Hi Doug,
I am running into a similar problem that Yuko describes in Martinos Tech message 5099 (wouldn't let me reply directly). Using Stable 4, I made an old analysis last spring and extracted ROI values, yielding significant differences between two conditions as expected. I believe I originally used mkanalysis-sess.new. I later redid some func2roi analyses after redoing mkanalysis-sess (rather than .new) and the significant difference was gone even though the data set and label were the same, so I tried to recreate the same old analysis with the .new added back in. However, I noticed that my roisummary-sess outputs are still different:
*Original --> ID100 104 104 395.4163 31.5163 258.00 3.000 -0.00 7 1 -0.1397 0.6388 3.3232 3.9929 -18.0485 -6.8824 -7.0568
mkanalysis-sess -inorm (autosets -TER to 0.15) = mkanalysis-sess (no -inorm) (autosets -TER to 0.15) --> ID100 104 104 431.8531 31.1137 258.00 0.150 -0.00 7 1 -1.6920 -1.7342 -1.6874 -2.2781 -1.9065 -1.6461 -2.1075
*mkanalysis-sess.new -TER 3.000 -inorm = mkanalysis-sess -TER 3.000 -inorm = mkanalysis-sess -TER 3.000 (no -inorm) --> ID100 104 104 431.296 31.1078 258.00 3.000 -0.00 7 1 -33.3915 -34.3076 -33.5928 -44.2051 -37.5009 -31.9037 -41.6321
mkanalysis-sess.new -TER 3.000 (no -inorm) --> ID100 104 104 233.0353 16.8080 258.00 3.000 -0.00 7 1 -18.0419 -18.5369 -23.8846 -20.2622 -17.2380 -22.4944
The two * analyses should be identical (all the same settings during mkanalysis-sess.new), but for some reason aren't.
Besides there being a large range among the values (-0.1397, -1.6920, -18.0419, -33.3915), in the previous extraction there were some positive values, whereas now there are mostly negative values. I noticed that the TER column was different between the original and the new one (both were originally nonspecified and used the autoset -TER of 3 for the original and 0.15 for the new), so I tried making a new analysis specifying TER as 3.000 in both mkanalysis-sess and mkanalysis-sess.new, as well as turning -inorm and -autostimdur on and off.
When using mkanalysis-sess I primarily get large negative values (-33.3915 -34.3076) and -inorm, -autostimdur, & -noautostimdur don't make much of a difference if any. If I change the -TER (3.000 0.0300 0.150) my values change, with larger negative values with 3.000 and smaller values with 0.0300. However, even with setting the TER to 3.000 as in the original analysis, my values do not come out the same as we originally found.
I was wondering, what is the difference between the mkanalysis-sess and mkanalysis-sess.new commands and why do they result in such different func2roi values? The volumetric maps look identical for all of the analyses, so the problem seems specific to the ROI signal extraction. Also, how do you decide what TER to use? Our TR = 3.000 s, TE = 0.0300 s, and the paradigm files were set up in intervals of 3.000 s. I'm not sure if the scan parameters' TE value is the same as the mkanalysis-sess TER setting.
We would like to compare ROI % signal change extracted from the same ROI from two different phases (day 1, day 2). However, if we use the original ROI extraction values, we get % Sig Ch values in the -0.3 to 0.3 range, whereas with the new ROI extraction values, we get % Sig Ch in the -0.008 to 0.01 range, and the scale of the values seem incomparable. Any idea what might be going on or how to get something more like the original values?
Thanks, Garth