The defaults in the SPM12b segmentation for the number of gaussians in the GM and WM class are currently =1.

They should be at least =2, I'm sure.

Thought I should point it out, and make sure you haven't made them =1 for good reason.

Best

Richard