Compute slice timings for multiband accelerated sequences

multiband_slice_times.py

def multiband_slice_times_oddshots(TR,slices,multiband,incrementslice):
    """ 
    Computes the slice timings of a multiband sequence.
    I think this is mostly right but I am not an expert, so please use with caution.
    And let me know if you spot errors. Thanks.
    
    This is based on the description here: 
    https://wiki.humanconnectome.org/download/attachments/40534057/CMRR_MB_Slice_Order.pdf?version=3&modificationDate=1386977824455&api=v2

    Arguments
    TR            : the TR duration (in ms)
    slices        : number of slices (e.g. 78)
    multiband     : multiband acceleration factor (e.g. 6 for 6x acceleration)
    incrementslice: increment slice parameter (default for Siemens is 2, check this in your sequence parameters)
    
    Returns
    a dict where the keys are the slice numbers (starting at 0) and the values are the acquisition times (in ms)
    """

    numberofshots = int(slices/multiband)
    
    assert numberofshots*multiband == slices   # number of slices must be a multiple of multiband factor
    assert incrementslice==2 # if not, might still work but use at your own risk
    assert numberofshots%2==1 # this function only works for an odd number of shots as far as I know
    
    slicetimes = {}
    overall_shot_n = 0
    for incrsl in range(incrementslice):

        # Compute how many shots we can fit in (starting from this incrementslice)
        for shot in range(int(ceil((numberofshots-incrsl)/incrementslice))):
            # The starting point for this shot
            startshot = incrsl + shot*incrementslice

            # Slices excited in this shot
            shotslices = startshot + arange(multiband)*numberofshots
            print(shotslices)

            for s in shotslices: slicetimes[s]=(float(TR)/numberofshots)*overall_shot_n
            overall_shot_n += 1

    return slicetimes

I had trouble extracting the slice timings from DICOM files acquired with a multiband accelerated fMRI sequence, so I wrote this little function that computes what the slice timings should be if I understand this correctly. Please use with caution.