Handcrafted fuzzy rules for tissue classification

This article proposes a handcrafted fuzzy rule-based system for segmentation and identification of different tissue types in magnetic resonance (MR) brain images. The proposed fuzzy system uses a combination of histogram and spatial neighborhood-based features. The intensity variation from one type of tissue to another is gradual at the boundaries due to the inherent nature of the MR signal (MR physics). A fuzzy rule-based approach is expected to better handle these variations and variability in features corresponding to different types of tissues. The proposed segmentation is tested to classify the pixels of the T2-weighted axial MR images of the brain into three primary tissue types: white matter, gray matter and cerebral-spinal fluid. The results are compared with those from manual segmentation by an expert, demonstrating good agreement between them.     

MRI gradient fields increase brain mannitol space

Following nephrectomy and intravenous injection of tritiated mannitol, adult male rats were exposed to magnetic resonance imaging (MRI) procedures at 1.5 T, 0.5 T, and 0.3 T. Compared to rats similarly handled but not exposed to MRI procedures, brain mannitol concentration, expressed as a percentage of mean body concentration, was significantly increased at 0.3 T and 0.5 T but not at 1.5 T. At 0.3 T, exposure to gradient-field fluctuations used for imaging increased brain mannitol concentration, but exposures to static main field and pulsed radiofrequency energies did not. Increased brain mannitol associated with gradient-field flux may reflect increased blood-brain barrier permeability or blood volume in brain. MRI effects on brain mannitol space are of uncertain clinical significance, but are consistent with prior evidence of an MRI-induced increase of brain capillary endothelial cell transport observed with horseradish peroxidase. Further studies are needed to confirm these findings and to explore the processes underlying changes in mannitol distribution related to MRI.     

Influence of thickening of the inner skull table on intracranial volume measurement in older people

Introduction

It is generally assumed that intracranial volume (ICV) remains constant after peaking in early adulthood. Thus ICV is used as a ‘proxy’ for original brain size when trying to estimate brain atrophy in older people in neuroimaging studies. However, physiological changes in the skull, such as thickening of the frontal inner table, are relatively common in older age and will reduce ICV. The potential influence that inner table skull thickening may have on ICV measurement in old age has yet to be investigated.

Methods

We selected 60 (31 males, 29 females) representative older adults aged 71.1–74.3 years from a community-dwelling ageing cohort, the Lothian Birth Cohort 1936. A semi-automatically derived current ICV measurement obtained from high resolution T1-weighted volume scans was compared to the estimated original ICV by excluding inner skull table thickening using expert manual image processing.

Results

Inner table skull thickening reduced ICV from an estimated original 1480.0 ml to a current 1409.1 ml, a median decrease of 7.3% (Z = − 6.334; p < 0.001), and this reduction was more prominent in women than men (median decrease 114.6 vs. 101.9 ml respectively). This led to potential significant underestimations of brain atrophy in this sample by 5.3% (p < 0.001) and obscured potential gender differences.

Conclusions

The effects of skull thickening are important to consider when conducting research in ageing, as they can obscure gender differences and result in underestimation of brain atrophy. Research into reliable methods of determining the estimated original ICV is required for research into brain ageing.     

Liver schwannoma: findings on MRI

Magnetic resonance imaging features of benign liver schwannoma in a 52-year-old woman are described. An oval shaped, 4.4×3.6×2.9-cm lesion was located in Segment 7 of the right hepatic lobe. The lesion was hypointense on T₁-weighted images and mixed hypo- and hyperintense on T₂-weighted images. On serial contrast-enhanced images, the lesion revealed gradually increasing centrilobular enhancement. The tumor was surgically removed thereafter.     

Magnetic properties of novel superparamagnetic MRI contrast agents based on colloidal nanocrystals

Novel systems based on colloidal magnetic nanocrystals (NCs), potentially useful as superparamagnetic (SP) contrast agents for magnetic resonance imaging (MRI) have been investigated. The NCs we have studied comprise organic-capped single-crystalline maghemite (γ-Fe₂O₃) cores possessing controlled sizes and shapes. We have comparatively examined spherical and tetrapod-like NCs, the latter being branched particles possessing four arms which depart out at tetrahedral angles from a central point. The as-synthesized NCs are passivated by hydrophobic surfactant molecules and thus are fully dispersible in nonpolar media only. The NCs have been made soluble in aqueous solution by applying a procedure based on the surface intercalation and coating with an amphiphilic polymer shell. NMR relaxivities R₁ and R₂ were compared with ENDOREM^®, one of the standard commercial SP-MRI contrast agent. We found that the spherical NCs exhibit R₁ and R₂ relaxivities slightly lower than those of ENDOREM^®, over the whole frequency range; on the contrary, tetrapods show relaxivities about one order of magnitude lower. The physical origin of such difference in relaxivities between tetrapod- and spheres-based nanostructures is under investigation and it is possibly related to different sources of the magnetic anisotropy.     

In vivo MRI reveals the dynamics of pathological changes in the brains of cathepsin D-deficient mice and correlates changes in manganese-enhanced MRI with microglial activation

Cathepsin D (CTSD; EC 3.4.23.5) is essential for normal development and/or maintenance of neurons in the central nervous system: its deficiency causes a devastating neurological disorder with severely shortened life span in man, sheep and mouse. Neuropathologically, the CTSD deficiencies are characterized by selective neuronal degeneration, gliosis and accumulation of autofluorescent proteinaceous storage material in neurons. Our aim was to study the dynamics behind the pathological alterations occurring in the brains of CTSD-deficient (CTSD-/-) mice by using in vivo magnetic resonance imaging (MRI) and histology. In order to do this, we measured T(2) signal intensity (SI), apparent diffusion coefficient, area and volume of multiple brain structures from MR images acquired using T(2)-, T(1)- and diffusion-weighted sequences at three time points during disease progression. MRI revealed no differences in the brains between CTSD-/- and control mice at postnatal day 15+/-1 (P15+/-1), representing an initial stage of the disease. In the intermediate stage of the disease, P19(+/-1), SI alterations in the thalami of the affected mice became evident in both T(1)- and T(2)-weighted images. The terminal stage of the disease, P25, was characterized by marked alterations in the T(2) SI, apparent diffusion coefficient and volume of multiple brain structures in CTSD-/- mice. In addition, manganese enhanced high-resolution T(1)-weighted 3D sequences (MEMRI) and histological stainings revealed that the hyperintense signal areas in MEMRI matched perfectly with areas of microglial activation in the brains of CTSD-/- mice at the terminal disease stage. In conclusion, the SI alterations in the thalami of CTSD-/- mice preceded other changes, and the degenerative process was greatly enhanced at the age P19(+/-1), leading to severely reduced brain volume in just 6 days.     

The use of multivariate MR imaging intensities versus metabolic data from MR spectroscopic imaging for brain tumour classification

This study investigated the value of information from both magnetic resonance imaging and magnetic resonance spectroscopic imaging (MRSI) to automated discrimination of brain tumours. The influence of imaging intensities and metabolic data was tested by comparing the use of MR spectra from MRSI, MR imaging intensities, peak integration values obtained from the MR spectra and a combination of the latter two. Three classification techniques were objectively compared: linear discriminant analysis, least squares support vector machines (LS-SVM) with a linear kernel as linear techniques and LS-SVM with radial basis function kernel as a nonlinear technique. Classifiers were evaluated over 100 stratified random splittings of the dataset into training and test sets. The area under the receiver operating characteristic (ROC) curve (AUC) was used as a global performance measure on test data. In general, all techniques obtained a high performance when using peak integration values with or without MR imaging intensities. For example for low- versus high-grade tumours, low- versus high-grade gliomas and gliomas versus meningiomas, the mean test AUC was higher than 0.91, 0.94, and 0.99, respectively, when both MR imaging intensities and peak integration values were used. The use of metabolic data from MRSI significantly improved automated classification of brain tumour types compared to the use of MR imaging intensities solely.     

Renal subcapsular pseudocyst: findings on MRI

Pancreatic pseudocysts are common complications of pancreatitis. They may occur in unusual places as a result of spread along the path of least resistance. Penetration of Gerota's fascia and spread into the renal subcapsular space are uncommon. Their appearance on magnetic resonance imaging has not been previously documented. Given the increased use of magnetic resonance cholangiopancreatography for the investigation of patients with pancreatitis, the recognition that a complicated renal cyst in this clinical context could represent a benign condition is important.     

Effects of random subject rotation on optimised diffusion gradient sampling schemes in diffusion tensor MRI

The choice of the number (N) and orientation of diffusion sampling gradients required to measure accurately the water diffusion tensor remains contentious. Monte Carlo studies have suggested that between 20 and 30 uniformly distributed sampling orientations are required to provide robust estimates of water diffusions parameters. These simulations have not, however, taken into account what effect random subject motion, specifically rotation, might have on optimised gradient schemes, a problem which is especially relevant to clinical diffusion tensor MRI (DT-MRI). Here this question is investigated using Monte Carlo simulations of icosahedral sampling schemes and in vivo data. These polyhedra-based schemes, which have the advantage that large N can be created from optimised subsets of smaller N, appear to be ideal for the study of restless subjects since if scanning needs to be prematurely terminated it should be possible to identify a subset of images that have been acquired with a near optimised sampling scheme. The simulations and in vivo data show that as N increases, the rotational variance of fractional anisotropy (FA) estimates becomes progressively less dependent on the magnitude of subject rotation (), while higher FA values are progressively underestimated as increases. These data indicate that for large subject rotations the B-matrix should be recalculated to provide accurate diffusion anisotropy information.     

Initial attempts at directly detecting alpha wave activity in the brain using MRI

The “direct detection” of neuronal activity by MRI could offer improved spatial and temporal resolution compared to the blood oxygenation level-dependent (BOLD) effect. Here we describe initial attempts to use MRI to detect directly the neuronal currents resulting from spontaneous alpha wave activity, which have previously been shown to generate the largest extracranial magnetic fields. Experiments were successfully carried out on four subjects at 3 T. A single slice was imaged at a rate of 25 images per second under two conditions. The first (in darkness with eyes-closed) was chosen to promote alpha wave activity, while the second (eyes-open viewing a visual stimulus) was chosen to suppress it. The fluctuations of the phase and magnitude of the resulting MR image data were frequency analysed, and tested for the signature of both alpha wave activity and neuronal activity evoked by the visual stimulus.

Regions were found that consistently showed elevated power in fluctuations of the phase of the MR signal, in the frequency range of alpha waves, during the eyes-closed condition. It was conservatively assumed that if oscillations occurred at the same frequency in the magnitude signal from the same region or at the same frequency in the phase or magnitude signal from other regions overlying large vessels or cerebrospinal fluid (CSF), then the phase changes were not due to neuronal activity related to alpha waves. Using these criteria the data obtained were consistent with direct detection of alpha wave activity in three of the four volunteers. No significant MR signal fluctuations due to evoked activity were identified.     

Fast MRI coil analysis based on 3-D electromagnetic and RF circuit co-simulation

To accelerate the analysis of a multi-element MRI coil, a two-way link is used between radiofrequency (RF) circuit and 3-D electromagnetic (EM) simulation tools. In this configuration, only one 3-D EM simulation is required to investigate the coil performance over a range of different tunings, saving considerable computation time. For the purpose of 3-D EM simulation, the coil feed networks and trim capacitors are substituted by 50 Ω ports. The entire coil was tuned in the RF circuit domain, and the near-field profiles of the electric and magnetic field components were then calculated, together with the specific energy absorption ratio (SAR) maps in the 3-D EM domain     

Automated classification of multispectral MR images using unsupervised constrained energy minimization based on fuzzy logic

Constrained energy minimization (CEM) has proven highly effective for hyperspectral (or multispectral) target detection and classification. It requires a complete knowledge of the desired target signature in images. This work presents “Unsupervised CEM (UCEM),” a novel approach to automatically target detection and classification in multispectral magnetic resonance (MR) images. The UCEM involves two processes, namely, target generation process (TGP) and CEM. The TGP is a fuzzy-set process that generates a set of potential targets from unknown information and then applies these targets to be desired targets in CEM. Finally, two sets of images, namely, computer-generated phantom images and real MR images, are used in the experiments to evaluate the effectiveness of UCEM. Experimental results demonstrate that UCEM segments a multispectral MR image much more effectively than either Functional MRI of the Brain's (FMRIB's) automated segmentation tool or fuzzy C-means does.     

Multispectral MR images segmentation based on fuzzy knowledge and modified seeded region growing

Magnetic resonance imaging (MRI) is a valuable diagnostic tool in medical science due to its capability for soft-tissue characterization and three-dimensional visualization. One potential application of MRI in clinical practice is brain parenchyma classification and segmentation. Based on fuzzy knowledge and modified seeded region growing, this work proposes a novel image segmentation method, called Fuzzy Knowledge-Based Seeded Region Growing (FKSRG), for multispectral MR images. In this work, fuzzy knowledge includes the fuzzy edge, fuzzy similarity and fuzzy distance, which are obtained from relationships between pixels in multispectral MR images and are applied to the modified seeded regions growing process. In conventional regions merging, the final number of regions is unknown. Therefore, a Target Generation Process is proposed and applied to support conventional regions merging, such that the FKSRG method does not over- or undersegment images. Finally, two image sets, namely, computer-generated phantom images and real MR images, are used in experiments to assess the effectiveness of the proposed FKSRG method. Experimental results demonstrate that the FKSRG method segments multispectral MR images much more effectively than the Functional MRI of the Brain Automated Segmentation Tool, K-means and Support Vector Machine methods.     

Transfer characteristics of arterial pulsatile force in regional intracranial tissue using dynamic diffusion MRI: A phantom study

Introduction

To clarify the mechanism underlying apparent diffusion coefficient (ADC) changes in regional intracranial tissue during the cardiac cycle, we investigated relationships among ADC changes, volume loading, and intracranial pressure using a hemodialyzer phantom in magnetic resonance imaging (MRI).

Materials and Methods

The hemodialyzer phantom consisted of hollow fibers (HF), the external space of HFs (ES), and a gateway of dialysis solution, filled with syrup solution and air. The high-volume and low-volume loadings were periodically applied to HFs by a to-and-fro flow pump, and syrup solution was permitted to enter or leave HFs during the volume loading cycles. ADC maps at each volume loading phase were obtained using ECG-triggered single-shot diffusion echo-planar imaging. Dynamic phase contrast MRI was also used to measure volume loading to the phantom. We compared the ADC changes, volume loading, and intracranial pressure in the phantom during the cardiac cycle.

Results

ADC changes synchronized significantly with absolute volumetric flow rate change. The maximum ADC change was higher in high-volume loading cycles than in low-volume loading cycles. Results showed that the water molecules in ES fluctuated according to the force transferred from HF to ES. ADC changes were dependent upon the volumetric flow rate during the cardiac cycle.

Conclusions

Our original phantom allowed us to clarify the mechanism underlying water fluctuations in intracranial tissues. Measurement of maximum changes in ADC is an effective method to define the transfer characteristics of the arterial pulsatile force in regional intracranial tissue.     

A realization of digital wireless transmission for MRI signals based on 802.11b

In this paper, a digital wireless transmission system based on 802.11b standard for magnetic resonance imaging (MRI) application is designed and built for the first time to eliminate the interference aroused by coil array cables. The analysis shows that the wireless receiver has a very high sensitivity to detect MRI signals. The modulation technique of differential quadrature phase shift keyed (DQPSK) can be applied to MRI data transmission with rate of 2 Mbps and bandwidth of 2 MHz. The bench test verifies that this wireless link has a dynamic range over 86 dB supporting up to 3 T MRI system data transmission. The 2D spin echo imaging of phantom is performed and the SNR of the image obtained by the wireless transmission can be comparable with that got by the coaxial cables.     

Brain tissue classification based on DTI using an improved Fuzzy C-means algorithm with spatial constraints

We present an effective method for brain tissue classification based on diffusion tensor imaging (DTI) data. The method accounts for two main DTI segmentation obstacles: random noise and magnetic field inhomogeneities. In the proposed method, DTI parametric maps were used to resolve intensity inhomogeneities of brain tissue segmentation because they could provide complementary information for tissues and define accurate tissue maps. An improved fuzzy c-means with spatial constraints proposal was used to enhance the noise and artifact robustness of DTI segmentation. Fuzzy c-means clustering with spatial constraints (FCM_S) could effectively segment images corrupted by noise, outliers, and other imaging artifacts. Its effectiveness contributes not only to the introduction of fuzziness for belongingness of each pixel but also to the exploitation of spatial contextual information. We proposed an improved FCM_S applied on DTI parametric maps, which explores the mean and covariance of the feature spatial information for automated segmentation of DTI. The experiments on synthetic images and real-world datasets showed that our proposed algorithms, especially with new spatial constraints, were more effective.     

An MRI denoising method using image data redundancy and local SNR estimation

This paper presents an LMMSE-based method for the three-dimensional (3D) denoising of MR images assuming a Rician noise model. Conventionally, the LMMSE method estimates the noise-less signal values using the observed MR data samples within local neighborhoods. This is not an efficient procedure to deal with this issue while the 3D MR data intrinsically includes many similar samples that can be used to improve the estimation results. To overcome this problem, we model MR data as random fields and establish a principled way which is capable of choosing the samples not only from a local neighborhood but also from a large portion of the given data. To follow the similar samples within the MR data, an effective similarity measure based on the local statistical moments of images is presented. The parameters of the proposed filter are automatically chosen from the estimated local signal-to-noise ratio. To further enhance the denoising performance, a recursive version of the introduced approach is also addressed. The proposed filter is compared with related state-of-the-art filters using both synthetic and real MR datasets. The experimental results demonstrate the superior performance of our proposal in removing the noise and preserving the anatomical structures of MR images.     

Automatic segmentation of brain MRI in high-dimensional local and non-local feature space based on sparse representation

Automatic extraction of the varying regions of magnetic resonance images is required as a prior step in a diagnostic intelligent system. The sparsest representation and high-dimensional feature are provided based on learned dictionary. The classification is done by employing the technique that computes the reconstruction error locally and non-locally of each pixel. The acquired results from the real and simulated images are superior to the best MRI segmentation method with regard to the stability advantages. In addition, it is segmented exactly through a formula taken from the distance and sparse factors. Also, it is done automatically taking sparse factor in unsupervised clustering methods whose results have been improved.     

Calculated signal-to-noise ratio of MRI detected with SQUIDs and Faraday detectors in fields from 10 microT to 1.5 T

We examine the calculated signal-to-noise ratio (SNR) achievable with different MRI detection modalities in precession fields ranging from 10 microT to 1.5 T. In particular, we compare traditional Faraday detectors with both tuned and untuned detectors based on superconducting quantum interference devices (SQUIDs). We derive general expressions for the magnetic field noise due to the samples and the detectors, and then calculate the SNR achievable for a specific geometry with each modality with and without prepolarization. We show that each of the three modalities is superior in one of the three field ranges. SQUID-based detection is superior to conventional Faraday detection for MRI in precession fields below 250 mT for a 65 mm diameter surface coil placed a distance of 25 mm from the voxel of interest embedded in a cylinder of tissue 50 mm tall and of radius 50 mm. This crossover field, however, is sensitive to the geometry.     

MRI allows for longitudinal quantitative analysis of body fat composition in rats: An analysis of sibutramine-associated changes at the group level

Purpose

Body fat distribution changes are associated with multiple alterations in metabolism. Therefore, the assessment of body fat compartments by MRI in animal models is a promising approach to obesity research. Standard T1-weighted (T1w) whole body MRI was used here to quantify different effects in the subcutaneous and visceral fat compartments in rats under treatment with an anorexiant.

Materials and methods

Twenty rats on a high caloric diet were investigated by the identical MRI protocol at baseline and after seven weeks. Ten rats received a treatment with sibutramine, 10 rats served as vehicle control group. To longitudinally assess body fat components, MRI analysis was used with two approaches: 2D slicewise graphic analysis (SGA) was compared with an automated 3D analysis algorithm (3DA).

Results

At the group level, fat volume differences showed a longitudinal increase of subcutaneous and visceral fat volumes for the control group, whereas the sibutramine group showed stable subcutaneous fat volumes and decrease in visceral fat volumes. SGA and 3DA volume determination showed significant correlations for subcutaneous fat volume (C = 0.85, p < 0.001), visceral fat volume (C = 0.87, p < 0.001), and total fat volume (C = 0.90, p < 0.001).

Conclusion

It could be demonstrated that computer-based analysis of T1w MRI could be used to longitudinally assess changes in body fat compartments in rats at the group level. In detail, it was possible to investigate the effect of sibutramine separate on the fat compartments in rats.