TY - GEN

T1 - Conceptual Parallels Between Stochastic Geometry and Diffusion-Weighted MRI

AU - Dela Haije, Tom

AU - Feragen, Aasa

N1 - Publisher Copyright: © 2021, The Author(s).

PY - 2021

Y1 - 2021

N2 - Diffusion-weighted magnetic resonance imaging (MRI) is sensitive to ensemble-averaged molecular displacements, which provide valuable information on e.g. structural anisotropy in brain tissue. However, a concrete interpretation of diffusion-weighted MRI data in terms of physiological or structural parameters turns out to be extremely challenging. One of the main reasons for this is the multi-scale nature of the diffusion-weighted signal, as it is sensitive to the microscopic motion of particles averaged over macroscopic volumes. In order to analyze the geometrical patterns that occur in (diffusion-weighted measurements of) biological tissue and many other structures, we may invoke tools from the field of stochastic geometry. Stochastic geometry describes statistical methods and models that apply to random geometrical patterns of which we may only know the distribution. Despite its many uses in geology, astronomy, telecommunications, etc., its application in diffusion-weighted MRI has so far remained limited. In this work we review some fundamental results in the field of diffusion-weighted MRI from a stochastic geometrical perspective, and discuss briefly for which other questions stochastic geometry may prove useful. The observations presented in this paper are partly inspired by the Workshop on Diffusion MRI and Stochastic Geometry held at Sandbjerg Estate (Denmark) in 2019, which aimed to foster communication and collaboration between the two fields of research.

AB - Diffusion-weighted magnetic resonance imaging (MRI) is sensitive to ensemble-averaged molecular displacements, which provide valuable information on e.g. structural anisotropy in brain tissue. However, a concrete interpretation of diffusion-weighted MRI data in terms of physiological or structural parameters turns out to be extremely challenging. One of the main reasons for this is the multi-scale nature of the diffusion-weighted signal, as it is sensitive to the microscopic motion of particles averaged over macroscopic volumes. In order to analyze the geometrical patterns that occur in (diffusion-weighted measurements of) biological tissue and many other structures, we may invoke tools from the field of stochastic geometry. Stochastic geometry describes statistical methods and models that apply to random geometrical patterns of which we may only know the distribution. Despite its many uses in geology, astronomy, telecommunications, etc., its application in diffusion-weighted MRI has so far remained limited. In this work we review some fundamental results in the field of diffusion-weighted MRI from a stochastic geometrical perspective, and discuss briefly for which other questions stochastic geometry may prove useful. The observations presented in this paper are partly inspired by the Workshop on Diffusion MRI and Stochastic Geometry held at Sandbjerg Estate (Denmark) in 2019, which aimed to foster communication and collaboration between the two fields of research.

UR - http://www.scopus.com/inward/record.url?scp=85102565854&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-56215-1_9

DO - 10.1007/978-3-030-56215-1_9

M3 - Article in proceedings

AN - SCOPUS:85102565854

SN - 9783030562144

T3 - Mathematics and Visualization

SP - 193

EP - 202

BT - Anisotropy Across Fields and Scales

A2 - Özarslan, Evren

A2 - Schultz, Thomas

A2 - Zhang, Eugene

A2 - Fuster, Andrea

PB - Springer

T2 - Workshop on Visualization and Processing of Anisotropy in Imaging, Geometry, and Astronomy, 2018

Y2 - 28 October 2018 through 2 November 2018

ER -