Hello again,
> Just to be follow up for clarity: when running DARTEL create template
> for each patient together with all 70 controls (70c+1p), I will have to
> do this for each patient separately, right? When for example one day by
> chance I have several (np) patients, it will not be possible to run them
> together (70c+np), but I need to run it each time again n * (70c+1p), right?
If you follow the logic I tried to lay out (I expect not everybody will
:) then yes.
> Using the SPM template or the IXI created template is equally unfair for
> all, true, but it actually means the systematic bias is just let in the
> dark and I have no guarantee that it is not affecting the results.
No, but you will be sure that it does not affect the difference you are
interested in. This is the main advantage.
> even don't know form which scanner, population etc. those templates are
> created, compared to the one created by my controls which correspond to
> the same scanner, the same age and sex of my patients... So I am not
> sure what is better!
Well, some of that information is provided in the databases, but yes,
which factor is more relevant in the individual setting cannot be
decided from afar. You could conceivably investigate this systematically
using f-tests on the covariate of interest, but it would take some time
to calculate all possible iterations. And of course, using the same
scanner and the same sequence is a strong argument.
Cheers,
Marko
> 2014-04-07 9:39 GMT+02:00 Marko Wilke <[log in to unmask]
> <mailto:[log in to unmask]>>:
>
> Hi,
>
> > Good point. Anyway, creating new DARTEL templates does not take
> too much
> > time, I saw.
>
> Well, you can see if this is the way you want to go. But then again, you
> could also use an externally-generated DARTEL template (such as the one
> coming with vbm8), or you could generate one from a control population
> that neither the patient nor your controls contributed to (such as the
> IXI dataset). This would then be "equally unfair" to both.
>
> > Sorry to insist, but what about my second question:
> >
> > Is there a recommended way how to normalize the whole T1
> image? Is
> > it automatically done when I select the initial T1 together
> with the
> > c1, c2, c3 images in the normalizeMNI step? I am not sure if the
> > image look OK like this.
>
> I do not use DARTEL much, and if so, from the command line, and not for
> the T1. Have you tried it to see what it looks like? What does the
> manual say? Can you take the deformation fields and apply them?
>
> Cheers,
> Marko
>
> > 2014-04-07 8:50 GMT+02:00 Marko Wilke
> <[log in to unmask]
> <mailto:[log in to unmask]>
> > <mailto:[log in to unmask]
> <mailto:[log in to unmask]>>>:
> >
> > Hi Markus,
> >
> > > Thanks for your diverse input ;-)
> >
> > That's science :)
> >
> > > @Marko: Yes, I plan to compare each patient to the healthy
> data
> > base of
> > > 70 patients (detection of local pathology).
> > [...]
> > > So is the difference of 1 additional patient in the 70
> other subjects
> > > really critical, given the smoothing, the SNR etc.?? Has
> this been
> > > evaluated?
> >
> > Not sure this has been evaluated rigorously, but my usual
> point is: what
> > do you reply to a reviewer who asks you to demonstrate why the
> > difference between subjects that you see in the end is not
> due to a
> > difference in data processing? And your answer is going to be
> "I hope it
> > is not" because you cannot prove it. If you have processed
> your controls
> > in a systematically different way than your patient, this is
> simply an
> > irreconcilable difference that you cannot make up. It is like
> scanning
> > children on one scanner and adults on another scanner and
> then finding
> > the effects of age. You cannot, as age is confounded by scanner
> > difference. In how far this effect is huge, I cannot tell
> you, but then
> > again, you do not usually look for huge effects in a patient
> anyway.
> >
> > Cheers,
> > Marko
> > --
> > ____________________________________________________
> > PD Dr. med. Marko Wilke
> > Facharzt für Kinder- und Jugendmedizin
> > Leiter, Experimentelle Pädiatrische Neurobildgebung
> > Universitäts-Kinderklinik
> > Abt. III (Neuropädiatrie)
> >
> >
> > Marko Wilke, MD, PhD
> > Pediatrician
> > Head, Experimental Pediatric Neuroimaging
> > University Children's Hospital
> > Dept. III (Pediatric Neurology)
> >
> >
> > Hoppe-Seyler-Str. 1
> > D - 72076 Tübingen, Germany
> > Tel. +49 7071 29-83416 <tel:%2B49%207071%2029-83416>
> <tel:%2B49%207071%2029-83416>
> > Fax +49 7071 29-5473 <tel:%2B49%207071%2029-5473>
> <tel:%2B49%207071%2029-5473>
> > [log in to unmask]
> <mailto:[log in to unmask]>
> > <mailto:[log in to unmask]
> <mailto:[log in to unmask]>>
> >
> > http://www.medizin.uni-tuebingen.de/kinder/epn/
> > ____________________________________________________
> >
> >
>
> --
> ____________________________________________________
> PD Dr. med. Marko Wilke
> Facharzt für Kinder- und Jugendmedizin
> Leiter, Experimentelle Pädiatrische Neurobildgebung
> Universitäts-Kinderklinik
> Abt. III (Neuropädiatrie)
>
>
> Marko Wilke, MD, PhD
> Pediatrician
> Head, Experimental Pediatric Neuroimaging
> University Children's Hospital
> Dept. III (Pediatric Neurology)
>
>
> Hoppe-Seyler-Str. 1
> D - 72076 Tübingen, Germany
> Tel. +49 7071 29-83416 <tel:%2B49%207071%2029-83416>
> Fax +49 7071 29-5473 <tel:%2B49%207071%2029-5473>
> [log in to unmask]
> <mailto:[log in to unmask]>
>
> http://www.medizin.uni-tuebingen.de/kinder/epn/
> ____________________________________________________
>
>
--
____________________________________________________
PD Dr. med. Marko Wilke
Facharzt für Kinder- und Jugendmedizin
Leiter, Experimentelle Pädiatrische Neurobildgebung
Universitäts-Kinderklinik
Abt. III (Neuropädiatrie)
Marko Wilke, MD, PhD
Pediatrician
Head, Experimental Pediatric Neuroimaging
University Children's Hospital
Dept. III (Pediatric Neurology)
Hoppe-Seyler-Str. 1
D - 72076 Tübingen, Germany
Tel. +49 7071 29-83416
Fax +49 7071 29-5473
[log in to unmask]
http://www.medizin.uni-tuebingen.de/kinder/epn/
____________________________________________________
|