Pub Date : 2023-01-01DOI: 10.1016/j.brain.2023.100064
Amedeo D'Angiulli , Matthew F. Kirby , Dao A.T. Pham , Gary Goldfield
Dipole source localization analysis (DSLA) of brain's event-related electrical potentials (ERPs) often presumes time constraints potentially too rigid to capture complex neural dynamics. We present a practical procedure (dynamically-guided DSLA) combining in a novel way well-established off-the-shelf modeling (Independent Component Analysis, and proprietary software modules running on MATLAB, such as FASTICA and EEG-Lab DIPFIT) with the cognitive modeling simulation framework tool known as Adaptive Control of Thought-Rational (ACT-R). The integration of these multiple methods can narrow down the time-windows of interest for DSLA more flexibly. As a demonstration, we used dynamically-guided DSLA to re-analyze cluster-level ERPs from a visual target detection task involving the participation of 26 preschool children. The key analytic features were dynamic ERP movies vis-à-vis validating ACT-R simulation of comparison adult data for the same task. Spatial topography for the six estimated sources did not differ significantly in children's and adult simulated data, which generally showed high fit (predicted R2 > 0.97). A control comparison using the static DSLA showed discrepant fits for two sources, suggesting that dynamic DSLA may offer higher discriminant reliability. Given its high validity, flexibility and relative user-friendliness, dynamically-guided DSLA seems useful for assessing developmental homology and may be suitable for a variety of clinical and experimental applications specifically involving neurodevelopmental data.
Statement of Significance
Accurately determining the location of neural activity observed via electroencephalogram remains a well-known challenge. Under a variety of conditions, conventional dipole source localization analysis methodologies can result in underqualified data. In this work we present a novel process, known as dynamically-guided DSLA, which demonstrates how pre-existing tools can be appropriated to facilitate the examination and analysis of neurological activity in preschool-aged children. Because the effects exerted by a stimulus or event on EEG signals can be linked to behaviors and actions, at different levels of physical mechanisms of different degree of complexity, this neuroimaging tool offers the opportunity to cut across multiple layers of physical systems underlying cognitive and emotional functions, and therefore can be leveraged to reach invaluable insights. We highlight how the proposed technique can help link the electrophysiology to underlying physical alteration (e.g., neurodevelopmental disease); and how the proposed combination of methodologies can help "reverse engineer" physical defects or anomalies (and their locations) to quantify the EEG measurements in terms of dynamic interactive physical phenomena (movie of topographically mapped brain activity), as opposed to just giving a number against a disease or identifying a brain
{"title":"Making movies of children's cortical electrical potentials: A practical procedure for dynamic source localization analysis with validating simulation","authors":"Amedeo D'Angiulli , Matthew F. Kirby , Dao A.T. Pham , Gary Goldfield","doi":"10.1016/j.brain.2023.100064","DOIUrl":"10.1016/j.brain.2023.100064","url":null,"abstract":"<div><p>Dipole source localization analysis (DSLA) of brain's event-related electrical potentials (<em>ERPs</em>) often presumes time constraints potentially too rigid to capture complex neural dynamics. We present a practical procedure (<em>dynamically-guided DSLA</em>) combining in a novel way well-established off-the-shelf modeling (Independent Component Analysis, and proprietary software modules running on MATLAB, such as FASTICA and EEG-Lab DIPFIT) with the cognitive modeling simulation framework tool known as Adaptive Control of Thought-Rational (ACT-R). The integration of these multiple methods can narrow down the time-windows of interest for DSLA more flexibly. As a demonstration, we used dynamically-guided DSLA to re-analyze cluster-level ERPs from a visual target detection task involving the participation of 26 preschool children. The key analytic features were dynamic ERP movies vis-à-vis validating ACT-R simulation of comparison adult data for the same task. Spatial topography for the six estimated sources did not differ significantly in children's and adult simulated data, which generally showed high fit (predicted R<sup>2</sup> > 0.97). A control comparison using the static DSLA showed discrepant fits for two sources, suggesting that dynamic DSLA may offer higher discriminant reliability. Given its high validity, flexibility and relative user-friendliness, dynamically-guided DSLA seems useful for assessing developmental homology and may be suitable for a variety of clinical and experimental applications specifically involving neurodevelopmental data.</p></div><div><h3>Statement of Significance</h3><p>Accurately determining the location of neural activity observed via electroencephalogram remains a well-known challenge. Under a variety of conditions, conventional dipole source localization analysis methodologies can result in underqualified data. In this work we present a novel process, known as dynamically-guided DSLA, which demonstrates how pre-existing tools can be appropriated to facilitate the examination and analysis of neurological activity in preschool-aged children. Because the effects exerted by a stimulus or event on EEG signals can be linked to behaviors and actions, at different levels of physical mechanisms of different degree of complexity, this neuroimaging tool offers the opportunity to cut across multiple layers of physical systems underlying cognitive and emotional functions, and therefore can be leveraged to reach invaluable insights. We highlight how the proposed technique can help link the electrophysiology to underlying physical alteration (e.g., neurodevelopmental disease); and how the proposed combination of methodologies can help \"reverse engineer\" physical defects or anomalies (and their locations) to quantify the EEG measurements in terms of dynamic interactive physical phenomena (movie of topographically mapped brain activity), as opposed to just giving a number against a disease or identifying a brain","PeriodicalId":72449,"journal":{"name":"Brain multiphysics","volume":"4 ","pages":"Article 100064"},"PeriodicalIF":0.0,"publicationDate":"2023-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42384061","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-01DOI: 10.1016/j.brain.2023.100077
Mohammadreza Ramzanpour , Bahram Jafari , Jeremy Smith , Jason Allen , Marzieh Hajiaghamemar
{"title":"Comprehensive study of sex-based anatomical variations of human brain and development of sex-specific brain templates","authors":"Mohammadreza Ramzanpour , Bahram Jafari , Jeremy Smith , Jason Allen , Marzieh Hajiaghamemar","doi":"10.1016/j.brain.2023.100077","DOIUrl":"https://doi.org/10.1016/j.brain.2023.100077","url":null,"abstract":"","PeriodicalId":72449,"journal":{"name":"Brain multiphysics","volume":"4 ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49776401","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-01DOI: 10.1016/j.brain.2023.100070
Martin Brennan, Peter Grindrod CBE
We consider a class of Kuramoto models, with an array of individual -dimensional clocks , coupled within a directed, range dependent, network. For each directed connection, a signal triggered at the sending clock incurs a (real valued) time delay before arriving at the receiving clock, where it induces an instantaneous phase reset affecting all -phases. Instantaneous phase resetting maps (PRMs) for -dimensional clocks have received little attention. The system may be treated as open and subject to periodic, or other types of, PRM forcing at any individual clock, as a result of external forcing stimuli. We show how the full system, with phase variables, responds to such stimuli, as the impact spreads across the entire network. Beyond simulations, we employ methods to reverse engineer the dynamical behaviour of the whole: estimating the intrinsic dimensions of the responses to different experiments; and by analysing pairwise comparisons between those responses. This shows that the system’s responses are governed by a hierarchy of internal dynamical modes, existing across both the phases and over time.
We argue that this Kuramoto system is a model for the human cortex, where each -dimensional clock models the dynamics of a single neural column, which contains 10,000 densely inter-connected neurons. The Kuramoto model exploits the natural network of networks architecture of the human cortex. An array of 1M such columns/clocks is at the 10B neuron scale of the human cortex. However its simulation is far more accessible than very large scale (VLS) simulations of neuron-to-neuron systems on supercomputers. The latent modes may have important implications for cognition (information processing) and for consciousness (the existence of internal phenomenological experiences). We argue that the existence of the latter plays a key role in preconditioning the former, reducing the decision sets and the cognitive load, and thus enabling a fast-thinking evolutionary advantage.
This is the first time that systems of -dimensional clocks ( 1), coupled via time-lagged PRMs, within range dependent networks, have been deployed to demonstrate the basic internal phenomenological elements (of consciousness) and their potential role within immediate cognition.
Statement of Significance: We argue that this Kuramoto system is a model for the human cortex, whe
{"title":"Generalised Kuramoto models with time-delayed phase-resetting for k-dimensional clocks","authors":"Martin Brennan, Peter Grindrod CBE","doi":"10.1016/j.brain.2023.100070","DOIUrl":"https://doi.org/10.1016/j.brain.2023.100070","url":null,"abstract":"<div><p>We consider a class of Kuramoto models, with an array of <span><math><mi>N</mi></math></span> individual <span><math><mi>k</mi></math></span>-dimensional clocks <span><math><mrow><mo>(</mo><mi>k</mi><mo>></mo><mn>1</mn><mo>)</mo></mrow></math></span>, coupled within a directed, range dependent, network. For each directed connection, a signal triggered at the sending clock incurs a (real valued) time delay before arriving at the receiving clock, where it induces an instantaneous phase reset affecting all <span><math><mi>k</mi></math></span>-phases. Instantaneous phase resetting maps (PRMs) for <span><math><mi>k</mi></math></span>-dimensional clocks have received little attention. The system may be treated as <em>open</em> and subject to periodic, or other types of, PRM forcing at any individual clock, as a result of external forcing stimuli. We show how the full system, with <span><math><mrow><mi>N</mi><mi>k</mi></mrow></math></span> phase variables, responds to such stimuli, as the impact spreads across the entire network. Beyond simulations, we employ methods to reverse engineer the dynamical behaviour of the whole: estimating the intrinsic dimensions of the responses to different experiments; and by analysing pairwise comparisons between those responses. This shows that the system’s responses are governed by a hierarchy of internal dynamical <em>modes</em>, existing across both the <span><math><mrow><mi>N</mi><mi>k</mi></mrow></math></span> phases and over time.</p><p>We argue that this Kuramoto system is a model for the human cortex, where each <span><math><mi>k</mi></math></span>-dimensional clock models the dynamics of a single <em>neural column</em>, which contains 10,000 densely inter-connected neurons. The Kuramoto model exploits the natural <em>network of networks</em> architecture of the human cortex. An array of <span><math><mrow><mi>N</mi><mo>=</mo></mrow></math></span>1M such columns/clocks is at the 10B neuron scale of the human cortex. However its simulation is far more accessible than very large scale (VLS) simulations of neuron-to-neuron systems on supercomputers. The latent modes may have important implications for cognition (information processing) and for consciousness (the existence of internal phenomenological experiences). We argue that the existence of the latter plays a key role in preconditioning the former, reducing the decision sets and the cognitive load, and thus enabling a fast-thinking evolutionary advantage.</p><p>This is the first time that systems of <span><math><mi>k</mi></math></span>-dimensional clocks (<span><math><mrow><mi>k</mi><mo>></mo></mrow></math></span> 1), coupled via time-lagged PRMs, within range dependent networks, have been deployed to demonstrate the basic internal phenomenological elements (of consciousness) and their potential role within immediate cognition.</p><p><strong>Statement of Significance</strong>: We argue that this Kuramoto system is a model for the human cortex, whe","PeriodicalId":72449,"journal":{"name":"Brain multiphysics","volume":"4 ","pages":"Article 100070"},"PeriodicalIF":0.0,"publicationDate":"2023-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49817687","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
An aneurysm is a medical condition where a section of the artery bulges out under high pressure. Patients suffering from aneurysm rupture have a mortality rate of around 20% and a morbidity rate of up to 40%. The present imaging methods, such as MRI and CT scans, only offer geometrical information on the aneurysm and cannot predict the risk of rupture associated with its progression. To address this gap, a novel computational modeling framework was developed to describe aneurysm growth and analyze the rupture risk under varying pressure loading conditions. The aneurysms were modeled at the vulnerable posterior cerebral artery (PCA) and posterior communicating artery (PCoA), extracted using image segmentation. Five different aneurysm diameters and two wall thicknesses were considered to simulate different phases of aneurysm progression. The realistic pressure loadings on the posterior cerebral arteries were described using three pressures (diastolic, systolic, and hypertensive), and the stress distributions across all models were evaluated to estimate the rupture risk. For PCA, the value of max. von-Mises stress varied between 5.334 MPa and 13.324 MPa for different models with wall thickness of 0.05 mm and from 2.579 MPa to 7.582 MPa for 0.1 mm wall thickness models. For PCoA, the value of max. von-Mises stress ranged from 2.073 MPa to 11.383 MPa for artery-aneurysm models with 0.075 mm thickness and from 2.817 MPa to 10.779 MPa for artery-aneurysm models with 0.15 mm thickness. It was found that the stress values on the aneurysm walls significantly varies with change in blood pressure and aneurysm diameter. An aneurysm with a large diameter and thin wall was also observed to pose a significant risk of rupture, particularly at high blood pressures. These results are expected to provide valuable information to the medical practitioners and help in the prediction of rupture risks using image analysis of aneurysm size and in making timely treatment decisions.
Statement of Significance
The points of significance of our work are:
•
A novel computational modeling framework to evaluate the aneurysm growth and analyze the rupture risk.
•
The realistic pressure loadings conditions (i.e., diastolic, systolic, and hypertensive) of the cardiac cycle were considered and the stress distributions were evaluated to estimate the rupture risk.
•
To date, such extensive research on cerebral aneurysms has not been reported. The results are anticipated to provide valuable information to the medical practitioners in predicting the rupture risks using structural parameters of the aneurysm.
{"title":"Biomechanical modeling of aneurysm in posterior cerebral artery and posterior communicating artery: Progression and rupture risk","authors":"Gurpreet Singh , Prem Nath Yadav , Shubham Gupta , Arnab Chanda","doi":"10.1016/j.brain.2023.100069","DOIUrl":"https://doi.org/10.1016/j.brain.2023.100069","url":null,"abstract":"<div><p>An aneurysm is a medical condition where a section of the artery bulges out under high pressure. Patients suffering from aneurysm rupture have a mortality rate of around 20% and a morbidity rate of up to 40%. The present imaging methods, such as MRI and CT scans, only offer geometrical information on the aneurysm and cannot predict the risk of rupture associated with its progression. To address this gap, a novel computational modeling framework was developed to describe aneurysm growth and analyze the rupture risk under varying pressure loading conditions. The aneurysms were modeled at the vulnerable posterior cerebral artery (PCA) and posterior communicating artery (PCoA), extracted using image segmentation. Five different aneurysm diameters and two wall thicknesses were considered to simulate different phases of aneurysm progression. The realistic pressure loadings on the posterior cerebral arteries were described using three pressures (diastolic, systolic, and hypertensive), and the stress distributions across all models were evaluated to estimate the rupture risk. For PCA, the value of max. von-Mises stress varied between 5.334 MPa and 13.324 MPa for different models with wall thickness of 0.05 mm and from 2.579 MPa to 7.582 MPa for 0.1 mm wall thickness models. For PCoA, the value of max. von-Mises stress ranged from 2.073 MPa to 11.383 MPa for artery-aneurysm models with 0.075 mm thickness and from 2.817 MPa to 10.779 MPa for artery-aneurysm models with 0.15 mm thickness. It was found that the stress values on the aneurysm walls significantly varies with change in blood pressure and aneurysm diameter. An aneurysm with a large diameter and thin wall was also observed to pose a significant risk of rupture, particularly at high blood pressures. These results are expected to provide valuable information to the medical practitioners and help in the prediction of rupture risks using image analysis of aneurysm size and in making timely treatment decisions.</p></div><div><h3>Statement of Significance</h3><p>The points of significance of our work are:</p><ul><li><span>•</span><span><p>A novel computational modeling framework to evaluate the aneurysm growth and analyze the rupture risk.</p></span></li><li><span>•</span><span><p>The realistic pressure loadings conditions (i.e., diastolic, systolic, and hypertensive) of the cardiac cycle were considered and the stress distributions were evaluated to estimate the rupture risk.</p></span></li><li><span>•</span><span><p>To date, such extensive research on cerebral aneurysms has not been reported. The results are anticipated to provide valuable information to the medical practitioners in predicting the rupture risks using structural parameters of the aneurysm.</p></span></li></ul></div>","PeriodicalId":72449,"journal":{"name":"Brain multiphysics","volume":"4 ","pages":"Article 100069"},"PeriodicalIF":0.0,"publicationDate":"2023-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49817686","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-01DOI: 10.1016/j.brain.2023.100073
Zhou Zhou , Xiaogai Li , Yuzhe Liu , Warren N. Hardy , Svein Kleiven
Traumatic brain injury (TBI) is an alarming global public health issue with high morbidity and mortality rates. Although the causal link between external insults and consequent brain injury remains largely elusive, both strain and strain rate are generally recognized as crucial factors for TBI onsets. With respect to the flourishment of strain-based investigation, ambiguity and inconsistency are noted in the scheme for strain rate calculation within the TBI research community. Furthermore, there is no experimental data that can be used to validate the strain rate responses of finite element (FE) models of the human brain. The current work presented a theoretical clarification of two commonly used strain rate computational schemes: the strain rate was either calculated as the time derivative of strain or derived from the rate of deformation tensor. To further substantiate the theoretical disparity, these two schemes were respectively implemented to estimate the strain rate responses from a previous-published cadaveric experiment and an FE head model secondary to a concussive impact. The results clearly showed scheme-dependent responses, both in the experimentally determined principal strain rate and model-derived principal and tract-oriented strain rates. The results highlight that cross-scheme comparison of strain rate responses is inappropriate, and the utilized strain rate computational scheme needs to be reported in future studies. The newly calculated experimental strain rate curves in the supplementary material can be used for strain rate validation of FE head models.
Statement of significance
–
Delineates a theoretical clarification of two algorithms for strain rate computation.
–
Highlights the strain rate responses directly depends on the computational schemes.
–
Presents experimental strain rate curves, serving as references for strain rate validation of finite element head models.
{"title":"Brain strain rate response: Addressing computational ambiguity and experimental data for model validation","authors":"Zhou Zhou , Xiaogai Li , Yuzhe Liu , Warren N. Hardy , Svein Kleiven","doi":"10.1016/j.brain.2023.100073","DOIUrl":"https://doi.org/10.1016/j.brain.2023.100073","url":null,"abstract":"<div><p>Traumatic brain injury (TBI) is an alarming global public health issue with high morbidity and mortality rates. Although the causal link between external insults and consequent brain injury remains largely elusive, both strain and strain rate are generally recognized as crucial factors for TBI onsets. With respect to the flourishment of strain-based investigation, ambiguity and inconsistency are noted in the scheme for strain rate calculation within the TBI research community. Furthermore, there is no experimental data that can be used to validate the strain rate responses of finite element (FE) models of the human brain. The current work presented a theoretical clarification of two commonly used strain rate computational schemes: the strain rate was either calculated as the time derivative of strain or derived from the rate of deformation tensor. To further substantiate the theoretical disparity, these two schemes were respectively implemented to estimate the strain rate responses from a previous-published cadaveric experiment and an FE head model secondary to a concussive impact. The results clearly showed scheme-dependent responses, both in the experimentally determined principal strain rate and model-derived principal and tract-oriented strain rates. The results highlight that cross-scheme comparison of strain rate responses is inappropriate, and the utilized strain rate computational scheme needs to be reported in future studies. The newly calculated experimental strain rate curves in the supplementary material can be used for strain rate validation of FE head models.</p></div><div><h3>Statement of significance</h3><p></p><ul><li><span>–</span><span><p>Delineates a theoretical clarification of two algorithms for strain rate computation.</p></span></li><li><span>–</span><span><p>Highlights the strain rate responses directly depends on the computational schemes.</p></span></li><li><span>–</span><span><p>Presents experimental strain rate curves, serving as references for strain rate validation of finite element head models.</p></span></li></ul></div>","PeriodicalId":72449,"journal":{"name":"Brain multiphysics","volume":"4 ","pages":"Article 100073"},"PeriodicalIF":0.0,"publicationDate":"2023-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49817688","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-01-01DOI: 10.1016/j.brain.2022.100049
Rodrigo Dalvit Carvalho da Silva , Thomas Richard Jenkyn , Victor Alexander Carranza
This paper introduces an orthogonal moment pre-processing method to enhance convolutional neural network outcomes for whole brain image segmentation in magnetic resonance images. The method implements kernel windows based on orthogonal moments to transform the original image into a modified version with orthogonal moment properties. The transformed image contains the optimal representation of the coefficients of the Legendre, Tchebichef and Pseudo-Zernike moments. The approach was evaluated on three distinct datasets; NFBS, OASIS, and TCIA, and obtained an improvement of 4.12%, 1.91%, and 1.05%, respectively. A further investigation employing transfer learning using orthogonal moments of various orders and repetitions, achieved an improvement of 9.86% and 7.76% on the NFBS and OASIS datasets, respectively, when trained using the TCIA dataset. In addition, the best image representations were used to compare different convolutional neural network architectures, including U-Net, U-Net++, and U-Net3+. U-Net3+ demonstrated a slight improvement over U-Net in an overall accuracy of 0.64 % for the original image and 0.33 % for the modified orthogonal moment image.
Statement of Significance
This manuscript introduces a method to initialize convolutional neural network using orthogonal moment filters for whole brain image segmentation in magnetic resonance images. Three orthogonal moments were selected and tests were performed in three distinct datasets. Also, the comparison of three different convolutional neural network (U-Net, U-Net++, and U-Net3+) were conducted. The use of an initial orthogonal moment filter for convolutional neural network in brain segmentation in magnetic resonance imaging achieved an improvement over conventional method. The findings in this study contribute to the long-standing search for the development of a pre-processing technique for whole brain segmentation in MRI.
本文介绍了一种正交矩预处理方法，以增强卷积神经网络在磁共振图像全脑图像分割中的效果。该方法实现了基于正交矩的核窗口，将原始图像转化为具有正交矩属性的改进图像。变换后的图像包含了Legendre、chebichef和Pseudo-Zernike矩系数的最优表示。该方法在三个不同的数据集上进行了评估;NFBS、OASIS和TCIA，分别获得4.12%、1.91%和1.05%的改善。在进一步的研究中，使用不同阶数和重复的正交矩进行迁移学习，在使用TCIA数据集训练时，NFBS和OASIS数据集的效率分别提高了9.86%和7.76%。此外，使用最佳图像表示来比较不同的卷积神经网络架构，包括U-Net, u - net++和U-Net3+。U-Net3+对原始图像的总体精度为0.64%，对改进的正交矩图像的总体精度为0.33%，比U-Net略有提高。本文介绍了一种利用正交矩滤波器初始化卷积神经网络的方法，用于磁共振图像的全脑图像分割。选择三个正交矩，在三个不同的数据集上进行测试。同时，对三种不同的卷积神经网络(U-Net、u - net++和U-Net3+)进行了比较。将初始正交矩滤波器应用于卷积神经网络在磁共振成像脑区分割中，实现了对传统方法的改进。本研究的发现有助于长期以来对MRI全脑分割预处理技术发展的研究。
{"title":"Enhanced pre-processing for deep learning in MRI whole brain segmentation using orthogonal moments","authors":"Rodrigo Dalvit Carvalho da Silva , Thomas Richard Jenkyn , Victor Alexander Carranza","doi":"10.1016/j.brain.2022.100049","DOIUrl":"10.1016/j.brain.2022.100049","url":null,"abstract":"<div><p>This paper introduces an orthogonal moment pre-processing method to enhance convolutional neural network outcomes for whole brain image segmentation in magnetic resonance images. The method implements kernel windows based on orthogonal moments to transform the original image into a modified version with orthogonal moment properties. The transformed image contains the optimal representation of the coefficients of the Legendre, Tchebichef and Pseudo-Zernike moments. The approach was evaluated on three distinct datasets; NFBS, OASIS, and TCIA, and obtained an improvement of 4.12%, 1.91%, and 1.05%, respectively. A further investigation employing transfer learning using orthogonal moments of various orders and repetitions, achieved an improvement of 9.86% and 7.76% on the NFBS and OASIS datasets, respectively, when trained using the TCIA dataset. In addition, the best image representations were used to compare different convolutional neural network architectures, including U-Net, U-Net++, and U-Net3+. U-Net3+ demonstrated a slight improvement over U-Net in an overall accuracy of 0.64 % for the original image and 0.33 % for the modified orthogonal moment image.</p></div><div><h3>Statement of Significance</h3><p>This manuscript introduces a method to initialize convolutional neural network using orthogonal moment filters for whole brain image segmentation in magnetic resonance images. Three orthogonal moments were selected and tests were performed in three distinct datasets. Also, the comparison of three different convolutional neural network (U-Net, U-Net++, and U-Net3+) were conducted. The use of an initial orthogonal moment filter for convolutional neural network in brain segmentation in magnetic resonance imaging achieved an improvement over conventional method. The findings in this study contribute to the long-standing search for the development of a pre-processing technique for whole brain segmentation in MRI.</p></div>","PeriodicalId":72449,"journal":{"name":"Brain multiphysics","volume":"3 ","pages":"Article 100049"},"PeriodicalIF":0.0,"publicationDate":"2022-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S2666522022000065/pdfft?md5=dc66cbee129394224b3c6fc7c58fe9d5&pid=1-s2.0-S2666522022000065-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"41520868","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-01-01DOI: 10.1016/j.brain.2022.100059
David B. MacManus , Mazdak Ghajari
Finite element (FE) brain models have revolutionised research on the biomechanics of traumatic brain injury (TBI). The accuracy and reliability of results arising from brain models depend equally on their geometric accuracy as the quality of the material properties used to describe the mechanical behaviour of brain. However, much of the literature on human brain tissues’ material properties have been performed at low strain rates and strains. This is particularly striking considering a large portion of the brain tissue mechanical characterisation literature is presented with a motivation of understanding brain tissues’ behaviour during TBI which occurs due to brain tissues’ exposure to large strains at high strain rates. Therefore, the aim of this review is to collate the mechanical characterisation studies on human brain tissue under conditions suitable for modelling TBI. We first review injury threshold studies and show that ≥20% strain at ≥10/s strain rate is a reasonable minimum threshold for producing injury to the brain. Using this threshold, we show that there are only five studies on the mechanical characterisation of human brain tissue under strains at strain rates relevant to TBI. These studies, provide material properties of human brain tissue at moderate and high rate loading, with only a recent study showing its region dependent characteristics. This review acts as a reference for scientists and engineers to select suitable material data when modelling human TBI. It also calls for more research to provide high fidelity material properties for modelling of TBI.
Statement of significance
The significance of this work is underscored by the reporting of brain tissues’ material properties in the context of traumatic brain injury (TBI) despite these properties having been characterised under strains and strain rates that are not relevant to TBI. This can result in inaccurate results if implemented in finite element brain models. Here, we address this problem by performing a review on the mechanical characterisation of human brain tissue under conditions that are suitable for modelling human TBI. Our findings show that there are only five studies on the mechanical characterisation of human brain tissue under strains at strain rate levels relevant to TBI. These results will allow researchers to select appropriate material properties for modelling human TBI providing more realistic behaviour of brain tissue in simulations. These results also provide minimum strain and strain rate values for mechanical characterisation experiments on brain tissue for TBI applications. Furthermore, our findings highlight the lack of suitable material properties of human brain tissue for modelling TBI and calls for more research into mechanical characterisation of human brain tissue under large strain at high strain rates.
{"title":"Material properties of human brain tissue suitable for modelling traumatic brain injury","authors":"David B. MacManus , Mazdak Ghajari","doi":"10.1016/j.brain.2022.100059","DOIUrl":"10.1016/j.brain.2022.100059","url":null,"abstract":"<div><p>Finite element (FE) brain models have revolutionised research on the biomechanics of traumatic brain injury (TBI). The accuracy and reliability of results arising from brain models depend equally on their geometric accuracy as the quality of the material properties used to describe the mechanical behaviour of brain. However, much of the literature on human brain tissues’ material properties have been performed at low strain rates and strains. This is particularly striking considering a large portion of the brain tissue mechanical characterisation literature is presented with a motivation of understanding brain tissues’ behaviour during TBI which occurs due to brain tissues’ exposure to large strains at high strain rates. Therefore, the aim of this review is to collate the mechanical characterisation studies on human brain tissue under conditions suitable for modelling TBI. We first review injury threshold studies and show that ≥20% strain at ≥10/s strain rate is a reasonable minimum threshold for producing injury to the brain. Using this threshold, we show that there are only five studies on the mechanical characterisation of human brain tissue under strains at strain rates relevant to TBI. These studies, provide material properties of human brain tissue at moderate and high rate loading, with only a recent study showing its region dependent characteristics. This review acts as a reference for scientists and engineers to select suitable material data when modelling human TBI. It also calls for more research to provide high fidelity material properties for modelling of TBI.</p></div><div><h3>Statement of significance</h3><p>The significance of this work is underscored by the reporting of brain tissues’ material properties in the context of traumatic brain injury (TBI) despite these properties having been characterised under strains and strain rates that are not relevant to TBI. This can result in inaccurate results if implemented in finite element brain models. Here, we address this problem by performing a review on the mechanical characterisation of human brain tissue under conditions that are suitable for modelling human TBI. Our findings show that there are only five studies on the mechanical characterisation of human brain tissue under strains at strain rate levels relevant to TBI. These results will allow researchers to select appropriate material properties for modelling human TBI providing more realistic behaviour of brain tissue in simulations. These results also provide minimum strain and strain rate values for mechanical characterisation experiments on brain tissue for TBI applications. Furthermore, our findings highlight the lack of suitable material properties of human brain tissue for modelling TBI and calls for more research into mechanical characterisation of human brain tissue under large strain at high strain rates.</p></div>","PeriodicalId":72449,"journal":{"name":"Brain multiphysics","volume":"3 ","pages":"Article 100059"},"PeriodicalIF":0.0,"publicationDate":"2022-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S2666522022000168/pdfft?md5=7fc489852e1e72c5bf0742d7e90b1429&pid=1-s2.0-S2666522022000168-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"43270579","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-01-01DOI: 10.1016/j.brain.2022.100056
Bo Xue , Xuejun Wen , Ram Kuwar , Dong Sun , Ning Zhang
Recent efforts in biomaterial-assisted brain tissue engineering suggest that match of mechanical properties of biomaterials to those of native brain tissue may be crucial for brain regeneration. In particular, the mechanical properties of native brain tissue vary as a function of age. To date, detailed characterization of age-dependent viscoelastic properties of brain tissue throughout the postnatal development to adulthood is only available at sparse age points in animal studies. To fill this gap, we have characterized the linear viscoelastic properties of the cerebral cortex in rats at well-spaced ages from postnatal day 4 to 4 months old, the age range that is widely used in neural regeneration studies. Using an oscillatory rheometer, the viscoelastic properties of rat cortical slices were measured independently by storage moduli (G′) and loss moduli (G″). The data demonstrated increases in both the storage moduli and the loss moduli of cortex tissue over post-natal age in rats. At all ages, the damping factor (G″/G′ ratio) remained constant at low oscillatory strain frequencies (<10 rad/s) before it started to decline at medium frequency range (10-100 rad/s). Such changes were not age-dependent. The stress-relaxation response increased over post-natal age, consistent with the increasing tissue stiffness. Taken together, our study demonstrates that age is a crucial factor determining the mechanical properties of the cerebral cortex in rats during early postnatal development. This data may provide the guidelines for age-specific biomechanics study of brain tissue and help to define the mechanical properties of biomaterials for biomaterial-assisted brain tissue regeneration studies.
Statement of significance
Studies about age-dependent viscoelastic properties of rat brain tissue throughout the postnatal development to adulthood is sparsely available. To fill up the gap of knowledge, in this study, we have characterized the age-dependent viscoelastic properties and the linear viscoelastic properties of the cerebral cortex throughout the postnatal development stage to adulthood in rats by measuring storage moduli (G′), loss moduli (G″), damping factor (G″/G′ ratio) and stress-relaxation response. We have found that age is a crucial factor determining the mechanical properties of the cerebral cortex in rats during early postnatal development. The findings of this study could provide guidelines for age-specific biomechanical study of brain tissue and help to define the mechanical properties of biomaterials for biomaterial-assisted brain tissue regeneration in experimental models in rats.
{"title":"Age-dependent viscoelastic characterization of rat brain cortex","authors":"Bo Xue , Xuejun Wen , Ram Kuwar , Dong Sun , Ning Zhang","doi":"10.1016/j.brain.2022.100056","DOIUrl":"10.1016/j.brain.2022.100056","url":null,"abstract":"<div><p>Recent efforts in biomaterial-assisted brain tissue engineering suggest that match of mechanical properties of biomaterials to those of native brain tissue may be crucial for brain regeneration. In particular, the mechanical properties of native brain tissue vary as a function of age. To date, detailed characterization of age-dependent viscoelastic properties of brain tissue throughout the postnatal development to adulthood is only available at sparse age points in animal studies. To fill this gap, we have characterized the linear viscoelastic properties of the cerebral cortex in rats at well-spaced ages from postnatal day 4 to 4 months old, the age range that is widely used in neural regeneration studies. Using an oscillatory rheometer, the viscoelastic properties of rat cortical slices were measured independently by storage moduli (G′) and loss moduli (G″). The data demonstrated increases in both the storage moduli and the loss moduli of cortex tissue over post-natal age in rats. At all ages, the damping factor (G″/G′ ratio) remained constant at low oscillatory strain frequencies (<10 rad/s) before it started to decline at medium frequency range (10-100 rad/s). Such changes were not age-dependent. The stress-relaxation response increased over post-natal age, consistent with the increasing tissue stiffness. Taken together, our study demonstrates that age is a crucial factor determining the mechanical properties of the cerebral cortex in rats during early postnatal development. This data may provide the guidelines for age-specific biomechanics study of brain tissue and help to define the mechanical properties of biomaterials for biomaterial-assisted brain tissue regeneration studies.</p></div><div><h3>Statement of significance</h3><p>Studies about age-dependent viscoelastic properties of rat brain tissue throughout the postnatal development to adulthood is sparsely available. To fill up the gap of knowledge, in this study, we have characterized the age-dependent viscoelastic properties and the linear viscoelastic properties of the cerebral cortex throughout the postnatal development stage to adulthood in rats by measuring storage moduli (G′), loss moduli (G″), damping factor (G″/G′ ratio) and stress-relaxation response. We have found that age is a crucial factor determining the mechanical properties of the cerebral cortex in rats during early postnatal development. The findings of this study could provide guidelines for age-specific biomechanical study of brain tissue and help to define the mechanical properties of biomaterials for biomaterial-assisted brain tissue regeneration in experimental models in rats.</p></div>","PeriodicalId":72449,"journal":{"name":"Brain multiphysics","volume":"3 ","pages":"Article 100056"},"PeriodicalIF":0.0,"publicationDate":"2022-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ftp.ncbi.nlm.nih.gov/pub/pmc/oa_pdf/a3/7b/nihms-1855863.PMC9757762.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"10750319","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-01-01DOI: 10.1016/j.brain.2022.100046
Songbai Ji , Shaoju Wu , Wei Zhao
Impact-induced brain strains are spatially rich and intrinsically dynamic. However, the dynamic information of brain strain is not typically used in any injury investigation. Here, we study the dynamic characteristics of maximum and minimum principal strain (maxPS and minPS) of the corpus callosum and highlight the significance of impact simulation time window. Three datasets are used: laboratory reconstructed National Football League (NFL; N=53), measured impacts from Stanford (SF; N=110) and Prevent Biometric (PB; N=314). Impact cases are discarded (by 20.8%, 11.8%, and 66.2%, respectively), when the simulation time window is considered inadequate to capture sufficient strain temporal responses. Fitted Gaussian peaks (with average relative root mean squared error of ∼5% and R2 >0.9) from all datasets have a similar median (15–18 ms) and inter-quantile range (5–9 ms) for the full width at half maximum (FWHM). FWHM significantly and negatively correlates with strain magnitude for NFL and SF, but not for PB. However, ratios between the largest minPS and maxPS magnitudes are similar across datasets (median of 0.5–0.6 with inter-quantile range of 0.2–0.7). Dynamic strain features improve injury prediction. This study motivates further development of advanced deep learning models to instantly estimate the complete details of spatiotemporal history of brain strains, beyond spatially detailed peak strains obtained at maximum values currently available. In addition, this study highlights the time lag between impact kinematics and corpus callosum strain deep in the brain, which has important implications for impact simulation and result interpretation as well as impact sensor designs in the future.
Statement of significance
•
First study to systematically characterize the temporal history of corpus callosum strain in contact sports head impact.
•
Allows to rapidly launch multiscale modeling of concussion in the corpus callosum without a costly whole brain model simulation.
•
Motivates further development of advanced deep learning models that will instantly reproduce the complete spatiotemporal details of strain in the entire brain.
•
Highlights the importance of sufficient impact simulation time window in order to capture the complete strain responses deep in the brain.
{"title":"Dynamic characteristics of impact-induced brain strain in the corpus callosum","authors":"Songbai Ji , Shaoju Wu , Wei Zhao","doi":"10.1016/j.brain.2022.100046","DOIUrl":"10.1016/j.brain.2022.100046","url":null,"abstract":"<div><p>Impact-induced brain strains are spatially rich and intrinsically dynamic. However, the dynamic information of brain strain is not typically used in any injury investigation. Here, we study the dynamic characteristics of maximum and minimum principal strain (maxPS and minPS) of the corpus callosum and highlight the significance of impact simulation time window. Three datasets are used: laboratory reconstructed National Football League (NFL; N=53), measured impacts from Stanford (SF; N=110) and Prevent Biometric (PB; N=314). Impact cases are discarded (by 20.8%, 11.8%, and 66.2%, respectively), when the simulation time window is considered inadequate to capture sufficient strain temporal responses. Fitted Gaussian peaks (with average relative root mean squared error of ∼5% and R<sup>2</sup> >0.9) from all datasets have a similar median (15–18 ms) and inter-quantile range (5–9 ms) for the full width at half maximum (FWHM). FWHM significantly and negatively correlates with strain magnitude for NFL and SF, but not for PB. However, ratios between the largest minPS and maxPS magnitudes are similar across datasets (median of 0.5–0.6 with inter-quantile range of 0.2–0.7). Dynamic strain features improve injury prediction. This study motivates further development of advanced deep learning models to instantly estimate the complete details of spatiotemporal history of brain strains, beyond spatially detailed peak strains obtained at maximum values currently available. In addition, this study highlights the time lag between impact kinematics and corpus callosum strain deep in the brain, which has important implications for impact simulation and result interpretation as well as impact sensor designs in the future.</p></div><div><h3>Statement of significance</h3><p></p><ul><li><span>•</span><span><p>First study to systematically characterize the temporal history of corpus callosum strain in contact sports head impact.</p></span></li><li><span>•</span><span><p>Allows to rapidly launch multiscale modeling of concussion in the corpus callosum without a costly whole brain model simulation.</p></span></li><li><span>•</span><span><p>Motivates further development of advanced deep learning models that will instantly reproduce the complete spatiotemporal details of strain in the entire brain.</p></span></li><li><span>•</span><span><p>Highlights the importance of sufficient impact simulation time window in order to capture the complete strain responses deep in the brain.</p></span></li></ul></div>","PeriodicalId":72449,"journal":{"name":"Brain multiphysics","volume":"3 ","pages":"Article 100046"},"PeriodicalIF":0.0,"publicationDate":"2022-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S266652202200003X/pdfft?md5=3bf2a50ac67bc1175a2b8da33406f64e&pid=1-s2.0-S266652202200003X-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"54405913","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号