Related Experiment Video
Updated: Oct 18, 2025
17:06
Co-analysis of Brain Structure and Function using fMRI and Diffusion-weighted Imaging
Published on: November 8, 2012
26.5K
Structural Brain Network Reproducibility: Influence of Different Diffusion Acquisition and Tractography
Pasquale Borrelli1, Carlo Cavaliere1, Marco Salvatore1
1IRCCS SDN, Napoli, Italy.
Brain Connectivity
|October 4, 2021
Summary
Brain network reproducibility varies with magnetic resonance imaging and tractography methods. Constrained spherical deconvolution with deterministic tractography offers high reproducibility for brain graph metrics.
Area of Science:
- Neuroimaging
- Computational Neuroscience
- Network Science
Background:
- Graph metrics analyze brain network topology.
- Variability in magnetic resonance imaging (MRI) acquisition and tractography reconstruction techniques affects results.
- Reproducibility of structural connectome graph metrics needs further characterization.
Purpose of the Study:
- To evaluate the influence of different diffusion MRI acquisition schemes, diffusion models, and tractography reconstruction approaches on the reproducibility of graph metrics.
- To assess the impact of these variations on both global and local graph metrics using test/retest data.
- To quantify reproducibility using intraclass correlation coefficient (ICC) and percentage relative standard deviation (pRSD).
Main Methods:
- Utilized Human Connectome Project test/retest data.
- Compared single and multishell diffusion acquisition schemes.
- Evaluated tensor and spherical deconvolution diffusion models.
- Assessed deterministic and probabilistic tractography reconstruction approaches.
- Calculated global and local graph metrics and their ICC.
Main Results:
- Different combinations of acquisition schemes, diffusion models, and tractography algorithms significantly affect graph metric reproducibility.
- Constrained spherical deconvolution (CSD) with deterministic tractography yielded high reproducibility (ICCs >0.75) and low pRSD.
- Probabilistic CSD with high b-value demonstrated the highest reproducibility.
- Streamline selection filters in CSD notably impacted reproducibility.
Conclusions:
- Test/retest reproducibility of brain network graph metrics is generally high but sensitive to the chosen acquisition and reconstruction pipeline.
- Findings impact the selection of MRI protocols and processing pipelines for reproducible brain network studies.
- Optimizing these choices is crucial for reliable topological analyses of brain networks.
