To install:

        Extract spmd99b.tgz; it will create a spmd99b directory.
	Add that directory to your Matlab path, either each time
	with
		addpath /my/path/to/spmd99b

	or by changing your MATLABPATH environmental variable
	(Linux/Unix)


To Use:

        FIRST NOTE:  SPMd-beta only works on models with independent
		     errors, that is *no* low-pass/temporal-smoothing
		     is permitted, and *no* AR(1)/intrinsic
		     autocorrelation is permitted. 

        1. Run "spmd"

        2. After cd-ing into a directory with completed SPM results,
           click 'Compute'.

           'Diagnosis Statistics', 'Studentized Residuals', 'Both' ?

                 It will ask you if you want to compute just
                 diagnostics or studentized residuals or both.
                 You can't do anything without diagnostics, but
                 the residuals take a while to create (take a
                 considerable amount of space, so you can skip
                 them initially, though they are highly recommended).

           'Use default (#) contrast?'

                 In the matlab window a default contrast will be
                 displayed; it will use to define a unit effect of
                 interest (See PCT, http://www.sph.umich.edu/fni-stat/PCT);
                 if that isn't the principal contrast of interest.

                 The diagnostic images will appear in your analysis
                 directory and a few .mat files will be created.

        3. Click 'Temporal summaries'

             Select the summaries you're interested in, or just click OK
             to get them all.

             Note the temporal cursor.

        4. Click 'Spatial summaries'

             Select the summaries you're interested in.


        5. 'Temporal Detail'

             For an interesting voxel, Right Click that voxel,
             *or* click 'Temporal Detail' button.

        6. 'Spatial Detail'

             For an interesting time point, click 'Spatial Detail'.
             This will bring up the residual images just before,
             during and after the point of interest. (if you didn't
             create residual images, you'll get a warning and
             will see the raw data images instead)'


Have fun!

We're looking forward to your feedback... don't be shy!  Email us
both at spmd-authors@umich.edu.

Tom Nichols
Wen-Lin Luo