Changed the call to diffusion_tensor.set() to the new specific function set_nonspin_params().
The code for setting parameters that are non-spin specific is very much dependent on the analysis
type, and hence this has been shifted into a call into specific_fns.
