A Jones matrix formalism for simulating three-dimensional polarized light imaging of brain tissue

TL;DR

This paper compares macroscopic and microscopic Jones matrix models for simulating 3D polarized light imaging of brain tissue, demonstrating improved fiber orientation estimation by considering myelin density.

Contribution

It introduces a Jones matrix formalism that incorporates microscopic fiber models and myelin density, enhancing simulation accuracy for 3D-PLI.

Findings

Macroscopic model reliably estimates fiber orientations at lower resolutions.

Higher resolution polarimeters can accurately resolve fiber bundles.

Considering myelin density significantly improves fiber orientation accuracy.

Abstract

The neuroimaging technique three-dimensional polarized light imaging (3D-PLI) provides a high-resolution reconstruction of nerve fibres in human post-mortem brains. The orientations of the fibres are derived from birefringence measurements of histological brain sections assuming that the nerve fibres - consisting of an axon and a surrounding myelin sheath - are uniaxial birefringent and that the measured optic axis is oriented in direction of the nerve fibres (macroscopic model). Although experimental studies support this assumption, the molecular structure of the myelin sheath suggests that the birefringence of a nerve fibre can be described more precisely by multiple optic axes oriented radially around the fibre axis (microscopic model). In this paper, we compare the use of the macroscopic and the microscopic model for simulating 3D-PLI by means of the Jones matrix formalism. The simulations show that the macroscopic model ensures a reliable estimation of the fibre orientations as long as the polarimeter does not resolve structures smaller than the diameter of single fibres. In the case of fibre bundles, polarimeters with even higher resolutions can be used without losing reliability. When taking the myelin density into account, the derived fibre orientations are considerably improved.