View source for FBIRN:October 2005 Calibration Discussion

Overall goals: Have BH and ASL calibration protocols ready to test on multi-site traveling subjects by this time next year.

'''Breathhold Development:''' 1) BH plan steps:

* a. '''Machinery and scans'''<nowiki>: Define a lowcost behavioral measure device, similar to Gary’s (by January) </nowiki><br />
** i. From the later ASL discussion: To be maximally useful both in the BH and the ASL scans, this device should continually measure the scanner pulses, the output of the breathhold belt, and a pulse-oximetry output. It will run through the site’s filter panel. <br />
* b. Poll what sites have what (several sites already have some capacity but it may not be adequate for breath*holding*. A/C filtering means that a constant pressure drops to zero.)
** i. Sites with other measures can do comparisons on a few subjects to make sure it’s working right.
* c. '''Roll this device out''' to: MGH, UCI, and test it on a few subjects. (before March)
** i. A traveling technician might be needed to set it up properly.
** ii. It needs to work at sites that already have some experience (e.g., MGH, NM, Duke, MN), and sites that don’t (Yale, UCI, UCLA, etc.)
* d. Analyses needed: We don’t need to collect new data per se; MIND '''data from MGH''' has 23 Sz and equivalent controls doing physiologically-measured BH and the Phase II SIRP is available.
** i. Randy’s already noted that Sz subjects do the breathing, but are not on time with the paradigm; we need the behavioral measures to know what they were actually doing, when, to get accurate BOLD assessments.
** ii. Gary, Greg B., and Randy to discuss this and how to test the calibration measure within this dataset. (Noting that this is not multi-site, though.) (by January)
* e. '''Other questions'''
** i. Resting data can be very useful in measuring scanner SNR. Lee is doing some work on the Phase I resting data and will keep us informed on that.
** ii. Certain specific subject characteristics are known to affect BOLD reactivity (e.g., caffeine, sleep levels, and smoking). Lee, Gary and Randy are already discussing smoking issues and how that affects BH as a calibration measure. Report on this in March.
* f. The protocol (16s on/off, etc.) is already pretty set. The question is how well the BH measures, including the behavioral measures, allow a cognitive task to be calibrated to reduce intersite and site x subject variance.
** i. Which cognitive task: To start with it could be a Phase II task; as time progresses over the next year we’d test it with the latest cognitive task as developed by the Cognitive Working Group.

'''ASL Development:''' 2) ASL development steps:

* a. '''UCSD, MN, MGH''', (and possibly Stanford) as the key sites. Duke should be included if possible for their GE ASL development.

* b. The idea is to measure baseline CBF, and *functional* CBF (e.g., collected during a SM or BH task). ASL however is much slower than the usual EPI.

* c. '''First question: Which ASL sequence to use'''?
** i. GE is not a problem.
** ii. Siemens is more of an issue. ASL is a patent (held by John Detry). There is a Siemens “Works in Progress” sequence which might work.
** iii. PASL vs CASL needs to be discussed and evaluated.
** iv. For ASL to work best, we need physiological monitoring (see BH discussion above).
** v. Whole brain ASL is possible but not common. More likely we’ll have 5-7 slices through the middle of the brain. The thickness can be 4 mm/1 mm gap as the functional scans are; we should align the slices so they are the same as a subset of the slices collected during the fMRI scans and cover the functional regions we are interested in. <br />
* d. '''Timeline:'''
** i. Collect all the possible sequences (a “laundry list”) for GE and Siemens.
** ii. Working with both GE and Siemens sites, implement the preferred sequences to determine comparability (spirals vs linear, e.g.) and slice parameters, etc.
** iii. ASAP: Identify whether R&D agreements are needed for any sites (particularly Siemens), and if so, who and how fast we can get one.
** iv. Collect ASL during rest and during BH task, using the same slice locations as during the regular fMRI cognitive task.
*** 1. Gary and Tom to discuss direct comparisons of the BH and ASL task.
** v. By March: Have the recommended GE and Siemens sequences that give the best images possible, tested at the key sites with those platforms.
** vi. After March:
*** 1. Test on Sz vs controls.
*** 2. Roll out to traveling subjects sites (MGH, BWH, Yale, Duke; UCSD, UCI, UCLA, Stanford) to make sure they all can run the sequences, etc. This is expected to take a while to work out the kinks at the new sites.
*** 3. Determine the best use of the ASL measures for calibration (Stats group)—as scalar, or through linear modeling? Baseline only or the functional ASL? Etc.<br />

'''Next Calibration call'''<nowiki>: Oct 28 at 11 am PDT. Meetings will resume every two weeks on Fridays. </nowiki>

----