Slice Timing

From BrainImagingCenter
Revision as of 03:37, 5 August 2009 by Yoh (Talk | contribs) (cut/paste Yu-Chien's email on the choice of the slicing order)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Interleaved Slicing

Thanks Dr. Elise Temple for Original email to the mrusers mailing list:

Many of our protocols are set up to acquire our EPI slices in an interleaved fashion (for all the reasons you have learned about why interleaved is good).

It seems that on the Philips machine when one has "interleaved" selected as an option for the slice order parameter, the acquisition is not doing what most of us would consider a traditional interleaved acquisition.

To have a ascending, odd first interleaved acquisition (e.g., 1 3 5 ..., 2 4 6 ...) one needs to have "default" selected in the slice order parameter under geometry of your protocol

If you have "interleaved" selected (which as near as I can tell most people do) the acquisition is set to maximize the time between adjacent slices.

What it does is - take the total number of slices you are acquiring. Take the square root of the number of your sleices and then round up to the nearest interger. Then acquire slices that far apart starting at 1.

For example - in our case we had 44 slices - so the square root of 44 is 6.6 so we round up to 7. the acquisition under interleaved is as follows: 1 8 15 22 29 36 43 2 9 16 23 30 37 44 3 10 17 24 31 38 4 11 18 25 32 39 5 12 19 26 33 40 6 13 20 27 34 41 7 14 21 28 35 42

Of course this is important fopr all of us when we report how our EPI was acquired, but it is mostly important for anyone doing slice timing - especially for rapid event related studies.

If you want to do interleaved, odd first, ascending you need to have "default" selected for this parameter.

The choice of slicing order

Thanks Dr. Yu-Chien Wu for followup email to the mrusers mailing list:

The reason to interleave the slice order is to avoid the "cross-talk" between contiguous slices. The "cross-talk" artifact comes from the imperfect slice selection pulse (or RF wave form). And it is impossible to impose a perfect slice selection pulse due to limited echo time.

Therefore, if one prefer shortest echo time (with "RF pulse transition" set as "norma"), I would recommend to "interleave" your slices as far apart as possible. Also, put a gap between slices will decrease cross-talk.

The sharp slice selection pulse is possible with the parameter, "RF pulse transition", set as "sharp". However, this option comes with a penalty of longer minimum echo-time.

With all the suggestions, however, one should balance what best suits the studies.