Effect of Data Acquisition and Analysis Method on Fiber Orientation Estimation in Diffusion MRI

Area of Science:

• Neuroimaging
• Diffusion MRI analysis
• White matter tractography

Background:

• Diffusion MRI enables non-invasive visualization of white matter microstructure.
• Accurate fiber orientation estimation is crucial for understanding brain connectivity.
• Evaluating different q-space sampling strategies and analysis models is essential for optimizing Diffusion MRI.

Purpose of the Study:

• To investigate the impact of single-shell q-space sampling on fiber orientation estimation.
• To compare the performance of various multi-fiber analysis methods in Diffusion MRI.
• To identify optimal methods for accurate white matter tractography.

Main Methods:

• Developed a simulation based on in-vivo Diffusion MRI data.
• Compared "ball-and-stick" model, constrained spherical deconvolution, and generalized Fourier transform approaches.
• Evaluated methods across varying numbers of angular diffusion-weighted samples (N=20-120) at b=1000s/mm² and SNR=18.

Main Results:

• Methods differed significantly in their ability to detect white matter fibers.
• The "ball-and-stick" model and spherical deconvolution demonstrated the lowest orientation error.
• These two methods achieved the highest rate of successful fiber detection.

Conclusions:

• The "ball-and-stick" model and spherical deconvolution are highly effective for fiber orientation estimation in Diffusion MRI.
• These methods offer improved accuracy and detection rates compared to other evaluated techniques.
• Findings support the use of these models for robust white matter analysis in clinical settings.