I would try option c first
c) don't do any slicetiming and see if the hrf derivative does the job
Den 13/01/2009 kl. 19.15 skrev Hamann, Stephan:
> Yes, thanks, I had wondered about whether including the temporal
> derivative would mop up enough of the small differences in timing
> enough to effectively make them negligible. I hope so -- it would be
> the easiest option to implement.
> From the other current thread discussing slice-specific design
> matrices, that might be an option in the future but perhaps it isn’t
> necessary for this situation.
> So, right now the two solutions we’re considering to deal with
> different slice durations are:
> (a) Run slice-timing “as-is”, ignoring the additional 20-30 msec per
> slice it takes to acquire each of the dozen or so inferior slices
> with the added-on asymmetric spin echo (ascending interleaved Trio
> with 38 slices with parallel imaging on) and then count on the extra
> basis function (temporal derivative) to mop up the residual error
> (along the lines you suggested).
> I’m tending towards this option. There is the minor remaining issue
> of which slice duration value to use in slice-timing (e.g., minimum,
> maximum, or average). The longer-duration slices are distributed
> across the TR (acquired at the start and middle of the TR) so
> perhaps using the average slice duration would help minimize the
> timing mismatch.
> (b) Modify the slice-timing routine to allow slice-specific
> shifting, accessing a user-defined array of slices and slice
> durations, and still add on the temporal derivative. If the temporal
> derivative is effective enough this would be more accurate but
> perhaps not worth the extra effort. Also, I’d imagine that a non-
> expert tinkering with the slice-timing routine could have some
> unforeseen consequences...
> On 1/8/09 7:21 AM, "Torben Ellegaard Lund" <[log in to unmask]>
> Dear Stephan
> This is an interesting problem, which I have been speculating about
> for some time now. I guess the slice timing procedure could be
> modified, but as we discussed earlier today slice specific
> designmatrices would be a better solution. As a workaround you could
> include a the hrf time derivative, this can account for small
> differences in arrival of the hrf. Note that while the individual
> slices take a different amount of time to acquire, the repetition time
> is the same for all slices.
> Hope this helps
> This e-mail message (including any attachments) is for the sole use of
> the intended recipient(s) and may contain confidential and privileged
> information. If the reader of this message is not the intended
> recipient, you are hereby notified that any dissemination,
> or copying of this message (including any attachments) is strictly
> If you have received this message in error, please contact
> the sender by reply e-mail message and destroy all copies of the
> original message (including attachments).