Re: trends in brain imaging

From: Eugene.Leitl@lrz.uni-muenchen.de
Date: Sat Jan 20 2001 - 05:31:28 MST


Andy Toth wrote:

> + \No, of course. What a strange thing to say.
>
> of course there is. what a strange thing to say.

Nyah-nyah. You're wroooong!

C'mon, what's the best current resoltion of MRI on
human head sized objects, about 250 um (0.25 mm) in
plane at 3 Tesla? I suggest you read

http://www.lems.brown.edu/vision/courses/medical-imaging/projects/Cochlea/final_project.html

to a nice intro of MRI state of the art and limitations.
Similiar limitations have xRay microscopy, and other noninvasive
methods. Invasive methods have no such limitations, of course.
CryoAFM can reach atomic resolution, but this has not yet been
tried on large biological specimens.

>> can you fit the beginning of a straight log plot to a straight log
> plot? brain-imaging is born.

I have no idea what you mean by this sentence. Brain imaging which
is worth the name (I could mention Herr Roentgen here) exists
about as long as we have intergrated circuits. Assuming early
MRI had cm resolution, progress to 0.25 mm does over the course
of few decades does not strike me as anything straight log plot
worthy. Unless you mean the number of voxels in the data set,
that would go up as n^3 with increased resolution. Of course,
we don't have the feature resolution going down linearly with
passage of time. High Tesla magnets and field homogeneity are
not an easy technology to scale. Plus, eventually the magnetic
field and the deposited RF energy will start harming the patient
(it already does, you have to program the pulse sequence properly,
and pushing a patient into a high Tesla magnet too rapidly will
cause convulsions due to induced currents).

> this question struck me. as though humanity will irrevocably begin to
> reach inside of itself (per the dictation of the recontextualization
> of moore's machine law).



This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:05:06 MST