A09: Using custom MRI from individual subjects
To visualize dipoles (DIPFIT plugin) or distributed source activity (LORETA plugin) within subject's individual MRI, it is first necessary to normalize them to the MNI brain template. The procedure is relatively simple and is explained below.
Normalizing an MR head image to the MNI brain template
Normalizing an anatomical MR brain image to the standard MNI brain is quite simple and does not require specialized knowledge about MRI or fMRI data. First, download SPM2, uncompress, add path in Matlab. Start spm by typing >> spm on the Matlab command line.
Next, click on the fMRI time series button. The following menu will appear (along with other screens).
Click on the display button and select your anatomical MR image file by clicking on it. Then press, Done as shown below.
A display screen as shown below will appear. Your MR image must be oriented as the one shown below. If it is otherwise oriented, use the edit boxes: Pitch, Roll and Yaw to rotate it in 3-D until it matches the template below. Then press the Reorient image button on the bottom.
The next step is to normalize your MRI to the standard MNI brain image. To do this, press the Normalize button in the main menu below. A blue tab will appear below the main menu. Select Determine Parameters & Write Normalized.
The following file selection window will appear. Select T1.mnc, the most common MRI format. (See the SPM manual for other formats). Then press Done.
Then reselect MRI data file (the same one you selected originally). Press Done.
Another file selection window will appear asking for the name of the file to write. Reselect once more the same MRI data file as the output file. SPM2 will not overwrite it but will use its name. Press Done. A final window will appear asking for data for a second subject. Simply press Done again. SPM2 will take a few minutes to normalize your MRI to the MNI brain. (It first segments your image and then compares the segmentation to the segmented MNI brain). The output file will have the same name as your original file, but with the prefix "w". You may use this file in the DIPFIT interface as the background image. An example is shown below:
Note that the resolution of this MR image is relatively low and appears noisy because of the renormalization process used in SPM2. If you use other software to perform the normalization, EEGLAB should be able to read the output image using the Fieldtrip function read_fcdc_mri() to import various MRI data formats.
Thanks to J-R Duann for helpful advice - Arnaud Delorme, March 2005.