tag:blogger.com,1999:blog-4402160631955197288.post1812688780944728620..comments2024-03-12T19:57:17.818-07:00Comments on practiCal fMRI: the nuts & bolts: Using GRAPPA for fMRI in the presence of subject motionpractiCal fMRIhttp://www.blogger.com/profile/07387300671699742416noreply@blogger.comBlogger7125tag:blogger.com,1999:blog-4402160631955197288.post-77469047503809556382011-08-24T10:16:16.447-07:002011-08-24T10:16:16.447-07:00....but if you can, once you've picked one of .......but if you can, once you've picked one of the two parameter sets you posted, try a couple of rs-fMRI runs on a typical patient (if you can) or on a volunteer who isn't an fMRI ringer and see if TSNR varies appreciably for flips of 50, 70 and 90 degrees. In your particular case, with higher than normal motion to be expected, it could make a bigger difference than I just suggested. <br /><br />Final word of warning: don't read too much into a small number of pilot runs. The differences should be relatively small. If they aren't, be suspicious and re-run!practiCal fMRIhttps://www.blogger.com/profile/07387300671699742416noreply@blogger.comtag:blogger.com,1999:blog-4402160631955197288.post-2906785007165570072011-08-24T10:09:48.150-07:002011-08-24T10:09:48.150-07:00Hi MCube, Both sets of parameters look reasonable....Hi MCube, Both sets of parameters look reasonable. There will of course be slightly more distortion in the longer echo spacing, but it's not a vast difference. Scanner stability could be slightly worse at the shorter echo spacing, offsetting the benefit. But I would be surprised if there was a large measurable difference either way, so go with the one which gives you the best TSNR, or the one that gives you the coverage you want.<br /><br />FA is still something that is under investigation for rs-fMRI. But at a TR of 3 sec the 90 deg flip shouldn't be a disaster, even if someone eventually shows that 50 or 60 deg is slightly better because of lower physiologic noise. I honestly wouldn't worry about it. However, I would very definitely be trying to measure respiration and heart rate during the acquisitions!! <br /><br />Cheers!practiCal fMRIhttps://www.blogger.com/profile/07387300671699742416noreply@blogger.comtag:blogger.com,1999:blog-4402160631955197288.post-4096606309212198932011-08-19T09:21:57.228-07:002011-08-19T09:21:57.228-07:00...oops, correction: we run VB17 I believe...oops, correction: we run VB17 I believeMCubehttps://www.blogger.com/profile/12390154180429458186noreply@blogger.comtag:blogger.com,1999:blog-4402160631955197288.post-73284817768280707412011-08-19T09:21:12.347-07:002011-08-19T09:21:12.347-07:00damn, I just got out of the MRI!
It would have bee...damn, I just got out of the MRI!<br />It would have been an excellent idea. Not sure if I will get to do another pilot that simply, but I see what you mean, a very good idea. I'll see if I can give it a shot.<br /><br />So this morning I tried on a TRIO the sequence you suggest in your posts (except for FoV and no gap):<br />Siemens 3 T Trio/TIM running VB15, 12-channel HEAD MATRIX coil<br />ep2d_bold pulse sequence<br />TR=3000 ms<br />TE=25 ms<br />FA=90<br />slice thickness = 2.5 mm<br />gap = 0.25 mm<br />52 interleaved slices<br />matrix=64x64<br />FOV=192x192mm<br />bandwidth=2056 Hz/pixel<br />echo spacing=0.55 ms<br /><br />and this rsfMRI sequence the siemens people loaded on there (notice the real high bandwidth)<br /><br />TR=3000<br />TE=30<br />FA=90<br />slice thickness = 3.00<br />49 interleaved slices<br />matrix=64x64<br />FOV=192x192 mm<br />bandwidth=2605 Hz/pixel<br />echo spacing=0.47 ms.<br /><br />I'll report on TSNR..<br /><br />As a sidepoint, any chance some day of a bit of discussion on how to actually pick the "best" FA? I mean, we all know it's the ernst equation, but then again that doesn't quite maximize the tissue contrast of interest, does it?<br /><br /> Thanks for the referencesMCubehttps://www.blogger.com/profile/12390154180429458186noreply@blogger.comtag:blogger.com,1999:blog-4402160631955197288.post-18812631697739869092011-08-19T01:57:00.756-07:002011-08-19T01:57:00.756-07:00Afterthought @MCube:
Dan Handwerker had a poster ...Afterthought @MCube:<br /><br />Dan Handwerker had a poster at HBM (or ISMRM, I forget) showing that flip angle can be reduced to net a benefit of TSNR in resting state fMRI, much as the Bandettini Lab had already demonstrated for task-based fMRI in:<br /><br />Physiological noise effects on the flip angle selection in BOLD fMRI.<br />Gonzalez-Castillo J, Roopchansingh V, Bandettini PA, Bodurka J.<br />Neuroimage. 2011 Feb 14;54(4):2764-78<br /><br />So if I were you I would do a couple of quick pilot acquisitions with 40 vs 90 degree flip and measure the TSNR in each. If the theory holds that the denominator is reduced via lower motion sensitivity (through overall lower image signal intensity) with the lower FA then in motion-vulnerable populations the benefit should be considerable.practiCal fMRIhttps://www.blogger.com/profile/07387300671699742416noreply@blogger.comtag:blogger.com,1999:blog-4402160631955197288.post-13277168832354331972011-08-19T01:33:08.400-07:002011-08-19T01:33:08.400-07:00You're welcome!
If you didn't already se...You're welcome! <br /><br />If you didn't already see it, the recent study from Randy Buckner's lab on gender effects in rsfMRI is especially pertinent:<br /><br />http://practicalfmri.blogspot.com/2011/08/resting-state-fmri-motion-confound-in.html<br /><br />No GRAPPA in that study, one can only wonder whether there would have been further differences, or perhaps less difference because of higher variance? That's rather the point: William of Ockam in da house!practiCal fMRIhttps://www.blogger.com/profile/07387300671699742416noreply@blogger.comtag:blogger.com,1999:blog-4402160631955197288.post-25671512282197592232011-08-18T19:42:47.356-07:002011-08-18T19:42:47.356-07:00Ben, thank you, your posts are such an amazing res...Ben, thank you, your posts are such an amazing resource. I am going to double up rsfMRI acquisitions using a Trio (you might remember from previous posts I mostly use an allegra..) and when our MR guru mentioned the use of R=2/GRAPPA I wondered whether the fact that my patients have a spontaneous tendency to "jitter" during acquisition might have been a problem. Luckily "Ben has the answer" :)<br /><br /> cheers<br /><br /> MCubeAnonymousnoreply@blogger.com