Microstructure Imaging in the Human Cortex: Comparative Systematic Analysis of Biophysical Models
Poster No:
1294
Submission Type:
Abstract Submission
Authors:
Santiago Mezzano1, Quentin Uhl1, Tommaso Pavan1, Hansol Lee2, Jasmine Nguyen-Duc1, Susie Huang2, Ileana Jelescu1
Institutions:
1Department of Radiology, Lausanne University Hospital (CHUV) and University of Lausanne (UNIL), Lausanne, Vaud, 2Department of Radiology, Massachusetts General Hospital, Boston, MA
First Author:
Santiago Mezzano
Department of Radiology, Lausanne University Hospital (CHUV) and University of Lausanne (UNIL)
Lausanne, Vaud
Department of Radiology, Lausanne University Hospital (CHUV) and University of Lausanne (UNIL)
Lausanne, Vaud
Co-Author(s):
Quentin Uhl
Department of Radiology, Lausanne University Hospital (CHUV) and University of Lausanne (UNIL)
Lausanne, Vaud
Department of Radiology, Lausanne University Hospital (CHUV) and University of Lausanne (UNIL)
Lausanne, Vaud
Tommaso Pavan
Department of Radiology, Lausanne University Hospital (CHUV) and University of Lausanne (UNIL)
Lausanne, Vaud
Department of Radiology, Lausanne University Hospital (CHUV) and University of Lausanne (UNIL)
Lausanne, Vaud
Jasmine Nguyen-Duc
Department of Radiology, Lausanne University Hospital (CHUV) and University of Lausanne (UNIL)
Lausanne, Vaud
Department of Radiology, Lausanne University Hospital (CHUV) and University of Lausanne (UNIL)
Lausanne, Vaud
Ileana Jelescu
Department of Radiology, Lausanne University Hospital (CHUV) and University of Lausanne (UNIL)
Lausanne, Vaud
Department of Radiology, Lausanne University Hospital (CHUV) and University of Lausanne (UNIL)
Lausanne, Vaud
Introduction:
Recent advances in gray matter (GM) biophysical models of diffusion aim to enhance specificity in characterizing the human cortex (Jelescu, 2022).
Neurite Exchange Imaging (NEXI) models neurites, extracellular water and the exchange between them, offering valuable insights into unmyelinated GM microstructure (Jelescu, 2022). Another model known as Soma and Neurite Density Imaging (SANDI) (Palombo, 2020), models somas and neurites but neglects exchange. SANDIX is an extension of SANDI that incorporates exchange between neurites and the extracellular space, at the cost of more complex fitting (Olesen, 2022). While the clinical translation of these models is at an early stage, the use of standardized protocols and open-source datasets can facilitate the evaluation of model performance. Here, we use the open-source Connectome Diffusion Microstructure Dataset (CDMD) (Tian, 2022) and the newly developed processing package Gray Matter Swiss Knife (Uhl, 2024) to estimate microstructure parameters in the cortical ribbon using these biophysical models, to evaluate their quality of fit and compare estimates between models, subjects and past works.
Neurite Exchange Imaging (NEXI) models neurites, extracellular water and the exchange between them, offering valuable insights into unmyelinated GM microstructure (Jelescu, 2022). Another model known as Soma and Neurite Density Imaging (SANDI) (Palombo, 2020), models somas and neurites but neglects exchange. SANDIX is an extension of SANDI that incorporates exchange between neurites and the extracellular space, at the cost of more complex fitting (Olesen, 2022). While the clinical translation of these models is at an early stage, the use of standardized protocols and open-source datasets can facilitate the evaluation of model performance. Here, we use the open-source Connectome Diffusion Microstructure Dataset (CDMD) (Tian, 2022) and the newly developed processing package Gray Matter Swiss Knife (Uhl, 2024) to estimate microstructure parameters in the cortical ribbon using these biophysical models, to evaluate their quality of fit and compare estimates between models, subjects and past works.
Methods:
The CDMD (N=26 healthy adults) was retrospectively analyzed. It included T1-weighted anatomical images (1-mm isotropic resolution), and diffusion-weighted images: multi-shell with 8 b-values ranging 50 to 17800 s/mm² (32-64 directions per shell), two diffusion times (Δ = 19 ms and 49 ms), gradient duration δ = 8 ms and at 2-mm isotropic resolution.
Diffusion data were pre-processed (MP-PCA denoising, Gibbs ringing, distortion, motion and eddy current corrections). Four biophysical models (NEXINPA (narrow pulse approximation, assuming δ~0), NEXIWP (wide pulse - accounting for the actual δ duration), SANDI, SANDIX) were fitted voxel-wise with the Gray Matter Swiss Knife package using Nonlinear Least Squares (NLS). We report average estimates in region of interests (ROIs) from the cortical DKT atlas (Tourville, 2012), which were segmented on the T1 image using FastSurfer (Henschel, 2020) and projected onto the diffusion native space using linear registration. To assess the quality of the fit, we compare measured vs. estimated signals as a function of b-value and diffusion time, and compute the corrected Akaike Information Criterion (AICc) for each model.
Diffusion data were pre-processed (MP-PCA denoising, Gibbs ringing, distortion, motion and eddy current corrections). Four biophysical models (NEXINPA (narrow pulse approximation, assuming δ~0), NEXIWP (wide pulse - accounting for the actual δ duration), SANDI, SANDIX) were fitted voxel-wise with the Gray Matter Swiss Knife package using Nonlinear Least Squares (NLS). We report average estimates in region of interests (ROIs) from the cortical DKT atlas (Tourville, 2012), which were segmented on the T1 image using FastSurfer (Henschel, 2020) and projected onto the diffusion native space using linear registration. To assess the quality of the fit, we compare measured vs. estimated signals as a function of b-value and diffusion time, and compute the corrected Akaike Information Criterion (AICc) for each model.
Results:
NEXINPA and NEXIWP yielded parameter estimates aligned with previous findings on the Connectom (Uhl, 2024a) and clinical scanners (Uhl, 2024b), even when derived from only two diffusion times rather than four. NEXIWP estimates for tex and Di were lower than NEXINPA ones (Fig. 1) and Di estimates appeared more biologically plausible, as Di cortical value from NEXINPA was very close to free water diffusivity (~3 μm²/ms). Despite these differences in quantitative estimates, both variants demonstrated similar patterns across the cortex and some clear white matter GM contrast (Fig. 2A).
SANDI and SANDIX parameter estimates were less consistent with a previous SANDI study on clinical scanners (Fig. 1, Fig. 2A) (Schiavi, 2023). In particular, we estimated lower neurite and soma fractions, while the soma radius was likely over-estimated. This suggests that the acquisition protocol (here with two diffusion times instead of a single one) and fitting procedure differences (here with NLS vs Convex Optimization) influence SANDI estimates substantially.
The visual fit quality and AICc which is inversely related to goodness of fit suggested that NEXINPA outperformed all other models (Fig. 2B), followed by SANDI and NEXI NEXIWP. SANDIX showed the poorest fit quality, likely due to the high parameter count and thus more complex fitting.
SANDI and SANDIX parameter estimates were less consistent with a previous SANDI study on clinical scanners (Fig. 1, Fig. 2A) (Schiavi, 2023). In particular, we estimated lower neurite and soma fractions, while the soma radius was likely over-estimated. This suggests that the acquisition protocol (here with two diffusion times instead of a single one) and fitting procedure differences (here with NLS vs Convex Optimization) influence SANDI estimates substantially.
The visual fit quality and AICc which is inversely related to goodness of fit suggested that NEXINPA outperformed all other models (Fig. 2B), followed by SANDI and NEXI NEXIWP. SANDIX showed the poorest fit quality, likely due to the high parameter count and thus more complex fitting.
Conclusions:
This study demonstrated the feasibility of estimating GM biophysical models of exchange using a multi-shell and two-diffusion time protocol, providing proof of concept for future clinical applications with shorter acquisition times.
Modeling and Analysis Methods:
Diffusion MRI Modeling and Analysis 1
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
Cortical Anatomy and Brain Mapping 2
Novel Imaging Acquisition Methods:
Diffusion MRI
Keywords:
Cortex
HIGH FIELD MR
Modeling
MRI
Open Data
Open-Source Code
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
Jelescu, I. O. (2022), ‘Neurite Exchange Imaging (NEXI): A minimal model of diffusion in gray matter with inter-compartment water exchange’, NeuroImage, vol. 256, p. 119277.
Klein, A., & Tourville, J. (2012), ‘101 labeled brain images and a consistent human cortical labeling protocol’, Frontiers in Neuroscience, vol. 6, p. 171.
Olesen, J. L.(2022), ‘Diffusion time dependence, power-law scaling, and exchange in gray matter’, NeuroImage, vol. 251, p. 118976.
Palombo, M.(2020), ‘SANDI: A compartment-based model for non-invasive apparent soma and neurite imaging by diffusion MRI’, NeuroImage, vol. 215, p. 116835.
Schiavi, S.(2023), ‘Mapping tissue microstructure across the human brain on a clinical scanner with soma and neurite density image metrics’, Human Brain Mapping, vol. 44, no. 13, pp. 4792–4811.
Tian, Q. (2022), ‘Comprehensive diffusion MRI dataset for in vivo human brain microstructure mapping using 300 mT/m gradients’, Scientific Data, vol. 9, no. 1, p. 7.
Uhl, Q.(2024), ‘Human gray matter microstructure mapped using Neurite Exchange Imaging (NEXI) on a clinical scanner’, bioRxiv.
Uhl, Q. (2024), ‘Quantifying human gray matter microstructure using neurite exchange imaging (NEXI) and 300 mT/m gradients’, Imaging Neuroscience, vol. 2, pp. 1–19.