Mapping the impact of nonlinear gradient fields with noise on diffusion MRI

Area of Science:

• Medical Imaging
• Neuroimaging
• Diffusion MRI

Background:

• Gradient nonlinearities in diffusion MRI introduce spatial distortions.
• These distortions can bias diffusion tensor information and tractography results.
• The impact of these distortions in the presence of noise requires further investigation.

Purpose of the Study:

• To investigate the impact of gradient nonlinearity correction in diffusion MRI under varying noise levels.
• To compare voxel-wise gradient table correction with direct signal scaling methods.
• To assess the effect of correction on key diffusion metrics.

Main Methods:

• Empirically derived gradient nonlinear fields were introduced at different signal-to-noise ratio (SNR) levels.
• Two experiments were conducted: tensor field simulation and brain simulation.
• Correction techniques evaluated included voxel-wise gradient table correction and direct signal scaling.
• Impact assessed using diffusion metrics: mean diffusivity (MD), fractional anisotropy (FA), axial diffusivity (AD), radial diffusivity (RD), and principal eigenvector (V1).

Main Results:

• Gradient nonlinearity correction does not introduce significant errors in linear systems.
• Correction does not adversely interact with noise in diffusion MRI data.
• Nonlinearity correction effectively mitigates distortions in typical SNR data.
• Both correction techniques performed similarly at SNR levels below 30.
• Greater impact of correction observed in regions with pronounced nonlinearities.

Conclusions:

• Gradient nonlinearity correction offers more benefits than adverse effects in diffusion MRI.
• Correction is recommended for analyses involving regions of interest with significant nonlinearities.
• The findings support the routine application of gradient nonlinearity correction in diffusion MRI processing.