I assume that you mean bringing scans from lots of subjects into a common alignment. I'd suggest the following paper as a starting point:
Klein, Arno, Jesper Andersson, Babak A. Ardekani, John Ashburner, Brian Avants, Ming-Chang Chiang, Gary E. Christensen et al. "Evaluation of 14 nonlinear deformation algorithms applied to human brain MRI registration."Neuroimage 46, no. 3 (2009): 786-802.
The old SPM algorithms did not do so well in the above comparison (ie the Normalise button in anything prior to SPM12). SPM's Dartel did well with some of the datasets, but less well on others where the SPM segmentation had struggled with scans that were too closely skull-stripped. A more recent evaluation of Dartel (with some of the same data as in Klein et al) is presented in:
Ashburner, John, and Karl J. Friston. "Diffeomorphic registration using geodesic shooting and Gauss–Newton optimisation." NeuroImage 55, no. 3 (2011): 954-967.
Best regards,
-John