History of MRI goes back to 1970 with low field MRI using resistive magnet such as 0.1T or 1 Kilo Gauss MRI (123). From the early and mid-1980, superconducting magnet became available and began to obtain clinically usable images. First 2.0T^* MRI became a reality in 1985 (KAIS 1985) and began to provide superior images with time limits that are acceptable to the clinical community and rapidly became the major diagnostic tool (45). In the mid-1990, sporadically 3.0T MRI became available and starts providing improved images in certain areas. Nearly two decades later, first experimental research proto type of ultra-high field 7.0T MRI became available in the research community and start providing markedly improved anatomical images hitherto unavailable with low field MRI (67).

History of modern MRI development is given in Figure 1 where the time and duration of the each modality remained and durations of the development periods are illustrated. As shown in Figure 2, one of the first 2.0T MRI magnet installed at the Korea Advanced Institute of Science (KAIS) in 1985 is shown together with the students and faculty who are watching the assembling the magnet to be installed at the laboratory (1985). Nearly two decades later in 2005, again one of the first 7.0T magnets for the installation at the Neuroscience Research Institute in Gachon Medical School in Incheon, Korea (2005) is shown in Figure 3.

This 7.0T MRI at NRI (2005), together with the other three 7.0T MRI installations at MGH Harvard (2005) and Magdeburg University in Germany (2005), and NIH (2005) were the four first groups started to develop the machine for the research and preclinical study. First research trials were mostly anatomical imaging of the human brain (67). To perform the imaging, first and most difficult task was to develop the suitable RF coils for the brain imaging within the limited time to avoid possible movement related artifacts as well as the dose uniformity to the subject under study (6). Several types of the RF coils developed during this period and used for the experiments are shown in Figure 4. Most of the experiments are performed with volunteers of various age categories from as young as 18 years old to 80 years old, both males and females. For the each experiment, we have obtained both Institutional Review Board (IRB) as well as permission from the Korea Food and Drug Administration with consent from the each volunteer. Quite large number of patients has been studied in collaborations with physicians who are in charge of the patients with consents from the patients who volunteered for the study. For the most of the study we have found no particular effects on both normal volunteers as well as patients except some who explained that there was minor dizziness during the study but no particular complains after the exam. It is generally felt that the 7.0T MRI is not health hazardous in any form except minor dizziness felt by patients during the study and agreed in general that there is no particular health hazard effect on human.

Among the many conventional imaging modality, diffusion MRI requires an additional care, for example data was obtained from a healthy adult young male of age 30 using a 7.0T research MRI scanner (Magnetom, Siemens, Erlangen, Germany) with 40 mT/m maximum gradient field. The specific imaging parameters used for the data acquisition were a single shot DW echo planar imaging with following parameters: TR/TE = 6000/83 ms; matrix = 128 · 128; 1.8mm isotropic resolution; 64 DW-directions, b = 0 and 2000 s/mm², 3 repeats; GRAPPA with factor 3 and 45 slices, band width: 1562 Hz/Px, and total acquisition time: 19 min 05 s (6 s · 65 · 3 repeats). For reference, we have also acquired T1-weighted MPRAGE images. Diffusion MRI data were further processed to correct for geometric distortions using the point spread function correction technique. Calculation of fiber orientation distribution was also performed using software package MRtrix (Brain Research Institute, Melbourne, Australia, http://www.brain.org.au/software/) (8).

Whole brain fiber tracking or the tractography was carried out with probabilistic streamline using the second order integration over fiber orientation distribution algorithm (9). From the reconstructed fiber-tracks, super-resolution TDI maps were generated by calculating the number of tracks in each element of a grid. One of the important new developments of the TDI is that the grid element can be made smaller than the original voxel size, often as much as a factor 10 or more depending on the statistical quality of the data (9). For the current study, we have chosen a 200 µm isotropic grid size, which is nearly 1000 times smaller compared with the original data in voxel size. Directionally encoded color (DEC) TDI maps (the super resolution equivalent of the DEC map in diffusion tensor imaging) were also used for each data set. The IRB of Gachon University of Medicine and Korea Food and Drug Administration approved the procedures of the present study. Participant provided written informed consent.

Three areas of the imaging experiments are focused on for the study, namely the anatomy, angiography, and tractography. First group of anatomical study results are shown in Figures 5,6,7 for the imaging of the in-vivo human brainstems at two areas. Figure 5 shows two sagittal view images of the brainstem areas at near the middle portion of the brain where the substantia nigra and hippocampus are located. As shown, at left side (x = -8 mm), we can see clearly the substantia nigra at the upper side of the pons with all the fine fiber tracts which appears to be the corticospinal tracts are clearly visible. In the right, at x = -12 mm, hippocampus as well as amygdala are all clearly visible with clarity never before possible with low field MRI, such as the conventional 1.5T or 3.0T. Figure 6 shows the expanded view of the midbrain area shown in Figure 5 with finer details such as the substantia nigra and nucleus accumbens. In Figure 7, a sagittal slice with expansion of the previous image shown in Figure 5, again we can see all the fine details of the hippocampus and amygdala together with the main body of the hippocampus and tail extended to the posterior aspect. For the first time, we can see clearly the separation between the head of hippocampus and amygdala complex. In Figure 8, an over view of the hippocampus with an expanded view image is shown which clearly visualizing the dentate gyrus connected from the head of the hippocampus to the tail with all the fine details. In Figure 9, an expanded view image of a coronal view of the hippocampus showing the substructures of the hippocampus such as the CA1, CA2, CA3, CA4, and dentate gyrus together with the subiculum and parahippocampal gyrus, the details never before visualized by the conventional low field MRI. In Figures 10 and 11, these two images demonstrate again one of the most clinically useful information, that is the early possible diagnosis of the Parkinson's disease, the possibility of directly observing the deterioration of the substantia nigra due to the Parkinsonism. In Figure 10, a sagittal view image and a cut view image in an axial direction (small one) of a healthy male are shown to visualize the substantia nigra, the main cortical gray matter believed to be involved in Parkinsonism. As seen, the substantia nigra is surrounded by the crus cerebri and more medially, the two red nuclei are clearly visible showing how the healthy substantia nigra should look like. In the subsequent figure (Fig. 11), a set of the same cut axial view images of a healthy control and a Parkinson's disease, respectively are shown. As it is clearly visible, the image of the Parkinson's disease case, the subject's substantia nigra is heavily distorted in the boundary of the substantia nigra and the crus cerebri compared with the smooth arch type shown in the left for the healthy normal subject. This simple and clearly differentiated observation of the Parkinsonism would be a useful tool for the differential diagnosis of the Parkinson's and Parkinson's related movement disorders, such as the supranuclear palsy (SNP) or the multiple system atrophy (MSA). In Figure 11, a comparison of the regions where the substantia nigra is located for Parkinson's and a healthy normal is shown. As seen two images are clearly different and Parkinson's is easily identified from that of the healthy. Note here, that the Parkinson's patient is a newly detected early Parkinsonism, the H&Y 1.

Last two decades or so, such as the Parkinson's diseases has been treated by surgical means such as the deep brain stimulation (DBS) technique, usually known as DBS. The surgical operation, however, requires an accurate information on the location of the involved nucleus or the grey matter which is sitting deep in the brain, the subthalamic nucleus or STN. Finding the location of the STN is a laborious and yet complicate and often difficult procedure to find the exact location of this small but important nucleus. 7.0T's high resolution capability offer great advantage in finding the nucleus or the target point precisely and rapidly. In Figure 12, a 7.0T MR image of coronal view is shown together with the anatomical details of the region where STN is located. From this coronal view image (Fig. 12), once STN target point is identified within the green elliptical circle, it is easy to locate it in an axial view via three dimensional coordinate translation as shown in Figure 13 where precise 3D locations of the target point in an axial view is shown for the DBS targeting of the STN.

In the areas of angiography, UHF MRI has another advantage in addition to the high resolution anatomical image, that is with increased T1 value with UHF MRI which helps us to make better contrast angiography thanks to the increased contrast between the steady state tissues in the imaging slice which are well saturated due to repeated excitation in contrast to that of the newly arriving fresh unsaturated blood which entered into the imaging slice, therefore, gives off better signals. In Figure 14, small vessel imaging of the lenticulostriate artery (LSA) region with 1.5T and 7.0T MR, respectively, are shown. As shown, small vessels in the LSA region in the low field MRI (1.5T) is completely lost while the small vessels in the LSA are clearly visible with 7.0T MR angiography, suggesting potentials of the 7.0T angiography in early and difficult small vessels detection for the strokes and other vascular abnormality. In Figure 15, an example of imaging of the middle cerebral artery (MCA) abnormality with newly generated collaterals. As shown, in 3.0T angiography, MCA occlusion is so severe, that it can be concluded that the MCA is virtually occluded while with 7.0T angiography, we can see all the collaterals which are developed over the time as a result of the MCA occlusion.

Finally, one of the most exciting and promising areas of UHF MRI such as 7.0T is in the areas of the neural fiber tracking using diffusion tensor imaging (DTI) or tractography (10). Up to now, most of the low field DTI suffered great deal with low resolution which involves many image distortion effects such as the "crossing" and "kissings" of the fiber tracts and bundles thereby severely disturbed formation of accurate tractography (10). Solution to this problem is the high resolution tractography or small voxel imaging which requires high signal to noise (SNR) ratio to fulfill the statistical requirement. In this respect, again 7.0T's high signal to noise ratio and high signal intensity allow us to improve image resolution so that small voxel imaging is possible, thereby, allows us to avoid the resolution degrading factors such as "crossing" and "kissings" and also allow us to manipulate the image by post image processing. Recently, Calamante et al. (8) have developed what is known as TDI, or track density imaging, a post signal processing of the data with voxel fractionation and deconvolution. Present tractography work is the result of the collaboration between the University of Melbourne group (89) who developed track density imaging algorithm, known as TDI and our 7.0T MRI DTI data.

In Figure 16, a comparison of the conventional DTI image and the new TDI +7.0T DTI image or tractography is shown. As shown, substantial improvement is seen (at right) compared with conventional data (8). In Figure 17, one example of the new 7.0T tractography of a coronal view image is shown with assignment of the fiber tracts such as the anterior thalamic radiation (ATR) and other small fiber tracts which would normally be difficult to visualize with conventional low field MR tractography. This remarkably improved resolution finally allow us to examine many new fiber tracts such as the sub-division of the classically known superior longitudinal fasciculus (SLF) hitherto unable to classify. In Figure 18, two coronal views of tractography are shown to compare the SLF, one at the left with the first original DTI tractography with proposed sub division of the SLF by Makris et al. (11) while, at the right, a new TDI +7.0T MR tractography is shown with new proposed subdivision scheme of the SLF (12). New SLF sub-division with corresponding sagittal view images are shown in Figure 19 to demonstrate the well separated individual sagittal view data by the newly proposed sub-division scheme of the SLF. From this new tractography, for the first time, we are able to clearly identify the individual sub-divisions of the SLF with which functional-connectivity map can be constructed or obtained (1012).

As shown in the result section, last decade was one of the most exciting period during which MRI research and developments made a remarkable progresses, both in the anatomical imaging as well as several other non-anatomical imaging arena such as angiography and tractography by extending the resolution limit down to sub-millimeter region, that is as small as one to two hundreds micron level in-vivo human imaging hitherto impossible to obtain with conventional low field MRI. It is also worth to point out that the UHF MRI appears no particular hazardous effect on human and safety.

More recently new UHF MRI systems without necessity of external shielding appeared and start providing easy installation without unnecessarily large space requirement like the old prototype research 7.0T system which requires as much as 500 tons of iron and large space for shielding and installation. It should also be noted that the last three to four decades of MRI development implies the future development of, such as 14.0T and beyond, is possible as technology develop. In Figure 20, future prospect of development of the UHF and EHF (Extreme High Field) MRI such as 14.0T is illustrated and the potential scientific progresses that can be gained are suggested. Similarly, in Figure 21, what has been progressed from the early start of 1970 by the group of University of Nottingham by Andrew et al. with 0.1T MRI in the mid 70's to modern 7.0T developments are noted. This historical development clearly suggests trend of future development of MRI in the field of UHF and EHF areas. In Korea, Cho et al. (3) at KAIS, Seoul, Korea, started developing 0.1T MRI as early as late 70's and succeeded in obtaining human images and progressed to 2.0T MRI in 1985, one of the first highest field MRI in the world.

Subsequent UHF development was followed after 20 years since the 2.0T MRI and installed one of the first 7.0T MRI in 2005 at Neuroscience Research Institute of Gachon Medical School, Incheon, Korea. Currently over 70 units of 7.0T MRI are installed at various research centers of the world and pursuing research and development not only for the preclinical research but also the basic MRI research. It is envisioned that, in near future, there will be 11.7T MRI for research (NIH etc.) as well as 10.4T (University of Minnesota) and will be added to the current 7.0T research which has nearly 10 years of research and preclinical experiences.

In conclusions, it is envisioned that, in near future, there will be more 7.0T MRIs for brain research and explosive clinical application research will also be developed, for example in the area of connectomics in neuroscience and clinical neurology and neurosurgery. It is also clear that the UHF and EHF MRI research and development will continue and start providing new insights to the research and future clinical applications such as the newly arising "Brain" and other Neuroscience related research much discussed today such as the USA's "Brain Initiative" and European Union's "Human Brain Project" as well as clinical arena such as the DBS development and its clinical applications (1314).