High Resolution Brain Imaging

T2*-weighted MRI at high magnetic field strength has recently been used to reveal cortical layer structures and white matter heterogeneity in vivo. Magnetic susceptibility differences have been widely thought to give rise to most of the contrast but the precise mechanisms underlying the contrast is still poorly understood. Here, we report an interesting finding from microscopic MRI and histological studies of white matter specimens of the human brain, which may provide further clues for better understanding of the mechanisms underlying the T2*-weighted contrast.

Cerebellar agenesis is a condition where the cerebellum fails to develop normally. Here we present data from two individuals acquired post-mortem from brains extracted in the 1940s showing high resolution anatomical and structural data with MPRAGE and DTI sequences.

The analysis of the human cerebral cortex and the measurement of its thickness based on MRI data provide insight into normal brain development and neurodegenerative disorders. Accurate and reproducible results of the cortical thickness measurement are desired. In addition to data processing tools, the quality (i.e. resolution) of the imaging data is evaluated. We thus compare ultra high resolution data acquired at 7T with 3T data for measuring the cortical thickness of the human brain.

There is increased interest in using ultra-high-field brain MRI to map intracortical structures. We simulated the effect on layer structure of small in-plane motions during data sampling, using 200 micron resolution ex vivo brain images. Such motions can easily introduce illusory structures, shown in images and cortical intensity profiles of human primary visual cortex, without and with motion corruption. Our simulations emphasize the crucial importance of appropriate motion correction of high resolution brain data.

The advent of high fields has made it possible to reconstruct the functional organization of ocular-dominance columns in the human cortex with sub-millimeter in-plane (2D) resolution. However, 2D-based imaging techniques necessarily use anisotropic spatial resolution and are restricted to subjects that have relatively flat regions of cortex. Using 3D EPI with sub-millimeter isotropic resolution at 7T and a differential ocular stimulation we found alternating activation patterns in V1 which may relate to the expected ocular-dominance column distribution. This suggests that at 7T, 3D EPI can offer an avenue for sub-millimeter isotropic mapping not limited by the underlying anatomy.

Many ophthalmological and neurological pathologies affect the optic nerve which provides the brain with retinal information. Revealing their manifestations with isotropic, high-resolution imaging of the optic nerve, the orbit and the chiasm may allow early and direct diagnosis of diseases that result in loss of visual function, partial or complete blindness. In this pilot study, we present isotropic, high-resolution optic nerve images which may be suitable for clinical applications.

Variations in magnetization transfer (MT) ratio across the cortex have been detected using high resolution MT scans at 7T and are assumed to correspond to variations in myelination, and variations in MT corresponding to the stria of Gennari have been detected on MT maps.

The pedunculopontine nucleus (PPN) is a potential target for deep brain stimulation to address symptoms of gait freezing and postural instability in Parkinson’s disease. Proton density-weighted (PD-w) MRI has been recommended to locate its position. Contrast and delineation of the PPN area in healthy subjects were improved by multi-echo 3D MRI at an increased resolution of 0.8 mm^3, and by using signal amplitude maps S0. These were calculated from a dual-angle FLASH protocol, thus eliminating the residual influence of T1 from the PD-w images. Usefulness for stereotactic planning was verified on two patients at 3T using a protocol of 4x4minutes.

Increased contrast and spatial resolution at 7 T permitted the reliable detection of internal architecture of the hippocampal formation. Submillimetric T2w images at 7 T consistently resolved the continuous white matter band, which separates deep portions of CA1-3 from CA4 and the dentate hilus. The resulting accuracy permitted intrahippocampal (subregional) volumetry. These preliminary results strongly support expectations that brain imaging at very high magnetic field may allow for a more accurate patient classification based on qualitative and quantitative information that is difficult or impossible to collect reliably at lower field.

Using a combination of 7 T field, B0-shim, high count receive-coil arrays and heavy T2* weighting we were able to depict hippocampal subfields down to 100 micron in 10/10 young adults in a clinically acceptable time of 14 minutes.

In the visual system the primary cortex area can robustly be identified by retinotopic mapping. In the auditory modality, a routine method to delineate the primary auditory cortex (PAC) area in individual human subjects is not available.

We developed a method to anatomically identify the PAC area on the basis of myelin content in single subjects by creating an artificial contrast using conventional T1 and T2 weighted imaging at 3 Tesla. Results show a region on the medial two thirds of Heschl’s gyrus that is very consistent to the probability map of the PAC defined in post-mortem brains.

2315. Differences in the Proportional Volume of Different Brain Regions Relative to the Whole Brain Size

Marcus Belke¹, David H. Salat², Enno Wehrmann¹, Katja Menzler¹, Ulrike Lengler³, Wolfgang H. Oertel¹, Felix Rosenow¹, Karsten Krakow³, Susanne Knake<sup>1

1</sup>Department of Neurology, Philipps-University Marburg, Marburg, Germany; ²Athinoula A. Martinos Center for Biomedical Imaging, Harvard Medical School, Boston, MA, United States; ³Brain Imaging Center Frankfurt, University of Frankfurt, Frankfurt, Germany

We investigated association between total intracranial volume (TIV) and the volume of several cortical, subcortical and white matter regions. After an automated parcellation of the brain, a slope was calculated, representing the proportional volume of each structure relative to the TIV. Cortical regions were particularly associated with TIV. The greatest slope of the subcortical regions was found for the brainstem. In a second test gender differences were investigated. Large differences were found between men and women when uncorrected volumes were compared. After correction for the influence of the TIV, no gender differences were found in any of the investigated regions.

Trigeminal autonomic cephalalgias include cluster headache, paroxysmal hemicrania and SUNCT. An earlier voxel-based morphometry (VBM) study pointed at the posterior inferior hypothalamus to be involved in CH, but results were never reproduced. In the current study we used state of the art whole-brain and regional VBM, and manual segmentation of the hypothalamus, in analyzing the brains of 151 subjects with TACs (n=70), migraine patients (n=33) and controls (n=48). We found the anterior part (but not the posterior part) of the hypothalamus, including the suprachiasmatic nucleus (“the biological clock”), to be larger in TACs compared to migraineurs and controls. Our results seem to be specific for TACs, and question the validity and/or relevance of the earlier finding, including its role in deep brain stimulation as treatment for intractable cluster headaches.

Quantitative MR techniques, such as accurate mapping of the longitudinal relaxation time and water content, have become more important in neurological research. The current T1 mapping methods are generally lengthy and not adequate in a clinical environment. Also, further acquisitions are usually required to obtain the brain tissue water content. Several factors, including RF field inhomogeneities and low SNR impair the accuracy of these methods. In this study, we present a modified two-acquisition SPGR method for simultaneous B1, T1, and M0 mapping with a 1-mm isotropic spatial resolution that covers the entire human brain in a clinically acceptable time.

Using 1.5 mm isotropic GE imaging of BOLD activation to a continuously rotating stimulus, we find individual voxels with significant orientation selectivity in human visual areas V1, V2, and V3.

The retina can be divided into seven cellular and synaptic layers. It has been shown that intraocular injection of manganese enhances contrast in the rat retina, revealing 7 layers with MRI. Gadolinium-DTPA is a T1 shortening contrast agent like manganese, but the localization of the two within in a tissue could be expected to be to differ, potentially leading to different layer-specific enhancement. In this study we used intraocular injection of gadolinium to provide unique layer enhancement in the rat retina. Gadolinium-enhanced MRI clearly resolved six retinal layers at 25x25 µm.

Knowledge of the relaxation times is not only necessary for sequence optimization; it may also be decisive to judge the advantages for ultra-high field MRI. Here, T₁, T₂ and T₂^* in the rat brain were measured at 16.4 T with a spatial resolution of 180 µm inplane. The relaxation times were quantified with high accuracy for 20 anatomical structures and maps were generated to display the spatial distribution of the relaxation times over the brain.

The study of species with unique behavioral and morphological specializations is critical when teasing apart evolutionary trends, yet becomes difficult, as often these species are extremely rare and invasive methodologies are impractical. This paper examines the use of MRI to obtain high-resolution image data in an important but damaged brain specimen of the whale shark, Rhincodon typus, wherein digital reconstruction allowed for non-invasive quantification of its brain organization. We will discuss the effectiveness of MRI as investigative tool for non-invasive visualization and quantification of the internal anatomy of fishes.

Magnetic resonance imaging is the only modality that can provide sufficient spatial resolution and image contrast to visualize Alzheimer’s amyloid plaques noninvasively. Previously Alzheimer’s amyloid plaques have been visualized in images acquired using spin-echo and gradient echo sequences at 7 T and 9.4 T. At high fields, it has been reported that the increased susceptibility-related contrast resulted in additional anatomical information, such as delineation of veins and iron-rich regions in human brain. In this study, we show that the susceptibility-induced contrast in gradient-echo phase images can improve detection of amyloid plaques.

Amyloid plaques are a marker of Alzheimer's disease which are traditionally detected as hypointense signals on T2*-weighted images due to the presence of iron. This study proposes a comparison between the images of an ex vivo APP/PS1 mouse brain obtained using a conventional T2* gradient echo sequence and a zoom adiabatic T2 spin echo sequence. This comparison, based on the ability of both sequences to allow successful plaques detection using an automatic home-made procedure, reveals that T2 contrast allows resolving amyloid plaques with a better specificity than T2* contrast, which is disturbed by the hypointense signals coming from blood vessels.

Rett Syndrome is an X-linked disorder, which primarily affects females, and is caused by mutations to the Mecp2 gene. A commonly used mouse model of RTT involves a truncation of the Mecp2 gene at codon 308. The purpose of this study was to examine the volume changes in the Mecp2³⁰⁸ Rett syndrome mouse model with high resolution MRI. Volume changes were found in many regions, for example, significant decreases were found in the cerebral cortex as well as increases in the cerebellar cortex and ventricles.

A transversal and longitudinal MRI study in patients with cervical dystonia using voxel-wise comparison of the local Gray Matter concentration.

In the present study the feasibility of imaging of the lumbar spine and its adjacent structures under healthy and under pathological conditions at 7 Tesla was investigated. A combination of a 3D-CISS and a 3D-VIBE sequence comprehended imaging of the vertebrae, the intervertebral discs, the bony neural foramina, the facet joints, the dural sac and the intraspinal portions of the spinal nerves.

Voxel-based morphometry was used to compare both gray and white matter volume in perinatally human immunodeficiency virus (HIV)-infected youth versus healthy controls. HIV patients had reduced gray matter volume in the bilateral caudate nucleus, left parietal lobe, but an increase of gray matter volume in the frontal lobe, posterior temporal lobe, and parietal lobe. Striking white matter volume reductions were found in the temporal lobe, pons, right pre-frontal area, corpus callosum and the junction of the thalamus and mid brain. These findings suggest the sensitivity of VBM in evaluating GM and WM abnormalities in perinatally HIV-infected youth.

Quantitative imaging of the R2* relaxation rate employing multiple echoes can be used to assess blood oxygenation and iron content in neural structures. However, R2* is not a strictly intrinsic tissue property, as it also depends on the spatial relationship between voxel geometry and background field inhomogeneities. These background field inhomogeneities cause additional signal decay. To investigate the influence of spatial resolution and smoothing on R2* values, we acquired images with high spatial resolution and applied spatial smoothing to the complex data, which simulates acquisition at lower spatial resolution, and to the magnitude data. We found that both changes in spatial resolution and spatial smoothing of magnitude of high resolution data leads to strong changes in R2*, which suggests that R2* values should be interpreted in the light of data acquisition parameters as well as data smoothing.

Deep brain stimulation (DBS), a surgical treatment involving the implantation of an electrode in the brain, is used for the treatment of patients with movement disorders. The success of this surgical technique is critically dependent on precise placement of the DBS electrode into the target structure. However, current clinical imaging methods lack the sensitivity for resolving and visualizing of the DBS target.

Here, using a combination of high magnetic field (7T) with susceptibility-weighted contrast resulted in a dramatically improved ability to identify and delineate anatomical architecture of deep brain structures that are FDA-approved DBS targets.

2330. Human T₂* and Phase Imaging at 9.4 T

Juliane Sabine Budde¹, Gunamony Shajan¹, Jens Hoffmann¹, Frank Muehlbauer¹, Kâmil Ugurbil², Rolf Pohmann<sup>1

1</sup>Max Planck Institute for Biological Cybernetics, Tuebingen, Germany; ²Center for Magnetic Resonance Research, University of Minnesota, Minneapolis, United States

Ultra-high static magnetic field causes higher susceptibility effects which yield shorter T2* values and larger variations of the image phase. In this work, we acquired highly detailed T2* maps showing internal structures. Mean T2* values for GM were estimated as 28ms±6ms and 20ms±4ms for WM. Phase images were post-processed to yield images with high tissue contrast between grey and white matter throughout the brain at a resolution of 200µm x 200µm x 1mm. Signal gain at ultra-high field allows for high resolution surface phase images of 130µm x 130µm in-plane resolution. In these, differences within grey matter are visible.

Seven patients with clinically and standard imaging-based proven ischemic or hemorrhagic stroke were scanned with magnetization prepared 3D FLAIR, combined time-of-flight inflow and multi-echo fast field echo (meFFE), T₁ 3D TFE, and DTI. 7.0 Tesla results were comparable to results of similar 1.5 Tesla sequences, but with better resolution and – in 3 out of 7 patients – additional information regarding underlying pathology. Furthermore, meFFE with 3 echoes was valuable in identification of microbleeds, microinfarcts and thrombus.