Hi Martyn,
you are right - the option is not visible yet from the batch interface
as I wanted to make sure there are no unforeseen issues such as this:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=spm;f129ded.1502
I would be grateful if you could test it (by editing
spm_cfg_factorial_design.m as you mention) and check that design matrix
and covariance components are defined as expected.
Also, as discussed extensively on this list, remember that an
alternative is to use a partitioned error model as described in this
document:
http://www.fil.ion.ucl.ac.uk/~wpenny/publications/rik_anova.pdf
Best regards,
Guillaume.
On 22/06/15 12:34, Martyn Mcfarquhar wrote:
> Hi Guillaume,
>
> I notice in the recent SPM changelog that v6470 now allows for N-way
> interactions in the Flexible Factorial module. This is a very welcome
> change, however it seems that only the spm_design_flexible.m file has
> been changed, meaning that it is still not possible to specify an
> interaction term higher than a 2-way in the GUI.
>
> I know from previous discussions on the list that this can be addressed
> by changing the spm_cfg_factorial_design.m file so that fnums.num (on
> line 579) is [Inf 1] rather than [2 1]. I can make the change myself for
> now, but would it be possible to include this in the next set of updates?
>
> Best wishes
> Martyn
>
> -------------------------------------------------
> Martyn McFarquhar, PhD
> Neuroscience and Psychiatry Unit
> G.803 Stopford Building
> The University of Manchester
> Oxford Road
> Manchester
> M13 9PL
>
> Tel: +44 (0) 161 275 1728
> -------------------------------------------------
--
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
|