Dear Christopher,
Sorry for the delay in responding, but we've had some internal discussion
of this and we are dubious about this approach. It would be a bit difficult
to run with the software and technically it would be more sensitive to the
model assumptions than usual analyses (having potential biases from
different processing) and so we are not recommending that this be done.
The best thing would be to obtain a set of scans from a control group with
your protocol and then run that through a standard FIRST vertex analysis
or look at volumes if you prefer.
All the best,
Mark
On 21 Feb 2011, at 17:16, Christopher Emerson wrote:
> Hello,
>
> I have a single, unique data set from a person with very high autobiographical memory. Because there are only a few people like this, I cannot perform a simple between groups comparison as could be done for a diseased vs. healthy study.
>
> Instead, I was possibly interested in using FIRST to segment the brain structures and compare these volumes with the model volumes (from the 336 subject database). However, in the FSL directory I only see .bmv files, and I am not sure how to interpret these in terms of, e.g. the output*_firstseg.nii.gz file I have. Is there a way to generate, e.g. a mean +/- std of the structure volume for the database values, or otherwise is there a way to compare this one unique dataset in a statistically meaningful way with the database?
>
> Thank you and best wishes,
> Christopher
|