first you may want to check out this:
In your case, you can still shift all your slices to the middle of your TR
of 5 secs. You would do this by using slicetimer's --tcustom and
acknowledging that your final slice was acquired at 2 secs (or 1.973 for
the matter) which is - in fractions of TR = 5secs - indeed 2/5.
Am 17.05.12 20:55 schrieb "Sean F Walsh" unter <[log in to unmask]>:
>Dear Professor Smith and Dr. Bartsch,
>Thanks a lot for the quick response. Would one of you mind dealing with
>one remaining doubt of mine?
>If I understand correctly without slice-timing correction FEAT will
>presume that all slices were taken after 2.5 seconds of each volume, when
>in fact each of the 37 slices were taken in the first 2 seconds. As
>slice-timing correction shifts "each time-series by an appropriate
>fraction of a TR relative to the middle of the TR period" I presume it
>will shift every time series by fractions of 5 seconds, which is not
>appropriate here. Is it necessary to input a slice timings file detailing
>the appropriate shift for each slice? I believe a value of 0.5 is no
>shift (and would thus represent 2.5 seconds in here).
>The slices were taken in interleaving sequence (0,2.....36, 1,3,.....35).
>Must I thus calculate each slice's timing by the following formulae?
>slice X timing for even slices = (((X/2)+1)/37)*(2/5)
>Where ((X/2)+1) will give each even slice its correct place in the order
>of acquisition from 1-37 (not sure if this should be 0-36 or 0.5 - 36.5
>instead). I then divide by 37 in order to give each slice its time in
>terms of 1 TR, with 0.5 being no shift and representing 2.5 seconds in
>this case. I then multiply by (2/5) to correct for the fact that all
>slices were taken in the first 2 seconds of the 5 second TR
>Similarly the formula for odd slices would be
>where (((x+1)/2)+19) assigns each odd slice its correct place in the
>order of acquisition, and once again dividing by 37 normalises this
>placing to give timings in terms of 1 TR, and multiplying by (2/5) takes
>into account the fact that all slices were taken in 2, not 5 seconds.
>Otherwise tips on how to modify the model times would be equally
>Many thanks in advance,