Thank you for all the clarifications, Steve!
-Yoshiko
On Mon, 21 Jan 2008 15:56:44 +0000, Steve Smith <[log in to unmask]> wrote:
> Hi,
>
> On 16 Jan 2008, at 01:43, Yoshiko Yamada wrote:
>
> > Hi Steve,
> >
> > Thank you for your response! I have further clarification questions
> > regarding your reply to my third question.
> >
> > 1) Do I select the “cluster” option under “Thresholding”? I tried
> > both “voxel” and “cluster”, but I get the web report table only when
> > I selected “cluster”, so I hope this is right. (By the web report
> > table, I mean the table you see in “cluster_zstat1_std.html”. Is
> > this right?)
>
> That's right - at present you only get the web table if you select
> cluster-thresholding, and the corrected p-values are per cluster.
>
> > 2) By loading the thresh_zstat image onto FSLView, should I be able
> > to change the threshold by changing the number number in the Min
> > (and Max) window just above the images? It doesn’t let me change
> > threshold in FSLView, so I’m assuming that this is not possible.
>
> Yes, that's right - and you'll be changing the Z threshold by changing
> the Min intensity value.
>
> > 3) In the Cluster List in the“cluster_zstat1_std.html”, I see Z-MAX.
> > Are these z-values directly translatable into p-values? In other
> > words, if I see Z-MAX > 2.3, are the corresponding p-values < .01?
>
> Not exactly, Z values (including z-max) refer to voxelwise Z values,
> whereas in the cluster table, the p-values refer to whole clusters.
>
> Cheers, Steve.
>
>
> >
> >
> > Thank you in advance for taking time to reply!
> >
> > -Yoshiko
> >
> >
> > On Tue, 15 Jan 2008 06:23:54 -0000, Stephen Smith <[log in to unmask]
> > > wrote:
> >> Hi,
> >>
> >>> I have a few basic questions for clarification about FEAT post-
> >>> stats.
> >>>
> >>> 1) At each level of analysis, FEAT gives you tstat.nii.gz and
> >>> zstat.nii.gz
> >>> in the stat folder. Are
> >>> activations you see in these files not corrected for multiple
> >>> comparisons?
> >>
> >> These t and z-statistic images not p-value images are not even
> >> thresholded, so no they are not corrected for multiple comparisons!
> >>
> >>> 2) Do you have to choose either the voxel-wise or cluster-wise
> >>> thresholding in the FEAT post-stats
> >>> tab to have correction applied?
> >>
> >> Indeed - that's right.
> >>
> >>> 3) Does FEAT give you a zstat.nii.gz file which contain stats that
> >>> are
> >>> unthreholded but corrected for
> >>> multiple comparisons that we can load onto FSLView and look at
> >>> activations
> >>> at different thresholds?
> >>
> >> yes - you can set the p-value threshold to 1 so that no
> >> thresholding is
> >> done - then you can look at the "thresh_zstat" images in
> >> conjunction with
> >> the web report tables to see the (GRF-corrected) p-values.
> >>
> >> Cheers.
> >>
> >>>
> >>> Thank you very much in advance for your information!
> >>>
> >>> -Yoshiko
> >>>
> >>>
> >>
> >
>
>
> ---------------------------------------------------------------------------
> Stephen M. Smith, Professor of Biomedical Engineering
> Associate Director, Oxford University FMRIB Centre
>
> FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
> +44 (0) 1865 222726 (fax 222717)
> [log in to unmask] http://www.fmrib.ox.ac.uk/~steve
> ---------------------------------------------------------------------------
>
|