Subject-specific Models Research Articles

Multi-Scale Masked Autoencoders for Cross-Session Emotion Recognition.

Affective brain-computer interfaces (aBCIs) have garnered widespread applications, with remarkable advancements in utilizing electroencephalogram (EEG) technology for emotion recognition. However, the time-consuming process of annotating EEG data, inherent individual differences, non-stationary characteristics of EEG data, and noise artifacts in EEG data collection pose formidable challenges in developing subject-specific cross-session emotion recognition models. To simultaneously address these challenges, we propose a unified pre-training framework based on multi-scale masked autoencoders (MSMAE), which utilizes large-scale unlabeled EEG signals from multiple subjects and sessions to extract noise-robust, subject-invariant, and temporal-invariant features. We subsequently fine-tune the obtained generalized features with only a small amount of labeled data from a specific subject for personalization and enable cross-session emotion recognition. Our framework emphasizes: 1) Multi-scale representation to capture diverse aspects of EEG signals, obtaining comprehensive information; 2) An improved masking mechanism for robust channel-level representation learning, addressing missing channel issues while preserving inter-channel relationships; and 3) Invariance learning for regional correlations in spatial-level representation, minimizing inter-subject and inter-session variances. Under these elaborate designs, the proposed MSMAE exhibits a remarkable ability to decode emotional states from a different session of EEG data during the testing phase. Extensive experiments conducted on the two publicly available datasets, i.e., SEED and SEED-IV, demonstrate that the proposed MSMAE consistently achieves stable results and outperforms competitive baseline methods in cross-session emotion recognition.

Renewing Iterative Self-Labeling Domain Adaptation With Application to the Spine Motion Prediction

The area of transfer learning comprises supervised machine learning methods that cope with the issue when the training and testing data have different input feature spaces or distributions. In this work, we propose a novel transfer learning algorithm called Renewing Iterative Self-labeling Domain Adaptation (Re-ISDA). This work is motivated by a cervical spine motion prediction problem, with the goal of predicting the cervical spine motion of different subjects using the measurements of “exterior features”. The joint distribution of the exterior features for each subject may vary with one’s distinct BMI, sex and other characteristics; the sample size of this problem is limited due to the high experimental cost. The proposed method is well suited for transfer learning problems with limited training samples. The learning problem is formulated as a dynamic programming model and the latter is then solved by an efficient greedy algorithm. Numerical studies show that the proposed method outperforms prevailing transfer learning methods. The proposed method also achieves high prediction accuracy for the cervical spine motion problem. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —This paper is motivated by the problem of using human body exterior features (e.g., contours, surface markers) acquired by computer vision or other measurement means to determine the underlying physics or mechanics. The goal is to build a subject-specific model to predict cervical spine motion based on surface-based measurements. Conventional supervised learning and transfer learning methods fail to give an accurate prediction due to the subject-to-subject variation and/or the small sample size of the training data. This paper proposes a novel transfer learning method called renewing iterative self-labeling domain adaptation to modify the possible error of the pseudo labels and thus make the training process more stable. The proposed method improves the accuracy by over <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$35\%$</tex-math> </inline-formula> on the traditional transfer learning methods. This method is applicable to limited-sample problems with different training and testing input distributions, and is particularly suitable for time-series data.

TFTL: A Task-Free Transfer Learning Strategy for EEG-based Cross-Subject & Cross-Dataset Motor Imagery BCI.

Motor imagery-based brain-computer interfaces (MI-BCIs) have been playing an increasingly vital role in neural rehabilitation. However, the long-term task-based calibration required for enhanced model performance leads to an unfriendly user experience, while the inadequacy of EEG data hinders the performance of deep learning models. To address these challenges, a task-free transfer learning strategy (TFTL) for EEG-based cross-subject & cross-dataset MI-BCI is proposed for calibration time reduction and multi-center data co-modeling. TFTL strategy consists of data alignment, shared feature extractor, and specific classifiers, in which the label predictor for MI tasks classification, as well as domain and dataset discriminator for inter-subject variability reduction are concurrently optimized for knowledge transfer from subjects across different datasets to the target subject. Moreover, only resting data of the target subject is used for subject-specific model construction to achieve task-free. We employed three deep learning methods (ShallowConvNet, EEGNet, and TCNet-Fusion) as baseline approaches to evaluate the effectiveness of the proposed strategy on five datasets (BCIC IV Dataset 2a, Dataset 1, Physionet MI, Dreyer 2023, and OpenBMI). The results demonstrate a significant improvement with the inclusion of the TFTL strategy compared to the baseline methods, reaching a maximum enhancement of 15.67% with a statistical significance (p=2.4e-5<0.05). Moreover, task-free resulted in MI trials needed for calibration being 0 for all datasets, which significantly alleviated the calibration burden for patients before usage. The proposed TFTL strategy effectively addresses challenges posed by prolonged calibration periods and insufficient EEG data, thus promoting MI-BCI from laboratory to clinical application.

Instantaneous Generation of Subject-Specific Finite Element Models of the Hip Capsule.

Subject-specific hip capsule models could offer insights into impingement and dislocation risk when coupled with computer-aided surgery, but model calibration is time-consuming using traditional techniques. This study developed a framework for instantaneously generating subject-specific finite element (FE) capsule representations from regression models trained with a probabilistic approach. A validated FE model of the implanted hip capsule was evaluated probabilistically to generate a training dataset relating capsule geometry and material properties to hip laxity. Multivariate regression models were trained using 90% of trials to predict capsule properties based on hip laxity and attachment site information. The regression models were validated using the remaining 10% of the training set by comparing differences in hip laxity between the original trials and the regression-derived capsules. Root mean square errors (RMSEs) in laxity predictions ranged from 1.8° to 2.3°, depending on the type of laxity used in the training set. The RMSE, when predicting the laxity measured from five cadaveric specimens with total hip arthroplasty, was 4.5°. Model generation time was reduced from days to milliseconds. The results demonstrated the potential of regression-based training to instantaneously generate subject-specific FE models and have implications for integrating subject-specific capsule models into surgical planning software.

Use of subject-specific models to detect fatigue-related changes in running biomechanics: a random forest approach.

Running biomechanics are affected by fatiguing or prolonged runs. However, no evidence to date has conclusively linked this effect to running-related injury (RRI) development or performance implications. Previous investigations using subject-specific models in running have demonstrated higher accuracy than group-based models, however, this has been infrequently applied to fatigue. In this study, two experiments were conducted to determine whether subject-specific models outperformed group-based models to classify running biomechanics during non-fatigued and fatigued conditions. In the first experiment, 16 participants performed four treadmill runs at or around the maximal lactate steady state. In the second experiment, nine participants performed five prolonged runs using commercial wearable devices. For each experiment, two segments were extracted from each trial from early and late in the run. For each participant, a random forest model was applied with a leave-one-run-out cross-validation to classify between the early (non-fatigued) and late (fatigued) segments. Additionally, group-based classifiers with a leave-one-subject-out cross validation were constructed. For experiment 1, mean classification accuracies for the single-subject and group-based classifiers were 68.2 ± 8.2% and 57.0 ± 8.9%, respectively. For experiment 2, mean classification accuracies for the single-subject and group-based classifiers were 68.9 ± 17.1% and 61.5 ± 11.7%, respectively. Variable importance rankings were consistent within participants, but these rankings differed from each participant to those of the group. Although the classification accuracies were relatively low, these findings highlight the advantage of subject-specific classifiers to detect changes in running biomechanics with fatigue and indicate the potential of using big data and wearable technology approaches in future research to determine possible connections between biomechanics and RRI.

Comparative evaluation of different spinal stability metrics

Direct in vivo measurements of spinal stability are not possible, leaving computational estimations (such as dynamic time series and structural analyses) as the feasible option. However, differences between different stability assessment approaches and metrics remain unclear. To explore this, we asked 32 participants to perform 35 cycles of repetitive lifts with and without load (4/2.6 kg for males/females). EMG signals and 3D kinematics were collected via 12 surface electrodes and 17 inertial sensors, and three dynamical stability measures were computed: short and long temporal and conventional maximum Lyapunov exponents (LyE) and maximum Floquet multipliers (FM). A dynamic subject-specific EMG-assisted musculoskeletal model computed four structural stability measures (critical muscle stiffness coefficient at which spine becomes unstable, average spine stiffness, minimum and geometric average of Hessian matrix eigenvalues). Across cycles, dynamical and structural stability outcomes varied noticeably. Temporal short-term LyE and all structural stability measures were more influenced by the cycle percentage (posture factor) than by phase (lifting, lowering) or load factor. The effect of all factors were non-significant for FM and long LyE, except for the posture on LyE-L with a small effect size. Pearson's correlations revealed a weak to moderate, or non-existent, correlation between structural and dynamical stability metrics, with small shared variances, underscoring their distinct and independent nature and theoretical foundations. Moreover, the low sensitivity of dynamic measures to posture and load factors, found in this study, calls for further examination. Considering the limitations and shortcomings of both dynamical and structural stability assessment approaches, there is a need for the development of improved musculoskeletal stability evaluation techniques.

Analysis of the pattern of microstructural changes in the brain after mTBI with diffusion tensor imaging and subject-specific FE models

Traumatic brain injury (TBI) is a major public health challenge. Up to 90 % of TBIs are on the mild spectrum of TBI (mTBI), where diagnosis is a major challenge. Majority of studies in this field have been conducted on human subjects, which inherently suffer from the lack of appropriate control group, selection bias, and individual differences in patients. To overcome these limitations, animal studies have been used as an alternative approach to provide deeper insights into the underlying mechanism related to the injury. Therefore our aim is to investigate various quantitative imaging biomarkers acquired from T1-W and diffusion tensor imaging (DTI) sequences to provide more information about the microstructural changes in the brain after mTBI. We then use this to generate subject-specific finite element models of the brain and examine how the changes in the brain material properties reflected in MR images affects strain distribution patterns on a subsequent head hit. Our study revealed a decrease in FA and an increase in diffusivity indices (MD, AD, RD) in the white matter tracts of the brain. This finding may represent the axonal damage, demyelination and gliosis after mild TBI, which have been shown in other animal and human studies. Moreover, our FE analysis showed that microstructural changes in the brain after mTBI might have weakened the structural integrity of the brain as the subsequent head hit led to wider and more severe brain deformations. SignificanceAnimal models have been used to investigate biomechanical and pathophysiological aspects of mild traumatic brain injuries in the past. Still, most of them used small animals such as rats and mice. These models provided valuable insight into the pathophysiology of mTBI, but their findings have limitations due to their inherent differences to human brains. We have developed a large animal model of mTBI with sheep brains by combining advanced MRI and finite element analysis as they mimic the human brain better. To the best of our knowledge, this study is the first mTBI neuroimaging study conducted on large animal brains to investigate the diffusional changes in the white matter tracts after mTBI. Our FE analysis revealed that such microstructural changes resulted in tissue softening as the extent of brain deformation increased on a subsequent head hit, indicating increased brain vulnerability after head impacts.

Sensitivity Analysis of Upper Limb Musculoskeletal Models During Isometric and Isokinetic Tasks.

Sensitivity coefficients are used to understand how errors in subject-specific musculoskeletal model parameters influence model predictions. Previous sensitivity studies in the lower limb calculated sensitivity using perturbations that do not fully represent the diversity of the population. Hence, the present study performs sensitivity analysis in the upper limb using a large synthetic dataset to capture greater physiological diversity. The large dataset (n = 401 synthetic subjects) was created by adjusting maximum isometric force, optimal fiber length, pennation angle, and bone mass to induce atrophy, hypertrophy, osteoporosis, and osteopetrosis in two upper limb musculoskeletal models. Simulations of three isometric and two isokinetic upper limb tasks were performed using each synthetic subject to predict muscle activations. Sensitivity coefficients were calculated using three different methods (two point, linear regression, and sensitivity functions) to understand how changes in Hill-type parameters influenced predicted muscle activations. The sensitivity coefficient methods were then compared by evaluating how well the coefficients accounted for measurement uncertainty. This was done by using the sensitivity coefficients to predict the range of muscle activations given known errors in measuring musculoskeletal parameters from medical imaging. Sensitivity functions were found to best account for measurement uncertainty. Simulated muscle activations were most sensitive to optimal fiber length and maximum isometric force during upper limb tasks. Importantly, the level of sensitivity was muscle and task dependent. These findings provide a foundation for how large synthetic datasets can be applied to capture physiologically diverse populations and understand how model parameters influence predictions.

The long head of the biceps tendon undergoes multiaxial deformation during shoulder motion

The long head biceps tendon (LHBT) is presumed a common source of shoulder joint pain and injury. Despite common LHBT pathologies, diagnosis and preferred treatment remain frequently debated. This Short Communication reports the development of a subject-specific finite element model of the shoulder joint based on one subject’s 3D reconstructed anatomy and 3D in vivo kinematics recorded from bone-fixed electromagnetic sensors. The primary purpose of this study was to use the developed finite element model to investigate the LHBT mechanical environment during a typical shoulder motion of arm raising. Furthermore, this study aimed to assess the viability of material models derived from uniaxial tensile tests for accurate simulation of in vivo motion. The findings of our simulations indicate that the LHBT undergoes complex multidimensional deformations. As such, uniaxial material properties reported in the existing body of literature are not sufficient to simulate accurately the in vivo mechanical behavior of the LHBT. Further experimental tests on cadaveric specimens, such as biaxial tension and combinations of tension and torsion, are needed to describe fully the mechanical behavior of the LHBT and investigate its mechanisms of injury.

An artificial neural network for full-body posture prediction in dynamic lifting activities and effects of its prediction errors on model-estimated spinal loads

Musculoskeletal models have indispensable applications in occupational risk assessment/management and clinical treatment/rehabilitation programs. To estimate muscle forces and joint loads, these models require body posture during the activity under consideration. Posture is usually measured via video-camera motion tracking approaches that are time-consuming, costly, and/or limited to laboratories. Alternatively, posture-prediction tools based on artificial intelligence can be trained using measured postures of several subjects performing many activities. We aimed to use our previous posture-prediction artificial neural network (ANN), developed based on many measured static postures, to predict posture during dynamic lifting activities. Moreover, effects of the ANN posture-prediction errors on dynamic spinal loads were investigated using subject-specific musculoskeletal models. Seven individuals each performed twenty-five lifting tasks while their full-body three-dimensional posture was measured by a 10-camera Vicon system and also predicted by the ANN as functions of the hand-load positions during the lifting activities. The measured and predicted postures (i.e., coordinates of 39 skin markers) and their model-estimated L5-S1 loads were compared. The overall root-mean-squared-error (RMSE) and normalized (by the range of measured values) RMSE (nRMSE) between the predicted and measured postures for all markers/tasks/subjects was equal to 7.4 cm and 4.1 %, respectively (R2 = 0.98 and p < 0.05). The model-estimated L5-S1 loads based on the predicted and measured postures were generally in close agreements as also confirmed by the Bland-Altman analyses; the nRMSE for all subjects/tasks was < 10 % (R2 > 0.7 and p > 0.05). In conclusion, the easy-to-use ANN can accurately predict posture in dynamic lifting activities and its predicted posture can drive musculoskeletal models.

Quantification of soft tissue artifacts using CT registration and subject-specific multibody modeling

The potential use of gait analysis for quantitative preoperative planning in total hip arthroplasty (THA) has previously been demonstrated. However, the joint kinematic data measured through this process tend to be unreliable for surgical planning due to distortions caused by soft tissue artifacts (STAs). In this study, we developed a novel motion capture framework by combining computed tomography (CT)-based postural calibration and subject-specific multibody dynamics modeling to prevent the effect of STAs in measuring hip kinematics. Three subjects with femoroacetabular impingement syndrome were recruited, and CT data for each patient were collected by attaching marker clusters near the hip. A subject-specific multibody hip joint model was developed based on reconstructed CT data. Spring–dashpot network calculations were performed to minimize the distance between the anatomical landmark and its corresponding infrared reflective marker. The STAs of the thigh was described as six degrees of freedom viscoelastic bushing elements, and their parameter values were identified via smooth orthogonal decomposition. Least squares optimization was used to modify the pelvic rotations to compensate for the rigid components of STAs. The results showed that CT-assisted motion tracking enabled the successful identification of STA influences in gait and squat positions. Furthermore, STA effects were found to alter maximal pelvis tilt and hip rotations during a squat. Compared to other techniques, such as dual fluoroscopic imaging, the adopted framework does not require additional medical imaging for patients undergoing robot-assisted THA surgery and is thus a practical way of evaluating hip joint kinematics for preoperative surgical planning.

Greater tuberosity morphology is altered in individuals with symptomatic isolated supraspinatus tendon tears.

In the presence of an isolated supraspinatus tear, the force applied to the greater tuberosity is reduced, which may lead to bony morphologic changes. Thus, diagnostic or surgical identification of landmarks to properly repair the torn tendon might be difficult if the anatomy of the greater tuberosity is altered. The objectives of the study were to assess the presence of the superior, middle, and inferior facets of the greater tuberosity in individuals with symptomatic isolated supraspinatus tendon tears and the associations between tear size, tear location and presence of each facet. Thirty-seven individuals with symptomatic isolated supraspinatus tendon tears were recruited to participate in the study. Individuals underwent a high-resolution computed tomography scan of the involved shoulder and images were segmented to generate subject specific models of each humerus. The vertices comprising each facet were identified; however, if even a single vertex comprising the facet was missing, the facet was considered altered. Percentage agreement for correctly identifying the presence of each facet was determined using 2 additional observers and 5 randomly selected humeri. Ultrasonography was performed to assess anterior-posterior (AP) tear size and tear location. Outcome parameters included presence of the superior, middle, and inferior facets; AP tear size; and tear location. Point-biserial correlations were used to determine the associations between AP tear size, tear location, and presence of the superior, middle, and inferior facets. Supraspinatus tear size and tear location was 13.1±6.1 mm (range, 1.9-28.3 mm) and 2.0±4.4 mm from the posterior edge of the long head of the biceps tendon (range, 0.0-19.0 mm), respectively. Overall, the superior, middle, and inferior facets were not altered in 24.3%, 29.7%, and 45.9% of individuals, respectively. Percentage agreement between observers was 83.4% on average. No associations were found between tear size or tear location and presence of the superior, middle, or inferior facet (P values ranged from .19 to .74). Individuals with symptomatic isolated supraspinatus tears experience significant alterations in the bony morphology of the greater tuberosity that were irrespective of supraspinatus tear size and location. This information is useful for radiologists and orthopedic surgeons as the altered anatomy may influence the ability to identify important anatomic landmarks during diagnostic imaging or surgical procedures.

Quantifying changes in shoulder orientation between the prone and supine positions from magnetic resonance imaging

BackgroundPredicting breast tissue motion using biomechanical models can provide navigational guidance during breast cancer treatment procedures. These models typically do not account for changes in posture between procedures. Difference in shoulder position can alter the shape of the pectoral muscles and breast. A greater understanding of the differences in the shoulder orientation between prone and supine could improve the accuracy of breast biomechanical models. Methods19 landmarks were placed on the sternum, clavicle, scapula, and humerus of the shoulder girdle in prone and supine breast MRIs (N = 10). These landmarks were used in an optimization framework to fit subject-specific skeletal models and compare joint angles of the shoulder girdle between these positions. FindingsThe mean Euclidean distance between joint locations from the fitted skeletal model and the manually identified joint locations was 15.7 mm ± 2.7 mm. Significant differences were observed between prone and supine. Compared to supine position, the shoulder girdle in the prone position had the lateral end of the clavicle in more anterior translation (i.e., scapula more protracted) (P < 0.05), the scapula in more protraction (P < 0.01), the scapula in more upward rotation (associated with humerus elevation) (P < 0.05); and the humerus more elevated (P < 0.05) for both the left and right sides. InterpretationShoulder girdle orientation was found to be different between prone and supine. These differences would affect the shape of multiple pectoral muscles, which would affect breast shape and the accuracy of biomechanical models.

Subject-specific one-dimensional fluid dynamics model of chronic thromboembolic pulmonary hypertension

Chronic thromboembolic pulmonary hypertension (CTEPH) develops due to the accumulation of blood clots in the lung vasculature that obstructs flow and increases pressure. The mechanobiological factors that drive progression of CTEPH are not understood, in part because mechanical and hemodynamic changes in the small pulmonary arteries due to CTEPH are not easily measurable. Using previously published hemodynamic measurements and imaging from a large animal model of CTEPH, we applied a subject-specific one-dimensional (1D) computational fluid dynamic (CFD) approach to investigate the impact of CTEPH on pulmonary artery stiffening, time-averaged wall shear stress (TAWSS), and oscillatory shear index (OSI) in extralobar (main, right, and left) pulmonary arteries and intralobar (distal to the extralobar) arteries. Our results demonstrate that CTEPH increases pulmonary artery wall stiffness and decreases TAWSS in extralobar and intralobar arteries. Moreover, CTEPH increases the percentage of the intralobar arterial network with both low TAWSS and high OSI, quantified by the novel parameter φ\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\varphi$$\\end{document}, which is related to thrombogenicity. Our analysis reveals a strong positive correlation between increases in mean pulmonary artery pressure (mPAP) and φ\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\varphi$$\\end{document} from baseline to CTEPH in individual subjects, which supports the suggestion that increased φ\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\varphi$$\\end{document} drives disease severity. This subject-specific experimental–computational framework shows potential as a predictor of the impact of CTEPH on pulmonary arterial hemodynamics and pulmonary vascular mechanics. By leveraging advanced modeling techniques and calibrated model parameters, we predict spatial distributions of flow and pressure, from which we can compute potential physiomarkers of disease progression. Ultimately, this approach can lead to more spatially targeted interventions that address the needs of individual CTEPH patients.

BORS/BJR TRAVELLING FELLOWSHIP: IMAGE-DRIVEN SUBJECT-SPECIFIC SPINE MODELS: DEVELOPING A NOVEL TOOL TO MEASURE IN VIVO LOADING

AbstractObjectivesSpinal disorders such as back pain incur a substantial societal and economic burden. Unfortunately, there is lack of understanding and treatment of these disorders are further impeded by the inability to assess spinal forces in vivo. The aim of this project is to address this challenge by developing and testing a novel image-driven approach that will assess the forces in an individual's spine in vivo by incorporating information acquired from multimodal imaging (magnetic resonance imaging (MRI) and biplane X-rays) in a subject-specific model.MethodsMagnetic resonance and biplane X-ray imaging are used to capture information about the anatomy, tissues, and motion of an individual's spine as they perform a range of everyday activities. This information is then utilised in a subject-specific computational model based on the finite element method to predict the forces in their spine. The project is also utilising novel machine learning algorithms and in vitro, six-axis mechanical testing on human, porcine and bovine samples to develop and test the modelling methods rigorously.Results &amp; DiscussionMRI sequences have been identified that provide high-quality image data and information on different tissue types which will be used to predict subject-specific disc properties. In-vivo protocols to capture motion analysis, EMG muscle activity, and video X-rays of the spine have been designed with planned data collection of 15 healthy volunteers. Preliminary modelling work has evaluated potential machine learning approaches and quantified the sensitivity of the models developed to material properties.ConclusionThe development and testing of these image-driven subject-specific spine models will provide a new tool for determining forces in the spine. It will also provide new tools for measuring and modelling spine movement and quantifying the properties of the spinal tissues.AcknowledgmentsFunding from the EPSRC: EP/V036602/1 (Meakin, Holsgrove &amp; Javadi) and EP/V032275/1 (Holt &amp; Williams).Declaration of Interest(b) declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported:I declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research project.

Impact of tricuspid annuloplasty device shape and size on valve mechanics—a computational study

BackgroundTricuspid valve disease significantly affects 1.6 million Americans. The gold standard treatment for tricuspid disease is the implantation of annuloplasty devices. These ring-like devices come in various shapes and sizes. Choices for both shape and size are most often made by surgical intuition rather than scientific rationale. MethodsTo understand the impact of shape and size on valve mechanics and to provide a rational basis for their selection, we used a subject-specific finite element model to conduct a virtual case study. That is, we implanted 4 different annuloplasty devices of 6 different sizes in our virtual patient. After each virtual surgery, we computed the coaptation area, leaflet end-systolic angles, leaflet stress, and chordal forces. ResultsWe found that contoured devices are better at normalizing end-systolic angles, whereas the one flat device, the Edwards Classic, maximized the coaptation area and minimized leaflet stress and chordal forces. We further found that reducing device size led to increased coaptation area but also negatively impacted end-systolic angles, stress, and chordal forces. ConclusionsBased on our analyses of the coaptation area, leaflet motion, leaflet stress, and chordal forces, we found that device shape and size have a significant impact on valve mechanics. Thereby, our study also demonstrates the value of simulation tools and device tests in “virtual patients.” Expanding our study to many more valves may, in the future, allow for universal recommendations.

Model-Based Assessment of Elastic Material Parameters in Rheumatic Heart Disease Patients and Healthy Subjects

Non-invasive measurements are important for the development of new treatments for heart failure, which is one of the leading causes of death worldwide. This study aimed to develop realistic subject-specific computational models of human biventricles using clinical data. Three-dimensional finite element models of the human ventricles were created using cardiovascular magnetic resonance images of rheumatic heart disease (RHD) patients and healthy subjects. The material parameter optimization uses inverse modeling based on the finite element method combined with the Levenberg–Marquardt method (LVM) by targeting subject-specific hemodynamics. The study of elastic myocardial parameters between healthy subjects and RHD patients showed an elevated stiffness in diseased hearts. In particular, the anisotropic material behavior of the healthy and diseased cardiac tissue significantly differed from one another. Furthermore, as the LVEF decreased, the stiffness and its orientation-dependent parameters increased. The simulation-derived LV myocardial circumferential and longitudinal stresses were negatively associated with the LVEF. The sensitivity analysis result demonstrated that the observed significant difference between the elastic material parameters of diseased and healthy myocardium was not exclusively attributable to an increased LVEDP in the diseased heart. These results could be applied to future computational studies for developing heart failure treatment.

Validation and evaluation of subject-specific finite element models of the pediatric knee

Finite element (FE) models have been widely used to investigate knee joint biomechanics. Most of these models have been developed to study adult knees, neglecting pediatric populations. In this study, an atlas-based approach was employed to develop subject-specific FE models of the knee for eight typically developing pediatric individuals. Initially, validation simulations were performed at four passive tibiofemoral joint (TFJ) flexion angles, and the resulting TFJ and patellofemoral joint (PFJ) kinematics were compared to corresponding patient-matched measurements derived from magnetic resonance imaging (MRI). A neuromusculoskeletal-(NMSK)-FE pipeline was then used to simulate knee biomechanics during stance phase of walking gait for each participant to evaluate model simulation of a common motor task. Validation simulations demonstrated minimal error and strong correlations between FE-predicted and MRI-measured TFJ and PFJ kinematics (ensemble average of root mean square errors < 5 mm for translations and < 4.1° for rotations). The FE-predicted kinematics were strongly correlated with published reports (ensemble average of Pearson's correlation coefficients (ρ) > 0.9 for translations and ρ > 0.8 for rotations), except for TFJ mediolateral translation and abduction/adduction rotation. For walking gait, NMSK-FE model-predicted knee kinematics, contact areas, and contact pressures were consistent with experimental reports from literature. The strong agreement between model predictions and experimental reports underscores the capability of sequentially linked NMSK-FE models to accurately predict pediatric knee kinematics, as well as complex contact pressure distributions across the TFJ articulations. These models hold promise as effective tools for parametric analyses, population-based clinical studies, and enhancing our understanding of various pediatric knee injury mechanisms. They also support intervention design and prediction of surgical outcomes in pediatric populations.

Unsupervised multi-source variational domain adaptation for inter-subject SSVEP-based BCIs

Steady-State Visual Evoked Potential-based Brain–Computer Interfaces (SSVEP-based BCIs) are widely used to decode electroencephalography (EEG) data, often employing subject-specific models. However, these methods demand a substantial volume of labeled data, which tends to induce visual fatigue in subjects. To address this challenge, an efficient approach is to transfer known subject knowledge through inter-subject methods using unlabeled data. However, the previous state-of-the-art unsupervised inter-subject model is limited by the feature extraction capability of spatial filter methods. To overcome this limitation, we propose a novel Multi-source Variational Domain Adaptation (MVDA) method, which incorporates a Filter-bank Dynamic Graph Convolutional Network (FBDGCN). The FBDGCN aims to extract distinctive graph features under different harmonics from offline labeled source and unlabeled target EEG data. Subsequently, the source and target graph features are jointly trained using MVDA for domain adaptation. With proposed variational inference mechanism, MVDA can constrain the distribution manifold and minimize the domain discrepancy. Since the performance of knowledge transfer heavily relies on the similarity between source and target domains, we also propose an unsupervised subject correlation analysis method to select the source subjects from the dataset. Our experimental results demonstrate that the MVDA method yields a higher information transfer rate (ITR) (189.87 ± 24.98 bits/min) and mean accuracy (92.11 ± 2.85%). The proposed model can be calibrated only by unlabeled target subject data to obtain a higher recognition accuracy, which is more applicable to real-world scenarios.

Brain decoding of the Human Connectome Project tasks in a dense individual fMRI dataset

Brain decoding aims to infer cognitive states from patterns of brain activity. Substantial inter-individual variations in functional brain organization challenge accurate decoding performed at the group level. In this paper, we tested whether accurate brain decoding models can be trained entirely at the individual level. We trained several classifiers on a dense individual functional magnetic resonance imaging (fMRI) dataset for which six participants completed the entire Human Connectome Project (HCP) task battery >13 times over ten separate fMRI sessions. We evaluated nine decoding methods, from Support Vector Machines (SVM) and Multi-Layer Perceptron (MLP) to Graph Convolutional Neural Networks (GCN). All decoders were trained to classify single fMRI volumes into 21 experimental conditions simultaneously, using ∼7 h of fMRI data per participant. The best prediction accuracies were achieved with GCN and MLP models, whose performance (57–67 % accuracy) approached state-of-the-art accuracy (76 %) with models trained at the group level on >1 K hours of data from the original HCP sample. Our SVM model also performed very well (54–62 % accuracy). Feature importance maps derived from MLP —our best-performing model— revealed informative features in regions relevant to particular cognitive domains, notably in the motor cortex. We also observed that inter-subject classification achieved substantially lower accuracy than subject-specific models, indicating that our decoders learned individual-specific features. This work demonstrates that densely-sampled neuroimaging datasets can be used to train accurate brain decoding models at the individual level. We expect this work to become a useful benchmark for techniques that improve model generalization across multiple subjects and acquisition conditions.