University of Southern California Dissertations and Theses

Shape, pose, and connectivity in subcortical networks across the human lifespan

(USC Thesis Other)

Shape, pose, and connectivity in subcortical networks across the human lifespan

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Shape, Pose, and Connectivity In Subcortical Networks Across the Human Lifespan YiLao FACULTY OF THE USC GRADUATE SCHOOL. UNIVERSITY OF SOUTHERN CALIFORNIA This dissertation is submitted for the degree of DoctorofPhilosophy(BIOMEDICALENGINEERING) May 2017 Committee Members Natasha Leporé Assistant Professor, Department of Radiology, Children’s Hospital Los Angeles, University of Southern California. Dave Saint-Amour Professeur, Département de psychologie, Faculté des sciences humaines, Université du Québec à Montréal. David Z. D’Argenio Professor, Department of Biomedical Engineering, Viterbi School of Engineering, University of Southern California. Krishna S. Nayak Professor, Department of Electrical Engineering, Viterbi School of Engineering, University of Southern California. Yaling Yang Assistant Professor, Department of Pediatrics, Children’s Hospital Los Angeles, University of Southern California. Yonggang Shi Assistant Professor, Department of Neurology, Keck School of Medicine, University of Southern California. TheresearchpresentedherewasperformedattheCIBORGLab (DepartmentofRadiology,Children’sHospitalLosAngeles,USC)in closecollaborationwithDepartmentofpsychology(Universitédu QuébecàMontréal),ComputerScienceandEngineeringLab(Schoolof Computing,Informatics,andDecisionSystemsEngineering,ASU)and Children’sHospitalofPittsburgh(DepartmentofRadiology). Abstract Brain development is an on-going process that spans the whole life, and exhibits position, shape, and connectivity changes due to genetics, exter- nal environment, experiences and aging. By rapidly unveiling the anatomy and function in the living brain, magnetic resonance imaging (MRI) tech- nology has fueled a scientific revolution in the neuroscience field. In par- allel with the advent of MRI, population-based post-processing algorithms facilitate the in-vivo detection and monitoring of brain alterations. Here, we describe a combined framework for the analysis of structural MRI and diffusion tensor imaging (DTI) that aims to depict subtle brain alterations in healthy and abnormal brain evolution throughout the lifespan. Firstly, a T1-MRI based analysis characterizing the brain subcortical gray matter morphometry and inter-structural correlation is shown, and applica- tions on premature neonates as well as children with chronic manganese (Mn) exposure are presented. Secondly, a T1-MRI based algorithm as- sessing the brain positional changes is described, and applications on pre- mature neonates as well as young adults with mild traumatic brain injury (mTBI) are presented. Finally, a novel T1 and DTI fusion method yield- ing an increasing detection power of white matter alterations is introduced, and applications on mTBI subjects as well as aging subjects with increased risk of cardiovascular diseases are presented. The presented work allows shape, positional, and connectivity exhibited in the developing brain to be assessed within a more comprehensive framework. Contents Contents iv List of Figures viii List of Tables xvi 1 Introduction 1 1.1 The Human Brain Across the Lifespan . . . . . . . . . . . . . . . . . . . 3 1.1.1 The Developing Brain . . . . . . . . . . . . . . . . . . . . . . . 3 1.1.2 The Aging Brain . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.1.3 External Disturbances and Diseases . . . . . . . . . . . . . . . . 7 1.2 Introduction To the Thesis . . . . . . . . . . . . . . . . . . . . . . . . . 8 2 Background and Significance 10 2.1 Shape Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.1.1 Significance of shape analysis . . . . . . . . . . . . . . . . . . . 11 2.1.2 Substructural Surface Registration . . . . . . . . . . . . . . . . . 13 2.1.3 Surface Multivariate Tensor-based Morphometry (mTBM) . . . . 15 2.1.4 Multivariate Statistical Analysis . . . . . . . . . . . . . . . . . . 16 2.2 The Relative Position Analysis . . . . . . . . . . . . . . . . . . . . . . . 17 2.2.1 Significance of the relative pose analysis . . . . . . . . . . . . . 17 2.2.2 Procrustes Alignment . . . . . . . . . . . . . . . . . . . . . . . . 20 2.3 The T1 and DTI Fusion Analysis . . . . . . . . . . . . . . . . . . . . . . 22 Contents v 2.3.1 Diffusion Tensor Imaging . . . . . . . . . . . . . . . . . . . . . 22 2.3.2 The significance of joint T1 and DTI analysis . . . . . . . . . . . 25 2.3.3 The joint T1 and DTI analysis . . . . . . . . . . . . . . . . . . . 27 3 Shape and Pose Analysis of the Thalamic and the Lentiform Nucleus in Preterm Neonates 30 3.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 3.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 3.3 Subjects and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 3.3.1 Subjects and preprocessing . . . . . . . . . . . . . . . . . . . . . 36 3.3.2 Surface based morphometry analysis . . . . . . . . . . . . . . . . 39 3.3.3 Relative pose analysis . . . . . . . . . . . . . . . . . . . . . . . 42 3.3.4 Correlation analysis . . . . . . . . . . . . . . . . . . . . . . . . 44 3.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 3.4.1 Surface based morphometry results . . . . . . . . . . . . . . . . 46 3.4.2 Relative pose analysis results . . . . . . . . . . . . . . . . . . . . 49 3.4.3 Correlation results . . . . . . . . . . . . . . . . . . . . . . . . . 50 3.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 3.5.1 Sensitivity of the our methodology . . . . . . . . . . . . . . . . . 57 3.5.2 Selective vulnerability of preterm birth . . . . . . . . . . . . . . 58 3.5.3 Neuro-developmental considerations . . . . . . . . . . . . . . . . 64 3.5.4 Contribution and limitation of the study . . . . . . . . . . . . . . 66 3.6 Funding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 4 Structural Changes of the Basal Ganglia in Children with Chronic Man- ganese Exposure 70 4.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 4.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 4.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 4.3.1 Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 4.3.2 V olume based analysis . . . . . . . . . . . . . . . . . . . . . . . 78 Contents vi 4.3.3 Surface based analysis . . . . . . . . . . . . . . . . . . . . . . . 78 4.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 4.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 5 Pose, Surface and Connectivity Changes of the Corpus Callosum in Colle- giate Contact Sport Athletes 91 5.1 Comparisons of the Relative Pose of the Corpus Callosum in Contact Sport Athletes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 5.1.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 5.1.2 Description of purpose . . . . . . . . . . . . . . . . . . . . . . . 93 5.1.3 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 5.1.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 5.1.5 New or breakthrough work to be presented . . . . . . . . . . . . 97 5.1.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 5.2 A T1 and DTI Fused 3D Corpus Callosum Analysis in Contact Sports Athletes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 5.2.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 5.2.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 5.2.3 Subjects and Methodologies . . . . . . . . . . . . . . . . . . . . 101 5.2.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 5.2.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 6 A study of brain white matter plasticity in early blinds using Tract Based Spatial Statistics and Tract Statistical Analysis 112 6.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 6.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 6.3 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 6.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 6.5 Discussion and conclusion . . . . . . . . . . . . . . . . . . . . . . . . . 118 Contents vii 7 Disentangling the primary brain effects of vascular risk factors from mild cognitive impairment in aging subjects. 120 7.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 7.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 7.3 Subjects and Methodologies . . . . . . . . . . . . . . . . . . . . . . . . 124 7.3.1 Data and Preprocessing . . . . . . . . . . . . . . . . . . . . . . . 124 7.3.2 Surface Based Sampling . . . . . . . . . . . . . . . . . . . . . . 126 7.3.3 Statistical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 127 7.3.4 Correlation Analysis . . . . . . . . . . . . . . . . . . . . . . . . 130 7.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 7.4.1 Correlation analysis results . . . . . . . . . . . . . . . . . . . . . 133 7.5 Discussion and Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . 137 7.5.1 Methodological Considerations . . . . . . . . . . . . . . . . . . 140 7.5.2 Anatomic and Functional Implications . . . . . . . . . . . . . . . 142 7.6 Limitations and Future directions . . . . . . . . . . . . . . . . . . . . . . 145 8 Conclusion and Outlook 147 8.1 Summary of Contributions . . . . . . . . . . . . . . . . . . . . . . . . . 147 8.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 References 154 List of Figures 1.1 subcortical structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.2 T1 weighted brain MRI scans from four healthy subjects in different age range (A: a neonate; B: a pre-school age children; C: a young adult; D: an elder subject). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.3 Figure of thesis organization. The presented thesis is organized according to the age ranges of the investigated populations, from neonates to elderly subjects. In the figure, chapters are connected with their corresponding populations, subject age ranges, brain structures, as well as algorithms used. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.1 (A) 3D visualization of the pose of mean shapes averaged from the preterm (red) and term groups (blue). More details can be found in Chapter 3. Ar- eas where the mean shapes of two groups are overlaid appear in purple. (B) 3D visualization of the pose shift in pre-season (blue)vs. postseason (red) contact sport players. More details can be found in Chapter 5. . . . . 19 2.2 Example maps of diffusion tensor derived metrics. Top row from left to right: λ 1 ,λ 2 , andλ 3 maps. Bottom row from left to right: color FA maps, FA maps, MD maps. All the maps are selected of the same subject, from the same axial slide (same location within the brain). . . . . . . . . . . . 24 2.3 subcortical connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 List of Figures ix 2.4 Overlaid T1 segmentations (A) and FA maps in the same subject scanned pre-season (B) versus post-season (C). In figure A, pre- and post-season CC segmentations are represented in yellow and blue respectively, with the overlaid areas appearing in orange. Comparing the three figures, cir- cled area shows both structural and diffusion changes, the area indicated by a triangle is structurally different but has an unchanged diffusion pat- tern, and vice versa for the area shown by the arrow [148]. . . . . . . . . 26 2.5 Traditional white matter analyses: voxel-wise analysis (A), skeleton based analysis (B), midline based analysis (C), mid-plane analysis (D). . . . . . 28 2.6 Surface of CC and the illustration of sampled voxels. The red line on the right side of the figure is the midline of the CC, and pink crosses are surface vertices. In the direction perpendicular to the midline and point- ing to each vertex, blue stars represent voxels projected to each of the vertices, and a mean index (FA, MD....) of projected voxels is assigned to the vertex for later statistics. . . . . . . . . . . . . . . . . . . . . . . . 29 3.1 Diagram of the combined shape and relative pose analysis. For simplic- ity, only the surfaces of thalamus are used for illustration. Top Row: the subcortical structures are first segmented from T1 images, and then re- constructed to 3D surface models; Middle Row: surface based morphom- etry and correlation analysis; Bottom Row: relative pose based statistics and correlation analysis. In the pipeline, letter a-g represent: a: T1-MR images; b: subcortical structure segmentation; c: 3D surface models; d: one-to-one correspondence obtained from registration; e: surface-based group differences results; f: surface based correlation results; g: pose parameters obtained from procrustes alignment; h: group differences; i: pose-based correlation results. . . . . . . . . . . . . . . . . . . . . . . . 37 List of Figures x 3.2 Thalamus and putamen segmentation. The rules of manual segmenta- tion were guided by anatomical references of the human striatum (Prensa et al., 2003; Morel et al., 2002) and thalamus (Niemann et al., 2000; Wiegell et al., 2003; Erbetta et al., 2009). Due to the lack of tissue con- trast on neonatal volumetric T1 images, it is difficult to accurately sepa- rate the globus pallidus from the putamen. Since our hypothesis is related to the ventral striatum, we included the globus pallidus in our putamen segmentation and 3D surface construction. For simplicity, we use the word ’putamen’ to represent the combined surface of the globus pallidus and putamen throughout the text. . . . . . . . . . . . . . . . . . . . . . . 40 3.3 Distribution of the parameters used in this study vs. post-conceptional age (PCA). Columns from left to right are: volumes vs. PCA; (Normal- ized) mean detJ vs. PCA; scale parameter (with mean pose subtracted) in log-Euclidean space (logS) vs. PCA. Rows from top to bottom represent parameters from: left thalamus, right thalamus, left putamen and right putamen. Scatter points in the figure represent observed results, while lines represent their corresponding linear regressions. In all the figures above, data from the preterm group are marked in red (circles and solid lines), while that from the term group are marked in blue (cross sym- bols and dash lines). Note that nearly all the parameters for the preterm group are distributed lower in the graphs, compared to that from the term controls, indicating a smaller size of structures in the preterm group. . . . 45 3.4 Thalamus statistical p-maps: Overall p-values are p=0.0035 for MAD + mTBM(a), p=0.0901 for MAD(b) and p=0.0683 for detJ(c). . . . . . . . 48 3.5 Vertex-wise correlation determinant (a) and radial distance (b) maps of the thalamus. Widespread shrinkage of preterm group is present in both thalami, and the clusters with significant preterm vs. term differences (seen in Fig. 3.4) are all falling on the preterm< term areas. . . . . . . . 48 List of Figures xi 3.6 Vertex-wise correlation determinant (a) and radial distance (b) maps of putamen. Widespread shrinkage of preterm group is present in both puta- men, and the clusters with significant preterm vs. term differences (seen in our previous publication [220]) are mostly falling on the preterm < term areas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 3.7 3D visualization of the pose of mean shapes averaged from the preterm (red) and term groups (blue). Areas where the mean shapes of two groups overlaid appear in purple. To better visualize the pose changes, enlarge- ments of locations (a), (b), (c) in the bottom right figure are presented in the top right, bottom left, and bottom right figures, respectively. Note the borders of these two structures: shifts of pose are evident on the left putamen (a), left thalamus (b), and right putamen (c), but are quite small in the right thalamus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 3.8 Vertex-wise correlation coefficient maps have been generated based on determinant (left) and radial distance (right), respectively. The upper row are displayed in a superior view, while the bottom ones are seen from the inferior one. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 3.9 P-maps of the corresponding to the correlation coefficients, derived from the determinant map (left), and radial distance maps (right). The upper row are displayed in a superior view, while the bottom are seen from the inferior one. In (A), overall p-values are p=0.0443 for the corre- lation between vertex-wise determinants on thalamus surfaces and puta- men volumes, p=0.0745 for the correlation between vertex-wise determi- nant on putamen surfaces and thalamus volumes. In (B), overall p-values are p=0.0624 for the correlation between vertex-wise radial distance on thalamus surfaces and putamen volumes, p=0.0285 for the correlation between vertex-wise radial distance on putamen surfaces and thalamus volume. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 List of Figures xii 3.10 The correlation between the thalamus and the putamen in the left (upper row) and right (bottom row) hemispheres, tested using pose parameters: logS(left column),||logR||(middle column),||logd||(right column). More details, such as correlation coefficients and p-values, are presented in Table. 3.2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 4.1 Illustration of segmentation and the corresponding 3D structures in axial and coronal views. Red, blue and yellow each represents the caudate, the putamen and the globus pallidus, respectively. None of the participants had visible hyper- or hypointensities in the basal ganglia. . . . . . . . . . 75 4.2 Mean ratio map of detJ between high-exposed and low-exposed groups are displayed in four views: superior (A), inferior (B), posterior (C), and anterior (D). Areas in red represent an increase of detJ in the high- exposed group compared with that in the low-exposed group, indicating an expansion of the corresponding areas. Areas in blue show a decrease detJ in the high-exposed group compared with that in the low-exposed group, indicating a contraction of the corresponding areas. . . . . . . . . 83 4.3 Statistical results of the multivariate analysis are displayed in four views: superior (A), inferior (B), posterior (C), and anterior (D). Areas in colors other than deep blue represent vertex-wise significances of the multivari- ate analyses (p<0.05). . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 4.4 Vertex-wise correlation coefficients between vertex-wise det J values with Santa Ana scores are displayed in four views: superior (A), inferior (B), posterior (C), and anterior (D). Areas in red colors represent a negative correlation between surface areas and motor performance, and vice versa for blue colors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 4.5 Correlation of vertex-wise detJ values with Santa Ana scores are dis- played in four views: superior (A), inferior (B), posterior (C), and an- terior (D). Areas in colors other than deep blue represent vertex-wise significances of the multivariate analyses (p <0.05). . . . . . . . . . . . 86 List of Figures xiii 5.1 3D visualization of the mean shape of all the CCs in the mean poses averaged from pre-season group (red) and post-season group (blue) re- spectively. A shift in pose is evident in terms of a rotation on the anterior and posterior ends of the CC, while less visible differences can be seen in size and transformation. . . . . . . . . . . . . . . . . . . . . . . . . . 107 5.2 In (a), from left to right: axial view, sagittal view, and coronal view. In (b), overlaid T1 segmentations (A) and FA maps in same subject scanned pre- (B) vs. post-season (C). In figure A, pre- and post-season CC seg- mentations are represented in yellow and blue, respectively, with the overlaid areas appearing in orange. Comparing the three figures, circled area shows both structural and diffusion changes, the area indicated by a triangle is structurally different but has an unchanged diffusion pattern, and vice versa for the area shown by the arrow. . . . . . . . . . . . . . . 108 5.3 Surface of CC for one of our subjects and the illustration of sampled vox- els. The red line on the right side of the figure is the midline of the CC, and pink crosses are surface vertices. In the direction perpendicular to the midline and pointing to each vertex, blue stars represent voxels pro- jected to each of the vertices, and a mean index (FA, MD....) of projected voxels is assigned to the vertex for later statistics. . . . . . . . . . . . . . 109 5.4 FA maps for the same subject sampled on the surface of the CC, using dif- ferent sampling distances. From left to right: R= 0.6mm 3 ,R= 0.8mm 3 , R= 1.0mm 3 . Note: to save computation time, results shown in this figure were obtained on a downsampled surface. All three results are grossly matched, while sampling using a larger radius leads to smoother regional changes. R= 0.6mm 3 was selected here. . . . . . . . . . . . . . . . . . . 109 5.5 Group difference results between pre- and post-season subjects obtained using centerline representations. Projections to the each of the centerline vertex were made through mean FA or MD of the perpendicular plane to the midline in the location of the given vertex. Midline vertices with a p-value< 0.05 are displayed in red. . . . . . . . . . . . . . . . . . . . . 110 List of Figures xiv 5.6 pre- versus post-season contact sports players using different measures. Vertex-wise corresponding p− values are displayed. In addition, whole structure corrected p− values are p=0.5061 for detJ, p= 0.2253 for (e1,e2,e3), p= 0.0797 for FA, p= 0.1181 for MD, p= 0.0352 for (λ 1 ,λ 2 ), p= 0.0136 for (λ 1 ,λ 2 , detJ). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 6.1 Statistical analysis of FA using TBSS(A)(C) and TSA(B)(D). For TBSS, results are shown as mean FA skeleton maps. Significant areas with p<0.01 are marked in red. For TSA, results are shown for the bilat- eral corticospinal tracts (CST), inferior fronto-occipital tracts (IFO), infe- rior longitudinal fasciculus tracts (ILF), superior longitudinal fasciculus tracts (SLF), uncinate tracts (UNC), and the corpus callosum (CC). Sig- nificant areas with p<0.01 are colored in dark blue and encircled with a white line. Figure A and B: results under the hypothesis of blind <sighted. Row C and D: results under the hypothesis of blind >sighted. . 117 7.1 Surface of CC and the illustration of sampled voxels. The red line on the right side of the figure is the midline of the CC, blue stars represent voxels within CC, and pink, yellow, as well as green crosses represent surface vertices. In the direction perpendicular to the midline and pointing to each vertex, voxels within pink, yellow, and green areas are projected to the vertices with the corresponding colors. Mean index (FA, MD....) of projected voxels is assigned to the vertex for later statistics. . . . . . . . 128 7.2 Group analysis of MCI-l vs. controls (1 st column), MCI-h vs. controls (2 nd column), and MCI-l vs. MCI-h (3 rd column ) using 5 different mea- sures: a) detJ; b)(s1,s2,s3); c) mean FA; d) (λ 1 ,λ 2 ); e) (λ 1 ,λ 2 ,s1,s2,s3). Vertex-wise corresponding p− values are color-coded according to the color bar in the upper left corner. P- maps are smoothed using heat ker- nel algorithm [47]. In addition, whole structure-wise corrected p− values are presented in Table. 7.1. . . . . . . . . . . . . . . . . . . . . . . . . . 131 List of Figures xv 7.3 Average map of detJ and mean FA between groups are color-coded ac- cording to the color bar in the upper left corner. When these results are compared with Fig. 7.2, we can see the main direction of change: nearly all the significance areas fell in the Controls > MCI-l, Controls> MCI-h, as well MCI-l> MCI-h areas. . . . . . . . . . . . . . . . . . . . . . . . 132 7.4 Vertex-wise significance results of correlation analyses between det J as well as mean FA vs. 5 neuropsychlogical scores. P- maps are smoothed using heat kernel algorithm [47]. . . . . . . . . . . . . . . . . . . . . . . 135 7.5 Vertex-wise correlation coefficient maps have been generated based on detJ (left column) and mean FA (right column), respectively. Compared this figure with Fig.7.4, we can see the direction of the correlation anal- yses: nearly all the significant regions represent positive correlations. . . . 136 List of Tables 3.1 P-value of statistical analyses on pose parameters: 6 sets of parame- ters characterizing relative pose of left thalamus (LTha), right thalamus (RTha), left putamen (LPuta), and right putamen (RPuta) are investigated here using univariate and multivariate analyses. Parameters are catego- rized aslogS,||logR||,||logd|| for univariate tests, and as (θ x ,θ y ,θ z ), (x, y, z), and a combination of 7 parameters for multivariate tests. All the p-values are obtained after permutation testing. Significant p-values (p < 0.05) are highlighted in cyan, while p-values that are interestingly low but failed to reach significance are highlighted in yellow. . . . . . . . . . 52 3.2 4 sets of correlation results based on pose parameters: logS,||logR||,||logd||, between 2 thalami and 2 putamen. Both coefficients and p values are provided in each of the test, with significant correlations highlighted in babypink. Here, we set the significance with p < 0.01. Note: here we investigated RPuta vs. LTha, and LPuta vs. RTha as an exploratory anal- ysis. Although RPuta vs. LTha, and LPuta vs. RTha are not directly connected to each other, they all have different levels of projections to the cortex, and interhemispheric callosal fibers may indirectly link them together (Gazzaniga, 2000). . . . . . . . . . . . . . . . . . . . . . . . . . 56 4.1 Demographics of our subjects. . . . . . . . . . . . . . . . . . . . . . . . 77 List of Tables xvii 4.2 Group difference results of traditional univariate volume based analyses (VBA), and multivariate tensor based morphometry (mTBM) analyses on left (l-), right (r-), and combined (c-) basal ganglia structures: Putamen (Puta), Globus Pallidus (GP), Caudate (Cau). All the p-values are cor- rected for multiple comparisons using structure-wide permutation testing with 10,000 permutations [150, 153]. Significance is set at p< 0.05, and significant values are marked using ∗∗ . Low p-values implying trends are marked using∗ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 5.1 P-value of statistical analyses on pose parameters: 13 sets of parame- ters characterizing the relative pose of the CC are investigated here using univariate and multivariate analyses. Parameters are categorized as logS, ||logR||,||logd||, θ x ,θ y ,θ z , x, y, and z for univariate tests, and as (θ x , θ y , θ z ), (x, y, z), (logS,||logR||,||logd||), and a combination of 7 parameters for multivariate tests. All the p-values are obtained after permutation testing. Significant p-values (p < 0.05) are highlighted in light cyan. Non significant, but interested low p-values are highlighted in light grey. . 96 7.1 Structure-wise corrected p− values for different measurements are dis- played. All the p-values were corrected using a permutation based anal- ysis with 10,000 permutations. Significance is set to p < 0.05, and is highlighted in light cyan. P-values implying trends are highlighted in light grey. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 7.2 Structure-wise corrected p− values for different measurements are dis- played. All the p-values were corrected using a permutation based anal- ysis with 10,000 permutations. Significance is set to p < 0.05, and is highlighted in light cyan. P-values implying trends are highlighted in light grey. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 Chapter 1 Introduction “Anymancould,ifheweresoinclined,bethesculptorofhisownbrain." -SantiagoRamónyCajal,1897 The earliest rigorous explorations of brain anatomy can be traced to the mid-16 th century, when Andreas Vesalius (1514-1564) published his great worksDeHumaniCor- poris Fabrica Libri Septem with detailed drawings of the human brain. One hundred years later, Thomas Willis (1621-1675) coined the term ‘neurology’. He hypothesized that the brain could be segregated into various functional ‘modules ’, as described in his book Cerebri Anatome. The investigation of brain anatomy and function proceeded slowly until the late 19 th century, when Pierre Broca (1824-1880) and Carl Wernicke (1848-1905) utilized the relation of brain dysfunction and brain lesions to find language areas. This accomplishment is often cited by neuroscientists as the first proof of brain 2 function localization. Studies of the brain are closely tied with brain dysfunctions, and much of our under- standing of the brain’s structure-function associations have come from analyses of brain lesions. Enlightened by Broca and Wernicke, our understanding of the human cerebral cortex and its functions had progressed considerably by the turn of the 20 th century. The next two major findings occurred in the early and middle 20 th century with the discov- ery by Gordon Holmes (1876-1965) [72] of the primary visual cortex from war injuries, and the association of the hippocampus with memory from the famous patient H.M., by Brenda Milner [216]. Over the years, various histochemical techniques were then devel- oped, and our knowledge of the human brain increased enormously at cellular, structural, and functional scales. However, the invasive methods of those times largely hinder the analysis of brain changes in longitudinal monitoring and large-scale studies. Further, given the uniqueness of each individual’s brain and the limited numbers of brain sam- ples, conclusions from postmortem investigations cannot easily be generalized to larger populations. The advent of magnetic resonance imaging (MRI) technology in the mid-1970s has fueled a scientific revolution in the field of neuroscience. Providing a range of soft tis- sue contrasts without ionizing radiation to the subjects, MRI enables the inspection of anatomy and function in the living brain, thus making the in-vivo analysis of brain al- terations possible. On top of this unprecedented imaging technique, a growing number of MRI post-processing algorithms are designed to tackle the difficulties of comparing brain anatomical and functional maps that vary substantially between individuals. These studies have paved the way for large-scale population based analyses. The present disser- 1.1 The Human Brain Across the Lifespan 3 tation is dedicated to structural and diffusion MR imaging analysis, and describes three post-processing algorithms to characterize brain morphological alterations. This chapter provides the neurological background for the studies performed, and introduces to the organization of this thesis. 1.1 The Human Brain Across the Lifespan 1.1.1 The Developing Brain The brain is the most complicated structure in the human body, and has three primary developmental divisions: prosencephalon, mesencephalon, and rhombencephalon. The prosencephalon, also referred as the forebrain, is the largest part of the three, and un- dertakes the high-level brain processing. The forebrain is often considered as two parts: the highly folded cortical areas on the surface - cerebral cortex -, and a variety of sub- cortical structures below the surface - the limbic system, basal ganglia, thalamus, and hypothalamus. The cerebral cortex can be further divided into frontal, temporal, parietal, and occipital lobes, each been involved in specific functions. The brain is mainly made up of two tissue types: the grey matter, areas in a pinkish grey color resulting from ac- cumulated cell bodies, and the white matter, areas in pinkish white color resulting from myelinated axons. In MRI, we can generally differentiate these two types, thus making subsequent analyses possible. Although it has long been viewed as physiologically static, the brain is constantly changing via numerous electrical and biochemical activities, and it grows and degen- erates throughout the lifespan. Starting from the 4th week post-conception, the brain 1.1 The Human Brain Across the Lifespan 4 (a) The human brain anatomy. Adapted from [lin]. (b) An illustration of the human brain sub-cortical structures. Adapted from [92]. Fig. 1.1 subcortical structures undergoes a major development period in the first 5 years of life, detailed pruning during the following decades, and gradually degenerates with aging. Studies have shown that brain maturation involves asynchronous but carefully programmed cascades of neuron proliferation, apoptosis, migration, and fine-tuning that varies largely by brain structures [240, 248]. The succession of events follows an inside-out rule that starts early from ventral and deep brain structures, such as the deepest layer of the neocortex and the tha- lamus, which are responsible for basic life-sustaining functions [240, 256]. On the other hand, the dorsal stream of the brain, such as the prefrontal cortex, may undergo constant changes that continue into late adolescence or even well into early adulthood [170, 248]. Zooming into one single brain region, especially the gray matter, the developmental time-course is usually non-linear. Cortical gray matter reaches a net maximum around age 7, which is followed by a decline into adulthood. However, growth rates vary from 1.1 The Human Brain Across the Lifespan 5 Fig. 1.2 T1 weighted brain MRI scans from four healthy subjects in different age range (A: a neonate; B: a pre-school age children; C: a young adult; D: an elder subject). one cortical region to another. For instance, the frontal and parietal lobe have grey mat- ter developmental peaks around 12 years of age, while this peak of grey matter volume presents 3-4 years later in the temporal lobe [82]. Subcortical deep gray matter follows similar developmental time courses as the cortex, but differ in terms of the exact turning- points between increase and reduction due to pruning. For instances, the striatum in- creases in volume until age 12, and then slightly decreases over time until it reaches it’s adult size around 20 years old; the thalamus has its developmental peak between 14 to 16 years of age; the pallidum reaches its maximum growth around age 8, followed by pro- gressive volume loss until early adulthood [203, 218]. On the other hand, white matter undergoes continuous increase with an approximately linear trajectory until 20 years old [118]. 1.1.2 The Aging Brain Like other organs of the human body, the brain ages. The earliest mental sign of ag- ing involves difficulties in word finding and increased efforts in new learning, while the 1.1 The Human Brain Across the Lifespan 6 physiological aspects of the brain aging process usually start years before functional de- creases. Although the theory that the apoptosis is the trigger of brain degeneration is still controversial [177, 196], decreases in the numbers of neurons and synapse as well as changes in transmitter levels and dendritic extent are inevitable consequences of brain aging [48, 228]. In-vivo brain imaging is readily able to reveal the brain anatomical al- terations associated with age, and as some of the earliest neuroimaging findings used CT or MRI. In particular, the enlargement of ventricles and cortical sulci are reported to be two overt indicators of aging [66, 181, 272]. To probe into the extent and starting age of brain anatomical alterations, a large-scale MRI study recruited subjects aged 18 to 97 and reported significant brain volume decline started from adulthood and continues into old age [74]. As the case in brain development, different types of tissue exhibit diverging trajec- tories in brain degeneration. In Fotenos’s study mentioned above [74], the gray matter gradually decreases in the early 30’s with a close to linear pattern, while the white mat- ter remains roughly constant and then decline at a slower rate compared to that of the gray matter. Apart from the global trends of brain atrophy, more dedicated study on par- cellated brain regions reveal a preferential influence of healthy aging on several specific brain regions. For instance, in a longitudinal study of 127 subjects with mean age at baseline of 50 years old, substantial five-year changes were reported in the caudate, the hippocampus, and the association cortices, while the entorhinal and the primary visual cortex were minimally affected [202]. In the same study, the shrinkage of certain brain regions, such as the hippocampus and the primary cortex was found to increase with age [202]. In agreement with this, another consecutive study on 138 elderly individuals aged 1.1 The Human Brain Across the Lifespan 7 over 65 identified accelerated changes in gray matter volumes in the frontal and parietal regions [67]. Understanding the heterogeneity of brain development and degeneration is of fundamental importance for identifying and tracking brain anomalies. 1.1.3 External Disturbances and Diseases The progress of brain maturation might be more complex than orchestrating a pre-programmed sequence of events. The whole process of brain development is also shaped by interac- tions with external environment and sometimes diseases. In many cases, brain subcortical regions are where neurophysiological disorders or diseases originate. For example, preterm birth causes a variety of complications, and hy- poxia is the most common one. Embedded in the cerebral vascular system, the mid brain location of the thalamus makes it the first to be affected in preterm birth associated hy- poxia. In patients with Alzheimer’s disease, neurofibrillary tangles and amyloid plaques accumulations usually start in hippocampus. Hippocampal atrophy is therefore identified as an important diagnostic marker for Alzheimer’s disease [60, 215]. Schizophrenia and Parkinson’s disease are commonly thought to involve miscommunication among differ- ent brain regions, and result from imbalanced dopamine activities [39, 80]. The basal ganglia, which gates the pathways to the cortex and modulates multiple functional do- mains, has been consistently reported to be altered in patients suffering schizophrenia and Parkinson’s disease [57, 96, 213]. The onset of external disturbances and diseases in different stages of the human lifes- pan has different power of sculpture on the trajectory of development. Exploring how these factors interact and influence the brain is of particular importance, in terms of un- 1.2 Introduction To the Thesis 8 Fig. 1.3 Figure of thesis organization. The presented thesis is organized according to the age ranges of the investigated populations, from neonates to elderly subjects. In the figure, chapters are connected with their corresponding populations, subject age ranges, brain structures, as well as algorithms used. derstanding neuro-substrate for associated disturbances or disorders as well as the process of the brain development itself. 1.2 Introduction To the Thesis The presented thesis is dedicated to the analysis of structural MRI and DTI, and aims to characterizing alterations in subcortical networks across the human lifespan. Specifically, we use shape, pose and connectivity measurements to pinpoint subcortical biomarkers in 1.2 Introduction To the Thesis 9 response to external disturbances including preterm birth, chronic manganese exposure, mild traumatic brain injury, as well as cardiovascular disease coupled to mild cognitive impairment. In this thesis, the chapters will be organized according to the age of investi- gated populations, as follows: In chapter 2, a brief background and the significance of subcortical shape, pose and fusion analyses are presented. In chapter 3, a structural MRI based brain subcortical structures analysis to charac- terize gray matter structural morphology, and a pose analysis assessing the positional changes are shown. These methods are applied on the thalamus and ventral striatum in premature neonates, to examine the brain deep gray matter alterations associated with prematurity. In chapter 4, the previously introduced structural MRI based surface morphometry analysis is applied in the basal ganglia, to investigate the brain anatomy association of chronic manganese waterborne exposure in school-age children. In chapter 5, a T1 and DTI fusion, with increased power to detect white matter al- terations is introduced. Both the newly introduced fusion method as well as previously introduced pose analysis are applied in young contact sport athletes, to analyze the corpus callosum changes linked to mild traumatic injury in one play season. In chapter 6, the previously introduced T1 and DTI fusion method is applied on the corpus callosum in an aging population dataset, to disentangling the effect of cardiovas- cular diseases from mild cognitive impairment. In chapter 7, conclusion and future directions are discussed. Chapter 2 Background and Significance Toward understanding how an external factor affects brain morphology and its long-term influence on functions and behavior, the first step, albeit the biggest challenge, is to find and to quantify brain alterations. In brain imaging analysis, the investigation of ’where the alterations are located in the brain, and how the onset areas change’ have a broad range of solutions. Intensity- based comparison is no doubt the most direct way. Signal intensity changes on T1 or T2 MR images have been long recognized as signs of brain lesions [77], and have been associated with a variety of neurological disorders[188]. With the introduction of im- age post-processing methods, notably image registration, cross sessional or longitudinal whole brain comparisons have become feasible. Some whole brain analysis methods are still widely used today, to name a few, voxel based analyses on T1 images or FA maps, or volume analyses over whole brain gray and white matter. Along with the growing un- derstanding of brain anatomy and pathology, more investigators have begun to shed light 2.1 Shape Analysis 11 on putative brain structures, to precisely probe brain alterations. To circumvent the in- efficiency of traditional whole brain algorithms in pinpointing brain changes, structural- specific analyses have emerged. Structural-specific analyses study particular essential properties of different brain regions, for example, the curvature and thickness for the cor- tex, the shape and relative position for the subcortical structures, and the connectivity for the white matter pathways. In striking contrast with the extensive investigation of cortical alterations in brain diseases and disorders, the relative changes in subcortical structures are much less con- sidered. The emphasis of this dissertation is to characterize the subcortical alterations in the brain, and in particular, here we will focus on describing the shape, pose, and connec- tivity changes. This chapter provides the significance and methodology background of the shape, pose, and connectivity analyses, which will be applied to characterize specific biomarkers to several medical conditions in later chapters of this thesis. 2.1 Shape Analysis 2.1.1 Significance of shape analysis V olume analysis is the most direct and intuitive way to depict the cell loss or increase in subcortical structures. Until recently, structural MRI studies have attempted to investigate the diseases’ associated subcortical structural impairments by comparing the volumes of the whole structure between healthy and affected populations [2, 11, 113, 280], or by dividing the structures into several subregions [15, 35, 56, 75, 193]. The whole volume based method facilitates an intuitive and coarse estimation of subcortical anatomy, but 2.1 Shape Analysis 12 has been ineffective in detecting subtle anatomical changes. Subdivision based studies are more tuned to the heterogeneity of structures, like the corpus callosum (CC) and the thalamus, but may easily be biased due to inconsistent classification, as well as arbitrary delineation of subdivisions. With the development of image registration algorithms, voxel-wise comparisons emerged and quickly drew broad attention as they provide increased sensitivity as compared to vol- ume based methods [16, 24]. However, it is common for voxel based analysis to detect scattered abnormalities fall into non-specific brain regions, thus cannot be well localized with subcortical structures. Moreover, voxel-wise algorithms are inevitably prone to reg- istration artifacts, which are non-negligible in subcortical tissue boundaries, for example, in regions close to lateral ventricles. To have a more complete view of subcortical structures and better depict ’where’ and ’how’ the brain region changes, researchers and scientists recently started to investigate each individual brain structure using shape analysis [242]. By being able to consider- ing the heterogeneity of subcortical structures and retaining high statistical power, shape analysis can describe morphological changes more precisely than merely ’reduced vol- ume in the basal ganglia’, ’dilated ventricles’, or ’significant alterations near the thalamus and putamen’. Preliminary results on adult populations have shown the feasibility and effectiveness of using univariate and multivariate morphometry measurements as dis- criminators of brain sub-cortical alterations [263, 266]. In this thesis, parameterized 3D representation based shape analysis will be exten- sively used in detecting disease or disorder associated subcortical alterations in popula- tions ranging from the neonates to older adults. In shape analysis, initial inputs are seg- 2.1 Shape Analysis 13 mentations of brain structures like the thalamus, the putamen, the caudate. Based on these binary segmentations, the corresponding 3D surface models are constructed, followed by surface based alignment and subsequent statistical analysis on structure-specific features. An overview of surface registration, surface multivariate tensor-based morphometry, and multivariate statistical analyses will be covered in the following sections. 2.1.2 Substructural Surface Registration In shape analysis, 3D surface parameterization is essential for visualization, registration, and statistical comparisons. Based on binary segmentations of targeted subcortical struc- tures, triangulated meshes are first constructed using a marching cube algorithm [161]. Frameworks based on surface meshes allows 2D coordinate grids to be mapped onto the structures’ surfaces, thus transforming the 3D problem into a 2D parameter domain. In particular, holomorphic differential based conformal mapping has been effectively used in parameterizing topologically complex structures such as the cortex and the ventri- cles, and cylinder-resembled structures such as the hippocampus, the caudate, and the thalamus [263, 266]. The computation is intrinsic and efficient, and the induced parame- terization ensures minimized angular distortions [260]. For most of the subcortical structures with elongated shape (i.e. the corpus callosum, and the caudate), mapping to a canonical space such as a sphere will introduce substantial distortion [242]. As a result, a key step before the conformal parameterization is topology optimization, in which two cuts are introduced at some geometrically extreme positions. These extremities are anatomically valid and stable across subjects, for example, the anterior and pulvinar for the thalamus, the anterior and posterior limb of internal capsule 2.1 Shape Analysis 14 for the putamen. The optimization step turns the surface model into a genus-zero surface with two open boundaries that resembles a cylinder. By the topology optimized model into a rectangle, the nonlinear conformal parameterization on 3D surfaces are efficiently transformed into a linear one on a 2D plane. To compare subcortical shape features between groups, parameterized vertices on surface models need to be mapped one-to-one across subjects. This is commonly done via constrained harmonic registration, in which the constrained harmonic map is computed as in [260]: Given a subject surface S s and a template surface S t , the corresponding conformal parameterizations areτ s :S s →ℜ 2 andτ t :S t →ℜ 2 . To compute a mapφ :S s →S t , one can first find a harmonic map between two parameter domains: τ :ℜ 2 →ℜ 2 , such that: τ◦ τ s (S s )=τ t (S t ),τ◦ τ s (∂S s )=τ t (∂S t ),∆τ = 0 Then φ may be indirectly calculated as φ =τ s ◦ τ◦ τ − 1 t . Because both τ s and τ t are conformal maps, τ is a harmonic map, the resulting φ should also be a harmonic map. Since certain landmark curves, such as the boundaries of thalami, need to be matched to guarantee the right correspondence during the process, the above mentioned algorithm is called the constrained harmonic registration. For structures that may vary a lot in shape between subjects, for example, neonatal subcortical structures in neonates with different post-conception age, a more optimal reg- istration algorithm is needed to enforce biologically meaningful match between surfaces and minimize distortions. In this case, a surface fluid registration algorithm can be used, 2.1 Shape Analysis 15 which treats the moving image as a viscous fluid to be transformed into the target, and the thorough description can be referred to [219, 220]. 2.1.3 Surface Multivariate Tensor-based Morphometry (mTBM) The aim of shape analysis is to compare intrinsic surface features of targeted subcortical structures. Unlike cortical surfaces, which have high curvature contrast, many subcortical structures have cylinder-like shape in which enlargement or shrinkage can be directly reflected by the radius of that location. Therefore, in conformally parametrized mesh surfaces, the most intuitive and efficient morphology measurement is medial axis distance (MAD). MAD is the distance between a surface vertex to the medial line of the structure, which can be obtained through averaging the iso-parametric curves on the surface. MAD provide a direct assessment of the thickness of a vertex, and may capture an intrinsic feature of the corresponding location in the normal direction. Registration induced deformations or their derivatives are alternative characteriza- tions of shape as they capture changes within surfaces. Supposeφ is a map from surface S 1 to the surface S 2 (φ :S 1 →S 2 ), the derivative map φ is the linear map between their tangent spaces. For triangle surface meshes, one surface face[v 1 ,v 2 ,v 3 ] and its mapped face in the transformed surface[w 1 ,w 2 ,w 3 ] can be isothermally embedded onto a plane, where their planar coordinates are also denoted byv i ,w i ,i= 1,2,3. The discrete deriva- tive mapJ can be explicitly calculated as in [219, 263, 265], by: J=dφ =[w 3 − w 1 ,w 2 − w 1 ][v 3 − v 1 ,v 2 − v 1 ] − 1 (2.1) 2.1 Shape Analysis 16 Functions of components from J, including detJ, trJ, are often used in group com- parisons. However, these univariate measures are limited in capturing shape changes. To incorporate the directional changes in group comparisons, multivariate comparisons based on deformation tensors have been introduced and applied in evaluating changes in volumes and surfaces [154, 263, 266]. Deformation tensor is defined as S= √ JJ T , which, in the case of above mentioned surface morphometry, represents a 2D ellipse with axes showing the extent and direction of changes between the two surfaces at that location. Since the deformation tensors are positive definite matrices, which do not form a vector space, a Log-Euclidean framework is commonly chosen to facilitate computations [153, 263, 266]. Using the inverse of the exponential map, the manifold-valued elements of deformation tensors are projected into the tangent plane at the origin where Euclidean space formula can be applied, thus simplifying computation of the statistics [12]. As a result, instead of using S, statistics are computed on the matrix logarithm log(S) in multivariate tensor based morphometry. 2.1.4 Multivariate Statistical Analysis Based on previously described surface measures (i.e. MAD, detJ, log(S)), group-wise comparisons can be computed using Student t- test or a multivariate extension of it - Hotelling T 2 test. Given n-dimensional vectors for two groups P i , i= 1,2,...,n p , C j , j = 1,2,...,n c , different of group means can be measured using Mahalanobis distance 2.2 The Relative Position Analysis 17 (M): M= n p n c n p +n c (P− C) − 1 ∑ (P− C) T (2.2) whereP andC are the means of groupP i andC j ,n p andn c are the numbers of individuals in each group, and∑ − 1 is the combined covariance matrix of the two groups [153, 263, 266]. A transformation of M follows F distribution with degrees of freedom n and n p + n c − n− 1, where statistical significance can be evaluated. In practice, a combination of MAD and log(S) gives higher statistic power, and often been referred to as MADMTBM. 2.2 The Relative Position Analysis 2.2.1 Significance of the relative pose analysis To anatomically define a subcortical structure, describing its location within the brain is as important, if not more critical, as describing its shape. In each stage of brain devel- opment, subcortical structures retain a relatively stable position within the brain. This is not surprising as when we try to locate a specific brain structure on the brain MRI, we sometimes locate its surrounding easy-to-identify structures first, and use it as a guide to pinpoint this structure. However, the position of a brain structure on the brain image is a combination of its relative position within the brain, compound by the subjects’ head size and the arbitrary location within the scanner. This last bit of information depends on external environment which are different from individual to individual, thus are irrelevant for across-subject investigation. After transforming subjects’ brain image into a standard 2.2 The Relative Position Analysis 18 coordinate system - a common template, the irrelevant external information is removed and each subject is left with a residual position. This residual position, representing the relative position within the brain, is especially important in depicting the developing or degeneration patterns of the brain. Upon external disturbances, subcortical positional shifts are likely to happen, sec- ondary to the volume and shape changes of one or more neighboring structures. For instance, in early stage of brain development, injuries in part of the brain may trigger a series of linked events. The disturbances involves multiple interconnected regions, and the resulting alterations on connections may lead to a tendency to become closer of farther away from neighboring structures. On the other hand, in the context of neurode- generation, one of the neuroimaging hallmark of Alzheimer’s disease is the increased volumes of the lateral ventricles. Accompanying this enlargement of ventricles in the mid-brain location, the surrounding gray matters are also anticipated to exhibit a posi- tional shift. Therefore, the relative position of subcortical structures, complimentary to shape measures, may also help to indicate brain abnormalities and potentially serve as a discriminator of different clinical populations. To characterize the relative pose of each subcortical structure, point distribution models from all the subjects are aligned with the mean shape of the corresponding structure. Although the relative poses of subcorti- cal structures are not computed with respect to each other, the mean shapes of various subcortical structures form a within-brain coordinate system, in which ’a relative pose between neighboring structures’ can be indirectly inferred. Related background of the relative pose computation using Procrustes alignment will be discussed in the next sec- tion. 2.2 The Relative Position Analysis 19 (a) Use of relative pose as a possible discriminator of thalamus and putamen alterations in preterm neonates [149]. (b) Use of relative pose as a possible discriminator of CC alterations in pre- v.s. post-season contact sports players [Lao et al.]. Fig. 2.1 (A) 3D visualization of the pose of mean shapes averaged from the preterm (red) and term groups (blue). More details can be found in Chapter 3. Areas where the mean shapes of two groups are overlaid appear in purple. (B) 3D visualization of the pose shift in pre-season (blue)vs. postseason (red) contact sport players. More details can be found in Chapter 5. 2.2 The Relative Position Analysis 20 2.2.2 Procrustes Alignment Procrustes alignment is an isomorphic transformation that optimally superimposes two or more shape representations. In full Procrustes alignment, the shape itself, i.e. angles and parallel lines, is preserved, while the size and the location is freely adjustable [68, 211]. The relative pose of each subcortical structure is obtained by a full Procrustes fit of a template shape to the corresponding shape representation, which is the point distribution model in the context of this thesis. The template shape is usually selected as the mean shape that minimized the Procrustes distances [28]: S m =arg min S∈Shape space N ∑ i=1 D 2 p (S i ,S) (2.3) Where D 2 p is the squared Procrustes distance - the distance after Procrustes fit, and can be computed iteratively [211]. The transformations are commonly centered according to the center of mass of the object, and the similarity transformation is estimated in terms of a uniform scale, a rota- tion and a translation in x, y and z directions [68]. To be more specific, a full Procrustes alignment is defined as [27], T(X)=(sRX,d), where s is the scalar scaling factor, R is a 3× 3 rotation matrix and d is the translation vector(x,y,z) T . To composite the three types of transformations in one representation that facilitating batching multiplication of the similarity transformations, the Procrustes transformation 2.2 The Relative Position Analysis 21 can be written as a matrix representation [28]: T =    sRX d 0 T 1    (2.4) The matrices form a Lie group, and the integration of multiple similarity transforma- tion can be written as matrix multiplication [27, 28]. In the case of similarity transforma- tion, the matrix is a group and a Riemannian manifold [28], thus the standard Euclidean space statistics can not be directly applied onto it. To overcome it, a log-Euclidean frame- work, as introduced in section 2.1.3 is used, in which computations with group elements are conducted by means of their logarithm representations [27]. For Lie group elements, exponential and logarithm mappings follows the standard matrix exponentials and loga- rithms as explained in [28]. In group studies, it is intuitive to define a mean. Similar to Euclidean space, here the mean in Lie group is chosen as the Karcher mean, which is the point that minimized the squared distance [126]. The mean pose can be calculated iteratively as [27, 192]: m k+1 =m k exp( 1 n n ∑ i=1 log(m − 1 k T i )). (2.5) where T i is the Procrustes transformation for the i th subject, k is the estimation step, and n is the total number of subjects. Using v i =log(m − 1 T i ), the mean is subtracted from the individual pose, which results in a residual pose for each subject. The residual pose belongs to an Euclidean space where norms denote the Euclidean distances between the group elements to the mean [27]. Subsequent univariate or multivariate statistics 2.1.4 can be computed on the residual pose which consists of 7 parameters: 1 scale scalar, 3 2.3 The T1 and DTI Fusion Analysis 22 rotation scalars and 3 translation scalars. 2.3 The T1 and DTI Fusion Analysis 2.3.1 Diffusion Tensor Imaging As a relatively new addition to structural MRI, diffusion tensor imaging (DTI) has been promising in characterizing brain white matter microstructure. In human brains, water molecule diffusion is not completely free due to the present of various tissue composi- tions, such as membranes, macromolecules and myelinated axons. Tissues with different levels of anisotropy may restrict the water molecule diffusion process differently, and DTI utilizes these different diffusion patterns to generate contrast. In the pulsed gradient spin echo sequence introduced by Stejska and Tanner [238, 239], the displacement of wa- ter molecules due to diffusion will lead to phase incoherence, which ultimately results in a signal loss. At a given voxel in 3D space, signal attenuation due to diffusion is denoted as [239]: S=S 0 exp − bD (2.6) with b=γ 2 G 2 δ 2 (∆− δ 3 ) (2.7) where γ is the gyromagnetic ratio for hydrogen nuclei (42 MHZ/Tesla), δ is the pulse duration, G is the diffusion gradient amplitude. (∆− δ 3 ) is the effective diffusion time, with ∆ being the separation between pulses and − δ 3 accounting for the spin diffusing 2.3 The T1 and DTI Fusion Analysis 23 while the gradients are turned on. Basser et al. modeled water molecule diffusions in the 3 dimensional space as a rank- 2 tensor D, also referred to as the diffusion tensor [17, 19, 194], which is represented as a symmetric positive semi-definite matrix with 6 degrees of freedom. Thus from Eq. 2.6, we have: ln S S 0 =− b      G 2 x G x G y G x G z G x G y G 2 y G y G z G x G z G y G z G 2 z           D xx D xy D xz D xy D yy D yz D xz D yz D zz      (2.8) To find D , a straightforward way is to linearize the system, and expand Eq. 2.8 as:        ln S 1 S 0 ln S 1 S 0 . . . ln S N S 0        =− b                G 2 1x G 1x G 1y G 1x G 1z G 1x G 1y G 2 1y G 1y G 1z G 1x G 1z G 1y G 1z G 2 1z G 2 2x G 2x G 2y G 2x G 2z G 2x G 2y G 2 2y G 2y G 2z G 2x G 2z G 2y G 2z G 2 2z . . . . . . . . . . . . . . . . . . G 2 Nx G Nx G Ny G Nx G Nz G Nx G Ny G 2 Ny G Ny G Nz G Nx G Nz G Ny G Nz G 2 Nz                            D xx D xy D xz D yy D yz D zz             (2.9) When the number of gradient directions N = 6, D has a unique solution. In clinical practice,N is commonly chosen⩾ 25 to compensate for the noise inS, in which cases,D can be efficiently estimated by least squares methods [33, 140]. To gain more geometric insights into the diffusion patterns,D is typically further factorized as: D=EΛE T (2.10) 2.3 The T1 and DTI Fusion Analysis 24 Fig. 2.2 Example maps of diffusion tensor derived metrics. Top row from left to right: λ 1 ,λ 2 , andλ 3 maps. Bottom row from left to right: color FA maps, FA maps, MD maps. All the maps are selected of the same subject, from the same axial slide (same location within the brain). where E represent orthogonal eigenvectors e i , i= 1,2,3, andΛ is a diagonal matrix of eigenvalues λ i , i= 1,2,3. Thus, a diffusion tensor can be geometrically interpreted as an ellipsoid at the corresponding location such that λ i and e i represent the extent and direction of diffusivity at that location. Based on the diffusion tensor model, a series of diffusion tensor derived metrics have been introduced into the evaluation of white matter integrity within the brain, among which the most used ones are the fractional anisotropy (FA) and mean diffusivity (MD): FA= s 3((λ 1 − λ) 2 +(λ 2 − λ) 2 +(λ 3 − λ) 2 ) 2(λ 2 1 +λ 2 2 +λ 2 3 ) (2.11) 2.3 The T1 and DTI Fusion Analysis 25 MD= λ 1 +λ 2 +λ 3 3 (2.12) 2.3.2 The significance of joint T1 and DTI analysis (a) Neural connectivity in intact brain revealed by light microscopy. Adapted from [134]. (b) Major white matter tracts. Adapted from [91]. Fig. 2.3 subcortical connectivity The human brain is a complex network exhibiting trillion of neural synaptical connec- tions (Fig.2.3a), and accommodating countless functioning processes at each second. At the mesoscale, the complexity of the brain is largely reflected by its intriguing white mat- ter pathways (Fig.2.3b). The corpus callosum (CC) is the largest white matter structure in the brain, and has its most densely packed bundle situated in the subcortical region. The midline location and its complex interconnections to a broad cortical domains make the CC an easy target for diseases and external disturbances. For instance, the vulner- ability of the CC to various of disorders has been detected in subjects with a huge age range, including prematurity in neonates [100, 185], traumatic brain injury (TBI) in chil- dren and adults [36, 78, 157, 269], mild cognitive impairment (MCI) in aging population [62, 108, 246, 277, 278], and many other disorders. 2.3 The T1 and DTI Fusion Analysis 26 Fig. 2.4 Overlaid T1 segmentations (A) and FA maps in the same subject scanned pre- season (B) versus post-season (C). In figure A, pre- and post-season CC segmentations are represented in yellow and blue respectively, with the overlaid areas appearing in orange. Comparing the three figures, circled area shows both structural and diffusion changes, the area indicated by a triangle is structurally different but has an unchanged diffusion pattern, and vice versa for the area shown by the arrow [148]. To characterize CC alterations, structural MRI is a typical choice and has been effec- tive in deciphering brain parenchyma loss [217, 245, 279], while DTI has been promising in characterizing white matter microstructure alterations [125, 246, 252, 277, 278, 281]. For white matter structures, parenchyma and diffuse injuries often occur concomitantly, thus morphological and diffusional changes may complementary to each other in de- ciphering brain alterations. As illustrated in Fig. 2.4, where displays both structural and diffusional data for one young athlete, scanned pre- and post-season, shape and FA are each shown to provide important information on pre- vs. post-season anatomical changes. Therefore, a joint analysis of diffusion and T1-weighted data may therefore pro- vide a more complete picture of CC changes due to disturbances or diseases. However, the vast majority of studies have been studying each aspect on their own [252, 277, 278], or by comparing them side-by-side [62, 246]. None have tried to truly combine these two features into one analysis. 2.3 The T1 and DTI Fusion Analysis 27 To fuse the shape and the diffusion aspects into one analysis, a model needs to be defined to carry both features. Group differences of the brain white matter, including CC, are frequently analyzed based on voxels, midlines, mid-planes, or an averaged index as a whole (Fig. 2.5). However, voxel-based methods give poor localization of differ- ences in anatomical regions compared to surface-based ones and may be contaminated by differently oriented tracts [187], while midline- or mid-plane based methods rely on assumptions that WM perpendicular to the mid-line or the mid-plane is uniformly dis- tributed. As shown in mTBM described in section 2.1.3, 3D representations may better localize injury, and may have higher statistical detection power to identify the neuro- circuits alterations underling the observed anatomical alterations. This is of particular importance for structure like the CC, as it is not a homogenous structure, in terms of fiber composition [3] as well as topographical distribution [191]. Enlightened by mTBM, a fused T1 and DTI on 3D structural models is proposed, and will be discussed in the next section. 2.3.3 The joint T1 and DTI analysis To conduct a joint T1 and DTI analysis, shape and diffusion features are sampled sepa- rately onto the surfaces, and then combined into a single vector for subsequent multivari- ate statistics. The sampling of shape features can be found in section 2.1.3. To project diffusion indices of each of the surface models onto its surface, center lines of all the 3D models are first calculated. Points in the center line are calculated as the center of each horizontal iso-parametric curve on the surface. For each surface vertex, the diffusion parameters are collected along its radius to the corresponding center point, as illustrated 2.3 The T1 and DTI Fusion Analysis 28 Fig. 2.5 Traditional white matter analyses: voxel-wise analysis (A), skeleton based anal- ysis (B), midline based analysis (C), mid-plane analysis (D). in Fig. 2.6. Specifically, for a surface vertex X and its corresponding center point M, a voxelP within the 3D representation can be assigned toX if it conforms to:            ( − −− → X− M)× ( − −− → P− M) ∥ − −− → X− M∥ ∥≤ R. ( − −− → X− P)· ( −−−→ P− M)≥ 0. (2.13) where R represents a pre-defined distance between P to the line of −−−→ X− M, and could be defined according to the intrinsic nature of the white matter structure. The criteria is to make sure each of the vertices has some voxels assigned, and to minimize overlap with neighboring vertices. Having assigned shape and diffusion features to each surface vertex, multivariate 2.3 The T1 and DTI Fusion Analysis 29 Fig. 2.6 Surface of CC and the illustration of sampled voxels. The red line on the right side of the figure is the midline of the CC, and pink crosses are surface vertices. In the direction perpendicular to the midline and pointing to each vertex, blue stars represent voxels projected to each of the vertices, and a mean index (FA, MD....) of projected voxels is assigned to the vertex for later statistics. analysis can be conducted on a series of combined features. For instance, Hotelling’s T 2 test can be performed on a combined univariate morphological measurement and a diffusion index (i.e. (detJ,FA)), or a combination of deformation tensor and diffusion tensor (i.e. (s 1 ,s 2 ,s 3 ,λ 1 ,λ 2 ). Note: λ 3 is usually susceptible to noise and may reduce detection power.) Chapter 3 Shape and Pose Analysis of the Thalamic and the Lentiform Nucleus in Preterm Neonates The work in this chapter has been adapted from the following publication: Yi Lao, Yalin Wang, Jie Shi, Rafael Ceschin, Marvin D. Nelson, Ashok Panigrahy, and Natasha Leporé, Thalamic alterations in preterm neonates and its relation to ventral stria- tum disturbances revealed by a combined shape and pose analysis, Brain Structure and Function, 221(1), 487-506. 3.1 Abstract 31 3.1 Abstract Finding the neuroanatomical correlates of prematurity is vital to understanding which structures are affected, and in designing treatments. Converging results reveal that tha- lamic abnormalities are important indicators of prematurity. However, little is known about the localization of the disturbance within the subnuclei of the thalamus, or on the association of altered thalamic development with other deep gray matter disturbances. Here, using brain structural MRI, we perform a novel combined shape and pose analysis of the thalamus and ventral striatum between 17 preterm and 19 term-born neonates. We detect statistically significant surface deformations and pose changes on the thalamus and putamen, successfully locating the alterations on specific regions such as the ventral and ventral-anterior thalamic nuclei, and for the first time, demonstrating the feasibility of us- ing relative pose parameters as indicators for prematurity in neonates. We also perform a set of correlation analyses between the thalamus and the ventral striatum, based on the surface and pose results. Our methods show that regional abnormalities of the thalamus are associated with alterations of the ventral striatum, possibly due to disturbed develop- ment of shared frontal-subcortical pathways. These findings point to potential anatomical substrates for the poor neurodevelopmental outcomes in the preterm population. 3.2 Introduction Neonates born prematurely are at risk for cognitive, behavioral and social problems in later life [37, 71, 101, 163, 205, 270]. In the last decade, the role of frontal-subcortical circuits have been put forward in mediating executive functions, emotions and motiva- 3.2 Introduction 32 tion, and have been investigated as a unifying framework in the studies of human neu- ropsychiatric disorders [55, 85, 170]. In prematurity studies, the damage to cerebral white matter, especially in the frontal lobes, has been long recognized as a major cause of premature related poor neurological outcomes. Neurological disorders like attentional deficit/hyperactivity disorder (ADHD), autism spectrum disorders (ASD) and learning disabilities, which are common in preterm subjects, have been linked to disturbances of brain functions that can be ultimately attributed to the dorsolateral prefrontal circuit (DLPC), the lateral orbitofrontal circuit (LOFC) and the oculomotor circuit [21, 121, 207, 231, 250]. The ventral striatum (putamen and globus pallidus) and the thalamus are paired deep gray matter structures involved in motor, cognitive, visual and sensory functioning, thus providing a potential seat for anatomical alterations underlying corre- sponding neurological dysfunctions. The vulnerability of these two structures to prema- turity or prematurity associated complications has been investigated in previous works, with reduced volume reported in both structures[2, 24, 25, 113, 130, 131, 180, 193, 220]. However, the investigation of neuroanatomical and functional correlations in the pres- ence of prematurity has been confined almost entirely to the cerebral cortex, while the role of subcortical structures within specific frontal-subcortical circuits is poorly under- stood. In addition, studies to date have not looked at the structural covariance between subcortical structures like the thalamus and the putamen in preterm neonates. As proposed by the concept of ‘encephopathy of prematurity’, the process of fetal brain development is accomplished by a series of complex and linked events, and dis- turbances or injuries in one part of the brain may lead to impaired functioning of other interconnected regions [99, 256]. Animal studies [94, 97] and diffusion tensor tractog- 3.2 Introduction 33 raphy techniques [137] provide descriptions of the topographical connections between subcortical areas and the cortex, and have showed that neuroanatomical subparcellations of deep gray matter structures can be linked to dedicated functional regions on the cor- tex. Given the functional and anatomical connections between the ventral striatum, the thalamus and the frontal lobes, a ‘trans-synaptical’ disturbance on the development of one or more in-circuit compartments may form a neuroanatomical substrate for circuit related neurological dysfunctions. Thus, a circuit wise investigation should not only cap- ture more fully the spectrum of deep gray matter alterations due to prematurity, but also reveal the particular relevance of deep gray matter abnormalities to potential neurological outcomes. As the ventral striatum and the thalamus are two of the subcortical structures most commonly affected by prematurity, analyzing their related vulnerability to prema- turity within the framework of frontal-subcortical circuits is crucial to our understanding of impaired brain development, and for guiding circuit-specific early interventions. Frontal-subcortical circuits mediate different functions via segregated but integrated pathways, with neuron fibers interdigitating specific nuclei of the thalamus and ventral striatum, rather than overlapping them [55]. As a result, to gain more insights on the altered development of frontal-subcortical circuits in preterm infants, anatomical abnor- malities need to be localized at the level of subdivisions of the thalamus and ventral striatum. However, the traditional volumetry method does not provide regional speci- ficity on the subcortical structures, while whole brain deformation-based morphometry [16, 24] findings consist in abnormalities that are not well localized within subcortical structures and, additionally, are prone to registration artifacts - a non-negligible effect in gray matter near the ventricles. This highlights the need for more sensitive methodolo- 3.2 Introduction 34 gies, and for a detailed assessment of the subtle morphometric structural changes caused by prematurity. In our previous work, we extended the traditional univariate surface-based morphom- etry [266] to a multivariate version that has shown greater statistical power to detect population differences in subcortical structures, both in neonates [220, 262] and in adults (e.g. [263, 266]). Complimentary to the surface based shape analysis, the relative pose of subcortical structures may also help to indicate abnormal growth of the brain[149]. The relative pose is the residual position of the subcortical structure (translation, rotation and scale relative to an average shape) after removing irrelevant external information such as the arbitrary location of patients’ head within the scanners and the size of the head. This information is especially important in depicting the development or degeneration pat- terns of the brain, when shifts of pose in different subcortical structures are more likely to happen, secondary to the volume and shape changes of neighboring structures. The relative pose is complementary to subcortical surface shape analyses, and the combined shape and pose results form a complete subcortical morphometry system. However, few attempts have been made to perform surface based statistics on subcortical structures in preterm neonates [220]. Moreover, there are few relative pose analyses in the field of neuroimaging [27, 28, 87], and to the best of our knowledge, none have focussed on neonates. Despite their joint involvement in fronto-subcortical circuits, the linked involvement of the thalamus and ventral striatum in preterm associated injury remains understud- ied. In the handful studies that integrate both deep gray matter structures, Lin et al. [159] reported reduced volume in both lentiform and thalamic nuclei in preterm neonates 3.2 Introduction 35 with periventricular leukomalacia (PVL). Srinivasan et al. [234] confirmed the reduced lentiform and thalamic volume in preterm subjects with nonfocal white matter abnormal- ities. However, these studies are not powerful enough to reveal regional disturbances on the gray matter structures, and none of them have attempted to examine the associated disturbance between different deep gray matter structures. In our recent study inves- tigating surface deformations of the ventral striatum [220], we found altered anterior and inferior part of the putamen in preterm newborns, but whether this involvement of the ventral striatum in prematurity is due to a secondary effect of injury to the thalamo- cortical projections, or to impaired primary putaminal functions remains unknown. Thus, analyzing the associated alterations within these two gray matter structures may allow a refinement of our model for the impaired developmental mechanisms of prematurity. Here, we focus on examining the effect of prematurity on the thalamus and on inves- tigating the linked disturbances in the ventral striatum. Our aim is two-fold: First, using brain structural MRI, we perform a novel pipeline that integrates multivariate surface- based morphometry and relative pose analyses, to fully capture the subtle alteration of deep gray matter development in preterm neonates. Here we focus more specifically on neonates with no visible evidence of white matter injury, to assess the power of our method to detect subtle abnormalities. Fig. 3.1 illustrates our surface generation and pose analysis pipelines. Secondly, we investigate the thalamic changes in relation to those of the ventral striatum using the information resulting from the combined shape and pose analysis. Our method extends knowledge gained from traditional volume based analyses of neonates in the literature, and provides anatomical evidence to the concept of ‘encephalopathy of prematurity’ as proposed by [256]. Our results indicate that sur- 3.3 Subjects and Methods 36 face morphology as well as pose information in subcortical structures may be sensitive indicators of brain injury. 3.3 Subjects and Methods Fig.3.1 illustrates our pipelines of point distribution model (PDM) generation, surface based morphometry and relative pose analysis. 3.3.1 Subjects and preprocessing Neonatal Data Our dataset comprises 17 premature neonates (gestational ages 25-36 weeks, 41.12± 5.08 weeks at scan time) prospectively recruited with normal MR scans and 19 healthy term born infants (gestational ages 37-40 weeks, 45.51± 5.40 weeks at scan time). T1- weighted MRI scans were acquired using a dedicated neonatal head coil on a 1.5T GE scanner using a coronal three-dimensional (3D) spoiled gradient echo (SPGR) sequence (TE =6 ms; TR= 25 ms, FOV= 18 cm; Matrix =256× 160), axial and sagittal T1-weighted FLAIR sequences (TE =7.4, TR= 2100; TI = 750; FOV= 20 cm; Matrix = 256× 160). The scans were optimized for the best grey and white matter contrast, both at the cortical and at the subcortical/deep grey level. The inclusion criteria for our preterm subjects were the following: 1) prematurity (less than 37 gestational weeks at birth), and 2) visually normal scans on conventional MR imaging. In terms of this particular cohort of preterm, there were no significant pregnancy complication (i.e. chorioamonitis) or postnatal complica- tions (NEC and sepsis) and there was no visible injury. Structural MRI were qualitatively 3.3 Subjects and Methods 37 Fig. 3.1 Diagram of the combined shape and relative pose analysis. For simplicity, only the surfaces of thalamus are used for illustration. Top Row: the subcortical structures are first segmented from T1 images, and then reconstructed to 3D surface models; Middle Row: surface based morphometry and correlation analysis; Bottom Row: relative pose based statistics and correlation analysis. In the pipeline, letter a-g represent: a: T1- MR images; b: subcortical structure segmentation; c: 3D surface models; d: one-to-one correspondence obtained from registration; e: surface-based group differences results; f: surface based correlation results; g: pose parameters obtained from procrustes alignment; h: group differences; i: pose-based correlation results. 3.3 Subjects and Methods 38 classified as controls by 2 board certified neonatal neuroradiologists. Preterm subjects were excluded based on chromosomal diseases or when their neurological examination revealed major abnormalities, or if they exhibited brain lesions including: (1) focal white matter necrosis as definite as cavitary/non-cavitary lesions (2) diffuse ventriculomegaly (3) significantly increased sub- arachnoid space and sulcal enlargement (4) diffuse excessive T2 hyperintensity by visual criteria. The institutional review board at our medical center approved the study protocol. 3D surface representations and registration All the T1-weighted MRI scans were first registered to a same template space through linear registration. Alignment quality was validated by superimposing images from dif- ferent subjects on top of each other. Irrelevant global pose difference induced by differ- ent locations and orientations during scans was factored out in this step. The thalami and putamen were then manually traced on linear registered T1 images by an experienced pediatric neuroradiologist using Insight Toolkit’s SNAP program [276], and the details of the segmentation can be found in Fig. 3.2 The intra-rater percentage overlap were 0.93 and 0.86 for measuring the thalamus and the putamen, respectively, in four participants at two subsequent times (two preterm and two term born participants) at two subsequent times spanning several months. 3D surface representations of the segmented structures were constructed and mesh grids were built on the surfaces using our in-house conformal mapping program [258]. We included the putamen and globus pallidus in our ventral striatal surface reconstruc- tion, due to the lack of tissue contrast on neonatal T1 images between these two struc- 3.3 Subjects and Methods 39 tures. Given the large variability of brain morphology in neonates, the registration should allow for large deformations. Here, we non-linearly registered all subjects using a new surface fluid registration algorithm [Shi et al.], which outperforms the more commonly used constrained harmonic registration, as compared in our previous work [220]. For the fluid registration, we used the subcortical surfaces we generated from one of the healthy controls. The one-to-one correspondence achieved between vertices allows us to accurately analyze localized information on the surfaces of subcortical structures. In the following methodology sections, we will use surface based morphometry and relative pose analyses to investigate the group-wise differences on these subcortical structures. Surface- and pose- based correlation analyses will also be used to test the association between thalamic and ventral striatum abnormalities. 3.3.2 Surface based morphometry analysis We perform a vertex-vise multivariate statistical analysis, including: 1. The distance ρ from a medial axis to a vertex on the surface (MAD) - which represents the thickness of the shape at each vertex [195, 247]. More precisely, the medial axis is computed using the center point on the iso-parametric curves, i.e. the iso-parametric curve is perpendicular to the medial axis, on the computed conformal grid [263], after whichρ is easily found at each vertex. 2. The logged deformation tensor (log √ JJ T , where J is the Jacobian matrix) [261], as in our prior multivariate tensor-based morphometry (mTBM) analyses [154, 261, 266], - which describes the directional differences in regional surface area between each subject and the template. 3.3 Subjects and Methods 40 Fig. 3.2 Thalamus and putamen segmentation. The rules of manual segmentation were guided by anatomical references of the human striatum (Prensa et al., 2003; Morel et al., 2002) and thalamus (Niemann et al., 2000; Wiegell et al., 2003; Erbetta et al., 2009). Due to the lack of tissue contrast on neonatal volumetric T1 images, it is difficult to accurately separate the globus pallidus from the putamen. Since our hypothesis is related to the ventral striatum, we included the globus pallidus in our putamen segmentation and 3D surface construction. For simplicity, we use the word ’putamen’ to represent the combined surface of the globus pallidus and putamen throughout the text. 3.3 Subjects and Methods 41 MTBM is complementary to MAD, as the radial distance primarily measures changes in thickness while the deformation tensors mainly capture changes in surface area, so we combine them in one multivariate 4x1 vector for analysis. Simpler univariate statistical analyses of ρ and of the determinant of the deformation tensor (detJ, the difference in surface area without directional information), as well as multivariate statistics on the de- formation tensors alone are also conducted for comparison and validation. As with our previous studies, here deformation tensors are expected to outperform univariate mea- sures in terms of detection power [263]. Multivariate Statistical Analysis Although the mean and distribution of age was roughly matched between preterm and term groups, subjects were scanned over an age range of 36-57 post-conception weeks. This variation of age is not negligible, especially for neonates whose brains change rapidly with age. The growth of the thalamus in neonates in terms of the volumetry and outgoing trajectory is approximately linear [16], thus we used linear regression to factor out the influence of age. While nonlinearity may present in brain development over years (Dosenbach, 2010), we are not aware of any references to nonlinear devel- opment in the first few weeks of life. As seen in Fig. 3.3, the data may be analyzed using linear approximations. Due to the limit number of subjects in our study, nonlin- ear approximations may not increase the accuracy of analysis, and will be subject to the over-fitting problem. In the future, when applying our methods to larger dataset, we will re-evaluate the influence of possible nonlinearities. Subsequent statistical analyses are performed on age-covaried data. 3.3 Subjects and Methods 42 T -tests are used for the univariate measures. For either of the multivariate tests (mTBM or MAD + mTBM), group statistics are computed using the Hotelling’s T 2 test [107] - the multivariate extension of the Student’st-test, as described in [154, 263, 266]. We run two permutation tests on the images: a vertex-based one that allows us not to assume a normal distribution, and one over the whole segmented image to correct for multiple comparisons [154, 183, 266]. The procedure is repeated 10,000 times. For the first permutation test, we obtain a null distribution of T 2 -values at each vertex to which we compare the T 2 -values from the real data. For the second one, we compute a single value over the whole image for each of the 10,000 permutations. The value we choose isN i p = number of p-values< 0.05 over all vertices in the image for permutationi. This provides us with a null distribution to which we compareN true p from the true labels. 3.3.3 Relative pose analysis Similarity transformation We compute the relative pose for each of the four structures (left thalamus (LTha ), right thalamus (RTha), left putamen (LPuta), and right putamen (RPuta)). This is done by full Procrustes alignment of each subject’s surface to a template shape - that is by rotating, translating and uniformly scaling the subject’s shape to align it with the template [68]. The Procrustes transformation can be written as [27, 28]: T =    sRX d 0 T 1    (3.1) 3.3 Subjects and Methods 43 where s is the scalar scaling factor, R is a 3× 3 rotation matrix, and d is the translation vector (x,y,z) T . The template is selected as the mean shape that minimizes Procrustes distances, and is computed iteratively [211]. All transformations are centered on the center of mass of the template. To simplify the computation of statistics, these trans- formations are then projected onto the tangent plane at the origin of the manifold of transformations, in which simple flat space calculations can be performed [12]. This is done using matrix logarithms, as in [28]. To define a center or representative object of the population under study (Bossa et al., 2011), we chose the Frechet mean as our mean pose. The mean pose m for the preterm and control groups can be calculated iteratively as [27, 126, 192]: m k+1 =m k exp( 1 n n ∑ i=1 log(m − 1 k T i )). (3.2) where T i is the Procrustes transformation for thei th subject, k is the estimation step, and n is the total number of subjects. After the subtraction of the mean from each subject’s individual pose, specifically using v i =log(m − 1 T i ), each subject is left with a residual pose. Statistics are computed on this residual pose which consists of 7 parameters: 1 scale, 3 rotations and 3 translations. Statistical analysis As was done in the surface based morphometry feature vectors, we used linear regression to factor out the influence of age on pose parameter. The distributions of pose parameters for 4 subcortical structures are shown in Fig. 3.3. To save space, only the data of the 3.3 Subjects and Methods 44 parameter logS are presented. Subsequent statistical analyses were performed on age- covaried data. Statistical comparisons between the two groups are performed via two methods: univariate t-tests for logS,||logR||,||logd||; Multivariate Hotelling’s T 2 -test, which is a multivariate generalization of thet-test, for 3 rotation parameters (θ x ,θ y ,θ z ), 3 translation parameters(x,y,z), as well as a combination of all 7 parameters. Considering the limited size of our dataset (36 subjects), a permutation test [183] is performed to avoid the normal distribution assumption. To do this, we randomly permute the labels of our subjects (preterm vs. term neonates), and generate t-values (for t-test) or F-values(for T 2 -test) for comparison. We use 10,000 permutations for each of the parameters to assemble a null distribution of non-parametric estimation fort- orF-values. 3.3.4 Correlation analysis To test the hypothesis that thalamic abnormality is associated with that of the putamen, two kinds of correlation tests are performed on our previously obtained surface based information and pose parameters. Surface based correlation To investigate the relationship between regional parameters on the thalamus and the puta- men a correlation analysis is performed between the determinant (surface area) or medial axis distance (thickness) of each vertex on the thalamus with the total volume of the two putamen. A similar test is performed on the putamen, covaried with the total volume of the two thalami. Similar to section 2.2, we run two permutation tests on the surface based correlations: 3.3 Subjects and Methods 45 Fig. 3.3 Distribution of the parameters used in this study vs. post-conceptional age (PCA). Columns from left to right are: volumes vs. PCA; (Normalized) mean detJ vs. PCA; scale parameter (with mean pose subtracted) in log-Euclidean space (logS) vs. PCA. Rows from top to bottom represent parameters from: left thalamus, right thalamus, left putamen and right putamen. Scatter points in the figure represent observed results, while lines represent their corresponding linear regressions. In all the figures above, data from the preterm group are marked in red (circles and solid lines), while that from the term group are marked in blue (cross symbols and dash lines). Note that nearly all the parameters for the preterm group are distributed lower in the graphs, compared to that from the term controls, indicating a smaller size of structures in the preterm group. 3.4 Results 46 a vertex-based one that allows us not to assume a normal distribution, and one over the whole segmented image to correct for multiple comparisons (Nichols and Holmes, 2001; Lepore et al., 2008; Wang et al., 2010a). The procedure is repeated 10,000 times. For the first permutation test, we obtain a null distribution of correlation coefficient r-values at each vertex to which we compare the r-values from the real data. For the second one, we compute a single value over the whole image for each of the 10,000 permutations. The value we choose isN p i= number of p-values<0.05 over all vertices in the image for permutation i. Pose based correlation In order to further explore the association between the thalamus and the putamen, in terms of the trend of changes in position, we run a series correlation tests on the pose parameters between the four structures: LPutavs.LTha , RPutavs.LTha , LPutavs. RTha , RPuta vs. RTha. The pose parameters that we test include: logS, ||logR||, ||logd||, representing the scale, total rotation, and total translation of each structure. 3.4 Results 3.4.1 Surface based morphometry results Fig.3.4 displays areas of significant abnormality on the thalamus when comparing the preterm vs. term groups. Broad areas of significance are seen in the anterior and ventro- anterior parts of the two thalami, while the main cluster resides in the anterior-dorsal part of the left thalamus. All four methods agree as to the location of clusters of significance, 3.4 Results 47 though the mTBM results are a bit noisier than those for MAD or the determinant of Jacobian (detJ). However, the overall p-value for MAD and detJ are above the threshold for statistical significance - which we set here at p t = 0.05, given our a priori hypothesis on the thalamus, while that for mTBM is much lower than p t . Both the FDR and per- mutation tests show that the combined analysis is the most powerful one. This is in line with our previous work with the putamen[220], in which significant regional differences are detected more extensively using multivariate measures. In order to see the direction of the changes, we also map R k detJ = 1 N p Σ N p i detJ k p i 1 N c Σ N c j detJ k c j (3.3) and R k ρ = 1 N p Σ N p i ρ k p i 1 N c Σ N c j ρ k c j (3.4) at each voxelk (see Fig. 3.5), whereJ k p i ,J k c i ,ρ k p i ,ρ k p i are the Jacobian matrices and radial distances for the preterm or term-born subjecti, respectively. N p andN c are the numbers of preterm and term-born subjects. The determinant of the Jacobian matrix indicates the difference in surface area of the region in the individual subject compared to the template, while the radius measures change along the surface normal direction (thickness). In both ratio maps, values smaller than 1 indicate trophic changes at a given vertex in preterm group, and vice-versa for values greater than 1. As expected, the majority of the surface area shows an atrophic tendency in the preterm group. The anterior side of both thalami is smaller in the preterm group, although the area does not reach significance on the right thalamus. Furthermore, the pulvinar is smaller on average in our preterm subjects 3.4 Results 48 Fig. 3.4 Thalamus statistical p-maps: Overall p-values are p=0.0035 for MAD + mTBM(a), p=0.0901 for MAD(b) and p=0.0683 for detJ(c). Fig. 3.5 Vertex-wise correlation determinant (a) and radial distance (b) maps of the thala- mus. Widespread shrinkage of preterm group is present in both thalami, and the clusters with significant preterm vs. term differences (seen in Fig. 3.4) are all falling on the preterm< term areas. compared to the controls, a region that has previously been implicated in prematurity related thalamic injury (see e.g. [180]). To get a more complete view of the ventral striatum, we also show the determinant and radius map for the putamen. In Fig. 3.6, we can also see a widespread atrophy in both of the putamen, with a more uniform distribution as compared to that of the thalamus. These results are consistent with neuroanatomy findings that the thalamus is more anatomically sub-dividable than other gray matter structures [256]. 3.4 Results 49 Fig. 3.6 Vertex-wise correlation determinant (a) and radial distance (b) maps of puta- men. Widespread shrinkage of preterm group is present in both putamen, and the clusters with significant preterm vs. term differences (seen in our previous publication [220]) are mostly falling on the preterm< term areas. 3.4.2 Relative pose analysis results All the p-values from previously described tests are presented in Table. 3.1. For the left thalamus, pose parameters representing scale and rotation show a significant differ- ence between the preterm and term groups, while no difference can be seen in translation parameters. It is also important to note that, apart from the difference detected in indi- vidual parameters, a combination of all 7 parameters also detects significant differences between the two groups in the left thalamus, indicating a possibility of using multivariate analysis of all pose parameters as the discriminant between the two populations. For the right thalamus, neither the individual nor combination of parameters detects any changes. For the left putamen, low p-values are seen in univariate scale and translation, as well as multivariate rotation parameters, but none of them reaches significance. For the right putamen, significant group difference are only seen in the scale parameter. These results are better visualized in Fig. 3.7, where mean shapes of preterm (repre- 3.4 Results 50 sented in red) vs. term (represented in blue) groups are overlaid in their corresponding mean pose. Areas where the mean shapes of two groups are overlaid appear in purple. Note the borders of these two structures: a shift in pose is evident on the left thalamus and putamen, while less visible variants appear on in right hemispheric structures. The left thalamus of the preterm group shows a smaller size, with the posterior part positioned more medially compared to the control group, which are consistent with the differences found in scale and rotation parameters. Compared to the obvious differences in size and shape, the shift of position of the left thalamus in these two groups are less evident, con- sistent with the results from our statistical tests. As for the left putamen, we can see a laterally shift in position along with a moderate difference in size, as is confirmed by the low p-values in all three sets of pose parameters. In the right hemisphere, the pose dif- ferences in both structures are less evident: mean shapes for the right thalamus from the two groups are mostly overlapped, in agreement with the relatively high p-value found in all pose parameters, while only a shrinkage in size is evident in the right putamen, which is validated by the significance in its scale parameter. 3.4.3 Correlation results Surface based correlation In Fig. 3.8 and 3.9, several subdivisions of the thalamus are significantly correlated with the volume of the putamen, such as the ventral anterior nucleus, medial dorsal nu- cleus and pulvinar. The same regions have been reported to be interconnected with the putamen and pallidum (also included in our putamen model) by a probabilistic tractogra- 3.4 Results 51 Fig. 3.7 3D visualization of the pose of mean shapes averaged from the preterm (red) and term groups (blue). Areas where the mean shapes of two groups overlaid appear in purple. To better visualize the pose changes, enlargements of locations (a), (b), (c) in the bottom right figure are presented in the top right, bottom left, and bottom right figures, respectively. Note the borders of these two structures: shifts of pose are evident on the left putamen (a), left thalamus (b), and right putamen (c), but are quite small in the right thalamus. 3.4 Results 52 Table 3.1 P-value of statistical analyses on pose parameters: 6 sets of parameters char- acterizing relative pose of left thalamus (LTha), right thalamus (RTha), left putamen (LPuta), and right putamen (RPuta) are investigated here using univariate and multivari- ate analyses. Parameters are categorized as logS, ||logR||, ||logd|| for univariate tests, and as (θ x ,θ y ,θ z ), (x, y, z), and a combination of 7 parameters for multivariate tests. All the p-values are obtained after permutation testing. Significant p-values (p< 0.05) are highlighted in cyan, while p-values that are interestingly low but failed to reach signifi- cance are highlighted in yellow. Pose Parameters LTha RTha LPuta RPuta logS 2.36e-02 3.37e-01 7.50e-02 3.28e-02 ||logR|| 7.80e-01 5.22e-01 1.92e-01 9.75e-01 ||logd|| 9.01e-01 3.66e-01 9.67e-02 4.41e-01 (θ x ,θ y ,θ z ) 1.06e-02 7.16e-01 6.71e-02 3.61e-01 (x,y,z) 8.26e-01 8.15e-01 8.71e-01 3.14e-01 All7parameters 2.16e-02 9.12e-01 1.90e-01 3.49e-01 phy study[65]. However, these regions have different levels of shrinkage as shown in the group average determinant and radius maps, but only the ventral anterior nucleus reach a significance in the group difference map, as shown in Fig. 3.4. As for the putamen, broad anterior and inferior areas, especially on the anterior puta- men, the globus pallidus internal (Gpi), and the globus pallidus external (Gpe), are sig- nificantly associated with the volume of the thalamus. These are also in line with the previously mentioned theoretical study [65], in which anterior putamen and pallidum, along with certain nuclei of the thalamus, are demonstrated to be concomitantly involved in several cortical-striatal-thalamo-cortical loops. Furthermore, these regions are largely overlapped with the significant group differenced areas, as previously found by our group [220]. This suggests that the vulnerability of the putamen to prematurity is possibly a secondary effect of the thalamic injury. 3.4 Results 53 Fig. 3.8 Vertex-wise correlation coefficient maps have been generated based on deter- minant (left) and radial distance (right), respectively. The upper row are displayed in a superior view, while the bottom ones are seen from the inferior one. 3.4 Results 54 Fig. 3.9 P-maps of the corresponding to the correlation coefficients, derived from the determinant map (left), and radial distance maps (right). The upper row are displayed in a superior view, while the bottom are seen from the inferior one. In (A), overall p-values are p=0.0443 for the correlation between vertex-wise determinants on thalamus surfaces and putamen volumes, p=0.0745 for the correlation between vertex-wise determinant on putamen surfaces and thalamus volumes. In (B), overall p-values are p=0.0624 for the correlation between vertex-wise radial distance on thalamus surfaces and putamen volumes, p=0.0285 for the correlation between vertex-wise radial distance on putamen surfaces and thalamus volume. 3.4 Results 55 Fig. 3.10 The correlation between the thalamus and the putamen in the left (upper row) and right (bottom row) hemispheres, tested using pose parameters: logS(left column), ||logR||(middle column),||logd||(right column). More details, such as correlation coeffi- cients and p-values, are presented in Table. 3.2. Pose based correlation As shown in Fig. 3.10, there are significant correlations( p< 0.01) between the left thala- mus and the left putamen in terms ofLogS and||logd||. A similar association exists in the right hemisphere, but to a lesser extent. Detailed correlation coefficients and their cor- responding p values are presented in Table3.2. All p-values are corrected using 10,000 permutations (see Sect. 3.3.3. Significant p-values (p< 0.01) are highlighted in pink. 3.5 Discussion 56 Table 3.2 4 sets of correlation results based on pose parameters: logS,||logR||,||logd||, between 2 thalami and 2 putamen. Both coefficients and p values are provided in each of the test, with significant correlations highlighted in babypink. Here, we set the signifi- cance with p< 0.01. Note: here we investigated RPuta vs. LTha, and LPuta vs. RTha as an exploratory analysis. Although RPuta vs. LTha, and LPuta vs. RTha are not directly connected to each other, they all have different levels of projections to the cortex, and interhemispheric callosal fibers may indirectly link them together (Gazzaniga, 2000). logS ||logR|| ||logd|| LPuta vs. LTha β=5.45, p=5.00e-04 β=3.94, p=1.55e-02 β=4.61, p=4.10e-03 RPuta vs. LTha β=6.13, p<1.00e-04 β=2.79, p=1.01e-01 β=3.58, p=3.50e-02 LPuta vs. RTha β=3.05, p=7.59e-02 β=4.19, p=1.25e-02 β=4.61, p=3.30e-03 RPuta vs. RTha β=3.61, p=3.12e-02 β=4.52, p=9.90e-03 β=3.74, p=2.59e-02 3.5 Discussion We applied a novel shape and pose combined analysis on the paired deep gray matter, the putamen (including the globus pallidus) and the thalamus, to compare the surface and relative pose changes in premature neonates to term born controls. Using surface based morphometry analysis, we detected widespread putaminal areas of significance group differences[220], as well as focal and regional involvement of anterior and ventral thalamic regions. All measures gave consistent results, and as expected, the combined (MAD-mTBM) statistic showed the highest detection power. Using relative pose anal- ysis, we also detect significant positional differences in the left thalamus and the right putamen, as well as trends in the left putamen. In addition, our data showed a significant structural covariance between thalamus and putamen, with regional involvement of the ventral anterior, ventral posterior, dorsal medial and posterior of the thalamus, as well as anterior and inferior areas of the putamen. The same regions have been demonstrated to have neuronal fibers interconnecting these two structures [65]. Moreover, we tested the 3.5 Discussion 57 correlation of the pose parameters between the thalamus and the putamen, and found a strong association of the pose changes between these two structures. Our findings further validate an intimate correlation of these two structures in the presence of prematurity, possibly underlying the disturbances of preterm delivery on the cortico-striato-pallido- thalamo-cortical circuits that involve both the ventral striatum and the thalamus. 3.5.1 Sensitivity of the our methodology Finding neuroanatomical indicators for abnormal development is necessary for timely therapeutic intervention, but is poorly achieved by current clinical measurements. Prior to this study, thalamic and putaminal volume reductions in preterm infants were reported by several volumetry and DBM studies [2, 24, 25, 113, 130, 131, 180]. While the majority of above prematurity studies focus on populations with PVL, poor neurodevelopmental outcomes are persistently presented in premature populations without focal brain injuries at birth [37, 95, 101]. In the few studies that zoom in on brain anatomic alterations in preterm populations without visible white matter injuries, Allin et al. [8] performed voxel based analysis on a cohort of adults with very low birth weight and found a significant correlation between increased lateral ventricle and decreased gray matter, including cau- date and thalamus, while no significant between-group difference in the whole brain grey matter was detected. Srinivasan et al. [234] reported reduced volume of the thalamus and lentiform nucleus in preterm infants, and the reduction persists, but to a lesser extent, when participants with supratentorial lesions were excluded. These findings indicate that neuroanatomical alterations caused by the prematurity itself, while less extensive than associated perinatal brain injuries, are greater than previously assumed. 3.5 Discussion 58 In this work, we excluded subjects with visible gray or white matter injuries and focused on analyzing the effects of prematurity itself. The vulnerability of the thalamus is confirmed by both the surface based statistic and relative pose analysis, and to be more specific, with the main surface differences seen in the anterior, ventral anterior(V A) nucleus of the thalamus (Fig. 3.4) and pose shift co-localized on the anterior/posterior poles (Fig. 3.7). In addition, significantly abnormalities of the anterior and inferior surface of the putamen are described in our previous publication[220], while significant pose differences on both putamen are first reported here. These findings provided new anatomical evidence for the importance of early detection of prematurity associated brain abnormalities and early clinical intervention. The relevance of detected abnormalities in this study to their underlying functional parcellation will be described in details in the next section. 3.5.2 Selective vulnerability of preterm birth Cortico-striatal-pallido-thalamo-cortical circuits Although the joint contribution of the thalamus and the putamen to altered brain develop- ment has not been described in previous prematurity studies, certain of their association features were proposed in [256] and connectivity based studies in health adults[65]. Here, the close relationship between the thalamus and putamen related to preterm birth is val- idated by the correlation on pose parameters (Table. 3.1), where scale and translation of the left putamen are significantly correlated with that of the left thalamus, and right tha- lamus and putamen are significantly associated through rotation. Additionally, as shown 3.5 Discussion 59 in our surface based correlation analysis, several regions on the surface of the thalamus are significantly correlated with reduced putaminal volume, such as the ventral anterior (V A), inferior medial to posterior area (potentially embedded centrum medianum(CM) and parafascicularis(Pf)), as well as pulvinar. For the putamen, clusters that are signifi- cantly correlated with thalamic volume are primarily located in the anterior and inferior regions of the putamen surface. The regions mentioned above have all been previously demonstrated to have interconnections between the thalamus and putamen [65, 172], and can be co-localized to several major frontal-subcortical circuits, including: the dorso- lateral prefrontal-subcortical circuit (DLPC), the lateral orbitofrontal-subcortical circuit (LOFC), the motor circuit, and the oculomotor circuit. All of the above have the partici- pation of both the ventral striatum and the thalamus. The DLPC and LOFC originate from the prefrontal cortex, outflow via the basal gan- glia, thalamus, and project back to the circuits’ origins. These closed loop connections can be further divided into ‘direct’ and ‘indirect’ pathways. The direct pathway relays inhibitory GABA fibers from the caudate nucleus, and output via the globus pallidus interna (Gpi) to the ventral anterior (V A) thalamus, while the indirect pathway alterna- tively unites the globus pallidus externa (Gpe), the subthalamic nucleus (STN), and the Gpi, before the projections return to the medial-dorsal (MD) thalamus [55, 170]. From our correlation results (shown in Fig. 3.8, 3.9), both Gpi and Gpe of the putamen are significantly associated with thalamic volume, and at the same time, V A of the thalamus presents significant correlations with putaminal volume. These are indirectly validating the existence of the DLPC and LOFC. Moreover, significant group differences between our preterm and term neonates are seen in the V A of the thalamus, as well as in the Gpi 3.5 Discussion 60 and Gpe of the putamen, indicating that the DLPC and LOFC are at risk in preterm birth. The DLPC has been identified as promoting executive function by substantial ani- mal and fMRI studies, while the LOFC is thought to be associated with social restraint and empathy [7, 55, 170]. Emerging studies have shown that prematurity is a risk factor leading to attentional deficit hyperactivity disorder (ADHD) [23, 205], autism spectrum disorders (ASD) [32, 158], and global cognitive abnormalities [37, 99]. These symptoms are usually considered as frontal lobe syndromes, and have been ultimately attributed to DLPC and LOFC dysfunctions. More specifically, ADHD is viewed as a disorder of cognitive and motivational regulation, and arises from an imbalance of the dual pathways system in- volving ventromedial prefrontal cortex and orbital frontal cortex [21, 231, 250]. ASD was demonstrated to link with impaired executive function[121, 207], reduced anterior thalamic radiation [44] and frontal cortical dysplasias[40]. The significantly shrinked and highly correlated Gpi, Gpe of the ventral striatum and V A of the thalamus, as detected by our study, provide additional anatomical support to the involvement of the DLPC, LOFC in preterm birth. Our results further suggest that the altered development of these frontal-subcortical circuits may form the neuroanatomical basis for poor cognitive and socialization outcomes often seen in adolescent born preterm. The motor and oculomotor circuits are two major frontal-subcortical circuits sub- serving motor functions, and united to the same subcortical members as DLPC/LOFC. In these two circuits, projections from the frontal cortex are topographically connected with the dorsal lateral putamen, Gpi, Gpe, V A and MD of the thalamus. As detected by our study, all the subcortical subparcellations mentioned above, except for the MD part 3.5 Discussion 61 of the thalamus, showed significant putaminal-thalamic correlations as well as significant group differences, making motor circuit and oculomotor circuit two putative locations for prematurity associated disturbances. The vulnerability of motor circuit and oculomotor circuit are also implied by clinical and research observations: faulty motor excitability and visuomotor coordination were observed in subject born very preterm[73]; delayed smooth pursuit movements were detected in preterm infants at 2 and 4 months corrected age [241]. These are all hallmarks of motor and oculomotor circuits disturbances. Our results add to this part of literature by providing in-vivo anatomical support to the in- volvement of motor and oculomotor circuits in preterm birth. Nevertheless, more direct connectivity research is needed in the future to substantiate our findings. In addition, it is important to note that MD of the thalamus, a major nucleus that has been previously demonstrated to be involved in all the frontal-subcortical circuits [55, 170], did not present much correlation with the ventral striatum volume from our results. This is not unexpected, as the MD is the major relay hub of the thalamus and is embedded in numerous cortical-subcortical circuits, and therefore, reciprocally con- nected to additional subcortical members and the cortex. The co-existence and interac- tion of different neuro-circuits make it difficult to segregate one from another [40, 84]. When considering the net effect of the potential hypoconnectivity from one circuit and wrongly wired neurons migrated from another, an increase in the latter would tend to balance a reduction in the former, thus hindering the detection of the correlation between MD thalamus and any specific members in the circuits, including putamen. Further sup- plementary studies, for instance a direct tractography study seeding on the putamen and MD of the thalamus would be a validation on the hypotheses derived here. 3.5 Discussion 62 Other circuits harboring the thalamus In Fig. 3.4, the preterm group presents an significant atrophy within the anterior part (AM, AD and V A) of the left thalamus, while only the V A is shown to be associated with putaminal disturbances, as described above. This suggests an influence of prematurity on some additional neuro-circuits encompassing the anterior nuclei of the thalamus, but without the participation of the ventral striatum. Unlike other association thalamic nu- clei (i.e. MD and pulvinar) that were extensively investigated in relation to psychiatric disorders based on animal models and MRI techniques, the role of the anterior thalamus is less described. Besides reciprocal connections with the cortex, the anterior nuclei of the thalamus also direct inputs from the mammillary bodies and the hippocampal formation into differ- ent parts of the cingulate cortex, and form part of the Delay-Brion circuit [172]. Aggleton and Brown [5], Aggleton et al. [6] proposed a hippocampal-anterior thalamic circuit me- diating recall or recollection, and Van Der Werf et al. [253] further confirmed the role of anterior thalamus in memory processing, from which amnesia following anteriorly lo- cated thalamic lesions resembles the symptoms seen in hippocampal lesions. Therefore, altered development of hippocampal-anterior thalamic circuit may be a putative basis for the anteriorly located thalamic atrophy shown in our study, and provides an explanation to the impaired memory and learning performance of ex-preterm subjects, as seen in their later life [113, 115, 193]. Our study provide new anatomical support to the hypothesis that there are additional open circuit elements, other than the four previously mentioned closed loop frontal-subcortical circuits, that are at risk in preterm birth. Again, further validation is required in the future, through including the hippocampus and cingulate 3.5 Discussion 63 cortex in the correlation analysis. Hemispheric asymmetrical vulnerability to preterm birth As detected by our surface based morphometry analysis, significant prematurity associ- ated changes are more prominent on the left thalamus, while fewer clusters with group differences are found on the right thalamus. Complementary to these results, the pose information we found in our study further confirms that the left thalamus and putamen may be more vulnerable to prematurity. Hemispheric asymmetry are well documented, with several higher cognitive func- tions, including language and auditory, lateralized in the cerebral cortex[110]. Several computed tomography based studies analyzing mood disorders revealed a higher fre- quency and severity of depression in patients with left hemisphere lesions, most notably within the dorsolateral prefrontal-subcortical circuit [55, 206, 236, 237]. In a segmenta- tion based MRI study investigating putamen and thalamus in patients with Alzheimer’s disease [59], left putaminal and thalamic volume changes were found to have a higher association with cognitive decline. However, to our knowledge, hemispheric asymmet- rical vulnerability in preterm population has not been reported yet. Our study supports the concept that there are regional thalamic asymmetries in the preterm that may be re- lated to subtle white matter injury, have prognostic significance, or be related to preterm birth itself. Nevertheless, longitudinal follow-up studies with specific neurological tests shall be conducted to confirm the association between asymmetrical abnormalities and lateralized cognitive abilities. 3.5 Discussion 64 3.5.3 Neuro-developmental considerations The process of brain development is a complex sequence, and highly depends on tim- ing. Apart from the general agreement that the earlier the preterm birth took place, the more severe the potential impact it will have on neurodevelopmental outcomes, more dedicated studies revealed a selective pattern of injury on brain structures with varied de- velopmental time windows. Thalamic abnormalities in preterm birth are often attributed to the effect of ischemia and hypoxia, sepsis, or undernutrition on thalamo-cortical sys- tem development[16, 49, 50]. Recently, special attention has also been paid to the linked disturbance of subplate zone. The subplate zone serves as a ‘waiting compartment’ di- recting the in-growth of thalamocortical and basal forebrain fibers to the cortical plate, and injury or disturbances in this transient zone may result in far-reaching developmental effects [256]. The subplate zone starts its formation as early as 13-15 weeks of gestation, and undergoes a major enlargement in the second trimester [178]. After 26/27 weeks of gestation, the subplate zone gradually dissolves, but remains visible in the prefrontal cortex in newborns[142]. The existence of the subplate zone in the frontal cortex can last until sixth postnatal months, and some researchers believe that individual subplate-like neurons may be present until adulthood [143]. The wide developmental time window of the prefrontal cortex provides a neurodevelopmental substrate for its high vulnerability to preterm birth, and this is further confirmed by the prominent involvement of frontal cortex detected in our study. The ingrowing of thalamocortical fibers from the medial dorsal nucleus (MD) of the thalamus into subplate zone starts as early as its formation period [178], and the MD neurons were found to gradually reduced in differentiation after 28 gestational weeks by 3.5 Discussion 65 a transient cholinesterase(ChE) staining study [141]. That is to say, MD is a putative location affected by preterm birth, and the influence might present in a greater extent in very preterm birth (VPB) subjects, those born before 28 weeks of gestation, as the MD neurons still undergo massive differentiation. This is of particular relevance to our study, because MD area of the thalamus in our preterm subjects (average gestational age is 32.8 weeks) is smaller on average, but failed to reach a significant preterm vs. term difference, as elsewhere presented in studies investigating subjects with VPB or different level of brain injuries[180]. In addition, although the group significant clusters limited within the anterior and ventral side of the thalamus, a widespread shrinkage in the preterm group is revealed in the average determinant and radius maps. As shown In Fig. 3.5, reduced volume of preterm group was presented in the majority of the thalamus nuclei, except for the left and right ventro-posterior nuclei. Ventro-posterior nuclei is involved in sensory input, which is considered to establish earlier in the brain development and therefore less vulnerable to preterm birth. Furthermore, by comparing the significant clusters in correlation analysis with that in group difference analysis, we can see a different pattern of behaviors on the thalamus and the putamen. For the putamen, significant group difference clusters are mostly overlaid with the area that significantly associated with thalamus atrophy, supporting the thoughts that the disturbances on the putamen are more a secondary effect of thalamic injury than a primary putaminal one. Whereas for the thalamus, significant group difference clusters do not fit well with the putaminal associated areas, and the only overlap confined to V A of the left thalamus. Other nuclei, like the anterior part of the thalamus, are not significantly 3.5 Discussion 66 correlated with putaminal volume, but also prone to atrophy in preterm infants. This re- veals a more complicated pattern of thalamic disturbance: the altered development of the thalamus in premature infants cannot be completely attributed to the disturbance of the closed frontal-subcortical loop, as other open elements to the frontal-subcortical circuits, for example the hippocampus, may also reciprocally connected with the thalamus and subject to preterm associated developmental alterations. Moreover, significant associa- tion between left thalamus and right putamen, as well as right thalamus and left through scale and rotation parameters may indicate the far-reaching secondary effects of thalamic alterations. These findings support the thought of V olpe [256], that thalamus undertakes duel type of injury during preterm birth, one from the preterm delivery associated disrup- tion on thalamo-cortical pathways, and one from cortico-basal ganglia circuits passing through the pallidum of the ventral striatum. As a result, to our knowledge, our current study for the first time provides neuroanatomical evidence to depict the associated, albeit not identical patterns of vulnerability of thalamus and putamen in relation to prematurity. 3.5.4 Contribution and limitation of the study The contributions of this study are threefold. First, we extend the localization power of subcortical structural analysis, and open the possibility to depict the effect of pre- maturity on the level of subnuclei of deep gray matter. Second, for the first time, we introduce the relative pose analysis into prematurity studies, and suggest that relative pose in subcortical structures is a useful indicator of brain injury, particularly in the pro- cess of development. In this paper, we used a similar method as in (Bossa et al., 2010) to represent the relative pose. For each sub-cortical structure, a point distribution model 3.5 Discussion 67 of each subject is aligned with the mean shape of the corresponding structure. This, however, is equivalent to computing the relative positions of subcortical structures with respect to each other. The mean shapes of different sub-cortical structures form a within- brain coordinate system. In this within-brain coordinate system, relative pose between neighboring structuresan be indirectly inferred. This is easy to see if, in addition, we define a fixed reference frame within the brain and compute the positions of the cen- ters of mass of the average template representing each of the subcortical structures. The Procrustes alignment used in this study is an easy and robust way to perform similarity transformation, and it is a typical method of choice when the shape is characterized as a labeled point set (Dryden and Mardia, 1998). It does not require input parameters, and thus involves minimal processing bias. Third, we investigate the associations between deep gray matter structures using the previously obtained surface and pose information, thus opening a new perspective of the investigation of specific and combined response of subcortical structures to brain disorders. The results presented in this paper may have implications for early prediction and long-term prognosis, which can be helped with im- plementing early specific behavior intervention for these preterm neonates, with the hope of reducing long term neurodevelopmental disabilities. There are several limitations in our study: First, we need to use age at scan time as a covariate in our group analysis, due to a slight mismatch of age in the preterm and term groups. The use of linear regression does not rule out the possibility of a non-linear pattern of growth against gestational age. Second, we assume each group is uniform after regressing out the effects of scan age, thus ex utero vs. in utero development is the only factor that causes the thalamus and putamen alterations in our baseline study, 3.5 Discussion 68 since all of our preterm subjects in this study are healthy with minimum therapeutic in- tervention. Due to the limitation of our dataset (only 17 preterm neonates), and the long time span of clinical data collecting, other factors like treatments or interventions were not investigated in this study. In our future studies with a larger dataset, we will group subjects in different subsets, to further investigate possible risk factors, such as degree of prematurity, different types of white matter injuries, birth weight, various treatments, etc. Third, the inter-structural association analysis is limited to the most commonly affected deep gray matter areas. However, the cortex is intimately connected with the putamen and thalamus through cortical-basal ganglia and thalamo-cortical pathways, and the in- volvement of cerebral cortex in prematurity is well accepted. Therefore, the investiga- tion of association of allied deep gray matter would be more comprehensive with the participant of the cerebral cortex. Moreover, the alteration in one brain structure may have effects on far-reaching regions in the brain, as revealed by our study. Other brain structures, especially the ones in the neighborhood of thalamus and ventral striatum, like lateral ventricles and hippocampus, may also be affected by prematurity. As a result, the inclusion of other sub-cortical structures may yield new insights to the linked brain developmental disturbances of prematurity. Fourth, there is difficult in distinguishing be- tween the margin of the globus pallidus and the putamen in the neonates, so we combined these two structures into one model in the analysis. This however may lead to erroneous or spurious measurements of surface and pose in the medial margins of these structures (the genu of the internal capsule is located between these structures). Fifth, the sub- nuclei of gray matter structures in this study (those of the thalamus, globus pallidus and putamen) were not differentiable in the T1 images, so the identification of sub-divisions 3.6 Funding 69 was based on visual comparisons with anatomical references. The inclusion of diffusion tensor imaging in future work will allow for connectivity-based parcellations, thus the frontal-subcortical connectivity will be addressed more directly in these diffusion tensor imaging based studies. In the future, we will also like to apply this data set to a larger group of preterm neonates which have different types of clinical risk factors to see how these risk factors affect subcortical structure development. For example, it will be interesting to divide the subjects into very preterm and late preterm birth groups, to investigate the influence of degree of prematurity on at-risk brain regions. It will also be useful to apply this methodology on preterm neonates with different degrees of perinatal white matter injury, which can give insight into the relationship between gray and white matter injuries. In addition, we will also apply the method to a longitudinal study of the same patients, thus allowing us to correlate the baseline results with long-term cognitive outcome, which will enrich our current project findings and yield important biomarkers of learning and behavioral deficits in premature children. 3.6 Funding This work was supported by the National Institutes of Health through NIH grant 5K23- NS063371 and grant R21EB012177. Chapter 4 Structural Changes of the Basal Ganglia in Children with Chronic Manganese Exposure The work in this chapter has been adapted from the following publication: Yi Lao, Laurie-Anne Dion, Gabriel Rocha, Guillaume Gilbert, Maryse Bouchard, Natasha Leporé, and Dave Saint-Amour, 3D Surface Analysis reveals the impact of childhood manganese exposure on the Basal Ganglia, in revision for Scientific Reports. 4.1 Abstract 71 4.1 Abstract Chronic manganese (Mn) exposure is associated with neuromotor and neurocognitive deficits, but the exact mechanism of Mn neurotoxicity is still unclear. With the advent of magnetic resonance imaging (MRI), in-vivo analysis of brain structures has become possible. Among different sub-cortical structures, the basal ganglia (BG) has been in- vestigated as a putative anatomical biomarker in MR-based studies of Mn toxicity. How- ever, previous investigations have yielded inconsistent results in terms of regional MR signal intensity changes. These discrepancies may be due to the subtlety of brain al- terations caused by Mn toxicity, coupled to analysis techniques that lack the requisite detection power. Here, based on brain MRI, we apply a 3D surface-based morphometry method on 3 bilateral basal ganglia structures in school-age children chronically exposed to Mn through drinking water to investigate the effect of Mn exposure on brain anatomy. Our method successfully pinpointed significant enlargement of many areas of the basal ganglia structures, preferentially affecting the putamen. Moreover, these areas showed significant correlations with fine motor performance, indicating a possible link between altered basal ganglia neurodevelopment and declined motor performance in high Mn ex- posed children. 4.2 Introduction Manganese (Mn) is an essential element that participates in daily metabolic activities in the human body, but that can be toxic when the dose exceeds Mn homeostasis [54, 93]. Studies have shown that Mn accumulation due to chronic occupational or environmental 4.2 Introduction 72 exposure may cause neuromotor and cognitive deficits [104, 171, 208, 209]. There are growing concerns about the risk of Mn childhood accumulation [29, 30, 64, 132, 133, 190, 267, 271], as the immaturity of the biliary system results in higher Mn retention in both young animals and humans [128, 129]. In particular, Wasserman et al. investigated 10-year old children and found that children consuming drinking water with a high con- centration of Mn had significantly lower IQ scores [267]. Bouchard et al. [30] further confirmed the adverse effects of Mn accumulation via drinking water on cognition, and these effects were observed in lower water Mn concentrations than in Wright et al. [271]. With the advent of MRI, efforts have been made to elucidate the neuro-substrates for Mn neurotoxicity. A brain diffusion tensor imaging (DTI) study of adult welders showed sig- nificant alterations of the corpus callosum (CC) and frontal white matter from long-term Mn exposure [135]. Task fMRI studies indicated additional brain activations in adult welders compared with controls, likely to ensure adequate motor performance and work- ing memory [42, 43]. However, the exact mechanisms of Mn neurotoxicity are still not fully understood, especially in cases of exposure to Mn through drinking water, and the determination of dose-response level of Mn toxicity remains elusive [93]. More dedi- cated studies on selective brain regions are needed to develop biomarkers of excessive Mn exposure and their association with clinical symptoms. The basal ganglia (BG) consists of several important subcortical gray matter nuclei, and is thought to be the intersection where action and cognition meet [127]. Due to its involvement in several motor and cognitive cortical-subcortical loops, BG dysfunctions often lead to disorders of motor initiation and inhibition, notably Parkinson’s disease (PD)[174, 200]. Mn intoxication patients often show symptoms that resemble those of 4.2 Introduction 73 PD, therefore the BG has been targeted as a putative location for studies on manganese toxicity [54, 227]. Among the several sub-nuclei of the BG, the globus pallidus is thought to be the most affected, but inconsistent results have been reported. For instance, Hauster et al. [102], Fell et al. [70], Kafritsa et al. [122], and Ikeda et al. [112] found T1 hy- perintensities in the globus pallidus in patients with elevated Mn blood concentrations due to chronic liver disease or long-term parenteral nutrition, while Shinotoh et al. [224], Kim et al. [136] as well as Aschner et al. [14] reported no BG T1 signal intensity alter- ations in patients with occupational Mn exposure or manganese intoxication secondary to iron deficiency anemia. In addition, Criswell et al. explored several BG subregions and found that intensity indices in the caudate, the anterior and posterior putamen and the globus pallidus were all correlated with the magnitude of Mn exposure [53]. More inter- estingly, the combined global basal ganglia intensity index was more correlated with Mn exposure than the widely accepted pallidal index. The discrepancy in previous studies may be due to varied Mn toxicity levels, and methodologies that are either not powerful enough or whose results vary with MRI scanner settings. Thus, higher detection power techniques are required to better understand the relative vulnerability of the BG nuclei to Mn toxicity. Here, for the first time, we apply a 3D surface-based morphometry analysis on 3 BG structures, including the putamen, the globus pallidus and the caudate, to investigate the neuroanatomic correlates of chronic childhood Mn exposure through drinking water. In the present study, we compare 10 high Mn exposure and 13 age matched low exposure school-age children, and aim to examine the differences associated with Mn exposure through drinking water. We previously assessed these children with the standard MRI 4.3 Methods 74 procedure for Mn exposure, that is, evaluation of MRI signal intensity in the globus pallidus. We did not observe any hyperintensities in the globus pallidus, as might be expected based on the literature on occupationally exposed individuals [119], but rather, as detailed in our previous work, we observed significant hypointensities in the group of children with higher Mn exposure [64]. The significance of this finding is not clear, but might indicate neurological damages since the hypointensities correlated with poorer motor function [64]. Our study may provide new biomarkers for Mn neurotoxicity as well as neuro-substrates for the impaired motor performance associated with excessive Mn exposure. 4.3 Methods Brain T1-weighted MR images of 10 (mean age: 12.5, SD: 1.30 years) children with chronic exposure to Mn and 13 (mean age: 11.9, SD: 1.9 years) age matched controls were acquired with a 3T Philips Achieva X system. T1-weighted images were acquired using a 3D spoiled gradient-echo with inversion recovery preparation (repetition time (TR) = 8.1 ms, echo time (TE) = 3.7 ms, inversion time (TI) = 1005 ms, acquisition matrix = 240 x 240 x 160, resolution = 1 mm x 1 mm x 1 mm, SENSE factor = 1.5). An 8-channel phased-array head coil was used for signal reception, and head motion was minimized using cushions. Sequence duration was 7.5 minutes. Selection criteria: For this present MRI investigation, carried out in 2010-2012, we selected children based on the concentration of Mn in the tap water of their residence, as measured in the initial study. We enrolled 13 children with low and 10 children with 4.3 Methods 75 Fig. 4.1 Illustration of segmentation and the corresponding 3D structures in axial and coronal views. Red, blue and yellow each represents the caudate, the putamen and the globus pallidus, respectively. None of the participants had visible hyper- or hypointensi- ties in the basal ganglia. high Mn concentration in their drinking water (< 30 and>100µg/L, respectively). For each participant, an informed written consent was obtained from a parent, along with a written assent from the child. All MRI imaging was performed in accordance with guidelines and regulations of the Department of Radiology of the Centre Hospitalier de l’Université de Montréal (CHUM), using procedures that were approved by the Sainte- Justine Hospital, Notre-Dame Hospital and Université du Québec à Montréal research ethics committees. Written informed consent was obtained from the subject prior to 4.3 Methods 76 the MR examination. The study was conducted in accordance with the Declaration of Helsinki. Mn concentration from the children in the low-exposure group ranged from 0.2 to 27 µg/L (median = 0.9, SD = 9) and those in the high-exposure group ranged from 103 to 264µg/L (median = 145, SD = 54). Children from the two exposure groups were comparable in terms of age, sex, HOME scores, and full-scale IQ scores [64], as shown in Table 4.1. 4.3.1 Preprocessing T1-weighted MR images for all the subjects are first bias corrected and skull stripped using the FSL software [117]. The preprocessed T1 data are linearly registered to one of the randomly chosen controls [116]. Three bilateral basal ganglia nuclei are manually traced on linearly aligned T1 images by a neuroradiology trainee using Insight Toolkit’s SNAP program [276]. The intra-rater percentage overlaps are 0.90 for the putamen, 0.91 for the globus pallidus, and 0.90 for the caudate. 3D surface representations of the 3 basal ganglia nuclei are constructed based on binary segmentations, and mesh grids are then built on the surfaces using our in-house conformal mapping program [264]. Subse- quently, constrained harmonic based registrations are performed between an intermediate surface and each of the surface models, to obtain a one-to-one correspondence between vertices [264]. Then, volume based analysis and surface based analysis are performed, as discussed in detail in the following section. 4.3 Methods 77 Table 4.1 Demographics of our subjects. Subjects Number Mn concentration Age Gender IQ Santa Ana scores (µg/L) Mn l-exposure 13 0.9± 9 12.5± 1.3 5M8F 101.3± 9.1 36.9± 7.6 Mn h-exposure 10 145± 54 11.9± 1.9 4M6F 105.0± 10.8 30.2± 5.9 Table 4.2 Group difference results of traditional univariate volume based analyses (VBA), and multivariate tensor based morphometry (mTBM) analyses on left (l-), right (r-), and combined (c-) basal ganglia structures: Putamen (Puta), Globus Pallidus (GP), Caudate (Cau). All the p-values are corrected for multiple comparisons using structure-wide per- mutation testing with 10,000 permutations [150, 153]. Significance is set at p < 0.05, and significant values are marked using ∗∗ . Low p-values implying trends are marked using∗ . Structures V olumes V olumes p values p values (high exposure group)/mm 3 (low exposure group)/mm 3 VBA mTBM l-Puta 5075.8± 216.6 4725.2± 648.3 0.1460 0.0067 ∗∗ r-Puta 4904.9± 311.5 4465.3± 1231.7 0.3686 0.1001 ∗ c-Puta 9980.7± 447.5 9190.5± 1607.9 0.1801 0.0121 ∗∗ l-GP 1298.2± 119.1 1221.2± 148.8 0.2338 0.0767 ∗ r-GP 1213.5± 133.1 1152.6± 116.1 0.1878 0.2453 c-GP 2511.7± 235.5 2373.9± 258.2 0.2046 0.1045 ∗ l-Cau 4156.4± 346.0 3921.4± 477.7 0.1759 0.1663 r-Cau 4431.9± 344.4 4234.9± 573.2 0.3124 0.1171 c-Cau 8588.2± 609.5 8156.3± 1037.0 0.2232 0.1396 4.3 Methods 78 4.3.2 Volume based analysis Among the various shape analysis methods on subcortical gray matter, the whole vol- ume based analysis is the most intuitive one. Although volume based comparisons are generally limited in terms of detection and localization, they may provide preliminary insights into which structures may be affected. Here, we calculate the volumes of the 3 bilateral basal ganglia nuclei, to detect gross volume changes prior to the fine grained surface based morphometry analysis. A univariatet- test is performed on the volumes of each structure, and 10,000 permutations are employed to avoid the normal distribution assumption [150, 153]. P- values for each of the tests are shown in Table 4.2. 4.3.3 Surface based analysis We used several shape measurements in our multivariate tensor based analysis. The first is medial axial distance (MAD), which represents the distance from a given surface vertex to the medial axis and represents the shape’s thickness at each point of its surface. The second is logged deformation tensor (logS). Specifically, the registration between each subject’s image and the template results in a displacement field ⃗ u. For each indexed vertex on subject’s surface, a Jacobian matrix J=Id+∇⃗ u is computed, where Id is the identity matrix. In tensor-based morphometry, vertex-wise Jacobian matrices, or their derived functions are typically used as measurements in group analyses. In particular, the deformation tensorS (S= √ JJ T ), which can be expressed as a 2D ellipse at the center of each grid cell, captures directional difference at the corresponding location on the surface area as compared to a template [153]. The use of S largely preserves the information in 4.3 Methods 79 shape changes, which is commonly lost when using the univariate determinant or trace of the Jacobian matrices (detJ ortrJ). To simplify computations, we use the logarithms of the deformation tensors (logS) instead of S, such that the projected elements form a vector space where standard Euclidean space formulae can be applied [13]. According to our previous studies on populations with various age ranges [150, 220, 264], the combination of MAD and logged deformation tensor (MADMTBM) has shown highest detection power compared to the widely used MAD or determinant based mor- phometry analyses (which detect surface area changes without the directional informa- tion). Thus, in this study, the surface based morphometry was mainly carried out through MADMTBM measures. For each vertex on the surface, the feature vector was defined as the combined logged deformation tensor and radial distance (4 x 1). To compare the direction of alterations, i.e. whether there is an enlargement or a shrinkage in the high Mn exposed group compared with low exposed controls, we map the ratio of the meandetJ of the two groups at each vertex. The ratio maps are shown in Fig. 4.2. Furthermore, a multivariate analysis based on the 4 feature values (MAD and 3 values of the deformation tensor matrix) is performed to test whether the alteration is significant or not. As previously stated, our subjects from the two exposure groups were comparable in terms of age, sex, home observation for measurement of the environment (HOME) scores, and full-scale IQ scores [64]. However, the effect of brain growth on BG mor- phometry from childhood into adolescence is not negligible [214, 218]. Given the rel- atively large age range (from 9 to 15 years old) in our dataset, we use linear regression to factor out the effect of age. Linear regressions are carried out for each of the channel 4.3 Methods 80 within the feature vector separately. Following this, Hotelling’s T 2 - tests are performed on the covaried feature vectors, as described in [107, 153, 264]. Given the limited sample size of our study, we conduct vertex-wise permutation tests to avoid assuming a normal distribution. To do this, we shuffle the group labels and compare T 2 - values obtained as such with T 2 - values from the real data. For each ver- tex, 10,000 permutations are performed to assemble a null distribution of T 2 - values [183]. Vertex-wise corrected p-value maps for BG are shown in Fig. 4.3. In addition, each structure has 15,000 vertices on its surface, thus is inevitably subjected to multiple comparison errors. As a result, we also conducted 10,000 permutations over the whole structure surface to obtain a single p- value corrected for multiple comparisons. The overall p-values after multiple comparison correction are shown in Table 4.2. Both cor- rection methods have been validated in previous multivariate tensor based morphometry studies with limited sample sizes [150, 153, 220]. Correlation Analysis Pearson’s correlations are performed between mTBM measures of detJ and motor per- formance. Specifically, we use the Santa Ana Pegboard Test from the Neurobehavioral Core Test Battery to access hand dexterity and coordination [268]. Santa Ana scores were collected in our previous epidemiological study [30], and have been shown to be most correlated to Mn exposure in the GP [64]. The vertex-wise correlation coefficients are shown in Fig. 4.4. 4.4 Results 81 4.4 Results Twenty-three children aged 9 to 15 (mean: 12.2 years) were grouped into 10 high Mn exposure and 13 age matched low exposure participants, according to the Mn concen- tration in their drinking water. Brain T1 MR images of all the subjects were acquired using a 3T Philips scanner. Three bilateral BG nuclei (the caudate, the putamen, and the globus pallidus) were manually traced on preprocessed and linearly aligned T1 images. In the presented dataset, no visible T1 intensity alterations were found in the BG in our participants. Thus, potential Mn depositions in the deep gray matter structures, if there were any, did not influence the differentiation of BG nuclei boundaries. Based on the binary segmentations, 3D surface mesh grids of the BG structures were constructed and registered using an in-house algorithm [264]. Manual segmentation of 3 bilateral BG nuclei and the corresponding 3D surfaces are shown in Fig. 4.1. This was followed by a traditional volume based analysis (VBA) and a multivariate tensor based morphometry (mTBM). Details can be found in the materials and methods section. Group analysis results using VBA and mTBM are shown in Table 4.2. Whole struc- ture volumes are computed to provide an intuitive comparison of structure sizes between the low and high Mn exposure groups, while mTBM is more sensitive to regional alter- ations in structures. As we can see from Table 4.2, all the investigated BG structures (the putamen, the globus pallidus and the caudate) have average volumes that are slightly larger in the high Mn exposure group compared with those from the low Mn exposure group. However, none of these differences reach statistical significance. As for surface based morphometry, after structure-wise multiple comparisons correction, the bilateral putamen shows significant differences between exposure groups, and the bilateral cau- 4.4 Results 82 date and globus pallidus both show similar trends. We also run structure-wise permuta- tion based statistical tests on both sides of the structures to correct for multiple compar- isons in an exploratory analysis. We detect significant differences in the left putamen and trends in the right putamen as well as the left globus pallidus. To visualize the type of alterations (i.e., enlargement or shrinkage) in surface based morphometry, we map the ratio of the mean determinants (detJ) of the two groups at each vertex (Fig. 4.2). Vertex-wise multivariate analyses based on the combination of 4 feature values (MAD and 3 eigenvalues of the logged deformation tensor) are also performed, and the corresponding p-value maps are shown in Fig. 4.3. While for several of the BG subnuclei (i.e., the right putamen and the left globus pallidus), structure-wise p-values in Table 4.2 only show trends, significant clusters with evident enlargements are present in several sub-regions on the nuclei’s surfaces. As shown in Figs. 4.2 and 4.3, several areas of the BG structures, notably the anterior aspect, are significantly enlarged. In the putamen, both the anterior and the posterior ends are larger in the high-exposure group, and broad surface areas show significant group differences in Fig. 4.3. In the globus pallidus, most of the significant clusters are larger in the high-exposure group, with the main significant clusters located in the left anterior side. In the caudate, we observe evidently larger anterior ends on the bilateral sides, and a smaller posterior end of the left side in the high-exposure group, though only few surface clusters reach significance. Vertex-wise detJ from mTBM are also correlated with motor performance, specif- ically the Santa Ana Pegboard Test developed to assess fine motor function [268]. In Fig. 4.4 and 4.5, structural enlargement, notably the anterior aspect of the left putamen 4.5 Discussion 83 and the left caudate are negatively correlated with motor performance. Moreover, areas significantly correlated with motor performance largely overlap those shown in the group difference maps in Fig. 4.2 and 4.3. Fig. 4.2 Mean ratio map of detJ between high-exposed and low-exposed groups are dis- played in four views: superior (A), inferior (B), posterior (C), and anterior (D). Areas in red represent an increase of detJ in the high-exposed group compared with that in the low-exposed group, indicating an expansion of the corresponding areas. Areas in blue show a decrease detJ in the high-exposed group compared with that in the low-exposed group, indicating a contraction of the corresponding areas. 4.5 Discussion Here, we applied a MRI based 3D mTBM analysis on BG subnuclei to investigate BG morphometry alterations in the developing brain in response to long-term Mn exposure from drinking water. Statistical comparisons based on BG subnuclei volume did not 4.5 Discussion 84 Fig. 4.3 Statistical results of the multivariate analysis are displayed in four views: supe- rior (A), inferior (B), posterior (C), and anterior (D). Areas in colors other than deep blue represent vertex-wise significances of the multivariate analyses (p <0.05). detect any significant alterations. Using mTBM, we successfully detected significant enlargement in the putamen and trends of enlargement in the left globus pallidus. In addition, we correlated fine motor performance, based on Santa Ana test scores, with regional surface measurements of detJ. Reduced motor performance was found to be significantly correlated with regional enlargement in anterior ends of the putamen, globus pallidus and caudate – the same areas that were also found to be significantly altered in the mTBM comparisons between Mn exposure groups. These results are in accordance with deficits in cognitive and neuromotor performance observed in these children from previous studies [64, 190], and with a study showing that enlarged anterior BGs are associated with lower behavioral performance in pre-adolescent children [214]. Our findings demonstrated significant regional enlargement in BG subnuclei of Mn 4.5 Discussion 85 Fig. 4.4 Vertex-wise correlation coefficients between vertex-wise det J values with Santa Ana scores are displayed in four views: superior (A), inferior (B), posterior (C), and anterior (D). Areas in red colors represent a negative correlation between surface areas and motor performance, and vice versa for blue colors. exposed children. Altered BG morphologies are significantly associated with poor motor performance in this population. Collectively, these results suggest that neuro-circuits within the motor loop are at risk in children with Mn exposure through drinking water. Although there is still controversy regarding the exact mechanisms of Mn toxicity, growing focus has been directed to dysfunctions of the basal ganglia motor pathways [51, 52, 224]. The striatum, including the caudate and the putamen, receive anatomi- cal projections from the motor cortex and substantia nigra pars compacta and send out- puts via the globus pallidus, forming programming motor loops. The striatum and the globus pallidus are thought to be the primary targets of Mn accumulation in the brain [54]. In particular, a positron emission tomography (PET) study on four patients with 4.5 Discussion 86 Fig. 4.5 Correlation of vertex-wise detJ values with Santa Ana scores are displayed in four views: superior (A), inferior (B), posterior (C), and anterior (D). Areas in colors other than deep blue represent vertex-wise significances of the multivariate analyses (p <0.05). chronic occupational Mn exposure proposed that Mn is likely to disrupt downstream BG dopaminergic output pathways [224]. Criswell et al. investigated 20 asymptomatic welders exposed to Mn fumes, and reported reduced 6-[18F]fluoro-L-DOPA (FDOPA) PET uptake predominantly in the anterior striatum [51]. A subsequent study on a patient with Mn accumulation in end-stage liver disease confirmed reduced FDOPA uptake in the caudate, and in the anterior and posterior putamen [52]. In addition, a voxel based morphometry study reported volume reduction co-located to the globus pallidus [41], while other T1 intensity based studies on the globus pallidus reported inconsistent find- ings [70, 102, 112, 122, 136, 224]. Our current study extends these findings on the Mn- induced neuroanatomical changes in the BG subnuclei to children with exposure to Mn 4.5 Discussion 87 present in drinking water. Here, we detected evident enlargements in BG subnuclei, no- tably the left caudate, the whole putamen and the globus pallidus. Concurrent alterations in all three regions may imply disturbances in their shared BG motor pathways. Support for the hypothesis of BG motor pathways disturbances also comes from our correlation analysis between BG shape measurements and fine motor performance. In a previous study in the same population [64], three different motor measurements were tested, and Santa Ana Pegboard Test performances were the ones most correlated with Mn exposure. In line with this, in the presented study, enlargements in the anterior cau- date and putamen were significantly correlated with Santa Ana scores. These are also consistent with the findings of Sandman et al. [214], in which anterior striatum changes were found to be significantly associated with poorer cognition in healthy pre-adolescent children. However, the current study is not powered to reveal disturbances in specific mo- tor pathways (i.e. direct/indirect motor pathways, or the nigrostiatal pathways.). Future investigations integrating high angular resolution diffusion - weighted imaging (HARDI) tractography analysis on selected region of interest will be needed to refine our current findings. While previous MRI studies of Mn neurotoxicity focused on occupational Mn ex- posure in adults, we are the first to investigate basal ganglia morphometry alterations in children with Mn exposure. Studies of Mn intoxication in adults have reported reduced gray matter volumes as well as reduced WM integrity [41]. For instance, in a whole brain voxel-based morphometry analysis [41], brain volumes, notably the globus pallidus and cerebellar regions, were found to be reduced in occupationally Mn exposed welders. A region of interest study that was also performed on welders showed reduced appar- 4.5 Discussion 88 ent diffusion coefficient values in the globus pallidus and anterior putamen [53]. The enlargement of bilateral BG structures in our study are opposite in direction to the reduc- tions found in adults. The discrepancy may be due to the different age of the included subjects, coupled with different Mn exposure routes [41]. The brain undergoes important development from childhood to adolescence, and in- volves a complex sequence of neuronal growth and pruning [81, 83]. The interplay of external disturbances and changes in normal brain growth encodes specific brain mor- phometry, making the onset age of Mn neurotoxicity an important factor to consider in interpreting the results. Prior imaging studies of gray matter maturation showed in- creased synaptic pruning and reduced gray matter density in adolescents’ brains [81, 203, 232, 233]. Consequences of disturbances of this process are not well understood, albeit certain association of enlargement and neurobehavioral measures were reported by some studies. For example, based on the correlation results from 50 pre-adolescent children, Sandman et al. proposed an enlargement of the basal ganglia as a possible bio-marker of developmental impairment, with all three basal ganglia nuclei involved, and the puta- men being preferentially affected [214]. In accordance with these findings, we detected significant enlargement in the basal ganglia, notably in the putamen. The enlargement of the caudate and putamen were also correlated with fine motor functions. Alternatively, this enlargement pattern may be specific to Mn toxicity through drink- ing water. Researches on risk assessments of inhaled and ingested Mn reported quantita- tive differences of Mn tissue uptake via different dose routes [10]. In particular, inhaled Mn contaminated dusts that bypass normal gastrointestinal control system likely bind to transferrin as trivalent Mn, and thus are hard to eliminate by the liver and are easily accu- 4.5 Discussion 89 mulated in the brain via transferrin receptors [10]. Thus, the occupational Mn exposure investigated in previous studies may have a longer period of tissue uptake, and may even- tually impact brain anatomy in a pattern different from other routes of Mn overexposure [51, 52]. In our previous study on the same population, T1 intensity indices were also found to be different from those of adult studies [64]. These results, collectively, may indicate alternative Mn neurotoxicity mechanisms in children with waterborne Mn expo- sure. However, these preliminary hypotheses need further validation, including pharma- cokinetic data on more subjects with different pathways of Mn overexposure. The limitations of our study are as follows: First, we relied on Mn concentration in drinking water to define exposure groups. Mn water concentration was much more associated with hair Mn concentration (an exposure biomarker to Mn) than dietary Mn intake in a larger study from which our current subjects were selected [30]. However, it is possible that additional Mn exposure also came from air or food. Second, we did not explore the possible influence of other contaminants present in water. However, including iron concentrations in the regression model did not reduced the association between Mn concentration and the lower IQ scores, whereas other metals had very low concentrations (e.g. lead, arsenic). Third, our sample spans a range of ages from 9-15 years old, a period at which the brain is still rapidly developing, requiring the use of linear regression to control for the effect of age. Shaw et al. mapped the trajectory of BG development in 220 healthy subjects, and found that the growth of globus pallidus is close to linear [218]. In the same study, the trajectory of the striatum was found to be approximately linear from 7 to 14 years old, and the growth rate slows down afterward. In the current study, we have a total of 23 subjects, and only one of our subjects was over 14 years old 4.5 Discussion 90 at the time of scan. In order to avoid the over-fitting problem using nonlinear regression with a limited sample size, we adopted a simple linear regression to factor out the effect of age. This is recognized as one of the caveats of this study, and shall be optimized with better regression models in future larger datasets. In the future, in addition to enrolling more subjects, we would also like to include more brain regions into our analysis. BG substructures investigated here are most likely the primary, but not the exclusive targets of Mn neurotoxicity. For example, alterations of the frontal cortex and corpus callosum have been reported in occupationally exposed adult welders [135]. These brain regions will be included in future. Chapter 5 Pose, Surface and Connectivity Changes of the Corpus Callosum in Collegiate Contact Sport Athletes The work in this chapter has been adapted from the following publications: Yi Lao, Niharika Gajawelli, Lauren Haas, Bryce Wilkins, Darryl Hwang, Sinchai Tsao, Yalin Wang, Meng Law, and Natasha Leporé, 3D Pre- vs. Post-Season Comparisons of Surface and Relative Pose of the Corpus Callosum in Contact Sport Athletes, SPIE Med- ical Imaging, 15 - 20 February (2014), San Diego, USA. Yi Lao, Meng Law, Jie Shi, Niharika Gajawelli, Lauren Haas, Yalin Wang, and Natasha Leporé, A T1 and DTI fused 3D Corpus Callosum analysis in pre- vs. post- season contact sports players, 10th International Symposium on Medical Information 5.1 Comparisons of the Relative Pose of the Corpus Callosum in Contact Sport Athletes 92 Processing and Analysis (SIPAIM), October 14-16, 2014, Cartagena, Columbia. 5.1 Comparisons of the Relative Pose of the Corpus Cal- losum in Contact Sport Athletes 5.1.1 Abstract Mild traumatic brain injury (MTBI) or concussive injury affects 1.7 million Americans annually, of which 300,000 are due to recreational activities and contact sports, such as football, rugby, and boxing[20]. Finding the neuroanatomical correlates of brain TBI non-invasively and precisely is crucial for diagnosis and prognosis. Several stud- ies have shown the influence of traumatic brain injury (TBI) on the integrity of brain WM [69, 169, 229]. The vast majority of these works focus on athletes with diagnosed concussions. However, in contact sports, athletes are subjected to repeated hits to the head throughout the season, and we hypothesize that these have an influence on white matter integrity. In particular, the corpus callosum (CC), as a small structure connecting the brain hemispheres, may be particularly affected by torques generated by collisions, even in the absence of full blown concussions. Here, we use a novel relative pose analysis, which consists of detecting the relative translation, rotation and scale between two groups, to investigate TBI related brain struc- tural changes between 9 pre-season and 9 post-season contact sport athlete MRIs 5.1 Comparisons of the Relative Pose of the Corpus Callosum in Contact Sport Athletes 93 5.1.2 Description of purpose TBI may lead to brain white matter injuries and disruptions, and thus temporarily or permanently impair brain function [78, 169, 229]. In-vivo imaging with MRI may be used to detect changes of brain structures in shape and relative position to better under- stand the impact of TBI on brain anatomy. However, to our knowledge, the relative pose of brain structures has not yet been studied in TBI. Additionally, we use a new surface based method, multivariate tensor-based morphometry (mTBM) [264] to study the full 3D structure of the CC. Being the largest WM structure in the brain, the CC bridges the left and right hemi- sphere of the brain through transverse WM fibers. The midline location and intricate pathways make CC particularly vulnerable to the shear of nerve fibers resulting from traumatic axonal injury [269]. Thus, the CC is likely a sensitive indicator of brain TBI, in terms of changes in size, shape and relative pose. Using brain structural MRI, we perform regional group comparisons of the relative pose of CC between pre-season and post-season contact sport athletes. Statistic analyses are conducted on a similarity trans- formation resulting pose parameters, indicating the changes on the relative position of CC within the brain. We hypothesize that there may be subtle positional differences between pre- and posts-season contact sport participants that may related to repeated physical impact. 5.1 Comparisons of the Relative Pose of the Corpus Callosum in Contact Sport Athletes 94 5.1.3 Method 18 T1-weighted MRI scans of collegiate contact sport athletes (9 pre-season scans and 9 post-season scans) were acquired on a 3T GE HDxT scanner. After pre-processing, in- cluding skull stripping and bias correction, the data were registered to the same template space through linear registration. The CCs were then manually traced on the linear regis- tered T1 images, and the intra-rater percentage overlap was 0.937, with four participants at two different intervals. Subsequently, 3D surface representations and mesh grids of the CC were constructed using our in-house conformal mapping program. One-to-one corre- spondence between vertices were obtained through a surface fluid registration algorithm [Shi et al.]. Relative pose analysis For each CC, the relative pose was obtained by a full Procrustes fit of a template shape to the PDM[68, 211]. The template shape was selected as the mean shape that minimized the Procrustes distances, and it was computed iteratively. The transformation rule of Procrustes alignment is defined as, T ⟨X⟩=(sRX, d), where s is the scalar scaling factor, R is a 3× 3 rotation matrix and d is the translation vector (x,y,z) T [27]. To form a Lie group with matrix multiplication [28], the matrix representation of the Procrustes transformation can be written as: T =    sRX d 0 T 1    (5.1) To simplify computations, all the parameters of the transformations were projected 5.1 Comparisons of the Relative Pose of the Corpus Callosum in Contact Sport Athletes 95 onto a log-Euclidean space. The mean pose in the log-Euclidean space can be calculated iteratively as in [192]: m k+1 =m k exp( 1 n n ∑ i=1 log(m − 1 k T i )). (5.2) After the subtraction of the mean (m) from each subject’s individual pose, specifically usingv i =log(m − 1 T i ), each subject is left with a residual pose. Statistics are computed on the residual pose which consists of 7 parameters: 1 scale scalar, 3 rotation scalars and 3 translation scalars. Statistical comparisons between the two groups were performed via two methods: univariatet-test for logS,||logR||,||logd||,θ x ,θ y ,θ z , x, y, z; multivariate Hotelling’s T 2 - test for 3 rotation parameters (θ x ,θ y ,θ z ), 3 translation parameters (x, y, z), a combination oflogS,||logR||, and||logd||, as well as a combination of all 7 parameters. Considering the limited size of our dataset, for each of the parameters, a permutation test[155, 183] with 10,000 random redistributions of the ’contact’ or ’non-contact’ labels was performed to avoid the normal distribution assumption. 5.1.4 Results Relative pose results All the p-values from previously described tests are presented in Table 1. As we can see from the table, pose parameters representing rotation show a significant difference between the pre-season and post-season groups, while no difference can be seen in scale and translation parameters. Additionally, a combination of all 7 parameters also detected 5.1 Comparisons of the Relative Pose of the Corpus Callosum in Contact Sport Athletes 96 Table 5.1 P-value of statistical analyses on pose parameters: 13 sets of parameters charac- terizing the relative pose of the CC are investigated here using univariate and multivariate analyses. Parameters are categorized as logS,||logR||,||logd||, θ x ,θ y ,θ z , x, y, and z for univariate tests, and as (θ x ,θ y ,θ z ), (x, y, z), (logS,||logR||,||logd||), and a combination of 7 parameters for multivariate tests. All the p-values are obtained after permutation testing. Significant p-values (p < 0.05) are highlighted in light cyan. Non significant, but interested low p-values are highlighted in light grey. pose parameters p-values pose parameters p-values logS 7.95e-01 ||logR|| 4.39e-02 θ x 1.51e-02 ||logd|| 4.70e-01 θ y 8.95e-01 (θ x ,θ y ,θ z ) 1.10e-01 θ z 1.63e-01 (x,y,z) 3.11e-01 x 5.73e-01 (logS,||logR||,||logd||) 2.49e-01 y 2.00e-01 All7paras 3.52e-02 z 1.27e-01 significant differences between the two groups, indicating a possibility of using multi- variate analysis of all pose parameters as the discriminant between the two populations. These results are better visualized in Fig. 5.1, which shows a superimposition of the mean poses of the pre-season group (red) and post-season group (blue) respectively. The CC of the post-season group showed an left leaning tendency compared to the pre-season group, which is consistent with the differences found in rotation parameters. Compared to the obvious rotational difference, the size and translation differences in these two groups are less evident, thus further validating lack of significance results from the size and translation parameters. 5.1 Comparisons of the Relative Pose of the Corpus Callosum in Contact Sport Athletes 97 5.1.5 New or breakthrough work to be presented We introduce a novel pose analysis system that integrates various brain surface process- ing techniques including parametrization and surface fluid registration and reports the subtle pose changes on subcortical structures. The obtained pose information is com- plementary to subcortical surface shape analyses. There is little research analyzing the relative pose of brain structures in general, and none in TBI. Our work may lead to new biomarkers for TBI. 5.1.6 Conclusions Here, we introduce a relative pose analyses into the TBI associated brain anatomy anal- yses. Our measurements successfully detected differences in the CC between pre-season and post-season contact sport participants, in terms of changes in both shape and pose. Our study supports the concept that there are TBI related brain structural changes in con- tact sport athletes that may be related to repeat physical impacts. Our results provide additional information in understanding the TBI patterns, and may serve as sensitive in- dicators of similar kinds of brain anatomy changes. In the future, we will recruit more subjects in our study and further validate these results. In addition, we will correlate our surface based morphometry and relative pose results with the neuropsychological testing scores of these subjects. 5.2 A T1 and DTI Fused 3D Corpus Callosum Analysis in Contact Sports Athletes 98 5.2 A T1 and DTI Fused 3D Corpus Callosum Analysis in Contact Sports Athletes 5.2.1 Abstract Sports related traumatic brain injury (TBI) is a worldwide public health issue, and dam- age to the corpus callosum (CC) has been considered as an important indicator of TBI. However, contact sports players suffer repeated hits to the head during the course of a season even in the absence of diagnosed concussion, and less is known about their effect on callosal anatomy. In addition, T1-weighted and diffusion tensor brain magnetic reso- nance images (DTI) have been analyzed separately, but a joint analysis of both types of data may increase statistical power and give a more complete understanding of anatom- ical correlates of subclinical concussions in these athletes. Here, for the first time, we fuse T1 surface-based morphometry and a new DTI analysis on 3D surface represen- tations of the CCs into a single statistical analysis on these subjects. Our new com- bined method successfully increases detection power in detecting differences between pre- vs. post-season contact sports players. Alterations are found in the ventral genu, isthmus, and splenium of CC. Our findings may inform future health assessments in con- tact sports players. The new method here is also the first truly multimodal diffusion and T1-weighted analysis of the CC in TBI, and may be useful to detect anatomical changes in the corpus callosum in other multimodal datasets. 5.2 A T1 and DTI Fused 3D Corpus Callosum Analysis in Contact Sports Athletes 99 5.2.2 Introduction Sports related traumatic brain injury (TBI) has been drawing broad attention in the last few years as short and long term consequences on athletes’ health have come to light. The corpus callosum (CC) is particularly at risk due to its midline location and intri- cate white matter pathways. Regional CC damage, including alterations of the genu and splenium, has been detected in MRI based studies [144, 166], and has been associated with post-concussion behavioral symptoms in moderate to severe TBI cases [144, 269]. However, less is known about the consequences of repeated head blows sustained by contact sports players even in the absence of diagnosed concussion. A couple of recent studies have begun to look at this issue through comparing whole brain diffusion maps [78] and regions-of-interest including the CC [168] in groups of athletes scanned pre- vs. post-season,. Finding sensitive indicators of CC alterations is of great importance for athletes’ health assessment and to the design of sports safety rules, with the aim of reducing long-term brain damage. From the above studies, it is clear that a better understanding of the alteration in fiber structure of the CC is much needed. However, current analysis methods that zoom in on specific brain white matter tracts may discard some of the information in the CC. This is because popular methods such as Tract-Based Spatial Statistics (TBSS) [230] and Tract Specific Analysis (TSA) [275] project the CC into midline or mid-plane sur- face. However, previous postmortem and probabilistic tractography studies have shown that the CC is not a homogenous structure, in terms of fiber composition [3] and to- pographical distribution[191]. Fig. 5.2a shows fractional anisotropy (FA) maps in 3 different orientations from one subject. As we can see, FA values of CC regions are not 5.2 A T1 and DTI Fused 3D Corpus Callosum Analysis in Contact Sports Athletes 100 uniformly distributed in all three views, thus dimension reduction methods may lead to reduced detection power (Fig. 5.5). Moreover, in addition to the more commonly studied anterior-posterior CC line, the dorsal-ventral organization of CC also presents important connections with the medial-lateral cortex [191]. Thus, 3D representations may better localize injury, and may have higher statistical detection power to identify neuro-circuits alterations underling the observed changes in behavior. In addition to the potential value of 3D surface representations to investigate diffusiv- ity changes in the CC, morphological alterations may provide complementary informa- tion in deciphering brain alterations due to trauma. This is illustrated in Fig. 5.2b, where both sets of data are overlaid and FA and shape are each shown to provide important in- formation on pre- vs. post-season anatomical changes. Impaired neurological abilities in TBI subjects have been attributed to both parenchyma and diffuse injuries, and the former is more easily seen on T1 images [90, 151], while the latter is better detected using DTI. For instance, in moderate to severe TBI studies, T1-based analyses found morphometric changes of several subcortical structures [106, 249]. However, parenchyma and diffuse injuries often occur concomitantly in white matter structures such as the CC, and a joint analysis of diffusion and T1-weighted data may therefore provide a more complete pic- ture of CC changes brought on by brain injury. As a result, an analysis based on a fusion of structural and diffusion information of brain regions is likely to yield higher detection power. Here we perform the first ever joint analysis of structural and diffusion MRI data on the CC of contact sports players scanned pre- vs. post-season. To do so, we design a new combined surface-based algorithm on 3D segmentations of the CCs. For each of the CCs, 5.2 A T1 and DTI Fused 3D Corpus Callosum Analysis in Contact Sports Athletes 101 a conformal grid is built on its surface and a constrained harmonic mapping is performed to match it to a common template. For the T1-weighted data, we statistically compare differences in local area and size via the vertex-wise deformation tensors from the surface registration. For the DTI data, we project the diffusion-derived information (FA, MD...) on the surface of the CC (as opposed to a medial axis [230] or mid-plane [275] as in popular methods). As only one of our subjects is known to have had a concussion during that particular season, we test whether repeated subclinical injury may lead to detectable changes in callosal anatomy. 5.2.3 Subjects and Methodologies Data and Preprocessing T1-weighted MRI and DTI scans of male collegiate contact sports athletes (19 set of scans in total:10 pre- and 9 post-season scans) were acquired on a 3T GE HDxT scanner. 8 subjects were scanned both pre- and post-season, while 2 subjects were scanned pre- season only, and one subject was scanned post-season only. Each of the DTI scans was obtained using a single shot echo-planar imaging sequence with a b-value of 1000s/mm 2 and 25 gradient directions. T1 images for each of the subjects were first skull stripped and bias corrected, and were then registered to a common template space through linear registration [116]. Lin- ear transformation matrices resulting from this step were saved for later use. The CCs were then manually traced on the linear registered T1 images, and the intra-rater percent- age overlap was 0.937, in four participants at two different times. 5.2 A T1 and DTI Fused 3D Corpus Callosum Analysis in Contact Sports Athletes 102 To combine the information from DTI and T1, we need to integrate the information in the same space. After preprocessing of DTI data, we transformed DT images from each subject to their corresponding T1 space, using linear registration between the b0- weighted and T1 images. Subsequently, the linear transformation matrices saved from the T1 registration were applied on the linearly aligned b0 images, to transform the dif- fusion information to the space of the T1 template. After each of the linear registrations, the diffusion tensors were resampled using the b0 transformation matrices, and rotated according to the underling anatomy. These steps are all achieved using MedINRIA [251]. 3D Representations and Registration Based on the CC segmentations on T1 images obtained in section 7.3.1, 3D surface rep- resentations and conformal mesh grids of the CC were constructed. Subsequently, one- to-one correspondences between vertices were obtained through a constrained harmonic algorithm [264]. After registration, a deformation tensor √ JJ T - whereJ is the Jacobian of the transformation from the registration - was computed at each vertex on the surface. Its determinant (detJ, the difference in surface area) and projection on the log-Euclidean space (log √ JJ T ) [12] were used in later statistical analysis [153, 264]. Statistical Analysis Our statistical analysis is performed using either the morphometry information only, the diffusion information only, or a combination of both. To project diffusion indices of each of the CCs onto its surface, we first calculated midlines of all the 3D CCs, and then collected diffusion parameters to each surface vertex along its corresponding radius to 5.2 A T1 and DTI Fused 3D Corpus Callosum Analysis in Contact Sports Athletes 103 the midline, specifically using: ∥ ( −−−→ X− M)× ( −−−→ P− M) ∥ −−−→ X− M∥ ∥≤ R. (5.3) and ( − −− → X− P)· ( −−−→ P− M)≥ 0. (5.4) HereX,M,P are the(x,y,z) coordinates of a vertex in the surface, the corresponding point of the vertex in the midline, and a voxel within the 3D representations, respectively, while R represents a pre-defined distance between P to the line of −−−→ X− M. Illustration of the sampling results can be found in Fig. 7.1. In our experience, some vertices will not be assigned any within-CC voxels if we use a R< 0.6mm 3 . To make sure each of the vertices has some voxels assigned, and to minimize overlap with neighboring vertices, we tested 3 R values: 0.6mm 3 , 0.8mm 3 and 1.0mm 3 . From Fig. 5.4, we choseR= 0.6mm 3 . Vertex-wise univariate analysis or multivariate analysis are performed based on the following variables: 1 Morphometry information: univariate detJ and multivariate(e1,e2,e3) from the logged deformation tensors (Fig. 7.2, 1 st and 2 nd row). 2 Diffusion information: mean FA and MD along the radius of CC for each vertex (Fig. 7.2, 3 th and 4 th row). 3 Diffusion information: multivariate λ 1 and λ 2 (Fig. 7.2, 5 th row). Note: we did not includeλ 3 . Being small, this value is susceptible to noise and may reduce detection 5.2 A T1 and DTI Fused 3D Corpus Callosum Analysis in Contact Sports Athletes 104 power. 4 A fusion of morphometry (detJ) and diffusion indices (λ 1 andλ 2 ) (Fig.7.2, 6 th row). Two types of permutation tests were performed: a vertex-based one to avoid the normal distribution assumption and one over the whole segmented image to correct for multiple comparisons [153, 183]. 5.2.4 Results 3D Diffusion Only Analyses Vertex-wise group differences results based on univariate FA and MD, as well as multivariate (λ 1 ,λ 2 ) are shown in Fig. 7.2, row 3 rd to 5 th . All methods give consistent results, with the main clusters located in the ventral genu, isth- mus, and splenium. The multivariate test of (λ 1 ,λ 2 ) shows a significant overall p-value (p=0.0352), while the univariate FA (p=0.0797) or MD (p=0.1181) only display trends. We also perform traditional centerline based methods on FA and MD as a comparison. As shown in Fig. 5.5, mean FA or MD over the perpendicular planes are assigned to each of the corresponding vertices in the centerline, where statistical analysis and permutation test are performed. From Fig. 7.2, significant areas from the centerline based methods are grossly matched with our 3D based method, but not surprisingly, fewer group differences are detected in the former. Fusion As shown in Fig. 7.2, the multivariate analysis of (λ 1 ,λ 2 ,detJ) shows consis- tent regions of significant group differences, and outperforms all the other univariate or multivariate methods in detection power (p=0.0136). 5.2 A T1 and DTI Fused 3D Corpus Callosum Analysis in Contact Sports Athletes 105 5.2.5 Discussion Methodological considerations Group differences of brain white matter (WM), including CC, are typically analyzed based on voxels, midlines, or midplanes. However, voxel-based methods give poor lo- calization of differences in anatomical regions compared to surface-based ones and may be contaminated by differently oriented tracts [187], while midline- or midplane-based methods rely on assumptions that WM perpendicular to the mid-line or the mid-plane are uniformly distributed. The method introduced in this paper uses clearly defined CC re- gions traced in T1 images, and largely preserved within tract information projected onto the surface of the CCs. In our dataset, our new method outperforms previous methods in specificity and detection power. Here, for the first time, we fuse the T1 based morphome- try information and DTI based diffusion information on 3D CCs, and successfully detect significant differences between pre- vs. post-season contact sports players. The fused method also outperforms univariate and multivariate 3D analyses based on the structural information only, or the diffusion information only in terms of detection power. Neuropathological considerations In moderate to severe TBI [269], the genu and splenium were reported to particularly vulnerable and highly associated with neurological outcomes. In sports related mild TBI, post-injury cognitive and affective symptoms, as well as impaired neurologic outcomes have also been reported [166, 167, 244]. However, the extent of brain vulnerability in contact sports players who did not suffer concussions is less understood, and none of the 5.2 A T1 and DTI Fused 3D Corpus Callosum Analysis in Contact Sports Athletes 106 studies to date have successfully located the injury in specific regions of CC [78, 168]. As shown in Fig. 7.2, the biggest clusters of significant differences between pre- and post-season contact sports players are located in the ventral genu and posterior areas in- cluding ventral isthmus and splenium. The ventral genu is neuroanatomically connected to lateral prefrontal cortex, which mediates executive functions and attention. Dam- age to this cortex region has been linked to impaired learning ability [170]. Posterior CC regions including the ventral isthmus and splenium are connected with the occipital, temporal and the posterior parietal lobe, and in particular are in charge of visual related tasks, including visual memory. The significantly altered ventral anterior and posterior regions detected by our fusion method are interestingly consistent with cognitive anal- yses on mild TBI contact sports players [244], in which impairments of visual working memory and altered dorsal lateral prefrontal cortex activation are observed. Our findings provide new neuroanatomic evidence for the diagnosis of mTBI, and may enhance our understanding of the neuroanatomical consequences mTBI related pathologies. 5.2 A T1 and DTI Fused 3D Corpus Callosum Analysis in Contact Sports Athletes 107 Fig. 5.1 3D visualization of the mean shape of all the CCs in the mean poses averaged from pre-season group (red) and post-season group (blue) respectively. A shift in pose is evident in terms of a rotation on the anterior and posterior ends of the CC, while less visible differences can be seen in size and transformation. 5.2 A T1 and DTI Fused 3D Corpus Callosum Analysis in Contact Sports Athletes 108 (a) FA value distribution for 3 orientations. (b) FA and shape information are complementary to each other. Fig. 5.2 In (a), from left to right: axial view, sagittal view, and coronal view. In (b), overlaid T1 segmentations (A) and FA maps in same subject scanned pre- (B) vs. post- season (C). In figure A, pre- and post-season CC segmentations are represented in yellow and blue, respectively, with the overlaid areas appearing in orange. Comparing the three figures, circled area shows both structural and diffusion changes, the area indicated by a triangle is structurally different but has an unchanged diffusion pattern, and vice versa for the area shown by the arrow. 5.2 A T1 and DTI Fused 3D Corpus Callosum Analysis in Contact Sports Athletes 109 Fig. 5.3 Surface of CC for one of our subjects and the illustration of sampled voxels. The red line on the right side of the figure is the midline of the CC, and pink crosses are surface vertices. In the direction perpendicular to the midline and pointing to each vertex, blue stars represent voxels projected to each of the vertices, and a mean index (FA, MD....) of projected voxels is assigned to the vertex for later statistics. Fig. 5.4 FA maps for the same subject sampled on the surface of the CC, using different sampling distances. From left to right: R= 0.6mm 3 , R= 0.8mm 3 , R= 1.0mm 3 . Note: to save computation time, results shown in this figure were obtained on a downsampled surface. All three results are grossly matched, while sampling using a larger radius leads to smoother regional changes. R= 0.6mm 3 was selected here. 5.2 A T1 and DTI Fused 3D Corpus Callosum Analysis in Contact Sports Athletes 110 Fig. 5.5 Group difference results between pre- and post-season subjects obtained using centerline representations. Projections to the each of the centerline vertex were made through mean FA or MD of the perpendicular plane to the midline in the location of the given vertex. Midline vertices with a p-value< 0.05 are displayed in red. 5.2 A T1 and DTI Fused 3D Corpus Callosum Analysis in Contact Sports Athletes 111 Fig. 5.6 pre- versus post-season contact sports players using different measures. Vertex- wise corresponding p− values are displayed. In addition, whole structure corrected p− values are p=0.5061 for detJ, p= 0.2253 for (e1,e2,e3), p= 0.0797 for FA, p= 0.1181 for MD, p= 0.0352 for (λ 1 ,λ 2 ), p= 0.0136 for (λ 1 ,λ 2 , detJ). Chapter 6 A study of brain white matter plasticity in early blinds using Tract Based Spatial Statistics and Tract Statistical Analysis The work in this chapter has been adapted from the following publication: Yi Lao, Yue Kang, Olivier Collignon, Caroline Brun, Shadi Bohlool Kheibai, Flamine Alary, James Gee, Marvin D. Nelson, Franco Leporé, Natasha Leporé, A study of brain white matter plasticity in early blinds using Tract Based Spatial Statistics and Tract Sta- tistical Analysis, NeuroReport, 2015, 26(18): 1151-1154. 6.1 Abstract 113 6.1 Abstract Blind individuals are known to exhibit brain white matter changes resulting from the in- terplay between visual deprivation and brain compensatory plasticity. Particularly, early onset blindness might trigger profound brain reorganizations that affects not only the vi- sual system, but also other sensory modalities. Diffusion tensor imaging (DTI) allows in- vivo visualization of brain white matter connectivity and brain white matter alterations. Among statistical approaches based on DTI, Tract-Based Spatial Statistics (TBSS) [230] is widely used because of its ability to automatically perform whole brain white mat- ter studies. Tract Specific Analysis (TSA) [275] is a more recent method that localizes changes in specific white matter bundles. Here we compare TBSS and TSA results on DTI scans from 12 early blind subjects (age of onset <5 years old) and 13 age matched sighted controls, with two aims: 1) to investigate white matter alterations associated with early visual deprivation; 2) to examine the relative sensitivity TSA, compared to TBSS, to both deficit and hypertrophy of white matter microstructures. Both methods give con- sistent results for regions of deficits. However, TBSS does not detect hypertrophy of white matter, while TSA shows a higher sensitivity in detecting subtle differences in white matter microstructures. 6.2 Introduction Blind individuals are subject to brain functional and anatomical reorganization, in re- sponse to the absence of visual inputs [31, 152, 156, 198, 199, 223, 225]. Functional and structural MRI studies have demonstrated that the timing of blindness onset influences 6.2 Introduction 114 the extent of brain alterations [31, 198, 223, 225]. As an example, a PET study revealed specific activation patterns of the visual cortex during Braille reading in congenital blind subjects, while the activation pattern differed in late blind subjects [31]. A structural MRI study found that volume gains in non-occipital areas are more widespread in early- onset subjects than late-onset blind subjects based on whole brain tensor-based morphometry [156]. With the advent of technique of diffusion tensor imaging (DTI), brain white matter integrity can be analyzed noninvasively at the micro scale, thus facilitating the in-vivo detection of white matter connectivity alterations in the human brain [8,9]. Among white matter dimensionality reduction methods, Tract-Based Spatial Statistics (TBSS) [230] is widely used because it is automatic and projects results on white matter tracts. In particular, Shu, et al. looked at subjects with early visual deprivation using TBSS, and reported significant white matter impairments in geniculocalcarine tract and its adjacent regions, while no white matter hypertrophy was mentioned [225]. Wang, et al. found both decreases and increases on the brain white matter skeleton of the blindness brain using TBSS, although the p-values for the increase changes are not corrected [257]. The slightly inconsistency in the reported results highlight the need for further investigation with higher detection power in this area. Tract Specific Analysis (TSA) [275] is a more recent semi-automated method that quantifies white matter changes in specific white matter bundles. Instead of looking at the whole brain white matter at once, TSA takes advantages of the thin sheet-like major tracts and zooms in on specific white matter bundles, thereby improving localization of the results [275]. Whereas TBSS algorithm registers the low-dimensional FA map, 6.3 Materials and Methods 115 TSA uses DTI-TK [274] to register the whole diffusion tensors. Another advantage of TSA is that with TBSS, a precise segmentation of the template is needed to determine specific locations of changes, whereas with bundle specific approaches such as TSA, the tract is manually defined and thus more anatomically correct. Here we hypothesize that the improved registration and localization in TSA will also lead to increased statistical detection power. In this paper, we compare the white matter alterations in early blind subjects using the TBSS and TSA approaches. Our purposes are two-fold: first, exploring the deficit and gain of white matter associated with early blindness, due to the interplay of sen- sory deprivation and brain compensatory; second, to examine the sensitivity of TSA, as compared to the more standard TBSS, to both deficit and hypertrophy of white matter microstructures. 6.3 Materials and Methods Our dataset consists of 12 early blind subjects (age: 40.1+/-10.2 years, 8 males, 4 fe- males) and 13 age-matched (age: 38.5+/-9.4 years, 8 males, 5 females) sighted controls. Brain DTI scans of the recruited subjects were acquired using a Siemens (Avanto) 1.5 Tesla scanner, single-shot echo-planar (EPI) diffusion tensor sequence, with an acquisi- tion time of approximately 8-12 minutes and 12 gradient directions. Imaging parameters were: TR=8000 ms, TE=92 ms, 2.5 mm inter-slice distance and 4 averages. DTI data from all the subjects are first preprocessed (including skull stripping, eddy current correction, and tensor estimation), and are then registered to the IXI adult tem- 6.4 Results 116 plate through DTI-TK [13]. Subsequently, statistical analyses on FA values are per- formed using TBSS and TSA. For TBSS, statistics are carried out on skeletonized mean FA image as a whole. For TSA, we focused on six major tracts: corpus callosum (CC), corticospinal tract (CST), fronto- occipital (IFO), inferior longitudinal fasciculus (ILF), superior longitudinal fasciculus (SLF), and uncinate (UNC)). In both cases, p values are corrected for multiple comparisons using 10,000 permutations, and the significance thresholds are set to p <0.01. 6.4 Results Statistical comparison of FA between early blind and sighted groups using TBSS, TBA approaches are shown in Fig.6.1. As for white matter deficit, both methods detect de- creased FA in early blind subjects in widespread areas of the brain white matter, in- cluding but not limited to the occipital lobe. These are consistent with the findings in [31, 156, 226, 257, 273]. On the other hand, under the hypothesis of FA greater in blind than sighted, no significant difference is detected using TBSS, while significantly increased white matter integrity was detected using TSA in parts of the CC and CST con- necting to the posterior parietal lobe. These are consistent with the tractography based findings of significantly increased mean FA of CST [273] and increased white matter volume in sub-cortical parietal lobe [156] in these subjects. 6.4 Results 117 Fig. 6.1 Statistical analysis of FA using TBSS(A)(C) and TSA(B)(D). For TBSS, results are shown as mean FA skeleton maps. Significant areas with p<0.01 are marked in red. For TSA, results are shown for the bilateral corticospinal tracts (CST), inferior fronto- occipital tracts (IFO), inferior longitudinal fasciculus tracts (ILF), superior longitudinal fasciculus tracts (SLF), uncinate tracts (UNC), and the corpus callosum (CC). Significant areas with p<0.01 are colored in dark blue and encircled with a white line. Figure A and B: results under the hypothesis of blind <sighted. Row C and D: results under the hypothesis of blind >sighted. 6.5 Discussion and conclusion 118 6.5 Discussion and conclusion In the present study, we found significant alterations in white matter microstructure in several brain regions between early blind and sighted group using TSA, consistent with the TBSS results. White matter deficits are detected using both TBSS and TSA in the congenital blindness subjects in broad brain regions, mainly distributed in the frontal and the occipital lobe. These are consistent with previous findings of impaired white matter integrity secondary to early visual deprivation, which extensively affects optic nerves and visual process involved neuro-connections [31, 156, 226, 257, 273]. On the other hand, white matter alteration resulted from CNS adaption are much more subtle and thus require more sensitive analysis approaches. In our experiment settings, TSA successfully pinpoints significantly strengthened white matter in the corpus callosum and CST, co- localized to the posterior parietal lobe, while TBSS failed to identify the alteration. Blind plasticity in early onset blind subjects is well accepted as a means of reor- ganizing functions of the visual system [31, 198, 223]. A few structural and diffusion MRI studies further confirm the brain microstructural alterations in the congenital blind brain in response to visual deprivation. For instance, Lepore, et al. looked at structural MRI scans of early blind subjects, and reported regional hypertrophy using tensor-based morphometry [156]. Yu, et al. calculated the FA values over the entire tractography de- rived cortical spinal tracts, and found significant higher FA values in early blind subjects compared to normal controls [273]. Wang, et al. reported more extensive white matter impairments in late blind subjects than congenital blind subjects using both TBSS and VBA, and found a trend of increased FA in bilateral CST using TBSS, with uncorrected p- values in these two groups compared to sighted controls [257]. 6.5 Discussion and conclusion 119 Here, the white matter hypertrophies detected by TSA are consistent with previous findings [152, 156, 257, 273]. Moreover, for the first time, white matter alterations are specifically located in the posterior parietal lobe. The involvement of parietal cortex in route planning is well documented [4, 109, 176]. In particular, the posterior parietal cortex (PPC), which receives the visual inputs from dorsal visual stream, has long been viewed as an essential part of spatial vision [109, 176]. In congenital blind subjects, in- creased functional connectivities between visual areas and the parietal cortex have been reported based on electrotactile stimulation of the tongue [31]. The significantly in- creased FA in white matter tracts towards the posterior parietal cortex further confirms the strengthened connectivity in response to the deprivation of sight. This may form new neuro-substrate for the compensatory adjustment in congenital blind subjects in the absence of visual inputs. Chapter 7 Disentangling the primary brain effects of vascular risk factors from mild cognitive impairment in aging subjects. The work in this chapter has been submitted to: Yi Lao, Binh Nguyen, Sinchai Tsao, Niharika Gajawelli, Meng Law, Helena Chui, Yalin Wang, and Natasha Leporé, A T1 and DTI fused 3D Corpus Callosum analysis in MCI subjects with high and low cardiovascular risk profile, In revision for Neuroimage:Clinical. 7.1 Abstract 121 7.1 Abstract Understanding the extent to which vascular disease and its risk factors are associated with prodromal dementia, notably Alzheimer’s disease (AD), may enhance predictive accuracy as well as guide early interventions. One promising avenue to determine this relationship consists of looking for reliable and sensitivein-vivo imaging methods capa- ble of characterizing the subtle brain alterations before the clinical manifestations. How- ever, little is known from the imaging perspective about how risk factors such as vascular disease influence AD progression. Here, for the first time, we apply an innovative T1 and DTI fusion analysis of 3D corpus callosum (CC) on mild cognitive impairment (MCI) populations with different levels of vascular profile, aiming to de-couple the vascular factor in the prodromal AD stage. Our new fusion method successfully increases the de- tection power for differentiating MCI subjects with high from low vascular risk profiles, as well as from healthy controls. MCI subjects with high and low vascular risk profiles showed differed alteration patterns in the anterior CC, which may help to elucidate the inter-wired relationship between MCI and vascular risk factors. 1, DTI, Corpus Callosum, cardiovascular disease, cognitive disorders, mild cognitive impairment, Alzheimer disease, dementia 7.2 Introduction Emerging evidence has shown that cardiovascular disease (CVD) and preclinical cardio- vascular risk factors are linked to the etiology of dementia, including Alzheimer disease (AD) [38, 88, 111, 123, 162, 182, 197, 254]. Specifically, some findings suggest a direct 7.2 Introduction 122 influence of vascular diseases in accelerating amyloid β accumulation [79, 111]. The entanglement of cardiovascular and neural factors is further evidenced by the recently hypothesized connection between the Locus Coeruleus (LC) and AD. In this scenario, AD is mediated by the integrated modulatory function of LC on the heart rate, attention memory and cognitive functions [Mather and Harley]. While an effective treatment for AD is still out of reach, there are established therapeutic strategies for CVD, and its risk factors are also clinical modifiable [46]. Therefore, disentangling the effects of CVD and its risk factors on the development of AD has implication for symptom management, and may potentially alter clinical outcomes for pre-dementia patients. In particular, differentiating the effects of different CVD profiles on the anatomy of the brain in mild cognitive impairment (MCI) - a precursor to AD and other types of dementia - would provide important insights into the effects of preventable CVD factors on the initial course of AD. Nevertheless, efforts aimed at differentiating vascular dis- eases from MCI report inconsistent results [103, 160, 184]. In particular, Hayden et al. identified a set of memory and executive tests in prodromal vascular dementia (VaD) that are distinguishable from prodromal AD [103]. Nordlund et al. confirmed the differences in executive function between MCI subjects with and without vascular disease, and also reported differences in speed, attention, and visuospatial functions in these two groups [184]. However, no differences between vascular and non-vascular types of MCI were found by other studies [160]. These discrepancies may partially be caused by different inclusion criteria for vascular disease (i.e. with or without stroke), coupled to analysis techniques that do not have the required detection sensitivity. This highlights the need for sensitive and reliable algorithms that can help in decoupling the vascular component 7.2 Introduction 123 of preclinical dementia, and thus aid in early diagnosis and therapeutic design. Being the largest white matter (WM) structure with a high demand of blood supply from several main arterial systems, the corpus callosum (CC) has been reported to be vulnerable to both MCI [62, 63, 246, 277] and vascular diseases [61, 76, 86]. Given the extensive connections between the CC and the cortex, regional disturbances of the CC may mirror dysfunctions of specific cortical domains. Therefore, anatomical alterations of the CC may serve as potential discriminators of the concurrent but possibly different effects of vascular and neurodegenerative components. Structural magnetic resonance imaging (MRI) is a typical choice for detecting CC anatomical alterations and has been effective in deciphering brain parenchyma loss [217, 245, 263, 279], while diffusion ten- sor imaging (DTI) has been promising in characterizing WM microstructure alterations [63, 124, 246, 277]. Parenchyma and diffuse injuries often occur concomitantly in WM structures such as the CC, and a joint analysis of diffusion and T1-weighted data may therefore provide a more complete picture of CC changes brought on by brain injury. However, to the best of our knowledge, all the above studies regarded each aspect on its own [63, 263, 277], or by comparing them side-by-side [62, 246]. None have tried to truly combine these two features into one analysis. Here, we perform an innovative, truly combined analysis of structural and diffusion MRI data on the 3D CC surfaces of MCI subjects. We group our MCI subjects into high and low vascular risk profiles (will be referred to as MCI-l and MCI-h groups in the re- mainder of the text), and conduct pairwise statistical analyses on the fused morphological and diffusion properties among these two MCI subgroups as well as on aging controls without cognitive impairment. Our aims are two-fold: 1) to test whether the vascular 7.3 Subjects and Methodologies 124 component affects distinct regional alterations that may help us to distinguish different MCI subtypes; 2) to further validate the feasibility and sensitivity of using our T1 and DTI fusion method to analyze subcortical alterations. 7.3 Subjects and Methodologies 7.3.1 Data and Preprocessing Fifty-eight subjects aged 66 to 89 were grouped based on their clinical dementia rating (CDR) and vascular risk profile into 15 MCI subjects with low vascular risk (76.40 ± 7.65 years, CDR=0.5, low Framingham cardiovascular risk profile (FCRP) scores), 18 MCI subjects with high vascular risk (78.39± 5.69 years, CDR = 0.5, high FCRP scores or had previous clinical diagnosis of myocardial infarction) and 25 healthy controls (76.68± 6.40 years). Subjects with confounding neurological conditions, such as stroke, were ex- cluded. Brain T1 and DT-MR scans of all the subjects were obtained using a 3T SIEMENS scanner. DTI data were acquired using an echo-planar imaging sequence, with a voxel size of 2× 2× 2 mm 3 , resolution of 128× 128× 60, b-value of 1000s/mm 2 , and 60 gradient directions. Anatomical data were acquired using a MPRAGE sequence, with a voxel size of 1× 1× 1mm 3 , resolution of 256× 256× 192, TE=2.98ms, TR=2500ms, and TI=1100ms. Each subject, in addition to being imaged via T1-MRI and DTI, was evaluated on the MMSE (Mini-Mental State Exam) as a marker of cognitive function, as well as a standardized battery of neuropsychological tests, consisting of MEMSC (verbal memory summary score) for verbal memory, NVMEMSC (non-verbal memory sum- mary score) for non-verbal memory, EXECSC (executive function summary score) to 7.3 Subjects and Methodologies 125 measure executive function, and GLOBSC (global cognition summary score) to assess global cognition. These measures have previously been described in the literature and are commonly used in neuropsychological assessments [179]. All the T1 data were preprocessed and linearly registered to the same template space - selected randomly from one of the controls that was previously transformed to MNI space [116]. On the linearly registered T1 images, each subject’s corpus callosum (CC) was manually traced on the mid-sagittal plane, and the lateral boundaries were determined where the CC starts to radiate into and merge with cerebral white matter. Subsequently, 3D surface representations and conformal mesh grids of the CC were constructed using an in-house conformal mapping program [264]. One-to-one correspondence between vertices were obtained through constrained harmonic based registration [264]. All the DTI data were first preprocessed, which included brain masking, eddy current correction, echo-planar imaging distortion correction, and tensor estimation. To truly integrate DTI and T1 information, we transformed the DT images from each subject to their corresponding T1 space, using linear registration between the b0-weighted and T1 images. The linear transformation matrices saved from the T1 registration were then applied on the linearly aligned b0 images, to transform the diffusion information to the space of the T1 template. After each of the linear registrations, the diffusion tensors were resampled using the b0 transformation matrices, and rotated according to the underlying anatomy. These steps were achieved using MedINRIA [251]. 7.3 Subjects and Methodologies 126 7.3.2 Surface Based Sampling After surface based registration in section 7.3.1, a deformation tensor √ JJ T - whereJ is the Jacobian of the transformation from the registration - was computed at each vertex on the surface. Its determinant (detJ, the difference in surface area) and projection on the log-Euclidean space (log √ JJ T ) [12] were used in later statistical analysis [153, 264]. To project diffusion indices of each of the CCs onto its surface, we first calculated midlines of all the 3D CCs, and then collected diffusion parameters to each surface vertex along its corresponding radius to the midline, specifically using: ∥ ( −−−→ X− M)× ( −−−→ P− M) ∥ −−−→ X− M∥ ∥≤ R. (7.1) and ( − −− → X− P)· ( −−−→ P− M)≥ 0. (7.2) HereX,M,P are the(x,y,z) coordinates of a vertex in the surface, the corresponding point of the vertex in the midline, and a voxel within the 3D representations, respectively, while R represents a pre-defined distance between P to the line of −−−→ X− M. According to our previous pilot study, we chose R= 0.6mm 3 to make sure each of the vertices has some voxels assigned, and to minimize overlap with neighboring vertices [148]. After the surface based registration in Section 7.3.1, a deformation tensor √ JJ T - whereJ is the Jacobian of the transformation from the registration - was computed at each vertex on the surface. Its determinant (detJ, the difference in surface area) and projection on the log-Euclidean space (log √ JJ T ) [12] were used in later statistical analysis [153, 264]. 7.3 Subjects and Methodologies 127 To project diffusion indices of each of the CCs onto its surface, we first calculated midlines of all the 3D CCs, and then collected diffusion parameters to each surface vertex along its corresponding radius to the midline, specifically using: ∥ ( −−−→ X− M)× ( −−−→ P− M) ∥ −−−→ X− M∥ ∥≤ R. (7.3) and ( − −− → X− P)· ( −−−→ P− M)≥ 0. (7.4) HereX,M,P are the(x,y,z) coordinates of a vertex in the surface, the corresponding point of the vertex in the midline, and a voxel within the 3D representations, respectively, while R represents a pre-defined distance between P to the line of −−−→ X− M. The sampling process can be more intuitively visualized in Fig.7.1. The first equation is used to assign the voxels close enough(≤ R) to the corresponding radius of the vertex, while the second equation constrains the sampling to the voxels within the space between the midpoint to the vertex (not in the prolongation direction). According to our previous pilot study, we chose R = 0.6 mm 3 to make sure each of the vertices has some voxels assigned, and to minimize overlap with neighboring vertices [148]. 7.3.3 Statistical Analysis Our statistical analyses were conducted using either the morphometry information, the diffusion information, or a combination of both. Vertex-wise univariate studentt− tests or multivariate Hotelling’sT 2 tests were performed based on the following variables: 1 Morphometry information: univariate detJ and multivariate(s1,s2,s3) from the logged 7.3 Subjects and Methodologies 128 Fig. 7.1 Surface of CC and the illustration of sampled voxels. The red line on the right side of the figure is the midline of the CC, blue stars represent voxels within CC, and pink, yellow, as well as green crosses represent surface vertices. In the direction perpendicular to the midline and pointing to each vertex, voxels within pink, yellow, and green areas are projected to the vertices with the corresponding colors. Mean index (FA, MD....) of projected voxels is assigned to the vertex for later statistics. deformation tensors, which are independent elements of the matrix (Fig. 7.2, 1 st and 2 nd row). 2 Diffusion information: univariate mean FA along the radius of the CC for each vertex and multivariate λ 1 and λ 2 (Fig. 7.2, 3 th and 4 th row). Note: we did not include λ 3 . Being small, this value is susceptible to noise and may reduce detection power. While this is fine for additive measures such as FA as the effect will be negligible (as the value is small), it is a much bigger issue when analyzing a multivariate vector of statistics, where each eigenvalue is treated as an independent measure. 3 A fusion of morphometry (s1,s2,s3) and diffusion indices (λ 1 and λ 2 ) (Fig.7.2, 5 th row). Here, one of the primary purposes is to determine a method that can sensitively de- tect underlying anatomical differences between our MCI groups. Therefore, our general 7.3 Subjects and Methodologies 129 criteria for measurements selection were: firstly, to use the most representative, or gen- erally used measurements in both shape and diffusion analysis for comparison, and sec- ondly, to compare these to what we hypothesized would be a joint structural and diffusion measure with enough sensitivity to detect subtle underlying differences between groups, based on ours and others prior studies. MD showed significant but less powerful results than FA, and similarly,(FA,s1,s2,s3), or(MD,s1,s2,s3) showed less significant results than (λ 1 ,λ 2 ,s1,s2,s3). Therefore, we only included the above five most representative univariate or multivariate measurements in our final analysis. Given the fact that our subjects were from a relatively large age range (66-89 years), we used linear regression to factor out the effect of age. For each feature value separately, we have: F =β 0 +β 1 ∗ age+β 2 ∗ group+error. (7.5) Where F is one of the features we previously obtained,β 0 ,β 1 ,β 2 are the correspond- ing correlation coefficients. Group are coded as dummy variable: 0 for controls, 1 for MCI-l group, and 2 for MCI-h group. All the following statistics were performed on linearly regressed features. For each of the tests, two types of permutations were performed: a vertex-based one to avoid the normal distribution assumption and one over the whole segmented im- age to correct for multiple comparisons, as described in [153, 183]. Permutation based corrections are independent of the distributions of the statistics, which is commonly non- parametric in voxel- or vertex-wise analyses. Furthermore, permutation based multiple comparison corrections are less stringent than conventional sequential correction meth- 7.4 Results 130 ods [22], as they do not assume independence of neighboring voxels or vertices, and they are widely accepted in brain image analyses [153, 201, 259]. In each of the permutation tests, 10,000 permutations were employed. 7.3.4 Correlation Analysis The imaging measurements were further evaluated with respect to overall mental status demonstrated by Mini Mental State Examination scores, as well as 4 domain specific neuropsychological measurements: MEMSC, NVMEMSC, EXECSC, GLOBSC. Four subjects with missing neuropsychological test data were excluded, leaving a total of 54 subjects for the correlation analysis. After controlling for age, Pearson’s correlation anal- yses were conducted to determine specific contributions of regional CC to cognitive per- formance, in terms of shape (represented by detJ) or WM integrity (represented by mean FA). To evaluate the association of neuropsychological performances with the combined feature of shape and diffusion properties of CC (represented by (λ 1 ,λ 2 ,s1,s2,s3)), we continued to perform a distance correlation−− a generalization to the classical bivariate measurements of dependance [243]. Similar to the permutation corrections employed in group-wise comparisons, 10,000 permutations were applied in each of the correlation tests, as described in [150, 183]. 7.4 Results Fig. 7.2 shows vertex-wise group differences among 3 groups based on 5 different mea- sures: detJ,(s1,s2,s3), mean FA, (λ 1 ,λ 2 ), and fused (λ 1 ,λ 2 ,s1,s2,s3). The corresponding 7.4 Results 131 Fig. 7.2 Group analysis of MCI-l vs. controls (1 st column), MCI-h vs. controls (2 nd column), and MCI-l vs. MCI-h (3 rd column ) using 5 different measures: a) detJ; b) (s1,s2,s3); c) mean FA; d) (λ 1 ,λ 2 ); e) (λ 1 ,λ 2 ,s1,s2,s3). Vertex-wise corresponding p− values are color-coded according to the color bar in the upper left corner. P- maps are smoothed using heat kernel algorithm [47]. In addition, whole structure-wise corrected p− values are presented in Table. 7.1. 7.4 Results 132 Fig. 7.3 Average map of detJ and mean FA between groups are color-coded according to the color bar in the upper left corner. When these results are compared with Fig. 7.2, we can see the main direction of change: nearly all the significance areas fell in the Controls > MCI-l, Controls> MCI-h, as well MCI-l> MCI-h areas. structure-wise corrected p− values are displayed in Table.7.1. The final statistical results on the CC surface in Fig. 7.2 are smoothed using heat kernel algorithm as described in [47]. The MCI-l group showed alterations spanning the midbody and the posterior surface of the CC as compared to controls, with significant structure-wise differences detected by detJ, (s1,s2,s3), and (λ 1 ,λ 2 ,s1,s2,s3) measurements; the MCI-h group presented broad areas of alterations mainly located in the dorsal anterior, mid-body, and splenium of the CC compared to controls, with significant structure-wise differences detected by det J, (s1,s2,s3), and (λ 1 ,λ 2 ,s1,s2,s3) measurements, as well as trends detected by mean FA and (λ 1 ,λ 2 ) measurements. For the MCI-h vs. MCI-l, the main clusters were located in the genu of the CC, and the fusion measurements reached structure-wise significance, while mean FA and (λ 1 ,λ 2 ) showed trends. It is important to note that up-sampling the 7.4 Results 133 Table 7.1 Structure-wise corrected p− values for different measurements are displayed. All the p-values were corrected using a permutation based analysis with 10,000 permuta- tions. Significance is set to p< 0.05, and is highlighted in light cyan. P-values implying trends are highlighted in light grey. Measurements MCI-l MCI-h MCI-l vs. ctls vs. ctls vs. MCI-h detJ 0.0309 0.0292 0.9901 (s1,s2,s3) 0.0425 0.0415 0.1550 FA 0.4810 0.0815 0.0813 (λ 1 ,λ 2 ) 0.3440 0.0731 0.0673 (λ 1 ,λ 2 ,s1,s2,s3) 0.0173 0.0153 0.0107 relatively low resolution DTI data resulted in same or similar diffusion indices appearing in surrounding vertices on the CC surface, thus causing the band-like areas shown in the significance map (Fig.7.2). To intuitively understand the direction of alterations, we also mapped the average maps of vertex-wise detJ and FA between 3 groups, as shown in Fig.7.5. Compared Fig.7.5 with Fig.7.2, we can see that nearly all the significance areas fell in the Controls > MCI-l, Controls > MCI-h, as well MCI-l > MCI-h areas. These findings point to shrinkage and reduced WM integrity in the CCs of MCI-l and MCI-h patients as com- pared to those of normal controls. 7.4.1 Correlation analysis results The vertex-wise significant p− map and correlation coefficients r map from Pearson’s correlation analyses between neuroanatomical measurements (detJ and mean FA) and 5 neuro-cognitive indices (MMSE, MEMSC, NVMEMSC, EXECSC, GLOBSC) are dis- 7.4 Results 134 Table 7.2 Structure-wise corrected p− values for different measurements are displayed. All the p-values were corrected using a permutation based analysis with 10,000 permuta- tions. Significance is set to p< 0.05, and is highlighted in light cyan. P-values implying trends are highlighted in light grey. Measurements detJ mean FA (λ 1 ,λ 2 ,s1,s2,s3) MMSE 0.5290 0.1639 0.7184 MEMSC 0.3322 0.0147 0.2886 NVMEMSC 0.2902 0.2736 0.3857 EXECSC 0.0411 0.0857 0.0066 GLOBSC 0.0131 0.0208 0.0048 played in Fig.7.4 and Fig.7.5. The corresponding structural-wise corrected p− values are shown in Table.7.2. For surface shape measurements, represented by detJ, significant regional correla- tions are seen in clusters mainly located in anterior and posterior CC, two of the five correlation tests (EXECSC and GLOBSC) hit structure-wise significances according to Table.7.2. As to WM integrity, represented by mean FA along radial direction, four of the five correlation tests (MMSE, MEMSC, EXECSC, and GLOBSC) showed anatom- ical meaningful correlations with WM integrity in the dorsal anterior CC. According to Table.7.2, EXECSC showed a structure-wise correlation with mean FA that repre- sented a trend (p = 0.0857), and 2 of the 5 measurements (MEMSC and GLOBSC) reached structure-wise significance. As to the combined shape and diffusion features (λ 1 ,λ 2 ,s1,s2,s3), areas of significant correlations are mainly consistent with those de- tected by shape and WM integrity separately, while nonverbal memory scores (NVMEMSC) showed anatomical meaningful correlations in the posterior CC, which were not fully captured by shape or diffusion feature based bivariate correlations. 7.4 Results 135 Fig. 7.4 Vertex-wise significance results of correlation analyses between det J as well as mean FA vs. 5 neuropsychlogical scores. P- maps are smoothed using heat kernel algorithm [47]. 7.4 Results 136 Fig. 7.5 Vertex-wise correlation coefficient maps have been generated based on det J (left column) and mean FA (right column), respectively. Compared this figure with Fig.7.4, we can see the direction of the correlation analyses: nearly all the significant regions represent positive correlations. 7.5 Discussion and Conclusion 137 7.5 Discussion and Conclusion In our study, the MCI-h group presented widespread atrophy and reduced WM integrity spanning nearly the whole CC as compared to controls, with the largest cluster located on the posterior end. In the group analysis between the MCI-l group and the controls, similar alterations were mainly shown in the middle to the posterior regions. When comparing the MCI-h and the MCI-l group, our fusion method detected significant disparities in the dorsal anterior CC. These findings together indicate a consistent influence of MCI on the midbody to the posterior end of the CC, and importantly, a distinct effect of cardiovascu- lar profile on the genu. Moreover, these same regions presented significant correlations with neurophysiological battery tests including MEMSC, EXECSC, and GLOBSC. Our findings provide important anatomical supports to the co-existence of MCI-subtypes and may yield new insights in the distinct role of cardiovascular components in the etiology of dementia. The T1 and DTI fusion analysis presented in this study yield higher sta- tistical detection power, and may provide a new direction in analyzing subcortical WM structures. Significance of the study In the past decades, MCI has draw increasing attention as a way to study the early evo- lution of AD, and as a potential target for early interventions. However, not all MCI patients will convert to AD as it is not a homogenous state and may also precede other types of dementia, such as vascular dementia (VaD). One of the main difficulties in accu- rately predicting the MCI - AD conversion is due to the co-morbidity and shared etiology with other types of disease [173, 189]. In particular, CVD – precursors of VaD – are 7.5 Discussion and Conclusion 138 also important risk factors of AD [182, 254]. Epidemiological studies have shown that cardiovascular risk factors such as hypertension, high cholesterol, diabetes are highly associated with cognitive decline and AD [58, 182, 235]. Nevertheless, no established mechanisms clarify how CVD participates in the development of AD, and whether there is a dissociable impact of CVD and cardiovascular risk factors (CRF). A considerable number of researchers suggested a selective cognitive decline pattern associated with vascular pathology, and efforts have been made to differentiate vascular disease from AD or MCI using cognitive performance. Ingles et al. investigated the neu- ropsychological performance in elderly subjects 5 years before diagnosis, and reported a selectively low abstract reasoning performance in subjects who evolved toward vascular cognitive impairment compared to those who converted to AD or remained normal[114]. Marra et al. reported executive functioning problems in the vascular form of MCI sub- jects, whereas the degenerative form of MCI were mainly impaired in episodic memory tasks [164]. Similar to these, greater impairments in executive function have been re- ported in MCI with a vascular component [89, 103, 184, 186]. However, there is no consensus on the executive dysfunction predominance of vascular pathology. A neuro- physiological study aiming to discriminate cerebrovascular disease from AD observed a slightly severe, but non-significant executive dysfunction than memory failure in autopsy- defined cerebrovascular disease group [204]. Moreover, a study comparing cognitive profiles in MCI subjects with different etiologies reported no differences of memory or executive function between the vascular and non-vascular types of MCI [160]. These inconsistencies hint at the limitation of using neuropsychological patterns only as disso- ciable features for vascular injury/dementia [204]. 7.5 Discussion and Conclusion 139 With the advent of MRI, multiple modalities such as arterial spin labeling (ASL), structural and diffusion MRI have been utilized to investigate the vascular pathology on brain anatomy. It is widely accepted that vascular disease or risk factors are associ- ated with an accelerated rate of cerebral atrophy [18, 120, 138] and decreased glucose metabolism [45], while the information as to whether these associations are independent of MCI is minimal. In the handful studies investigating vascular pathology in the context of MCI or AD, mixed results were reported. Specifically, a volume based T1-weighted MRI study on CC in AD, VaD as well as mild ambiguous subjects reported significantly smaller anterior and posterior CC regions in the AD group, significantly smaller ante- rior CC regions in the VaD group, and no difference in sub-clinical dementia group as compared to controls, while no differences between VaD and AD groups were detected [98]. A region of interest based DTI study on MCI subjects reported decreased WM in- tegrity in selected frontal, temporal, parietal lobe regions as well as the corpus callosum in both groups of patients with and without subcortical vascular changes, while the WM alterations in the centrum semiovale and parietal lobe were believed to be more asso- ciated with the vascular pathology [222]. A T1-MRI based study on cortical thickness and grey matter (GM) volume in MCI subjects with different levels of cardiovascular file profile observed an association between elevated vascular risk factors and atrophy in the temporal and parietal lobe - the same regions affected by AD [38]. Difficulties inherent in diagnosing and AD and VaD, and different inclusion crite- ria of vascular diseases clouded the interpretation of these studies. Moreover, limited statistical power of volume based methodologies and measurements focusing on single modality measurement further reduced the sensitivity of these studies to the potential 7.5 Discussion and Conclusion 140 neuroanatomical alterations lurking in pre-clinical stages. Therefore, in-vivo measure- ments with higher sensitivity are highly desired to further explore the concurrent but possibly distinct effects of CVD and MCI on the brain. In this work, we focused on neu- rodegenerative patterns in pre-dementia stages, and excluded compounding conditions such as stroke that directly alter brain anatomy. In the present study, the joint T1 and DTI measurement in 3D CC had successfully pinpointed dorsal anterior CC regions that sig- nificantly differed between MCI-l and MCI-h group of subjects. These findings provided new anatomical evidence for the distinctive impact of vascular pathology before clinical magnification of dementia, thus of great importance in early preventive intervention and in guiding therapeutical design. The sensitivity of our methodology and the relevance of detected anatomic alterations and the corresponding neuroanatomic and functional implications will be described in details in the next sections. 7.5.1 Methodological Considerations Postmortem and probabilistic tractography studies have shown that the CC is not a ho- mogenous structure, in terms of fiber composition [3] and topographical distribution [191]. Group differences of brain WM, including the CC, are typically analyzed based on whole structure volume or some anatomically motivated partitions, voxels, midlines, or midplanes. The whole volume based method facilitates an intuitive and coarse esti- mation of CC anatomy [11, 280], but has been ineffective in detecting subtle anatomical changes. Studies based on subdivisions of the CC are more tuned to the heterogeneity of CC, but may easily be biased due to inconsistent classification (i.e. partitioning into 3, 5 or 7 compartments), as well as arbitrary delineation of subdivisions [15, 75, 210]. V oxel- 7.5 Discussion and Conclusion 141 based methods give poor localization of differences in anatomical regions compared to surface-based ones and may be contaminated by differently oriented tracts [187], while midline- or midplane-based methods rely on assumptions that WM perpendicular to the mid-line or the mid-plane is uniformly distributed. The method introduced in this paper uses clearly defined CC regions traced in T1 images, that are largely preserved within tract information projected onto the surface of the corpus callosa. The 3D representa- tions may better localize injury in heterogeneous CC, and may have higher statistical detection power to identify the neuro-circuit alterations underlying the observed anatom- ical alterations. The vulnerability of the CC in MCI has been reported in both structural and diffu- sion studies [62, 108, 246, 252, 277, 278, 281]. Structural MRI is a typical choice and has been effective in deciphering brain parenchyma loss [217, 245, 279], while DTI has been promising in characterizing white matter microstructure alterations [125, 246, 252, 277, 278, 281]. These previous studies have been analyzing the brain parenchyma or its diffusion properties on their own [252, 277, 278], or by comparing them side-by-side [62, 210, 246]. To the best of our knowledge, none have tried to truly combine these two features into one statistical analysis. As shown in our study, measurements in both structural and diffusion aspects have given significant between group differences, confirming the concomitantly occurred parenchyma and diffuse injuries in CC. Group analyses based on structural information (detJ and (s1,s2,s3)) have successfully detected alterations in the mid-body to the posterior end, while group analyses based on diffusion information (mean FA and (λ 1 ,λ 2 )) are more sensitive to alterations in the anterior and the posterior ends of CC. Here, for the first 7.5 Discussion and Conclusion 142 time, we fuse the T1-based morphometry information and DTI-based diffusion infor- mation into one, single analysis. In all three group-wise analyses, the fused method successfully outperforms analyses based on structural information or diffusion informa- tion alone. Moreover, in group comparisons between MCI-h and MCI-l group, only the fusion method reached overall significance, while no significance is detected if T1 and DTI measures are considered separately. These results show the feasibility of using the T1 and DTI fusion method to increase detection power. 7.5.2 Anatomic and Functional Implications The corpus callosum spans the midline of the brain and possesses numerous connections to surrounding structures. At its most anterior end, the genu, WM tracts innervate the frontal lobes, and infarction of the genu has been reported to result in frontal lobe dys- function [34, 146, 175]. The splenium, at the posterior end, lies in close proximity to the hippocampus through the amygdala [145], and alterations of the splenium are often as- sociated with impairments in memory and visual perception [139, 212]. In our cohort of subjects, these anatomic-functional relationships have been further validated in the cor- relation analysis of regional CC FA values with 5 neuropsychological tests. As shown in Fig. 7.4 and Fig.7.5, executive functioning, verbal memory, and global cognitive profile, which are higher brain functions that are extensively involved frontal networks [55, 105], showed significant associations with dorsal anterior CC, while nonverbal memory, which is highly correlated with hippocampus functioning [26], selectively correlates with the ventral posterior CC. Thus, the anatomy of the dorsal anterior CC is more predictive of frontal lobe in- 7.5 Discussion and Conclusion 143 volved executive and verbal memory functions, while the posterior CC is more associ- ated with temporal and parietal nonverbal memory. Taken together, the constellation of group-wise analysis and anatomical-neurophysiological correlations imply a main effect of MCI on medial to posterior cortex involved memory functions, while CVD and its risk factors add to the symptoms through frontal connections. Comparing to controls, CCs in the MCI-h group presented similar but more exten- sive alterations than those from the MCI-l group, suggesting an ’interactive’ impact of the vascular and neurodegenerative factors on brain morphometry. These are generally in line with anatomical findings showing associations between vascular brain injuries or risk factors and aggregated brain atrophy, especially in the parietal and temporal lobe [38, 255]. In terms of the comparison within MCI subgroups, significant differences resided in the dorsal anterior CC, implying an ’additive’ effect of vascular pathology on the brain frontal network that is differentiable from the non-vascular neurodegeneration influences. The frontal lobe hypo-perfusion detected by ASL-MRI has been reported to be associated with worse executive and memory function [9]. Vascular risk factors mea- sured by FCRP and high-density lipoprotein cholesterol are found to link with thinner frontotemporal cortex [255]. In our study, the distinctive impact of the cardiovascular factor in the genu of CC is consistent with the anatomic relationship between cognitive profile and the frontal lobe[255], and provides further neuroanatomic evidence support- ing the neuropsychological findings of the selective executive dysfunction of vascular pathology [103, 164, 184]. However, our implications for a vascular associated frontal lobe dominated cognitive functions needs to be interpreted with caution. As detected by the T1 and DTI joint analy- 7.5 Discussion and Conclusion 144 sis, broader areas including anterior, mid-body, and posterior CC showed different levels of alterations in the MCI-h group, while only the anterior regions reached group-wise significance. Hence, the significant alterations detected in genu between MCI with high and low vascular types does not mean that the anterior CC is the only region involved in vascular pathology. For instance, the conclusion in [255] is that vascular risk fac- tors ’interact’ with neurodegeneration factors in temporal and posterior lobe, and cause additional adverse effect on the frontal lobe. Further, reduced executive or global cog- nitive functioning implied by more severely altered genu does not necessarily lead to the assertion that cardiovascular factors impair executive functioning more severely than nonverbal memory. Previously, to validate the executive predominance model of vascu- lar pathology, [204] hypothesized a lower executive performance than episodic memory performance in cases with autopsy-defined cerebrovascular diseases, but failed to detect a statistical significant differences between the two tests. Our study provides a potential interpretation to previous results that vascular pathology may accelerating the deterio- ration of multiple cognitive domain, with its influence on the frontal involved network especially dissociable from non-vascular neurodegeneration factors. Our current study extend the database of vascular pathology on the brain in pre- dementia stages, and suggests an ’additive’, albeit not ’dominant’ effect of vascular as- sociated impact on the frontal lobe, which may eventually lead to the refinement of the widely accepted frontal predominancy theory. Our findings provide new neuroanatomi- cal substrate of vascular contributions to cognitive impairment before the magnification of dementia, which may serve as a new biomarker that helps clinical diagnostic and ther- apeutical design. 7.6 Limitations and Future directions 145 7.6 Limitations and Future directions There are also several limitations of the presented study. First, due to the limited size of our cohort, we merged subjects with high FCRP and subjects with histories of my- ocardial infarction into one single group - the MCI-h group. We hope to enroll more subjects in the future, and refine our vascular model. Second, due to the large age range within our cohort, we used linear regression to factor out the effect of age. The effect of age on brain anatomy in elderly subjects has been widely accepted, and we did observe a significant linear relationship between age the surface diffusion indices as shown in Fig.7.4. However, the use of linear regression does not rule out the possible existence of nonlinear relationship between age and brain anatomy. Third, here we only included the most representative univariate or multivariate measurements in our statistical analyses. Nonetheless, our methods can also be applied to other shape or diffusion measurements like thickness, mean diffusivity (MD), radial diffusivity (RD), and alike, as well as com- binations among these. Fourth, it would be desirable to include other factors, like gender, gene, education and ethnicity in future analyses to derive a more comprehensive model. In the future, we would also like to extend this method to to subdivide the CC into functionally or anatomically relevant regions, to further strengthen the interpretation of our results. For example, we could use a probabilistic tractography to the cortex or functional-based partition on the CC subregions [191], especially where showed signif- icant group differences, to investigate the association between regional CC alterations with disturbances in specific cortical domains. In addition, it would be important to track the mental status of patients, to see whether any of them transform into clinically diag- nosed dementia. This may provide additional insight into the contribution of the vascular 7.6 Limitations and Future directions 146 component in the conversion to Alzheimer’s disease or other types of dementia. Chapter 8 Conclusion and Outlook 8.1 Summary of Contributions This dissertation has focused on structural and diffusion MRI based shape, pose and connectivity analysis. As extensions to the established shape analysis on point distribu- tion models, the relative pose analysis and the joint shape and diffusion analysis have been developed. These methods have also been applied to characterize brain subcorti- cal alterations associated with diseases and disturbances, including prematurity, chronic waterborne manganese exposure, sports related mild traumatic brain injury, early onset blindness, as well as vascular disease and mild cognitive impairment. In these appli- cations, the proposed methods have been validated in subjects with a wide age range across the lifespan: from neonates, children, young adults, to the aging population. The shape, pose and the joint T1 and DTI analysis have successfully pinpointed anatomically meaningful brain changes in pre-clinical conditions, advancing the prognosis and early 8.1 Summary of Contributions 148 diagnosis of the associated diseases. The following publications have been made while completing this thesis: Journal publications Yi Lao, Binh Nguyen, Sinchai Tsao, Niharika Gajawelli, Meng Law, Helena Chui, Yalin Wang, and Natasha Lepore, A T1 and DTI fused 3D Corpus Callosum analysis in MCI subjects with high and low cardiovascular risk profile, In revision for NeuroImage:Clinical. Yi Lao, Laurie-Anne Dion, Gabriel Rocha, Guillaume Gilbert, Maryse Bouchard, Natasha Lepore, and Dave Saint-Amour, Mapping the impact of childhood manganese exposure on the basal ganglia, Sci- entific Reports, in revision. Yi Lao, Yue Kang, Olivier Collignon, Caroline Brun, Shadi Bohlool Kheibai, Flamine Alary, James Gee, Marvin D. Nelson, Franco Lepore, Natasha Lepore, A study of brain white matter plasticity in early blinds using Tract Based Spatial Statistics and Tract Statistical Analysis, NeuroReport, 2015, 26(18): 1151- 1154. Jie Shi, Olivier Collignon, Liang Xu, Gang Wang, Yue Kang, Franco Lepore, Yi Lao, Anand Joshi, Natasha Lepore, and Yalin Wang, Impact of early and late visual deprivation on the structure of the corpus callosum: A study Combining Thickness profile with Surface Tensor-based morphometry, Neuroinformat- ics 13.3 (2015): 321-336. Yi Lao, Yalin Wang, Jie Shi, Rafael Ceschin, Marvin D. Nelson, Ashok Panigrahy, and Natasha Lep- ore, Thalamic alterations in preterm neonates and its relation to ventral striatum disturbances revealed by a combined shape and pose analysis, Brain Structure and Function, 221(1), 487-506. 8.1 Summary of Contributions 149 Niharika Gajawelli, Yi Lao, Michael Apuzzo, Russ Romano, Charles Liu, Sinchai Tsao, Darryl Hwang, Bryce Wilkins, Natasha Lepore, Meng Law, Neuroimaging changes in the brain in Contact vs. Non-contact sport athletes using Diffusion Tensor Imaging, World neurosurgery, 2013, 80(6): 824-828. Jie Shi, YalinWang, Rafael Ceschin, Xing An, Yi Lao, Douglas Vanderbilt, Marvin D. Nelson, Ashok Panigrahy, Natasha Lepore, A multivariate surface-based analysis of the putamen in premature newborns, PlosOne, 2013, 8(7): e66736. Peer-reviewed conference papers Natasha Lepore and Fernando Yepes, Yi Lao, Ashok Panigrahy, Rafael Ceschin, Subhashree Ravichan- dran, Marvin D. Nelson, Pierre Fillard, Template-Based Tractography for Clinical Neonatal Diffusion Imaging Data, SPIE Medical Imaging, 4 - 9 February (2012), San Diego, USA, Oral Presentation. Yi Lao* and Fernando Yepes*, G Soria, P Fillard, A Planas, MD Nelson, C Justicia, N Lepormproving the CNR of High Spatial Resolution Small Animal Diffusion Tensor Imaging at 7T, International Confer- ence on Image Processing (ICIP), Orlando, FL, USA, September 20-October 3 (2012). Yi Lao, Jie Shi, Ashok Panigrahy, Rafeal Ceschin, Darryl Hwang, M.D. Nelson, Yalin Wang, and Natasha Lepore, Statistical analysis of relative pose of the thalamus in preterm neonates, MICCAI work- shop on Clinical Image-based Procedures: Translational Research in Medical Imaging, September 22, 2013, Nagoya, Japan. One of the finalists for best paper. Sinchai Tsao, Niharika Gajawelli, Peter A. Michels, Darryl Hwang, Yi Lao, Fernando Yepes, Vidya Rajagopalan, Meng Law, and Natasha LeporCA-based Multi-Fiber DWI Tractography in Neurosurgical Planning, MICCAI workshop on DTI challenge, September 22, 2013, Nagoya, Japan. N Gajawelli, S Tsao, D Hwang, Y Lao, F Yepes, V Rajagopalan, M Law, N Lepore , ICA-Based 8.1 Summary of Contributions 150 Diffusion Tensor Imaging in Neurosurgical Planning, MICCAI DTI Tractography Challenge, September 22, 2013, Nagoya, Japan. Liang Xu, Olivier Collignon, Gang Wang, Yue Kang, Franco Lepore, Jie Shi, Yi Lao, Anand Joshi, Natasha Lepore, Yalin Wang, Combining Thickness Information with Surface Tensor-based Morphometry for the 3D Statistical Analysis of the Corpus Callosum, 4th MICCAI workshop on Mathematical Founda- tions of Computational Anatomy (MFCA), September 22, 2013, Nagoya, Japan. Yi Lao, Niharika Gajawelli, Lauren Haas, Bryce Wilkins, Darryl Hwang, Sinchai Tsao, Yalin Wang, Meng Law, and Natasha LeporD Pre- vs. Post-Season Comparisons of Surface and Relative Pose of the Corpus Callosum in Contact Sport Athletes, SPIE Medical Imaging, 15 - 20 February (2014), San Diego, USA. Yi Lao, Meng Law, Jie Shi, Niharika Gajawelli, Lauren Haas, Yalin Wang, and Natasha Lepor T1 and DTI fused 3D Corpus Callosum analysis in pre- vs. post-season contact sports players. In Tenth International Symposium on Medical Information Processing and Analysis (SIPAIM 2014) (pp. 92870O- 92870O). International Society for Optics and Photonics. Jie Shi, Yalin Wang, Yi Lao, Rafael Ceschin, Liang Mi, Marvin D. Nelson, Ashok Panigrahy, Natasha Lepore. Abnormal ventricular development in preterm neonates with visually normal MRIs. In 11th International Symposium on Medical Information Processing and Analysis (SIPAIM 2015) (pp. 96810H- 96810H). International Society for Optics and Photonics. Fernando Yepes-C, Rebecca Johnson, Yi Lao, Darryl Hwang, Julie Coloigner, Felix Yap, Desai Bushan Philip Cheng, Inderbir Gill, Vinay Duddalwar, Natasha Lepore, The 3D EdgeRunner Pipeline: a novel shape-based analysis for neoplasms characterization, Proc. SPIE 9788, Medical Imaging 2016: Biomedical Applications in Molecular, Structural, and Functional Imaging, 97882N (March 29, 2016); doi:10.1117/12.2217238. 8.1 Summary of Contributions 151 Yaqiong Chai, Yi Lao, Yicen Li, Chaoran Ji, Sharon O’Neil, Yalin Wang, Natasha Lepore*, and John Wood*, Multivariate surface-based analysis of corpus callosum in patients with sickle cell dis- ease, SIPAIM, 12th International Symposium on Medical Information Processing and Analysis (SIPAIM), Tandil, Argentina. Conference abstracts Yi Lao, Pierre Fillard, Fernando Yepes, Ashok Panigrahy , Rafael Ceschin , Subhashree Ravichan- dran , Marvin D. Nelson , and Natasha Leporecovery of Developmental Marker Tracts from Low-resolution Clinical Neonatal DTI Data, Human Brain Mapping (OHBM), 10-14 June (2012). Yi Lao, Yue Kang, Caroline Brun, Franco Leporlivier Collignon, Flamine Alary, James Gee, Marvin D. Nelson, Natasha LeporBSS vs. TSA: Comparison on White Matter Structure, the Saban research insti- tute 17th annual poster session, June 4, 2012. Yi Lao, Caroline Brun, Yue Kang, Franco Leporlivier Collignon, Flamine Alary, James Gee, Marvin D. Nelson, Natasha Leporifferences between TBSS and TSA: a study of brain white matter plasticity in early blinds, Society for Neuroscience, 13-17 October (2012). Liang Xu, Olivier Collignon, Gang Wang, Yue Kang, Franco Lepore, Jie Shi, Yi Lao, Anand Joshi, Yalin Wang, and Natasha Lepore, 3D anatomy of the corpus callosum in early- and late-blind subjects from surface multivariate tensor-based morphometry, cognitive neuroscience society (CNS), 13-16 April (2013). Yi Lao, Laurie Anne Dion, Fernando Yepes, Guillaume Gilbert, Maryse Bouchard, Dave Saint Amour, and Natasha Leporract Specific Analysis reveals the impact of childhood manganese exposure on the cor- pus callosum, ISMRM 2013, 20-26 April (2013). 8.1 Summary of Contributions 152 LA Dion, Y Lao, N Lepore, F Yepes, G Gilbert, M Bouchard, D Saint-Amour, Diffusion tensor imaging correlates of memory performance in children exposed to manganese through drinking water, Federation of European Neuroscience (FENS), October 13-17, 2013. Yi Lao, Yalin Wang, Jie Shi, Rafael Ceschin, Marvin D. Nelson, Ashok Panigrahy, and Natasha Lepor 3D surface based correlation analysis of the putamen and thalamus in premature neonates, ISMRM 2014, 10-16 May, Milan, Italy. Yaqiong Chai, Mary Nelson, Yi Lao, Natasha Leporrack Specific Analysis on Children with Brain Tumors Treated with Surgery and Chemotherapy, OHBM 2014, 8-12 June, Hamburg, Germany. Yi Lao, Niharika Gajawelli, Lauren Hass, Darryl Hwang, Sinchai Tsao, Yalin Wang, Meng Law and Natasha Lepore, Joint diffusion and structural analysis of brain MRI data, 12th annual congress of the SBMT, 6-8 March, Los Angeles, USA. Invited Abstract. Yi Lao, Laurie-Anne Dion, Gabriel Rocha, Yalin Wang, Guillaume Gilbert, Maryse Bouchard, Natasha Lepore and Dave Saint-Amour, 3D Surface Analysis of the basal ganglia in children exposed to manganese from drinking water, OHBM 2015, 14-18 June, Hawaii, USA. J Shi, O Collignon, L Xu, G Wang, Y Kang, F Lepore, Y Lao, AA Joshi, Y Wang and N Lepore, Effects of Visual Deprivation on the Corpus Callosum Shape Morphometry, OHBM 2015, 14-18 June, Hawaii, USA. J Shi, Y Wang, Y Lao, R Ceschin, N Gajawelli, MD Nelson, A Panigrahy, N Lepore, Surface based analysis of the ventricles in premature newborns, OHBM 2015, 14-18 June, Hawaii, USA. Yi Lao, Binh Nguyen, Sinchai Tsao, Niharika Gajawelli, Meng Law, Helena Chui, Yalin Wang, and 8.2 Future Work 153 Natasha Lepore, A T1 and DTI fused 3D Corpus Callosum analysis in MCI subjects with high and low cardiovascular risk profile, ISMRM 2016, 07-13 May, Singapore. 8.2 Future Work The majority of the analyses in this dissertation are limited to identifying local anatomical alterations. Although structural-wise correlations between several subcortical structures have been investigated, the patterns of how the brain structures/regions vary with each other in the diseases investigated has not been thoroughly explored. The rapid growth of multi-modal MRI facilitates reliable extraction of structural connections and functional co-activations. These measures probe into the connections between distal brain regions and would be invaluable in deciphering the brain communication system across the lifes- pan. Brain connectivity analysis based on these measures will be an interesting future direction to work on, and would undoubtedly enrich our understanding of how brain networks are disrupted and reorganized under external insults. References [lin] http://diagramreview.com/brain-anatomy-and-diagram/brain-anatomy-detail/. [2] Abernethy, L. J., Cooke, R. W., and Foulder-Hughes, L. (2004). Caudate and hip- pocampal volumes, intelligence, and motor impairment in 7-year-old children who were born preterm. Pediatricresearch, 55(5):884–893. [3] Aboitiz, F., Scheibel, A. B., Fisher, R. S., and Zaidel, E. (1992). Fiber composition of the human corpus callosum. Brainresearch, 598(1):143–153. [4] Addis, D. R., Wong, A. T., and Schacter, D. L. (2007). Remembering the past and imagining the future: Common and distinct neural substrates during event construc- tion and elaboration. Neuropsychologia, 45(7):1363–1377. [5] Aggleton, J. P. and Brown, M. W. (1999). Episodic memory, amnesia, and the hippocampal-anterior thalamic axis. BehavioralandBrainSciences, 22(3):425–444. [6] Aggleton, J. P., O’Mara, S. M., Vann, S. D., Wright, N. F., Tsanov, M., and Erichsen, J. T. (2010). Hippocampal–anterior thalamic pathways for memory: uncovering a net- work of direct and indirect actions. EuropeanJournalofNeuroscience, 31(12):2292– 2307. [7] Alexander, G. E., Crutcher, M. D., and DeLong, M. R. (1991). Basal ganglia- thalamocortical circuits: parallel substrates for motor, oculomotor,“prefrontal” and “limbic” functions. Progressinbrainresearch, 85:119–146. [8] Allin, M., Henderson, M., Suckling, J., Nosarti, C., Rushe, T., Fearon, P., Stewart, A. L., Bullmore, E., Rifkin, L., and Murray, R. (2004). Effects of very low birth- weight on brain structure in adulthood. DevelopmentalMedicine&ChildNeurology, 46(1):46–53. [9] Alosco, M. L., Gunstad, J., Jerskey, B. A., Xu, X., Clark, U. S., Hassenstab, J., Cote, D. M., Walsh, E. G., Labbe, D. R., Hoge, R., et al. (2013). The adverse effects References 155 of reduced cerebral perfusion on cognition and brain structure in older adults with cardiovascular disease. Brainandbehavior, 3(6):626–636. [10] Andersen, M. E., Gearhart, J. M., and Clewell III, H. (1999). Pharmacokinetic data needs to support risk assessments for inhaled and ingested manganese. Neurotoxicol- ogy, 20(2):161–172. [11] Ardekani, B. A., Bachman, A. H., Figarsky, K., and Sidtis, J. J. (2014). Corpus callosum shape changes in early alzheimer’s disease: an mri study using the oasis brain database. BrainStructureandFunction, 219(1):343–352. [12] Arsigny, V ., Commowick, O., Pennec, X., and Ayache, N. (2006a). A log-euclidean framework for statistics on diffeomorphisms. In Medical Image Computing and Computer-AssistedIntervention–MICCAI2006, pages 924–931. Springer. [13] Arsigny, V ., Fillard, P., Pennec, X., and Ayache, N. (2006b). Log-euclidean metrics for fast and simple calculus on diffusion tensors. Magnetic resonance in medicine, 56(2):411–421. [14] Aschner, J. L., Anderson, A., Slaughter, J. C., Aschner, M., Steele, S., Beller, A., Mouvery, A., Furlong, H. M., and Maitre, N. L. (2015). Neuroimaging identifies in- creased manganese deposition in infants receiving parenteral nutrition. TheAmerican journalofclinicalnutrition, 102(6):1482–1489. [15] Bachman, A. H., Lee, S. H., Sidtis, J. J., and Ardekani, B. A. (2014). Corpus callosum shape and size changes in early alzheimer’s disease: A longitudinal mri study using the oasis brain database. JournalofAlzheimer’sDisease, 39(1):71–78. [16] Ball, G., Boardman, J. P., Rueckert, D., Aljabar, P., Arichi, T., Merchant, N., Gou- sias, I. S., Edwards, A. D., and Counsell, S. J. (2012). The effect of preterm birth on thalamic and cortical development. CerebralCortex, 22(5):1016–1024. [17] Bammer, R. (2003). Basic principles of diffusion-weighted imaging. European journalofradiology, 45(3):169–184. [18] Barnes, J., Carmichael, O. T., Leung, K. K., Schwarz, C., Ridgway, G. R., Bartlett, J. W., Malone, I. B., Schott, J. M., Rossor, M. N., Biessels, G. J., et al. (2013). Vascular and alzheimer’s disease markers independently predict brain atrophy rate in alzheimer’s disease neuroimaging initiative controls. Neurobiology of aging, 34(8):1996–2002. References 156 [19] Basser, P. J., Mattiello, J., and LeBihan, D. (1994). Mr diffusion tensor spectroscopy and imaging. Biophysicaljournal, 66(1):259. [20] Bazarian, J. J., Zhu, T., Blyth, B., Borrino, A. o., and Zhong, J. (2012). Subject- specific changes in brain white matter on diffusion tensor ima ging after sports-related concussion. Magneticresonanceimaging, 30(2):171–180. [21] Bechara, A., Tranel, D., and Damasio, H. (2000). Characterization of the decision-making deficit of patients with ventromedial prefrontal cortex lesions. Brain, 123(11):2189–2202. [22] Benjamini, Y . and Hochberg, Y . (1995). Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the royal statistical society.SeriesB(Methodological), pages 289–300. [23] Bhutta, A. T., Cleves, M. A., Casey, P. H., Cradock, M. M., and Anand, K. (2002). Cognitive and behavioral outcomes of school-aged children who were born preterm. JAMA:thejournaloftheAmericanMedicalAssociation, 288(6):728–737. [24] Boardman, J. P., Counsell, S. J., Rueckert, D., Olga, K., Bhatia, K. K., Aljabar, P., Hajnal, J., Allsop, J. M., Rutherford, M. A., and Edwards, A. D. (2006). Abnormal deep grey matter development following preterm birth detected using deformation- based morphometry. Neuroimage, 32(1):70–78. [25] Boardman, J. P., Craven, C., Valappil, S., Counsell, S. J., Dyet, L., Rueckert, D., Aljabar, P., Rutherford, M. A., Chew, A., Allsop, J. M., et al. (2010). A common neonatal image phenotype predicts adverse neurodevelopmental outcome in children born preterm. Neuroimage, 52(2):409–414. [26] Bonner-Jackson, A., Mahmoud, S., Miller, J., and Banks, S. J. (2015). Verbal and non-verbal memory and hippocampal volumes in a memory clinic population. Alzheimer’sresearch&therapy, 7(1):1–10. [27] Bossa, M., Zacur, E., and Olmos, S. (2011). Statistical analysis of relative pose information of subcortical nuclei: application on adni data. Neuroimage, 55(3):999– 1008. [28] Bossa, M. N. and Olmos, S. (2006). Statistical model of similarity transforma- tions: Building a multi-object pose model of brain structures. In IEEE Comput. Soc. WorkshopMath.MethodsBiomed.ImageAnal, page 59. References 157 [29] Bouchard, M. (2007). Hair manganese and hyperactive behaviors: Pilot study of school-age children exposed through tap water. Epidemiology, 18(5):S164–S165. [30] Bouchard, M. F., Sauvé, S., Barbeau, B., Legrand, M., Brodeur, M.-È., Bouffard, T., Limoges, E., Bellinger, D. C., and Mergler, D. (2011). Intellectual impairment in school-age children exposed to manganese from drinking water. Environmentalhealth perspectives, 119(1):138. [31] Büchel, C., Price, C., Frackowiak, R., and Friston, K. (1998). Different activation patterns in the visual cortex of late and congenitally blind subjects.Brain, 121(3):409– 419. [32] Buchmayer, S., Johansson, S., Johansson, A., Hultman, C. M., Sparén, P., and Cnat- tingius, S. (2009). Can association between preterm birth and autism be explained by maternal or neonatal morbidity? Pediatrics, 124(5):e817–e825. [33] Budil, D. E., Lee, S., Saxena, S., and Freed, J. H. (1996). Nonlinear-least- squares analysis of slow-motion epr spectra in one and two dimensions using a mod- ified levenberg–marquardt algorithm. Journal of Magnetic Resonance, Series A, 120(2):155–189. [34] Buklina, S. (2005). The corpus callosum, interhemisphere interactions, and the function of the right hemisphere of the brain. Neuroscience and behavioral physiol- ogy, 35(5):473–480. [35] Byne, W., Buchsbaum, M. S., Mattiace, L. A., Hazlett, E. A., Kemether, E., El- hakem, S. L., Purohit, D. P., Haroutunian, V ., and Jones, L. (2002). Postmortem assessment of thalamic nuclear volumes in subjects with schizophrenia. American JournalofPsychiatry, 159(1):59–65. [36] Caeyenberghs, K., Leemans, A., Coxon, J., Leunissen, I., Drijkoningen, D., Geurts, M., Gooijers, J., Michiels, K., Sunaert, S., and Swinnen, S. P. (2011). Bimanual coordination and corpus callosum microstructure in young adults with traumatic brain injury: a diffusion tensor imaging study. Journalofneurotrauma, 28(6):897–913. [37] Caravale, B., Tozzi, C., Albino, G., and Vicari, S. (2005). Cognitive development in low risk preterm infants at 3–4 years of life. Archives of Disease in Childhood-Fetal andNeonatalEdition, 90(6):F474–F479. [38] Cardenas, V . A., Reed, B., Chao, L. L., Chui, H., Sanossian, N., DeCarli, C. C., Mack, W., Kramer, J., Hodis, H. N., Yan, M., et al. (2012). Associations among References 158 vascular risk factors, carotid atherosclerosis, and cortical volume and thickness in older adults. Stroke, 43(11):2865–2870. [39] Carlsson, M. and Carlsson, A. (1990). Interactions between glutamatergic and monoaminergic systems within the basal ganglia-implications for schizophrenia and parkinson’s disease. Trendsinneurosciences, 13(7):272–276. [40] Casanova, M. F., El-Baz, A. S., Kamat, S. S., Dombroski, B. A., Khalifa, F., El- nakib, A., Soliman, A., Allison-McNutt, A., and Switala, A. E. (2013). Focal corti- cal dysplasias in autism spectrum disorders. Actaneuropathologicacommunications, 1(1):67. [41] Chang, Y ., Jin, S.-U., Kim, Y ., Shin, K. M., Lee, H. J., Kim, S. H., Ahn, J.-H., Park, S.-J., Jeong, K. S., Weon, Y . C., et al. (2013). Decreased brain volumes in manganese-exposed welders. Neurotoxicology, 37:182–189. [42] Chang, Y ., Lee, J.-J., Seo, J.-H., Song, H.-J., Kim, J.-H., Bae, S.-J., Ahn, J.-H., Park, S.-J., Jeong, K. S., Kwon, Y . J., et al. (2010a). Altered working memory process in the manganese-exposed brain. Neuroimage, 53(4):1279–1285. [43] Chang, Y ., Song, H.-J., Lee, J.-J., Seo, J.-H., Kim, J.-H., Lee, H. J., Kim, H. J., Kim, Y ., Ahn, J.-H., Park, S.-J., et al. (2010b). Neuroplastic changes within the brains of manganese-exposed welders: recruiting additional neural resources for successful motor performance. Occupationalandenvironmentalmedicine. [44] Cheon, K.-A., Kim, Y .-S., Oh, S.-H., Park, S.-Y ., Yoon, H.-W., Herrington, J., Nair, A., Koh, Y .-J., Jang, D.-P., Kim, Y .-B., et al. (2011). Involvement of the anterior tha- lamic radiation in boys with high functioning autism spectrum disorders: a diffusion tensor imaging study. Brainresearch, 1417:77–86. [45] Chételat, G., Landeau, B., Salmon, E., Yakushev, I., Bahri, M. A., Mézenge, F., Perrotin, A., Bastin, C., Manrique, A., Scheurich, A., et al. (2013). Relationships between brain metabolism decrease in normal aging and changes in structural and functional connectivity. Neuroimage, 76:167–177. [46] Chui, H. C. (2006). Vascular cognitive impairment: today and tomorrow. Alzheimer’s&Dementia, 2(3):185–194. [47] Chung, M. K., Robbins, S. M., Dalton, K. M., Davidson, R. J., Alexander, A. L., and Evans, A. C. (2005). Cortical thickness analysis in autism with heat kernel smoothing. NeuroImage, 25(4):1256–1265. References 159 [48] Coleman, P. D. and Flood, D. G. (1987). Neuron numbers and dendritic extent in normal aging and alzheimer’s disease. Neurobiologyofaging, 8(6):521–545. [49] Counsell, S. J., Allsop, J. M., Harrison, M. C., Larkman, D. J., Kennea, N. L., Kapellou, O., Cowan, F. M., Hajnal, J. V ., Edwards, A. D., and Rutherford, M. A. (2003). Diffusion-weighted imaging of the brain in preterm infants with focal and diffuse white matter abnormality. Pediatrics, 112(1):1–7. [50] Counsell, S. J., Shen, Y ., Boardman, J. P., Larkman, D. J., Kapellou, O., Ward, P., Allsop, J. M., Cowan, F. M., Hajnal, J. V ., Edwards, A. D., et al. (2006). Axial and radial diffusivity in preterm infants who have diffuse white matter changes on magnetic resonance imaging at term-equivalent age. Pediatrics, 117(2):376–386. [51] Criswell, S., Perlmutter, J., Videen, T., Moerlein, S., Flores, H., Birke, A., and Racette, B. (2011). Reduced uptake of [18f] fdopa pet in asymptomatic welders with occupational manganese exposure. Neurology, 76(15):1296–1301. [52] Criswell, S. R., Perlmutter, J. S., Crippin, J. S., Videen, T. O., Moerlein, S. M., Flores, H. P., Birke, A. M., and Racette, B. A. (2012a). Reduced uptake of fdopa pet in end-stage liver disease with elevated manganese levels. Archives of neurology, 69(3):394–397. [53] Criswell, S. R., Perlmutter, J. S., Huang, J. L., Golchin, N., Flores, H. P., Hob- son, A., Aschner, M., Erikson, K. M., Checkoway, H., and Racette, B. A. (2012b). Basal ganglia intensity indices and diffusion weighted imaging in manganese-exposed welders. Occupationalandenvironmentalmedicine, pages oemed–2011. [54] Crossgrove, J. and Zheng, W. (2004). Manganese toxicity upon overexposure.NMR inBiomedicine, 17(8):544–553. [55] Cummings, J. L. (1993). Frontal-subcortical circuits and human behavior. Archives ofneurology, 50(8):873. [56] Danos, P., Baumann, B., Bernstein, H.-G., Stauch, R., Krell, D., Falkai, P., and Bogerts, B. (2002). The ventral lateral posterior nucleus of the thalamus in schizophre- nia: a post-mortem study. PsychiatryResearch: Neuroimaging, 114(1):1–9. [57] Davie, C. A. (2008). A review of parkinson’s disease. British medical bulletin, 86(1):109–127. [58] de Bruijn, R. F. and Ikram, M. A. (2014). Cardiovascular risk factors and future risk of alzheimer’s disease. BMCmedicine, 12(1):130. References 160 [59] De Jong, L., Van Der Hiele, K., Veer, I., Houwing, J., Westendorp, R., Bollen, E., De Bruin, P., Middelkoop, H., Van Buchem, M., and Van Der Grond, J. (2008). Strongly reduced volumes of putamen and thalamus in alzheimer’s disease: an mri study. Brain, 131(12):3277–3285. [60] De Leon, M., George, A., Golomb, J., Tarshish, C., Convit, A., Kluger, A., De Santi, S., Mc Rae, T., Ferris, S., Reisberg, B., et al. (1997). Frequency of hippocampal formation atrophy in normal aging and alzheimer’s disease. Neurobiology of aging, 18(1):1–11. [61] Delano-Wood, L., Bondi, M. W., Jak, A. J., Horne, N. R., Schweinsburg, B. C., Frank, L. R., Wierenga, C. E., Delis, D. C., Theilmann, R. J., and Salmon, D. P. (2010). Stroke risk modifies regional white matter differences in mild cognitive impairment. Neurobiologyofaging, 31(10):1721–1731. [62] Di Paola, M., Di Iulio, F., Cherubini, A., Blundo, C., Casini, A., Sancesario, G., Passafiume, D., Caltagirone, C., and Spalletta, G. (2010). When, where, and how the corpus callosum changes in mci and ad a multimodal mri study. Neurology, 74(14):1136–1142. [63] Dimitra, S., Verganelakis, D., Gotsis, E., Toulas, P., Papatriantafillou, J., Kara- georgiou, C., Thomaides, T., Kapsalaki, E., Hadjigeorgiou, G., and Papadimitriou, A. (2013). Diffusion tensor imaging (dti) in the detection of white matter lesions in pa- tients with mild cognitive impairment (mci). Acta Neurologica Belgica, 113(4):441– 451. [64] Dion, L.-A., Bouchard, M. F., Sauvé, S., Barbeau, B., Tucholka, A., Major, P., Gilbert, G., Mergler, D., and Saint-Amour, D. (2016). Mri pallidal signal in children exposed to manganese in drinking water. NeuroToxicology. [65] Draganski, B., Kherif, F., Klöppel, S., Cook, P. A., Alexander, D. C., Parker, G. J., Deichmann, R., Ashburner, J., and Frackowiak, R. S. (2008). Evidence for segre- gated and integrative connectivity patterns in the human basal ganglia. TheJournalof Neuroscience, 28(28):7143–7152. [66] Drayer, B. P. (1988). Imaging of the aging brain. part i. normal findings. Radiology, 166(3):785–796. [67] Driscoll, I., Davatzikos, C., An, Y ., Wu, X., Shen, D., Kraut, M., and Resnick, S. (2009). Longitudinal pattern of regional brain volume change differentiates normal aging from mci. Neurology, 72(22):1906–1913. References 161 [68] Dryden, I. and Mardia, K. (1998). Statisticalanalysisofshape. Wiley. [69] Duda, J., Avants, B., Kim, J., Zhang, H., ˜ Patel, S., Whyte, J., and Gee, J. (2008). Multivariate analysis of thalamo-cortical connectivity loss in tbi. In Computer Vi- sion and Pattern Recognition Workshops, 2008. CVPRW’08. IEEE Computer Society Conferenceon, pages 1–8. IEEE. [70] Fell, J., Meadows, N., Khan, K., Long, S., Milla, P., Reynolds, A., Quaghebeur, G., and Taylor, W. (1996). Manganese toxicity in children receiving long-term parenteral nutrition. TheLancet, 347(9010):1218–1221. [71] Fily, A., Pierrat, V ., Delporte, V ., Breart, G., Truffert, P., et al. (2006). Factors associated with neurodevelopmental outcome at 2 years after very preterm birth: the population-based nord-pas-de-calais epipage cohort. Pediatrics, 117(2):357–366. [72] Fishman, R. S. (1997). Gordon holmes, the cortical retina, and the wounds of war. Documentaophthalmologica, 93(1-2):9–28. [73] Flamand, V . H., Nadeau, L., and Schneider, C. (2012). Brain motor excitability and visuomotor coordination in 8-year-old children born very preterm. Clinical Neuro- physiology, 123(6):1191–1199. [74] Fotenos, A. F., Snyder, A., Girton, L., Morris, J., and Buckner, R. (2005). Norma- tive estimates of cross-sectional and longitudinal brain volume decline in aging and ad. Neurology, 64(6):1032–1039. [75] Frederiksen, K. S., Garde, E., Skimminge, A., Ryberg, C., Rostrup, E., Baaré, W. F., Siebner, H. R., Hejl, A.-M., Leffers, A.-M., and Waldemar, G. (2011). Corpus callo- sum atrophy in patients with mild alzheimer’s disease. Neurodegenerative Diseases, 8(6):476–482. [76] Friedman, J. I., Tang, C. Y ., de Haas, H. J., Changchien, L., Goliasch, G., Dabas, P., Wang, V ., Fayad, Z. A., Fuster, V ., and Narula, J. (2014). Brain imaging changes associated with risk factors for cardiovascular and cerebrovascular disease in asymp- tomatic patients. JACC:CardiovascularImaging, 7(10):1039–1053. [77] Fujioka, M., Taoka, T., Hiramatsu, K.-I., Sakaguchi, S., and Sakaki, T. (1999). De- layed ischemic hyperintensity on t1-weighted mri in the caudoputamen and cerebral cortex of humans after spectacular shrinking deficit. Stroke, 30(5):1038–1042. References 162 [78] Gajawelli, N., Lao, Y ., Apuzzo, M. L., Romano, R., Liu, C., Tsao, S., Hwang, D., Wilkins, B., Lepore, N., and Law, M. (2013). Neuroimaging changes in the brain in contact versus noncontact sport athletes using diffusion tensor imaging. World neurosurgery, 80(6):824–828. [79] Garcia-Alloza, M., Gregory, J., Kuchibhotla, K. V ., Fine, S., Wei, Y ., Ayata, C., Frosch, M. P., Greenberg, S. M., and Bacskai, B. J. (2011). Cerebrovascular lesions induce transientβ-amyloid deposition. Brain, 134(12):3694–3704. [80] Gibb, W. (1989). Neuropathology in movement disorders. Journal of Neurology, Neurosurgery&Psychiatry, 52(Suppl):55–67. [81] Giedd, J. N. (2004). Structural magnetic resonance imaging of the adolescent brain. AnnalsoftheNewYorkAcademyofSciences, 1021(1):77–85. [82] Giedd, J. N., Blumenthal, J., Jeffries, N. O., Castellanos, F. X., Liu, H., Zijdenbos, A., Paus, T., Evans, A. C., and Rapoport, J. L. (1999). Brain development during childhood and adolescence: a longitudinal mri study.Natureneuroscience, 2(10):861– 863. [83] Gogtay, N., Giedd, J. N., Lusk, L., Hayashi, K. M., Greenstein, D., Vaituzis, A. C., Nugent, T. F., Herman, D. H., Clasen, L. S., Toga, A. W., et al. (2004). Dynamic mapping of human cortical development during childhood through early adulthood. Proceedings of the National Academy of Sciences of the United States of America, 101(21):8174–8179. [84] Golden, J. A. and Harding, B. N. (2010). Cortical malformations: unfolding polymicrogyria. NatureReviewsNeurology, 6(9):471–472. [85] Goldman-Rakic, P. S. (1987). Circuitry of primate prefrontal cortex and regulation of behavior by representational memory. ComprehensivePhysiology. [86] Gons, R. A., van Oudheusden, L. J., de Laat, K. F., van Norden, A. G., van Uden, I. W., Norris, D. G., Zwiers, M. P., van Dijk, E., and de Leeuw, F.-E. (2012). Hy- pertension is related to the microstructure of the corpus callosum: the run dmc study. JournalofAlzheimer’sDisease, 32(3):623–631. [87] Gorczowski, K., Styner, M., Jeong, J. Y ., Marron, J., Piven, J., Hazlett, H. C., Pizer, S. M., and Gerig, G. (2010). Multi-object analysis of volume, pose, and shape using statistical discrimination. Pattern Analysis and Machine Intelligence, IEEE Transac- tionson, 32(4):652–661. References 163 [88] Gorelick, P. B., Scuteri, A., Black, S. E., DeCarli, C., Greenberg, S. M., Iadecola, C., Launer, L. J., Laurent, S., Lopez, O. L., Nyenhuis, D., et al. (2011). Vascular contributions to cognitive impairment and dementia a statement for healthcare pro- fessionals from the american heart association/american stroke association. Stroke, 42(9):2672–2713. [89] Graham, N., Emery, T., and Hodges, J. (2004). Distinctive cognitive profiles in alzheimer’s disease and subcortical vascular dementia. Journal of Neurology, Neuro- surgery&Psychiatry, 75(1):61–71. [90] Granacher Jr, R. P. (2007). Traumaticbraininjury: Methodsforclinicalandforen- sicneuropsychiatricassessment. CRC Press. [91] Gray, H. and Lewis, W. (1918). Anatomy of the human body. 20th.LeaandFebiger. [92] Greenfield, S., Association, R. D., et al. (1996). The human mind explained: An owner’sguidetothemysteriesofthemind. Henry Holt. [93] Greger, J. (1998). Nutrition versus toxicology of manganese in humans: evaluation of potential biomarkers. Neurotoxicology, 20(2-3):205–212. [94] Groenewegen, H. J. and Berendse, H. W. (1994). The specificity of the ‘nonspe- cific’midline and intralaminar thalamic nuclei. Trendsinneurosciences, 17(2):52–57. [95] Grunau, R. E., Whitfield, M. F., and Davis, C. (2002). Pattern of learning disabilities in children with extremely low birth weight and broadly average intelligence. Archives ofpediatrics&adolescentmedicine, 156(6):615. [96] Gur, R. E., Maany, V ., Mozley, P. D., Swanson, C., Bilker, W., and Gur, R. C. (1998). Subcortical mri volumes in neuroleptic-naive and treated patients with schizophrenia. AmericanJournalofPsychiatry, 155(12):1711–1717. [97] Haber, S. N. (2003). The primate basal ganglia: parallel and integrative networks. Journalofchemicalneuroanatomy, 26(4):317–330. [98] Hallam, B. J., Brown, W. S., Ross, C., Buckwalter, J. G., Bigler, E. D., Tschanz, J. T., Norton, M. C., Welsh-Bohmer, K. A., and Breitner, J. C. (2008). Regional atrophy of the corpus callosum in dementia. Journal of the International Neuropsy- chologicalSociety, 14(03):414–423. References 164 [99] Hart, A. R., Whitby, E. W., Griffiths, P. D., and Smith, M. F. (2008). Magnetic resonance imaging and developmental outcome following preterm birth: review of current evidence. DevelopmentalMedicine&ChildNeurology, 50(9):655–663. [100] Hasegawa, T., Yamada, K., Morimoto, M., Morioka, S., Tozawa, T., Isoda, K., Murakami, A., Chiyonobu, T., Tokuda, S., Nishimura, A., et al. (2011). Develop- ment of corpus callosum in preterm infants is affected by the prematurity: in vivo assessment of diffusion tensor imaging at term-equivalent age. Pediatric research, 69(3):249–254. [101] Hattie, J. (2008). Visiblelearning: Asynthesisofover800meta-analysesrelating toachievement. Routledge. [102] Hauser, R. A., Zesiewicz, T. A., Rosemurgy, A. S., Martinez, C., and Olanow, C. (1994). Manganese intoxication and chronic liver failure. Annals of neurology, 36(6):871–875. [103] Hayden, K., Warren, L., Pieper, C., Østbye, T., Tschanz, J., Norton, M., Breitner, J., and Welsh-Bohmer, K. (2005). Identification of vad and ad prodromes: the cache county study. Alzheimer’s&Dementia, 1(1):19–29. [104] Haynes, E. N., Sucharew, H., Kuhnell, P., Alden, J. y., Barnas, M., Wright, R. O., Parsons, P. J., Aldous, K. e. M., Praamsma, M. L., Beidler, C., et al. (2015). Man- ganese exposure and neurocognitive outcomes in rural school-age chi ldren: the com- munities actively researching exposure study (ohio, usa). Environmental health per- spectives, 123(10):1066. [105] Hoffmann, M. (2013). The human frontal lobes and frontal network systems: an evolutionary, clinical, and treatment perspective. ISRNneurology, 2013. [106] Holli, K. K., Harrison, L., Dastidar, P., Wäljas, M., Liimatainen, S., Luukkaala, T., Öhman, J., Soimakallio, S., and Eskola, H. (2010). Texture analysis of mr images of patients with mild traumatic brain injury. BMCmedicalimaging, 10(1):8. [107] Hotelling, H. (1931). The generalization of student’s ratio. TheAnnalsofMathe- maticalStatistics, 2(3):360–378. [108] Hu, Z., Wu, L., Jia, J., and Han, Y . (2014). Advances in longitudinal studies of amnestic mild cognitive impairment and alzheimer’s disease based on multi-modal mri techniques. Neurosciencebulletin, 30(2):198–206. References 165 [109] Husain, M. and Nachev, P. (2007). Space and the parietal cortex. Trendsincogni- tivesciences, 11(1):30–36. [110] Hutsler, J. and Galuske, R. A. (2003). Hemispheric asymmetries in cerebral corti- cal networks. Trendsinneurosciences, 26(8):429–435. [111] Iadecola, C. (2013). The pathobiology of vascular dementia. Neuron, 80(4):844– 866. [112] Ikeda, S., Yamaguchi, Y ., Sera, Y ., Ohshiro, H., Uchino, S., Yamashita, Y ., and Ogawa, M. (2000). Manganese deposition in the globus pallidus in patients with biliary atresia. Transplantation, 69(11):2339–2343. [113] Inder, T. E., Warfield, S. K., Wang, H., Hüppi, P. S., and V olpe, J. J. (2005). Abnor- mal cerebral structure is present at term in premature infants. Pediatrics, 115(2):286– 294. [114] Ingles, J. L., Boulton, D. C., Fisk, J. D., and Rockwood, K. (2007). Preclinical vascular cognitive impairment and alzheimer disease neuropsychological test perfor- mance 5 years before diagnosis. Stroke, 38(4):1148–1153. [115] Isaacs, E. B., Lucas, A., Chong, W. K., Wood, S. J., Johnson, C. L., Marshall, C., Vargha-Khadem, F., and Gadian, D. G. (2000). Hippocampal volume and everyday memory in children of very low birth weight. Pediatricresearch, 47(6):713–720. [116] Jenkinson, M., Bannister, P., Brady, M., and Smith, S. (2002). Improved opti- mization for the robust and accurate linear registration and motion correction of brain images. Neuroimage, 17(2):825–841. [117] Jenkinson, M., Beckmann, C. F., Behrens, T. E., Woolrich, M. W., and Smith, S. M. (2012). Fsl. Neuroimage, 62(2):782–790. [118] Jernigan, T. L. and Tallal, P. (1990). Late childhood changes in brain morphology observable with mri. DevelopmentalMedicine&ChildNeurology, 32(5):379–385. [119] Jiang, Y ., Zheng, W., Long, L., Zhao, W., L ˜ i, X., Mo, X., Lu, J., Fu, X., Li, W., Liu, S. o., et al. (2007). Brain magnetic resonance imaging and manganese concentrations in red b lood cells of smelting workers: search for biomarkers of manganese exposure. Neurotoxicology, 28(1):126–135. References 166 [120] Jochemsen, H. M., Muller, M., Visseren, F. L., Scheltens, P., Vincken, K. L., Mali, W. P., van der Graaf, Y ., and Geerlings, M. I. (2013). Blood pressure and progression of brain atrophy: the smart-mr study. JAMAneurology, 70(8):1046–1053. [121] Joseph, R. M. and Tager-Flusberg, H. (2004). The relationship of theory of mind and executive functions to symptom type and severity in children with autism. Devel- opmentandpsychopathology, 16(01):137–155. [122] Kafritsa, Y ., Fell, J., Long, S., Bynevelt, M., Taylor, W., and Milla, P. (1998). Long term outcome of brain manganese deposition in patients on home parenteral nutrition. Archivesofdiseaseinchildhood, 79(3):263–265. [123] Kalaria, R. N., Akinyemi, R., and Ihara, M. (2012). Does vascular pathology contribute to alzheimer changes? Journal of the neurological sciences, 322(1):141– 147. [124] Kantarci, K., Petersen, R. C., Boeve, B. F., Knopman, D. S., Weigand, S. D., O’Brien, P. C., Shiung, M. M., Smith, G. E., Ivnik, R. J., Tangalos, E. G., et al. (2005). Dwi predicts future progression to alzheimer disease in amnestic mild cogni- tive impairment. Neurology, 64(5):902–904. [125] Karas, G., Sluimer, J., Goekoop, R., Van Der Flier, W., Rombouts, S., Vrenken, H., Scheltens, P., Fox, N., and Barkhof, F. (2008). Amnestic mild cognitive impair- ment: structural mr imaging findings predictive of conversion to alzheimer disease. AmericanJournalofNeuroradiology, 29(5):944–949. [126] Karcher, H. (1977). Riemannian center of mass and mollifier smoothing. Commu- nicationsonpureandappliedmathematics, 30(5):509–541. [127] Kawagoe, R., Takikawa, Y ., and Hikosaka, O. (1998). Expectation of reward modulates cognitive signals in the basal ganglia. Natureneuroscience, 1(5):411–416. [128] Keen, C., Ensunsa, J., and Clegg, M. (1999). Manganese metabolism in animals and humans including the toxicity of manganese. Metal ions in biological systems, 37:89–121. [129] Keen, C. L. (1986). The effect of age on manganese uptake and retention from milk and infant formulas in rats. J.Nutr, 116(3):395–402. [130] Kesler, S. R., Ment, L. R., V ohr, B., Pajot, S. K., Schneider, K. C., Katz, K. H., Ebbitt, T. B., Duncan, C. C., Makuch, R. W., and Reiss, A. L. (2004). V olumetric References 167 analysis of regional cerebral development in preterm children. Pediatric neurology, 31(5):318–325. [131] Kesler, S. R., Reiss, A. L., V ohr, B., Watson, C., Schneider, K. C., Katz, K. H., Maller-Kesselman, J., Silbereis, J., Constable, R. T., Makuch, R. W., et al. (2008). Brain volume reductions within multiple cognitive systems in male preterm children at age twelve. TheJournalofpediatrics, 152(4):513–520. [132] Khan, K., Factor-Litvak, P., Wasserman, G. A., Liu, X., Ahmed, E., Parvez, F., Slavkovich, V ., Levy, D., Mey, J., van Geen, A., et al. (2011). Manganese exposure from drinking water and children’s classroom behavior in bangladesh. Environmental healthperspectives, 119(10):1501. [133] Khan, K., Wasserman, G. A., Liu, X., Ahmed, E., Parvez, F., Slavkovich, V ., Levy, D., Mey, J., van Geen, A., Graziano, J. H., et al. (2012). Manganese exposure from drinking water and children’s academic achievement. Neurotoxicology, 33(1):91–97. [134] Kim, S.-Y ., Chung, K., and Deisseroth, K. (2013). Light microscopy mapping of connections in the intact brain. Trendsincognitivesciences, 17(12):596–599. [135] Kim, Y ., Jeong, K. S., Song, H.-J., Lee, J.-J., Seo, J.-H., Kim, G.-C., Lee, H. J., Kim, H. J., Ahn, J.-H., Park, S.-J., et al. (2011). Altered white matter microstructural integrity revealed by voxel-wise analysis of diffusion tensor imaging in welders with manganese exposure. Neurotoxicology, 32(1):100–109. [136] Kim, Y ., Park, J. K., Choi, Y ., Yoo, C.-I., Lee, C. R., Lee, H., Lee, J.-H., Kim, S.-R., Jeong, T.-H., Yoon, C. S., et al. (2005). Blood manganese concentration is elevated in iron deficiency anemia patients, whereas globus pallidus signal intensity is minimally affected. Neurotoxicology, 26(1):107–111. [137] Klein, J. C., Rushworth, M. F., Behrens, T. E., Mackay, C. E., de Crespigny, A. J., D’Arceuil, H., and Johansen-Berg, H. (2010). Topography of connections between human prefrontal cortex and mediodorsal thalamus studied with diffusion tractogra- phy. Neuroimage, 51(2):555–564. [138] Kloppenborg, R., Nederkoorn, P., Grool, A., Vincken, K., Mali, W., Vermeulen, M., van der Graaf, Y ., Geerlings, M., Group, S. S., et al. (2012). Cerebral small-vessel disease and progression of brain atrophy the smart-mr study. Neurology, 79(20):2029– 2036. References 168 [139] Knyazeva, M. G. (2013). Splenium of corpus callosum: patterns of interhemi- spheric interaction in children and adults. Neuralplasticity, 2013. [140] Koay, C. G., Chang, L.-C., Carew, J. D., Pierpaoli, C., and Basser, P. J. (2006). A unifying theoretical and algorithmic framework for least squares methods of estima- tion in diffusion tensor imaging. JournalofMagneticResonance, 182(1):115–125. [141] Kostovic, I. and Goldman-Rakic, P. S. (1983). Transient cholinesterase staining in the mediodorsal nucleus of the thalamus and its connections in the developing human and monkey brain. JournalofComparativeNeurology, 219(4):431–447. [142] Kostovi´ c, I., Judaš, M., and Sedmak, G. (2011). Developmental history of the subplate zone, subplate neurons and interstitial white matter neurons: relevance for schizophrenia. InternationalJournalofDevelopmentalNeuroscience, 29(3):193–205. [143] Kostovi´ c, I., Lukinovi´ c, N., Judaš, M., Bogdanovi´ c, N., Mrzljak, L., Zeˇ cevi´ c, N., and Kubat, M. (1989). Structural basis of the developmental plasticity in the human cerebral cortex: the role of the transient subplate zone. Metabolic brain disease, 4(1):17–23. [144] Kraus, M. F., Susmaras, T., Caughlin, B. P., Walker, C. J., Sweeney, J. A., and Little, D. M. (2007). White matter integrity and cognition in chronic traumatic brain injury: a diffusion tensor imaging study. Brain, 130(10):2508–2519. [145] Kretschmann, H.-J., Weinrich, W., and Fiekert, W. (1998). Neurofunctional sys- tems: 3Dreconstructionswithcorrelatedneuroimaging. Univ of California Press. [146] Krupa, K. and Bekiesinska-Figatowska, M. (2013). Congenital and acquired ab- normalities of the corpus callosum: A pictorial essay. BioMedresearchinternational, 2013. [Lao et al.] Lao, Y ., Gajawelli, N., Haas, L., Wilkins, B., Hwang, D., Tsao, S., Wang, Y ., Law, M., and Leporé, N. 3d pre-vs. post-season comparisons of surface and relative pose of the corpus callosum in contact sport athletes. In Proc. of SPIE Vol, volume 9034, pages 90344E–1. [148] Lao, Y ., Law, M., Shi, J., Gajawelli, N., Haas, L., Wang, Y ., and Leporé, N. (2015). A t1 and dti fused 3d corpus callosum analysis in pre-vs. post-season contact sports players. In Tenth International Symposium on Medical Information Processing and Analysis, pages 92870O–92870O. International Society for Optics and Photonics. References 169 [149] Lao, Y ., Shi, J., Wang, Y ., Ceschin, R., Hwang, D., Nelson, M., Panigrahy, A., and Lepore, N. (2013). Statistical analysis of relative pose of the thalamus in preterm neonates. MICCAIworkshoponClinicalImage-basedProcedures. [150] Lao, Y ., Wang, Y ., Shi, J., Ceschin, R., Nelson, M. D., Panigrahy, A., and Leporé, N. (2014). Thalamic alterations in preterm neonates and their relation to ventral stria- tum disturbances revealed by a combined shape and pose analysis. Brain Structure andFunction, pages 1–20. [151] Lebby, P. C. (2013). Brain Imaging: A Guide for Clinicians. Oxford University Press. [152] Lee, A., Lepore, N., Lepore, F., Alary, F., V oss, P., Chou, Y ., Brun, C., Barysheva, M., Toga, A., and PM, T. (2007). Brain differences visualized in the blind using tensor manifold statistics and diffusion tensor imaging. Frontiers in the convergence ofBioscienceandInformationTechnologies(FBIT),Jeju-Do,Korea,October11-13. [153] Lepore, F., Brun, C., Chou, Y .-Y ., Chiang, M.-C., Dutton, R. A., Hayashi, K. M., Luders, E., Lopez, O. L., Aizenstein, H. J., Toga, A. W., et al. (2008a). Generalized tensor-based morphometry of hiv/aids using multivariate statistics on deformation ten- sors. MedicalImaging,IEEETransactionson, 27(1):129–141. [154] Lepore, N., Brun, C., Chou, Y .-Y ., Chiang, M.-C., Dutton, R. A., Hayashi, K. M., Luders, E., Lopez, O. L., Aizenstein, H. J., Toga, A. W., et al. (2008b). Generalized tensor-based morphometry of hiv/aids using multivariate statistics on deformation ten- sors. MedicalImaging,IEEETransactionson, 27(1):129–141. [155] Lepore, N., Brun, C., Chou, Y .-Y ., Chiang, M.-C., Dutton, R. A., Hayashi, K. M., Luders, E., Lopez, O. L., Aizenstein, H. J., Toga, A. W., et al. (2008c). Generalized tensor-based morphometry of hiv/aids using multivariate statistics on deformation ten- sors. MedicalImaging,IEEETransactionson, 27(1):129–141. [156] Leporé, N., V oss, P., Leporé, F., Chou, Y .-Y ., Fortin, M., Gougous, F., Lee, A. D., Brun, C., Lassonde, M., Madsen, S. K., Toga, A. W., and Thompson, P. M. (2010). Brain structure changes visualized in early- and late-onset blind subjects.NeuroImage, 49(1):134–140. [157] Levin, H. S., Benavidez, D., Verger-Maestre, K., Perachio, N., Song, J., Mendel- sohn, D., and Fletcher, J. (2000). Reduction of corpus callosum growth after severe traumatic brain injury in children. Neurology, 54(3):647–647. References 170 [158] Limperopoulos, C., Bassan, H., Sullivan, N. R., Soul, J. S., Robertson, R. L., Moore, M., Ringer, S. A., V olpe, J. J., and du Plessis, A. J. (2008). Positive screening for autism in ex-preterm infants: prevalence and risk factors. Pediatrics, 121(4):758– 765. [159] Lin, Y ., Okumura, A., Hayakawa, F., Kato, T., Kuno, K., and Watanabe, K. (2001). Quantitative evaluation of thalami and basal ganglia in infants with periventricular leukomalacia. DevelopmentalMedicine&ChildNeurology, 43(7):481–485. [160] Loewenstein, D. A., Acevedo, A., Agron, J., Issacson, R., Strauman, S., Crocco, E., Barker, W. W., and Duara, R. (2006). Cognitive profiles in alzheimer’s disease and in mild cognitive impairment of different etiologies. Dementiaandgeriatriccognitive disorders, 21(5-6):309–315. [161] Lorensen, W. E. and Cline, H. E. (1987). Marching cubes: A high resolution 3d surface construction algorithm. In ACM Siggraph Computer Graphics, volume 21, pages 163–169. ACM. [162] Luchsinger, J. A., Tang, M.-X., Stern, Y ., Shea, S., and Mayeux, R. (2001). Dia- betes mellitus and risk of alzheimer’s disease and dementia with stroke in a multiethnic cohort. Americanjournalofepidemiology, 154(7):635–641. [163] Marlow, N., Wolke, D., Bracewell, M. A., and Samara, M. (2005). Neurologic and developmental disability at six years of age after extremely preterm birth. New EnglandJournalofMedicine, 352(1):9–19. [164] Marra, C., Ferraccioli, M., Gabriella Vita, M., Quaranta, D., and Gainotti, G. (2011). Patterns of cognitive decline and rates of conversion to dementia in patients with degenerative and vascular forms of mci. Current Alzheimer Research, 8(1):24– 31. [Mather and Harley] Mather, M. and Harley, C. W. The locus coeruleus: Essential for maintaining cognitive function and the aging brain. [166] Mayer, A., Ling, J., Mannell, M., Gasparovic, C., Phillips, J., Doezema, D., Re- ichard, R., and Yeo, R. (2010). A prospective diffusion tensor imaging study in mild traumatic brain injury. Neurology, 74(8):643–650. [167] McAllister, T., Flashman, L., Maerlender, A., Greenwald, R., Beckwith, J., Toste- son, T., Crisco, J., Brolinson, P., Duma, S. M., Duhaime, A.-C., et al. (2012). Cog- nitive effects of one season of head impacts in a cohort of collegiate contact sport athletes. Neurology, 78(22):1777–1784. References 171 [168] McAllister, T. W., Ford, J. C., Flashman, L. A., Maerlender, A., Greenwald, R. M., Beckwith, J. G., Bolander, R. P., Tosteson, T. D., Turco, J. H., Raman, R., et al. (2014). Effect of head impacts on diffusivity measures in a cohort of collegiate contact sport athletes. Neurology, 82(1):63–69. [169] McKee, A. C., Stein, T. D., Nowinski, C. J., Stern, . R. A., Daneshvar, D. H., Alvarez, V . E., Lee, H.-S., Hãll, G., Wojtowicz, S. M., Baugh, C. M., et al. (2013). The spectrum of disease in chronic traumatic encephalopathy. Brain, 136(1):43–64. [170] Mega, M. S. and Cummings, J. L. (1994). Frontal-subcortical circuits and neu- ropsychiatric disorders. TheJournalofNeuropsychiatryandClinicalNeurosciences. [171] Mergler, D., Huel, G., Bowler, R., Iregren, A., Belanger, . S., Baldwin, M., Tardif, R., Smargiassi, A., and Martin, L. (1994). Nervous system dysfunction among work- ers with long-term exposure to ma nganese. Environmentalresearch, 64(2):151–180. [172] Metzger, C. D., van der Werf, Y . D., and Walter, M. (2013). Functional mapping of thalamic nuclei and their integration into cortico-striatal-thalamo-cortical loops via ultra-high resolution imaging—from animal anatomy to in vivo imaging in humans. Frontiersinneuroscience, 7. [173] Meyer, J. S., Xu, G., Thornby, J., Chowdhury, M. H., and Quach, M. (2002). Is mild cognitive impairment prodromal for vascular dementia like alzheimer’s disease? Stroke, 33(8):1981–1985. [174] Middleton, F. A. and Strick, P. L. (2000). Basal ganglia output and cognition: evidence from anatomical, behavioral, and clinical studies. Brain and cognition, 42(2):183–200. [175] Miller, B. L. and Cummings, J. L. (2007). The human frontal lobes: Functions anddisorders. Guilford Press. [176] Mishkin, M., Ungerleider, L. G., and Macko, K. A. (1983). Object vision and spatial vision: two cortical pathways. Trendsinneurosciences, 6:414–417. [177] Morrison, J. H. and Hof, P. R. (1997). Life and death of neurons in the aging brain. Science, 278(5337):412–419. [178] Mrzljak, L., Uylings, H., Kostovic, I., and van Eden, C. G. (1988). Prenatal devel- opment of neurons in the human prefrontal cortex: I. a qualitative golgi study. Journal ofcomparativeneurology, 271(3):355–386. References 172 [179] Mungas, D., Reed, B. R., and Kramer, J. H. (2003). Psychometrically matched measures of global cognition, memory, and executive function for assesment of cog- nitive decline in older persons. Neuropsychology, 17(3):380. [180] Nagasunder, A., Kinney, H., Blüml, S., Tavaré, C., Rosser, T., Gilles, F., Nelson, M., and Panigrahy, A. (2011). Abnormal microstructure of the atrophic thalamus in preterm survivors with periventricular leukomalacia. AmericanJournalofNeuroradi- ology, 32(1):185–191. [181] Nagata, K., Basugi, N., Fukushima, T., Tango, T., Suzuki, I., Kaminuma, T., and Kurashina, S. (1987). A quantitative study of physiological cerebral atrophy with aging. Neuroradiology, 29(4):327–332. [182] Newman, A. B., Fitzpatrick, A. L., Lopez, O., Jackson, S., Lyketsos, C., Jagust, W., Ives, D., DeKosky, S. T., and Kuller, L. H. (2005). Dementia and alzheimer’s disease incidence in relationship to cardiovascular disease in the cardiovascular health study cohort. JournaloftheAmericanGeriatricsSociety, 53(7):1101–1107. [183] Nichols, T. E. and Holmes, A. P. (2001). Nonparametric permutation tests for functional neuroimaging: a primer with examples. Human brain mapping, 15(1):1– 25. [184] Nordlund, A., Rolstad, S., Klang, O., Lind, K., Hansen, S., and Wallin, A. (2007). Cognitive profiles of mild cognitive impairment with and without vascular disease. Neuropsychology, 21(6):706. [185] Nosarti, C., Rushe, T. M., Woodruff, P. W., Stewart, A. L., Rifkin, L., and Murray, R. M. (2004). Corpus callosum size and very preterm birth: relationship to neuropsy- chological outcome. Brain, 127(9):2080–2089. [186] Nyström, O., Wallin, A., and Nordlund, A. (2015). Mci of different etiologies differ on the cognitive assessment battery.ActaNeurologicaScandinavica, 132(1):31– 36. [187] O’Donnell, L. J., Westin, C.-F., and Golby, A. J. (2009). Tract-based morphometry for white matter group analysis. Neuroimage, 45(3):832–844. [188] Oh, S.-H., Lee, K.-Y ., Im, J.-H., and Lee, M.-S. (2002). Chorea associated with non-ketotic hyperglycemia and hyperintensity basal ganglia lesion on t1-weighted brain mri study: a meta-analysis of 53 cases including four present cases. Journal oftheneurologicalsciences, 200(1):57–62. References 173 [189] Ott, A., Breteler, M., Van Harskamp, F., Claus, J. J., Van Der Cammen, T. J., Grobbee, D. E., and Hofman, A. (1995). Prevalence of alzheimer’s disease and vascu- lar dementia: association with education. the rotterdam study. Bmj, 310(6985):970– 973. [190] Oulhote, Y ., Mergler, D., Barbeau, B., Bellinger, D. C., Bouffard, T., Brodeur, M.-È., Saint-Amour, D., Legrand, M., Sauvé, S., and Bouchard, M. F. (2014). Neu- robehavioral function in school-age children exposed to manganese in drinking water. [191] Park, H.-J., Kim, J. J., Lee, S.-K., Seok, J. H., Chun, J., Kim, D. I., and Lee, J. D. (2008). Corpus callosal connection mapping using cortical gray matter parcellation and dt-mri. Humanbrainmapping, 29(5):503–516. [192] Pennec, X., Fillard, P., and Ayache, N. (2006). A riemannian framework for tensor computing. InternationalJournalofComputerVision, 66(1):41–66. [193] Peterson, B. S., V ohr, B., Staib, L. H., Cannistraci, C. J., Dolberg, A., Schnei- der, K. C., Katz, K. H., Westerveld, M., Sparrow, S., Anderson, A. W., et al. (2000). Regional brain volume abnormalities and long-term cognitive outcome in preterm in- fants. JAMA:thejournaloftheAmericanMedicalAssociation, 284(15):1939–1947. [194] Pierpaoli, C., Jezzard, P., Basser, P. J., Barnett, A., and Di Chiro, G. (1996). Dif- fusion tensor mr imaging of the human brain. Radiology, 201(3):637–648. [195] Pizer, S. M., Fritsch, D. S., Yushkevich, P. A., Johnson, V . E., and Chaney, E. L. (1999). Segmentation, registration, and measurement of shape variation via image object shape. MedicalImaging,IEEETransactionson, 18(10):851–865. [196] Pollack, M., Phaneuf, S., Dirks, A., and Leeuwenburgh, C. (2002). The role of apoptosis in the normal aging brain, skeletal muscle, and heart. Annals of the New YorkAcademyofSciences, 959(1):93–107. [197] Posner, H. B., Tang, M.-X., Luchsinger, J., Lantigua, R., Stern, Y ., and Mayeux, R. (2000). The relationship of hypertension in the elderly to ad, vascular dementia, and cognitive function. Neurology, 58(8):1175–1181. [198] Ptito, M., Moesgaard, S. M., Gjedde, A., and Kupers, R. (2005). Cross-modal plasticity revealed by electrotactile stimulation of the tongue in the congenitally blind. Brain, 128(3):606–614. [199] Ptito, M., Schneider, F. C., Paulson, O. B., and Kupers, R. (2008). Alterations of the visual pathways in congenital blindness. ExpBrainRes, 187:41–49. References 174 [200] RAFAL, R. D., WALKER, J. A., POSNER, M. I., and FRIEDRICH, F. J. (1984). Cognition and the basal ganglia separating mental and motor components of perfor- mance in parkinson’s disease. Brain, 107(4):1083–1094. [201] Rajagopalan, V ., Scott, J., Habas, P. A., Kim, K., Rousseau, F., Glenn, O. A., Barkovich, A. J., and Studholme, C. (2012). Mapping directionality specific volume changes using tensor based morphometry: an application to the study of gyrogenesis and lateralization of the human fetal brain. Neuroimage, 63(2):947–958. [202] Raz, N., Lindenberger, U., Rodrigue, K. M., Kennedy, K. M., Head, D., Williamson, A., Dahle, C., Gerstorf, D., and Acker, J. D. (2005). Regional brain changes in aging healthy adults: general trends, individual differences and modifiers. Cerebralcortex, 15(11):1676–1689. [203] Raznahan, A., Shaw, P. W., Lerch, J. P., Clasen, L. S., Greenstein, D., Berman, R., Pipitone, J., Chakravarty, M. M., and Giedd, J. N. (2014). Longitudinal four- dimensional mapping of subcortical anatomy in human development. Proceedingsof theNationalAcademyofSciences, 111(4):1592–1597. [204] Reed, B. R., Mungas, D. M., Kramer, J. H., Ellis, W., Vinters, H. V ., Zarow, C., Jagust, W. J., and Chui, H. C. (2007). Profiles of neuropsychological impairment in autopsy-defined alzheimer’s disease and cerebrovascular disease. Brain, 130(3):731– 739. [205] Reijneveld, S. A., De Kleine, M., van Baar, A. L., Kollée, L. A., Verhaak, C. M., Verhulst, F. C., and Verloove-Vanhorick, S. P. (2006). Behavioural and emotional problems in very preterm and very low birthweight infants at age 5 years. Archivesof DiseaseinChildhood-FetalandNeonatalEdition, 91(6):F423–F428. [206] Robinson, R. G., Kubos, K. L., Starr, L. B., Rao, K., and Price, T. R. (1984). Mood disorders in stroke patients: importance of location of lesion. Brain, 107(1):81–93. [207] Robinson, S., Goddard, L., Dritschel, B., Wisley, M., and Howlin, P. (2009). Ex- ecutive functions in children with autism spectrum disorders. Brain and Cognition, 71(3):362–368. [208] Rodier, J. (1955). Manganese poisoning in moroccan miners. British journal of industrialmedicine, 12(1):21. [209] Roels, H., Bowler, R., Kim, Y ., Claus Henn, B., Mergler, D., Hoet, P., Gocheva, V ., Bellinger, D., Wright, R., Harris, M., et al. (2012). Manganese exposure and References 175 cognitive deficits: a growing concern for manganese neurotoxicity. Neurotoxicology, 33(4):872–880. [210] Rosas, H. D., Lee, S. Y ., Bender, A. C., Zaleta, A. K., Vangel, M., Yu, P., Fis- chl, B., Pappu, V ., Onorato, C., Cha, J.-H., et al. (2010). Altered white matter mi- crostructure in the corpus callosum in huntington’s disease: implications for cortical “disconnection”. Neuroimage, 49(4):2995–3004. [211] Ross, A. (2004). Procrustes analysis. Course report, Department of Computer ScienceandEngineering,UniversityofSouthCarolina. [212] RUDGE, P. and WARRINGTON, E. K. (1991). Selective impairment of memory and visual perception in splenial tumours. Brain, 114(1):349–360. [213] Saint-Cyr, J. A., Taylor, A., and Nicholson, K. (1994). Behavior and the basal ganglia. Advancesinneurology, 65:1–28. [214] Sandman, C. A., Head, K., Muftuler, L. T., Su, L., Buss, C., and Davis, E. P. (2014). Shape of the basal ganglia in preadolescent children is associated with cogni- tive performance. NeuroImage. [215] Scheltens, P., Fox, N., Barkhof, F., and De Carli, C. (2002). Structural magnetic resonance imaging in the practical assessment of dementia: beyond exclusion. The LancetNeurology, 1(1):13–21. [216] Scoville, W. B. and Milner, B. (1957). Loss of recent memory after bilateral hippocampal lesions. Journalofneurology,neurosurgery,andpsychiatry, 20(1):11. [217] Serra, L., Cercignani, M., Lenzi, D., Perri, R., Fadda, L., Caltagirone, C., Macaluso, E., and Bozzali, M. (2010). Grey and white matter changes at different stages of alzheimer’s disease. JournalofAlzheimer’sdisease, 19(1):147–159. [218] Shaw, P., De Rossi, P., Watson, B., Wharton, A., Greenstein, D., Raznahan, A., Sharp, W., Lerch, J. P., and Chakravarty, M. M. (2014). Mapping the development of the basal ganglia in children with attention-deficit/hyperactivity disorder. Journal of theAmericanAcademyofChild&AdolescentPsychiatry, 53(7):780–789. [219] Shi, J., Thompson, P. M., Gutman, B., Wang, Y ., Initiative, A. D. N., et al. (2013a). Surface fluid registration of conformal representation: Application to detect disease burden and genetic influence on hippocampus. NeuroImage, 78:111–134. References 176 [220] Shi, J., Wang, Y ., Ceschin, R., An, X., Lao, Y ., Vanderbilt, D., Nelson, M. D., Thompson, P. M., Panigrah˜ y, A., and Leporé, N. (2013b). A multivariate surface- based analysis of the putamen in premature newb orns: Regional differences within the ventral striatum. PloSone, 8(7):e66736. [Shi et al.] Shi, J., Wang, Y ., Ceschin, R., An, X., Nelson, M. D., Panigrahy, A., and Leporé, N. Surface fluid registration and multivariate tensor-based morphometry in newborns-the effects of prematurity on the putamen. [222] Shim, Y . S., Yoon, B., Shon, Y .-M., Ahn, K.-J., and Yang, D.-W. (2008). Differ- ence of the hippocampal and white matter microalterations in mci patients according to the severity of subcortical vascular changes: neuropsychological correlates of dif- fusion tensor imaging. Clinicalneurologyandneurosurgery, 110(6):552–561. [223] Shimony, J., Burton, H., Epstein, A., Mclaren, D., Sun, S., and Snyder, A. (2006). Diffusion tensor imaging reveals white matter reorganization in early blind humans. CerebralCortex, 16:1653–1661. [224] Shinotoh, H., Snow, B., Chu, N.-S., Huang, C.-C., Lu, C.-S., Lee, C., Takahashi, H., and Calne, D. (1997). Presynaptic and postsynaptic striatal dopaminergic func- tion in patients with manganese intoxication a positron emission tomography study. Neurology, 48(4):1053–1055. [225] Shu, N., Liu, Y ., Li, J., Li, Y ., Yu, C., and Jiang, T. (2009a). Altered anatomical network in early blindness revealed by diffusion tensor tractography. PLoSOne, 4(9). [226] Shu, N., Liu, Y ., Li, J., Li, Y ., Yu, C., and Jiang, T. (2009b). Altered anatomi- cal network in early blindness revealed by diffusion tensor tractography. PloS one, 4(9):e7228. [227] Sigel, A., Sigel, H., and Sigel, R. K. (2006). Neurodegenerative Diseases and MetalIons: MetalIonsinLifeSciences, volume 9. John Wiley & Sons. [228] Šimi´ c, G., Kostovi´ c, I., Winblad, B., and Bogdanovi´ c, N. (1997). V olume and number of neurons of the human hippocampal formation in normal aging and alzheimer’s disease. JournalofComparativeNeurology, 379(4):482–494. [229] Singh, M., Jeong, J., Hwang, D., Sungkarat, W. t., and Gruen, P. (2010). Novel dif- fusion tensor imaging methodology to detect and quantify inju red regions and affected brain pathways in traumatic brain injury. Magneticresonanceimaging, 28(1):22–40. References 177 [230] Smith, S. M., Jenkinson, M., Johansen-Berg, H., Rueckert, D., Nichols, T. E., Mackay, C. E., Watkins, K. E., Ciccarelli, O., Cader, M. Z., Matthews, P. M., et al. (2006). Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage, 31(4):1487–1505. [231] Sonuga-Barke, E. J. (2002). Psychological heterogeneity in ad/hd—a dual path- way model of behaviour and cognition. Behaviouralbrainresearch, 130(1):29–36. [232] Sowell, E. R., Peterson, B. S., Thompson, P. M., Welcome, S. E., Henkenius, A. L., and Toga, A. W. (2003). Mapping cortical change across the human life span. Natureneuroscience, 6(3):309–315. [233] Sowell, E. R., Thompson, P. M., Tessner, K. D., and Toga, A. W. (2001). Map- ping continued brain growth and gray matter density reduction in dorsal frontal cortex: Inverse relationships during postadolescent brain maturation. The Journal of Neuro- science, 21(22):8819–8829. [234] Srinivasan, L., Dutta, R., Counsell, S. J., Allsop, J. M., Boardman, J. P., Ruther- ford, M. A., and Edwards, A. D. (2007). Quantification of deep gray matter in preterm infants at term-equivalent age using manual volumetry of 3-tesla magnetic resonance images. Pediatrics, 119(4):759–765. [235] Stampfer, M. (2006). Cardiovascular disease and alzheimer’s disease: common links. Journalofinternalmedicine, 260(3):211–223. [236] Starkstein, S. E., Robinson, R. G., Berthier, M. L., Parikh, R. M., and Price, T. R. (1988). Differential mood changes following basal ganglia vs thalamic lesions. ArchivesofNeurology, 45(7):725. [237] Starkstein, S. E., Robinson, R. G., and Price, T. R. (1987). Comparison of cor- tical and subcortical lesions in the production of poststroke mood disorders. Brain, 110(4):1045–1059. [238] Stejskal, E. (1965). Use of spin echoes in a pulsed magnetic-field gradient to study anisotropic, restricted diffusion and flow. The Journal of Chemical Physics, 43(10):3597–3603. [239] Stejskal, E. O. and Tanner, J. E. (1965). Spin diffusion measurements: spin echoes in the presence of a time-dependent field gradient. The journal of chemical physics, 42(1):288–292. References 178 [240] Stiles, J. and Jernigan, T. L. (2010). The basics of brain development. Neuropsy- chologyreview, 20(4):327–348. [241] Strand-Brodd, K., Ewald, U., Grönqvist, H., Holmström, G., Strömberg, B., Grön- qvist, E., von Hofsten, C., and Rosander, K. (2011). Development of smooth pur- suit eye movements in very preterm infants: 1. general aspects. Acta Paediatrica, 100(7):983–991. [242] Styner, M., Oguz, I., Xu, S., Brechbühler, C., Pantazis, D., Levitt, J. J., Shenton, M. E., and Gerig, G. (2006). Framework for the statistical shape analysis of brain structures using spharm-pdm. Theinsightjournal, (1071):242. [243] Székely, G. J., Rizzo, M. L., et al. (2009). Brownian distance covariance. The annalsofappliedstatistics, 3(4):1236–1265. [244] Talavage, T., Nauman, E., Breedlove, E., Yoruk, U., Dye, A., Morigaki, K., Feuer, H., and Leverenz, L. (2013). Functionally-detected cognitive impairment in high school football players without clinically-diagnosed concussion. Journal of neuro- trauma. [245] Teipel, S. J., Bayer, W., Alexander, G. E., Teichberg, D., Kulic, L., Schapiro, M. B., Möller, H.-J., Rapoport, S. I., and Hampel, H. (2002). Progression of corpus callosum atrophy in alzheimer disease. Archivesofneurology, 59(2):243–248. [246] Teipel, S. J., Meindl, T., Grinberg, L., Grothe, M., Cantero, J. L., Reiser, M. F., Möller, H.-J., Heinsen, H., and Hampel, H. (2011). The cholinergic system in mild cognitive impairment and alzheimer’s disease: an in vivo mri and dti study. Human brainmapping, 32(9):1349–1362. [247] Thompson, P. M., Hayashi, K. M., de Zubicaray, G. I., Janke, A. L., Rose, S. E., Semple, J., Hong, M. S., Herman, D. H., Gravano, D., Doddrell, D. M., et al. (2004). Mapping hippocampal and ventricular change in alzheimer disease. Neuroimage, 22(4):1754–1766. [248] Toga, A. W., Thompson, P. M., and Sowell, E. R. (2006). Mapping brain matura- tion. Focus. [249] Tomaiuolo, F., Carlesimo, G., Di Paola, M., Petrides, M., Fera, F., Bonanni, R., Formisano, R., Pasqualetti, P., and Caltagirone, C. (2004). Gross morphology and morphometric sequelae in the hippocampus, fornix, and corpus callosum of patients References 179 with severe non-missile traumatic brain injury without macroscopically detectable le- sions: a t1 weighted mri study. Journal of Neurology, Neurosurgery & Psychiatry, 75(9):1314–1322. [250] Toplak, M. E., Jain, U., and Tannock, R. (2005). Executive and motivational pro- cesses in adolescents with attention-deficit-hyperactivity disorder (adhd). Behavioral andBrainFunctions, 1(1):8. [251] Toussaint, N., Souplet, J., and P., F. (2007). Medinria: Medical image navigation and research tool by inria. Proc. of MICCAI’07 Workshop on Interaction in medical imageanalysisandvisualization. [252] Ukmar, M., Makuc, E., Onor, M., Garbin, G., Trevisiol, M., and Cova, M. (2008). Evaluation of white matter damage in patients with alzheimer’s disease and in patients with mild cognitive impairment by using diffusion tensor imaging. La radiologia medica, 113(6):915–922. [253] Van Der Werf, Y . D., Jolles, J., Witter, M. P., and Uylings, H. (2003). Contributions of thalamic nuclei to declarative memory functioning. Cortex, 39(4):1047–1062. [254] Vermeer, S. E., Prins, N. D., den Heijer, T., Hofman, A., Koudstaal, P. J., and Breteler, M. M. (2003). Silent brain infarcts and the risk of dementia and cognitive decline. NewEnglandJournalofMedicine, 348(13):1215–1222. [255] Villeneuve, S., Reed, B. R., Madison, C. M., Wirth, M., Marchant, N. L., Kriger, S., Mack, W. J., Sanossian, N., DeCarli, C., Chui, H. C., et al. (2014). Vascular risk and aβ interact to reduce cortical thickness in ad vulnerable brain regions. Neurology, 83(1):40–47. [256] V olpe, J. J. (2009). Brain injury in premature infants: a complex amalgam of destructive and developmental disturbances. TheLancetNeurology, 8(1):110–124. [257] Wang, D., Qin, W., Liu, Y ., Zhang, Y ., Jiang, T., and Yu, C. (2013). Altered white matter integrity in the congenital and late blind people. Neuralplasticity, 2013. [258] Wang, Y . (2011d). Multivariate surface-based subcortical structure morphometry system. http://gsl.lab.asu.edu/conformal.htm. [259] Wang, Y ., Chan, T. F., Toga, A. W., and Thompson, P. M. (2009). Multivariate tensor-based brain anatomical surface morphometry via holomorphic one-forms. In InternationalConferenceonMedicalImageComputingandComputer-AssistedInter- vention, pages 337–344. Springer. References 180 [260] Wang, Y ., Lui, L. M., Gu, X., Hayashi, K. M., Chan, T. F., Toga, A. W., Thompson, P. M., and Yau, S.-T. (2007). Brain surface conformal parameterization using riemann surface structure. IEEEtransactionsonmedicalimaging, 26(6):853–865. [261] Wang, Y ., Panigrahy, A., Ceschin, R., Liu, S., Thompson, P. M., and Lepore, N. (2011a). Surface morphometry of subcortical structures in premature neonates. In Proc.Intl.Soc.Mag.Reson.Med, volume 19, page 2585. [262] Wang, Y ., Panigrahy, A., Shi, J., Ceschin, R., and Nelson, M. (2011b). Multivariate tensor based morphometry on premature neonates: a pilot study. Proceedings of the MICCAIworkshoponImageAnalysisofHumanBrainDevelopment(IAHBD2011). [263] Wang, Y ., Song, Y ., Rajagopalan, P., An, T., Liu, K., Chou, Y .-Y ., Gutman, B., Toga, A. W., and Thompson, P. M. (2011c). Surface-based TBM boosts power to detect disease effects on the brain: An n= 804 ADNI study. Neuroimage, 56(4):1993– 2010. [264] Wang, Y ., Song, Y ., Rajagopalan, P., An, T., Liu, K., Chou, Y .-Y ., Gutman, B., Toga, A. W., and Thompson, P. M. (2011d). Surface-based TBM boosts power to detect disease effects on the brain: An N= 804 ADNI study. Neuroimage, 56(4):1993– 2010. [265] Wang, Y ., Yin, X., Zhang, J., Gu, X., Chan, T., Thompson, P. M., and Yau, S.-T. (2008). Brain mapping with the ricci flow conformal parameterization and multi- variate statistics on deformation tensors. In 2ndMICCAIWorkshoponMathematical FoundationsofComputationalAnatomy, pages 36–47. [266] Wang, Y ., Zhang, J., Gutman, B., Chan, T. F., Becker, J. T., Aizenstein, H. J., Lopez, O. L., Tamburo, R. J., Toga, A. W., and Thompson, P. M. (2010a). Multi- variate tensor-based morphometry on surfaces: Application to mapping ventricular abnormalities in HIV/AIDS. NeuroImage, 49(3):2141–2157. [267] Wasserman, G. A., Liu, X., Parvez, F., Ahsan, H. b., Levy, D., Factor-Litvak, P., Kline, J., van Geen, A. a., Slavkovich, V ., Lolacono, N. J., et al. (2006). Water manganese exposure and children’s intellectual function in araih azar, bangladesh. Environmentalhealthperspectives, pages 124–129. [268] WHO Office of Occupational Health (1986). Operational guide for the who neu- robehavioral test battery (ntcb). WorldHealthOrganization(WHO),Geneva. References 181 [269] Wilde, E. A., Chu, Z., Bigler, E. D., Hunter, J. V ., Fearing, M. A., Hanten, G., Newsome, M. R., Scheibel, R. S., Li, X., and Levin, H. S. (2006). Diffusion tensor imaging in the corpus callosum in children after moderate to severe traumatic brain injury. Journalofneurotrauma, 23(10):1412–1426. [270] Wood, N. S., Marlow, N., Costeloe, K., Gibson, A. T., and Wilkinson, A. R. (2000). Neurologic and developmental disability after extremely preterm birth. New EnglandJournalofMedicine, 343(6):378–384. [271] Wright, R. O., Amarasiriwardena, C., Woolf, A. D., Jim, R., and Bellinger, D. C. (2006). Neuropsychological correlates of hair arsenic, manganese, and cadmium lev- els in school-age children residing near a hazardous waste site. Neurotoxicology, 27(2):210–216. [272] Yamaura, H., Ito, M., Kubota, K., and Matsuzawa, T. (1980). Brain atrophy dur- ing aging: a quantitative study with computed tomography. Journal of gerontology, 35(4):492–498. [273] Yu, C., Shu, N., Li, J., Qin, W., Jiang, T., and Li, K. (2007). Plasticity of the corticospinal tract in early blindness revealed by quantitative analysis of fractional anisotropy based on diffusion tensor tractography. NeuroImage, 36:411–417. [274] Zhang, H., Avants, B., Yushkevich, P., Woo, J., Wang, S., McCluskey, L., Elman, L., Melhem, E., and Gee, J. (2007a). High-dimensional spatial normalization of diffu- sion tensor images improves the detection of white matter differences in amyotrophic lateral sclerosis. IEEETransactionsonMedicalImaging-SpecialIssueonComputa- tionalDiffusionMRI, 26(11):1585–1597. [275] Zhang, H., Awate, S. P., Das, S. R., Woo, J. H., Melhem, E. R., Gee, J. C., and Yushkevich, P. A. (2010). A tract-specific framework for white matter morphome- try combining macroscopic and microscopic tract features. Medical image analysis, 14(5):666–673. [276] Zhang, H., Yushkevich, P. A., Alexander, D. C., Gee, J. C., et al. (2006). De- formable registration of diffusion tensor mr images with explicit orientation optimiza- tion. Medicalimageanalysis, 10(5):764–785. [277] Zhang, Y ., Schuff, N., Camacho, M., Chao, L. L., Fletcher, T. P., Yaffe, K., Wool- ley, S. C., Madison, C., Rosen, H. J., Miller, B. L., et al. (2013). Mri markers for mild cognitive impairment: Comparisons between white matter integrity and gray matter volume measurements. PloSone, 8(6):e66367. References 182 [278] Zhang, Y ., Schuff, N., Jahng, G.-H., Bayne, W., Mori, S., Schad, L., Mueller, S., Du, A.-T., Kramer, J., Yaffe, K., et al. (2007b). Diffusion tensor imaging of cingulum fibers in mild cognitive impairment and alzheimer disease. Neurology, 68(1):13–19. [279] Zhu, M., Gao, W., Wang, X., Shi, C., and Lin, Z. (2012). Progression of corpus callosum atrophy in early stage of alzheimer’s disease: Mri based study. Academic radiology, 19(5):512–517. [280] Zhu, M., Wang, X., Gao, W., Shi, C., Ge, H., Shen, H., and Lin, Z. (2013). Corpus callosum atrophy and cognitive decline in early alzheimer’s disease: Longitudinal mri study. Dementiaandgeriatriccognitivedisorders, 37(3-4):214–222. [281] Zhuang, L., Wen, W., Zhu, W., Trollor, J., Kochan, N., Crawford, J., Reppermund, S., Brodaty, H., and Sachdev, P. (2010). White matter integrity in mild cognitive impairment: a tract-based spatial statistics study. Neuroimage, 53(1):16–25.

Abstract (if available)

Abstract Brain development is an on-going process that spans the whole life, and exhibits position, shape, and connectivity changes due to genetics, external environment, experiences and aging. By rapidly unveiling the anatomy and function in the living brain, magnetic resonance imaging (MRI) technology has fueled a scientific revolution in the neuroscience field. In parallel with the advent of MRI, population-based post-processing algorithms facilitate the in-vivo detection and monitoring of brain alterations. Here, we describe a combined framework for the analysis of structural MRI and diffusion tensor imaging (DTI) that aims to depict subtle brain alterations in healthy and abnormal brain evolution throughout the lifespan. ❧ Firstly, a T1-MRI based analysis characterizing the brain subcortical gray matter morphometry and inter-structural correlation is shown, and applications on premature neonates as well as children with chronic manganese (Mn) exposure are presented. Secondly, a T1-MRI based algorithm assessing the brain positional changes is described, and applications on premature neonates as well as young adults with mild traumatic brain injury (mTBI) are presented. Finally, a novel T1 and DTI fusion method yielding an increasing detection power of white matter alterations is introduced, and applications on mTBI subjects as well as aging subjects with increased risk of cardiovascular diseases are presented. The presented work allows shape, positional, and connectivity exhibited in the developing brain to be assessed within a more comprehensive framework.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Characterization of the brain in early childhood

PDF

Characterization of lenticulostriate arteries using high-resolution black blood MRI as an early imaging biomarker for vascular cognitive impairment and dementia

PDF

Unveiling the white matter microstructure in 22q11.2 deletion syndrome with diffusion magnetic resonance imaging

PDF

Pattern detection in medical imaging: pathology specific imaging contrast, features, and statistical models

PDF

Susceptibility-weighted MRI for the evaluation of brain oxygenation and brain iron in sickle cell disease

PDF

Correction, coregistration and connectivity analysis of multi-contrast brain MRI

PDF

Pediatric magnetic resonance image processing: applications to posterior fossa cancer and normal development

PDF

Diffusion MRI of the human brain: signal modeling and quantitative analysis

PDF

Functional connectivity analysis and network identification in the human brain

PDF

Signal processing methods for brain connectivity

PDF

Mapping water exchange rate across the blood-brain barrier

PDF

Methods for improving reliability and consistency in diffusion MRI analysis

PDF

Seeing sleep: real-time MRI methods for the evaluation of sleep apnea

PDF

Measuring functional connectivity of the brain

PDF

Relationships between lifetime chronic stress exposure, vascular risk, cognition, and brain structure in HIV

PDF

3D vessel mapping techniques for retina and brain as an early imaging biomarker for small vessel diseases

PDF

Characterization of visual cortex function in late-blind individuals with retinitis pigmentosa and Argus II patients

PDF

Toward understanding speech planning by observing its execution—representations, modeling and analysis

PDF

Neuroimaging in complex polygenic disorders

PDF

Cross-sectional association of blood pressure, antihypertensive medications, MRI volumetric measures and cognitive function scores in an aging population

Asset Metadata

Creator Lao, Yi (author)

Core Title Shape, pose, and connectivity in subcortical networks across the human lifespan

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Biomedical Engineering

Publication Date 05/02/2017

Defense Date 11/17/2016

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag brain,connectivity,human lifespan,MRI,neuroimaging,OAI-PMH Harvest,pose,shape,subcortical structures

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor D’Argenio, David Z. (committee chair), Lepore, Natasha (committee chair), Nayak, Krishna S. (committee member), Shi, Yonggang (committee member), Yang, Yaling (committee member)

Creator Email yilao1987@gmail.com,ylao@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-368368

Unique identifier UC11255952

Identifier etd-LaoYi-5303.pdf (filename),usctheses-c40-368368 (legacy record id)

Legacy Identifier etd-LaoYi-5303.pdf

Dmrecord 368368

Document Type Dissertation

Rights Lao, Yi

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA