Hi Jim,
you can try and run featquery on the lower-level feat dirs (rather
than the higher-level gfeat directory) using the ROI as a mask. You
can do this easily for all the lower-level feat directory with a
single featquery call - that way you'll automatically get the maximum
Z stats and a lot of other info for each of the runs.
cheers
christian
On 8 Jan 2007, at 18:31, James N. Porter wrote:
> Hello Doug-
>
> Well, I suppose the problem is I'm unsure of the formula used to
> calculate z-stats from intensity values. Since FEAT doesn't know
> what my ROIs are, I assume that my N-size/population is all the
> voxels in the brain, but is mu the mean value of all brain voxels,
> or does FEAT use info from FAST to know when a voxel is labeled as
> grey or white matter and use the appropriate mean greyscale value?
>
> Basically, my issue is that when I run FEAT and take a look at, for
> example, cluster_zstat1.html, then I get a great table of clusters
> and voxel counts and max z-stats. However, these clusters are
> distributed across the brain, and I want to know specifically what
> are the max z-stats for each individual in a particular predefined
> cortical region. Yes, the table includes xyz coordinates and I
> could locate each cluster individually, but that's not a feasible
> option across dozens of subjects with multiple analyses. So, in an
> attempt to streamline things I've run reg-feat2anat and associated
> processes to create binary masks of particular cortical regions,
> and hoped to obtain the same cluster list & local maxima tables
> found in cluster_zstat1.html, but constrained by my ROIs via
> masking in Featquery or pulling rows out of mri_segstats.
>
> I suppose if I knew what FEAT was using as mu, sigma, and N, then I
> could use the max intensity value output to calculate my own max z-
> scores.
>
> Thanks for your assistance.
>
> Jim Porter
> TRiCAM Lab Coordinator
> Elliott Hall N437
> 612.624.3892
> www.psych.umn.edu/research/tricam
>
>
> Doug Greve wrote:
>> Hi James,
>>
>> mri_segstats will include the max for each segmentation. Is this not
>> what you want? Or is it a format issue?
>>
>> doug
>>
>>
>>
>>
>>
>> On Sat, 6 Jan 2007, James Porter wrote:
>>
>>> Hello-
>>>
>>> I'm hoping to obtain outputs of activation z-statistics by ROI.
>>> I've used
>>> FreeSurfer to create ROI masks for all of my subjects, and run
>>> Featquery to
>>> obtain information about the ROI. However, I was disappointed to
>>> see the
>>> output options do not give me the ability to generate tables of
>>> max z-stats,
>>> as the first-level FEAT output does.
>>>
>>> Likewise, I've used FreeSurfer's mri_segstats program to create
>>> stats tables
>>> delineated by ROI, but it gives similar output as Featquery.
>>>
>>> What path should I have gone down to obtain z-stat tables within
>>> masked
>>> ROIs? My first thought right now is to run a first-level FEAT
>>> analysis on
>>> the masked functional volumes, but I'm willing to bet that you
>>> all have a
>>> better solution.
>>>
>>> Thanks for your help.
>>>
>>>
>>
--
Christian F. Beckmann
Oxford University Centre for Functional
Magnetic Resonance Imaging of the Brain,
John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
Email: [log in to unmask] - http://www.fmrib.ox.ac.uk/~beckmann/
Phone: +44(0)1865 222551 Fax: +44(0)1865 222717
|