Hi FSL experts,
In regards to using phase partial fourier, in a different post I was told that it does not affect the echo spacing time (use protocol ES), however I am not 100% sure this is true after reading this post. I have an EPI sequence which is collected WITHOUT acceleration (GRAPPA/iPAT), but DOES have a phase partial fourier factor of 6/8. My "reconstructed" matrix is 76x76, however when I look at the dicom header I see that only 57 lines are actually collected in the PE direction (76 * 6/8=57). When I use the equation in the post :
Echo Spacing = 1 / [(0019,1028) * (0051,100b component #1)]
>>> where
>>> 0019 1028: BandwidthPerPixelPhaseEncode -- BW in PE direction (Hz)
>>> and
>>> 0051 100b: AcquisitionMatrixText (gives size of *reconstructed* matrix)
ES=1/(28.604*76)=0.00046s which matches up with the echo spacing in the protocol sheet.
However since we are using partial fourier the time is in reality reduced so shouldn't the effective echo spacing time match the # of lines of k space as opposed to the "reconstructed" image. By this I mean ES=1/(28.604*57)? This would seem to be in line with how SPM's fieldmap accounts for partial fourier (they use total readout time so 0.00046s* 57= 0.02622 seconds actual readout time) whereas 0.00046s*76=0.03496 which would suggest a longer readout time than actual.
My thoughts are that we would need to account for partial fourier (6/8) into our effective echo spacing by using ES=ES*PF = 0.00046* 6/8 so that our total time is accurate. When looking at the fugue.cc code it seems that dwell time is used in calculations with fmap dimensions (76x76x42) and so if we do not account for the partial fourier in our dwell time we supply, then this may not be taken into account anywhere else. Does this seem accurate or am I missing something?
Thanks,
Ajay
|