Computational diffusion MRI : 13th international workshop, CDMRI 2022, held in conjunction with MICCAI 2022, Singapore, Singapore, September 22, 2022 : proceedings

material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Preface 6 Organization 8 Contents 10 Data Preprocessing 12 Slice Estimation in Diffusion MRI of Neonatal and Fetal Brains in Image and Spherical Harmonics Domains Using Autoencoders 13 1 Introduction 14 2 Methodology 15 2.1 Materials 15 2.2 Model 15 3 Results 18 3.1 DWI Assessment 18 3.2 FA and MD in Newborns 19 3.3 Qualitative Results of FA and MD in Fetuses 19 4 Conclusion 21 References 21 Super-Resolution of Manifold-Valued Diffusion MRI Refined by Multi-modal Imaging 24 1 Introduction 24 2 Background and Methods 25 2.1 Background 25 2.2 Proposed Method 27 3 Experiments 29 3.1 Data 29 3.2 Implementation and Training Details 29 3.3 Models and Evaluation Metrics 30 4 Results 31 4.1 Proposed Model Performance 31 4.2 Model Ablation Results 32 5 Discussion 34 References 34 .28em plus .1em minus .1emLossy Compression of Multidimensional Medical Images Using Sinusoidal Activation Networks: An Evaluation Study 36 1 Introduction 36 2 Methodology 38 3 Results 40 4 Discussion and Conclusion 45 References 46 Correction of Susceptibility Distortion in EPI: A Semi-supervised Approach with Deep Learning 48 1 Introduction 49 2 Background 50 2.1 Distortion Model 50 2.2 Distortion Correction Using Image Registration 50 3 Method 51 3.1 Model Architecture 51 3.2 Models 52 4 Evaluations 53 4.1 Datasets 54 4.2 Models 55 4.3 Assessment Metrics 55 5 Results 55 6 Discussion 57 7 Conclusion 58 References 58 The Impact of Susceptibility Distortion Correction Protocols on Adolescent Diffusion MRI Measures 60 1 Introduction 60 2 Methods 62 2.1 Study Participants and MRI Data 62 2.2 dMRI Preprocessing and Subsampling 62 2.3 dMRI Models and Regional Measures 63 2.4 Statistics 64 3 Results 64 3.1 DTI and NODDI Map Comparisons 64 3.2 DTI and NODDI Fit Evaluations 64 3.3 GAM Age Associations 66 4 Discussion 66 References 70 Signal Representations 72 Diffusion MRI Fibre Orientation Distribution Inpainting 73 1 Introduction 73 2 Methods 74 2.1 The Human Connectome Project Dataset 74 2.2 Data Preprocessing 75 2.3 3D FOD Inpainting Framework 76 2.4 Feature Encoding Stage 76 2.5 Order-Wise Coefficient Decoders 77 2.6 Implementation Details 77 3 Experimental Results 78 3.1 Inpainting Quality Analysis 78 3.2 Connectome Matrix Analysis 80 4 Conclusion 82 References 82 Fitting a Directional Microstructure Model to Diffusion-Relaxation MRI Data with Self-supervised Machine Learning 85 1 Introduction 85 2 Methods 86 2.1 Microstructure Model 86 2.2 Combined T1-Diffusion in Vivo Data 87 2.3 Simulated Data 88 2.4 Non-linear Least Squares Fitting 88 2.5 Self-supervised Model Fitting 89 3 Results 89 3.1 Simulated Data 89 3.2 Real Data 90 4 Discussion 90 5 Conclusion 94 References 94 Stepwise Stochastic Dictionary Adaptation Improves Microstructure Reconstruction with Orientation Distribution Function Fingerprinting 97 1 Introduction 97 2 Methods 98 2.1 Biophysical Diffusion Model 98 2.2 Orientation Distribution Function Fingerprinting 99 2.3 Stepwise Stochastic Adaptation of a Dictionary 99 2.4 Data 100 2.5 Evaluation 101 3 Results 101 4 Discussion 104 5 Conclusions 107 References 107 How Can Spherical CNNs Benefit ML-Based Diffusion MRI Parameter Estimation? 109 1 Introduction 109 2 ML Solutions to the dMRI Parameter Estimation Problem and the Theoretical Benefits of S-CNNs 110 2.1 Fully-Connected Networks 111 2.2 Spherical CNNs 111 3 Experiments 112 3.1 Experiment 1 112 3.2 Experiment 2 114 4 Results and Discussion 115 4.1 For Experiment 1 115 4.2 For Experiment 2 116 5 Conclusion 119 References 119 Tractography and WM Pathways 121 DC2U-Net: Tract Segmentation in Brain White Matter Using Dense Criss-Cross U-Net 122 1 Introduction 122 2 Methods 123 2.1 Dense Criss-Cross U-Net 123 2.2 Dense Criss-Cross Attention (DCCA) Block 125 2.3 Deeply Supervised Loss Function 126 3 Experiments 126 3.1 Dataset and Implementation Details 126 3.2 Results 127 3.3 Ablation Analysis of DC2U-Net 127 4 Conclusion 130 References 130 Clustering in Tractography Using Autoencoders (CINTA) 132 1 Introduction 132 1.1 Related Work 133 2 Material and Methods 134 3 Experiments 135 4 Results 136 5 Discussion 138 6 Conclusion 139 A Appendix 139 A.1 Misclassified Streamlines 139 A.2 Time Computational Requirements 140 References 141 Tractometric Coherence of Fiber Bundles in DTI 144 1 Introduction 144 2 Theory 145 3 Experiments 147 4 Discussion 149 References 153 Author Index 156 This book constitutes the proceedings of the International Workshop on Computational Diffusion MRI, CDMRI 2022, which was held 22 September 2022, in conjunction with MICCAI 2022. The 12 full papers included were carefully reviewed and selected for inclusion in the book. The papers were organized in topical sections as follows: Data processing, Signal representations, Tractography and WM pathways.