Technological Development of 7T MRI
7T MRI Scan Environment
In comparison with conventional MRI scans, 7T MRI scans provide higher soft-tissue contrast and sensitivity. However, unlike 3T or lower MRI scans, 7T MRI requires special attention because it is sensitive to background magnetic field and transmission inhomogeneity. To address this issue, the 7T MRI measurement environment was optimized, and a basic standard operating procedure applicable to a wide variety of experiments was formulated. Building on these efforts, appropriate measurement systems were established for use in the following projects: Brain Mapping by Integrated Neurotechnologies for Disease Studies (Brain/MINDS), Strategic Research Program for Brain Sciences (SRPBS [Nou Puro]) subarea entitled “Research on the Nervous System for Decision-Making and Behavioral Selection That Enables Flexible Adaptation to the Environment” (hereinafter referred to as “Decision Making”), and bidirectional collaborative research. In structural brain imaging, a large number of radiofrequency pulse train is frequently employed. Therefore, transmit field (B1+) inhomogeneity creates significant technical problems. To address these issues, magnetization-prepared 2 rapid gradient echo (MP2RAGE) sequences were introduced, and three-dimensional T1-weighted imaging with a spatial resolution of 0.75 × 0.75 × 0.75 mm3 was adopted as the standard protocol. For fMRI scans, adequate spatial resolutions (1–2 mm) should be achieved to allow analysis of gray matter thickness of several millimeters. Such high resolution is necessary to achieve segmentation along the cortical depth (layers) for segregating input and output activity and to prevent artifacts from large veins running along the cortical surface. Consequently, the multiband technique was adopted and optimized. This enabled whole-brain scans at a spatial resolution of 1.2 × 1.2 ×1.2 mm3 and a temporal resolution of approximately 2.0 s.
The echo-planar imaging (EPI), which is widely employed in 7T fMRI, is susceptible to image distortions, and therefore requires correction by appropriate image registration. High-resolution scans suffer from geometric distortions and other flaws. Conventional post-processing and analytical procedures cannot remove those artifacts, and therefore fail to take advantage of the high resolution. In order to overcome these challenges and optimize the high-resolution 7T MRI scan, the Human Connectome Project (HCP) minimal preprocessing pipelines developed by the Washington University and the University of Minnesota were introduced into the cluster systems of our laboratory. As a result, a maximum displacement of 15 mm (Euclidean distance) was successfully corrected in the orbitofrontal cortex, inferior temporal lobes, and other areas susceptible to large magnetic field inhomogeneity. Regarding surface-based analysis, extracortical signals arising from mis-registration between the functional and structural scans were reduced by 22.3% in the central sulcus, visual cortex, and other regions even without severe geometric distortions. In the absence of appropriate registration, high-precision 7T MRI scan data may lead researchers to inaccurate conclusions. Given the increasing popularity and spatial precision of surface-based mapping in recent years, the importance of appropriate correction techniques cannot be overstated because small localization errors may be wrongly judged as meaningful differences. At present, our standard surface-based mapping pipeline can be applied to 7T fMRI and diffusion MRI measurements.
High-Resolution MRI Research
Using the measurement systems mentioned above, 1.2-mm isotropic fMRI, 1.05-mm isotropic diffusion MRI, 200-μm in-plane resolution quantitative susceptibility mapping, and cortical myelin mapping based on T1- and T2-weighted MRI were performed. When the relative myelin content estimated from MRI images was compared with individual digit areas of fMRI in the primary somatosensory cortex (S1), fine-grained digit representations were successfully obtained in individual participants’ S1; consequently, there was no need to apply the group level analysis. The pattern of finger localization identified in individual participants showed an excellent inter-subject reproducibility. Notably, the second digit area was located above the boundary of myelin gradient in most participants.
Presently, activities are in progress to achieve a submillimeter resolution fMRI measurement and monitoring neural activities in different cortical layers (depths). The findings reported above are the results of joint research with Prof. Essa Yacoub of the University of Minnesota (Minneapolis, MN), and Prof. Matt Glasser of Washington University (St. Louis, MO). In this study, Project Assistant Prof. Sho Sugawara played a leading role. These results were presented at an international academic conference. Follow-up experiments are underway as part of the following Japan Agency for Medical Research and Development (AMED) projects: Brain/MINDS (Kakushin Nou), SRPBS (Nou Puro), and Brain/MINDS Beyond (Kokusai Nou).
MR Spectroscopy Research
Increasing the static magnetic field strength improves the sensitivity and spectral resolution of the magnetic resonance spectroscopy (MRS). Therefore, the 7T MRS has the potential to differentiate adjacent metabolite peaks that the 3T MRS cannot. The frequency resolution necessary to detect differences in the chemical shifts of metabolites is the highest in the direction of frequency encoding of spatial position. Therefore, in order to adjust for the resulting deviations in spatial metabolite distribution, chemical shift imaging (CSI) technique was introduced. Currently, efforts are ongoing to apply the CSI technique to the measurement of N-acetyl aspartate, gamma-aminobutyric acid (GABA), glutamic acid, and other neurochemical markers. In addition, a novel MRS measurement technique for monitoring glucose levels has been developed in cooperation with Siemens K.K.. In contrast to the conventional positron emission tomography (PET) approaches, in which an extrinsic marker of glucose uptake is administered, this technique enables identification of intrinsic glucose molecules in local brain tissue.
MRI Research Using Japanese Monkeys
Resting-State fMRI Measurement of Biomarkers for Functional Recovery
It remained unknown how the large-scale brain circuit of the macaque monkey changes during functional recovery following corticospinal tract injury. In this study using 3T-MRI (Allegra), macaque monkeys were subjected to spinal cord injury, and the process of large-scale circuit reorganization associated with functional recovery was investigated using resting-state MRI under anesthesia. In cooperation with Siemens, we developed a novel MRI pulse sequence protocol that optimizes resting-state fMRI, diffusion MRI, and high-resolution structural MRI measurements of the macaque monkey. This protocol included in-house modified pulse sequence for correction of image distortions. In addition, optimal anesthetic conditions were investigated using healthy monkeys, and we identified the isoflurane anesthesia parameters that ensured stable resting-state brain activity. Using these pulse sequence and anesthetic conditions, 16 healthy monkeys were subjected to MRI, and the functional brain network was identified with high intersession, intrasession, and interindividual reproducibility.
Next, six Japanese monkeys trained on a precision grip task underwent unilateral two-thirds resection of the vertebral column below the cervical level to make a spinal cord injury model. Their hand motor function was evaluated every week from baseline through 4 months post-surgery. These animals also underwent resting-state fMRI, diffusion MRI, and structural MRI scans under anesthesia. Three of them were administered anti-repulsive guidance molecule neutralizing antibodies to accelerate functional recovery (the “antibody” group).
Functional recovery was evaluated based on the success rate and time to precision grip task. After the surgery, the antibody group of monkeys exhibited a tendency to perform the first grip motion at an earlier time point than the control group. The antibody group shortened the time to precision grip success at a higher rate than the control group, and their time to precision grip success was significantly shorter at 4 months post-surgery. Resting-state fMRI scans of the antibody-group monkeys revealed an increase in the network centrality measure (Eigenvector centrality) in the parietal lobe and surrounding regions during the period of rapid early-stage recovery. This measure remained high in the monkeys with sustained functional recovery, and was considered as a candidate parameter for functional recovery. Currently, research is ongoing to adapt the HCP protocols to the MRI analysis in monkeys. This will enable us to conduct additional studies to analyze the structural and diffusion MRI scans.
7T MRI Measurement Systems for Interspecies Comparison
Research was conducted to establish the 7T MRI system for comparison between humans and nonhuman primates, based on the 3T MRI systems. After discussions and simulations with Dr. Yoshihiko Kawabata (Takashima Seisakusho Co., Ltd., Tokyo, Japan) and Dr. Takuya Hayashi (Riken, Hyogo, Japan), a special 24-channel array head coil was designed for use with monkeys. At present, the performance of this coil is being evaluated while the measurement sequence is being optimized. This coil will be applied to studies in the AMED Brain/MINDS (Kakushin Nou) and Brain/MINDS Beyond (Kokusai Nou) projects.