Estimation of Bounded and Unbounded Trajectories in Diffusion MRI

Area of Science:

• Biophysics
• Neuroimaging
• Computational Neuroscience

Background:

• Diffusion MRI (dMRI) is crucial for brain tissue microstructure analysis.
• Understanding the relationship between dMRI signals and acquisition parameters requires analyzing the autocorrelation function of diffusing spins.
• Current models may not fully capture diffusion in restricted cellular environments.

Purpose of the Study:

• To develop a model for the autocorrelation function of diffusion in restricted spaces.
• To utilize the multivariate Ornstein-Uhlenbeck process for modeling bounded diffusion trajectories.
• To establish a framework for extracting brain tissue-specific measures from dMRI data.

Main Methods:

• Modeling the matrix-valued exponential autocorrelation function using the multivariate Ornstein-Uhlenbeck process.
• Analyzing the relationship between model parameters and apparent axon radius.
• Developing a frequency-domain model for dMRI signals.
• Applying a two-compartment model (bounded and unbounded diffusion) to ex-vivo monkey brain data.

Main Results:

• Demonstrated that diffusion autocorrelation in restricted spaces can be approximated by exponential functions.
• Provided a detailed analysis linking model parameters to time-dependent apparent axon radius.
• Presented a general frequency-domain model for dMRI signals.
• Successfully modeled diffusion in the corpus-callosum of a monkey brain.

Conclusions:

• The Ornstein-Uhlenbeck process effectively models diffusion in bounded spaces, crucial for dMRI.
• This approach enhances the extraction of microstructural information from dMRI.
• The developed model offers a novel perspective for analyzing complex diffusion processes in biological tissues.