Comparison of modelled diffusion-derived electrical conductivities found using magnetic resonance imaging

Area of Science:

• Neuroimaging
• Biophysics

Background:

• Magnetic resonance (MR)-based electrical conductivity imaging is a novel contrast mechanism for disease diagnosis.
• Conductivity tensor imaging (CTI) reconstructs low-frequency conductivity images using MR diffusion microstructure imaging data.

Purpose of the Study:

• To investigate the variability in conductivity predictions across different diffusion microstructure models.
• To evaluate the alignment of these predictions with experimental observations.

Main Methods:

• Extracted microstructure parameters from public diffusion databases for three models: Neurite Orientation Dispersion and Density Imaging (NODDI), Soma and Neurite Density Imaging (SANDI), and Spherical Mean technique (SMT).
• Calculated conductivity predictions for gray matter (GM) and white matter (WM) tissues using each model.
• Assessed the range and bilateral consistency of predicted conductivities.

Main Results:

• Significant variability in conductivity estimates was observed across the NODDI, SANDI, and SMT models.
• Each model predicted distinct conductivity values for GM and WM tissues.
• SMT provided estimates closer to experimental values, though no model aligned with spectroscopic models.

Conclusions:

• Substantial discrepancies exist in tissue conductivity estimates from different diffusion models, posing challenges for CTI model selection.
• SMT shows better alignment with experimental results, but other models might offer superior tissue discrimination.