Are there any systems, which exploit any sort of "reverse" MRI technique?

We can create magnetic field so that it has required value only in desired location (say, cancer tumor) and then pass powerful radiowaves of required frequencies. Only desired locations of body will absorb the power and we will have "directed microwave oven", i.e. will be able to destruct required tissues deep inside body.

  • $\begingroup$ This doesn't make sense to me; how can you create a localised magnetic field? And why would it even be necessary?--we don't need magnetic fields to make the body absorb radiation, we can even target radiation to specific parts using standard radiotherapy techniques. $\endgroup$
    – lemon
    Aug 11, 2016 at 10:46
  • $\begingroup$ In an MRI experiment up and down states are almost equally occupied with a difference on the order of one in $10^6$, or so, so you can't really exploit it to make a strong physical difference. $\endgroup$
    – CuriousOne
    Aug 11, 2016 at 19:48

2 Answers 2


People in my research field of ultracold atomic physics do use MRI techniques in this way. The fundamental idea (which you have hit on) is that the magnetic field gradient gives spatially dependent Larmor frequencies. Thus the radiowaves/microwaves you send in have spatial selectivity.

Here is a paper in which researchers hold atoms in a 1D optical lattice (lattice sites space by ~1 $\mu m$) inside of an optical cavity (cavity axis is the $z$-axis). Atoms are spin aligned along the $z$ axis along which there is a constant bias field $B_0 \hat{z}$. A magnetic field gradient $\frac{dB_z}{dz}$ is then turned on so that atoms in neighboring lattice sites have different Larmor frequencies. RF is then used to flip spins locally.

The atoms exhibit give a spin dependent phase shift to the circularly polarized light inside the cavity. If the atoms are pointed along the cavity axis they put a phase shift onto the light of $\phi_+$ and if they are anti-aligned with the cavity axis they put a phase shift $\phi_-$. Thus, depending on whether atoms are spin up or spin down the cavity has a differential phase shift $\phi_+ - \phi_- = \Delta \phi$. This differential phase shift manifests as a spin dependent resonances frequency of the optical cavity. Thus, by probing the resonance frequency of the optical cavity one can determine how many of the spins are spin up vs spin down and this is what is done in the experiment to show that spatially resolved spin flips (MRI) is being performed.

One could imagine selectively flipping certain spins within the cavity and using a spin-selective optical field to "blow away" atoms you do not want in the picture. This would be a direct analogy to the "MRI" knife to which the OP refers.

Such techniques are prolific in atomic physics but I don't know if they have been utilized in medical treatments.


I am not aware of such a concept that creates a certain magnetic field and then excites the spins there with the right frequency. That also seems to be quite challenging, since it is not so easy to form spaitally fitting magnetic fields.

However, there is a concept that uses a technique called parallel transmission (pTX), but in an alternative (therapeutic) fashion. However, this is still a pure research topic and not in wide use up to my knowledge. You could search for Thoralf Niendorf and his research on this subject.

In a nutshell:

Original intention of pTX

At ultra-high magnetic fields, such as 7T, MRI becomes very challenging, since the excitation wave lengths of the radio-frequency (RF) wave becomes so short that it is in the order of the spatial dimensions of the human body that you would like to measure. Hence, you are facing interference effects resulting in non-uniform excitation of the tissue and therefore varying image intensity, which is undesired. A solution to this problem is to use multiple RF transmitters that can transmit RF pulses independently (meaning different amplitude an phase). With an array of e.g. 8 transmit channels, the area can be excited much more homogeniously.

Therapeutic version

You can acquire an image of the tissue that contains a tumor or something alike. Then you can base treatment planning on this image, i.e. defining organs at risk and the target area. You then use the pTX technique but not to excite the magnetization homogeniously, but inhomogeniously. This means: Making sure that the RF waves interfere constructively in the treatment area, and destructively in the other areas, especially the organs at risk. By this technique you apply the maximum RF energy to the tumor and hence having the most destructive effect there.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.