It depends on the actual covariates and whether you used DOSS or DODS.
The default is DODS, which is a full interaction model. Assuming that
state is a categorical variable with 2 levels, you'd have 4 regressors
(gender by state) + 4 more regressors (gender by state by age). If the
SPSS model did not have 8 regressors, then it is not the same.
doug
On 1/30/15 1:26 PM, maaike rive wrote:
How should I have described the model?
------------------------------------------------------------------------
Date: Fri, 30 Jan 2015 13:09:40 -0500
From: gr...@nmr.mgh.harvard.edu
To: freesurfer@nmr.mgh.harvard.edu
Subject: Re: [Freesurfer] average values per cluster
>>>I told him it the model is a GLM with age, gender and state as
covariates
This is not a sufficient description. Several design matrices could be
derived from this description.
On 1/30/15 11:58 AM, maaike rive wrote:
The p value from SPSS is 0.156. I checked with a statistician,
just to be sure, but he says the SPSS model is correct... (he
doesn't know anything about freesurfer and FSGD files, but I told
him it the model is a GLM with age, gender and state as covariates).
------------------------------------------------------------------------
Date: Fri, 30 Jan 2015 11:46:37 -0500
From: gr...@nmr.mgh.harvard.edu <mailto:gr...@nmr.mgh.harvard.edu>
To: freesurfer@nmr.mgh.harvard.edu
<mailto:freesurfer@nmr.mgh.harvard.edu>
Subject: Re: [Freesurfer] average values per cluster
You should report the clusterwise p-value. The p-value that you
have computed is not interpretable. If it did not come out
significant, then it would be worrisome, but it is only a check
and cannot be used for anything. What p-value did you get from
SPSS? I'd be curious to know what the source of the descrpancy is
since this has happened several times now.
doug
On 1/30/15 3:25 AM, maaike rive wrote:
Hi Douglas,
I used the matlab code and the p value is signifiant, although
less so than the p-value I get from the clusterwise statistics
(0.0027 vs 0.00010). (Indeed I used an abs threshold for the
clusterwise statistics).
Which p-value should I report?
Thanks,
Maaike
------------------------------------------------------------------------
From: r_maa...@hotmail.com <mailto:r_maa...@hotmail.com>
To: freesurfer@nmr.mgh.harvard.edu
<mailto:freesurfer@nmr.mgh.harvard.edu>
Date: Thu, 29 Jan 2015 19:48:44 +0100
Subject: Re: [Freesurfer] average values per cluster
No not yet; I will do so!
> Date: Thu, 29 Jan 2015 13:44:33 -0500
> From: gr...@nmr.mgh.harvard.edu
<mailto:gr...@nmr.mgh.harvard.edu>
> To: freesurfer@nmr.mgh.harvard.edu
<mailto:freesurfer@nmr.mgh.harvard.edu>
> Subject: Re: [Freesurfer] average values per cluster
>
>
> The t is correct. When the contrast matrix only has one row,
then the F
> is just an unsigned t. I don't know anything about SPSS so I
can't tell
> from what you have sent whether it is the same model or not.
Try to get
> SPSS to output the design matrix. Did you try the matlab
code below?
>
> On 01/29/2015 01:40 PM, maaike rive wrote:
> > Hi Douglas,
> >
> > I checked but as far as I see it I used the same models. I
attached
> > the SPSS model and output as wel as the FSGD file and
contrast file
> > (to test a diagnosis x age interaction). Now that I come
to think of
> > it, I think the reason for the discrepant findings is that
I used a
> > t-contrast instead of an F-contrast for the interaction in
Freesurfer.
> > Apologies I did not think about this earlier. But maybe
there is
> > something else I did completely wrong.
> >
> > What do you think?
> >
> > If it is indeed the t vs F contrast, than how should I
specify that
> > the contrast is an F-contrast? Is it ok to just add
another line to
> > the contrastfile (so
> > 0 0 0 0 0 0 0 0 1 1 1 1 -1 -1 -1 -1
> > 0 0 0 0 0 0 0 0 -1 -1 -1 -1 1 1 1 1)?
> >
> > Thanks again,
> >
> > Maaike
> >
> > > Date: Mon, 26 Jan 2015 11:23:59 -0500
> > > From: gr...@nmr.mgh.harvard.edu
<mailto:gr...@nmr.mgh.harvard.edu>
> > > To: freesurfer@nmr.mgh.harvard.edu
<mailto:freesurfer@nmr.mgh.harvard.edu>
> > > Subject: Re: [Freesurfer] average values per cluster
> > >
> > >
> > > We get these kind of reports occasionally. When I ask
people to confirm
> > > that they use exactly the same design matrix in SPSS, I
never hear
> > back,
> > > so I assume that it gets resolved. So please check. The
other thing you
> > > can do is to run in matlab, something like
> > >
> > > cd glmdir/contrast
> > > X = load('Xg.dat');
> > > y = load('ocn.dat');
> > > C = load('C.dat');
> > > [beta rvar] = fast_glmfit(y,X);
> > > [F p] = fast_fratio(beta,X,rvar,C);
> > > p will be the p-value
> > >
> > > If you used an unsigned cluster-forming threshold (ie,
abs), then it is
> > > possible that some of the voxels are pos and some are
neg so that they
> > > average out
> > >
> > > doug
> > >
> > >
> > >
> > >
> > > On 01/26/2015 09:03 AM, maaike rive wrote:
> > > > Dear Freesurfer experts,
> > > >
> > > > Sorry to bother you again, but I have two more
questions about
> > > > extracting (thickness/surface/GI) values from a
certain cluster.
> > > >
> > > > As I understood, the abs.y.ocn.dat file gives the
average values
> > for a
> > > > given significant cluster (e.g. a cluster where there
is a
> > significant
> > > > AxB interaction).
> > > >
> > > > I may be completely misunderstanding things, but if I
use these
> > values
> > > > in SPSS for further statistics and test the same
interaction (AxB),
> > > > than according to SPSS this interaction is /not
/significant
> > > > (corrected for the same covariates as in the FSGD file).
> > > >
> > > > Could you tell me what is going wrong here? I do not
trust my
> > results now.
> > > >
> > > > Furthermore, is it possible (and if so, how?) to
extract the average
> > > > values of exactly the same cluster, but in an
independent group not
> > > > used in the analysis, for post-hoc comparisons in SPSS?
> > > >
> > > > Thanks,
> > > >
> > > > Maaike
> > > >
> > > >
> > > > _______________________________________________
> > > > Freesurfer mailing list
> > > > Freesurfer@nmr.mgh.harvard.edu
<mailto:Freesurfer@nmr.mgh.harvard.edu>
> > > >
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
> > >
> > > --
> > > Douglas N. Greve, Ph.D.
> > > MGH-NMR Center
> > > gr...@nmr.mgh.harvard.edu <mailto:gr...@nmr.mgh.harvard.edu>
> > > Phone Number: 617-724-2358
> > > Fax: 617-726-7422
> > >
> > > Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting
> > > FileDrop: https://gate.nmr.mgh.harvard.edu/filedrop2
> > > www.nmr.mgh.harvard.edu/facility/filedrop/index.html
<http://www.nmr.mgh.harvard.edu/facility/filedrop/index.html>
> > > Outgoing:
ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
> > >
> > > _______________________________________________
> > > Freesurfer mailing list
> > > Freesurfer@nmr.mgh.harvard.edu
<mailto:Freesurfer@nmr.mgh.harvard.edu>
> > > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
> > >
> > >
> > > The information in this e-mail is intended only for the
person to
> > whom it is
> > > addressed. If you believe this e-mail was sent to you in
error and
> > the e-mail
> > > contains patient information, please contact the
Partners Compliance
> > HelpLine at
> > > http://www.partners.org/complianceline . If the e-mail
was sent to
> > you in error
> > > but does not contain patient information, please contact
the sender
> > and properly
> > > dispose of the e-mail.
> > >
> >
> >
> > _______________________________________________
> > Freesurfer mailing list
> > Freesurfer@nmr.mgh.harvard.edu
<mailto:Freesurfer@nmr.mgh.harvard.edu>
> > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
>
> --
> Douglas N. Greve, Ph.D.
> MGH-NMR Center
> gr...@nmr.mgh.harvard.edu <mailto:gr...@nmr.mgh.harvard.edu>
> Phone Number: 617-724-2358
> Fax: 617-726-7422
>
> Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting
> FileDrop: https://gate.nmr.mgh.harvard.edu/filedrop2
> www.nmr.mgh.harvard.edu/facility/filedrop/index.html
<http://www.nmr.mgh.harvard.edu/facility/filedrop/index.html>
> Outgoing:
ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
>
> _______________________________________________
> Freesurfer mailing list
> Freesurfer@nmr.mgh.harvard.edu
<mailto:Freesurfer@nmr.mgh.harvard.edu>
> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
_______________________________________________ Freesurfer
mailing list Freesurfer@nmr.mgh.harvard.edu
<mailto:Freesurfer@nmr.mgh.harvard.edu>
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person
to whom it is addressed. If you believe this e-mail was sent
to you in error and the e-mail contains patient information,
please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was
sent to you in error but does not contain patient information,
please contact the sender and properly dispose of the e-mail.
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu>
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
_______________________________________________ Freesurfer mailing
list Freesurfer@nmr.mgh.harvard.edu
<mailto:Freesurfer@nmr.mgh.harvard.edu>
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer The
information in this e-mail is intended only for the person to whom
it is addressed. If you believe this e-mail was sent to you in
error and the e-mail contains patient information, please contact
the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to
you in error but does not contain patient information, please
contact the sender and properly dispose of the e-mail.
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu>
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
_______________________________________________ Freesurfer mailing
list Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer The
information in this e-mail is intended only for the person to whom it
is addressed. If you believe this e-mail was sent to you in error and
the e-mail contains patient information, please contact the Partners
Compliance HelpLine at http://www.partners.org/complianceline . If the
e-mail was sent to you in error but does not contain patient
information, please contact the sender and properly dispose of the e-mail.
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
