Abstract:Time- or frequency-dependent (“restricted”) diffusion potentially provides useful information about cellular-scale structures in the brain but is challenging to interpret because of the intravoxel tissue heterogeneity. Frequency-dependence was recently incorporated in a multidimensional diffusion-relaxation correlation MRI framework relying on tensor-valued diffusion encoding at multiple frequencies, as well as variable echo and repetition times, to give model-free characterization of intravoxel heterogeneity by Monte Carlo data inversion into nonparametric distributions of frequency-dependent diffusion tensors and nuclear relaxation rates. While microimaging equipment with high-amplitude magnetic field gradients allows exploration of frequencies from tens to hundreds of Hz, clinical scanners with more moderate gradient capabilities are limited to a narrow frequency range, which may be insufficient to observe effects of restricted diffusion for brain tissues. We here investigate the effects of including or omitting frequency-dependence in the data inversion from isotropic and anisotropic liquids, excised tumor tissue, ex vivo mouse brain, and in vivo human brain. For microimaging measurements covering a wide frequency range, from 35 to 320 Hz at -values over 4·10 sm , the inclusion of frequency-dependence drastically reduces the fit residuals and avoids bias in the diffusion metrics for tumor and brain voxels with micrometer-scale structures. Conversely, for the case of in vivo human brain investigated in the narrow frequency range from 5 to 11 Hz at = 3·10 sm , analyses with and without inclusion of frequency-dependence yield similar fit residuals and diffusion metrics for all voxels. These results indicate that frequency-dependent inversion may be generally applied to diffusion-relaxation correlation MRI data with and without observable effects of restricted diffusion.

What problem does this paper attempt to address?

The problem this paper attempts to address is whether frequency dependence is a meaningful parameter in multidimensional diffusion-relaxation related magnetic resonance imaging (MRI), or if it merely leads to overfitting. Specifically, the authors explore whether frequency dependence should be included when performing data inversion over different frequency ranges (broadband and narrowband), and how this choice affects fitting residuals and diffusion metrics. ### Background and Problem - **Time or Frequency Dependent Diffusion**: This type of diffusion can provide useful information about the microstructural scale of brain tissue, but it is challenging to interpret due to tissue heterogeneity within a voxel. - **Frequency Dependent Diffusion**: Recently incorporated into the multidimensional diffusion-relaxation related MRI framework, it allows for model-free characterization of intra-voxel heterogeneity through tensor-valued diffusion encoding and variable echo and repetition times. - **Clinical Scanner Limitations**: Due to limited gradient capabilities, clinical scanners can only operate within a narrow frequency range, which may be insufficient to observe the effects of restricted diffusion. ### Research Objectives - **Validate Frequency Dependence**: Investigate whether frequency dependence should be included when performing data inversion over different frequency ranges (from 35 to 320 Hz covered by microimaging devices, to 5 to 11 Hz covered by clinical scanners). - **Evaluate Impact**: Analyze the impact of frequency dependence on fitting residuals and diffusion metrics, particularly in microimaging measurements and in vivo human brain measurements. ### Methods - **Experimental Design**: Conduct experiments using both microimaging systems and clinical systems, covering different frequency ranges. - **Data Inversion**: Employ Monte Carlo data inversion methods to generate non-parametric distributions of frequency-dependent diffusion tensors and nuclear relaxation rates. - **Comparative Analysis**: Compare inversion results with and without frequency dependence, assessing changes in fitting residuals and diffusion metrics. ### Results - **Microimaging Data**: In the broadband frequency range (35 to 320 Hz), including frequency dependence significantly reduced fitting residuals and avoided biases in diffusion metrics of micron-scale structures within tumor and brain tissue voxels. - **Clinical Data**: In the narrow frequency range (5 to 11 Hz), inversion results with and without frequency dependence showed similar fitting residuals and diffusion metrics across all voxels. ### Conclusion - **Importance of Frequency Dependence**: Frequency-dependent inversion can be universally applied to diffusion-relaxation related MRI data, regardless of whether the effects of restricted diffusion can be observed. - **Recommendation**: When interpreting results of frequency-dependent water populations, consider fully or partially mixed scenarios, and explicitly include frequency dependence in data inversion to accommodate cases where restricted mechanisms fall within the used frequency window.

Frequency-dependence in multidimensional diffusion-relaxation correlation MRI of the brain: overfitting or meaningful parameter?

Analysis of the effects of noise, DWI sampling, and value of assumed parameters in diffusion MRI models

In vivo disentanglement of diffusion frequency‐dependence, tensor shape, and relaxation using multidimensional MRI

The variability of multidimensional diffusion-relaxation MRI estimates in the human brain

Model-free approach to the interpretation of restricted and anisotropic self-diffusion in magnetic resonance of biological tissues

Resolving orientation-specific diffusion-relaxation features via Monte-Carlo density-peak clustering in heterogeneous brain tissue

In vivo observation and biophysical interpretation of time-dependent diffusion in human cortical gray matter

Diffusion dispersion imaging: Mapping oscillating gradient spin‐echo frequency dependence in the human brain

Angular Resolution Enhancement of Diffusion MRI Data Using Inter-Subject Information Transfer.

Improving Estimation of Fiber Orientations in Diffusion MRI Using Inter-Subject Information Sharing

Intrinsic Frequencies of the Resting-State fMRI Signal: The Frequency Dependence of Functional Connectivity and the Effect of Mode Mixing

Reproducibility of the Standard Model of diffusion in white matter on clinical MRI systems

Diffusion Dispersion Imaging: Mapping OGSE Frequency Dependence in the Human Brain

Multifractal Long-Range Dependence Pattern of Functional Magnetic Resonance Imaging in the Human Brain at Rest.

In vivo rat-brain mapping of multiple gray matter water populations using nonparametric D(omega)-R1-R2 distributions MRI

In-vivo characterization of magnetic inclusions in the subcortex from non-exponential transverse relaxation decay

Universal power-law scaling of water diffusion in human brain defines what we see with MRI

Robust Fiber Orientation Distribution Function Estimation Using Deep Constrained Spherical Deconvolution for Diffusion MRI

Unique information from common diffusion MRI models about white-matter differences across the human adult lifespan

Generalisation of continuous time random walk to anomalous diffusion MRI models with an age-related evaluation of human corpus callosum

Sensitivity of quantitative diffusion MRI tractography and microstructure to anisotropic spatial sampling