Single-step nonlinear diffusion tensor estimation in the presence of microscopic and macroscopic motion

Area of Science:

• Medical Imaging
• Neuroimaging
• Biophysics

Background:

• Patient motion introduces significant artifacts in diffusion tensor imaging (DTI), compromising the reliability of diffusion tensor estimation.
• Existing motion correction methods primarily address minor physiological movements, leaving a gap in correcting for gross patient motion.

Purpose of the Study:

• To develop a general mathematical framework for correcting gross patient motion in multishot, multicoil DTI scans.
• To account for both rotational and translational motion and changes in diffusion-encoding direction.

Main Methods:

• A novel signal model was developed incorporating rotational and translational patient motion.
• A nonlinear least-squares formulation was derived from the signal model.
• Diffusion tensors were estimated using a nonlinear conjugate gradient algorithm.

Main Results:

• The proposed algorithm demonstrated superior performance in correcting for gross motion compared to conventional tensor estimation methods.
• Validation was performed using both phantom simulations and in vivo studies.
• The method effectively mitigates artifacts caused by significant patient movement.

Conclusions:

• The developed framework provides a robust solution for correcting gross patient motion in DTI.
• This advancement enhances the reliability and accuracy of diffusion tensor estimation in challenging imaging conditions.
• The algorithm has significant implications for clinical DTI applications where motion is a concern.