Why do I get a Inf/NaN value when I create a 4D double array?

Dear experts,

I am trying to calculate an overlap between two masks in nifti format in a loop. The standard mask is binary, the looped one should contain numbers between 0:6.

ID_CODES=dir([wd '*.MRI.1']);        %dir of all subjects
ROImask=spm_read_vols(spm_vol('/Data/users/gulizar/verbgen_results/img/outmask_seg_all.nii'));     #an image/'mask', containing numbers between 0-6. 
for i = 1:length(ID_CODES)       % Loop over subjects
dp=fullfile('/Data/users/*/*/',ID_CODES(i).name, '/Images/wseg.nii'); %binary image that differs between the subjects 
seg(:,:,:,i)=spm_read_vols(spm_vol(dp)); %spm_read_vols simply reads the file (I don't have the image processing toolbox)
end
segf(:,:,:,:)=flipud(seg(:,:,:,:));      %I do this, because for some reason my files are mirrored
seg2(:,:,:,:)=griddedInterpolant(segf(:,:,:,:));
seg3(:,:,:,:)=seg2({linspace(1, size(segf(:,:,:,:),1),51),linspace(1,size(segf(:,:,:,:),2),75),linspace(1,size(segf(:,:,:,:),3),55),linspace(1,size(segf(:,:,:,:),4),156)}); %changes the dimension of each binary image, so that it is the same as the ROImask dimensions
for i= 1:length(ID_CODES)
locseg(:,:,:,i)=ROImask(:,:,:).*seg3(:,:,:,i);  %multiply ROImask with binary mask, for each subject
location(:,:,:,i)=mode(locseg(:,:,:,i));      %most frequent values in overlap between the masks, for each subject; output must be a number between 0-6.
end

When I do this, my output contains a lot of NaN values. In debugging mode I saw this happens early on while defining the variable 'seg'.

>> any(isfinite(seg(:))) --> ans = 1

Any idea why I get NaN values in my output, or an Inf value in 'seg' in the first place? How can I fix this? I tried changing Inf and NaN values to zero after defining seg, but somehow this does not do the trick.

Also I am pretty new in scripting, so any help is very much appreciated!