|
Size: 2086
Comment:
|
← Revision 13 as of 2011-05-19 15:29:07 ⇥
Size: 2375
Comment: the latest is FsFastTutorialV5.1
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 1: | Line 1: |
| Describe FsFastTutorial/800 Normalize signal intensity here. | ## page was renamed from FsFastTutorial/800_Normalize_signal_intensity <<Navigation(slides)>> |
| Line 3: | Line 4: |
| 6. NORMALIZE SIGNAL INTENSITY USING INORM-SESS The inorm-sess command generates a file containing mean intensity data for the entire functional volume for each run, whether raw or motion corrected. Intensity normalization is a slight misnomer of this process since the global mean of the fMRI signal inside the tissue is calculated by segmentation of tissue from air. The number is later used to rescale the data so that when intersubject averaging is done, all subjects have the same global mean. This step is also optional, but it is recommended if you are averaging data across subjects. Run inormsess on Bert, even though you won t be averaging data across subjects. To normalize the intensity of Bert s motion corrected data, cd to the Study Directory and type: inorm-sess sf sessid df sesspar motioncor Normalization takes about ten minutes to complete. OUTPUT: After running the signal intensity command separately for each functional run subdirectory, you should see four new files in each run subdirectory called: | Important note: THIS COMMAND IS NO LONGER NEEDED. When you run [[spatialsmooth-sess]] it automatically calls [[inorm-sess]] for you. If you won't be running [[spatialsmooth-sess]], that is the only time you need to consider running IntensityNormalization. |
| Line 5: | Line 6: |
| fmc.meanval; fmc.report; fmc.twf-over; and fmc.twf-under. attachment:book.jpg |
The [[inorm-sess]] command generates a file containing mean intensity data for the entire functional volume for each run, whether raw or motion corrected. IntensityNormalization is a slight misnomer of this process since the global mean of the fMRI signal inside the tissue is calculated by segmentation of tissue from air. The number is later used to rescale the data so that when intersubject averaging is done, all subjects have the same global mean. This step is also optional, but it is recommended if you are averaging data across subjects. |
| Line 8: | Line 8: |
| 1) In the text file called fmc.meanval in 007 you should see the number: 402.142299, which is the global mean. 2) The fmc.report file contains many statistics relating to the cleanliness of your data, the most important of which are: OV Zmax and OU Mean. The Zmax statistic is an indication of spiking or instability. The OU Mean statistic is a measure of distortion. These are meant to be indicators of a possible problem during scanning. However, it is difficult to assign hard thresholds below which the user doesn't need to investigate the integrity of the data. The user should examine these statistics for all the sessions in a study in order to determine appropriate levels. As a rule of thumb, a Zmax over 3.5 or an OU Mean under 25 should trigger a closer look at that run. For backwards compatibility, inorm also reports the same statistics produced by an older program called "stackfix". | Run [[inorm-sess]] on Bert, even though you will not be averaging data across subjects. To normalize the intensity of Bert motion corrected data, cd to the Study Directory and type: {{{ inorm-sess -sf sessid -df sesspar -motioncor }}} Normalization takes about ten minutes to complete on very slow computers. OUTPUT: After running the signal intensity command separately for each functional run subdirectory, you should see four new files in each run subdirectory called: fmc.meanval; fmc.report; fmc.twf-over; and fmc.twf-under. In the text file called fmc.meanval in 007 you should see the number: 402.142299, which is the global mean. The fmc.report file contains many statistics relating to the cleanliness of your data, the most important of which are: OV Zmax and OU Mean. The Zmax statistic is an indication of spiking or instability. The OU Mean statistic is a measure of distortion. These are meant to be indicators of a possible problem during scanning. However, it is difficult to assign hard thresholds below which the user doesn't need to investigate the integrity of the data. The user should examine these statistics for all the sessions in a study in order to determine appropriate levels. As a rule of thumb, a Zmax over 3.5 or an OU Mean under 25 should trigger a closer look at that run. For backwards compatibility, inorm also reports the same statistics produced by an older program called "stackfix". <<Navigation(slides)>> |
| Slideshow ^ |< << Slide 7 of 14 >> >| |
Important note: THIS COMMAND IS NO LONGER NEEDED. When you run spatialsmooth-sess it automatically calls inorm-sess for you. If you won't be running spatialsmooth-sess, that is the only time you need to consider running IntensityNormalization.
The inorm-sess command generates a file containing mean intensity data for the entire functional volume for each run, whether raw or motion corrected. IntensityNormalization is a slight misnomer of this process since the global mean of the fMRI signal inside the tissue is calculated by segmentation of tissue from air. The number is later used to rescale the data so that when intersubject averaging is done, all subjects have the same global mean. This step is also optional, but it is recommended if you are averaging data across subjects.
Run inorm-sess on Bert, even though you will not be averaging data across subjects. To normalize the intensity of Bert motion corrected data, cd to the Study Directory and type:
inorm-sess -sf sessid -df sesspar -motioncor
Normalization takes about ten minutes to complete on very slow computers.
OUTPUT: After running the signal intensity command separately for each functional run subdirectory, you should see four new files in each run subdirectory called:
fmc.meanval; fmc.report; fmc.twf-over; and fmc.twf-under.
In the text file called fmc.meanval in 007 you should see the number: 402.142299, which is the global mean. The fmc.report file contains many statistics relating to the cleanliness of your data, the most important of which are: OV Zmax and OU Mean. The Zmax statistic is an indication of spiking or instability. The OU Mean statistic is a measure of distortion. These are meant to be indicators of a possible problem during scanning. However, it is difficult to assign hard thresholds below which the user doesn't need to investigate the integrity of the data. The user should examine these statistics for all the sessions in a study in order to determine appropriate levels. As a rule of thumb, a Zmax over 3.5 or an OU Mean under 25 should trigger a closer look at that run. For backwards compatibility, inorm also reports the same statistics produced by an older program called "stackfix".
| Slideshow ^ |< << Slide 7 of 14 >> >| |