SPM toolbox: MNI Align

SPM12 toolbox to align an image to SPM's template space.

This toolbox uses the initial step from Segment under the hood.

This registration algorithm is more robust than Coregister when applied to tissue probability maps, and handles affine transforms.

Installation

1. Install SPM12
2. Copy or link the spm_mni_align folder under spm12/toolbox

GUI usage

1. Launch spm.
2. Open the Batch interface.
3. Load the toolbox from SPM -> Tools -> MNI Align. Usage is simlar to that of Coregister.
   • Estimate: align images to the template and update the header.
   • Reslice: reslice already registered images to the grid of another image.
   • Estimate and Reslice: align images to the template, then reslice them to the space of the template.

Note: By default, Estimate does not modify the header of the input files in-place but instead writes a new file.
To recover the in-place behaviour of Coregister, set the Filename Prefix option to an empty string.

Script usage

Estimate

matlabbatch{1}.spm.tools.align.estimate.source = {'<PATH TO FIRST IMAGE>'};  % image used in registration
matlabbatch{1}.spm.tools.align.estimate.other = {
                                                 '<PATH TO OTHER IMAGE 1>'
                                                 '<PATH TO OTHER IMAGE 2>'
                                                 '<PATH TO OTHER IMAGE 3>'
                                                 ...
                                                 };  % estimated transform is applied to these images too
matlabbatch{1}.spm.tools.align.estimate.ref = {
                                               fullfile(spm('dir'), 'tpm', 'TPM.nii,1')
                                               fullfile(spm('dir'), 'tpm', 'TPM.nii,2')
                                               fullfile(spm('dir'), 'tpm', 'TPM.nii,3')
                                               fullfile(spm('dir'), 'tpm', 'TPM.nii,4')
                                               fullfile(spm('dir'), 'tpm', 'TPM.nii,5')
                                               fullfile(spm('dir'), 'tpm', 'TPM.nii,6')
                                               };  % tissue probability map defining the template space
matlabbatch{1}.spm.tools.align.estimate.eoptions.rigid   = 0;      % 0 = affine, 1 = rigid
matlabbatch{1}.spm.tools.align.estimate.eoptions.sep    = [8 2];   % Distance between sampled points in mm
matlabbatch{1}.spm.tools.align.estimate.eoptions.fwhm   = [32 1];  % Fudge factor in mm (higher = stiff)
matlabbatch{1}.spm.tools.align.estimate.eoptions.prefix = 'e';     % Prefix of the output image
matlabbatch{1}.spm.tools.align.estimate.eoptions.dir    = {''};    % Output folder. Empty '' = same as input

Reslice

matlabbatch{1}.spm.tools.align.write.ref             = {'<PATH TO TARGET IMAGE>'};  % defines the resliced space
matlabbatch{1}.spm.tools.align.write.source          = {'<PATH TO SOURCE IMAGE>'};  % image to reslice
matlabbatch{1}.spm.tools.align.write.roptions.interp = 4;         % Interpolation order
matlabbatch{1}.spm.tools.align.write.roptions.wrap   = [0 0 0];   % Dimensions with circulant boundaries 
matlabbatch{1}.spm.tools.align.write.roptions.mask   = 0;         % Mask out voxels where at least one image is out-of-bounds
matlabbatch{1}.spm.tools.align.write.roptions.prefix = 'r';       % Prefix of the output image
matlabbatch{1}.spm.tools.align.write.roptions.dir    = '';        % Output folder. Empty '' = same as input

Estimate & Reslice

matlabbatch{1}.spm.tools.align.estwrite.source = {'<PATH TO FIRST IMAGE>'};  % image used in registration
matlabbatch{1}.spm.tools.align.estwrite.other = {
                                                 '<PATH TO OTHER IMAGE 1>'
                                                 '<PATH TO OTHER IMAGE 2>'
                                                 '<PATH TO OTHER IMAGE 3>'
                                                 ...
                                                 };  % estimated transform is applied to these images too
matlabbatch{1}.spm.tools.align.estwrite.ref = {
                                               fullfile(spm('dir'), 'tpm', 'TPM.nii,1')
                                               fullfile(spm('dir'), 'tpm', 'TPM.nii,2')
                                               fullfile(spm('dir'), 'tpm', 'TPM.nii,3')
                                               fullfile(spm('dir'), 'tpm', 'TPM.nii,4')
                                               fullfile(spm('dir'), 'tpm', 'TPM.nii,5')
                                               fullfile(spm('dir'), 'tpm', 'TPM.nii,6')
                                               };  % tissue probability map defining the template space
matlabbatch{1}.spm.tools.align.estwrite.eoptions.rigid  = 0;       % 0 = affine, 1 = rigid
matlabbatch{1}.spm.tools.align.estwrite.eoptions.sep    = [8 2];   % Distance between sampled points in mm
matlabbatch{1}.spm.tools.align.estwrite.eoptions.fwhm   = [32 1];  % Fudge factor in mm (higher = stiff)
matlabbatch{1}.spm.tools.align.estwrite.eoptions.prefix = 'e';     % Prefix of the output estimated image
matlabbatch{1}.spm.tools.align.estwrite.eoptions.dir    = {''};    % Output folder (estimate). Empty '' = same as input
matlabbatch{1}.spm.tools.align.estwrite.roptions.interp = 4;       % Interpolation order
matlabbatch{1}.spm.tools.align.estwrite.roptions.wrap   = [0 0 0]; % Dimensions with circulant boundaries 
matlabbatch{1}.spm.tools.align.estwrite.roptions.mask   = 0;       % Mask out voxels where at least one image is out-of-bounds
matlabbatch{1}.spm.tools.align.estwrite.roptions.prefix = 'r';     % Prefix of the output resliced image
matlabbatch{1}.spm.tools.align.estwrite.roptions.dir    = '';      % Output folder (reslice). Empty '' = same as input

License

GNU-GPL 3