3 means the 2-D arrays are being stacked along the 3rd dimension. To visualize, imagine that the arrays are gray scale images in the x-y plane. Now you use cat() to stack them vertically along the third, Z, dimension. See image below, and attached demo script.
The 3-D image above would represent a color image.
Another case could be where you have lots of images, like slices from a CT or MRI scan. These images, stacked along the 3rd dimension would represent a volumetric images.
If the number were 4, then it could mean that you're stacking 3-D (color) images along the 4th dimension (representing time) to create a movie.
Does that explain it sufficiently to you?
