Article
Computer Science, Artificial Intelligence
Jascha Achterberg, Danyal Akarca, D. J. Strouse, John Duncan, Duncan E. Astle
Summary: This article investigates the resource limitations, structural and functional features, and optimization process of brain networks. By introducing spatially embedded recurrent neural networks (seRNNs), it is found that they share similar features with primate cerebral cortices, revealing the intertwined nature of common structural and functional brain motifs attributed to basic biological optimization processes.
NATURE MACHINE INTELLIGENCE
(2023)
Article
Anatomy & Morphology
Yu-Lin Li, Mou-Xiong Zheng, Xu-Yun Hua, Xin Gao, Jia-Jia Wu, Chun-Lei Shan, Jun-Peng Zhang, Dong Wei, Jian-Guang Xu
Summary: This study aimed to investigate the consistency and diversity between metabolic and structural brain networks at individual level in healthy Chinese population. The F-18-FDG PET and T1-weighted images were used to construct individual brain networks, and graph theoretical analysis was conducted to analyze their topological properties. The results showed that the metabolic and structural networks exhibited small-world architecture, with differences in their assortativity, efficiency, and hierarchy.
BRAIN STRUCTURE & FUNCTION
(2023)
Article
Neurosciences
Farras Abdelnour, Michael Dayan, Orrin Devinsky, Thomas Thesen, Ashish Raj
Summary: The study found that the relationship between brain function and structure in patients with temporal lobe epilepsy and in normal individuals is similar, suggesting that the brain reconfigures and rewires fine-scale connectivity under temporal lobe epilepsy conditions.
Article
Automation & Control Systems
Gilberto Diaz-Garcia, Gabriel Narvaez, Luis Felipe Giraldo, Jairo Giraldo, Alvaro A. Cardenas
Summary: The consensus problem is relevant to various fields and requires addressing implementation issues and malicious agents in the design of a consensus system. This article introduces a formulation to design the topology of a consensus network, improving its resilience to attacks while maintaining sparsity and consistency with the structural relations between agents.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Neurosciences
Janina Wilmskoetter, Xiaosong He, Lorenzo Caciagli, Jens H. Jensen, Barbara Marebwa, Kathryn A. Davis, Julius Fridriksson, Alexandra Basilakos, Lorelei P. Johnson, Chris Rorden, Danielle Bassett, Leonardo Bonilha
Summary: This study used network control theory to evaluate aphasia recovery after stroke. By reconstructing the whole-brain connectome, the researchers found that regional controllability measures were associated with treatment outcomes. The inferior frontal gyrus was the strongest predictor of recovery, outperforming traditional graph theory and demographic measures.
JOURNAL OF NEUROSCIENCE
(2022)
Article
Multidisciplinary Sciences
Beatriz Garcia Santa Cruz, Jan Slter, Gemma Gomez-Giro, Claudia Saraiva, Sonia Sabate-Soler, Jennifer Modamio, Kyriaki Barmpa, Jens Christian Schwamborn, Frank Hertel, Javier Jarazo, Andreas Husch
Summary: The study focuses on addressing the challenges of data acquisition and image analysis in complex disease research. By combining traditional computer vision methods with deep learning, the research team successfully trained a deep learning network and improved the segmentation quality using automatically generated labels. The user-friendly graphical interface allows researchers to evaluate and correct the predictions. Furthermore, the study demonstrates the feasibility of training a deep learning solution on a large dataset of noisy labels.
SCIENTIFIC REPORTS
(2022)
Article
Computer Science, Artificial Intelligence
Ana P. Millan, Joaquin J. Torres, Samuel Johnson, J. Marro
Summary: The study shows that a transient period of relatively high synaptic connectivity is critical for the development of the system under noise circumstances, enabling the recovery of stored memories. Intermediate synaptic densities provide optimal developmental paths with minimum energy consumption, and the transient heterogeneity in the network ultimately determines its evolution.
Article
Neurosciences
Daniel Kristanto, Andrea Hildebrandt, Werner Sommer, Changsong Zhou
Summary: Cognitive neuroscience assumes that different mental abilities correspond to separable brain subnetworks. This study used a bottom-up approach to investigate the association between structural and functional brain subnetworks and domain-specific cognitive abilities. The findings suggest that domain-specific abilities rely on specific combinations of brain subnetworks.
Article
Multidisciplinary Sciences
Madison Lewis, Tales Santini, Nicholas Theis, Brendan Muldoon, Katherine Dash, Jonathan Rubin, Matcheri Keshavan, Konasale Prasad
Summary: This study investigated the structural covariance network (SCN) of first-episode antipsychotic-naive psychosis (FEAP) using graph theoretical methods. The results showed that FEAP patients had higher betweenness centrality (BC) and lower degree in all three morphometric features, suggesting lower network resilience. Moreover, the disintegration of the network with fewer attacks was associated with greater negative symptom severity.
SCIENTIFIC REPORTS
(2023)
Article
Multidisciplinary Sciences
Sergiu P. Pasca, Paola Arlotta, Helen S. Bateup, J. Gray Camp, Silvia Cappello, Fred H. Gage, Juergen A. Knoblich, Arnold R. Kriegstein, Madeline A. Lancaster, Guo-Li Ming, Alysson R. Muotri, In-Hyun Park, Orly Reiner, Hongjun Song, Lorenz Studer, Sally Temple, Giuseppe Testa, Barbara Treutlein, Flora M. Vaccarino
Summary: The nomenclature of human multicellular models of nervous system development and disease, including organoids, assembloids, and transplants, is clarified and provided as a basic framework to facilitate progress and improve communication with the scientific community and the public. These models derived from human pluripotent stem cells or primary tissue have the potential to provide insights into the unique development of the human nervous system and the progression and treatment of nervous system disorders.
Article
Automation & Control Systems
Emma Tegling, Bassam Bamieh, Henrik Sandberg
Summary: We examine distributed consensus in networks with integrator dynamics of order two or higher (n > 2). The paper shows that standard consensus algorithms encounter scale fragilities as the network grows. For high-order agents (n > 3), consensus cannot be achieved in networks of any size using fixed-gain algorithms. For second-order consensus (n = 2), the same scale fragility is observed in directed graphs with a complex Laplacian eigenvalue approaching the origin. The results are proven using Routh-Hurwitz criteria and apply to general directed network graphs. Various classes of graphs subject to these scale fragilities are surveyed, their scaling constants are discussed, and it is shown that a sub-linear scaling of nodal neighborhoods can overcome the issue.
Article
Computer Science, Artificial Intelligence
Lu Zhang, Li Wang, Dajiang Zhu
Summary: Understanding the relationship between brain structure and function is crucial for uncovering organizational principles of the human brain. This study proposes a method called MGCN-GAN, which combines generative adversarial networks and graph convolutional networks, to infer individual structural connectivity based on functional connectivity. Experimental results demonstrate that the model can generate reliable individual structural connectivity across different individuals.
MEDICAL IMAGE ANALYSIS
(2022)
Article
Computer Science, Interdisciplinary Applications
Berardino Barile, Aldo Marzullo, Claudio Stamile, Francoise Durand-Dubief, Dominique Sappey-Marinier
Summary: This study introduces a framework based on generative adversarial network to create synthetic structural brain networks in Multiple Sclerosis (MS). The quality of generated data is comparable to real data, and augmenting the existing dataset with generated samples leads to an improvement in classification performance.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2021)
Article
Neuroimaging
James J. Gugger, Nishant Sinha, Yiming Huang, Alexa E. Walter, Cillian Lynch, Priyanka Kalyani, Nathan Smyk, Danielle Sandsmark, Ramon Diaz-Arrastia, Kathryn A. Davis
Summary: Traumatic brain injury (TBI) leads to diffuse axonal injury and maladaptive changes in network function, resulting in incomplete recovery and persistent disability. Normative modeling can capture network deviations in TBI patients and predict post-TBI symptoms and functional status. Structural network deviation scores could be useful for targeted therapies in clinical trials.
NEUROIMAGE-CLINICAL
(2023)
Article
Neurosciences
Ramana V. Vishnubhotla, Yi Zhao, Qiuting Wen, Jonathan Dietrich, Gregory M. Sokol, Senthilkumar Sadhasivam, Rupa Radhakrishnan
Summary: This study examines brain structural connectivity in infants with prenatal opioid exposure using diffusion tensor imaging (DTI) and graph theoretical modeling. The results show alterations in brain structural connections in these infants. Long-term clinical outcomes related to these findings may be assessed in longitudinal follow-up studies.
FRONTIERS IN NEUROSCIENCE
(2022)
Article
Acoustics
L. Joyeux, J. van der Merwe, M. Aertsen, P. A. Patel, A. Khatoun, M. G. M. C. Mori da Cunha, S. De Vleeschauwer, J. Parra, E. Danzer, M. McLaughlin, D. Stoyanov, T. Vercauteren, S. Ourselin, E. Radaelli, P. de Coppi, F. Van Calenbergh, J. Deprest
Summary: This study aimed to investigate the neurostructural and neurofunctional efficacy of watertight prenatal spina bifida aperta (SBA) repair in a fetal lamb model. The results showed that lambs with watertight repair achieved better neuroprotection and improved brain and spinal cord structure and function compared to lambs without repair. This research has important clinical implications for improving neuroprotection in fetal centers.
ULTRASOUND IN OBSTETRICS & GYNECOLOGY
(2023)
Article
Clinical Neurology
Roberta Messina, Carole H. Sudre, Diana Y. Wei, Massimo Filippi, Sebastien Ourselin, Peter J. Goadsby
Summary: The objective of this study was to identify MRI biomarkers that distinguish between migraine and cluster headache patients, as well as investigate shared imaging features. Clinical, functional, and structural MRI data were collected from 20 migraineurs, 20 cluster headache patients, and 15 healthy controls. Support vector machine algorithms were used to classify headache patients from controls, and regional differences and associations with clinical characteristics were examined. The results showed that MRI could accurately classify headache patients from controls, with accuracies of 80% for all headache patients, 89% for migraine, and 98% for cluster headache. The bilateral hypothalamic and periaqueductal gray functional networks were found to be important in classifying both migraine and cluster headache patients. The presence of restlessness was the most important clinical feature in distinguishing between the two groups.
ANNALS OF NEUROLOGY
(2023)
Article
Medicine, General & Internal
Alex Tsui, Petru-Daniel Tudosiu, Mikael Brudfors, Ashwani Jha, Jorge Cardoso, Sebastien Ourselin, John Ashburner, Geraint Rees, Daniel Davis, Parashkev Nachev
Summary: By studying a group of older patients, it was found that multimodal predictive models based on machine learning can accurately predict long-term mortality. Extracranial bone and soft tissue features contribute more to mortality prediction than intracranial features.
Article
Computer Science, Artificial Intelligence
Reuben Dorent, Aaron Kujawa, Marina Ivory, Spyridon Bakas, Nicola Rieke, Samuel Joutard, Ben Glocker, Jorge Cardoso, Marc Modat, Kayhan Batmanghelich, Arseniy Belkov, Maria Baldeon Calisto, Jae Won Choi, Benoit M. Dawant, Hexin Dong, Sergio Escalera, Yubo Fan, Lasse Hansen, Mattias P. Heinrich, Smriti Joshi, Victoriya Kashtanova, Hyeon Gyu Kim, Satoshi Kondo, Christian N. Kruse, Susana K. Lai-Yuen, Hao Li, Han Liu, Buntheng Ly, Ipek Oguz, Hyungseob Shin, Boris Shirokikh, Zixian Su, Guotai Wang, Jianghao Wu, Yanwu Xu, Kai Yao, Li Zhang, Sebastien Ourselin, Jonathan Shapey, Tom Vercauteren
Summary: Domain Adaptation has recently gained attention in the medical imaging community. To overcome the limitations of existing techniques, this article presents a large-scale multi-class benchmark for unsupervised cross-modality domain adaptation. The benchmark focuses on the segmentation of two key brain structures in vestibular schwannoma cases.
MEDICAL IMAGE ANALYSIS
(2023)
Article
Computer Science, Artificial Intelligence
Guilherme Pombo, Robert Gray, M. Jorge Cardoso, Sebastien Ourselin, Geraint Rees, John Ashburner, Parashkev Nachev
Summary: CounterSynth is a conditional generative model that uses diffeomorphic deformations to induce label driven, biologically plausible changes in volumetric brain images. The model is designed to synthesise counterfactual training data augmentations for discriminative modelling tasks affected by data imbalance, distributional instability, confounding, or underspecification, and exhibit unfair performance across different subpopulations. Through various evaluations, we demonstrate that CounterSynth achieves state-of-the-art improvements in fidelity and equity compared to current solutions. The source code for CounterSynth is available at https://github.com/guilherme-pombo/CounterSynth.
MEDICAL IMAGE ANALYSIS
(2023)
Article
Computer Science, Interdisciplinary Applications
Jieqing Jiao, Fiona Heeman, Rachael Dixon, Catriona Wimberley, Isadora Lopes Alves, Juan Domingo Gispert, Adriaan A. Lammertsma, Bart N. M. van Berckel, Casper da Costa-Luis, Pawel Markiewicz, David M. Cash, M. Jorge Cardoso, Sebastien Ourselin, Maqsood Yaqub, Frederik Barkhof
Summary: This paper introduces a Python-based software package called NiftyPAD for versatile analyses of dynamic PET data. NiftyPAD can handle dual-time window scans, pharmacokinetic modelling, and PET data-based motion correction. It produces comparable results with established software packages and has the advantages of multi-platform usage, modular setup, and lightweight.
Article
Robotics
Zicong Wu, Mikel De Iturrate Reyzabal, S. M. Hadi Sadati, Hongbin Liu, Sebastien Ourselin, Daniel Leff, Robert K. K. Katzschmann, Kawal Rhode, Christos Bergeles
Summary: Soft robots that grow through eversion/apical extension are capable of navigating fragile environments inside the human body. This letter presents a physics-based model of a miniature steerable eversion growing robot. The robot's growing, steering, stiffening, and interaction capabilities are demonstrated. The study investigates the interaction between a steerable catheter and a growing sheath, and the behavior of the growing robot under different pressures and external forces. The simulations conducted within the SOFA framework align with extensive experimentation using a physical robot setup, showing a mean absolute error of 10-20% between simulation and experimental results for curvature values.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Article
Engineering, Biomedical
Charlie Budd, Luis C. Garcia-Peraza C. Herrera, Martin Huber, Sebastien Ourselin, Tom Vercauteren
Summary: This paper addresses the estimation problem of endoscopic content area and proposes two algorithms based on edge detection and circle fitting. A dataset of manually annotated and pseudo-labelled content areas is provided for research. The proposed algorithm shows significant improvement in both accuracy and computational time compared to state-of-the-art methods.
COMPUTER METHODS IN BIOMECHANICS AND BIOMEDICAL ENGINEERING-IMAGING AND VISUALIZATION
(2023)
Article
Engineering, Biomedical
Luis C. Garcia Peraza C. Herrera, Conor Horgan, Sebastien Ourselin, Michael Ebner, Tom Vercauteren
Summary: Traditional RGB imaging poses challenges in visual discrimination of clinical tissue types, while hyperspectral imaging (HSI) provides rich spectral information beyond three-channel RGB imaging. Our study examines the performance of deep learning image segmentation methods when trained on HSI and RGB images, as well as HSI and RGB pixels. Using the Oral and Dental Spectral Image Database (ODSI-DB) with 215 manually segmented dental reflectance spectral images, we emphasize the significance of spectral resolution, range, and spatial information for the development and application of clinical HSI.
COMPUTER METHODS IN BIOMECHANICS AND BIOMEDICAL ENGINEERING-IMAGING AND VISUALIZATION
(2023)
Article
Clinical Neurology
T. Deprest, L. Fidon, F. De Keyzer, M. Ebner, J. Deprest, P. Demaerel, L. De Catte, T. Vercauteren, S. Ourselin, S. Dymarkowski, M. Aertsen
Summary: This study tested an algorithm for segmenting abnormal fetal brains and found that it achieved good results in fetuses with severe brain abnormalities. However, it is necessary to include rare cases in the current dataset and quality control measures should be implemented to prevent occasional errors.
AMERICAN JOURNAL OF NEURORADIOLOGY
(2023)
Article
Computer Science, Information Systems
Nicole Lai-Tan, Marios G. G. Philiastides, Fahim Kawsar, Fani Deligianni
Summary: Music therapy is a successful intervention that improves patient outcomes in neurological and mood disorders, without adverse effects. The interaction of music with brain networks explains its efficacy in motor rehabilitation, emotional regulation, and cardiovascular health. The potential for personalized and automated music selection processes to enhance quality of life and reduce stress requires further exploration.
IEEE PERVASIVE COMPUTING
(2023)
Article
Clinical Neurology
William Coath, Marc Modat, M. Jorge J. Cardoso, Pawel J. A. Markiewicz, Christopher A. D. Lane, Thomas D. Parker, Ashvini M. Keshavan, Sarah M. E. Buchanan, Sarah E. J. Keuss, Matthew J. Harris, Ninon Burgos, John Dickson, Anna L. Barnes, David L. Thomas, Daniel B. Beasley, Ian B. Malone, Andrew Wong, Kjell A. Erlandsson, Benjamin A. Thomas, Michael Scholl, Sebastien Ourselin, Marcus C. Richards, Nick C. M. Fox, Jonathan M. M. Schott, David M. Cash
Summary: The Centiloid scale aims to harmonize Aβ PET measures across different analysis methods. This study investigated the Centiloid transformation with PET/MRI data, finding that the transformation is valid but further understanding of the effects of acquisition or biological factors on using white matter as a reference is needed.
ALZHEIMER'S & DEMENTIA: DIAGNOSIS, ASSESSMENT & DISEASE MONITORING
(2023)
Article
Computer Science, Interdisciplinary Applications
Jianghao Wu, Guotai Wang, Ran Gu, Tao Lu, Yinan Chen, Wentao Zhu, Tom Vercauteren, Sebastien Ourselin, Shaoting Zhang
Summary: Domain Adaptation is crucial for deep learning models in medical image segmentation to handle testing images from new target domains. Source-Free Domain Adaptation (SFDA) is an appealing approach for efficient adaptation to the target domain without source-domain data. However, existing SFDA methods suffer from limited performance due to lack of sufficient supervision with unavailable source-domain images and unlabeled target-domain images. In this study, we propose a novel Uncertainty-aware Pseudo Label guided SFDA method for medical image segmentation, which improves performance by enhancing diversity in the target domain and using reliable pseudo labels.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2023)
Article
Computer Science, Artificial Intelligence
Mark S. Graham, Petru-Daniel Tudosiu, Paul Wright, Walter Hugo Lopez Pinaya, Petteri Teikari, Ashay Patel, Jean-Marie U-King-Im, Yee H. Mah, James T. Teo, Hans Rolf Jager, David Werring, Geraint Rees, Parashkev Nachev, Sebastien Ourselin, M. Jorge Cardoso
Summary: In this study, several segmentation methods with uncertainty were evaluated for the task of segmenting bleeds in 3D CT of the head. The results showed that these models can fail catastrophically in the far out-of-distribution domain, often providing highly confident but incorrect predictions. A method using a latent transformer model for out-of-distribution detection was proposed, which could identify images that are both far and near out-of-distribution, as well as provide spatial maps highlighting the regions considered to be out-of-distribution. Furthermore, a strong relationship between an image's likelihood and the quality of a model's segmentation on it was found, demonstrating the viability of this approach for filtering out unsuitable images.
MEDICAL IMAGE ANALYSIS
(2023)