Evaluation of fiber tracking from subsampled q-space data in diffusion spectrum imaging

Area of Science:

• Neuroimaging
• Diffusion Spectrum Imaging (DSI)
• Computational Neuroscience

Background:

• Diffusion spectrum imaging (DSI) enables detailed mapping of complex white matter architecture, including crossing and touching fiber bundles within a single voxel.
• Acquiring comprehensive DSI data requires extensive q-space sampling, leading to prolonged scan times.
• Exploiting q-space symmetry offers a potential strategy to reduce DSI acquisition duration without compromising data quality.

Purpose of the Study:

• To investigate the impact of subsampling q-space data on diffusion spectrum imaging (DSI) tractography accuracy.
• To determine the minimum number of q-space samples required for reliable fiber pathway reconstruction.
• To evaluate the trade-off between reduced scan time and tractography fidelity.

Main Methods:

• Comparison of fiber pathways for five major white matter tracts across three subsampled DSI datasets (derived from 257 q-space samples).
• Quantitative analysis using the coefficient of variation to assess the effect of subsampling on tractography results across 10 subjects.
• Evaluation of the influence of threshold angles on tractography outcomes and the orientation distribution function (ODF).
• Adaptation of a streamline tractography method to mitigate curvature issues and local truncation errors.

Main Results:

• Subsampling DSI data to 129 and 198 q-space points yielded results highly similar to the full dataset when using a threshold angle between 41° and 45°.
• Quantitative and qualitative analyses indicated that reduced q-space sampling (129 points) is sufficient for accurate fiber pathway mapping.
• The study identified optimal threshold angles for tractography with subsampled data.

Conclusions:

• Diffusion spectrum imaging (DSI) tractography can be effectively performed using significantly reduced q-space sampling, specifically 129 points.
• This reduction in sampling offers a substantial advantage in scan time while maintaining high fidelity in fiber pathway reconstruction.
• The findings suggest that 129 q-space points are adequate for reliable DSI tractography, balancing efficiency and accuracy.