Scanning Papers

Useful Papers - fMRI Scanning

Also check out DesignPapers, JitterPapers, FmriPhysicsPapers and PhysiologyPapers...

Primary:

Aguirre et. al (2002), "Experimental design and the relative sensitivity of BOLD and perfusion fMRI," NeuroImage 15, 488-500 PDF

Summary: Aguirre et. al experimentally examine the noise profile and power profile of perfusion imaging, demonstrating that perfusion should show increased power relative to BOLD when task-switching frequencies go below blocks of around 60 sec, as well as showing the relative greater across-subject noise for BOLD, and the lack of autocorrelation in perfusion noise.

Bottom line: Try using perfusion imaging for extremely long block-length experiments, or possibly for any experiment with a group analysis.

Glover & Law (2001), "Spiral in/out BOLD fMRI for increased SNR and reduced susceptibility artifacts," Magnetic Resonance in Medicine 46, 515-522 PDF

Summary: Gary describes his trademark spiral in-out sequence, a combination of spiral-in and spiral-out images weighted in a variety of ways, and shows that its use significantly increases SNR over traditional spiral methods and greatly reduces drop-out due to magnetic susceptibility.

Bottom line: If your institution supports it, spiral in-out beats traditional methods senseless for reducing drop-out due to susceptibility and should be used.

Supplementary:

Preston et. al (2004), "Comparison of spiral-in/out and spiral-out BOLD fMRI at 1.5T and 3T," NeuroImage 21, 291-301 PDF

Summary: Preston et. al compare a set of tasks known to activate susceptibility-heavy areas using spiral in-out imaging at both 1.5T and 3T, so show that where traditional spiral methods can actually increase dropout at higher field strengths, spiral in-out gets better activation volumes and less dropout at 3T than at 1.5T and confirms spiral in-out is better for a variety of real experimental tasks than traditional spiral.

Bottom line: No, really, we mean it. You should definitely use spiral in-out, even if you have a giant magnet.

Yang. et al (2000), "A CBF-based event-related brain activation paradigm: characterization of impulse-response function and comparison to BOLD," NeuroImage 12, 287-297 PDF

Summary: Yang et. al describe ways to substantially improve the effective TR of perfusion imaging (although some schemes don't work with all experimental designs), and use event-related paradigms to describe the shape of the impulse response function with perfusion, which is similar to that in BOLD, but precedes it in both rising and falling. The perfusion IRF shows similar nonlinearities to the BOLD IRF.

Bottom line: Perfusion can be effectively used in an event-related setting, and the hemodynamic response in perfusion precedes that in BOLD on both onset and offset.

Follow-Up:

Glover (1999), "On signal to noise ratio tradeoffs in fMRI," monograph. PDF

Summary: Gary discusses various tradeoffs about SNR, such as length of TR and number of interleaves, and walks through the consequences of changing those parameters both in mathematical models and in single-subject data.

Bottom line: Use the smallest TR possible; the decrease in SNR in individual image is more than offset by the increase in image number. Fewer interleaves are better, as is higher field strength.

Glover handout: Summarizes the above monograph, with a little bit extra about slice thickness. Handout

Philippe handout: A nice primer on k-space and perfusion, and some good page-long summaries on the primary articles. Handout

Constable & Spencer (2001), "Repetition time in echo planar functional MRI," Magnetic Resonance in Medicine 46, 748-755. PDF

Summary: Authors attempt to determine, from theoretical and empirical footings, the optimal, RT for a given experimental paradigm, and come to the conclusion that shorter TRs (1-1.5 sec) vs. longer (3-4 sec) offer greater statistical power, at the expense of some other tradeoffs.


CategoryPapers