As I said, this isn’t trivial and would require some manual manipulation of the data using Matlab and SPM commands/functions– there is no way to do this with mouse clicks using a UI (at least not that I’m aware). Again, the best and safest thing to do is to try to re-export the scan from the scanner. If that’s not possible or doesn’t solve the problem. The steps I would take are:

1)      load the raw data into a 3D-matrix with 

V=spm_vol(spm_select()); IMG=spm_read_vols(V);

Your variable IMG will now be a 3D matrix with the dimensions of your scan (e.g. 240x240x180). Individual slices are the third dimension in this matrix.

2)      Identify which slice is wrong

imagesc(IMG(:,:,z));   ß here you need to vary the value of z until you find the problematic slice

3)      save this slice into a new 2D matrix:

SLICE=IMG(:,:,z);  

4)      shift the data in this slice until it seems to fit with the two neighbouring slices:

SLICEshifted=circshift(SLICE,[xOffset yOffset]);

You need to determine the values of xOffset and yOffset yourself, either by visual comparison of the neighbouring slices, or by identifying a feature in the image that runs mostly laterally through all slices. This will likely involve a good deal of trial-and-error.

5)      Once the slice is shifted into the proper position, it needs to be re-inserted into the 3D matrix and then saved as nifti image:

IMG(:,:,z)=SLICEshifted;
spm_write_vol(V,IMG);

Optionally, before running spm_write_vol, you could change the file name it will be saved under, by changing the content of V.fname:

V.fname=’manipulated_image.nii’;

A few other comments: the above procedure assumes that the problematic slice is only shifted with respect to the other slices, not rotated. If it looks as if it is also rotated, or otherwise deformed, you need different procedures, and are likely out of luck. Also, depending on what you want to do with the file, I would always treat this file with some suspicion, even after it’s been “fixed”. Estimating normalization parameters is probably fine, but I’m not sure if would use this file for VBM type analyses.

Best wishes,

Andreas