Mapping neurite density with Rician noise modeling
Poster No:
1285
Submission Type:
Abstract Submission
Authors:
Naofumi Yoshida1,2, Katsutoshi Murata3, Yuta Urushibata3, Shinsuke Koike4, David Van Essen5, Matthew Glasser5,6, Stamatios Sotiropoulos7,8, Toshinori Hirai2, Hui Zhang9, Takuya Hayashi1,10
Institutions:
1Laboratory for Brain Connectomics Imaging, RIKEN Center for Biosystems Dynamics Research, Kobe, Japan, 2Department of Diagnostic Radiology, Graduate School of Medical Sciences, Kumamoto University, Kumamoto, Japan, 3Siemens Healthcare K.K., Tokyo, Japan, 4Graduate School of Arts and Sciences, The University of Tokyo, Tokyo, Japan, 5Department of Neurobiology, Washington University in St. Louis, St. Louis, MO, 6Department of Radiology, Washington University in St. Louis, St. Louis, MO, 7Sir Peter Mansfield Imaging Centre, School of Medicine, University of Nottingham, Nottingham, United Kingdom, 8Wellcome Centre for Integrative Neuroimaging - Centre for Functional Magnetic Resonance Imaging of the Brain (FMRIB), University of Oxford, Oxford, United Kingdom, 9Centre for Medical Image Computing, UCL Department of Computer Science, University College London, London, United Kingdom, 10Department of Brain Connectomics, Kyoto University Graduate School of Medicine, Kyoto, Japan
First Author:
Naofumi Yoshida
Laboratory for Brain Connectomics Imaging, RIKEN Center for Biosystems Dynamics Research|Department of Diagnostic Radiology, Graduate School of Medical Sciences, Kumamoto University
Kobe, Japan|Kumamoto, Japan
Laboratory for Brain Connectomics Imaging, RIKEN Center for Biosystems Dynamics Research|Department of Diagnostic Radiology, Graduate School of Medical Sciences, Kumamoto University
Kobe, Japan|Kumamoto, Japan
Co-Author(s):
Matthew Glasser
Department of Neurobiology, Washington University in St. Louis|Department of Radiology, Washington University in St. Louis
St. Louis, MO|St. Louis, MO
Department of Neurobiology, Washington University in St. Louis|Department of Radiology, Washington University in St. Louis
St. Louis, MO|St. Louis, MO
Stamatios Sotiropoulos
Sir Peter Mansfield Imaging Centre, School of Medicine, University of Nottingham|Wellcome Centre for Integrative Neuroimaging - Centre for Functional Magnetic Resonance Imaging of the Brain (FMRIB), University of Oxford
Nottingham, United Kingdom|Oxford, United Kingdom
Sir Peter Mansfield Imaging Centre, School of Medicine, University of Nottingham|Wellcome Centre for Integrative Neuroimaging - Centre for Functional Magnetic Resonance Imaging of the Brain (FMRIB), University of Oxford
Nottingham, United Kingdom|Oxford, United Kingdom
Toshinori Hirai
Department of Diagnostic Radiology, Graduate School of Medical Sciences, Kumamoto University
Kumamoto, Japan
Department of Diagnostic Radiology, Graduate School of Medical Sciences, Kumamoto University
Kumamoto, Japan
Hui Zhang
Centre for Medical Image Computing, UCL Department of Computer Science, University College London
London, United Kingdom
Centre for Medical Image Computing, UCL Department of Computer Science, University College London
London, United Kingdom
Takuya Hayashi
Laboratory for Brain Connectomics Imaging, RIKEN Center for Biosystems Dynamics Research|Department of Brain Connectomics, Kyoto University Graduate School of Medicine
Kobe, Japan|Kyoto, Japan
Laboratory for Brain Connectomics Imaging, RIKEN Center for Biosystems Dynamics Research|Department of Brain Connectomics, Kyoto University Graduate School of Medicine
Kobe, Japan|Kyoto, Japan
Introduction:
Rician noise in magnetic resonance imaging (MRI) introduces positive bias in signal measurements. It potentially compromises estimation with multi-compartment models like the neurite orientation dispersion and density (NODDI) model (Zhang et al., 2012) of diffusion MRI. The noise modelling was built in the original NODDI but the efficiency of this modelling has not been thoroughly explored. Here, we investigated the impact of explicit Rician noise modeling for handling low signal–to-noise ratio (SNR) to estimate NODDI parameters using both simulated and in vivo data in human. We also examined how these methods affect the cortical surface mapping of NODDI parameters (Fukutomi et al., 2018).
Methods:
Simulations were performed using the synthetic data for gray matter (GM) (d∥=1.1 μm2⁄ms, diso=3 μm2⁄ms) and white matter (WM) (d∥=1.7 μm2⁄ms, diso=3 μm2⁄ms) conditions. The synthetic data were created using values of fic=0.29, fiso=0.05, and κ=0.17 for GM, and fic=0.6, fiso=0.04, and κ=0.05 for WM and acquisition parameters of b=1000, 2000, 3000 s/mm2, 500 directions of diffusion gradients. Rician noise was added at variable degree with signal to noise ratio (SNR) ranging from 1 to infinity. The probability distribution of signal with Rician noise is modeled by the following equation: p(r|S,σ) = r⁄σ2⋅exp(-(r2+S2)⁄2σ2)⋅I0(rS⁄σ2) where p, r, S, and I0 are probability, measured signal, true signal, standard deviation of Gaussian distribution in the complex plane, and 0th order Bessel function of 1st kind, respectively. We compared the NODDI fitting with and without Rician noise modeling using Maximum Likelihood Estimate (MLE) and Markov Chain Monte Carlo (MCMC) approaches. Cortical surface mapping of NODDI (Fukutomi et al., 2018) was performed for the diffusion MRI from Brain/MINDS-Beyond Human Brain Mapping project (BMB-HBM) (Koike et al., 2021) and Human Connectome Project (HCP) (Van Essen et al., 2013). The effect of the noise modeling in NODDI was also compared with a widely used denoising of diffusion MRI magnitude with Marchenko-Pastur PCA (MPPCA) (Veraart et al., 2016), followed by NODDI without noise modeling. NODDI matlab toolbox (Zhang et al., 2012) and CUDA Diffusion Modeling Toolbox (cuDIMOT) (Hernandez-Fernandez et al., 2019) were used for calculating the NODDI.
Results:
In simulations, NODDI with Rician noise modeling showed optimum estimation of fic in both GM and WM simulations (Fig. 1A,B). The optimality was also confirmed in the orientation dispersion (kappa and orientation index). In addition, the MCMC and MLE showed comparable performance, consistent with a previous report (Hernandez-Fernandez et al., 2019). In cortical surface mapping of the neurite density index (NDI) of NODDI, Rician noise model (Fig. 2B) reduced overestimation biases in the standard NODDI (Fig. 2A, arrows), particularly in low temporal SNR (tSNR) regions (Fig. 2C), such as the ventromedial frontal and ventral temporal cortices in average maps from BMB-HBM data (N=100). The model showed slightly higher correlation of NDI with T1w/T2w myelin map (R=0.77, p<0.01) than in those without the model (R=0.75, p<0.01) across parcellations of Glasser et al., (Glasser et al., 2016) with SNR>17 (Fukutomi et al., 2018). The magnitude-domain MPPCA denoising followed by NODDI did not reduce overestimation bias in the low tSNR areas, as expected from a recent study (Manzano Patron et al., 2024), and resulted in lower correlation with T1w/T2w myelin maps (R=0.57, p<0.01).
·Figure 1
·Figure 2
Conclusions:
NODDI with Rician noise modeling significantly improved the neurite index estimation compared with the standard approach. The model provided a more biologically plausible distribution of the neurite properties in the cerebral cortex than the standard NODDI and magnitude-domain denoising, effectively reducing Rician noise-related biases in challenging brain regions.
Modeling and Analysis Methods:
Bayesian Modeling
Diffusion MRI Modeling and Analysis 1
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
Cortical Anatomy and Brain Mapping
Cortical Cyto- and Myeloarchitecture 2
Keywords:
Cortex
MRI
Myelin
Neuron
Open-Source Software
Other - NODDI, Rician noise, Neurite Density
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.
2. Glasser, M. F., Coalson, T. S., Robinson, E. C., Hacker, C. D., Harwell, J., Yacoub, E., Ugurbil, K., Andersson, J., Beckmann, C. F., Jenkinson, M., Smith, S. M., & Van Essen, D. C. (2016). A multi-modal parcellation of human cerebral cortex. Nature, 536(7615), 171–178.
3. Hernandez-Fernandez, M., Reguly, I., Jbabdi, S., Giles, M., Smith, S., & Sotiropoulos, S. N. (2019). Using GPUs to accelerate computational diffusion MRI: From microstructure estimation to tractography and connectomes. NeuroImage, 188, 598–615.
4. Koike, S., Tanaka, S. C., Okada, T., Aso, T., Yamashita, A., Yamashita, O., Asano, M., Maikusa, N., Morita, K., Okada, N., Fukunaga, M., Uematsu, A., Togo, H., Miyazaki, A., Murata, K., Urushibata, Y., Autio, J., Ose, T., Yoshimoto, J., Araki, T., Brain/MINDS Beyond Human Brain MRI Group (2021). Brain/MINDS beyond human brain MRI project: A protocol for multi-level harmonization across brain disorders throughout the lifespan. NeuroImage. Clinical, 30, 102600.
5. Van Essen, D. C., Smith, S. M., Barch, D. M., Behrens, T. E., Yacoub, E., Ugurbil, K., & WU-Minn HCP Consortium (2013). The WU-Minn Human Connectome Project: an overview. NeuroImage, 80, 62–79.
6. Zhang, H., Schneider, T., Wheeler-Kingshott, C. A., & Alexander, D. C. (2012). NODDI: practical in vivo neurite orientation dispersion and density imaging of the human brain. NeuroImage, 61(4), 1000–1016.
7. Veraart, J., Fieremans, E., & Novikov, D. S. (2016). Diffusion MRI noise mapping using random matrix theory. Magnetic resonance in medicine, 76(5), 1582–1593.
8. Manzano-Patron, J. P., Moeller, S., Andersson, J. L. R., Ugurbil, K., Yacoub, E., & Sotiropoulos, S. N. (2024). Denoising diffusion MRI: Considerations and implications for analysis. Imaging Neuroscience, 2, 1–29.