Analysis of errors in diffusion kurtosis imaging caused by slice crosstalk in simultaneous multi-slice imaging

Area of Science:

• Magnetic Resonance Imaging
• Diffusion MRI
• Biomedical Engineering

Background:

• Simultaneous multi-slice (SMS) imaging accelerates diffusion MRI acquisition.
• Imperfect slice separation in SMS MRI causes signal leakage, introducing bias.
• This bias can affect diffusion model parameter estimation, particularly for higher-order models like kurtosis.

Purpose of the Study:

• To quantify the effects of slice crosstalk on diffusion kurtosis parameters.
• To evaluate two SMS reconstruction techniques (slice GRAPPA, split-slice GRAPPA) and an alternative acquisition approach.
• To assess the impact of varying slice pairings on diffusion kurtosis metrics.

Main Methods:

• Two SMS reconstruction algorithms (slice GRAPPA, split-slice GRAPPA) were evaluated.
• An alternative acquisition strategy involved randomizing slice pairings across diffusion-weighted images.
• Simulations and volunteer experiments were conducted to analyze diffusion kurtosis metrics.

Main Results:

• Simulation results suggested potential benefits from randomizing slice pairings.
• Experimental factors confounded the advantages of randomized slice pairings in practice.
• Volunteer studies revealed significant metric errors in brain white matter using SMS acquisitions and reconstructions.

Conclusions:

• Slice crosstalk in SMS MRI significantly impacts diffusion kurtosis parameter estimation.
• Current SMS reconstruction techniques and acquisition strategies may not fully mitigate these errors.
• Further research is needed to optimize SMS acquisition and reconstruction for accurate diffusion kurtosis imaging.