Hi Katie, this is a problem in version 4 in which the scaling is a little weird. It has no effect on the statistics (ie, p-values), but when you extract the percent signal change, they appear to be very low. This has been fixed in 5.X. For 4.X, you will need to compute a scaling factor. This can be done in matlab with the following commands:

cd session/bold/analysis x = load('X'); tirf = x.flac0.ev(3).tirf; Xirf = x.flac0.ev(3).Xirf; scalefOld = sum(Xirf .* repmat(tirf,[1 size(Xirf,2)])); TR = x.flac0.TR; dt = tirf(2)-tirf(1); scalefNew = sum(Xirf)*TR/dt;

You would then multiply your % signal change that you get from the ROI analysis by scalefNew.

doug

Katie Bettencourt wrote:

Hi all,

I'm doing ROI analysis on functional data with an anatomical label, and while the activity looks strong and robust in tksurfer, the % signal changes I am getting are very low (and much lower than we've previously found), and I'm not sure what's going on.

I'm running Freesurfer 4.5 on red hat linux.

The design is a blocked design with 4 non-null conditions, baseline is fixation. Attached is a picture of the activity I get for 2 subjects for on of the conditions vs fixation (baseline) at a p<0.05 threshold with the ROI outlined in black as well as the output from roisummary-sess for these two subjects in that roi. I calculate % signal change based of the output from roisummary-sess, dividing each condition by the baseline in that file and then multiplying by 100. % signal changes are listed below, along with all commands run.

% signal change: 110222SKJ_connected 110222SP_connected 0.000988998467052 0.000533852720367 0.00105349836708 0.000556814127695 0.000408499366826 0.000327200054419 0.000440749316839 0.000436266739225

commands: mkanalysis-sess -analysis connect -TR 2 -paradigm connect.dat -designtype blocked -funcstem fmc -motioncor -runlistfile connect_runs.txt -inorm -tpexclude tpexclude.dat -nconditions 4 -timewindow 20 -gammafit 2.25 1.25 -noautostimdur

mkcontrast-sess -analysis connect -contrast connected_vs_unconnected -a 1 -c 2 mkcontrast-sess -analysis connect -contrast connected_vs_noise -a 1 -c 3 mkcontrast-sess -analysis connect -contrast connected_vs_phase -a 1 -c 4 mkcontrast-sess -analysis connect -contrast connected_vs_fix -a 1 -c 0 mkcontrast-sess -analysis connect -contrast unconnected_vs_noise -a 2 -c 3 mkcontrast-sess -analysis connect -contrast unconnected_vs_phase -a 2 -c 4 mkcontrast-sess -analysis connect -contrast unconnected_vs_fix -a 2 -c 0 mkcontrast-sess -analysis connect -contrast noise_vs_phase -a 3 -c 4 mkcontrast-sess -analysis connect -contrast noise_vs_fix -a 3 -c 0 mkcontrast-sess -analysis connect -contrast shape_vs_noise -a 1 -a 2 -c 3 mkcontrast-sess -analysis connect -contrast shape_vs_phase -a 2 -a 1 -c 4 mkcontrast-sess -analysis connect -contrast shape_vs_noisephase -a 2 -a 1 -c 3 -c 4 mkcontrast-sess -analysis connect -contrast act_vs_fix -a 2 -a 1 -a 3 -a 4 -c 0

selxavg3-sess -sf connect-sess -df connect.dir -analysis connect

#run ROIs setenv anal connect foreach roi ( infIPS supIPS loc ) foreach h (lh rh) func2roi-sess -roidef ${roi}-${anal}-${h} -analysis $anal -anatlabel ips_loc/${roi}-${h} -sf connect-sess -df connect.dir roisummary-sess -sumfile ips-rois/${anal}_${roi}-${h}.dat -roidef ${roi}-${anal}-${h} -analysis ${anal} -sf connect-sess -df connect.dir python roi_reader.py -r ${roi} -a ${anal} -z ${h} -f ips-rois -c 4 end end

Thanks,

Katie

---

---

---

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer