From Diffusion MRI to Brain Connectomics - Modeling in Computational Biology and Biomedicine - page 191

In 1946 Bloch [ 12 ] and Purcell [ 56 ] independently extended the techniques
established by Rabi. They successfully demonstrated the magnetic resonance effect
in liquids and solids. Bloch and Purcell shared the Nobel prize of 1952 in
physics, and NMR was established. In his seminal paper of 1950 [ 31 ]ErwinHahn
proposed the spin echo experiment, which used a combination of 90 ◦ and 180 ◦
electromagnetic or radio frequency pulses to filter out effects of magnetic field
inhomogeneities in the measurement of the transverse signal. Further works of
Herman Carr and Purcell in 1954 [ 14 ] led to the full development of the radio
frequency pulse technique introduced by Hahn. This formed the foundations of
NMR.
It must be noted at this point that both the papers of Hahn [ 31 ]andCarr
and Purcell [ 14 ] critically point out the observed effects of diffusion of the spin
bearing nuclei in magnetic resonance experiments with a succession of radio
frequency pulses. Although these papers generally perceive the diffusion effect as
an unfortunate phenomenon resulting in a loss of signal, Carr and Purcell [ 14 ] in fact
demonstrate that diffusion can be directly measured from NMR and go on to actually
measure the diffusion constant of water at 25 ◦ C. This forms the corner-stone of
diffusion NMR.
Although NMR became a well established technique for studying various
materials, it took almost three decades since the experiments of Bloch and Purcell
in 1946, for MRI to be invented. NMR by itself is capable of examining a single
spin ensemble or a tiny region of a sample, but it can't image the whole sample to
recreate a 2D slice or a 3D volumetric image necessary to study entire biological
samples like the human body. Paul Lauterbur in 1973 [ 37 ] proposed the use of
magnetic gradient fields to spatially encode the positions or voxel regions of the spin
ensembles. This was a remarkable invention, which made it possible to reconstruct
entire slice or volumetric images from NMR data. Spatial encoding was improved in
terms of frequency encoding by Richard Ernst in 1978, and phase encoding by Bill
Edelstein in 1980 using pulsed gradients. In 1977 Peter Mansfield [ 42 ] developed
the mathematical framework for rapidly switching gradients for spatial encoding,
greatly speeding up the process of reconstructing images of an entire biological
sample. This is known as echo planar imaging (EPI). Lauterbur and Mansfield were
jointly awarded the Nobel prize in medicine in 2003 for making MRI possible.
Thus modern MRI was developed from the phenomenon of NMR coupled with
the method of spatial encoding.

6.3 Nuclear Magnetic Resonance and Diffusion

The principles of NMR are based on spin , a fundamental quantum characteristic
possessed by electrons, protons, and neutrons, like electrical charge and mass. Spins
come in multiples of and can be positive or negative. In grouped particles,

e.g. atomic nuclei, opposite spin-signs can pair up to eliminate the total spin
of the group. But the net spin of unpaired particles or atomic nuclei imparts a

1 / 2