Does the MR signal persist forever?
No. In the sample there are minute variations in magnetic field strength. This causes the different spins to experience slightly different magnetic fields (and therefore different Larmor frequencies). Because the spins are precessing at different frequencies, they loss their phase coherence, and the signal decays. This process is called transverse relaxation (because it happens while the spins are in the transverse plane), and is characterized by an exponential time constant, T2. For most biological metabolites, T2 is on the order of tens to hundreds of ms. The spins also lose phase coherence (dephase) because of an inhomogeneous B0 field. This occurs in part because an imperfectly engineered magnet and also because putting the sample into the magnet disrupts its homogeneity somewhat. Thus there is a faster effective loss of signal, which is characterized by the time constant T2*.
