Reslice Module

The function reslice_image is used to deform images aligned to the individual anatomical image into SUIT space. We assume you have produced in the previous steps an isolation mask (sub-ex_T1w_cerebellum_dseg.nii.gz) and a deformation file (sub-ex_T1w_to-SUIT_mode-image_xfm.nii.gz).

[1]:

# Import necessary packages
from nilearn import plotting
import SUITPy as suit
import nibabel as nib

Reslice image without mask

The reslice function takes at the minimum 2 input parameters: The list of source images (the images that you want to reslice) and the deformation map.

[5]:

output_img = suit.reslice_image(source_image = 'sub-ex_T1w.nii.gz',
        deformation = 'sub-ex_T1w_to-SUIT_mode-image_xfm.nii.gz')

[7]:

plotting.plot_anat(output_img,cut_coords=(-20,-60,-35))

[7]:

<nilearn.plotting.displays._slicers.OrthoSlicer at 0x3104a18d0>

../_images/tutorials_5.reslice_example_4_1.png

To save the resliced image, use:

nib.save(output_img, "img1.nii")

Reslice image with mask

For cerebellar data analysis, we strongly recommend to mask functional data when reslicing. This prevents un-wanted spill-over of neocortical signal into the cerebellum during subsequent smoothing.

[9]:

output_img = suit.reslice_image(source_image = 'sub-ex_T1w.nii.gz',
        deformation = 'sub-ex_T1w_to-SUIT_mode-image_xfm.nii.gz',
        mask='sub-ex_T1w_cerebellum_dseg.nii.gz')

[10]:

plotting.plot_anat(output_img)

[10]:

<nilearn.plotting.displays._slicers.OrthoSlicer at 0x36a5a00d0>

../_images/tutorials_5.reslice_example_8_1.png

Specify voxel size

By default, the new image will be resliced into the same resolution and bounding box as template (1mm resolution for SUIT). For functional data we often would like a different resolution. For this, you can simply specify the voxel size.

[11]:

output_img = suit.reslice_image(source_image = 'sub-ex_T1w.nii.gz',
        deformation = 'sub-ex_T1w_to-SUIT_mode-image_xfm.nii.gz',
        mask='sub-ex_T1w_cerebellum_dseg.nii.gz',
        voxelsize=(2,2,2))

[12]:

plotting.plot_anat(output_img)

[12]:

<nilearn.plotting.displays._slicers.OrthoSlicer at 0x36b669690>

../_images/tutorials_5.reslice_example_11_1.png

[13]:

# Voxel size has been changed.
output_img.shape

[13]:

(71, 48, 44)

Specify output dimensions and affine transformation matrix

To fully specify the output dimensions of the data, you can also specify an affine transformation and a image dimensions

[14]:

output_img = suit.reslice_image(source_image = 'sub-ex_T1w.nii.gz',
        deformation = 'sub-ex_T1w_to-SUIT_mode-image_xfm.nii.gz',
        mask='sub-ex_T1w_cerebellum_dseg.nii.gz',
        imagedim = [100,50,30],
        affine = output_img.affine)

[15]:

plotting.plot_anat(output_img)
output_img.shape

[15]:

(100, 50, 30)

../_images/tutorials_5.reslice_example_15_1.png

Reslice atlas to individual space using the inverse deformation

If you want to bring a group template of atlas into the space of an individual, you need the inverse deformation file. To save that inverse deformation set write_inv_deformation to True.

[ ]:

# Set write_inv_deformation=True to save the inverse deformation file.
results = suit.normalize(source_file = 'sub-ex_T1w.nii.gz',
                 mask_file = 'sub-ex_T1w_cerebellum_dseg.nii.gz',
                 write_inv_deformation=True,
                 verbose = 0)

The inverse deformation is saved as sub-ex_T1w_from-SUIT_mode-image_xfm.nii.gz.

We can then reslice any atlas, such as the Buckner networks, from SUIT into the individual space.

Note: because the atlas is a label file, we need to use nearest-neighbour interpolation here

[ ]:

# reslice the Buckner atlas from SUIT space to individual space using the inverse deformation.
buckner_img = suit.reslice_image('Buckner_17Networks.nii',
                   deformation = results['inv_deformation'],
                   interp=0)

[ ]:

plotting.plot_img(buckner_img,display_mode='y',threshold=0.1, cut_coords=[-53.0], bg_img='sub-ex_T1w.nii.gz',colorbar=False,cmap='tab20')

<nilearn.plotting.displays._slicers.YSlicer at 0x3249da610>

../_images/tutorials_5.reslice_example_21_1.png

Cerebellar Atlas

Using deformation maps from SUIT(Matlab)

The reslice module can also be used to reslice images from individual that was normalized with the MATLAB version of the SUIT toolbox. We assume that you have run suit_normalize_dartel.m (see here for documentation), The reslice module requires the result of the normalization step as a non-linear deformation file. In Matlab you can produce this file using the following command:

suit_save_darteldef(<name>, 'wdir', <dir>)

Where <name> is the name of the original anatomical and <dir> the suit working directory (defaults to current directory). This function produces y_<name>_suitdef.nii as an output. This image can then be used instead of the *_xfm.nii.gz file in SUITPy to reslice anatomical and functional images into atlas space.