Hi Jenifer,

Below are some answers:

Quoting "Juranek, Jenifer " Jenifer.Juranek@uth.tmc.edu:

Hi,

We've run LGI and have looked at the overlay.

1.  Is there a txt-type file containing the actual LGI values as
   

computed across aggregates of vertices? Noticed that the recon-all.log file does contain a quantitative value summarizing an average of each hemisphere.

Yes, the ?h.pial_lgi is a curv file containing one lgi value per cortical vertex. You can easily convert it in a text file using mri_convert, or read it in Matlab.

2.   Is there a way to plot the computed LGI info along the
   

rostral-caudal extent of each hemisphere?

What kind of rostro-caudal extent are you thinking about? Something like in the Zilles et al paper in 1988? These kind of rostro-caudal made sense before the 3D, when GI was computed on 2D coronal sections. As you had one value per slice with, representing them along the rostro-caudal extent was useful to get more precision on the localization of the differences. I would say that the vertex-wise approach is more sensitive to local differences that only the rostro-caudal axis... That said, if you have a rationale for that, everything is possible in Matlab!

3.  Best method to compare across subjects (or groups)? Seems like
   

qdec is based on vertex-wise analyses, whereas LGI uses aggregates of vertices.

In the validation paper, we compared group on a vertex-wise basis, exactly as you would do it for cortical thickness measurements (i.e. using Qdec). As a first step before the vertex-wise, you can also take the average lGI values provided in the recon_all.log in order to compare them between groups and see the direction of the difference.

Hope it helps,

Marie

Many Thanks for any assistance anyone can provide

Jenifer

~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~