Do you feel discouraged and downtrodden every time you sit through an MRI physics lecture? Do you feel the need to close your eyes and forget the past as hallucinations of spinning arrows and frequency space dance through your mind? If so, go see your doctor... Hallucinations are bad. But after you see your doctor, turn the lights up in the reading room, buckle in and join us as we discuss MRI Physics, because it's the lecture series the networks don't want you to see!.
Like our videos? Consider supporting us on our quest to bring truth to MRI physics by donating below: www.buymeacoffee.com/lightsonrads
Follow Dr. TE on: Instagram: instagram.com/lightsonradiology Twitter: @DrTEMD Reddit: Dr_TE_MD
Email your questions or requests to: lightsonradiology@proton.me
Thank you! My brain hurts equally as much making them haha, if you liked this one then definitely check out the latest on Diffusion Tensor Imaging which is a continuation of this one: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-mbpUalR_Z3E.html
This is the best MRI explanation I’ve ever seen and it helped! Hope more people could see this so we don’t have to suffer from profs pretending they know mri
This is great. One thing that helped things click was to spend some time thinking about "why" an electric field generates a magnetic field. A moving electron alone actually does not produce a magnetic field. It requires electrons and protons to produce a magnetic field. As the electrons move relative to protons (ions) there is a relativistic charge (per unit volume) difference between the positive and the negative charges. This causes any external charges to feel a force we know as the magnetic field. The magnetic field is a relativistic correction to the electrostatic field. The metals in MRI machines have more freely floating electrons due to their internal lattices, which makes them generate stronger fields.
Always love hearing about the relativistic and quantum mechanics underpinning these simplified classical explanations! If you like getting into the weeds, the latest lecture is probably the most technical yet. Check it out here if you're interested! ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-mbpUalR_Z3E.html
One additional thing you may have noticed in practice but which I forgot to add to the video... On some MRI exams, you may see DTI was performed instead of DWI, yet you still see an Isotropic or Trace image produced. From the video, hopefully you can now see DTI is somewhat of an overkill version of DWI. These "cheap" DTI sequences typically use the minimum 6 gradients needed to perform DTI imaging and generate FA images yet are not accurate enough to use for Tractography. However, you can still calculate a Trace image from the images produced by the diagonal gradients (Dxx, Dyy, Dzz), as well as an isotropic image (geometric mean) of the images produced from all gradients applied and thus you clinically treat these just the same as the Trace or Isotropic images produced from a standard DWI sequence.
Hi, Thanks a lot... these video's are really helpful. Can you please clarify, why the RF pulse is applied before phase or frequency gradient? Logically RF pulse should be applied after the slice, phase and frequency is selected right?
So think of the RF pulse as the initial step to energize the system. This is what gets all those protons precessing together around the B0/Z axis. The slice select gradient is applied at the same time as the RF pulse so that we only energize the slice that we want to produce an image of (explained in ttps://ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-v8jW8K1y-KE.html). It is only AFTER our RF pulse where the protons are precessing together that we can apply our frequency and phase encoding gradients (further explained in the frequency encoding lecture ttps://ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-DYj1SLNppQM.html and phase encoding lecture ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-nFDzXvjF7gg.html). Hope this helps!
Doesn't the precession happen as soon as the protons enter the magnetic field? A combination of the field b0 and the spin angular momentum of the protons cause them to precess. The RF pulse just tips the alignment of that precession and causes phase coherence. In other words how would the lamor frequency work if the protons did not already precess at that frequency?
These are good questions, and as always with MRI physics there are layers of ever finer details you will find the more specific you try to understand it. When we talk about precession in MRI physics, what we're really referring to is the precession of the net magnetic vector. Yes, the individual protons precess about their axis when put into a magnetic field. You don't even need an MRI machine, we're in a magnetic field right now on Earth. But does this generate signal in our MRI machine? So we need to get these individual precessing protons aligned, and with enough energy from our magnetic field, we can cause more to align with the field than the random orientations caused by the background kinetic energy. This is what builds our net magnetic vector, and it doesn't necessarily mean all the individual protons are in phase with each other, but there is a net magnetic vector due to their alignment. The RF pulse then tips all of these aligned protons off the B0 axis, and they precess together at the larmor frequency, generating our signal. It is the dephasing of the orientation that causes our signal loss. Check out this article for more information: www.mriquestions.com/how-does-b1-tip-m.html
Lectures up until now were great but this one sucks. The animations are just too hard to understand, also you said the refocusing pulse is ninety degrees to reverse the direction of the rotation of the magnetic vector you have to flip 180 degrees, so more hydrogen ions generally pointing opposite the B zero field instead of in the same direction.
Ha I will be the first to admit my animations are rudimentary and not the best, and this is a very difficult subject so not quite sure how to present this in a different way but the key point is if we're able to invert the spins with a strong enough RF pulse, the receiver coil will see a reversal of the spins that causes rephasing of our spins and signal. The screw example at the end is the most tangible example of this I could think of. This is just to get a grasp on what's going on before proceeding to the pulse sequences, starting with the spin-echo sequence ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-vK6PeCPpOLY.html where we flush out this idea out a little more so maybe check out that video and see if it helps further clear up these concepts.
This is because we have changed the magnetic field across the x-axis so that the CSF voxel see a different field strength than the fat voxel, and remember that the Larmor frequency is directly related to the magnetic field strength for each voxel. Hope this helps!
This perplexing, I got to know how to calculate it when there is only a phsse shift, but how do we go calculating it when there is also a frequency shift at the time we measured the signal
The frequency localizing gradient is not a shift but an encoding of different frequencies along the x-axis. Since we create this gradient magnetic field, we know which frequencies should be where spatially along the frequency encoding axis, so we when we break our raw signal down into frequencies via the Fourier Transform, we get this information. Check out the dedicated frequency encoding lecture here for further info if you haven't already: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-DYj1SLNppQM.html
Thanks for this amazing set of lectures; Question... in the Noa image when the hand pokes the vectors I understand in whats its said that only after the poking of the protons start to precess, or that move already exist before the stimulation?
Hello, thanks for commenting and great question! While the individual protons will have some baseline spin like all atoms and particles, they are not precessing all together as one. We need to impart them with some angular energy with our RF pulse to get them precessing as one unit together. Hope this helps!
Well, now, you just had to go and introduce FLAIR. As I'm sure you must be aware, that now means you're contractually obliged to cover *all* the common inversion recovery sequences.
Great video, could you make a video about difference between monopolar and bipolar dwi and on the meaning of 3 scan 4scan 3diagonal gradients (siemens) Thank you!!
Will definitely put it on my radar, thanks for watching and commenting! Nearing the final stages of the next lecture on DTI, should be a good one so stay tuned!
I noticed that the T1 relaxation times are constant. So does the MRI tech choose the TR time only to control TE or does the tech choose both TR and TE? I'm taking the ARRT test in December 2024
In general, the chosen TR time will affect T1 image contrast, and the chosen TE time will affect T2 image contrast and these values will remain constant for that sequence. For most sequences, these parameters are all ready chosen and optimized so the technologist simply picks the appropriate "protocol" to run. However, you can certainly manually go in and change the values of TE and TR, albeit this will change the contrast of the picture and you'll likely get an angry call from your local radiologist :)
Great video series! Ok, so if I understand correctly the magnetization recovery process can be described (omitting constants) by a differential equation f'(t)=1-f(t) where t is time, f'(t) is the re-magnetization rate which is proportional to 1-f(t) i.e. the difference missing to "fully magnetized", where f(t) is the current magnetization proportion (1 meaning 100% magnetized). Because the constant 1 vanishes when taking the derivative f'(t) must be -f(t) which means it must be the exponential function with negative exponent: f'(t) = -e^-t and f(t) = e^-t, because the exponential function is the only function that is its own derivative. And to account for the re-magnetization speed or respectively the re-magnetization time we introduce a scaling factor 1/T1. Which gives us the full formula 1 - e^(-t/T1). And because this function approaches 1 i.e. full re-magnetization only in the limit t -> infinity, we choose some finite time instead. With t/T1 = 0 meaning full de-magnetization the next obvious choice is to choose t/T1 = 1 i.e. t = T1 which gives us 1-e^-1 = 1-0.37 = 0.63 approximately. Phew, I think I got the math part 🙂
Thanks for commenting! I actually did a deep dive on this 63% and did a little segment on it in the Q&A video here, check if out if you'd like to learn more ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-3Zjm1wuFo6M.html
Local AM radio stations operate between 540 and 1700 kHz, not 68 MHz. FM radio stations operate between 88 and 108 MHz which is slightly higher than your example of a 1.5T magnetic field.
Thanks for the comment. This (should have been) already corrected with a caption that pops up during this segment saying the band is analog TV and into the FM range.
