How to generate a 3d ROI sphere in a specific location within a 3d file (python / nilearn)

Question

My aim is to mask a 3D image at a given location with a 3d spherical mask. I want to extract all of the values of the image within this 3D mask

I am currently using python's nilearn.image.NiftiSpheresMasker function to achieve this in the following way:

central_coordinates=(x,y,z)
masker = NiftiSpheresMasker(central_coordinates, radius=rad)
values=masker.fit_transform(resampled_image)

However, this only returns the mean signal within the ROI. Is there a way to either:

Extract the mask that this function uses (so that I can obtain the values manually)
Use an alternative method to achieve this 3D ROI

P.s I have already tried using FSL's fslmaths -kernel which works for cubes but not spheres.

What does your 3d image look like? Is it just a 3d numpy array, a 3-tuple list or something else? — mmensing, Apr 05 '17 at 15:55

score 1 · Accepted Answer · answered Apr 06 '17 at 20:48

1

I managed to solve this by using the function create_sphere from the package nltools

answered Apr 06 '17 at 20:48

Joe

1 Answers1