Automated model-based tissue classification of MR images of the brain

Multi-atlas learner fusion: An efficient segmentation approach for large-scale data

We propose multi-atlas learner fusion (MLF), a framework for rapidly and accurately replicating the highly accurate, yet computationally expensive, multi-atlas segmentation framework based on fusing local learners. In the largest whole-brain multi-atlas study yet reported, multi-atlas segmentations are estimated for a training set of 3464 MR brain images. Using these multi-atlas estimates we (1) estimate a low-dimensional representation for selecting locally appropriate example images, and (2) build AdaBoost learners that map a weak initial segmentation to the multi-atlas segmentation result. Thus, to segment a new target image we project the image into the low-dimensional space, construct a weak initial segmentation, and fuse the trained, locally selected, learners. The MLF framework cuts the runtime on a modern computer from 36 h down to 3-8 min - a 270× speedup - by completely bypassing the need for deformable atlas-target registrations. Additionally, we (1) describe a technique for optimizing the weak initial segmentation and the AdaBoost learning parameters, (2) quantify the ability to replicate the multi-atlas result with mean accuracies approaching the multi-atlas intra-subject reproducibility on a testing set of 380 images, (3) demonstrate significant increases in the reproducibility of intra-subject segmentations when compared to a state-of-the-art multi-atlas framework on a separate reproducibility dataset, (4) show that under the MLF framework the large-scale data model significantly improve the segmentation over the small-scale model under the MLF framework, and (5) indicate that the MLF framework has comparable performance as state-of-the-art multi-atlas segmentation algorithms without using non-local information.

Comparison of accelerated T1-weighted whole-brain structural-imaging protocols

Imaging in neuroscience, clinical research and pharmaceutical trials often employs the 3D magnetisation-prepared rapid gradient-echo (MPRAGE) sequence to obtain structural T1-weighted images with high spatial resolution of the human brain. Typical research and clinical routine MPRAGE protocols with ~1mm isotropic resolution require data acquisition time in the range of 5-10min and often use only moderate two-fold acceleration factor for parallel imaging. Recent advances in MRI hardware and acquisition methodology promise improved leverage of the MR signal and more benign artefact properties in particular when employing increased acceleration factors in clinical routine and research. In this study, we examined four variants of a four-fold-accelerated MPRAGE protocol (2D-GRAPPA, CAIPIRINHA, CAIPIRINHA elliptical, and segmented MPRAGE) and compared clinical readings, basic image quality metrics (SNR, CNR), and automated brain tissue segmentation for morphological assessments of brain structures. The results were benchmarked against a widely-used two-fold-accelerated 3T ADNI MPRAGE protocol that served as reference in this study. 22 healthy subjects (age=20-44yrs.) were imaged with all MPRAGE variants in a single session. An experienced reader rated all images of clinically useful image quality. CAIPIRINHA MPRAGE scans were perceived on average to be of identical value for reading as the reference ADNI-2 protocol. SNR and CNR measurements exhibited the theoretically expected performance at the four-fold acceleration. The results of this study demonstrate that the four-fold accelerated protocols introduce systematic biases in the segmentation results of some brain structures compared to the reference ADNI-2 protocol. Furthermore, results suggest that the increased noise levels in the accelerated protocols play an important role in introducing these biases, at least under the present study conditions.

Assessing atrophy measurement techniques in dementia: Results from the MIRIAD atrophy challenge

Structural MRI is widely used for investigating brain atrophy in many neurodegenerative disorders, with several research groups developing and publishing techniques to provide quantitative assessments of this longitudinal change. Often techniques are compared through computation of required sample size estimates for future clinical trials. However interpretation of such comparisons is rendered complex because, despite using the same publicly available cohorts, the various techniques have been assessed with different data exclusions and different statistical analysis models. We created the MIRIAD atrophy challenge in order to test various capabilities of atrophy measurement techniques. The data consisted of 69 subjects (46 Alzheimer's disease, 23 control) who were scanned multiple (up to twelve) times at nine visits over a follow-up period of one to two years, resulting in 708 total image sets. Nine participating groups from 6 countries completed the challenge by providing volumetric measurements of key structures (whole brain, lateral ventricle, left and right hippocampi) for each dataset and atrophy measurements of these structures for each time point pair (both forward and backward) of a given subject. From these results, we formally compared techniques using exactly the same dataset. First, we assessed the repeatability of each technique using rates obtained from short intervals where no measurable atrophy is expected. For those measures that provided direct measures of atrophy between pairs of images, we also assessed symmetry and transitivity. Then, we performed a statistical analysis in a consistent manner using linear mixed effect models. The models, one for repeated measures of volume made at multiple time-points and a second for repeated “direct” measures of change in brain volume, appropriately allowed for the correlation between measures made on the same subject and were shown to fit the data well. From these models, we obtained estimates of the distribution of atrophy rates in the Alzheimer's disease (AD) and control groups and of required sample sizes to detect a 25% treatment effect, in relation to healthy ageing, with 95% significance and 80% power over follow-up periods of 6, 12, and 24 months. Uncertainty in these estimates, and head-to-head comparisons between techniques, were carried out using the bootstrap. The lateral ventricles provided the most stable measurements, followed by the brain. The hippocampi had much more variability across participants, likely because of differences in segmentation protocol and less distinct boundaries. Most methods showed no indication of bias based on the short-term interval results, and direct measures provided good consistency in terms of symmetry and transitivity. The resulting annualized rates of change derived from the model ranged from, for whole brain: − 1.4% to − 2.2% (AD) and − 0.35% to − 0.67% (control), for ventricles: 4.6% to 10.2% (AD) and 1.2% to 3.4% (control), and for hippocampi: − 1.5% to − 7.0% (AD) and − 0.4% to − 1.4% (control). There were large and statistically significant differences in the sample size requirements between many of the techniques. The lowest sample sizes for each of these structures, for a trial with a 12 month follow-up period, were 242 (95% CI: 154 to 422) for whole brain, 168 (95% CI: 112 to 282) for ventricles, 190 (95% CI: 146 to 268) for left hippocampi, and 158 (95% CI: 116 to 228) for right hippocampi. This analysis represents one of the most extensive statistical comparisons of a large number of different atrophy measurement techniques from around the globe. The challenge data will remain online and publicly available so that other groups can assess their methods.

•

We compared numerous brain atrophy measurement techniques using multiple metrics.

•

Each participant produced measures on the exact same dataset, blinded to disease.

•

A central statistical analysis using linear mixed effect models was performed.

•

Head to head comparisons for each region were performed using sample size estimates.

•

Brain and ventricle measures were more consistent across groups than for hippocampi.

Fat may affect magnetic resonance signal intensity and brain tissue volumes

Obesity/overweight is reported to affect MR-measured brain tissue volume and white matter (WM) signal intensity. This study investigated possible effects of fat on these measures, using pig fat on three participants at a 4T magnet. Grey matter volumes in the presence of fat were lower than baseline measures. Total WM volumes in the presence of fat were higher than baseline measures. WM hypo-intensities on T1-weighted images were higher in the presence of fat than baseline measures. Therefore physical effects of head fat of obese/overweight individual may at least, partly contribute to the association of obesity/overweight with MR structural measures.

Fractal dimension analysis of cerebellum in Chiari Malformation type I

Chiari Malformation type I (CM-I) is a serious neurological disorder that is characterized by hindbrain herniation. Our aim was to evaluate the usefulness of fractal analysis in CM-I patients. To examine the morphological complexity features of this disorder, fractal dimension (FD) of cerebellar regions were estimated from magnetic resonance images (MRI) of 17 patients with CM-I and 16 healthy control subjects in this study. The areas of white matter (WM), gray matter (GM) and cerebrospinal fluid (CSF) were calculated and the corresponding FD values were computed using a 2D box-counting method in both groups. The results indicated that CM-I patients had significantly higher (p<0.05) FD values of GM, WM and CSF tissues compared to control group. According to the results of correlation analysis between FD values and the corresponding area values, FD and area values of GM tissues in the patients group were found to be correlated. The results of the present study suggest that FD values of cerebellar regions may be a discriminative feature and a useful marker for investigation of abnormalities in the cerebellum of CM-I patients. Further studies to explore the changes in cerebellar regions with the help of 3D FD analysis and volumetric calculations should be performed as a future work.

Tissue-specific sparse deconvolution for brain CT perfusion

Enhancing perfusion maps in low-dose computed tomography perfusion (CTP) for cerebrovascular disease diagnosis is a challenging task, especially for low-contrast tissue categories where infarct core and ischemic penumbra usually occur. Sparse perfusion deconvolution has been recently proposed to effectively improve the image quality and diagnostic accuracy of low-dose perfusion CT by extracting the complementary information from the high-dose perfusion maps to restore the low-dose using a joint spatio-temporal model. However the low-contrast tissue classes where infarct core and ischemic penumbra are likely to occur in cerebral perfusion CT tend to be over-smoothed, leading to loss of essential biomarkers. In this paper, we propose a tissue-specific sparse deconvolution approach to preserve the subtle perfusion information in the low-contrast tissue classes. We first build tissue-specific dictionaries from segmentations of high-dose perfusion maps using online dictionary learning, and then perform deconvolution-based hemodynamic parameters estimation for block-wise tissue segments on the low-dose CTP data. Extensive validation on clinical datasets of patients with cerebrovascular disease demonstrates the superior performance of our proposed method compared to state-of-art, and potentially improve diagnostic accuracy by increasing the differentiation between normal and ischemic tissues in the brain.

Shape index distribution based local surface complexity applied to the human cortex

The quantification of local surface complexity in the human cortex has shown to be of interest in investigating population differences as well as developmental changes in neurodegenerative or neurodevelopment diseases. We propose a novel assessment method that represents local complexity as the difference between the observed distributions of local surface topology to its best-fit basic topology model within a given local neighborhood. This distribution difference is estimated via Earth Move Distance (EMD) over the histogram within the local neighborhood of the surface topology quantified via the Shape Index (SI) measure. The EMD scores have a range from simple complexity (0.0), which indicates a consistent local surface topology, up to high complexity (1.0), which indicates a highly variable local surface topology. The basic topology models are categorized as 9 geometric situation modeling situations such as crowns, ridges and fundi of cortical gyro and sulci. We apply a geodesic kernel to calculate the local SI histrogram distribution within a given region. In our experiments, we obtained the results of local complexity that shows generally higher complexity in the gyral/sulcal wall regions and lower complexity in some gyral ridges and lowest complexity in sulcal fundus areas. In addition, we show expected, preliminary results of increased surface complexity across most of the cortical surface within the first years of postnatal life, hypothesized to be due to the changes such as development of sulcal pits.

Hierarchical max-flow segmentation framework for multi-atlas segmentation with Kohonen self-organizing map based Gaussian mixture modeling

The incorporation of intensity, spatial, and topological information into large-scale multi-region segmentation has been a topic of ongoing research in medical image analysis. Multi-region segmentation problems, such as segmentation of brain structures, pose unique challenges in image segmentation in which regions may not have a defined intensity, spatial, or topological distinction, but rely on a combination of the three. We propose a novel framework within the Advanced segmentation tools (ASETS)(2), which combines large-scale Gaussian mixture models trained via Kohonen self-organizing maps, with deformable registration, and a convex max-flow optimization algorithm incorporating region topology as a hierarchy or tree. Our framework is validated on two publicly available neuroimaging datasets, the OASIS and MRBrainS13 databases, against the more conventional Potts model, achieving more accurate segmentations. Each component is accelerated using general-purpose programming on graphics processing Units to ensure computational feasibility.

Reproducibility of multiphase pseudo-continuous arterial spin labeling and the effect of post-processing analysis methods

Arterial spin labeling (ASL) is an emerging MRI technique for non-invasive measurement of cerebral blood flow (CBF). Compared to invasive perfusion imaging modalities, ASL suffers from low sensitivity due to poor signal-to-noise ratio (SNR), susceptibility to motion artifacts and low spatial resolution, all of which limit its reliability. In this work, the effects of various state of the art image processing techniques for addressing these ASL limitations are investigated. A processing pipeline consisting of motion correction, ASL motion correction imprecision removal, temporal and spatial filtering, partial volume effect correction, and CBF quantification was developed and assessed. To further improve the SNR for pseudo-continuous ASL (PCASL) by accounting for errors in tagging efficiency, the data from multiphase (MP) acquisitions were analyzed using a novel weighted-averaging scheme. The performances of each step in terms of SNR and reproducibility were evaluated using test-retest ASL data acquired from 12 young healthy subjects. The proposed processing pipeline was shown to improve the within-subject coefficient of variation and regional reproducibility by 17% and 16%, respectively, compared to CBF maps computed following motion correction but without the other processing steps. The CBF measurements of MP-PCASL compared to PCASL had on average 23% and 10% higher SNR and reproducibility, respectively.

Fully automated whole-head segmentation with improved smoothness and continuity, with theory reviewed

Individualized current-flow models are needed for precise targeting of brain structures using transcranial electrical or magnetic stimulation (TES/TMS). The same is true for current-source reconstruction in electroencephalography and magnetoencephalography (EEG/MEG). The first step in generating such models is to obtain an accurate segmentation of individual head anatomy, including not only brain but also cerebrospinal fluid (CSF), skull and soft tissues, with a field of view (FOV) that covers the whole head. Currently available automated segmentation tools only provide results for brain tissues, have a limited FOV, and do not guarantee continuity and smoothness of tissues, which is crucially important for accurate current-flow estimates. Here we present a tool that addresses these needs. It is based on a rigorous Bayesian inference framework that combines image intensity model, anatomical prior (atlas) and morphological constraints using Markov random fields (MRF). The method is evaluated on 20 simulated and 8 real head volumes acquired with magnetic resonance imaging (MRI) at 1 mm3 resolution. We find improved surface smoothness and continuity as compared to the segmentation algorithms currently implemented in Statistical Parametric Mapping (SPM). With this tool, accurate and morphologically correct modeling of the whole-head anatomy for individual subjects may now be feasible on a routine basis. Code and data are fully integrated into SPM software tool and are made publicly available. In addition, a review on the MRI segmentation using atlas and the MRF over the last 20 years is also provided, with the general mathematical framework clearly derived.

Automatic segmentation and volumetry of multiple sclerosis brain lesions from MR images

The location and extent of white matter lesions on magnetic resonance imaging (MRI) are important criteria for diagnosis, follow-up and prognosis of multiple sclerosis (MS). Clinical trials have shown that quantitative values, such as lesion volumes, are meaningful in MS prognosis. Manual lesion delineation for the segmentation of lesions is, however, time-consuming and suffers from observer variability. In this paper, we propose MSmetrix, an accurate and reliable automatic method for lesion segmentation based on MRI, independent of scanner or acquisition protocol and without requiring any training data. In MSmetrix, 3D T1-weighted and FLAIR MR images are used in a probabilistic model to detect white matter (WM) lesions as an outlier to normal brain while segmenting the brain tissue into grey matter, WM and cerebrospinal fluid. The actual lesion segmentation is performed based on prior knowledge about the location (within WM) and the appearance (hyperintense on FLAIR) of lesions. The accuracy of MSmetrix is evaluated by comparing its output with expert reference segmentations of 20 MRI datasets of MS patients. Spatial overlap (Dice) between the MSmetrix and the expert lesion segmentation is 0.67 ± 0.11. The intraclass correlation coefficient (ICC) equals 0.8 indicating a good volumetric agreement between the MSmetrix and expert labelling. The reproducibility of MSmetrix' lesion volumes is evaluated based on 10 MS patients, scanned twice with a short interval on three different scanners. The agreement between the first and the second scan on each scanner is evaluated through the spatial overlap and absolute lesion volume difference between them. The spatial overlap was 0.69 ± 0.14 and absolute total lesion volume difference between the two scans was 0.54 ± 0.58 ml. Finally, the accuracy and reproducibility of MSmetrix compare favourably with other publicly available MS lesion segmentation algorithms, applied on the same data using default parameter settings.

Transfer learning improves supervised image segmentation across imaging protocols

The variation between images obtained with different scanners or different imaging protocols presents a major challenge in automatic segmentation of biomedical images. This variation especially hampers the application of otherwise successful supervised-learning techniques which, in order to perform well, often require a large amount of labeled training data that is exactly representative of the target data. We therefore propose to use transfer learning for image segmentation. Transfer-learning techniques can cope with differences in distributions between training and target data, and therefore may improve performance over supervised learning for segmentation across scanners and scan protocols. We present four transfer classifiers that can train a classification scheme with only a small amount of representative training data, in addition to a larger amount of other training data with slightly different characteristics. The performance of the four transfer classifiers was compared to that of standard supervised classification on two magnetic resonance imaging brain-segmentation tasks with multi-site data: white matter, gray matter, and cerebrospinal fluid segmentation; and white-matter-/MS-lesion segmentation. The experiments showed that when there is only a small amount of representative training data available, transfer learning can greatly outperform common supervised-learning approaches, minimizing classification errors by up to 60%.

LOGISMOS-B for Primates: Primate Cortical Surface Reconstruction and Thickness Measurement

Cortical thickness and surface area are important morphological measures with implications for many psychiatric and neurological conditions. Automated segmentation and reconstruction of the cortical surface from 3D MRI scans is challenging due to the variable anatomy of the cortex and its highly complex geometry. While many methods exist for this task in the context of the human brain, these methods are typically not readily applicable to the primate brain. We propose an innovative approach based on our recently proposed human cortical reconstruction algorithm, LOGISMOS-B, and the Laplace-based thickness measurement method. Quantitative evaluation of our approach was performed based on a dataset of T1- and T2-weighted MRI scans from 12-month-old macaques where labeling by our anatomical experts was used as independent standard. In this dataset, LOGISMOS-B has an average signed surface error of 0.01 ± 0.03mm and an unsigned surface error of 0.42 ± 0.03mm over the whole brain. Excluding the rather problematic temporal pole region further improves unsigned surface distance to 0.34 ± 0.03mm. This high level of accuracy reached by our algorithm even in this challenging developmental dataset illustrates its robustness and its potential for primate brain studies.

Bayesian segmentation of brainstem structures in MRI

In this paper we present a method to segment four brainstem structures (midbrain, pons, medulla oblongata and superior cerebellar peduncle) from 3D brain MRI scans. The segmentation method relies on a probabilistic atlas of the brainstem and its neighboring brain structures. To build the atlas, we combined a dataset of 39 scans with already existing manual delineations of the whole brainstem and a dataset of 10 scans in which the brainstem structures were manually labeled with a protocol that was specifically designed for this study. The resulting atlas can be used in a Bayesian framework to segment the brainstem structures in novel scans. Thanks to the generative nature of the scheme, the segmentation method is robust to changes in MRI contrast or acquisition hardware. Using cross validation, we show that the algorithm can segment the structures in previously unseen T1 and FLAIR scans with great accuracy (mean error under 1mm) and robustness (no failures in 383 scans including 168 AD cases). We also indirectly evaluate the algorithm with a experiment in which we study the atrophy of the brainstem in aging. The results show that, when used simultaneously, the volumes of the midbrain, pons and medulla are significantly more predictive of age than the volume of the entire brainstem, estimated as their sum. The results also demonstrate that the method can detect atrophy patterns in the brainstem structures that have been previously described in the literature. Finally, we demonstrate that the proposed algorithm is able to detect differential effects of AD on the brainstem structures. The method will be implemented as part of the popular neuroimaging package FreeSurfer.

MRI predicts remission at 1 year in first-episode schizophrenia in females with larger striato-thalamic volumes

BACKGROUND/AIMS

The Remission in Schizophrenia Working Group has defined remission as 'a low-mild symptom intensity level, maintained for a minimum of 6 months, where such symptoms do not affect an individual's behaviour' [Andreasen et al.: Am J Psychiatry 2005;162:441-449]. Since brain morphology relates to symptomatology, treatment and illness progression, MRI may assist in predicting remission.

METHODS

Thirty-nine patients newly diagnosed with DSM-IV schizophrenia underwent MRI brain scan prior to antipsychotic exposure. The Global Assessment of Functioning (GAF) score was entered into a voxel-based analysis to evaluate its relationship with cerebral grey matter volume from the baseline MRI. We entered age, total intracranial volume and intake GAF score as co-variates. Males and females were analysed separately because gender is a potent determinant of outcome.

RESULTS

Males had lower GAF scores than females, both at intake and at 1 year. Males comprised only 40% (12 out of 39) of the early remission group. For females only, early remission was strongly and positively correlated with bilateral lentiform and striatal volumes. For males, there was no such relationship.

CONCLUSION

Larger striato-thalamic volume correlated with early remission in females only. These baseline MRI findings were unlikely to be confounded by antipsychotic treatment and chronicity. These brain morphological markers show gender dimorphism and may assist in the prediction of early remission in newly diagnosed schizophrenia.

Asymmetric mixture model with simultaneous feature selection and model detection

A mixture model based on the symmetric Gaussian distribution that simultaneously treats the feature selection, and the model detection has recently received great attention for pattern recognition problems. However, in many applications, the distribution of the data has a non-Gaussian and nonsymmetric form. This brief presents a new asymmetric mixture model for model detection and model selection. In this brief, the proposed asymmetric distribution is modeled with multiple student's- t distributions, which are heavily tailed and more robust than Gaussian distributions. Our method has the flexibility to fit different shapes of observed data, such as non-Gaussian and nonsymmetric. Another advantage is that the proposed algorithm, which is based on the variational Bayesian learning, can simultaneously optimize over the number of the student's- t distribution that is used to model each asymmetric distribution, the number of components, and the saliency of the features. Numerical experiments on both synthetic and real-world datasets are conducted. The performance of the proposed model is compared with other mixture models, demonstrating the robustness, accuracy, and effectiveness of our method.

Abnormalities in cortical gray matter density in borderline personality disorder

BACKGROUND

Borderline personality disorder (BPD) is a chronic condition with a strong impact on patients' affective, cognitive and social functioning. Neuroimaging techniques offer invaluable tools to understand the biological substrate of the disease. We aimed to investigate gray matter alterations over the whole cortex in a group of Borderline Personality Disorder (BPD) patients compared to healthy controls (HC).

METHODS

Magnetic resonance-based cortical pattern matching was used to assess cortical gray matter density (GMD) in 26 BPD patients and in their age- and sex-matched HC (age: 38 ± 11; females: 16, 61%).

RESULTS

BPD patients showed widespread lower cortical GMD compared to HC (4% difference) with peaks of lower density located in the dorsal frontal cortex, in the orbitofrontal cortex, the anterior and posterior cingulate, the right parietal lobe, the temporal lobe (medial temporal cortex and fusiform gyrus) and in the visual cortex (P<0.005). Our BPD subjects displayed a symmetric distribution of anomalies in the dorsal aspect of the cortical mantle, but a wider involvement of the left hemisphere in the mesial aspect in terms of lower density. A few restricted regions of higher density were detected in the right hemisphere. All regions remained significant after correction for multiple comparisons via permutation testing.

CONCLUSIONS

BPD patients feature specific morphology of the cerebral structures involved in cognitive and emotional processing and social cognition/mentalization, consistent with clinical and functional data.

Automated segmentation of breast in 3-D MR images using a robust atlas

This paper presents a robust atlas-based segmentation (ABS) algorithm for segmentation of the breast boundary in 3-D MR images. The proposed algorithm combines the well-known methodologies of ABS namely probabilistic atlas and atlas selection approaches into a single framework where two configurations are realized. The algorithm uses phase congruency maps to create an atlas which is robust to intensity variations. This allows an atlas derived from images acquired with one MR imaging sequence to be used to segment images acquired with a different MR imaging sequence and eliminates the need for intensity-based registration. Images acquired using a Dixon sequence were used to create an atlas which was used to segment both Dixon images (intra-sequence) and T1-weighted images (inter-sequence). In both cases, highly accurate results were achieved with the median Dice similarity coefficient values of 94% ±4% and 87 ±6.5%, respectively.

A preliminary examination of cortical neurotransmitter levels associated with heavy drinking in posttraumatic stress disorder

Posttraumatic stress disorder (PTSD) patients have low cortical concentrations of γ-aminobutyric acid (GABA) and elevated glutamate (Glu) as measured by proton magnetic resonance spectroscopy ((1)H MRS). Alcohol use disorder (AUD) is highly comorbid with PTSD, but the neurobiological underpinnings are largely unknown. We wanted to determine if PTSD patients with AUD have normalized cortical GABA and Glu levels in addition to metabolite alterations common to AUD. We compared brain metabolite concentrations in 10 PTSD patients with comorbid AUD (PAUD) with concentrtations in 28 PTSD patients without AUD and in 20 trauma-exposed controls (CON) without PTSD symptoms. We measured concentrations of GABA, Glu, N-acetylaspartate (NAA), creatine- (Cr) and choline-containing metabolites (Cho), and myo-Inositol (mI) in three cortical brain regions using (1)H MRS and correlated them with measures of neurocognition, insomnia, PTSD symptoms, and drinking severity. In contrast to PTSD, PAUD exhibited normal GABA and Glu concentrations in the parieto-occipital and temporal cortices, respectively, but lower Glu and trends toward higher GABA levels in the anterior cingulate cortex (ACC). Temporal NAA and Cho as well as mI in the ACC were lower in PAUD than in both PTSD and CON. Within PAUD, more cortical GABA and Glu correlated with better neurocognition. Heavy drinking in PTSD is associated with partially neutralized neurotransmitter imbalance, but also with neuronal injury commonly observed in AUD.

Robust whole-brain segmentation: application to traumatic brain injury

We propose a framework for the robust and fully-automatic segmentation of magnetic resonance (MR) brain images called "Multi-Atlas Label Propagation with Expectation-Maximisation based refinement" (MALP-EM). The presented approach is based on a robust registration approach (MAPER), highly performant label fusion (joint label fusion) and intensity-based label refinement using EM. We further adapt this framework to be applicable for the segmentation of brain images with gross changes in anatomy. We propose to account for consistent registration errors by relaxing anatomical priors obtained by multi-atlas propagation and a weighting scheme to locally combine anatomical atlas priors and intensity-refined posterior probabilities. The method is evaluated on a benchmark dataset used in a recent MICCAI segmentation challenge. In this context we show that MALP-EM is competitive for the segmentation of MR brain scans of healthy adults when compared to state-of-the-art automatic labelling techniques. To demonstrate the versatility of the proposed approach, we employed MALP-EM to segment 125 MR brain images into 134 regions from subjects who had sustained traumatic brain injury (TBI). We employ a protocol to assess segmentation quality if no manual reference labels are available. Based on this protocol, three independent, blinded raters confirmed on 13 MR brain scans with pathology that MALP-EM is superior to established label fusion techniques. We visually confirm the robustness of our segmentation approach on the full cohort and investigate the potential of derived symmetry-based imaging biomarkers that correlate with and predict clinically relevant variables in TBI such as the Marshall Classification (MC) or Glasgow Outcome Score (GOS). Specifically, we show that we are able to stratify TBI patients with favourable outcomes from non-favourable outcomes with 64.7% accuracy using acute-phase MR images and 66.8% accuracy using follow-up MR images. Furthermore, we are able to differentiate subjects with the presence of a mass lesion or midline shift from those with diffuse brain injury with 76.0% accuracy. The thalamus, putamen, pallidum and hippocampus are particularly affected. Their involvement predicts TBI disease progression.

Partial volume effect modeling for segmentation and tissue classification of brain magnetic resonance images: A review

Quantitative analysis of magnetic resonance (MR) brain images are facilitated by the development of automated segmentation algorithms. A single image voxel may contain of several types of tissues due to the finite spatial resolution of the imaging device. This phenomenon, termed partial volume effect (PVE), complicates the segmentation process, and, due to the complexity of human brain anatomy, the PVE is an important factor for accurate brain structure quantification. Partial volume estimation refers to a generalized segmentation task where the amount of each tissue type within each voxel is solved. This review aims to provide a systematic, tutorial-like overview and categorization of methods for partial volume estimation in brain MRI. The review concentrates on the statistically based approaches for partial volume estimation and also explains differences to other, similar image segmentation approaches.

Partial volume estimation in brain MRI revisited

We propose a fast algorithm to estimate brain tissue concentrations from conventional T1-weighted images based on a Bayesian maximum a posteriori formulation that extends the "mixel" model developed in the 90's. A key observation is the necessity to incorporate additional prior constraints to the "mixel" model for the estimation of plausible concentration maps. Experiments on the ADNI standardized dataset show that global and local brain atrophy measures from the proposed algorithm yield enhanced diagnosis testing value than with several widely used soft tissue labeling methods.

New partial volume estimation methods for MRI MP2RAGE

Magnetic resonance imaging (MRI) is commonly used as a medical iagnosis tool, especially for brain applications. Some limitations affecting image quality include receive field (RF) inhomogeneity and partial volume (PV) effects which arise when a voxel contains two different tissues, introducing blurring. The novel Magnetization-Prepared 2 Rapid Acquisition Gradient Echoes (MP2RAGE) provides an image robust to RF inhomogeneity. However, PV effects are still an issue for automated brain quantification. PV estimation methods have been proposed based on computing the proportion of one tissue with respect to the other using linear interpolation of pure tissue intensity means. We demonstrated that this linear model introduces bias when used with MP2RAGE and we propose two novel solutions. The PV estimation methods were tested on 4 MP2RAGE data sets.

Structural brain differences in alcohol-dependent individuals with and without comorbid substance dependence

BACKGROUND

Over 50% of individuals with alcohol use disorders (AUD) also use other substances; brain structural abnormalities observed in alcohol dependent individuals may not be entirely related to alcohol consumption. This MRI study assessed differences in brain regional tissue volumes between short-term abstinent alcohol dependent individuals without (ALC) and with current substance use dependence (polysubstance users, PSU).

METHODS

Nineteen, one-month-abstinent PSU and 40 ALC as well as 27 light-drinkers (LD) were studied on a 1.5 T MR system. Whole brain T1-weighted images were segmented automatically into regional gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) volumes. MANOVA assessed group differences of intracranial volume-normalized tissue volumes of the frontal, parietal, occipital, and temporal lobes and regional subcortical GM volumes. The volumetric measures were correlated with neurocognitive measures to assess their functional relevance.

RESULTS

Despite similar lifetime drinking and smoking histories, PSU had significantly larger normalized WM volumes than ALC in all lobes. PSU also had larger frontal and parietal WM volumes than LD, but smaller temporal GM volumes and smaller lenticular and thalamic nuclei than LD. ALC had smaller frontal, parietal, and temporal GM, thalamic GM and cerebellar volumes than LD. ALC had more sulcal CSF volumes than both PSU and LD.

CONCLUSION

One-month-abstinent ALC and PSU exhibited different patterns of gross brain structural abnormalities. The larger lobar WM volumes in PSU in the absence of widespread GM volume loss contrast with widespread GM atrophy in ALC. These structural differences may demand different treatment approaches to mitigate specific functionally relevant brain abnormalities.

Genetic and behavioral determinants of hippocampal volume recovery during abstinence from alcohol

Alcohol-dependent individuals (ALC) have smaller hippocampi and poorer neurocognition than healthy controls. Results from studies on the association between alcohol consumption and hippocampal volume have been mixed, suggesting that comorbid or premorbid factors (i.e., those present prior to the initiation of alcohol dependence) determine hippocampal volume in ALC. We aimed to characterize the effects of select comorbid (i.e., cigarette smoking) and premorbid factors (brain-derived neurotrophic factor [BDNF] genotype [Val66Met rs6265]) on hippocampal volume in an ALC cohort followed longitudinally into extended abstinence. One hundred twenty-one adult ALC in treatment (76 smokers, 45 non-smokers) and 35 non-smoking light-drinking controls underwent quantitative magnetic resonance imaging, BDNF genotyping, and neurocognitive assessments. Representative subgroups were studied at 1 week, 1 month, and at an average of 7 months of abstinence. ALC had smaller hippocampi than healthy controls at all time points. Hippocampal volume at 1 month of abstinence correlated with lower visuospatial function. Smoking status did not influence hippocampal volume or hippocampal volume recovery during abstinence. However, only BDNF Val homozygotes tended to have hippocampal volume increases over 7 months of abstinence, and Val homozygotes had significantly larger hippocampi than Met carriers at 7 months of abstinence. These findings suggest that BDNF genotype, but not smoking status or measures of drinking severity, regulate functionally relevant hippocampal volume recovery in abstinent ALC. Future studies aimed at exploring genetic determinants of brain morphometry in ALC may need to evaluate individuals during extended abstinence after the acute environmental effects of chronic alcohol consumption have waned.

Three-dimensional brain magnetic resonance imaging segmentation via knowledge-driven decision theory

Brain tissue segmentation on magnetic resonance (MR) imaging is a difficult task because of significant intensity overlap between the tissue classes. We present a new knowledge-driven decision theory (KDT) approach that incorporates prior information of the relative extents of intensity overlap between tissue class pairs for volumetric MR tissue segmentation. The proposed approach better handles intensity overlap between tissues without explicitly employing methods for removal of MR image corruptions (such as bias field). Adaptive tissue class priors are employed that combine probabilistic atlas maps with spatial contextual information obtained from Markov random fields to guide tissue segmentation. The energy function is minimized using a variational level-set-based framework, which has shown great promise for MR image analysis. We evaluate the proposed method on two well-established real MR datasets with expert ground-truth segmentations and compare our approach against existing segmentation methods. KDT has low-computational complexity and shows better segmentation performance than other segmentation methods evaluated using these MR datasets.

BOOST: a supervised approach for multiple sclerosis lesion segmentation

BACKGROUND

Automatic multiple sclerosis lesion segmentation is a challenging task. An extensive analysis of the most recent techniques indicates an improvement of the results obtained when using prior knowledge and contextual information.

NEW METHOD

We present BOOST, a knowledge-based approach to automatically segment multiple sclerosis lesions through a voxel by voxel classification. We used the Gentleboost classifier and a set of features, including contextual features, registered atlas probability maps and an outlier map.

RESULTS

Results are computed on a set of 45 cases from three different hospitals (15 of each), obtaining a moderate agreement between the manual annotations and the automatically segmented results.

COMPARISON WITH EXISTING METHOD(S)

We quantitatively compared our results with three public state-of-the-art approaches obtaining competitive results and a better overlap with manual annotations. Our approach tends to better segment those cases with high lesion load, while cases with small lesion load are more difficult to accurately segment.

CONCLUSIONS

We believe BOOST has potential applicability in the clinical practice, although it should be improved in those cases with small lesion load.

Automatic whole brain MRI segmentation of the developing neonatal brain

Magnetic resonance (MR) imaging is increasingly being used to assess brain growth and development in infants. Such studies are often based on quantitative analysis of anatomical segmentations of brain MR images. However, the large changes in brain shape and appearance associated with development, the lower signal to noise ratio and partial volume effects in the neonatal brain present challenges for automatic segmentation of neonatal MR imaging data. In this study, we propose a framework for accurate intensity-based segmentation of the developing neonatal brain, from the early preterm period to term-equivalent age, into 50 brain regions. We present a novel segmentation algorithm that models the intensities across the whole brain by introducing a structural hierarchy and anatomical constraints. The proposed method is compared to standard atlas-based techniques and improves label overlaps with respect to manual reference segmentations. We demonstrate that the proposed technique achieves highly accurate results and is very robust across a wide range of gestational ages, from 24 weeks gestational age to term-equivalent age.

The isometric log-ratio transform for probabilistic multi-label anatomical shape representation

Sources of uncertainty in the boundaries of structures in medical images have motivated the use of probabilistic labels in segmentation applications. An important component in many medical image segmentation tasks is the use of a shape model, often generated by applying statistical techniques to training data. Standard statistical techniques (e.g., principal component analysis) often assume data lies in an unconstrained vector space, but probabilistic labels are constrained to the unit simplex. If these statistical techniques are used directly on probabilistic labels, relative uncertainty information can be sacrificed. A standard method for facilitating analysis of probabilistic labels is to map them to a vector space using the LogOdds transform. However, the LogOdds transform is asymmetric in one of the labels, which skews results in some applications. The isometric log-ratio (ILR) transform is a symmetrized version of the LogOdds transform, and is so named as it is an isometry between the Aitchison geometry, the inherent geometry of the simplex, and standard Euclidean geometry. We explore how to interpret the Aitchison geometry when applied to probabilistic labels in medical image segmentation applications. We demonstrate the differences when applying the LogOdds transform or the ILR transform to probabilistic labels prior to statistical analysis. Specifically, we show that statistical analysis of ILR transformed data better captures the variability of anatomical shapes in cases where multiple different foreground regions share boundaries (as opposed to foreground-background boundaries).

Automated MRI cerebellar size measurements using active appearance modeling

Although the human cerebellum has been increasingly identified as an important hub that shows potential for helping in the diagnosis of a large spectrum of disorders, such as alcoholism, autism, and fetal alcohol spectrum disorder, the high costs associated with manual segmentation, and low availability of reliable automated cerebellar segmentation tools, has resulted in a limited focus on cerebellar measurement in human neuroimaging studies. We present here the CATK (Cerebellar Analysis Toolkit), which is based on the Bayesian framework implemented in FMRIB's FIRST. This approach involves training Active Appearance Models (AAMs) using hand-delineated examples. CATK can currently delineate the cerebellar hemispheres and three vermal groups (lobules I-V, VI-VII, and VIII-X). Linear registration with the low-resolution MNI152 template is used to provide initial alignment, and Point Distribution Models (PDM) are parameterized using stellar sampling. The Bayesian approach models the relationship between shape and texture through computation of conditionals in the training set. Our method varies from the FIRST framework in that initial fitting is driven by 1D intensity profile matching, and the conditional likelihood function is subsequently used to refine fitting. The method was developed using T1-weighted images from 63 subjects that were imaged and manually labeled: 43 subjects were scanned once and were used for training models, and 20 subjects were imaged twice (with manual labeling applied to both runs) and used to assess reliability and validity. Intraclass correlation analysis shows that CATK is highly reliable (average test-retest ICCs of 0.96), and offers excellent agreement with the gold standard (average validity ICC of 0.87 against manual labels). Comparisons against an alternative atlas-based approach, SUIT (Spatially Unbiased Infratentorial Template), that registers images with a high-resolution template of the cerebellum, show that our AAM approach offers superior reliability and validity. Extensions of CATK to cerebellar hemisphere parcels are envisioned.

Comparison of MR-less PiB SUVR quantification methods

(11)C-Pittsburgh compound B (PiB) is a positron emission tomography (PET) tracer designed to bind to amyloid-β (Aβ) plaques, one of the hallmarks of Alzheimer's disease (AD). The potential of PiB as an early marker of AD led to the increasing use of PiB in clinical research studies and development of several F-18-labeled Aβ radiotracers. Automatic quantification of PiB images requires an accurate parcellation of the brain's gray matter (GM). Typically, this relies on a coregistered magnetic resonance imaging (MRI) to extract the cerebellar GM, compute the standardized uptake value ratio (SUVR), and provide parcellation and segmentation for quantification of regional and global SUVR. However, not all subjects can undergo MRI, in which case, an MR-less method is desirable. In this study, we assess 3 PET-only quantification methods: a mean atlas, an adaptive atlas, and a multi-atlas approaches on a database of 237 subjects having been imaged with both PiB PET and MRI. The PET-only methods were compared against MR-based SUVR quantification and evaluated in terms of correlation, average error, and performance in classifying subjects with low and high Aβ deposition. The mean atlas method suffered from a significant bias between the estimated neocortical SUVR and the PiB status, resulting in an overall error of 5.6% (R(2) = 0.98), compared with the adaptive and multi-atlas approaches that had errors of 3.06% and 2.74%, respectively (R(2) = 0.98), and no significant bias. In classifying PiB-negative from PiB-positive subjects, the mean atlas had 10 misclassified subjects compared with 0 for the adaptive and 1 for the multi-atlas approach. Overall, the adaptive and the multi-atlas approaches performed similarly well against the MR-based quantification and would be a suitable replacements for PiB quantification when no MRI is available.

A Bayesian bounded asymmetric mixture model with segmentation application

Segmentation of a medical image based on the modeling and estimation of the tissue intensity probability density functions via a Gaussian mixture model has recently received great attention. However, the Gaussian distribution is unbounded and symmetrical around its mean. This study presents a new bounded asymmetric mixture model for analyzing both univariate and multivariate data. The advantage of the proposed model is that it has the flexibility to fit different shapes of observed data such as non-Gaussian, nonsymmetric, and bounded support data. Another advantage is that each component of the proposed model has the ability to model the observed data with different bounded support regions, which is suitable for application on image segmentation. Our method is intuitively appealing, simple, and easy to implement. We also propose a new method to estimate the model parameters in order to minimize the higher bound on the data negative log-likelihood function. Numerical experiments are presented where the proposed model is tested in various images from simulated to real 3- D medical ones.

False discovery rate control in magnetic resonance imaging studies via Markov random fields

Magnetic resonance imaging (MRI) is widely used to study population effects of factors on brain morphometry. Inference from such studies often require the simultaneous testing of millions of statistical hypotheses. Such scale of inference is known to lead to large numbers of false positive results. Control of the false discovery rate (FDR) is commonly employed to mitigate against such outcomes. However, current methodologies in FDR control only account for the marginal significance of hypotheses, and are not able to explicitly account for spatial relationships, such as those between MRI voxels. In this article, we present novel methods that incorporate spatial dependencies into the process of controlling FDR through the use of Markov random fields. Our method is able to automatically estimate the relationships between spatially dependent hypotheses by means of maximum pseudo-likelihood estimation and the pseudo-likelihood information criterion. We show that our methods have desirable statistical properties with regards to FDR control and are able to outperform noncontexual methods in simulations of dependent hypothesis scenarios. Our method is applied to investigate the effects of aging on brain morphometry using data from the PATH study. Evidence of whole brain and component level effects that correspond to similar findings in the literature is found in our investigation.

Automated fetal brain segmentation from 2D MRI slices for motion correction

Motion correction is a key element for imaging the fetal brain in-utero using Magnetic Resonance Imaging (MRI). Maternal breathing can introduce motion, but a larger effect is frequently due to fetal movement within the womb. Consequently, imaging is frequently performed slice-by-slice using single shot techniques, which are then combined into volumetric images using slice-to-volume reconstruction methods (SVR). For successful SVR, a key preprocessing step is to isolate fetal brain tissues from maternal anatomy before correcting for the motion of the fetal head. This has hitherto been a manual or semi-automatic procedure. We propose an automatic method to localize and segment the brain of the fetus when the image data is acquired as stacks of 2D slices with anatomy misaligned due to fetal motion. We combine this segmentation process with a robust motion correction method, enabling the segmentation to be refined as the reconstruction proceeds. The fetal brain localization process uses Maximally Stable Extremal Regions (MSER), which are classified using a Bag-of-Words model with Scale-Invariant Feature Transform (SIFT) features. The segmentation process is a patch-based propagation of the MSER regions selected during detection, combined with a Conditional Random Field (CRF). The gestational age (GA) is used to incorporate prior knowledge about the size and volume of the fetal brain into the detection and segmentation process. The method was tested in a ten-fold cross-validation experiment on 66 datasets of healthy fetuses whose GA ranged from 22 to 39 weeks. In 85% of the tested cases, our proposed method produced a motion corrected volume of a relevant quality for clinical diagnosis, thus removing the need for manually delineating the contours of the brain before motion correction. Our method automatically generated as a side-product a segmentation of the reconstructed fetal brain with a mean Dice score of 93%, which can be used for further processing.

Automatic segmentation for detecting uterine fibroid regions treated with MR-guided high intensity focused ultrasound (MR-HIFU)

PURPOSE

Up to 25% of women suffer from uterine fibroids (UF) that cause infertility, pain, and discomfort. MR-guided high intensity focused ultrasound (MR-HIFU) is an emerging technique for noninvasive, computer-guided thermal ablation of UFs. The volume of induced necrosis is a predictor of the success of the treatment. However, accurate volume assessment by hand can be time consuming, and quick tools produce biased results. Therefore, fast and reliable tools are required in order to estimate the technical treatment outcome during the therapy event so as to predict symptom relief.

METHODS

A novel technique has been developed for the segmentation and volume assessment of the treated region. Conventional algorithms typically require user interaction ora priori knowledge of the target. The developed algorithm exploits the treatment plan, the coordinates of the intended ablation, for fully automatic segmentation with no user input.

RESULTS

A good similarity to an expert-segmented manual reference was achieved (Dice similarity coefficient = 0.880 ± 0.074). The average automatic segmentation time was 1.6 ± 0.7 min per patient against an order of tens of minutes when done manually.

CONCLUSIONS

The results suggest that the segmentation algorithm developed, requiring no user-input, provides a feasible and practical approach for the automatic evaluation of the boundary and volume of the HIFU-treated region.

Brain MR image segmentation with spatial constrained K-mean algorithm and dual-tree complex wavelet transform

In brain MR images, the noise and low-contrast significantly deteriorate the segmentation results. In this paper, we propose an automatic unsupervised segmentation method integrating dual-tree complex wavelet transform (DT-CWT) with K-mean algorithm for brain MR image. Firstly, a multi-dimensional feature vector is constructed based on the intensity, the low-frequency subband of DT-CWT and spatial position information. Then, a spatial constrained K-mean algorithm is presented as the segmentation system. The proposed method is validated by extensive experiments using both simulated and real T1-weighted MR images, and compared with the state-of-the-art algorithms.

MBIS: multivariate Bayesian image segmentation tool

We present MBIS (Multivariate Bayesian Image Segmentation tool), a clustering tool based on the mixture of multivariate normal distributions model. MBIS supports multichannel bias field correction based on a B-spline model. A second methodological novelty is the inclusion of graph-cuts optimization for the stationary anisotropic hidden Markov random field model. Along with MBIS, we release an evaluation framework that contains three different experiments on multi-site data. We first validate the accuracy of segmentation and the estimated bias field for each channel. MBIS outperforms a widely used segmentation tool in a cross-comparison evaluation. The second experiment demonstrates the robustness of results on atlas-free segmentation of two image sets from scan-rescan protocols on 21 healthy subjects. Multivariate segmentation is more replicable than the monospectral counterpart on T1-weighted images. Finally, we provide a third experiment to illustrate how MBIS can be used in a large-scale study of tissue volume change with increasing age in 584 healthy subjects. This last result is meaningful as multivariate segmentation performs robustly without the need for prior knowledge.

Bounded asymmetrical Student's-t mixture model

The finite mixture model based on the Student's-t distribution, which is heavily tailed and more robust than the Gaussian mixture model (GMM), is a flexible and powerful tool to address many computer vision and pattern recognition problems. However, the Student's-t distribution is unbounded and symmetrical around its mean. In many applications, the observed data are digitalized and have bounded support. The distribution of the observed data usually has an asymmetric form. A new finite bounded asymmetrical Student's-t mixture model (BASMM), which includes the GMM and the Student's-t mixture model (SMM) as special cases, is presented in this paper. We propose an extension of the Student's-t distribution in this paper. This new distribution is sufficiently flexible to fit different shapes of observed data, such as non-Gaussian, nonsymmetric, and bounded support data. Another advantage of the proposed model is that each of its components can model the observed data with different bounded support regions. In order to estimate the model parameters, previous models represent the Student's-t distributions as an infinite mixture of scaled Gaussians. We propose an alternate approach in order to minimize the higher bound on the data negative log-likelihood function, and directly deal with the Student's-t distribution. As an application, our method has been applied to image segmentation with promising results.

LOGISMOS-B: layered optimal graph image segmentation of multiple objects and surfaces for the brain

Automated reconstruction of the cortical surface is one of the most challenging problems in the analysis of human brain magnetic resonance imaging (MRI). A desirable segmentation must be both spatially and topologically accurate, as well as robust and computationally efficient. We propose a novel algorithm, LOGISMOS-B, based on probabilistic tissue classification, generalized gradient vector flows and the LOGISMOS graph segmentation framework. Quantitative results on MRI datasets from both healthy subjects and multiple sclerosis patients using a total of 16,800 manually placed landmarks illustrate the excellent performance of our algorithm with respect to spatial accuracy. Remarkably, the average signed error was only 0.084 mm for the white matter and 0.008 mm for the gray matter, even in the presence of multiple sclerosis lesions. Statistical comparison shows that LOGISMOS-B produces a significantly more accurate cortical reconstruction than FreeSurfer, the current state-of-the-art approach (p << 0.001). Furthermore, LOGISMOS-B enjoys a run time that is less than a third of that of FreeSurfer, which is both substantial, considering the latter takes 10 h/subject on average, and a statistically significant speedup.