This tutorial aims at introducing the molecular mechanisms behind the relaxation times, and possible pitfalls in their experimental determination. The focus of the contribution is on the field-dependence of relaxation times, the importance of parasitic effects, and on addressing non-exponential signal behavior in a quantitative manner.
Relaxation times constitute the main contrast parameters in MRI. Apart from T1 and T2, further parameters such as T1ρ can be readily implemented on clinical scanners and can be made available for contrast. However, the use of relaxation times frequently involves a number of misconceptions as well as subtleties considering the execution of the standard pulse sequences. This tutorial will provide background information about how to extract the maximum from experimental data:
First of all, relaxation times are field-dependent – “the T1” does not exist. All relaxation times are expressed by the molecular reorientation spectrum of water (or fat), with additional contributions possibly stemming from contrast agents, oxygen, but also interfaces and molecular or magnetization exchange. To this day, only crude approximations of the actual field dependence of these molecular processes exist. While there is empirical knowledge about the increase of T1 with magnetic field strength in most biological tissue, the situation is more complex with T2 which has a weak field dependence but where susceptibility differences may become dominating. In this case, one needs to distinguish the true T2 from additional effects which are more suitably expressed by T2eff or T2*. In particular for T2, the experiment determines the outcome of the contrast parameter – pulse separations in an echo experiment and the orientation dependence of relaxation of water in the vicinity of aligned structures can alter the result by several orders of magnitude. Likewise, T1ρ, which has gained popularity in recent years, depends on the power of the lock pulses, but the excitation bandwidth of these pulses needs to be taken into account, especially for voxels with a large chemical shift distribution such as coexisting water and fat. A further aspects of vast potential is the distribution of relaxation times. Subtle deviations from exponential decay or build-up functions hint at structural or dynamic heterogeneities in the voxel, and ineffective exchange between different molecular environments. While an approximation by exponential fits will be a sufficiently good description of the contrast parameter, the degree of non-exponentiality contains important information about the tissue properties. Suggestions will be given how to extract this information, and where is the limit of distinguishing between two different but similar relaxation times.