Article
Computer Science, Information Systems
Jiaofen Nan, Junya Su, Jincan Zhang
Summary: This paper proposes a technique of human brain image registration based on tissue morphology in vivo, which aims to address the problems of previous image registration. Different feature points, including those at the boundary of different brain tissues and those of the maximum or minimum from the original image, are extracted and combined. The correct matching pairs of feature points are used to generate the model parameters of spatial transformation, and the brain image registration is completed by combining interpolation techniques. The proposed method outperforms other algorithms in terms of quantitative indicators and spatial location, size, appearance contour, and registration details.
Article
Cardiac & Cardiovascular Systems
Odgerel Baasan, Omar Freihat, David U. Nagy, Szimonetta Lohner
Summary: This study aimed to assess the methodological quality of published cardiovascular clinical research trials. The results showed that almost two-thirds of the RCTs had high or unclear risk of bias, indicating a need for more rigorous trial planning and conduct.
FRONTIERS IN CARDIOVASCULAR MEDICINE
(2022)
Article
Engineering, Biomedical
Juan Carlos Vizcarra, Erik A. Burlingame, Clemens B. Hug, Yury Goltsev, Brian S. White, Darren R. Tyson, Artem Sokolov
Summary: Emerging multiplexed imaging platforms offer the ability to observe an increasing number of molecular markers at subcellular resolution and track the dynamic evolution of tumor cellular composition. However, the computational methods for handling and analyzing the data generated by these platforms have not kept pace with their rapid development.
COMPUTERIZED MEDICAL IMAGING AND GRAPHICS
(2022)
Review
Biology
Hanguang Xiao, Xufeng Xue, Mi Zhu, Xin Jiang, Qingling Xia, Kai Chen, Huanqi Li, Li Long, Ke Peng
Summary: This review discusses the development and applications of lung image registration, with a focus on deep learning-based methods and their different supervision types. The review also provides a comprehensive analysis of evaluation metrics, loss functions, and publicly available datasets for lung image registration.
COMPUTERS IN BIOLOGY AND MEDICINE
(2023)
Article
Computer Science, Artificial Intelligence
Ruixing Zhang, Tao Yao, Lianshan Yan
Summary: This paper proposes a self-supervised feature point detection and description network called ACPoint, which uses an asymmetric convolution module to train three convolution branches simultaneously. It enhances the backbone of the square convolution from both horizontal and vertical directions to improve the representation of local features. Based on the ACPoint network, a cross-resolution image-matching method is proposed. Experimental results show that the proposed network model achieves higher localization accuracy and homography estimation ability on the HPatches dataset.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2023)
Article
Engineering, Biomedical
Yahya Moshaei-Nezhad, Juliane Mueller, Martin Oelschlaegel, Matthias Kirsch, Ronald Tetzlaff
Summary: The purpose of this study was to compare and analyze six automatic intensity-based registration methods for intraoperative infrared thermography and visible light imaging. The results showed that the combination of normalized intensity, mutual information measure, and one-plus-one evolutionary optimizer with Demon registration achieved better accuracy and performance compared to other methods.
INTERNATIONAL JOURNAL OF COMPUTER ASSISTED RADIOLOGY AND SURGERY
(2022)
Article
Multidisciplinary Sciences
Javier Perez de Frutos, Andre Pedersen, Egidijus Pelanis, David Bouget, Shanmugapriya Survarachakan, Thomas Lango, Ole-Jakob Elle, Frank Lindseth
Summary: This study aims to improve convolutional neural network-based image-to-image deformable registration for abdominal imaging through different training strategies, loss functions, and transfer learning schemes. Results showed that using segmentations for registration guidance in the training step and fine-tuning the pretrained model improved performance without impacting runtime. A proposed augmentation layer and loss layer further enhanced the training process. Future work should validate the framework on different datasets to assess its value.
Article
Chemistry, Analytical
Chengjia Wang, Guang Yang, Giorgos Papanastasiou
Summary: In this study, an unsupervised deep learning registration method is proposed to accurately model affine and non-rigid transformations. The method utilizes bi-directional cross-modality image synthesis and an inverse-consistency loss to address inverse consistency. The model, named FIRE, shows improved performances in multi-modality brain MRI and intra-modality cardiac MRI data experiments. It also demonstrates efficient and topology-preserving image registration directly in the training phase.
Article
Biochemistry & Molecular Biology
Aleix Martinez, Jean-Karim Heriche, Maria Calvo, Christian Tischer, Amaia Otxoa-de-Amezaga, Jordi Pedragosa, Anna Bosch, Anna M. Planas, Valerie Petegnief
Summary: This study developed free access image analysis scripts to quantify microglia morphologies and phagocytosis. The researchers found that inflammation is critical in promoting phenotypical changes in microglia. The method is versatile and useful for studying the relationship between microglia sub-populations and environmental changes.
Article
Computer Science, Artificial Intelligence
Wenzhe Yin, Jan-Jakob Sonke, Efstratios Gavves
Summary: Deformable image registration is crucial in medical image analysis. This paper proposes PC-Reg, a pyramidal Prediction and Correction method for deformable registration, to tackle large deformations. PC-Reg treats multi-scale registration as solving an ordinary differential equation and progressively predicts and corrects the deformation vector field. Experimental results demonstrate that PC-Reg significantly improves registration for both large and small deformations.
MEDICAL IMAGE ANALYSIS
(2023)
Article
Engineering, Biomedical
Soumick Chatterjee, Himanshi Bajaj, Istiyak H. Siddiquee, Nandish Bandi Subbarayappa, Steve Simon, Suraj Bangalore Shashidhar, Oliver Speck, Andreas Nuernberger
Summary: Image registration is widely used in computer vision applications, with deep learning techniques being successfully applied to medical image registration. This paper extends the Voxelmorph approach to improve performance and address specific problems in medical image registration. It integrates multi-scale UNet supervision, self-constructing graph network (SCGNet) for structural co-relations, and cycle consistency loss for inverse-consistency of deformations. The proposed method achieves significant improvements over existing techniques in brain MRI registration, with higher Dice scores compared to ANTs and VoxelMorph.
COMPUTERIZED MEDICAL IMAGING AND GRAPHICS
(2023)
Review
Public, Environmental & Occupational Health
Amanda Diaz-Garcia, Marvin Franke, Rocio Herrero, David Daniel Ebert, Cristina Botella
Summary: The study found that overall effects of online resilience training were not significant, but in studies with a clear assessment theory, there were promising effects on enhancing resilience.
EUROPEAN JOURNAL OF PUBLIC HEALTH
(2021)
Article
Computer Science, Interdisciplinary Applications
Enrique Bermejo, Kei Taniguchi, Yoshinori Ogawa, Ruben Martos, Andrea Valsecchi, Pablo Mesejo, Oscar Ibanez, Kazuhiko Imaizumi
Summary: This study introduced an automatic method for annotating 3D surface skull models, validated through analyzing inter- and intra-observer variability and visual assessment of results. The automatic method proved to be reliable, accurate, robust, and reproducible compared to manual landmarking tasks.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2021)
Article
Multidisciplinary Sciences
Yanling Chi, Yuyu Xu, Huiying Liu, Xiaoxiang Wu, Zhiqiang Liu, Jiawei Mao, Guibin Xu, Weimin Huang
Summary: This work proposed an innovative deep registration pipeline, called KidneyRegNet, for 3D CT and 2D U/S kidney scans of free breathing. The pipeline consists of a feature network and a 3D-2D CNN-based registration network. The feature network utilizes handcrafted texture feature layers to reduce the semantic gap, while the registration network employs an encoder-decoder structure with feature-image-motion (FIM) loss for hierarchical regression. The experiment showed promising results in kidney registration with mean contour distances between 0.94-1.28mm.
SCIENTIFIC REPORTS
(2023)
Editorial Material
Medicine, General & Internal
Julia H. Littell, Dennis M. Gorman
Summary: This article highlights methodological issues in the Campbell Collaboration's new review on school-based anti-bullying interventions, including protocol deviations, inadequate documentation of search strategies, inconsistent reports on the number of included studies, and undisclosed risk of bias ratings. The authors emphasize the importance of transparent methods and adherence to relevant standards in systematic reviews for enhancing the trustworthiness of results and conclusions.
SYSTEMATIC REVIEWS
(2022)
Correction
Mathematical & Computational Biology
John Muschelli, Adrian Gherman, Jean-Philippe Fortin, Brian Avants, Brandon Whitcher, Jonathan D. Clayden, Brian S. Caffo, Ciprian M. Crainiceanu
Article
Critical Care Medicine
James R. Stone, Brian B. Avants, Nicholas J. Tustison, Eric M. Wassermann, Jessica Gill, Elena Polejaeva, Kristine C. Dell, Walter Carr, Angela M. Yarnell, Matthew L. LoPresti, Peter Walker, Meghan O'Brien, Natalie Domeisen, Alycia Quick, Claire M. Modica, John D. Hughes, Francis J. Haran, Carl Goforth, Stephen T. Ahlers
JOURNAL OF NEUROTRAUMA
(2020)
Article
Critical Care Medicine
Anupa Ambili Vijayakumari, Drew Parker, Yusuf Osmanlioglu, Jacob A. Alappatt, John Whyte, Ramon Diaz-Arrastia, Junghoon J. Kim, Ragini Verma
Summary: This study utilized FW DTI to estimate free water volume fraction in moderate-to-severe TBI patients and found its association with injury severity and long-term outcomes. The study demonstrated that MVF at 3 months significantly predicted functional outcome, executive function, and processing speed at 12 months, and was correlated with injury severity. These findings are a significant step towards using MVF as a biomarker for personalized therapy and rehabilitation planning in TBI patients.
JOURNAL OF NEUROTRAUMA
(2021)
Article
Critical Care Medicine
Linda Xu, Jeffrey B. Ware, Junghoon J. Kim, Pashtun Shahim, Erika Silverman, Brigid Magdamo, Cian Dabrowski, Leroy Wesley, My Duyen Le, Justin Morrison, Hannah Zamore, Cillian E. Lynch, Dmitriy Petrov, H. Isaac Chen, James Schuster, Ramon Diaz-Arrastia, Danielle K. Sandsmark
Summary: Imaging detection of brain perfusion alterations after traumatic brain injury (TBI) may provide prognostic insights. The study found that TBI participants had globally elevated CBF in both acute and chronic phases, with focal hypo- and hyperperfusion present in brain tissue. Furthermore, acute elevation in CBF post-TBI may play a reparative role and be associated with better clinical outcomes in the long term.
JOURNAL OF NEUROTRAUMA
(2021)
Article
Radiology, Nuclear Medicine & Medical Imaging
Nicholas J. Tustison, Talissa A. Altes, Kun Qing, Mu He, G. Wilson Miller, Brian B. Avants, Yun M. Shim, James C. Gee, John P. Mugler, Jaime F. Mata
Summary: This study compared histogram-based and image-based algorithms for segmentation of hyperpolarized gas lung images. The results showed that the image-based convolutional neural network outperformed histogram-based methods in terms of stability and precision.
MAGNETIC RESONANCE IN MEDICINE
(2021)
Article
Psychology, Clinical
Lindsay D. Nelson, Mark D. Kramer, Keanan J. Joyner, Christopher J. Patrick, Murray B. Stein, Nancy Temkin, Harvey S. Levin, John Whyte, Amy J. Markowitz, Joseph Giacino, Geoffrey T. Manley
Summary: Through analyzing neuropsychiatric symptoms in patients with traumatic brain injury (TBI) and orthopedic-injured trauma controls, it was found that there are distinct dimensions of symptoms, including internalizing factors (depression, anxiety, fear) and somatic symptoms (sleep, physical, pain). The study suggests that brain injury, especially milder forms, may exacerbate these symptoms.
JOURNAL OF ABNORMAL PSYCHOLOGY
(2021)
Article
Neurosciences
Andrew A. Chen, Joanne C. Beer, Nicholas J. Tustison, Philip A. Cook, Russell T. Shinohara, Haochang Shou
Summary: In order to address discrepancies in neuroimaging data acquired from different research sites, efforts have been made to harmonize the data by removing site-related effects in the mean and variance. Additionally, the utilization of machine learning in neuroimaging has become increasingly popular, providing improved sensitivity and specificity due to modeling the joint relationship across brain measurements. Researchers have proposed a novel method called Correcting Covariance Batch Effects (CovBat) that removes site effects in mean, variance, and covariance, demonstrating successful harmonization of within-site correlation matrices and accurate disease group prediction after harmonization.
HUMAN BRAIN MAPPING
(2022)
Letter
Respiratory System
Mu He, Kun Qing, Nicholas J. Tustison, Zach Beaulac, Tabitha W. King, Thomas B. Huff, Mikell Paige, Kranthikiran Earasi, Roselove Nunoo-Asare, Sarah Struchen, Marie Burdick, Zhimin Zhang, Alan Ropp, Grady W. Miller, James Patrie, Jaime F. Mata, John P. Mugler, Yun Michael Shim
INTERNATIONAL JOURNAL OF CHRONIC OBSTRUCTIVE PULMONARY DISEASE
(2021)
Letter
Rehabilitation
John Whyte, Jarrad Van Stan, Lyn Turkstra
ARCHIVES OF PHYSICAL MEDICINE AND REHABILITATION
(2022)
Editorial Material
Clinical Neurology
John Whyte
DEVELOPMENTAL MEDICINE AND CHILD NEUROLOGY
(2022)
Article
Neurosciences
Yusuf Osmanlioglu, Drew Parker, Jacob A. Alappatt, James J. Gugger, Ramon R. Diaz-Arrastia, John Whyte, Junghoon J. Kim, Ragini Verma
Summary: Traumatic brain injury (TBI) is a major public health problem characterized by the shearing of axons across the white matter, leading to cognitive deficits. Assessing network-wide structural connectivity disruptions in TBI is necessary for personalized treatment and rehabilitation planning. A novel connectomic measure called network normality score (NNS) captures the integrity of structural connectivity in TBI patients by leveraging the diffuseness of axonal injury and the heterogeneity of the disease.
HUMAN BRAIN MAPPING
(2022)
Article
Biochemistry & Molecular Biology
Kun Qing, Talissa A. Altes, John P. Mugler III, Jaime F. Mata, Nicholas J. Tustison, Kai Ruppert, Juliana Bueno, Lucia Flors, Yun M. Shim, Li Zhao, Joanne Cassani, William G. Teague, John S. Kim, Zhixing Wang, Iulian C. Ruset, F. William Hersman, Borna Mehrad
Summary: The use of hyperpolarized xenon-129 gas as a contrast agent in lung MRI imaging is a promising technique for assessing lung function and detecting impaired lung physiology in UIP patients.
Article
Medicine, General & Internal
Jay Hegde, Nicholas J. Tustison, William T. Parker, Fallon Branch, Nathan Yanasak, Lorie A. Stumpo
Summary: During medical image analysis, aligning a given image of the body part to a representative standard is useful. While brain templates have had a significant impact on brain image analysis, templates for human hands do not exist. This study reports the construction of an anatomical template for healthy adult human hands using T1-weighted magnetic resonance images.
Article
Neuroimaging
Naomi L. Gaggi, Jeffrey B. Ware, Sudipto Dolui, Daniel Brennan, Julia Torrellas, Ze Wang, John Whyte, Ramon Diaz-Arrastia, Junghoon J. Kim
Summary: Traumatic brain injury (TBI) is associated with alterations in cerebral blood flow (CBF), which may underlie functional disability and precipitate TBI-induced neurodegeneration. Using arterial spin labeled (ASL) perfusion magnetic resonance imaging (MRI), longitudinal CBF changes in 29 patients with moderate-severe TBI (msTBI) were examined at 3, 6, and 12 months post-injury. The results showed lower mean CBF in gray matter (GM) in the TBI group compared to healthy controls (HC) at 6 and 12 months post-injury. Regional decreases in CBF were observed from 3 to 6 months post-injury in the TBI group, indicating stabilization of hypoperfusion. CBF changes were correlated with change in executive function from 3 to 6 months post-injury in TBI patients.
NEUROIMAGE-CLINICAL
(2023)
Article
Neurosciences
Adam Kimbler, Dana L. McMakin, Nicholas J. Tustison, Aaron T. Mattfeld
Summary: The development of the hippocampal formation during puberty plays a critical role in negative overgeneralization, a common feature of anxiety disorders. This study investigated the relationship between mnemonic generalization and pubertal maturity using MRI scans and recognition tasks. The findings suggest a developmental balance between hippocampal functioning and its connections with other regions, with maturational differences potentially contributing to negative overgeneralization during peripuberty.