High-resolution Brain Imaging Research Articles

Automated subcortical segmentation using FIRST: Test-retest reliability, interscanner reliability, and comparison to manual segmentation

Multiple techniques exist for the automated segmentation of magnetic resonance images (MRIs). The validity of these techniques can be assessed by evaluating test-retest reliability, interscanner reliability, and consistency with manual segmentation. We evaluate these measures for the FSL/FIRST subcortical segmentation tool. We retrospectively analyzed 190 MRI scans from 87 subjects with mood or anxiety disorders and healthy volunteers scanned multiple times on different platforms (N = 56) and/or the same platform (N = 45, groups overlap), and 146 scans from subjects who underwent both high-resolution and whole brain imaging in a single session, for comparison with manual segmentation of the hippocampus. The thalamus, caudate, putamen, hippocampus, and pallidum were reliably segmented in different sessions on the same scanner (Intraclass correlation coefficient (ICC) > 0.83 scanners and diagnostic groups pooled). In these regions, the range of between platform reliabilities were lower (0.527 < ICC < 0.953), although values below 0.7 were due to systematic differences between platforms or low reliability in the hippocampus between eight- and single-channel coil platforms. Accumbens and amygdala segmentations were generally unreliable within and between scanning platforms. ICC values for hippocampal volumes between automated and manual segmentations were acceptable (ICC > 0.7, groups pooled), and both methods detected significant differences between genders. In addition, FIRST segmentations were consistent with manual segmentations (in a subset of images; N = 20) in the left caudate and bilateral putamen. This retrospective analysis assesses realistic performance of the algorithm in conditions like those found in multisite trials or meta-analyses. In addition, the inclusion of psychiatric patients establishes reliability in subjects exhibiting volumetric abnormalities, validating patient studies.

Spin-echo magnetic resonance spectroscopic imaging at 7 T with frequency-modulated refocusing pulses.

Two approaches to high-resolution SENSE-encoded magnetic resonance spectroscopic imaging (MRSI) of the human brain at 7 Tesla (T) with whole-slice coverage are described. Both sequences use high-bandwidth radiofrequency pulses to reduce chemical shift displacement artifacts, SENSE-encoding to reduce scan time, and dual-band water and lipid suppression optimized for 7 T. Simultaneous B0 and transmit B1 mapping was also used for both sequences to optimize field homogeneity using high-order shimming and determine optimum radiofrequency transmit level, respectively. One sequence ("Hahn-MRSI") used reduced flip angle (90°) refocusing pulses for lower radiofrequency power deposition, while the other sequence used adiabatic fast passage refocusing pulses for improved sensitivity and reduced signal dependence on the transmit-B1 level. In four normal subjects, adiabatic fast passage-MRSI showed a signal-to-noise ratio improvement of 3.2±0.5 compared to Hahn-MRSI at the same spatial resolution, pulse repetition time, echo time, and SENSE-acceleration factor. An interleaved two-slice Hahn-MRSI sequence is also demonstrated to be experimentally feasible.

Mathematical Methods for Images and Surfaces 2011

Mathematical Methods for Images and Surfaces 2011

Cortical Surface Reconstruction from High-Resolution MR Brain Images.

Reconstruction of the cerebral cortex from magnetic resonance (MR) images is an important step in quantitative analysis of the human brain structure, for example, in sulcal morphometry and in studies of cortical thickness. Existing cortical reconstruction approaches are typically optimized for standard resolution (~1 mm) data and are not directly applicable to higher resolution images. A new PDE-based method is presented for the automated cortical reconstruction that is computationally efficient and scales well with grid resolution, and thus is particularly suitable for high-resolution MR images with submillimeter voxel size. The method uses a mathematical model of a field in an inhomogeneous dielectric. This field mapping, similarly to a Laplacian mapping, has nice laminar properties in the cortical layer, and helps to identify the unresolved boundaries between cortical banks in narrow sulci. The pial cortical surface is reconstructed by advection along the field gradient as a geometric deformable model constrained by topology-preserving level set approach. The method's performance is illustrated on exvivo images with 0.25–0.35 mm isotropic voxels. The method is further evaluated by cross-comparison with results of the FreeSurfer software on standard resolution data sets from the OASIS database featuring pairs of repeated scans for 20 healthy young subjects.

Fast high-resolution brain imaging with balanced SSFP: Interpretation of quantitative magnetization transfer towards simple MTR

Magnetization transfer (MT) reflects the exchange of magnetization between protons bound to macromolecules, such as lipids and proteins, and protons in free liquid, and thus might be an early marker for subtle and undetermined pathologic changes in tissue. Detailed analysis of the entire MT phenomenon, however, commonly requires extensive data acquisition and scanning time, and hence is only of limited clinical interest. Therefore, in practice, magnetization transfer effects are commonly confined into a simple ratio measure, the so-called magnetization transfer ratio (MTR), calculated from a MT-weighted and a non-MT-weighted image. However, subtle physiologic and pathologic changes in tissue, invaluable for specific diagnostic imaging, may be lost since MTR-values depend not only on quantitative magnetization transfer (qMT) parameters but also on sequence parameters and relaxation properties. In order to evaluate and assess the diagnostic specificity of MTR versus qMT, high-resolution whole brain MT data was collected from twelve healthy volunteers using balanced steady-state free precession (bSSFP). In contrast to common MT imaging based on spoiled gradient echo (SPGR) sequences, whole brain qMT imaging can be performed with MT-sensitized bSSFP within a clinically feasible acquisition time. Hence, MT-sensitized bSSFP provides access to both MTR and qMT parameters within a clinical setting. The reliability and possible diagnostic value of MTR are analyzed for twelve white matter (WM) and eleven gray matter (GM) structures of the normal appearing brain. Strong correlations were found within and between longitudinal and transverse relaxation times (T1, T2) and MT parameters (ratio between macromolecular and water protons, F, and magnetization exchange rate, kf), whereas weaker correlations were observed between MTR-values and relaxation times or MT parameters. Structures with highly similar MTR-values, such as the crus cerebri and the anterior commissure in the WM, or the pallidum and the amygdala in the GM, however, were also found that showed significant differences in most quantitative parameters. This observation was confirmed from simulations revealing that the overall effect on MTR from an increase (decrease) in relaxation times may be counterbalanced with a decrease (increase) in MT parameters. These findings corroborate the expectation that qMT is superior to MTR imaging, especially for the evaluation and assessment of pathologic or physiological changes in healthy and pathologic brain tissue.

Ultra-Wideband Sensors for Improved Magnetic Resonance Imaging, Cardiovascular Monitoring and Tumour Diagnostics

The specific advantages of ultra-wideband electromagnetic remote sensing (UWB radar) make it a particularly attractive technique for biomedical applications. We partially review our activities in utilizing this novel approach for the benefit of high and ultra-high field magnetic resonance imaging (MRI) and other applications, e.g., for intensive care medicine and biomedical research. We could show that our approach is beneficial for applications like motion tracking for high resolution brain imaging due to the non-contact acquisition of involuntary head motions with high spatial resolution, navigation for cardiac MRI due to our interpretation of the detected physiological mechanical contraction of the heart muscle and for MR safety, since we have investigated the influence of high static magnetic fields on myocardial mechanics. From our findings we could conclude, that UWB radar can serve as a navigator technique for high and ultra-high field magnetic resonance imaging and can be beneficial preserving the high resolution capability of this imaging modality. Furthermore it can potentially be used to support standard ECG analysis by complementary information where sole ECG analysis fails. Further analytical investigations have proven the feasibility of this method for intracranial displacements detection and the rendition of a tumour’s contrast agent based perfusion dynamic. Beside these analytical approaches we have carried out FDTD simulations of a complex arrangement mimicking the illumination of a human torso model incorporating the geometry of the antennas applied.

Heterogeneous patterns of tissue injury in NARP syndrome

Point mutations at m.8993T>C and m.8993T>G of the mtDNA ATPase 6 gene cause the neurogenic weakness, ataxia and retinitis pigmentosa (NARP) syndrome, a mitochondrial disorder characterized by retinal, central and peripheral neurodegeneration. We performed detailed neurological, neuropsychological and ophthalmological phenotyping of a mother and four daughters with NARP syndrome from the mtDNA m.8993T>C ATPase 6 mutation, including 3-T brain MRI, spectral domain optical coherence tomography (SD-OCT), adaptive optics scanning laser ophthalmoscopy (AOSLO), electromyography and nerve conduction studies (EMG-NCS) and formal neuropsychological testing. The degree of mutant heteroplasmy for the m.8993T>C mutation was evaluated by real-time allele refractory mutation system quantitative PCR of mtDNA from hair bulbs (ectoderm) and blood leukocytes (mesoderm). There were marked phenotypic differences between family members, even between individuals with the greatest degrees of ectodermal and mesodermal heteroplasmy. 3-T MRI revealed cerebellar atrophy and cystic and cavitary T2 hyperintensities in the basal ganglia. SD-OCT demonstrated similarly heterogeneous areas of neuronal and axonal loss in inner and outer retinal layers. AOSLO showed increased cone spacing due to photoreceptor loss. EMG-NCS revealed varying degrees of length-dependent sensorimotor axonal polyneuropathy. On formal neuropsychological testing, there were varying deficits in processing speed, visual–spatial functioning and verbal fluency and high rates of severe depression. Many of these cognitive deficits likely localize to cerebellar and/or basal ganglia dysfunction. High-resolution retinal and brain imaging in NARP syndrome revealed analogous patterns of tissue injury characterized by heterogeneous areas of neuronal loss.Electronic supplementary materialThe online version of this article (doi:10.1007/s00415-010-5775-1) contains supplementary material, which is available to authorized users.

Introduction

Introduction

Anatomical Global Spatial Normalization

Anatomical global spatial normalization (aGSN) is presented as a method to scale high-resolution brain images to control for variability in brain size without altering the mean size of other brain structures. Two types of mean preserving scaling methods were investigated, “shape preserving” and “shape standardizing”. aGSN was tested by examining 56 brain structures from an adult brain atlas of 40 individuals (LPBA40) before and after normalization, with detailed analyses of cerebral hemispheres, all gyri collectively, cerebellum, brainstem, and left and right caudate, putamen, and hippocampus. Mean sizes of brain structures as measured by volume, distance, and area were preserved and variance reduced for both types of scale factors. An interesting finding was that scale factors derived from each of the ten brain structures were also mean preserving. However, variance was best reduced using whole brain hemispheres as the reference structure, and this reduction was related to its high average correlation with other brain structures. The fractional reduction in variance of structure volumes was directly related to ρ 2, the square of the reference-to-structure correlation coefficient. The average reduction in variance in volumes by aGSN with whole brain hemispheres as the reference structure was approximately 32%. An analytical method was provided to directly convert between conventional and aGSN scale factors to support adaptation of aGSN to popular spatial normalization software packages.

A non-local approach for image super-resolution using intermodality priors☆

Image enhancement is of great importance in medical imaging where image resolution remains a crucial point in many image analysis algorithms. In this paper, we investigate brain hallucination (Rousseau, 2008), or generating a high-resolution brain image from an input low-resolution image, with the help of another high-resolution brain image. We propose an approach for image super-resolution by using anatomical intermodality priors from a reference image. Contrary to interpolation techniques, in order to be able to recover fine details in images, the reconstruction process is based on a physical model of image acquisition. Another contribution to this inverse problem is a new regularization approach that uses an example-based framework integrating non-local similarity constraints to handle in a better way repetitive structures and texture. The effectiveness of our approach is demonstrated by experiments on realistic Brainweb Magnetic Resonance images and on clinical images from ADNI, generating automatically high-quality brain images from low-resolution input.

V3D enables real-time 3D visualization and quantitative analysis of large-scale biological image data sets.

The V3D system provides three-dimensional (3D) visualization of gigabyte-sized microscopy image stacks in real time on current laptops and desktops. Combined with highly ergonomic features for selecting an X, Y, Z location of an image directly in 3D space and for visualizing overlays of a variety of surface objects, V3D streamlines the on-line analysis, measurement, and proofreading of complicated image patterns. V3D is cross-platform and can be enhanced by plug-ins. We built V3D-Neuron on top of V3D to reconstruct complex 3D neuronal structures from large brain images. V3D-Neuron enables us to precisely digitize the morphology of a single neuron in a fruit fly brain in minutes, with about 17-fold improvement in reliability and 10-fold savings in time compared to other neuron reconstruction tools. Using V3D-Neuron, we demonstrated the feasibility of building a 3D digital atlas of neurite tracts in the fruit fly brain.

Segregation analysis in mesial temporal lobe epilepsy with hippocampal atrophy

Itiswellknownthatamongepilepsieswithfocalseizureonset, temporal lobe epilepsies (TLEs) are the mostcommon forms (Gloor, 1991). Mesial temporal lobe epi-lepsy (MTLE) is one type of TLE frequently associatedwith mesial temporal sclerosis (MTS), a neuropathologicabnormality that can be diagnosed in vivo by high-resolution brain imaging as hippocampal atrophy (HA)and abnormal signal intensity (Berkovic et al., 1991; Cen-desetal.,1993a).MTS,whichischaracterizedbyselectiveneuronal loss and gliosis in regions of hippocampus andhilus, has been associated with predisposing environmen-talfactors,suchasprolongedfebrileseizures(FS)inchild-hood(Abou-Khaliletal.,1993;Cendesetal.,1993b).Familial forms of TLE have been identified, suggestingthatgenetic factorscouldbeinvolvedinthedeterminationof different types of TLE (Berkovic et al., 1996; Ottmanetal.,1995;Gambardellaetal.,2000).Familialrecurrenceof MTLE associated with HA has also been reported(Cendes et al., 1998; Fernandez et al., 1998; Kobayashiet al., 2001). Extensive phenotypic studies in large MTLEfamilies showed a certain degree of variability in clinicalandimagingfindings,butmostaffectedindividualshaveabenign course of the disease. In these families HA wasobservedinpatientswhohadrefractoryseizures,requiringsurgical treatment (Kobayashi et al., 2003a), as well as inindividuals who had only a single partial seizure and,unexpectedly, in 34% of asymptomatic first-degree rela-tives of patients (Kobayashi et al., 2002, 2003b). Theseinitial observations suggested that genetic factors mayplay a role in MTLE associated with HA. In order to fur-ther investigate this issue we performed complex segrega-tion analysis in a sample of nuclear families of probandswithMTLEandHA.

Mix-ratio sampling: Classifying multiclass imbalanced mouse brain images using support vector machine

Support Vector Machine (SVM) is a classifier designed to achieve optimized classification accuracy. It has been applied to numerous applications associated with images. Yet challenges remain when applying SVM on segmenting mouse brain images. This is due to the fact that each high-resolution mouse brain image is a very large data set and it is a multiclass classification problem with extremely imbalanced data size for different classes. To address these issues, a mix-ratio sampling approach for SVM is proposed which determines various over-sampling ratios for different minority classes. In addition, to improve the imaging classification accuracy, spatial information is incorporated into the classification problem. Five mouse Magnetic Resonance Microscopy (MRM) images are collected to test the accuracy of classifying 21 brain structures. The first comparison experiment demonstrates the SVM with mix-ratio sampling method relieves the imbalance problem for multiclass more effectively and efficiently than the SVM with simple over-sampling method. In the second comparison experiment, another classifier, Artificial Neural Network (ANN) is used to compare against SVM based on the same mix-ratio sampled data and the results indicate that SVM shows better classification performance than ANN. Thirdly, the cross validation is conducted to demonstrate SVM with mix-ration sampling can classify multiclass imbalanced data with high accuracy.

Great Expectations

A convergence of diverse disciplines is fueling the expansion of neuroscience research, and underpinning this growth are increasingly advanced technologies such as high-performance computation, simulation and neuron modeling, real-time and high-resolution brain imaging, advanced recording techniques, and detailed neuron mapping. This special issue of Science gives a well-deserved nod to the technological innovations that have improved our understanding of the development, function, and disorders of the nervous system, and that will continue to open up many new questions about the human brain.

Gray matter abnormalities in subjects at ultra-high risk for schizophrenia and first-episode schizophrenic patients compared to healthy controls

Neuroimaging studies have revealed gray matter abnormalities in schizophrenia in various regions of the brain. It is, however, still unclear whether such abnormalities are already present in individuals at ultra-high risk (UHR) for transition into psychosis. We investigated this issue using voxel-based morphometry of structural magnetic resonance images (MRI) and compared UHR patients with first-episode patients with schizophrenia and healthy controls. Gray matter volume maps from high-resolution MR T1-weighted whole brain images were analyzed in a cross-sectional study in 30 UHR patients, 23 first-episode schizophrenic patients and 29 controls. UHR patients showed significantly lower gray matter volume in the cingulate gyrus bilaterally, in the right inferior frontal and right superior temporal gyrus, as well as in the left and right hippocampus in comparison to healthy subjects. First-episode patients with schizophrenia showed smaller gray matter volume in the cingulate cortex bilaterally, in the left orbitofrontal gyrus, in the right inferior frontal and superior temporal gyrus, in the right temporal pole, in the left and right hippocampus, in the left parahippocampus, left amygdala, and in the left fusiform gyrus compared to the UHR patients. This study provides further evidence that gray matter brain volume, especially in the anterior cingulate cortex, is already reduced in the prodromal state of schizophrenia.

Early parental care is important for hippocampal maturation: Evidence from brain morphology in humans

The effects of early life experience on later brain structure and function have been studied extensively in animals, yet the relationship between childhood experience and normal brain development in humans remains largely unknown. Using a unique longitudinal data set including ecologically valid in-home measures of early experience during childhood (at age 4 and 8 years) and high-resolution structural brain imaging during adolescence (mean age 14 years), we examined the effects on later brain morphology of two dimensions of early experience: parental nurturance and environmental stimulation. Parental nurturance at age 4 predicts the volume of the left hippocampus in adolescence, with better nurturance associated with smaller hippocampal volume. In contrast, environmental stimulation did not correlate with hippocampal volume. Moreover, the association between hippocampal volume and parental nurturance disappears at age 8, supporting the existence of a sensitive developmental period for brain maturation. These findings indicate that variation in normal childhood experience is associated with differences in brain morphology, and hippocampal volume is specifically associated with early parental nurturance. Our results provide neuroimaging evidence supporting the important role of warm parental care during early childhood for brain maturation.

The DRD2 TaqIA polymorphism associated with changed midbrain volumes in healthy individuals

The human DRD2 gene is located on chromosome 11q22-q23 and contains one specific functional polymorphism called TaqIA, which characteristically presents two alleles referred to as A1 and A2. Evidence indicates that the A1 allele impacts brain dopaminergic function and may confer an increased risk of developing Parkinson's disease. However, possible morphological changes underlying such genetic variant remain to be clarified. The aim of this study was to provide an in vivo demonstration of changes in brain structures associated with the TaqIA polymorphism of the DRD2 gene. Optimized voxel-based morphometry (VBM) was applied to high-resolution MR brain images of 70 healthy controls divided into two groups according to their DRD2 genotype (A1/A2, n = 15; A2/A2, n = 55). Compared with individuals' homozygous for the A2 allele, the A1 carriers had significantly smaller areas of a specific part of the midbrain, encompassing the substantia nigra bilaterally. Our findings showed an association of the DRD2 TaqIA polymorphism with the changed volumes of a specific subcortical region strongly involved in the dopaminergic system.

High-resolution imaging of the large non-human primate brain using microPET: a feasibility study

The neuroanatomy and physiology of the baboon brain closely resembles that of the human brain and is well suited for evaluating promising new radioligands in non-human primates by PET and SPECT prior to their use in humans. These studies are commonly performed on clinical scanners with 5 mm spatial resolution at best, resulting in sub-optimal images for quantitative analysis. This study assessed the feasibility of using a microPET animal scanner to image the brains of large non-human primates, i.e. papio hamadryas (baboon) at high resolution. Factors affecting image accuracy, including scatter, attenuation and spatial resolution, were measured under conditions approximating a baboon brain and using different reconstruction strategies. Scatter fraction measured 32% at the centre of a 10 cm diameter phantom. Scatter correction increased image contrast by up to 21% but reduced the signal-to-noise ratio. Volume resolution was superior and more uniform using maximum a posteriori (MAP) reconstructed images (3.2–3.6 mm3 FWHM from centre to 4 cm offset) compared to both 3D ordered subsets expectation maximization (OSEM) (5.6–8.3 mm3) and 3D reprojection (3DRP) (5.9–9.1 mm3). A pilot 18F-2-fluoro-2-deoxy-d-glucose ([18F]FDG) scan was performed on a healthy female adult baboon. The pilot study demonstrated the ability to adequately resolve cortical and sub-cortical grey matter structures in the baboon brain and improved contrast when images were corrected for attenuation and scatter and reconstructed by MAP. We conclude that high resolution imaging of the baboon brain with microPET is feasible with appropriate choices of reconstruction strategy and corrections for degrading physical effects. Further work to develop suitable correction algorithms for high-resolution large primate imaging is warranted.

Hemodynamic responses in cortex investigated with optical imaging methods. Implications for functional brain mapping

During the last 20 years, optical imaging methods – either alone or in combination with other recording techniques – has proven a fruitful approach to explore both the physiological and the functional aspects of activity-evoked hemodynamic responses in cortex. One of the main advantages of optical imaging consists in its high spatio-temporal resolution (in the order of few microns and milliseconds), allowing not only to unambiguously distinguish between activity patterns relating to the underlying functional architecture and those originating from the activation of medium/large blood vessels, but also to investigate the various activity-evoked hemodynamic processes at very fine detail. Here, we briefly review the principal findings obtained by optical imaging about the spatio-temporal properties of the various hemodynamic responses in cortex, i.e., changes in blood-oxygenation, blood-volume, and, to some extent, blood-flow. We also discuss the implications of those findings for non-invasive high-resolution functional brain imaging, in particular for fMRI. Finally, we underscore the importance of novel approaches for high-resolution blood-flow imaging, in the context of the need to gather information at fine spatial detail about the blood-flow response, necessary to constrain the multiple free parameters of hemodynamic response models.

Magnetron Surface Coil for Brain MR Imaging

A resonator surface coil was developed for magnetic resonance imaging of the brain and tested on a clinical imager. This resonator design was based on the cavity magnetron with an 8 slot-and-hole configuration. High-resolution brain images were obtained from a water-filled phantom and from a healthy volunteer brain. To compare coil performance, SNR-vs.-depth plots were computed for a single-loop coil and the magnetron prototype from phantom images. These experimentally acquired profiles show an important improvement in SNR. Thus, the magnetron surface coil can generate brain images with a high resolution and penetration capacity. The high sensitivity of this coil makes it a good candidate to be used in multicoil imaging sequences.