MRI measures the signal from millimeter-sized voxels, inside which billions of spins interact through complex biological processes. The classical Bloch equation, derived in 1946, does not account for this heterogeneity, which is why quantitative parameters, such as relaxation times, depend on the measurement protocol. Existing approaches add new compartments or exchange pathways, leading to increased model complexity. This work stands apart from neighboring research: nearby are studies on quantum sensors, magnon traps, and magnetic fields of stars, but none of them address an effective theory of MRI in tissues. The authors propose a different path: rather than complicating the microscopic description, they build a phenomenological model of macroscopic dynamics, following the spirit of Bloch himself.
The authors consider the dynamics of magnetization after the radiofrequency excitation and gradient fields are turned off, in the so-called rotating frame. Instead of the standard linear relaxation term, they introduce a general nonlinear vector field that must satisfy four physical requirements: equivariance under rotations about the longitudinal axis, uniqueness and global asymptotic stability of thermal equilibrium, and strict dissipativity, that is, monotonic decrease of the Lyapunov function (the squared distance to equilibrium) along trajectories. Using methods from dynamical systems and group theory, they fully characterize admissible relaxation fields: they prove that these fields have a special structure depending on radial and longitudinal coordinates, and they derive conditions on the functions defining the field. They then provide a constructive procedure for generating a wide class of admissible models, including examples with non-exponential longitudinal relaxation and coupled transverse-longitudinal dynamics.
The main result is a theorem on the complete characterization of admissible relaxation fields. It is shown that any such field has the form of a sum of a diagonal part, governing decay, and a skew-symmetric part, responsible for rotation about the longitudinal axis; both parts depend only on the radial coordinate and the longitudinal component of magnetization. The admissibility conditions include smoothness on the symmetry axis, uniqueness of the zero of the longitudinal function, negativity of coefficients at the equilibrium point, and a key dissipativity inequality. The authors also give an explicit construction: by choosing positive functions a and c, a rotation function β, and a coupling function ψ, one can build an infinite family of admissible nonlinear models. The classical Bloch model is obtained as a special case with β and ψ zero and constant a=1/T2, c=1/T1. Examples demonstrate non-exponential recovery of longitudinal magnetization, characteristic of magnetization transfer effects, and coupled relaxation, where the rate of transverse decay depends on the longitudinal deviation.
This work differs from neighboring studies in that it does not add microscopic details but builds an effective theory at the macroscopic level, preserving the dimensionality of the classical Bloch equation. This is a fundamentally different approach to describing MRI in heterogeneous media: instead of increasing the number of parameters, it restricts the class of admissible dynamics by physical principles. The obtained characterization provides a rigorous basis for developing new relaxation models that may reduce the protocol dependence of quantitative parameters. The theorem also shows that admissible nonlinearities are not arbitrary: they must have a specific structure, which narrows the search space and guides future experiments.
The proposed framework opens the way to systematically constructing phenomenological models of MRI in tissues. Immediate next steps include selecting specific parameterizations from the admissible class, calibrating them on experimental data, and testing for protocol independence. The theory could be extended to non-exponential transverse relaxation, observed in tissues with high iron content. One could also relax or modify the original physical requirements to cover other scenarios, such as the effects of diffusion or exchange.
The theory may impact quantitative MRI and relaxometry by providing new models for describing signals in biological tissues. It also sets a direction for developing MRI data analysis methods that are robust to protocol choice.
The next step will be a numerical implementation of admissible models and comparison of their predictions with real MRI data for various tissues. Then one can investigate which nonlinear terms best describe observed deviations from classical exponential relaxation.
In the field of MRI, the problem of protocol independence of quantitative parameters remains unsolved: T1 and T2 measurements depend on the pulse sequence. This work offers a mathematical apparatus for constructing models that may reduce this dependence, but experimental verification is still ahead. Another open question is the connection between effective parameters and tissue microstructure; the theory does not yet provide such a link, but it creates a foundation for finding it.
This section is addressed to the author of the paper and their colleagues. Neighbouring works were found by meaning-based search across our archive, not by word overlap. We do not review: we show the field around the work.
What the work does
The authors construct an effective theory of magnetization relaxation by replacing the linear Bloch term with a general nonlinear field that satisfies rotational symmetry, equilibrium stability, and dissipativity. They obtain a full characterization of admissible fields and a constructive method for building them, with the classical Bloch model as a special case.
Where its strength is
The strength lies in the rigorous mathematical characterization of admissible nonlinear relaxation fields, which provides a foundation for the phenomenological description of MRI signals in heterogeneous tissues without complicating microscopic models.
Where it could go next
Worth trying
Why this may matter
If nonlinear relaxation models prove robust to variations in protocols, they could yield more reproducible quantitative biomarkers in MRI, reducing dependence on measurement parameters.
Nearby works in our archive
The archive is sparse around this work: the closest review is only 0.49, against a usual 0.64. It looks like few have drilled nearby.
The classical Bloch equation was proposed by Felix Bloch in 1946 as a phenomenological description, without derivation from microscopic physics. The new theory follows the same spirit: it postulates physical principles and derives admissible forms of dynamics, rather than trying to model every spin.