are not complete sulci... Sulcus-wide meaurements are made by the morphometry processes in Morphologist: they basically regroup the former measurements at the scale of a sulcus (or a relation between two sulci), combining them the right way.

You have (very succinct) help on the morphometric measurements from the morphometry process in BrainVisa, as "viewers" for morphometric models (the eye button). A bit more precisely (but still shortly),
* size is the volume of voxels forming the sulcus object. Such voxels are (mostly) in the CSF inside folds and do not represent a physical cerebral structure. Moreover it depends on the voxel size, its orientation etc, so it is not a robust measurement, so we have replaced it by the sulcus mesh area in later models. And... I have relalized just yesterday that there was a but in the size properties (oups)... I have fixed it for the next release but it's unreliable for now.
* moments are invariant moments, derived from physical moments: inertia etc, and made invariant to position, orientation, scale. They have been used to characterize the shape of an object, but have been proved somewhat tricky because some of them sometimes present discontinuities (or more likely drastic variations).
* the hull is the brain hull, the external brain surface not going into sulci. The hull_normal for a sulcus is the (average) normal to the external brain surface at the positions of its connection with the given sulcus. It represents roughly the orientation of the depth of the sulcus.
* normal it the (average) normal direction to the sulcus surface: it's more or less orthogonal to the former.
* direction weight is the number of voxels for which the direction has been averaged. It is only useful when combining with other sulci parts, to get a correctly weighted sulcus average.
* direction (if I remember) is the general orientation vector for the fold voxels (I think it's the principal inertia axis direction). It may differ a little bit from the direction of the connection line with the hull (hull_junction).

But once again, you should not use these measures directly from the sulcus graph, they are meant to be recombined for several fold parts (that's whar the morphometry process does), and have been designed primarily for the sulci recognition algorithms.


Re: Need help to interpret Cortical fold morphometric attributes

Posted: Thu Apr 07, 2016 3:01 pm
by ruslana

Great thank for your reply.
It helps a lot!!!