Hi
the easiest way to see how things work in featquery is to check the
log files within the featqury directory - this will tell you exactly
what steps are being performed. %BOLD are always calculated from the
cope/pe images which are relative to the EVs (and the haemodynamics
defined within) and not relative to data derived peri-stimulus plots
- hence you should expect to see differences.
hth
Christian
On 3 Jul 2008, at 14:52, Betina Ip wrote:
> Dear FSL list
>
> In featquery, I would like to use the average peristimulus time
> course of the mask
> (ps_tsplotc_zstat1.txt ), and convert the reduced raw data to % BOLD
> change. What is the
> respective baseline ?
>
> Various posting on fsl list suggest to use either a baseline of
> 10000, as the first level data is
> grand mean scaled to that, or better, to use the
> filtered_func_data.nii.gz. Does this mean the
> baseline is obtained by
>
> (1) using the mask in lowres space to obtain the mean timecourse of
> the filt_func_data.nii.gz
> using fslmeants? However the output is not the same as the
> mean_mask_ts.txt in the featquery
> directory.
>
> Or, do I use the mean_func.nii.gz in the .feat directory, and just
> extract the baseline value using
> fslstats -M without any mask ?
>
> (2) then, the % BOLD change is calculated by
>
> (ps_tsplotc_zstat1.txt*height of regressor)/ baselineofROI
>
> (3) finally, with the correct conversion the min, max and mean %
> BOLD change of the peristimulus
> timecourses should be close to those reported as PE-values in the
> featquery stats table. Is that
> correct ?
>
> Many thanks for your information!
>
> Betina
|