Automatic segmentation of the caudate nucleus from human brain MR images

Whole brain segmentation with full volume neural network

Whole brain segmentation is an important neuroimaging task that segments the whole brain volume into anatomically labeled regions-of-interest. Convolutional neural networks have demonstrated good performance in this task. Existing solutions, usually segment the brain image by classifying the voxels, or labeling the slices or the sub-volumes separately. Their representation learning is based on parts of the whole volume whereas their labeling result is produced by aggregation of partial segmentation. Learning and inference with incomplete information could lead to sub-optimal final segmentation result. To address these issues, we propose to adopt a full volume framework, which feeds the full volume brain image into the segmentation network and directly outputs the segmentation result for the whole brain volume. The framework makes use of complete information in each volume and can be implemented easily. An effective instance in this framework is given subsequently. We adopt the 3D high-resolution network (HRNet) for learning spatially fine-grained representations and the mixed precision training scheme for memory-efficient training. Extensive experiment results on a publicly available 3D MRI brain dataset show that our proposed model advances the state-of-the-art methods in terms of segmentation performance.

TASH: Toolbox for the Automated Segmentation of Heschl's gyrus

Auditory cortex volume and shape differences have been observed in the context of phonetic learning, musicianship and dyslexia. Heschl’s gyrus, which includes primary auditory cortex, displays large anatomical variability across individuals and hemispheres. Given this variability, manual labelling is the gold standard for segmenting HG, but is time consuming and error prone. Our novel toolbox, called ‘Toolbox for the Automated Segmentation of HG’ or TASH, automatically segments HG in brain structural MRI data, and extracts measures including its volume, surface area and cortical thickness. TASH builds upon FreeSurfer, which provides an initial segmentation of auditory regions, and implements further steps to perform finer auditory cortex delineation. We validate TASH by showing significant relationships between HG volumes obtained using manual labelling and using TASH, in three independent datasets acquired on different scanners and field strengths, and by showing good qualitative segmentation. We also present two applications of TASH, demonstrating replication and extension of previously published findings of relationships between HG volumes and (a) phonetic learning, and (b) musicianship. In sum, TASH effectively segments HG in a fully automated and reproducible manner, opening up a wide range of applications in the domains of expertise, disease, genetics and brain plasticity.

Finding of increased caudate nucleus in patients with Alzheimer's disease

OBJECTIVES

A recently published study using an automated MRI volumetry method (NeuroQuant®) unexpectedly demonstrated larger caudate nucleus volume in patients with Alzheimer's disease dementia (AD) compared to patients with subjective and mild cognitive impairment (SCI and MCI). The aim of this study was to explore this finding.

MATERIALS & METHODS

The caudate nucleus and the hippocampus volumes were measured (both expressed as ratios of intracranial volume) in a total of 257 patients with SCI and MCI according to the Winblad criteria and AD according to ICD-10 criteria. Demographic data, cognitive measures, and APOE-ɛ4 status were collected.

RESULTS

Compared with non-dementia patients (SCI and MCI), AD patients were older, more of them were female, and they had a larger caudate nucleus volume and smaller hippocampus volume (P<.001). In multiple linear regression analysis, age and female sex were associated with larger caudate nucleus volume, but neither diagnosis nor memory function was. Age, gender, and memory function were associated with hippocampus volume, and age and memory function were associated with caudate nucleus/hippocampus ratio.

CONCLUSIONS

A larger caudate nucleus volume in AD patients was partly explained by older age and being female. These results are further discussed in the context of (1) the caudate nucleus possibly serving as a mechanism for temporary compensation; (2) methodological properties of automated volumetry of this brain region; and (3) neuropathological alterations. Further studies are needed to fully understand the role of the caudate nucleus in AD.

Diffuse optical tomography using semiautomated coregistered ultrasound measurements

Diffuse optical tomography (DOT) has demonstrated huge potential in breast cancer diagnosis and treatment monitoring. DOT image reconstruction guided by ultrasound (US) improves the diffused light localization and lesion reconstruction accuracy. However, DOT reconstruction depends on tumor geometry provided by coregistered US. Experienced operators can manually measure these lesion parameters; however, training and measurement time are needed. The wide clinical use of this technique depends on its robustness and faster imaging reconstruction capability. This article introduces a semiautomated procedure that automatically extracts lesion information from US images and incorporates it into the optical reconstruction. An adaptive threshold-based image segmentation is used to obtain tumor boundaries. For some US images, posterior shadow can extend to the chest wall and make the detection of deeper lesion boundary difficult. This problem can be solved using a Hough transform. The proposed procedure was validated from data of 20 patients. Optical reconstruction results using the proposed procedure were compared with those reconstructed using extracted tumor information from an experienced user. Mean optical absorption obtained from manual measurement was 0.21±0.06  cm-1 for malignant and 0.12±0.06  cm-1 for benign cases, whereas for the proposed method it was 0.24±0.08  cm-1 and 0.12±0.05  cm-1, respectively.

Alzheimer's disease detection via automatic 3D caudate nucleus segmentation using coupled dictionary learning with level set formulation

BACKGROUND AND OBJECTIVE

This paper presents a novel method for Alzheimer's disease classification via an automatic 3D caudate nucleus segmentation.

METHODS

The proposed method consists of segmentation and classification steps. In the segmentation step, we propose a novel level set cost function. The proposed cost function is constrained by a sparse representation of local image features using a dictionary learning method. We present coupled dictionaries: a feature dictionary of a grayscale brain image and a label dictionary of a caudate nucleus label image. Using online dictionary learning, the coupled dictionaries are learned from the training data. The learned coupled dictionaries are embedded into a level set function. In the classification step, a region-based feature dictionary is built. The region-based feature dictionary is learned from shape features of the caudate nucleus in the training data. The classification is based on the measure of the similarity between the sparse representation of region-based shape features of the segmented caudate in the test image and the region-based feature dictionary.

RESULTS

The experimental results demonstrate the superiority of our method over the state-of-the-art methods by achieving a high segmentation (91.5%) and classification (92.5%) accuracy.

CONCLUSIONS

In this paper, we find that the study of the caudate nucleus atrophy gives an advantage over the study of whole brain structure atrophy to detect Alzheimer's disease.

Fully automated segmentation of the pons and midbrain using human T1 MR brain images

Purpose

This paper describes a novel method to automatically segment the human brainstem into midbrain and pons, called LABS: Landmark-based Automated Brainstem Segmentation. LABS processes high-resolution structural magnetic resonance images (MRIs) according to a revised landmark-based approach integrated with a thresholding method, without manual interaction.

Methods

This method was first tested on morphological T1-weighted MRIs of 30 healthy subjects. Its reliability was further confirmed by including neurological patients (with Alzheimer's Disease) from the ADNI repository, in whom the presence of volumetric loss within the brainstem had been previously described. Segmentation accuracies were evaluated against expert-drawn manual delineation. To evaluate the quality of LABS segmentation we used volumetric, spatial overlap and distance-based metrics.

Results

The comparison between the quantitative measurements provided by LABS against manual segmentations revealed excellent results in healthy controls when considering either the midbrain (DICE measures higher that 0.9; Volume ratio around 1 and Hausdorff distance around 3) or the pons (DICE measures around 0.93; Volume ratio ranging 1.024–1.05 and Hausdorff distance around 2). Similar performances were detected for AD patients considering segmentation of the pons (DICE measures higher that 0.93; Volume ratio ranging from 0.97–0.98 and Hausdorff distance ranging 1.07–1.33), while LABS performed lower for the midbrain (DICE measures ranging 0.86–0.88; Volume ratio around 0.95 and Hausdorff distance ranging 1.71–2.15).

Conclusions

Our study represents the first attempt to validate a new fully automated method for in vivo segmentation of two anatomically complex brainstem subregions. We retain that our method might represent a useful tool for future applications in clinical practice.

Segmentation of the striatum from MR brain images to calculate the 99mTc-TRODAT-1 binding ratio in SPECT images

Quantification of regional ^99mTc-TRODAT-1 binding ratio in the striatum regions in SPECT images is essential for differential diagnosis between Alzheimer's and Parkinson's diseases. Defining the region of the striatum in the SPECT image is the first step toward success in the quantification of the TRODAT-1 binding ratio. However, because SPECT images reveal insufficient information regarding the anatomical structure of the brain, correct delineation of the striatum directly from the SPECT image is almost impossible. We present a method integrating the active contour model and the hybrid registration technique to extract regions from MR T1-weighted images and map them into the corresponding SPECT images. Results from three normal subjects suggest that the segmentation accuracy using the proposed method was compatible with the expert decision but has a higher efficiency and reproducibility than manual delineation. The binding ratio derived by this method correlated well (R² = 0.76) with those values calculated by commercial software, suggesting the feasibility of the proposed method.

Anatomizing the “King of Neurosciences”

The human nucleus accumbens (NA), a major part of the ventral striatum, is the area of continuity between the putamen and head of the caudate nucleus. It consists of two parts, a shell laterally and a core medially. The first is mainly connected to the limbic system and the second to the extrapyramidal motor system. The NA, a major pleasure center of the human brain, acts as a limbic-motor interface and is involved in several cognitive, emotional and psychomotor functions. It has a modulating function in the amygdala-basal ganglia-prefrontal cortex circuit. It is considered as the neural interface between motivation and action. Further, it is a principal modulator of the reward circuits and supplies motor expression to emotional responses. Such a clinical significance could easily explain the intense work taking place in the respective field of basic research. Its exceptional clinical importance justifies the title of the “King of Neurosciences” for this nucleus. Purpose of this editorial is to review the “informational paths” left behind by the few researchers who tried to explore the architecture (gross anatomy) of this ‘kingdom’. The first anatomical study focused on this nucleus came from Neto et al. The most extensive study of the NA gross, imaging, stereotactic and neurosurgical anatomy so far, came from the research efforts of Mavridis et al.

Preliminary evidence of abnormal white matter related to the fusiform gyrus in Williams syndrome: a diffusion tensor imaging tractography study

Williams syndrome (WS) is a genetic condition caused by a hemizygous microdeletion on chromosome 7q11.23. WS is characterized by a distinctive social phenotype composed of increased drive toward social engagement and attention toward faces. In addition, individuals with WS exhibit abnormal structure and function of brain regions important for the processing of faces such as the fusiform gyrus. This study was designed to investigate if white matter tracts related to the fusiform gyrus in WS exhibit abnormal structural integrity as compared to typically developing (TD; age matched) and developmentally delayed (DD; intelligence quotient matched) controls. Using diffusion tensor imaging data collected from 40 (20 WS, 10 TD and 10 DD) participants, white matter fibers were reconstructed that project through the fusiform gyrus and two control regions (caudate and the genu of the corpus callosum). Macro-structural integrity was assessed by calculating the total volume of reconstructed fibers and micro-structural integrity was assessed by calculating fractional anisotropy (FA) and fiber density index (FDi) of reconstructed fibers. WS participants, as compared to controls, exhibited an increase in the volume of reconstructed fibers and an increase in FA and FDi for fibers projecting through the fusiform gyrus. No between-group differences were observed in the fibers that project through the control regions. Although preliminary, these results provide further evidence that the brain anatomy important for processing faces is abnormal in WS.

A fully-automatic caudate nucleus segmentation of brain MRI: application in volumetric analysis of pediatric attention-deficit/hyperactivity disorder

Background

Accurate automatic segmentation of the caudate nucleus in magnetic resonance images (MRI) of the brain is of great interest in the analysis of developmental disorders. Segmentation methods based on a single atlas or on multiple atlases have been shown to suitably localize caudate structure. However, the atlas prior information may not represent the structure of interest correctly. It may therefore be useful to introduce a more flexible technique for accurate segmentations.

Method

We present Cau-dateCut: a new fully-automatic method of segmenting the caudate nucleus in MRI. CaudateCut combines an atlas-based segmentation strategy with the Graph Cut energy-minimization framework. We adapt the Graph Cut model to make it suitable for segmenting small, low-contrast structures, such as the caudate nucleus, by defining new energy function data and boundary potentials. In particular, we exploit information concerning the intensity and geometry, and we add supervised energies based on contextual brain structures. Furthermore, we reinforce boundary detection using a new multi-scale edgeness measure.

Results

We apply the novel CaudateCut method to the segmentation of the caudate nucleus to a new set of 39 pediatric attention-deficit/hyperactivity disorder (ADHD) patients and 40 control children, as well as to a public database of 18 subjects. We evaluate the quality of the segmentation using several volumetric and voxel by voxel measures. Our results show improved performance in terms of segmentation compared to state-of-the-art approaches, obtaining a mean overlap of 80.75%. Moreover, we present a quantitative volumetric analysis of caudate abnormalities in pediatric ADHD, the results of which show strong correlation with expert manual analysis.

Conclusion

CaudateCut generates segmentation results that are comparable to gold-standard segmentations and which are reliable in the analysis of differentiating neuroanatomical abnormalities between healthy controls and pediatric ADHD.

Anatomy of the human nucleus accumbens: a combined morphometric study

PURPOSE

The human nucleus accumbens (NA), which belongs to the basal ganglia of the brain, is the main part of the ventral striatum. The purpose of our clinically oriented anatomical-radiologic study was to provide anatomical and imaging data of the human NA, primarily useful to neurosurgeons.

METHODS

For our imaging study, we used cerebral magnetic resonance images (MRIs) from 26 neurosurgical patients (52 NAs). The material for our anatomic study consisted of 32 cerebral hemispheres (32 NAs) from 18 normal human brains which we have in our department (Department of Anatomy) from cadaver donors. We measured and analyzed the dimensions of the NA at specific clinically important transverse, coronal and sagittal levels.

RESULTS

The human NA suffers from age-related but no side- or sex-related morphometric changes. In surgically important stereotactic levels this nucleus is easily identifiable on MRIs.

CONCLUSIONS

We present an anatomic guide of the NA from carefully measured data of our extensive and combined study and we hope that our work will be really helpful to neuroscientists interested in the NA.

A time-saving and facilitating approach for segmentation of anatomically defined cortical regions: MRI volumetry

In this study, we present an accurate, reliable, robust, and time-efficient technique for a semi-automatic segmentation of neuroanatomically defined cortical structures in Magnetic Resonance Imaging (MRI) scans. It involves manual drawing of the border of a region of interest (ROI), supported by three-dimensional (3D) visualization techniques (rendering), and a subsequent automatic tracing of the gray matter voxels inside the ROI by means of an automatic tissue classifier. The approach has been evaluated on a set of MRI scans of 75 participants selected from the Maastricht Aging Study (MAAS) and applied to cortical brain structures for both the left and right hemispheres, viz., the inferior prefrontal cortex (PFC); the orbital PFC; the dorsolateral PFC; the anterior cingulate cortex; and the posterior cingulate cortex. The use of a 3D surface-rendered brain can be rotated in any direction was invaluable in identifying anatomical landmarks on the basis of gyral and sulcal topography. This resulted in a high accuracy (anatomical correctness) and reliability: the intra-rater intra-class correlation coefficient (ICC) was between 0.96 and 0.99. Furthermore, the obtained time savings were substantial, i.e., up to a factor of 7.5 compared with fully manual segmentations.

Putaminal volume and diffusion in early familial Creutzfeldt-Jakob disease

BACKGROUND

The putamen is centrally implicated in the pathophysiology of Creutzfeldt-Jakob Disease (CJD). To our knowledge, its volume has never been measured in this disease. We investigated whether gross putaminal atrophy can be detected by MRI in early stages, when the diffusion is already reduced.

METHODS

Twelve familial CJD patients with the E200K mutation and 22 healthy controls underwent structural and diffusion MRI scans. The putamen was identified in anatomical scans by two methods: manual tracing by a blinded investigator, and automatic parcellation by a computerized segmentation procedure (FSL FIRST). For each method, volume and mean Apparent Diffusion Coefficient (ADC) were calculated.

RESULTS

ADC was significantly lower in CJD patients (697+/-64 microm(2)/s vs. 750+/-31 microm(2)/s, p<0.005), as expected, but the volume was not reduced. The computerized FIRST delineation yielded comparable ADC values to the manual method, but computerized volumes were smaller than manual tracing values.

CONCLUSIONS

We conclude that significant diffusion reduction in the putamen can be detected by delineating the structure manually or with a computerized algorithm. Our findings confirm and extend previous voxel-based and observational studies. Putaminal volume was not reduced in our early-stage patients, thus confirming that diffusion abnormalities precede detectible atrophy in this structure.

Three-dimensional coupled-object segmentation using symmetry and tissue type information

This paper presents an automatic method for segmentation of brain structures using their symmetry and tissue type information. The proposed method generates segmented structures that have homogenous tissues. It benefits from general symmetry of the brain structures in the two hemispheres. It also benefits from the tissue regions generated by fuzzy c-means clustering. All in all, the proposed method can be described as a dynamic knowledge-based method that eliminates the need for statistical shape models of the structures while generating accurate segmentation results. The proposed approach is implemented in MATLAB and tested on the Internet Brain Segmentation Repository (IBSR) datasets. To this end, it is applied to the segmentation of caudate and ventricles three-dimensionally in magnetic resonance images (MRI) of the brain. Impacts of each of the steps of the proposed approach are demonstrated through experiments. It is shown that the proposed method generates accurate segmentation results that are insensitive to initialization and parameter selection. The proposed method is compared to four previous methods illustrating advantages and limitations of each method.

Early white-matter abnormalities of the ventral frontostriatal pathway in fragile X syndrome

AIM

Fragile X syndrome is associated with cognitive deficits in inhibitory control and with abnormal neuronal morphology and development.

METHOD

In this study, we used a diffusion tensor imaging (DTI) tractography approach to reconstruct white-matter fibers in the ventral frontostriatal pathway in young males with fragile X syndrome (n=17; mean age 2y 9mo, SD 7mo, range 1y 7mo-3y 10mo), and two age-matched comparison groups: (1) typically developing (n=13; mean age 2y 3mo, SD 7mo, range 1y 7mo-3y 6mo) and (2) developmentally delayed (n=8; mean age 3y, SD 4mo, range 2y 9mo-3y 8mo).

RESULTS

We observed that young males with fragile X syndrome exhibited increased density of DTI reconstructed fibers than those in the typically developing (p=0.001) and developmentally delayed (p=0.001) groups. Aberrant white-matter structure was localized in the left ventral frontostriatal pathway. Greater relative fiber density was found to be associated with lower IQ (Mullen composite scores) in the typically developing group (p=0.008).

INTERPRETATION

These data suggest that diminished or absent fragile X mental retardation 1 protein expression can selectively alter white-matter anatomy during early brain development and, in particular, neural pathways. The results also point to an early neurobiological marker for an important component of cognitive dysfunction associated with fragile X syndrome.

An evaluation of four automatic methods of segmenting the subcortical structures in the brain

The automation of segmentation of subcortical structures in the brain is an active research area. We have comprehensively evaluated four novel methods of fully automated segmentation of subcortical structures using volumetric, spatial overlap and distance-based measures. Two methods are atlas-based - classifier fusion and labelling (CFL) and expectation-maximisation segmentation using a brain atlas (EMS), and two incorporate statistical models of shape and appearance - profile active appearance models (PAM) and Bayesian appearance models (BAM). Each method was applied to the segmentation of 18 subcortical structures in 270 subjects from a diverse pool varying in age, disease, sex and image acquisition parameters. Our results showed that all four methods perform on par with recently published methods. CFL performed better than the others according to all three classes of metrics. In summary over all structures, the ranking by the Dice coefficient was CFL, BAM, joint EMS and PAM. The Hausdorff distance ranked the methods as CFL, joint PAM and BAM, EMS, whilst percentage absolute volumetric difference ranked them as joint CFL and PAM, joint BAM and EMS. Furthermore, as we had four methods of performing segmentation, we investigated whether the results obtained by each method were more similar to each other than to the manual segmentations using Williams' Index. Reassuringly, the Williams' Index was close to 1 for most subjects (mean=1.02, sd=0.05), indicating better agreement of each method with the gold standard than with the other methods. However, 2% of cases (mainly amygdala and nucleus accumbens) had values outside 3 standard deviations of the mean.

Non-rigid registration based segmentation of brain subcortical structures using a priori knowledge

Segmentation of the brain internal structures is an important and a challenging task due to their complex shapes, partial volume effects, low contrasts and anatomical variability between subjects. In this paper we propose a new non-rigid registration method that automatically segments the deep brain internal structures from brain MRI images. An atlas of the structures is used as a priori knowledge, which is modeled as a shape representation. By integrating the shape knowledge into a classical intensity based non-rigid registration algorithm, the proposed segmentation method allows to ameliorate the results in the case of low contrast on the boundaries of the structures. The shape model is based on distance representation obtained from the atlas. The segmentation of brain subcortical structures is performed on real MRI images and the obtained results are very encouraging.

Automatic segmentation of magnetic resonance images using a decision tree with spatial information

Here we proposed an automatic segmentation method based on a decision tree to classify the brain tissues in magnetic resonance (MR) images. Two types of data - phantom MR images obtained from IBSR (http://www.cma.mgh.harvard.edu/ibsr) and simulated brain MR images obtained from BrainWeb (http://www.bic.mni.mcgill.ca/brainweb) - were segmented using an automatic decision tree algorithm to obtain images with improved visual rendition. Spatial information on the general gray level (G), spatial gray level (S), and two-dimensional wavelet transform (W) was combined in-plane in two coordinate systems (Euclidean coordinates (x, y) or polar coordinates (r, theta)). The decision tree was constructed based on a binary tree with nodes created by splitting the distribution of input features of the tree. The spatial information obtained from MR images with different noise levels and inhomogeneities were segmented to compare whether the use of a decision tree improved the identification of human anatomical structures in a neuroimage. The average accuracy rates of segmentation for phantom images with a noise variation of 15 gray levels were 0.9999 and 0.9973 with spatial information (G, x, y, r, theta) and (S, x, y, r, theta), respectively, and 0.9999 and 0.9819 with spatial information (G, x, y, S, r, theta) and (W, x, y, G, r, theta). The average accuracy rates of segmentation for simulated MR images with a noise level of 5% were 0.9532 and 0.9439 with spatial information (G, x, y, r, theta) and (S, x, y, r, theta), respectively, and 0.9446 and 0.9287 with spatial information (G, x, y, S, r, theta) and (W, x, y, G, r, theta). The accuracy rates of segmentation were highest for both simulated phantom and brain MR images, having the lowest noise levels, from a reduction of overlapping gray levels in the images. The accuracies of segmentation were higher when the spatial information included the general gray level than when it included the spatial gray level, which in turn were higher than when it included the wavelet transform. Furthermore, the performance of segmentation was also evaluated with a boundary detection methodology that is based on the Hausdorff distance to compare with the mean computer to observer difference (COD) and mean interobserver difference (IOD) for gray matter (GM), white matter (WM), and all areas (ALL) from images segmented using the decision tree. The values of mean COD are similar and around 12mm for GM segmented using the decision tree. Our segmentation method based on a decision tree algorithm presented an easy way to perform automatic segmentation for both phantom and tissue regions in brain MR images.

Morphometric spatial patterns differentiating boys with fragile X syndrome, typically developing boys, and developmentally delayed boys aged 1 to 3 years

CONTEXT

Brain maturation starts well before birth and occurs as a unified process with developmental interaction among different brain regions. Gene and environment play large roles in such a process. Studies of individuals with genetic disorders such as fragile X syndrome (FXS), which is a disorder caused by a single gene mutation resulting in abnormal dendritic and synaptic pruning, together with healthy individuals may provide valuable information.

OBJECTIVE

To examine morphometric spatial patterns that differentiate between FXS and controls in early childhood.

DESIGN

A cross-sectional in vivo neuroimaging study.

SETTING

Academic medical centers.

PARTICIPANTS

A total of 101 children aged 1 to 3 years, comprising 51 boys with FXS, 32 typically developing boys, and 18 boys with idiopathic developmental delay.

MAIN OUTCOME MEASURES

Regional gray matter volume as measured by voxel-based morphometry and manual tracing, supplemented by permutation analyses; regression analyses between gray and white matter volumes, IQ, and fragile X mental retardation protein level; and linear support vector machine analyses to classify group membership.

RESULTS

In addition to aberrant brain structures reported previously in older individuals with FXS, we found reduced gray matter volumes in regions such as the hypothalamus, insula, and medial and lateral prefrontal cortices. These findings are consistent with the cognitive and behavioral phenotypes of FXS. Further, multivariate pattern classification analyses discriminated FXS from typical development and developmental delay with more than 90% prediction accuracy. The spatial patterns that classified FXS from typical development and developmental delay included those that may have been difficult to identify previously using other methods. These included a medial to lateral gradient of increased and decreased regional brain volumes in the posterior vermis, amygdala, and hippocampus.

CONCLUSIONS

These findings are critical in understanding interplay among genes, environment, brain, and behavior. They signify the importance of examining detailed spatial patterns of healthy and perturbed brain development.

FreeSurfer-initiated fully-automated subcortical brain segmentation in MRI using Large Deformation Diffeomorphic Metric Mapping

Fully-automated brain segmentation methods have not been widely adopted for clinical use because of issues related to reliability, accuracy, and limitations of delineation protocol. By combining the probabilistic-based FreeSurfer (FS) method with the Large Deformation Diffeomorphic Metric Mapping (LDDMM)-based label-propagation method, we are able to increase reliability and accuracy, and allow for flexibility in template choice. Our method uses the automated FreeSurfer subcortical labeling to provide a coarse-to-fine introduction of information in the LDDMM template-based segmentation resulting in a fully-automated subcortical brain segmentation method (FS+LDDMM). One major advantage of the FS+LDDMM-based approach is that the automatically generated segmentations generated are inherently smooth, thus subsequent steps in shape analysis can directly follow without manual post-processing or loss of detail. We have evaluated our new FS+LDDMM method on several databases containing a total of 50 subjects with different pathologies, scan sequences and manual delineation protocols for labeling the basal ganglia, thalamus, and hippocampus. In healthy controls we report Dice overlap measures of 0.81, 0.83, 0.74, 0.86 and 0.75 for the right caudate nucleus, putamen, pallidum, thalamus and hippocampus respectively. We also find statistically significant improvement of accuracy in FS+LDDMM over FreeSurfer for the caudate nucleus and putamen of Huntington's disease and Tourette's syndrome subjects, and the right hippocampus of Schizophrenia subjects.