MICRA: Microstructural image compilation with repeated acquisitions.

Area of Science:

• Neuroimaging
• Biomedical Engineering
• Radiology

Background:

• Limited availability of datasets for simultaneous test-retest reliability assessment of multiple microstructural MRI contrasts.
• Need for robust data to validate advanced neuroimaging analysis pipelines.

Purpose of the Study:

• To provide a unique, multi-contrast microstructural MRI dataset for assessing test-retest reliability.
• To serve as a resource for the neuroimaging community to validate analysis methods and estimate statistical power.

Main Methods:

• Acquisition of multi-contrast microstructural MRI data (diffusion, relaxometry, quantitative magnetization transfer) on an ultra-strong gradient 3T Connectom scanner.
• Collection of 5 repeated scan sets from 6 healthy participants.
• Computation of raw data and microstructure maps.

Main Results:

• Demonstration of high reproducibility for several microstructural MRI measures across scan sessions using intra-class correlation coefficients and coefficients of variation.
• Calculation of a priori sample sizes for statistical power across different white matter pathways and microstructure measures.
• Voxel-wise repeatability assessment of microstructural maps across the whole brain white matter.

Conclusions:

• The MICRA dataset exhibits high reproducibility, supporting its utility for reliability testing and power estimations in neuroimaging research.
• This resource will benefit researchers evaluating the robustness of their analysis pipelines and planning future studies.
• The dataset facilitates specialized analyses to inform current and future research in brain microstructure.