X-Message-Number: 5227 From: (Brian Wowk) Newsgroups: sci.cryonics Subject: Re: High definition MRI Date: 20 Nov 1995 19:34:31 GMT Message-ID: <> References: <48pjlo$> In <48pjlo$> yvan Bozzonetti <> writes: >There is a simple way to upgrade the signal to noise level: Take data for a >longer time. With a N-times longer duration you get a square root better S/noise >ratio. If you go from 100 seconds to 10^8 s, you have a 1000 times improvement. Of course. I was refuting the original poster's assertion that computer power increase (not scan time increase) can increase SNR. >There is an even better solution: Use a high polarization gas, for example >hyperpolarized He3 or Xe and get a millions times stronger signal. I attended a plenary lecture on this subject in Nice, France this past summer. It is indeed very interesting work. However it's not going to revolutionize SNR in MRI. Yes, laser excitation can be used to prepare He3 or Xe nuclei to give a million times more signal than H1 nuclei (protons). However there are 1000 times more protons in blood than dissolved He3 or Xe, so there is only a ~1000 times signal increase. However this increase is only available *for one excitation*. Hyperpolarized nuclei do not undergo T1 recovery like protons. One 90 degree excitation will destroy the hyperpolarization forever, ending the imaging experiment. This means that hyperpolarized gas imaging must use very small flip angles, which further reduces the SNR to about the same as ordinary protons in tissue. Hyperpolarized gas allows lungs to be imaged in a new way, but it's not going to substantially increase resolution of brain MRI. ---Brian Wowk Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=5227