On Mon, 14 Mar 2005 21:16:54 -0600, Tim Williams wrote:
"Dave Hinz" wrote in message
...
Most MRI systems are 1.5T. Some of the newest ones are 4T, but those
are few and far between.

Really? Then were do I keep reading 10-20T from?

Dunno. Maybe for in-vitro use, for NMR spectroscopy use, but not for human
scanning - I think 4T or 5T is about the upper limit, even for research
sites. Something about signals being induced into nerves at some
point, from what I recall, or at least a concern of that. It's also
a bitch to site even a 4T magnet, the fringe fields get pretty expensive
to manage.

When I worked in MR Engineering, we, well, were playing around with
magnets, oddly enough. The permanent magnets in a Hawk hard drive
(1 gigabyte full height - these were old) came in above the measurement
capability of our gauss meter, which went up to above 4T. So, at
the surface, it's more than a 4T permanent magnet, in those drives.

Huh. Anyone know offhand what the top field strength of the toughest
magnetic materials is these days? I always thought magnets topped out
around iron's saturation levels (1-2T)...

These weren't iron, they were chrome-plated rare earth magnets of some
sort. Strong enough that they held strong through my hand. Hint:
don't let 'em go around and hit each other, because blood blisters will
result (based on a sample of one).

It's strong, but the field isn't anywhere near even (parts per billion
is needed), or large enough.

As I recall, MRI's use a Helmholtz coil (I think), basically two short
solenoids seperated by a distance.

Yes, there's an inner coil and a counter-wound outer coil.

The space between them has a relatively
constant field. Not perfectly constant, so they must have to use extra
coils and tweaks either way, no?

Yes, the GE magnets have 18 "shim coils", first, second, and third-order
corrections. The simplest is just X/Y/Z; get the overall gradient
from say front to back the same. Second-order does opposing corners,
third order starts getting strange. Some delightful math to calculate
how to get the shims in order, as they all interact heavily.

Um, no. RF Heating is a real and measurable phenomenon in MRI imaging.
The "SAR" (specific absorption ratio) calculates just how much RF
you can pump into the patient, based on their mass.

I'll admit you've got me on the RF part... All I know is the field aligns
hydrogen atoms

Well, the protons in them, but yes,

(and maybe others),

There are a few others but the frequencies get strange. I think all the
diagnostic work is using hydrogen.

then RF is applied and re-radiated...how
the hell they decipher the 10^20llions of hydrogen atoms spatially is beyond
me.

Well, there are gradient aplifiers which make the RF pulse only resonate
one location cleanly (this is wrong but close enough to make the point).
The RF hits the protons there, they flip (usually to 90 degrees to
get the most signal). As they re-align to the magnetic field, they disturb
their electrons, which give off an echo of the RF pulse.

The echo is detected with the same RF coil (goes from transmit to receive
very quickly), which gets fed into a 2-dimensional Fast Fourier Transform,
breaks it into frequencies which tells you a number of things, and then
does it again with more lines of data (each of which also came from a FFT),
and FFT's _that_ into the image. Sort of.

Also, RF coupling to the body coil can increase this;

Basically induction heating a person? Or more likely capacitively (B and E
are related anyway, so who cares) given the lossy nature of flesh.

Right. It's usually localized to, say, the arms where they're
closest to the coil. A person inside an antenna acts as an RF "slug"
(heh) and both lowers the frequency of, and widens the peak of, how the
antenna is tuned.

It's neat seeing how these scanners are made. There are guys on
the manufacturing floor using technology that hasn't changed in
generations, or even centuries. Mix that with superconductors chilled
with liquid helium, and there's a ton of interesting stuff going on.