Thank you Michael.
However, when I look at the log/code it seems like the median is calculated from the motion corrected volumes, and then applied to the smoothed volumes. Would it not be more appropriate to calculate the median from the smoothed volumes? It will not make any big difference, but it seems more correct to calculate the median from the smoothed volumes.
# Calculate median from motion corrected volumes
/home/andek/fsl/bin/fslstats prefiltered_func_data_mcf -k mask -p 50
1570.906006
# Apply intensity normalisation to smoothed volumes, scaling = 10000/1570.9
/home/andek/fsl/bin/fslmaths prefiltered_func_data_smooth -mul 6.36575324164 prefiltered_func_data_intnorm
Cheers,
Anders
________________________________________
Från: FSL - FMRIB's Software Library [[log in to unmask]] för Michael Harms [[log in to unmask]]
Skickat: den 27 november 2012 16:58
Till: [log in to unmask]
Ämne: Re: [FSL] How to calculate normalisation factor?
Hi Anders,
For a question like this, your best bet is to track down the underlying
commands within $FSLDIR/tcl/featlib.tcl.
In this case, searching that file for "intnorm" quickly brings you to the
following section
---- SNIP ----
#{{{ intensity normalization
set normmean 10000
if { $fmri(norm_yn)} {
fsl:exec "${FSLDIR}/bin/fslmaths $funcdata -inm $normmean
prefiltered_func_data_intnorm"
set ps "$ps; multiplicative mean intensity normalization of the volume
at each timepoint"
} else {
set scaling [ expr $normmean / $median_intensity ]
fsl:exec "${FSLDIR}/bin/fslmaths $funcdata -mul $scaling
prefiltered_func_data_intnorm"
set ps "$ps; grand-mean intensity normalisation of the entire 4D
dataset by a single multiplicative factor"
}
set funcdata prefiltered_func_data_intnorm
#}}}
---- END SNIP -----
and it is relatively straightforward to trace back from there to see how
$median_intensity gets defined depending on the particular options that
you are using within FEAT.
cheers,
-MH
--
Michael Harms, Ph.D.
-----------------------------------------------------------
Conte Center for the Neuroscience of Mental Disorders
Washington University School of Medicine
Department of Psychiatry, Box 8134
660 South Euclid Ave. Tel: 314-747-6173
St. Louis, MO 63110 Email: [log in to unmask]
On 11/27/12 3:33 PM, "Anders Eklund" <[log in to unmask]> wrote:
>Dear FSL experts,
>
>I'm looking at the log from a first level analysis and notice that a
>normalisation of the intensity seems to be performed. Does anyone know
>how the normalisation factor (6.3657 in this case) is calculated?
>
>fslmaths prefiltered_func_data_smooth -mul 6.36575324164
>prefiltered_func_data_intnorm
>
>/Anders
|