# How to get anatomical masks?

I recently read a paper called "An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions". I want to use the method of this paper to apply to my dataset in order to get anatomical masks. I already installed fsl on my Linux system. But I still cannot fully understand this paper. Specifically, I want to get Temporal Occipital Fusiform gyrus (TOFC) and the lateral occipital Cortex (LOC) region of interest. Anyone can help me?

I'm a TMS, but not an fMRI, researcher but I have experience using anatomical masks. I can guide you through the steps to make an anatomical mask, but unfortunately I don't have any code to offer you.

First of all you need a set of x, y and z co-ordinates that specify where TOFC or LOC are. One way to find these is to find a paper that reports the x,y,z values corresponding to a voxel that contains the peak BOLD signal change co-ordinates within the LOC or TOFC. An alternatively way, might be to find some anatomical studies that provide these co-ordinates for you. I cannot recommend one or the other to you; this is a decision for you to make based on the analyses you are doing.

Once you have found some LOC or TOFC co-ordinates you are happy with, you can create a mask based on them using FSLView (type in fslview in the Linux terminal to open it).

First of all, open FSLView and open one of the standardized brain images. You can enter your TOFC or LOC co-ordinates (in MNI or Talairach space - I'd recommend MNI) in FSL using the x, y and z boxes in the bottom left corner. This will move a crosshair on top of the peak voxel. Then click file - create mask. This will enable you to use the toolbar, on this toolbar you will see an icon shaped like a pencil. Use it to mark voxel the crosshair is pointing at, making it sure it corresponds to the voxels you entered earlier. Do this separately for both site

You can then save this you will have an image containing an anatomical mask indicating where LOC and TOFC are. If this is unclear, try this blog - an fMRI researcher in my lab found it really useful.