Identification and Quantification of Precursory Changes of Rheumatoid Vasculitis in the Dorsalis Pedis Artery.

OBJECTIVE: Rheumatoid arthritis (RA) is a systemic connective tissue autoimmune disease that can infiltrate arterial walls. The delay in diagnosis and treatment of rheumatoid vasculitis (RV) in patients with RA may lead to irreversible damage to the arterial walls of small-to-medium vessels, which has serious and devastating consequences, most notably lung and cardiac damage. In this work an ultrasound image-based biomarker was developed to detect precursory changes in RV. METHODS: The ground truth was initiated from a medical diagnosis of RA, with arterial wall thickening of the proximal dorsalis pedis artery (DPA) indicating precursory changes of RV identified with ultrasound scanning. Ultrasound images of the DPA from 49 healthy subjects in the control group and 46 patients in the RA group were obtained. In total, 187 texture features were extracted from the images, followed by principal component analysis and linear discriminant analysis. RESULTS: The proposed biomarker detected a significant difference between the two groups (p = 5.74 × 10-18) with an area under the receiver operating characteristic curve of 0.85. Ten major textural features contributing most heavily to the biomarker were identified, with these textures being consistent with clinical observations of RV identified in previous studies. Interscan reproducibility was assessed by computing the biomarker twice based on repeated scans of each ankle. High interscan reproducibility was demonstrated by a strong and significant Pearson's coefficient (r = 0.85, p < 0.01) between the two repeated measurements of the proposed biomarker. CONCLUSION: The proposed biomarker can discriminate image textural differences seen in images acquired from RA patients, demonstrating precursory changes in RV compared with healthy controls. The major discriminative features identified in this study may facilitate the early identification and treatment of RV.

Concerto of movement: how expertise shapes the synergistic control of upper limb muscles in complex motor tasks with varying tempo and dynamics.

Objective. This research aims to reveal how the synergistic control of upper limb muscles adapts to varying requirements in complex motor tasks and how expertise shapes the motor modules.Approach. We study the muscle synergies of a complex, highly skilled and flexible task-piano playing-and characterize expertise-related muscle-synergy control that permits the experts to effortlessly execute the same task at different tempo and force levels. Surface EMGs (28 muscles) were recorded from adult novice (N= 10) and expert (N= 10) pianists as they played scales and arpeggios at different tempo-force combinations. Muscle synergies were factorized from EMGs.Main results. We found that experts were able to cover both tempo and dynamic ranges using similar synergy selections and achieved better performance, while novices altered synergy selections more to adapt to the changing tempi and keystroke intensities compared with experts. Both groups relied on fine-tuning the muscle weights within specific synergies to accomplish the different task styles, while the experts could tune the muscles in a greater number of synergies, especially when changing the tempo, and switch tempo over a wider range.Significance. Our study sheds light on the control mechanism underpinning expertise-related motor flexibility in highly skilled motor tasks that require decade-long training. Our results have implications on musical and sports training, as well as motor prosthetic design.

Rectified Latent Variable Model-Based EMG Factorization of Inhibitory Muscle Synergy Components Related to Aging, Expertise and Force-Tempo Variations.

Muscle synergy has been widely acknowledged as a possible strategy of neuromotor control, but current research has ignored the potential inhibitory components in muscle synergies. Our study aims to identify and characterize the inhibitory components within motor modules derived from electromyography (EMG), investigate the impact of aging and motor expertise on these components, and better understand the nervous system's adaptions to varying task demands. We utilized a rectified latent variable model (RLVM) to factorize motor modules with inhibitory components from EMG signals recorded from ten expert pianists when they played scales and pieces at different tempo-force combinations. We found that older participants showed a higher proportion of inhibitory components compared with the younger group. Senior experts had a higher proportion of inhibitory components on the left hand, and most inhibitory components became less negative with increased tempo or decreased force. Our results demonstrated that the inhibitory components in muscle synergies could be shaped by aging and expertise, and also took part in motor control for adapting to different conditions in complex tasks.

Ultrasound imaging of the dorsalis pedis artery as an early indicator of the precursory changes for rheumatoid vasculitis: A case series.

Introduction: Clinical verification of rheumatoid vasculitis (RV) persists as a mid-to-late diagnosis with medical imaging or biopsy. Early and subclinical presentations of RV, in particular, can remain underdiagnosed in the absence of adequate diagnostic testing. In this study, the research demonstrated the precursory changes for RV in patients with rheumatoid arthritis (RA) using non-invasive ultrasound imaging of a peripheral vessel. Method: Six participants were recruited: three participants with (RA) and three age- and gender-matched healthy controls. All participants completed a Foot Health Survey Questionnaire (FHSQ), and participants with RA completed a Rheumatoid Arthritis Disease Activity Index-5 (RADAI-5). Bilateral B-mode and Doppler ultrasound of the dorsalis pedis artery (DPA) was performed. The degree of inflammation, lumen and artery diameters, lumen diameter-to-artery diameter ratio and peak systolic velocity in the proximal DPA were compared between the two groups. Results: The mean RADAI-5 score (5.4 ± 0.8 out of 10) indicated moderate disease activity amongst participants with RA. Inflammation was observed in the DPA wall in all participants with RA, compared to no inflammation observed in the control group (Friedmans two-way analysis: χ2 = 15.733, P = 0.003). Differences between groups for inflammation, lumen diameter and lumen diameter-to-artery diameter ratio were found (P < 0.034), without differences for artery diameter and peak systolic velocity (P > 0.605). DPA wall inflammation did not correlate with FHSQ scores (r = -0.770, P = 0.073). Conclusion: Despite moderate RA disease activity, this is the first study to demonstrate the use of ultrasound to observe inflammation in small vessel disease. Our findings suggest ultrasound imaging may be a viable screening tool to demonstrate arterial wall inflammation, indicating the precursory changes of RV.

Tracking Bilateral Lower Limb Kinematics of Distance Runners on Treadmill Using a Single Inertial Measurement Unit.

Distance running related injuries are common, and many ailments have been associated with faulty posture. Conventional measurement of running kinematics requires sophisticated motion capture system in laboratory. In this study, we developed a wearable solution to accurately predict lower limb running kinematics using a single inertial measurement unit placed on the left lower leg. The running data collected from participants was used to train a model using long short-term memory (LSTM) neural networks with an inter-subject approach that predicted lower limb kinematics with an average accuracy of 80.2%, 85.8%, and 69.4% for sagittal hip, knee and ankle joint angles respectively for the ipsilateral limb. A comparable accuracy range was observed for the contralateral limb. The average RMSE (root mean squared error) of sagittal hip, knee and ankle were 8.76°, 13.13°, and 9.67° respectively for the ipsilateral limb. Analysis of contralateral limb kinematics was performed. The model established in this study can be used as a monitoring device to track essential running kinematics in natural running environments. Besides, the wearable solution can be an integral part of a real-time gait retraining biofeedback system for injury prevention and rehabilitation.

Comparing the Graphical Features of Simple Artificial Neural Networks and Cortical Development.

Brain development is characterized by changes in connections and information processing complexity. These changes inspire the training process of artificial neural network (ANN), which requires adjusting the neuron weights and biases to enhance efficiency in performing a specific task. In this work, we found affinities in the ratio of positive and negative weights in simple ANNs during training with that of excitatory and inhibitory synapses in the cortex. Additionally, we present a graphical representation of simple ANNs formed by pruning unimportant weights and aligning neurons and connections of different layers. Our findings suggest a strong relationship between the accuracy of simple neural network and graphical representation features, with graphical features at the inflection point resembling the graphical representation of the cortex.

DLATA: Deep Learning-Assisted transformation alignment of 2D brain slice histology.

Accurate alignment of brain slices is crucial for the classification of neuron populations by brain region, and for quantitative analysis in in vitro brain studies. Current semi-automated alignment workflows require labor intensive labeling of feature points on each slice image, which is time-consuming. To speed up the process in large-scale studies, we propose a method called Deep Learning-Assisted Transformation Alignment (DLATA), which uses deep learning to automatically identify feature points in images after training on a few labeled samples. DLATA only requires approximately 10% of the sample size of other semi-automated alignment workflows. Following feature point recognition, local weighted mean method is used as a geometrical transformation to align slice images for registration, achieving better results with about 4 fewer pixels of error than other semi-automated alignment workflows. DLATA can be retrained and successfully applied to the alignment of other biological tissue slices with different stains, including the typically challenging fluorescent stains. Reference codes and trained models for Nissl-stained coronal brain slices of mice can be found at https://github.com/ALIGNMENT2023/DLATA.

Reliability of Ultrasound Imaging in Examining Integrity and Inflammation in the Dorsalis Pedis Artery.

OBJECTIVE: Assessment of small vessel inflammatory diseases such as rheumatoid vasculitis is challenging. Small arteries such as the dorsalis pedis artery (DPA) are difficult to assess for changes in the arterial wall with medical imaging. Ultrasound imaging is a viable tool for examining the integrity and inflammatory changes in the arterial wall; however, no empirical data on its reliability have been described. METHODS: We measured the intra- and inter-rater reliability of ultrasound measurements across five parameters evaluating arterial integrity of the proximal DPA in participants with and without small vessel disease. We recruited 10 participants with rheumatoid arthritis and 10 healthy controls. Two sonographers using ultrasound independently measured DPA lumen diameter, artery diameter, lumen-to-arterial diameter ratio, arterial Doppler velocity and inflammatory changes in the proximal wall of the DPA. The intraclass correlation coefficient (ICC) was used to evaluate 95% confidence intervals within and between raters. Bland-Altman analyses were used to assess limits of agreement and were compared with minimal clinically important differences (MCID). RESULTS: Four of five selected parameters were found to have excellent intra- and inter-rater reliability within and between raters (ICC = 0.903-0.996). Acceptable reliability was found for measurement of arterial blood flow velocity within raters (ICC = 0.815-0.909), but not between raters (ICC = 0.634). Standard mean errors in all parameters were within minimal clinically important differences. CONCLUSION: Ultrasound imaging has been found to be a reliable method of assessment of arterial integrity and inflammation of the proximal DPA in people with small vessel disease. Evaluation of arterial blood flow velocity requires cautious interpretation.

An iterative neural processing sequence orchestrates feeding.

Feeding requires sophisticated orchestration of neural processes to satiate appetite in natural, capricious settings. However, the complementary roles of discrete neural populations in orchestrating distinct behaviors and motivations throughout the feeding process are largely unknown. Here, we delineate the behavioral repertoire of mice by developing a machine-learning-assisted behavior tracking system and show that feeding is fragmented and divergent motivations for food consumption or environment exploration compete throughout the feeding process. An iterative activation sequence of agouti-related peptide (AgRP)-expressing neurons in arcuate (ARC) nucleus, GABAergic neurons in the lateral hypothalamus (LH), and in dorsal raphe (DR) orchestrate the preparation, initiation, and maintenance of feeding segments, respectively, via the resolution of motivational conflicts. The iterative neural processing sequence underlying the competition of divergent motivations further suggests a general rule for optimizing goal-directed behaviors.

Power Law in Deep Neural Networks: Sparse Network Generation and Continual Learning With Preferential Attachment.

Training deep neural networks (DNNs) typically requires massive computational power. Existing DNNs exhibit low time and storage efficiency due to the high degree of redundancy. In contrast to most existing DNNs, biological and social networks with vast numbers of connections are highly efficient and exhibit scale-free properties indicative of the power law distribution, which can be originated by preferential attachment in growing networks. In this work, we ask whether the topology of the best performing DNNs shows the power law similar to biological and social networks and how to use the power law topology to construct well-performing and compact DNNs. We first find that the connectivities of sparse DNNs can be modeled by truncated power law distribution, which is one of the variations of the power law. The comparison of different DNNs reveals that the best performing networks correlated highly with the power law distribution. We further model the preferential attachment in DNNs evolution and find that continual learning in networks with growth in tasks correlates with the process of preferential attachment. These identified power law dynamics in DNNs can lead to the construction of highly accurate and compact DNNs based on preferential attachment. Inspired by the discovered findings, two novel applications have been proposed, including evolving optimal DNNs in sparse network generation and continual learning tasks with efficient network growth using power law dynamics. Experimental results indicate that the proposed applications can speed up training, save storage, and learn with fewer samples than other well-established baselines. Our demonstration of preferential attachment and power law in well-performing DNNs offers insight into designing and constructing more efficient deep learning.

Age-Related Modifications of Muscle Synergies and Their Temporal Activations for Overground Walking.

Healthy ageing modifies neuromuscular control of human overground walking. Previous studies found that ageing changes gait biomechanics, but whether there is concurrent ageing-related modulation of neuromuscular control remains unclear. We analyzed gait kinematics and electromyographic signals (EMGs; 14 lower-limb and trunk muscles) collected at three speeds during overground walking in 11 healthy young adults (mean age of 23.4 years) and 11 healthy elderlies (67.2 years). Neuromuscular control was characterized by extracting muscle synergies from EMGs and the synergies of both groups were k -means-clustered. The synergies of the two groups were grossly similar, but we observed numerous cluster- and muscle-specific differences between the age groups. At the population level, some hip-motion-related synergy clusters were more frequently identified in elderlies while others, more frequent in young adults. Such differences in synergy prevalence between the age groups are consistent with the finding that elderlies had a larger hip flexion range. For the synergies shared between both groups, the elderlies had higher inter-subject variability of the temporal activations than young adults. To further explore what synergy characteristics may be related to this inter-subject variability, we found that the inter-subject variance of temporal activations correlated negatively with the sparseness of the synergies in elderlies but not young adults during slow walking. Overall, our results suggest that as humans age, not only are the muscle synergies for walking fine-tuned in structure, but their temporal activation patterns are also more heterogeneous across individuals, possibly reflecting individual differences in prior sensorimotor experience or ageing-related changes in limb neuro-musculoskeletal properties.

Does Meta-Learning Improve EEG Motor Imagery Classification?

Deep learning has been applied to enhance the performance of EEG-based brain-computer interface applications. However, the cross-subject variations in EEG signals cause domain shifts and negatively affect the model performance and generalization. Meta-learning algorithms have shown fast new domain adaption in various fields, which may help solve the domain shift problems in EEG. Reptile, with satisfactory performance and low computational costs, stands out from other existing meta-learning algorithms. We integrated Reptile with a deep neural network as Reptile-EEG for the EEG motor imagery tasks, and compared Reptile-EEG with other state-of-the-art models in three motor imagery BCI benchmark datasets. Results show that Reptile-EEGdoes not outperform simple training of deep neural networks in motor imagery BCI tasks.

Sensor-Based Gait Retraining Lowers Knee Adduction Moment and Improves Symptoms in Patients with Knee Osteoarthritis: A Randomized Controlled Trial.

The present study compared the effect between walking exercise and a newly developed sensor-based gait retraining on the peaks of knee adduction moment (KAM), knee adduction angular impulse (KAAI), knee flexion moment (KFM) and symptoms and functions in patients with early medial knee osteoarthritis (OA). Eligible participants (n = 71) with early medial knee OA (Kellgren-Lawrence grade I or II) were randomized to either walking exercise or gait retraining group. Knee loading-related parameters including KAM, KAAI and KFM were measured before and after 6-week gait retraining. We also examined clinical outcomes including visual analog pain scale (VASP) and Knee Injury and Osteoarthritis Outcome Score (KOOS) at each time point. After gait retraining, KAM1 and VASP were significantly reduced (both Ps < 0.001) and KOOS significantly improved (p = 0.004) in the gait retraining group, while these parameters remained similar in the walking exercise group (Ps ≥ 0.448). However, KAM2, KAAI and KFM did not change in both groups across time (Ps ≥ 0.120). A six-week sensor-based gait retraining, compared with walking exercise, was an effective intervention to lower medial knee loading, relieve knee pain and improve symptoms for patients with early medial knee OA.

Classification of runners' performance levels with concurrent prediction of biomechanical parameters using data from inertial measurement units.

Identification of runner's performance level is critical to coaching, performance enhancement and injury prevention. Machine learning techniques have been developed to measure biomechanical parameters with body-worn inertial measurement unit (IMU) sensors. However, a robust method to classify runners is still unavailable. In this paper, we developed two models to classify running performance and predict biomechanical parameters of 30 subjects. We named the models RunNet-CNN and RunNet-MLP based on their architectures: convolutional neural network (CNN) and multilayer perceptron (MLP), respectively. In addition, we examined two validation approaches, subject-wise (leave-one-subject-out) and record-wise. RunNet-MLP classified runner's performance levels with an overall accuracy of 97.1%. Our results also showed that RunNet-CNN outperformed RunNet-MLP and gradient boosting decision tree in predicting biomechanical parameters. RunNet-CNN showed good agreement (R2 > 0.9) with the ground-truth reference on biomechanical parameters. The prediction accuracy for the record-wise method was better than the subject-wise method regardless of biomechanical parameters or models. Our findings showed the viability of using IMUs to produce reliable prediction of runners' performance levels and biomechanical parameters.

Right Temporal Oscillations of Infants in Relation to Contingent Learning.

Contingent learning is an agent for infants to explore the environment, which enhances the maturation of different developmental domains. This paper presents one of the first to investigate neural activities related to contingent learning of infants by analyzing their motor response that could elicit an audio-visual feedback. Three different kinds of motor response of infants were investigated, including unilateral kicks, synchronized kicks, and alternate kicks. Electroencephalographic (EEG) signals of infants were recorded before the motor experiments. Higher theta band power and lower upper beta power at the right temporal lobe of infants predicted a higher ratio of total unilateral kicks and a lower ratio of synchronized kicks at the later acquisition stage of the experiment. As contingent learning could be reflected by specific motor response in relation to the audio-visual stimuli, the results suggested that right temporal oscillations could predict different levels of contingent learning of infants.

Machine Learning-Driven Drug Discovery: Prediction of Structure-Cytotoxicity Correlation Leads to Identification of Potential Anti-Leukemia Compounds.

In vitro cytotoxicity screening is a crucial step of anticancer drug discovery. The application of deep learning methodology is gaining increasing attentions in processing drug screening data and studying anticancer mechanisms of chemical compounds. In this work, we explored the utilization of convolutional neural network in modeling the anticancer efficacy of small molecules. In particular, we presented a VGG19 model trained on 2D structural formulae to predict the growth-inhibitory effects of compounds against leukemia cell line CCRF-CEM, without any use of chemical descriptors. The model achieved a normalized RMSE of 15.76% on predicting growth inhibition and a Pearson Correlation Coefficient of 0.72 between predicted and experimental data, demonstrating a strong predictive power in this task. Furthermore, we implemented the Layer-wise Relevance Propagation technique to interpret the network and visualize the chemical groups predicted by the model that contribute to toxicity with human-readable representations.Clinical relevance-This work predicts the cytotoxicity of chemical compounds against human leukemic lymphoblast CCRF-CEM cell lines on a continuous scale, which only requires 2D images of the structural formulae of the compounds as inputs. Knowledge in the structure-toxicity relationship of small molecules will potentially increase the hit rate of primary drug screening assays.

An investigation of in-ear sensing for motor task classification.

Objective.Our study aims to investigate the feasibility of in-ear sensing for human-computer interface.Approach.We first measured the agreement between in-ear biopotential and scalp-electroencephalogram (EEG) signals by channel correlation and power spectral density analysis. Then we applied EEG compact network (EEGNet) for the classification of a two-class motor task using in-ear electrophysiological signals.Main results.The best performance using in-ear biopotential with global reference reached an average accuracy of 70.22% (cf 92.61% accuracy using scalp-EEG signals), but the performance in-ear biopotential with near-ear reference was poor.Significance.Our results suggest in-ear sensing would be a viable human-computer interface for movement prediction, but careful consideration should be given to the position of the reference electrode.

Comparing biological and artificial vision systems: Network measures of functional connectivity.

Advances in Deep Convolutional Neural Networks (DCNN) provide new opportunities for computational neuroscience to pose novel questions regarding the function of biological visual systems. Some attempts have been made to utilize advances in machine learning to answer neuroscientific questions, but how to appropriately make comparisons between the biological systems and artificial neural network structure is an open question. This analysis quantifies network properties of the mouse visual system and a common DCNN model (VGG16), to determine if this comparison is appropriate. Utilizing weighted graph-theoretic measures of node density (weighted node-degree), path length, local clustering coefficient, and betweenness, differences in functional connectivity patterns in the modern artificial computer vision system and the biological vision system are quantified. Results show that the mouse exhibits network measure distributions more similar to Poisson than normal, whereas the VGG16 exhibits network measure distributions with a more Gaussian shape than the sampled biological network. The artificial network shows higher density measures and shorter path lengths in comparison to the biological network. These results show that training a VGG16 for an object recognition task is unlikely to produce a network whose functional connectivity is similar to the mammalian visual system.

Real-Time Estimation of Knee Adduction Moment for Gait Retraining in Patients With Knee Osteoarthritis.

Previous clinical studies have reported that gait retraining is an effective non-invasive intervention for patients with medial compartment knee osteoarthritis. These gait retraining programs often target a reduction in the knee adduction moment (KAM), which is a commonly used surrogate marker to estimate the loading in the medial compartment of the tibiofemoral joint. However, conventional evaluation of KAM requires complex and costly equipment for motion capture and force measurement. Gait retraining programs, therefore, are usually confined to a laboratory environment. In this study, machine learning techniques were applied to estimate KAM during walking with data collected from two low-cost wearable sensors. When compared to the traditional laboratory-based measurement, our mobile solution using artificial neural network (ANN) and XGBoost achieved an excellent agreement with R2 of 0.956 and 0.947 respectively. With the implementation of a real-time audio feedback system, the present algorithm may provide a viable solution for gait retraining outside laboratory. Clinical treatment strategies can be developed using the continuous feedback provided by our system.

Multimodal hand gesture recognition using single IMU and acoustic measurements at wrist.

To facilitate hand gesture recognition, we investigated the use of acoustic signals with an accelerometer and gyroscope at the human wrist. As a proof-of-concept, the prototype consisted of 10 microphone units in contact with the skin placed around the wrist along with an inertial measurement unit (IMU). The gesture recognition performance was evaluated through the identification of 13 gestures used in daily life. The optimal area for acoustic sensor placement at the wrist was examined using the minimum redundancy and maximum relevance feature selection algorithm. We recruited 10 subjects to perform over 10 trials for each set of hand gestures. The accuracy was 75% for a general model with the top 25 features selected, and the intra-subject average classification accuracy was over 80% with the same features using one microphone unit at the mid-anterior wrist and an IMU. These results indicate that acoustic signatures from the human wrist can aid IMU sensing for hand gesture recognition, and the selection of a few common features for all subjects could help with building a general model. The proposed multimodal framework helps address the single IMU sensing bottleneck for hand gestures during arm movement and/or locomotion.