Step 4: Normalize functional images and anatomy to MNI template

Method I.

This is the most typical method, which tends to work for most subjects and has 2 parts.

Part - 1 Normalize the inplane anatomy to the MNI structural template

Click on “NORMALIZE”

Select Normalize: Estimate & Write

Under Data , add as many new subjects as you have subjects/job (probably 1)

For Source image enter the coregistered structural image. This should probably be named rS.img.

Leave Source Weighting Image blank. This is usually used for abnormal or lesioned brains.

For Images to Write select the source image, rS.img This means that the parameters generated by the normalization of rS to the template will also be applied to rS.

Under Estimation Options select Template image.

You can find template images at usr/local/fmri_progs/spm5/templates/

Select the T1.nii image, if the inplane is T1-weighted or T2.nii if it is T2-weighted. T1 structural images look dark where the gray matter should be bright where white matter should be. If the structural images has the opposite contrast (bright where gray matter should be and dark where white matter should be), it was probably collected using a T2 acquisition sequence.

There are a number of other standard templates, that are provided within the same directory, and should have been downloaded automatically when SPM was installed.

Leave Template Weighting Image blank.

Under Writing options -> Wrapping, select “no wrap”

For Writing Options ->voxel size , enter [2 2 4] if you acquired your images on the UBC 3T scanner. Normalization reslices all files into images having the voxel size specified in your defaults. The default defaults’ voxel size is 2×2x2 mm. However, this may result in rather large files (the size of each image file depends on the number of voxels in it, and the smaller the voxel size, the more voxels there will be). Given that typical acquisition protocols give relatively low spatial resolution in the through-plane direction (the direction of acquisition), it is usually a good idea to reslice into images of voxel size 4 mm in the through-plane direction. For examples, in the case of axial acquisition, the voxel size would be 2×2x4 mm; for coronal acquisition, the voxel size would be 2×4x2 mm, etc. See Setting SPM Defaults for more information.

click Run. to proceed

This step will find the transformation that maps your inplane anatomy (and functionals) into MNI space. In addition to the normalized rS file (which produces a file named wrS), a new sn.mat file will be produced in the anatomy directory, containing the transformation parameters. These will be later applied to all functionals.

The results from this stage are displayed in the SPM Graphics window. At the bottom of this window, the template images is displayed and the normalized images is displayed on the right. As in the results from coregistration, you can click in different locations of the images to examine whether the outlines of the two images, or particular brain structures (e.g., the caudate) coincide.

__________________________________
Part - 2 Normalize all functionals using the parameters estimated in Part -1

click NORMALIZE

Select Normalize:write

Add one subject under Data.

Under parameter file, select the parameter file you created in part one. This file has encoded the instructions for normalizing your structural image into MNI space. It should be named something like rS_sn.mat. We will now use these parameters to normalize your functional images into MNI space.

Under Images to Write, select all time-slicing- and motion-corrected images from all sessions of the subject. These should all be have names in the form of aV???.img.

For Writing Options ->voxel size , again enter [2 2 4].

Run .

Method II.

This method normalizes the functional images directly to the functional MNI template (EPI.nii) provided in SPM and has 2 parts. This could be useful when the inplane anatomical image has failed to normalize well, because of unusual shape, excessive skull thickness, or some other problem. Note that Method III may be inappropriate if there is a large signal drop-off in parts of the functional images due to susceptibility artifacts.

__________________________________
click NORMALIZE

Select Normalize:write

Add one subject under Data.

For Source image, selectsession1/meanaV001.img

Leave Source Weighting Image blank.

For Images to write, select session1/meanaV001.img,

For Template images, slect usr/local/fmri_progs/spm5/templates/EPI.nii

Leave Template Weighting Image blank.

Under Writing options -> Wrapping, select “no wrap”

For Writing Options ->voxel size , enter [2 2 4] if you acquired your images on the UBC 3T scanner.

Run

This step will produce a normalized wmeanaV001.img file in session1/, as well as a file with the transformation parameters - meanaV001_sn.mat. Display the new wmeanax001.img to see how the normalization worked.

__________________________________
Part - 2 Normalize all functionals using the parameters esimated in Part - 1

click NORMALIZE

Select Normalize:write

Under Parameter file , select the parameter you created in Part1, named session 1/meanax001_sn.mat.

For Images to write, select all of your functional images

session1/ax*.img,

session2/ax*.img, etc

Run