FRESH

Hacker News

Sharper MRI scans may be on horizon thanks to new physics-based model

157 points by hhs

by uecker

1 subcomments

While this work is great, this will not directly lead to "sharper MRI scans". This is about better modelling of NMR signals, which may eventually lead to better MRI, but it is still pretty far away from imaging. If you want how we use simpler signal models in physics-based reconstruction to improve MR images, you can look at our paper: https://doi.org/10.1098/rsta.2020.0196

by iandanforth

8 subcomments

FYI if you're getting a contrast MRI in the near future, avoid vitamin c. https://hscnews.unm.edu/news/unm-scientists-discover-how-nan...

by Animats

2 subcomments

Here's the paper.[1] No paywall this way; U.S. Government funded research. The paper claims an associated Github repository but there is no obvious link. There's no imagery in the paper, just the development of the math. So this may or may not help much.
[1] https://www.osti.gov/pages/servlets/purl/3001752

by Insanity

2 subcomments

After reading the article I get what it’s saying.. but isn’t any MRI _technically_ physics-based.

by ForOldHack

0 subcomment

"However Gd is also retained in the brain, bone, skin, and other tissues in patients with normal renal function, and the presence of Gd can persist months to years after the last administration of a GBCA."
This is therapy. May persist... I want to throw this here, because there is a lot of discussion here by not only a few patients but a few very knowledgeable researchers which is surprising for such an extremely niche field, and the discussion is amazing.

by ForOldHack

1 subcomments

I suspected this was blahtering AI, and this confirms it: "By solving this equation, they were able to capture the full spectrum of molecular motion and relaxation." The equation is solved buddy, and has been since Plank solved it. It's the discreet calculation of its values give is a digital image that can be contrast enhanced by differing molecules, the physics has not changed, and the equation has not changed. <Downvote>