Hi,

I am new to dcm2niix, so it might be a silly question. I found when I converted dicom into nifti, there was only one output nifti file. I wonder how to run realignment in SPM since I only got one nifti file. Btw, does scm2niix correct slice timing during conversion? I appreciate it much if someone could help me answer these questions!

best,
Shengjie

In SPM, you can select individual 3D volumes from a 4D dataset by setting the volume selector in the SPM image selection box (the box just below the filter). As noted in my tutorial, you can type "1:9999" into this box to see the first 9999 volumes of a file. The screenshot below is from the dataset included with the tutorial:
 http://people.cas.sc.edu/rorden/tutorial/index.html

Hi Chris,

Thanks for your explanation!

I have a following question about slice timing correction with multi-band EPIs. My EPI's TR is 1.62 sec and multiband factor is 3. After conversion, I checked generated JSON file, and found in slice timing tag there was an array with 57 time points. However, in 3d format of my EPI, it was 307 files but why only 57 time points were given? Then I tried another group of EPI with different numbers of 3d nii file, but got the same number (57) of time points in slice timing tag. I also noticed in this array, first 19 values were the same as second 19 ones and the last 19 ones, eg 1,2,...19,1,2...19,1,2...19.

Is 57 a default number of slice timing values?  if I wanna run slice timing correction in SPM12, is it correct to use this 57-value array, or I should use 19 values as slice timing factor, or I should skip slice timing correction due to this problem? I searched many online websites but still can't figure out why.

Thanks so much and apologize for my questions!

Best,
Shengjie

Multi-band allows you to acquire a 3D volume of the brain more rapidly by acquiring multiple distant 2D slices simultaneously. In your example, your 3D volume has 57 slices, acquired in ascending order with slice 1, 20 and 39 acquired simultaneously. This is revealed in the "SliceTiming" BIDS tag which generates one entry for each of your 57 slices, revealing the time of acquisition of that slice (in seconds) relative to the first slice in that volume. It sounds like you acquired 307 volumes with a TR of 1.62s. Therefore, if your first 3D volume was acquired at time = 0, the acquisition of the second volume began 1.62s later. Since the TR is constant (as changing the TR will change your image contrast, e.g. T1 and T2 signal), my software stores just one value in the BIDS file "RepetitionTime": 1.62,. Therefore, the SliceTiming Tag reveals the timing of individual 2D slices that compose a 3D volume, and the RepetitionTime Tag reveals the timing between successive 3D volumes that compose a 4D volume. You can use the SliceTiming tag to perform Slice Timing Correction, but with multiband the TRs tend to be very brief and therefore the temporal derivative should be able to soak up most/all the variability described by slice timing differences. I encourage you to look at the PowerPoint files from my course for more details.