Optimizing the intrinsic parallel diffusivity in NODDI: An extensive empirical evaluation.

Area of Science:

• Neuroimaging
• Biomedical Engineering
• Computational Neuroscience

Background:

• Neurite Orientation Dispersion and Density Imaging (NODDI) is a key tool for quantifying brain microstructure.
• NODDI assumes a fixed neurite intrinsic parallel diffusivity (d∥) of 1.7 μm²⋅ms⁻¹, an assumption with unexplored implications.
• This study critically evaluates the impact of this fixed d∥ value on NODDI model performance.

Purpose of the Study:

• To investigate the optimality of the default d∥ = 1.7 μm²⋅ms⁻¹ in the NODDI model.
• To assess the influence of varying d∥ on microstructural parameter estimation across different age groups, sexes, and tissue types.
• To determine if alternative d∥ values enhance NODDI model accuracy.

Main Methods:

• Model residuals were used to assess d∥ optimality across a range of 0.5–3.0 μm²⋅ms⁻¹.
• Analyses were conducted across white matter and gray matter, varying age (infancy to adulthood), sex, and diffusion-weighting protocols.
• The sensitivity of estimated NODDI parameters to d∥ variations was also examined.

Main Results:

• The default d∥ = 1.7 μm²⋅ms⁻¹ is optimal for adult white matter but suboptimal for gray matter, where lower values are preferred.
• Significantly lower optimal d∥ values were found for both white and gray matter in infants.
• Diffusion protocol influenced optimal d∥ minimally, with no significant sex-based differences observed.

Conclusions:

• The standard d∥ value of 1.7 μm²⋅ms⁻¹ in NODDI is not universally optimal, particularly for gray matter and infant brain tissues.
• Adjusting d∥ based on tissue type and age is crucial for accurate microstructural parameter estimation.
• The findings necessitate re-evaluation of the fixed d∥ assumption in NODDI analyses for specific populations.