Dear Alex,

At 10:44 AM 6/9/99 +0100, Alex leff wrote:
| 1) in spm99b results section there is no 'maxima' button. Are there
| plans to introduce one as I found it very helpful.

The "maxima" button in SPM94/5/6/7 results section jumped the cursor to the
nearest local maxima and then printed an exhaustive summary of local maxima
in the suprathreshold cluster containing this local maxima.

In SPM99 results, the cluster p-values button does much the same thing: The
cursor jumps to the nearest suprathreshold voxel, and the 16 highest local
maxima within the current suprathreshold cluster are surmised in the table.

The 16 is a hardcoded parameter. You can ask for more (once in the results
section) by running the function called by the "cluster" button directly,
with different parameters:
	spm_list('Cluster',SPM,VOL,4,20)
...would show at most 20 local maxima more than 4mm apart.

To jump to the nearest local maxima before using the "cluster" p-values
button, right-click on the MIP and select the "jump to nearest local
maxima" option.

Alternatively, clicking on the co-ordinates in a table of p-values and
statistics will jump the cursor to that location.

| 2) On a similar note, spm96 used to print out the adjusted rCBF voxel
| values (presumably the raw values that t-statistics and Z scores were
| derived from) on the matlab window when the button 'plot' was pressed in
| results. spm99b does not do this. Is this something that is also to be
| added in, or has it been taken away for a reason, or can one indeed get
| access to these values by entering a code (as was the case with spm95)?

SPM99's plotting routine (spm_graph.m) returns four variables to the base
Matlab workspace: Y, y, beta & SE. The values are dependent on what you
plot, and whether there is saved raw data for the voxel in question.
Generically, they are: Y - fitted data; y - adjusted data; beta - parameter
estimates; SE - standard error of parameter estimates.

Type the corresponding letter in the Maltab command window to see the
values.

Hope this helps,

-andrew

PS: (I see Dave McGonigle has also answered :-)


+- Dr Andrew Holmes [log in to unmask]
|  ___  __  __ Wellcome Department of Cognitive Neurology             |
| (  _)(  )(  )    Functional Imaging Laboratory,            Stats &  |
|  ) _) )(  )(__   12 Queen Square,                          Systems  |
| (_)  (__)(____)  London. WC1N 3BG. England, UK                      |
+------------------------------------- http://www.fil.ion.ucl.ac.uk/ -+



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