Step 2: Realignment (motion correction)

Realignment registers all images from a given subject. We use it to generate parameter files which can later be used to correct for head motion. Performing this step will also create a mean image from your set of functional images, which we can later use to match functional to structural images.

Realignment produces text files with the estimated realignment (or motion) parameters for each session. These are the rp_ax001.txt stored in each session’s directory. They contain 6 columns and each row corresponds to a x-file. The columns are the estimated translations in mm (”right”, “forward”, “up”) and the estimated rotations in rad (”pitch”, “roll”, “yaw”) that are needed to shift each V-file. These text files can be used later at the statistics stages, to enter the estimated motion parameters as user-specified regressors in the design matrix.

The realignment stage will also produce a .ps file, which is automatically displayed when realignment finishes, and which contains plots of the above estimates. It is important to examine these plots carefully after realignment finished. The first of the two plots, titled “translation”, contains the estimated motion in the x-direction (left-to-right, blue line), y-direction (front-back, green line), and z-direction (bottom-top, red line). SPM successfully corrects for small amount of motion 1-3 mm in any direction, but if a subject moved more than 6 mm in any one direction in the course of the entire experiment, motion artifacts may be expected in the results later on. The “rotation” plot is a little less intuitive, but motion of more than 5 degrees in any direction can be considered excessive.

Motion artifacts are likely to appear as activations outlining the edges of the brain. If excessive motion is present, and such artifacts are observed, the motion parameters may be entered as regressors of no interest in the regression equation. However, on certain occasion, even this step is unable to correct for motion, or it sometimes causes all activations to “disappear” (in this case, correlation between motion and task may be present). As a last resort, it may be necessary to exclude subjects with excessive motion or motion correlated with the experimental paradigm.

Click on the “ REALIGN” button. Select Realign.

Select Realign: estimate and reslice.

Selecting estimate only will cause all files to be realigned by creating *.mat files that will contain realignment transformations that need to be applied to the corresponding ax-files. No new files will be produced (i.e., the the image files will not be resliced).

Selecting reslice only will cause new rax*.img files to be produced. The ax*.img you import will be transformed according to their corresponding ax*.mat files (given that they exist) and the resulting images will be written out as rax*.img files with the usual corresponding rax*.hdr files. No rax*.mat files will be created.

Selecting estimate & reslice will both realign the selected files, and will produce new files

NB: Realignment works in two stages. First, the first files (*x001.img) from each session are realigned to the the first file of the first session. Second, within each session, the second, third, etc… (2..n) images are realigned to the first image. As a consequence, after realignment, all files are realigned to the first files from the first session. Because of this algorithm, motion correction works both between and within session. This means that it is important to select the files for the different sessions separately (i.e., SPM needs to know which file belongs to which session), as indicated below.

Typically, the images will be resliced at the next stage (normalization). Again, the fewer time you reslice the images, the more resolution is preserved. In this case, we chose “Coregister and Reslice”, however, in order to be able to create a new mean image.

under Data, create as many new sessions as you have scanning sessions.

For each session, specify time-corrected images corresponding to that session. These should be prefixed by ‘aV’. For instance, for the first session, select all functional images from session 1 that are prefixed by aV.

Usually, we keep leave the estimation options at their default value.

For the Reslice Options:

Under Reslice Images select Mean image only.

Selecting all images (1..n) will produce as many r*x*.img files as *x*.img files are selected. If normalization will not be performed, the images should probably be resliced as this stage.

Selecting images 2..n will produce new r*x*.img corresponding to all *x*.img imported, except the first *x*.img from the first session. Since every files in the sequence is being aligned to this first file, strictly speaking, it does not need to be transformed or resliced. For practical reasons, however, it is better if all files have the same prefix (which is not going to be the case if this option is selected).

Selecting all images + mean image will produce as many r*x*.img files as *x*.img files are selected, and in addition, a new mean_x001.img files will be produced in the first session’s directory. The latter is the mean image of all files selected for realignment and it can be used with subsequent coregistration and normalization steps.

Selecting mean image only will produce only one new mean_x001.img file, without reslicing the selected *x*.img files. If normalization will be performed next, there is no need to reslice the ax*img files at this stage. The transformation parameters aved in the ax*.mat file to each image will be applied to the images during the normalization stages. This way, the images will be reslice only once. In general, unnecessary reslicing should be avoided, because every time it is performed, the images lose some spatial resolution.