FYI...we did code a version of this in SPM5 and used it for some multisite
studies, would be easy to modify for SPM8. Happy to share.
Regards,
Vince
> -----Original Message-----
> From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] On
> Behalf Of Elizabeth Liddle
> Sent: Saturday, March 17, 2012 7:43 AM
> To: [log in to unmask]
> Subject: Re: [SPM] Missing voxels
>
> Yes, that seems like a useful approach. I've been experimenting along
those lines!
>
> -----Original Message-----
> From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] On
> Behalf Of Jason Steffener
> Sent: 17 March 2012 13:07
> To: [log in to unmask]
> Subject: Re: [SPM] Missing voxels
>
> Hello.
> I agree that this is an important issue that needs more discussion.
>
> I have often thought that a degrees of freedom map could be used. So
instead of
> assuming you have a fixed dof across the brain one would have dof that
vary. This of
> course complicates interpretation but results could be reported based on
height of
> significance for all voxels with dof greater than some value. And as
research studies are
> starting to use larger and larger number of subjects, I feel this approach
is starting to
> become more viable.
>
> Now how this would affect cluster extents and thresholds is another issue.
> This approach is possible, but only with some coding.
>
> Jason
>
>
> On Sat, Mar 17, 2012 at 8:47 AM, Elizabeth Liddle
> <[log in to unmask]> wrote:
> > When a study involves a large number of participants, the cumulative
> > effect of a few voxels masked out in each subject at 2nd level
> > analysis can lead to a very small group mask, even though for any one
> > voxel only a small number of subjects may have missing data. To make
> > matters worse, often missing voxels are towards the edge of the image,
> > meaning that the result composite mask is lacking the regions of
> > greatest interest - the cortex near the skull.
> >
> >
> >
> > One solution, obviously, is to use an explicit mask, and another is to
> > over-ride the default threshold of .8 for voxel inclusion at first
> > level, but these solutions means that voxels that are, by definition,
> > noisy, will be included. Do people on the list have any solutions to
this problem?
> >
> >
> >
> > Elizabeth
> >
> >
> >
> >
> _____________________________________________________________________
> _
> > _______
> >
> > Dr Elizabeth Liddle
> >
> > Lecturer in Translational Mental Health
> >
> > Division of Psychiatry
> >
> > E Floor, South Block
> >
> > Queen's Medical Centre
> >
> > University of Nottingham
> >
> > NG11 2UH
> >
> > UK
> >
> > Tel: +44 (0)115 74 84013
> >
> >
> >
> >
> > This message and any attachment are intended solely for the addressee
> > and may contain confidential information. If you have received this
> > message in error, please send it back to me, and immediately delete
> > it. Please do not use, copy or disclose the information contained in
> > this message or in any attachment. Any views or opinions expressed by
> > the author of this email do not necessarily reflect the views of the
University of
> Nottingham.
> >
> > This message has been checked for viruses but the contents of an
> > attachment may still contain software viruses which could damage your
computer
> system:
> > you are advised to perform your own checks. Email communications with
> > the University of Nottingham may be monitored as permitted by UK
legislation.
|