Dear Seung-Goo,
it is indeed related to modulation. Modulated data => differences in volume. Unmodulated => differences in density. For details see Ashburner & Friston (2000, Neuroimage, "Voxel-based morphometry- the methods"), section "The Effect of Spatial Normalization", and Good et al. (2001, Neuroimage "A Voxel-Based Morphometric Study of Ageing in 465 Normal Adult Human Brains").
Some interesting papers about modulation and other preprocessing aspects in clinical context might be Keller et al. (2004, Neuroimage), Eckert et al. (2006, Neuroimage), Henley et al. (2010, AJNR) and probably various others. For example I vaguely recall a review or meta-analysis focusing on VBM in schizophrenia and typical results for modulated an unmodulated data. There's also a recent paper by Radua et al. (2014, Neuroimage), stating that modulation leads to a somewhat decreased sensitivity.
Hope this helps,
Helmut
|