There may be a few other things that are also worth taking into account:

	1) Also use a symmetric weighting image for the spatial normalisation.
	   This can be specified by modifying the spm_defaults.m  file at
	   around line 191 so that it uses a weighting image that has also
	   been left-right flipped and averaged.

	2) Remember that rotations, zooms and translations are also included
	   in the deformation fields - not just shape.  In order to discount
	   the effects of subject positioning in the scanner, you will need
	   to somehow normalise the deformation fields for subject position.
	   Also think about whether you want to normalise out head size from
	   the deformations.

	3) The deformation fields can be constructed from a linear combination
	   of basis functions.  When using DCT basis functions, some of the
	   coefficient describe symmetric deformations, whereas others
	   describe anti-symmetric deformations.  I'm not sure how you would
	   use this information for the kind of analysis you are planning, but
	   it may be worth thinking about.
	   
	4) There is a hidden option within SPM99, that although not yet fully
	   tested, is nevertheless there in anticipation of doing multi-variate
	   stats.  It computes a voxelwise Wilk's Lambda, which is transformed
	   to a Chi^2 statistic.  For the random field theory, the transformed
	   statistic field is assumed to be a Chi^2 field  (I think), with the
	   corrected p values based on these assumptions.  The multivariate
	   version is accessed by typing the following into Matlab (after
	   invoking SPM):
	   	spm_spm_ui
	   
	   Note that a Wilk's Lambda field transformed to a Chi^2 field is
	   not exactly the same as a pure Chi^2 field.  The corrections for
	   numbers of resels is therefore not exact, but I'm sure it's still
	   very close.  Although random field theory has been worked out for
	   Hotelling's T^2 fields (Cao, J. and Worsley, K.J. (1999). The
	   detection of local shape changes via the geometry of Hotelling's
	   T^2 fields. Annals of Statistics, 27, 925-942), I don't think it
	   has yet been done for Wilk's Lambda fields.

	5) If I was to this form of analysis, I would use flipped and unflipped
	   deformation fields, rather than using a deformation field for each
	   hemisphere.
	   
	6) Remember that this type of analysis only shows where the relative
	   position of structures is different.  It does not directly localise
	   regions with different shapes.  A simpler analysis could be based
	   on the regional volume changes estimated from the deformation fields.
	   Volume changes are encoded by the Jacobian determinants of the
	   deformation field.  It may also be worth taking a look at:
	   httP://www.math.mcgill.ca/chung/volumechange/volumechange.html

Best of luck,
-John

| I am thinking of a morphological analysis using deformation fields obtained
| with John's m-file spm_write_defs.m. The outline of the analysis is similar
| to that reported by Gaser, etal.(NeuroImage 10, 107-113, 1999). But I am
| plannig to add one more independent variable, hemisphere. The main objective
| of the analysis is to detect stuructural differences not only among groups but
| also between hemisperes (and their interactions). The point of my idea is as
| follows;
| 
|   Normalization: Nonlinear normalization will be performed using a SYMMETRICAL
|                           T1 template created by averaging the original
| template with itself
|                           L-R flipped.
| 
|   Separation: Deformation images (y1.img, y2.img, and y3.img) will be
| separated
|                      into two hemispheres.
| 
|   Statistics: 3-variate 2-way MANOVA will be performed.
|                     Independent variables: Group (3 categories defined by
| cognitive abilities)
|                                                          and
|                                                         Hemispere (L and R).
|                     Dependent variables: Deformation values stored in
| y1.img, y2.img, and y3.img
|                                                       for each hemisphere of
| each subject.
| 
|                     # of subjects: 14(group1), 16(group2), and 24(group3).
| 
| I am interested in the main effect of group and hemisphere, and interactions
| of them.
| A simpler design in my mind is making L-R subtraction images for deformation
| fields
| and performing 3-variate 1-way MANOVA using Group as an independent variable.
| (I guess this is equivalent to the interactions in the above 2-way analysis.)
| 
| Is this analysis valid?  Are there any better designs?
| 
| Any comments/suggestions would be appreciated.
| Regards,
| 
| Kota KATANODA
| Dept. Cognitive Neurosci, Fac. Med.
| Univ. Tokyo.
| 
| 
| 
| 
| 
| 
| 



%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%