J Stroke Cerebrovasc Dis. 2019 Apr;28(4):1113-1125.

Unraveling Specific Brain Microstructural Damage in Moyamoya Disease Using Diffusion Magnetic Resonance Imaging and Positron Emission Tomography

Shoko Hara, Masaaki Hori, Ryo Ueda, Shihori Hayashi, Motoki Inaji, Yoji Tanaka, Taketoshi Maehara, Kenji Ishii, Shigeki Aoki, Tadashi Nariai.

Correspondence to: Shoko Hara; Department of Neurosurgery, Tokyo Medical and Dental University; Department of Radiology, Juntendo University, Tokyo Japan.

Email: shara.nsrg@tmd.ac.jp

Research gate: https://www.researchgate.net/profile/Shoko_Hara

ORCiD: https://orcid.org/0000-0003-1097-2738

Publons: https://publons.com/author/1354354/shoko-hara#profile



Background and Purpose- Chronic ischemia may induce brain microstructural damage and lead to neurocognitive dysfunction in patients with Moyamoya disease. We applied neurite orientation dispersion and density imaging (NODDI) and 15O-gas positron emission tomography (PET) to elucidate the specific ischemic brain microstructural damage of MMD in the cortex and the white matter.

Materials and Methods- Thirty-one patients (16-63years old, 9 males) and 20 age- and sex-matched normal controls were enrolled in this study. NODDI evaluates quantitative parameters reflecting neurite and axonal density, network complexity and the interstitial fluid in all participants. Of 31 patients, 12 newly diagnosed patients were evaluated with PET, also. We evaluated correlations between the microstructural parameters of NODDI and the hemodynamic and metabolic parameters of PET, the relationship between NODDI and clinical severity of each hemisphere (Normal, Asymtpomatic, Symptomatic, and Infarcted) as well as neurocognitive performance.

Results- All NODDI parameters significantly correlated with PET parameters (absolute r = 0.46-0.83, P ≤ .048) and clinical severity (P < .001), suggesting that neurite and axonal density and network complexity decreased, and the interstitial fluid increased, as the ischemic burden became severe. NODDI parameters reflecting neurite and axonal density and network complexity significantly correlated with neurocognitive profiles (r = 0.36-0.64, P ≤ .048), but the interstitial fluid component did not.

Conclusions- Chronic ischemia in patients with Moyamoya disease may induce decreased neurite and axonal density, simplified network complexity, and may lead to neurocognitive dysfunction. The increased interstitial fluid accompanying hemodynamic impairment may not be identical to the decreased neurite density and might be driven by another mechanism.

Keywords: Moyamoya disease; ischemia; cerebral blood flow and metabolism; diffusion MRI; magnetic resonance imaging; cognition; cognitive impairment


Supplemental Information

In this study, to evaluate microstructure of patients with Moyamoya disease, we used neurite orientation dispersion and density imaging (NODDI)1, a model-based analytical method for multishell diffusion MRI that assesses quantitative parameters directly reflecting brain microstructures. In the NODDI model, three microstructural environments are distinguished (intracellular, extracellular, and cerebrospinal fluid compartments), and the Watson distribution is adopted to model highly dispersed neuritic structures, such as dendritic trees in the gray matter. Using NODDI analysis, we can obtain three parameters that directly reflects brain microstructures:  intracellular volume fraction (Vic) represents the density of axons and dendrites based on intracellular diffusion; the orientation dispersion index (OD) represents the dispersion of axons and dendrites in the intracellular component; and isotropic volume fraction (Viso) reflects the interstitial fluid component in the brain parenchyma. As shown in Fig. 1, NODDI-derived microstructural parameters significantly correlated with hemodynamic and metabolic parameters obtained with 15O-gas positiron emission tomography (PET).



Fig. 1. A representative case of a patient with unilateral Moyamoya disease (reprinted from Fig. 1 in this manuscript). As the PET parametrical images of the lower column suggested, this patient suffers characteristic misery perfusion in the left hemisphere. In this affected hemisphere, the orientation dispersion index (OD) that represents network complexity is decreased, and isotropic volume fraction (Viso) that reflects parenchymal water component is increased (arrow). The decrease of the intracellular volume fraction (Vic) in the white matter that reflects axonal density is also suggested (arrow).


We found that OD values in the cortex may reveal novel characteristics of brain microstructures that conventional diffusion MRI has failed to recognize. Because of the reported decrease of fractional anisotropy (FA) obtained by traditional diffusion tensor imaging (DTI) and negative correlation between FA and OD, we presumed OD would be increased in patients with Moyamoya disease, as the randomized, disorganized and highly disrupted network structure as a consequence of chronic ischemic damage. However, we found OD is decreased in patients with Moyamoya disease, and OD show negative correlation between the severities of ischemic burden. Moreover, cortical OD values shows the strongest correlation between CBF (r=0.73; P<.001) and MTT (r=-0.82; P<.001) that was superior to traditional DTI parameters such as FA (Supplemental Table IV available at https://ars.els-cdn.com/content/image/1-s2.0-S1052305718307389-mmc1.docx).  The decrease of OD might reflect the simplified network complexity, such as the shrinkage in dendritic processes observed in chronic ischemic animals2.


In the white matter, we found that Vic that reflects axonal density showed strongest correlation between PET-measured hemodynamic parameters, and that white matter Vic showed strongest correlation with neurocognitive performance, especially the Processing Speed Index (r=0.64; P<.001). Therefore, in the subsequent study, patients with Moyamoya disease without significant brain damage were evaluated using whole-brain voxel analysis to clarify the anatomical and parametrical correlation between NODDI parameters and neurocognitive function3. The strongest correlation with neurocognitive function was observed in Vic in the white matter, especially in the posterior part of the brain (available at https://doi.org/10.1161/STROKEAHA.118.022367).


Interestingly, the Viso that reflects parenchymal water component, correlated well with the hemodynamic parameters of PET as well as the Vic and OD but not at all with neurocognitive function. These differences suggest that the increase in Viso is not identical to the decrease in Vic and not due to the decrease of dendrite and axonal density. A possible and attractive explanation is that the increased parenchymal water is due to a dysfunction in the glymphatic system, which is driven by arterial pulsatility4 and is speculated to be impaired in patients with Moyamoya disease. In some patients, we actually observed the increase of Viso has been subsided after the restoration of hemodynamic impairment (Fig. 2, modified from https://doi.org/10.2463/mrms.ci.2018-0088).



Fig. 2. Postoperative decrease of isotropic volume fraction (Viso) that reflects parenchymal water component in a patient with Moyamoya disease (modified from Fig. 1 in Reference 3). After the successful indirect bypass surgery that dramatically improved the decreased cerebral blood flow in the right hemisphere, the increase of Viso, has completely disappeared. (Please note that the Viso images shown here are smoothed images as shown in Fig. 1).


In addition to the neurocognitive performance evaluated in the studies described above, we also evaluated the relationship between Big Five personality trait and microstructural parameters of NODDI in 29 patients with Moyamoya disease and 26 normal controls6. The effect of Moyamoya disease to the personality has been rarely reported, and the relationship with brain microstructure has never been investigated before. By this evaluation, we found significant positive correlation between Openness to Experience and the OD reflecting network complexity in the right posterior part of the brain. Our results might suggest the chronic ischemic microstructural damage also affect the personality of the patients with Moyamoya disease (available at http://dx.doi.org/10.26044/ecr2019/C-0038).


Currently, as a new project, we are working on myelin imaging using magnetization transfer saturation (MTsat) method on patients with Moyamoya disease, and postoperative changes of NODDI parameters in patients with Moyamoya disease. So far, we found that patients with Moyamoya disease might have myelin damage as well as axonal damage compared to normal controls, and the myelin content might correlated with neurocognitive tests7. Also, we found the increase of parenchymal water component (Viso) is reversible5, but the postoperative change was less unclear in other parameters (OD and Vic). Further evaluation is required to establish the clinical significance of microstructural evaluation using NODDI in patients with Moyamoya disease.



I, Shoko Hara, the project leader of microstructural imaging of Moyamoya disease, would like to thank all coauthors and coworkers who always help my research. In addition to the researchers shown in Fig. 3, I would like to thank Syo Murata, Ryo Ueda, Yuuki Takenaka, Masami Goto for analytical guidance, all members of the Juntendo Neuro meetings for always inspiring me and giving me ideas of research, staffs of Tokyo Medical Clinic and Tokyo Metropolitan Institute of Gerontology for image acquisition, Masako Akiyama for statistical guidance, and Maki Mukawa for gene analysis.




Fig. 3. Researchers and cooperators regarding this research project.



  1. Zhang H, Schneider T, Wheeler-Kingshott C, Alexander NODDI: practical in vivo neurite orientation dispersion and density imaging of the human brain. Neuroimage. 2012; 61:1000–1016.
  2. Farkas, P.G. Luiten, F. Bari. Permanent, bilateral common carotid artery occlusion in the rat: a model for chronic cerebral hypoperfusion-related neurodegenerative diseases. Brain Res Rev. 2007;54:162-180
  3. Hara S, Hori K, Aoki S, Nariai T et al. Microstructural Damage in Normal-Appearing Brain Parenchyma and Neurocognitive Dysfunction in Adult Moyamoya Disease. 2018;49:2504–2507. https://doi.org/10.1161/STROKEAHA.118.022367
  4. J. Iliff, M. Wang, D.M. Zeppenfeld, et al.Cerebral arterial pulsation drives paravascular CSF-interstitial fluid exchange in the murine brain. J Neurosci. 2013;33:18190-18199
  5. Hara S, Hori K, Aoki S, Nariai T et al. Regression of White Matter Hyperintensity after Indirect Bypass Surgery in a Patient with Moyamoya Disease. Magn Res Med Sci. 2018. https://doi.org/10.2463/mrms.ci.2018-0088
  6. Hara S, Hori K, Aoki S, Nariai T. Microstructural correlates of personality in patients with Moyamoya disease measured by neurite orientation dispersion and density imaging. ECR 2019. E-poster: C-0038.  http://dx.doi.org/10.26044/ecr2019/C-0038
  7. Hara S, Hori K, Aoki S, Nariai T et al. Myelin imaging may reveal microstructural damage caused by chronic ischemia correlated with neurocognitive dysfunction in patients with moyamoya disease. Joint Annual Meeting ISMRM-ESMRMB. Electronic Poster Session Neuro: 4806. https://www.ismrm.org/18/program_files/EP17.htm#sub1