Measuring Strain in Diffusion-Weighted Data Using Tagged Magnetic Resonance Imaging.

Area of Science:

• Medical Imaging
• Biomechanics
• Computational Anatomy

Background:

• Accurate in vivo strain measurement in deforming organs is crucial for motion analysis using medical imaging.
• Dynamic magnetic resonance (MR) imaging provides motion data but lacks intrinsic fiber direction information.
• Diffusion-weighted MR imaging offers fiber tractography but yields static images incompatible with dynamic data.

Purpose of the Study:

• To develop and validate an algorithm workflow for estimating tissue strain within the diffusion MR space by correlating tagged dynamic MR images.
• To analyze human tongue deformations during speech and protrusion using this novel method.

Main Methods:

• A deep neural network segmented synthetic cine dynamic MR sequences from tagged data to define tongue shapes.
• Spatiotemporal motion fields were derived from tagged MR analysis.
• Diffeomorphic registration transformed motion fields into diffusion MR space for strain computation along muscle fiber directions.

Main Results:

• The algorithm successfully estimated strain values in the diffusion MR space for human tongue deformations.
• Strain patterns were analyzed in seven internal tongue muscles across 78 time volumes during speech and protrusion.
• Compression and stretching patterns revealed unique muscle behaviors and potential activation patterns.

Conclusions:

• This method integrates dynamic and diffusion MR imaging to compute muscle-specific strain, overcoming limitations of individual techniques.
• The findings provide insights into tongue muscle mechanics during speech and voluntary movements.
• This approach has potential applications in speech pathology and understanding organ biomechanics.