3) Sept 13, 2012
Ted Trouard talked about DTI physics. He had borrowed some of his ppt slides, so I'm not able to post them here. I'll do my best to summarize (but take my understanding with a grain of salt).
Ted introduced the term "rotational mobility". The rotational mobility of the molecules is NOT the same thing as the spin of the protons.
However, rotational mobility of the water molecules, like translation of the molecules, affects the total dti signal.
Rotational mobility can be measured separately from translation.
b-weighting Diffusion imaging requires that we apply 2 gradients. We must carefully choose the gradient strength, the length of time each gradient is applied, and "big delta" (the length of time between the first gradient and second gradient). If we change delta, then we change the time molecules have to reach barriers and bounce off of them. Interpreting the effect of different deltas on dti images is difficult, so it is better to use gradient strength changes to implement different b-weighting. Ted also suggested that because the length of delta has multiple effects, it is really important for researchers to publish that time in their DTI specs. Apparently, people do not always do this, and they should.
How is TE related to big delta? Big delta is not exactly the same as TE, but it is definitely tied to it. Big delta is always a little less than TE, but because we are always keeping TE to a minimum, and the separation and duration of the gradients determine that minimum, the TE is somewhat close to big delta.
Why set b-weighting=1000? Ted showed us a graph of how signal in water vs tissue changes as b-value changes. If b-weighting is too high, the total signal drops off so that you cannot distinguish the signal from the noise. If b-weighting is too low, you can't see the diffusion translation effect very well. b=1000 turns out to be a comfortable balance that lets you see tissue differences and the direction of water molecule movement without losing the signal in the noise near the bottom of the graph.
How many B0 images? The recommendation from Jones 1999 is to collect 1 B0 image for every 8-10 diffusion directions. The point of having multiple B0 images is that you are going to compare B0 images to the B-weighted images for calculating various diffusion images. If you collect more B0 images, then you can average across them, and get a more accurate estimate of the "true" B0 image. This will make your diffusion calculations more accurate.
What if all the B0 images are at the beginning of the scan? Because the reason you acquire multiple B0 images is that you want to average and get an accurate estimation of B0, it does not matter when you gather them (there isn't necessarily much benefit to gathering them at different points in the scan.
Why Single Shot Spin Echo? Because DTI is so super sensitive to motion, it is important to gather each image in a single shot.
NEX (NEX has nothing to do with single shot): NEX=The number of excitations. If you use 2 NEX, then each of your final images is an average of 2 excitations (an average of 2 single shot spin echos).
DWI (Diffusion Weighted Imaging) vs DTI (Diffusion Tensor Imaging)
DWI is used in clinical scanning because it is fast and identifies acute ischemic stroke really effectively. DWI only requires 3 volumes, because you are not interested in the direction of water molecule movement...but only how far on average they've moved. Interestingly, we don't really understand why diffusion is reduced in an area of ischemic stroke. It may have to do with cell death (nothing is moving around in the dead cells any more)...but the mechanism is not understood.
DTI requires 6 directions (at least) to characterize the direction as well as amount of diffusion. This has to do with the number of axes you need to measure the orientation of a 3D ellipse in space (6).
Suggested Reading (Not cheap, but worth owning if you are serious about DTI): Introduction to Diffusion Tensor Imaging