[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/sci/ - Science & Math


View post   

File: 25 KB, 255x255, Magnetic-Resonance.jpg [View same] [iqdb] [saucenao] [google]
2455830 No.2455830 [Reply] [Original]

<span class="math">{\bf TL;DR}[/spoiler] Don't be a dick and just read the text.

Magnetic resonance imaging appreciation thread. I’ll be posting a bunch of different techniques that all exploit the principle of magnetic resonance, but first I’ll explain the basics a bit. (note that it’s considerably more complex than explained below, but this is 4chan after all, so hopefully everyone can understand the principles).

So basically the idea is that particles like protons have nuclear spin. When a very strong magnetic field (called the B0 field) is applied, the axis of spin can be aligned with- or a against the field. Because this is a quantum effect, there are no intermediate alignments. The balance between particles aligned with and aligned against is dependent on the strength of the magnetic field. The two types of alignment differ in their energy state, so with an increasing B0 field strength the net energy difference between the total aligned particles also increases.

It is possible to stimulate the nuclei using electromagnetic radiation with a frequency that is proportional to the strength of the magnetic field (the Lamor frequency, which is radiofrequency in magnetic resonance imaging). This adds energy to the particles in the lowest energy state, so that they flip their alignment temporarily. After a very short period of time they flip back, emitting electromagnetic radiation.

[continues]

>> No.2455833

[continued]

In order to get spatial information of where this electromagnetic radiation is coming from, three gradient magnetic fields are sequentially applied, so that space is divided up in blocks (voxels) with their own field strength. Remember that the excitation of the atomic nuclei is dependent on the strength of the magnetic field, so particles within each voxel are stimulated to a different extent, and consequentially emit electromagnetic radiation of a different frequency when the excitation decays. This makes it possible to measure the amount of specific molecules that have magnetic resonance properties (such as protons) within living tissue.

In the next few posts I’ll explain a bunch of different ways of how we can exploit this.

>> No.2455834
File: 15 KB, 403x308, brein1.jpg [View same] [iqdb] [saucenao] [google]
2455834

Because the decay of excitation is dependent upon the surroundings of the nuclei, different types of tissue show a different speed of decay. This makes it possible to image these tissue types. Depending on the pulse sequence and the contrast used (what these terms mean exactly are a bit too complicated to explain here), different tissue types have different grayscale intensities. The image here shows a T1 contrast (of my own brain), which typically shows white matter as white, gray matter as gray, and cerebrospinal fluid as black.

This makes it possible to measure tissue type specific volumes. A technique that’s used to index gray matter volume is called voxel based morphometry (VBM). It provides a measure of cortical volume.

>> No.2455835
File: 203 KB, 1021x1125, DTI-sagittal-fibers.jpg [View same] [iqdb] [saucenao] [google]
2455835

Because the molecules within the measured space obviously don’t sit still while they’re being measured (they diffuse), the decay signal changes with diffusion not precisely aligned with the gradient field (remember that decay properties are dependent on the magnetic field?). In a space of unrestricted diffusion (such as the ocean), motion is isotropic in all directions. However when diffusion space is restricted (such as in the fibers of white matter) diffusion is relatively anisotropic in one direction. That means signal decay is less strong when the gradient magnetic field is aligned with the fiber tract.

We can exploit this principle by sequentially applying a whole bunch of gradient fields with different directions, and measuring anisotropic diffusion. The data can be used to (probabilistically) reconstruct white matter tracts and their level of coherence (picture related). This technique is called diffusion tensor imaging.

>> No.2455837
File: 17 KB, 360x360, BOLD MRI.jpg [View same] [iqdb] [saucenao] [google]
2455837

Remember that the decay of excitation is dependent upon the surroundings. The decay can be caused by spin-spin interactions, but also by local inhomogeneities of the magnetic field. One element that causes such inhomogeneities in particular is iron. It just so happens, the complex which carries oxygen in our blood (hemoglobin) has four iron atoms in it. When oxygen is bound to it, these four iron atoms are bent inwards of the molecule. When oxygen is used (in neural tissue for instance) they are directed outwards, causing a distortion in the local magnetic field, and thus affect the decay properties of the surrounding protons.

We can exploit this principle by measuring not just spin-spin interaction decay, but also decay due to inhomogeneities of the magnetic field. From this we can reconstruct the relative proportion of oxyhemoglobin and deoxyhemoglobin. This can serve as a measure of neural activity, and thus be used to image in vivo brain function, which is why it is sometimes referred to as functional magnetic resonance imaging, or fMRI in short. Actually BOLD-MRI is a subset of fMRI, which will be explained in the next post.

>> No.2455840

MOAR ...

tnx OP ..

>> No.2455842

Protons are the most readily used types of particles that have magnetic resonance properties. However, there are many more types, and every type has its own specific emitted EM spectrum signature. Somewhat recently a number of molecules which have paramagnetic properties have been developed to bind to individual neurotransmitters. Injecting these substances can enables the imaging of specific neurotransmitter function (such as GABA, pic related), with a spatial and temporal resolution which is greater than that of other such techniques, such as positron emission tomography.

(sadly, I don’t have a picture available of this.)

In conclusion, MRI is an extremely versatile technique which enables us to learn a great deal about not just human brain function, but about the entire body (note that the techniques described above are only a subset of the ones available).

I think it deserves some appreciation.

>> No.2455850

FUCK YEAH NMR.
*brofist*
Although we are doing material research with it.

>> No.2455852

have fun reading

>> No.2455858

Also, DTI is my favorite <3

>> No.2455860

bumping a science thread since it looks like there's a horde of schoolchildren on sci today

>> No.2455861

physics graduate beginning a job in magnetic resonance in 5 days reporting in

cool thread, op

>> No.2455863

>>2455860
>>2455861
Thanks guys

>> No.2455864
File: 7 KB, 181x175, notread.jpg [View same] [iqdb] [saucenao] [google]
2455864

>> No.2455865

your loss brah

>> No.2455868

>>2455861 here
OP, you should look into what DNP has to offer the world of medical imaging, some interesting stuff there

>> No.2455871

>>2455868
Feel free to add to the thread!

>> No.2455877

>>2455871
i don't have time at the moment, but if you are interested checkout http://www.pnas.org/content/106/14/5453.full

or just search along the lines of "medical imaging: hyperpolarized contrast agents"

>> No.2455884

Do you think 3D real time high detail MRIs will be possible in a 10 years or so?

>> No.2455889

>>2455884
they are on the order of magnitude of real time. the reason patients have to wait a week or two is the time it takes doctors to analyse the scans themselves.

do you mean how long until the MRI software will be able to automatically pick up on things that currently takes time for doctors to check over?

>> No.2455890

>>2455877
Awesome. I had actually never heard of this technique.

>> No.2455894

>>2455884
Depending on what you would consider high detail, it's already available.

>> No.2455910

A+ OP

>> No.2455928
File: 26 KB, 400x400, what the fuck am I reading.png [View same] [iqdb] [saucenao] [google]
2455928

my head is not big enough for this information

>> No.2455956

Bump. Didn't do all that typing for nothing :_;

>> No.2455965

>>2455889
I mean as in. Completely 3d and moving and shit. Showing the brain activity traversing through different neural paths. Can they already do that?

>> No.2455974

It's somewhat disappointing to see the troll threads get 100+ posts within ten minutes, and real science threads (such as the above) get a few interested people, but that interest eventually dies off and before you know it it's at the bottom of page 15.

>somewhat
Scratch that, it's VERY disappointing.
/b/ truly has spread like cancer throughout the website.

Though I'd love to hear more behind the discovery or the physics behind MRI scanning. Anything to educate myself.

>> No.2455976

Sorry for what must be a dumb question, but how to NMR scans not trigger the same effects as transcranial magnetic stimulators? Are they so much weaker?

>> No.2455984

>>2455974

Eh, /sci/ has always been like that. People don't come here to learn about interesting science, but to flaunt their own intellectual superiority in the safe knowledge that nobody knows they're actually pathetic losers IRL.

That said, fuck year MRI. This is relevant to my interests and possibly future studies.

>> No.2455993

>>2455965
Well, the 3D part might be a bit problematic, but real time 2D MRI is certainly available. I think you'll find the following video / article both interesting and weird:
http://www.newscientist.com/article/dn17662-human-sex-from-the-inside-out.html

>brain activity traversing through different neural paths.
This is something for the relatively far future. Using NMR it's possible to track action potentials through very large axons in the leg for instance, but not in real time. In the brain this would be even more complicated because of the vast complexity of the neural architecture involved.

>> No.2455996

>>2455974
I appreciate it :)

>> No.2456001

>>2455976
TMS uses magnetic induction to trigger neural activity. The B0 field of an MRI scanner is a lot stronger than a TMS coil, however it's static. Fast movement through the field could potentially lead to the same type of activation, but in practice that doesn't happen.

>> No.2456006

started reading! wish me luck!

bump for science

>> No.2456009

>>2456001
Ahh, interesting.

>> No.2456014

>>2456009
Does it make sense though? (I'm pretty bad at explaining these things in simple terms, sorry about that)

>> No.2456021

>>2456006
>started reading! wish me luck!
Good luck! Let me know if you have any questions.

>> No.2456027

I don't know shit about neuroscience or whatever, but I understood the main idea of each paragraph and let me tell ya, it's pretty interesting.
Thanks for sharing bro.

PS: Yes, DTI seems kinda fun lol.

>> No.2456029

>>2455834
>The image here shows a T1 contrast (of my own brain)
wat

how?

>> No.2456035

>>2456027
>Thanks for sharing bro.
No problem!

>>2456029
It's my job. I go in the scanner myself when pilot testing a new behavioral task.

>> No.2456045
File: 39 KB, 504x378, mri-cocaine.jpg [View same] [iqdb] [saucenao] [google]
2456045

I have a set of MRI scans of a friend's pelvis... she trusted me with them, so I can't share them here, but it's fucking awesome.

I did my research paper on the MRI... 12 pages of awesome, that's what it was. Pic related, here's the effect of cocaine.

>> No.2456054

>>2456045
If I may ask, where did you get the image?

>> No.2456062
File: 25 KB, 801x801, b1.png [View same] [iqdb] [saucenao] [google]
2456062

>>2456054
That image was just from an image search. Why?

>> No.2456063

Read every post. Fuck yeah MRI.
>I think it deserves some appreciation.
Well, didn't they get the Nobel prize for developing it?

>> No.2456066

>>2456063
>Well, didn't they get the Nobel prize for developing it?

http://en.wikipedia.org/wiki/Paul_Lauterbur
http://en.wikipedia.org/wiki/Peter_Mansfield

Yep, they did.

>> No.2456069

>>2456062
Ah I see. Well, usually when data is presented overlaid on a whole brain like that you should be suspicious of it. It's a way to mask there is something wrong with the data, because it sort of hides the exact anatomical placement of the activity. The blobs aren't activity directly on the cortex, but also protruding into white matter.

>> No.2456072

>>2456066

And also for the development of NMR:
http://en.wikipedia.org/wiki/Felix_Bloch
http://en.wikipedia.org/wiki/Edward_Mills_Purcell

>> No.2456078

>>2456063
>>2456066
Well I guess they did get plenty of appreciation already, but still, fucking awesome :)

>> No.2456095
File: 308 KB, 800x388, 1268859709104.png [View same] [iqdb] [saucenao] [google]
2456095

>> No.2456096

>>2456095
Fucking lol. Yeah they can be dangerous if you're not careful.

>> No.2456106

>>2456096

DNP guy from a while ago here. What do you operate at, 1-3 Tesla? I do love hearing people (with no scientific education) say things like "magnestism is bad for you". I've stuck my arm and head in the bore of a 14.1 Tesla magnet - i'm yet to turn into Magneto.

>> No.2456114

>>2456106
Regular old 3T.
>I do love hearing people (with no scientific education) say things like "magnestism is bad for you".
My mother is one of those people literally afraid of big magnets. She thinks it causes your spiritual energy to be disrupted or something.
>I've stuck my arm and head in the bore of a 14.1 Tesla magnet - i'm yet to turn into Magneto.
Lol, was it a scanner bore, or just a big ass magnet?

>> No.2456241

>>2456114
Bruker 500 MHz Magnet (I suppose that qualifies as big-ass)

>> No.2456242

>>2456045
stfu and post them

>> No.2456265

>>2456095

THAT PICTURE IS OLD AS SHIT

LONG BEFORE SCI

BORING

PENIS

>> No.2456275

Research paper fag here again, I got clocked with a line-drive to the eye and got an MRI... they let me see my scans, but they said it was against the rules to show me. Any idea why that might be?

>> No.2456281

>>2456275
They might not want you looking at them until a radiologist has looked at them in case you go all hypochondriac on them.

>> No.2456282

>>2456275
probably the same reasons they won't show you your full CT/PET scans, only sections which they cut out after the analysis. it's to stop you going "oh whats that?? maybe you should check that out too..."

>> No.2456290

>>2456265
It was an epic long thread here on /sci/ you newfag moron

>> No.2456325

>>2456275
I don't know about hospitals, but in research institutions the usual policy is not to show people their scans in case of accidental findings. Sometimes (rarely) actual abnormalities are found, but people often get worried over nothing. So yeah, pretty much this: >>2456282

>> No.2456334

>>2456325
are people with larger forehead-ear distance smarter? Ive heard that the frontal parts of the brain are responsible for advanced functions

>> No.2456344

>>2456265

Confirmed for post of the year.

>> No.2456350

>>2456334
There's probably a correlation there, but it's not necessarily meaningful. To simplify it a bit (a lot), the frontal parts of the brain exert control over the rest. To some extent it modulates functions in other areas of the brain. You could call this advanced, because humans are the species which has the relative size of the frontal cortex, and phylogenetically it is fairly new.

>> No.2456407

how to bold?

>> No.2456422

>>2456407
Use LaTeX

>> No.2456439

>>2456282

this guy is totally right, but remember if you ask, they have to show you, its the law that you have the right to all your medical information

>> No.2456472

I appreciate this thread, thus bump.

>> No.2456496

Disappointing lack of research results is disappointing.

>> No.2456505

>>2456472
same. we need more of these

>> No.2456510

>>2456496
Leave it up to the faggots to discredit a great thread.

>> No.2456517

>>2456496
>implying you read it all

>> No.2456552

let's keep this shit on the front page

>> No.2456611

>>2456496
This thread was meant to explain the principle of MRI, and to show what can be done with it.

If you want specific research, just ask what you want and I'll look something up.

>> No.2456724

How about something of the distinctions between brainhalfs and how they interact?

Supposedly right half is creative and left is logical. It wasn't disproved was it?

Once my class had an excercise. First we we told to draw something. Then we we told to copy it. The latter drawings turned out remarkably worse.

>> No.2456792

>>2456724
Sorry, I completely forgot about my thread (just had dinner)

What you're referring to is more a popular science notion. There's no clear distinction between the hemispheres on the basis of logic or creativity. There is some lateralization of functions like language, and spatial reasoning, and I think this is where the idea came from. If you want actual research results you'll have to be a bit more specific.

>> No.2456845

bumping one last time and then I'm off

>> No.2456852

>>2456290

Epic thread for the troll that trolled you all.

>> No.2456886

ACTUAL SCIENCE!!? ON MY /SCI/!? I must be dreaming :3

>> No.2456887
File: 44 KB, 496x384, 1281856977083.jpg [View same] [iqdb] [saucenao] [google]
2456887

I read that Sony had developed extremely high resolution MRI technology a few years back. Possibly several times more accurate than the status quo if I remember correctly.

What is the current state of MRI resolution? When will we be able to observe individual neurons with some degree of accuracy?

>> No.2456920

Bumping for justice.

>> No.2456956

>>2456887
Well the spatial resolution of the scanners themselves isn't really the problem. Human MRI goes up to 11 tesla field strength, that gives you <1mm^3 voxels. The problem is the measure of neural activity is indirect, and dependent on blood oxygenation, and thus the local capillary bed. In principle we can distinguish between the cortical layers with the current scanners of only 3T. Even if you'd be able to measure voxels smaller than a cubic micrometer (which is possible with ultra high resolution MRI used on small rodents) that won't really benefit the spatial resolution in locating neural activity.

We'd have to come up with some way of measuring neural activity directly, and non-invasively. Using magnetic resonance this would be pretty difficult.

I'm off now. Thanks to everyone contributing!

>> No.2456963

>>2456956
Thank you for the enlightenment CNS.

>> No.2456987

This thread... whoa

>> No.2457009

>>2456956
> that gives you <1mm^3 voxels
That's not enough.

We need MRI that can see the connectivity and structures of individual neurons and individual synapses. Hell, it should see the structure of membrane proteins for optimal performance.

Why? Mind uploading. Not feasible today, but when exaflop supercomputers start to pop up(2020, in nine years, or less) it will start to be feasible to simulate activity patterns in large brain structures. If we had synapse resolution today we could probably do virtual experimental neuroscience on fly brains, yielding new insights.

>> No.2457040

LAtest development in MRI: SQUIDs, or Superconducting Quantum Interference Devices. Basically, supersensitive magnetometers made using Josephson Junctions. Can give current MRI resolutions using the earth's magnetic field in place of the main field (you still need gradient coils, but they can be weaker), or much higher resolutions at much higher refresh rates at existing field strengths.

>> No.2457094

>>2457040
lolwat MEG doesn't use magnetic resonance

>> No.2457118
File: 48 KB, 900x300, PBF020-Skub.gif [View same] [iqdb] [saucenao] [google]
2457118

>>2457094
Fuck of MEGfag, this is an MRI thread.

>> No.2457134
File: 2.64 MB, 2848x4288, NIMH_MEG.jpg [View same] [iqdb] [saucenao] [google]
2457134

>>2457040
READY FOR TAKEOFF

>> No.2457140

>>2457118
blame the guy posting about SQUIDS

>> No.2457159

>>2457140
I AM the guy posting about SQUIDs. Yes, they are used in MEG, but their appliance to MRI is not unusual. It's rather common for small sample, often inorganic, MRI (e.g. structural analysis of composite material samples). There's more to MRI than full-body machines in hospitals.

>> No.2457162

>>2457094
ololol it's not even an imaging method either

>> No.2457170

>>2457159
I stand corrected.

>> No.2458255

Let's put this bitch back up on the front page so the amerifags can benefit.

>> No.2458277

FUCK YEAH

>> No.2458294

You are now aware that MRI is the single most over-used method in all of neuroscience.

>> No.2458308

>>2458294
+88 posts? Surprising for an actual science thread.

>> No.2458314

>>2458294
with good reason brah... with good reason.

CAPTCHA: giveriv neuronal
how relevant

>> No.2458338

ARCHIVE

>> No.2458354

Too much reading for the americunts?

>> No.2458364

>>2458354
I theory (guess)

>> No.2458367

>>2458354
USA: BOOKS ARE FOR FAGGOTS

>> No.2458444

tits

>> No.2458454

DNPfag reporting in again

>>2456887
pretty much what CNS said, although i deal with powder samples/solutions (NMR spectroscopy, not MRI). given a reasonable sample size, i can detect differences in molecular structure (say for a molecule of 20 proton components) up to a sensitivity of 1:10^9

>> No.2458965

I graduated from a pretty good research university with a B.A. in biochemistry in May of 2009.

What would I have to do to become an MRI technician (or whatever they call it)?

>> No.2459085

thank you OP. I wish all /sci/ threads were like this.

>> No.2459236

Bumping this awesome /sci/ thread.

Thank you CNS

>> No.2459261

MIR incress you geting cancer more then 50%.
so unless you have somthing really wroung with you do NOT have one of these done to you.

some if the test comes back - theres a good chance down the road you will have cancer from the MIR test

>> No.2459280

posting in an awesome thread.

>> No.2459367

>>2456956
Another problem with the MRI is that the temporal resolution isn't so great because it relies on oxygen levels in the brain. It takes time for that oxygen to reach the brain. Even though it's just seconds, that's a long time considering how fast neurons can fire.

>> No.2459379

>>2459261

... what?

>> No.2459501

Bump for America. I like it here.

>> No.2459735

>>2458965

Does anybody want to answer my question?

CNS?

>> No.2461340

Let's put this bitch back up on the front page

>> No.2461347

>>2459261
lol moron
MRI don't give you cancer... CT scans do because they work with x-rays, but nothing near 50% increased chance

>> No.2461394

Toasting in epic bread. Awesome science is awesome

>> No.2461428

If I was a mod I'd make this thread a sticky ;_;

>> No.2461462

>>2455833
quick question
i thought space wasn't composed of voxels, or that space isn't voxelated, in the sense that information is stored in two dimensions (holographic principle) and not in three dimensions.

>> No.2461485

>>2461462
I think OP's gone my friend

>> No.2461722

Wow this thread went on a lot longer than I expected.
>>2461485
Back again ;)
>>2461462
For all intents and purposes the space we observe is three dimensional. It can be divided up into voxels of local homogeneous field strength manually.
>>2459367
Correct you are. The temporal resolution is probably a bigger drawback than the spatial resolution.
>>2458965
What do you mean by MRI technician? Do you want to fix or design new MRI devises, or do you want to work as a lab technician operating an MRI scanner?

>> No.2461790

OP I LOVE YOU THIS IS THE FIRST TIME I SEE THE TECHNIQUE CLEARLY EXPLAINED IN A SHORT AND DIRECT MANNER

please make more of such threads.

>> No.2461795

>>2461790
<3

>> No.2462046

ima bumin this shit

>> No.2462055

>>2461722
>for all intents and purposes

idiot, it's 'for all intensive purposes.' 'intents' and 'purposes' mean the same thing and that's just redundant, I don't know why this error is so common.

>> No.2462074

>>2462055
English isn't my native language, so lets refrain from calling names shall we? Lets see you learn my language.

Besides:
http://www.thefreedictionary.com/for+all+intents+and+purposes
Also, http://en.wiktionary.org/wiki/for_all_intents_and_purposes

>> No.2462095

>>2462055
>'intents' and 'purposes' mean the same thing
That's the point ignoramus

>> No.2462120

>>2462074

I know this isn't the purpose of the thread, but could you perhaps tell me which is more relevant to graduate neuroscience studies, calculus or physics? I really can only take one of them before I graduate and I wasn't sure which looks better for applications and which, if either is actually used.

>> No.2462121
File: 44 KB, 594x446, 1295653518884.jpg [View same] [iqdb] [saucenao] [google]
2462121

OP is a gentleman and a scholar.

>> No.2462140
File: 1016 KB, 470x394, KobeBlackMamba.png [View same] [iqdb] [saucenao] [google]
2462140

>>2462074
>>2462095
Are you guys seriously going to fall for that? It's the trolling flavor of the month.

>> No.2462145

{\bf Intredasting}

>> No.2462154

>>2462120
Probably physics, but you won't need a whole lot of physics anyway. Chem and biology are the two most fundamental disciplines to master. Only if you want to learn the intricacies of molecular dynamics of electrophysiology, or if you want to learn how to properly use neuroimaging techniques will physics be somewhat useful. Calc can be handy if you want to get into multivariate statistics and such, but all these things are probably for grad-school. I'd say physics, but it really doesn't matter all that much.

>> No.2462157

>>2462121
<3

>> No.2462178

>>2462154

Thanks, man. Keep up the good work, your threads are always a pleasure to read.

>> No.2462184

>>2462178
No problem, and that's good to hear!

>> No.2462217

Thank you for this. I am an RT(R) and considering on moving into advanced modalities. I was torn between CT and MR but I think you just sold me!

>> No.2462361

bump

>> No.2462846

>What are you doing, thread? You're good, you don't belong to page 7.

http://neurodojo.blogspot.com/2011/01/scanning-salmon-smelling-streams.html

>> No.2463557

>>2462846
>http://neurodojo.blogspot.com/2011/01/scanning-salmon-smelling-streams.html
Great article. Would read again.

>> No.2463568
File: 206 KB, 469x550, 1296488933436.jpg [View same] [iqdb] [saucenao] [google]
2463568

http://www.jsur.org/ar/jsur_ben102010.pdf
>Journal of Serendipitous and Unexpected Results
That has to be the biggest troll journal out there.

>> No.2463679

>>2461722
Yes, I understand that the "fields" may be voxelated, but the amount of information that can be stored in an area is not voxelated.

>> No.2463935

>>2455834
>The brain of an anon.
>:O

>> No.2464081
File: 16 KB, 200x263, Magnets.jpg [View same] [iqdb] [saucenao] [google]
2464081

>> No.2464087

tl;ra
>too long. read anyway

>> No.2464111

>>2464081
I can't believe this is the first trollmagnet reference in the entire thread.

>> No.2464308

bump

>> No.2464703

>>2461722
I mean a lab technician who operates the devices.