Some unknown points about MRI

We all might have heard of the term MRI very often. Some of us would have had their MRI done as well. MRI is mainly used in clinical setting as well as in research setting as a variation - fMRI or functional MRI.  Magnetic Resonance Imaging is based on the principle of Nuclear Magnetic Resonance. I recently had the privilege of attending a nice lecture by Ramesh Venkatesan (GE Health Care). I’ll share with everyone some of the important points of the talk (while they are fresh in my mind!).

The talk started with the basic concept of NMR. Although NMR is a very general in the sense NMR signal can be generated from any atom. However, MRI uses H atom from the water molecule as its signal. In water, the Oxygen atom is the more electronegative atom which polarizes the Hydrogen atom and imparts it a slight positive charge. Due to this charge, nucleus has a magnetic moment. However, in normal circumstances these magnetic moments are randomly oriented and thus cancel each other. When an external magnetic field like the one used in a MRI scanner is applied, the magnetic moment of hydrogen nuclei is aligned parallel to the external field. 

Some of the interesting points that I learnt from the lecture were:

       ·         Fringe effects at the ends of the scanner bore. The magnetic fields are about twice as strong at the ends of the bore because of the property of magnetic field lines at the ends. 

       ·          Larger bore diameters can reduce such effects but a larger bore will mean that the magnetic field lines will travel to much more space and will need more containing.

      ·         When an external field B₀ is applied, we can use a rotating reference and get on the hydrogen atom to understand the effect of any other magnetic field applied (such as a gradient)

      ·         By controlling the time for which a gradient in horizontal direction is applied, we can set the flip angle. 

      ·         A gradient is used to figure out the source of the signal. If no gradient is present, on a RF pulse signals from all the regions will be present. To resolve this, a gradient field is used. 

      ·         Since a gradient field is used, the RF pulse frequency will be different for different positions (basic criterion for resonance).

      ·         Instead of a single RF pulse, a bandwidth (which is +/- 1 kHz ) is used to select a volume of the tissue with a certain thickness.

      ·         To fill up the k-space, a phase encoding is applied in one direction (a small gradient in that direction). This gradient is in addition to the z gradient already present. Read out is taken by keeping a fixed gradient in one direction (say y) and in x direction, another gradient is applied.

      ·         To better sample the k-space, blips or small magnetic field changes in the phase encoding direction are applied at the end of readout. So with only one shot of pulse, we can acquire full k-space.

      ·         Fat suppression or water only pulses are important for EPI images because of chemical shift properties.