A community effort in SARS-CoV-2 drug discovery.

The COVID-19 pandemic continues to pose a substantial threat to human lives and is likely to do so for years to come. Despite the availability of vaccines, searching for efficient small-molecule drugs that are widely available, including in low- and middle-income countries, is an ongoing challenge. In this work, we report the results of an open science community effort, the "Billion molecules against COVID-19 challenge", to identify small-molecule inhibitors against SARS-CoV-2 or relevant human receptors. Participating teams used a wide variety of computational methods to screen a minimum of 1 billion virtual molecules against 6 protein targets. Overall, 31 teams participated, and they suggested a total of 639,024 molecules, which were subsequently ranked to find 'consensus compounds'. The organizing team coordinated with various contract research organizations (CROs) and collaborating institutions to synthesize and test 878 compounds for biological activity against proteases (Nsp5, Nsp3, TMPRSS2), nucleocapsid N, RdRP (only the Nsp12 domain), and (alpha) spike protein S. Overall, 27 compounds with weak inhibition/binding were experimentally identified by binding-, cleavage-, and/or viral suppression assays and are presented here. Open science approaches such as the one presented here contribute to the knowledge base of future drug discovery efforts in finding better SARS-CoV-2 treatments.

DENVIS: Scalable and High-Throughput Virtual Screening Using Graph Neural Networks with Atomic and Surface Protein Pocket Features.

Computational methods for virtual screening can dramatically accelerate early-stage drug discovery by identifying potential hits for a specified target. Docking algorithms traditionally use physics-based simulations to address this challenge by estimating the binding orientation of a query protein-ligand pair and a corresponding binding affinity score. Over the recent years, classical and modern machine learning architectures have shown potential for outperforming traditional docking algorithms. Nevertheless, most learning-based algorithms still rely on the availability of the protein-ligand complex binding pose, typically estimated via docking simulations, which leads to a severe slowdown of the overall virtual screening process. A family of algorithms processing target information at the amino acid sequence level avoid this requirement, however, at the cost of processing protein data at a higher representation level. We introduce deep neural virtual screening (DENVIS), an end-to-end pipeline for virtual screening using graph neural networks (GNNs). By performing experiments on two benchmark databases, we show that our method performs competitively to several docking-based, machine learning-based, and hybrid docking/machine learning-based algorithms. By avoiding the intermediate docking step, DENVIS exhibits several orders of magnitude faster screening times (i.e., higher throughput) than both docking-based and hybrid models. When compared to an amino acid sequence-based machine learning model with comparable screening times, DENVIS achieves dramatically better performance. Some key elements of our approach include protein pocket modeling using a combination of atomic and surface features, the use of model ensembles, and data augmentation via artificial negative sampling during model training. In summary, DENVIS achieves competitive to state-of-the-art virtual screening performance, while offering the potential to scale to billions of molecules using minimal computational resources.

Discrete Action Control for Prosthetic Digits.

We aim to develop a paradigm for simultaneous and independent control of multiple degrees of freedom (DOFs) for upper-limb prostheses. To that end, we introduce action control, a novel method to operate prosthetic digits with surface electromyography (EMG) based on multi-output, multi-class classification. At each time step, the decoder classifies movement intent for each controllable DOF into one of three categories: open, close, or stall (i.e., no movement). We implemented a real-time myoelectric control system using this method and evaluated it by running experiments with one unilateral and two bilateral amputees. Participants controlled a six-DOF bar interface on a computer display, with each DOF corresponding to a motor function available in multi-articulated prostheses. We show that action control can significantly and systematically outperform the state-of-the-art method of position control via multi-output regression in both task- and non-task-related measures. Using the action control paradigm, improvements in median task performance over regression-based control ranged from 20.14% to 62.32% for individual participants. Analysis of a post-experimental survey revealed that all participants rated action higher than position control in a series of qualitative questions and expressed an overall preference for the former. Action control has the potential to improve the dexterity of upper-limb prostheses. In comparison with regression-based systems, it only requires discrete instead of real-valued ground truth labels, typically collected with motion tracking systems. This feature makes the system both practical in a clinical setting and also suitable for bilateral amputation. This work is the first demonstration of myoelectric digit control in bilateral upper-limb amputees. Further investigation and pre-clinical evaluation are required to assess the translational potential of the method.

Recalibration of myoelectric control with active learning.

Introduction: Improving the robustness of myoelectric control to work over many months without the need for recalibration could reduce prosthesis abandonment. Current approaches rely on post-hoc error detection to verify the certainty of a decoder's prediction using predefined threshold value. Since the decoder is fixed, the performance decline over time is inevitable. Other approaches such as supervised recalibration and unsupervised self-recalibration entail limitations in scaling up and computational resources. The objective of this paper is to study active learning as a scalable, human-in-the-loop framework, to improve the robustness of myoelectric control. Method: Active learning and linear discriminate analysis methods were used to create an iterative learning process, to modify decision boundaries based on changes in the data. We simulated a real-time scenario. We exploited least confidence, smallest margin and entropy reduction sampling strategies in single and batch-mode sample selection. Optimal batch-mode sampling was considered using ranked batch-mode active learning. Results: With only 3.2 min of data carefully selected by the active learner, the decoder outperforms random sampling by 4-5 and ~2% for able-bodied and people with limb difference, respectively. We observed active learning strategies to systematically and significantly enhance the decoders adaptation while optimizing the amount of training data on a class-specific basis. Smallest margin and least confidence uncertainty were shown to be the most supreme. Discussion: We introduce for the first time active learning framework for long term adaptation in myoelectric control. This study simulates closed-loop environment in an offline manner and proposes a pipeline for future real-time deployment.

Co-creation and User Perspectives for Upper Limb Prosthetics.

People who either use an upper limb prosthesis and/or have used services provided by a prosthetic rehabilitation centre, experience limitations of currently available prosthetic devices. Collaboration between academia and a broad range of stakeholders, can lead to the development of solutions that address peoples' needs. By doing so, the rate of prosthetic device abandonment can decrease. Co-creation is an approach that can enable collaboration of this nature to occur throughout the research process. We present findings of a co-creation project that gained user perspectives from a user survey, and a subsequent workshop involving: people who use an upper limb prosthesis and/or have experienced care services (users), academics, industry experts, charity executives, and clinicians. The survey invited users to prioritise six themes, which academia, clinicians, and industry should focus on over the next decade. The prioritisation of the themes concluded in the following order, with the first as the most important: function, psychology, aesthetics, clinical service, collaboration, and media. Within five multi-stakeholder groups, the workshop participants discussed challenges and collaborative opportunities for each theme. Workshop groups prioritised the themes based on their discussions, to highlight opportunities for further development. Two groups chose function, one group chose clinical service, one group chose collaboration, and another group chose media. The identified opportunities are presented within the context of the prioritised themes, including the importance of transparent information flow between all stakeholders; user involvement throughout research studies; and routes to informing healthcare policy through collaboration. As the field of upper limb prosthetics moves toward in-home research, we present co-creation as an approach that can facilitate user involvement throughout the duration of such studies.

Intramuscular EMG For Abstract Myoelectric Control: A Proof Of Concept Study.

Myoelectric prostheses are commonly controlled by surface EMG. Many control algorithms, including the user learning-based control paradigm abstract control, benefit from independent control signals. Measuring at the surface of the skin reduces the signal independence through cross talk. To increase the number of independent signals, intramuscular EMG recordings might be a viable alternative for myoelectric control. This proof of concept study investigated if real time abstract myoelectric control is possible with intramuscular measurements. Six participants performed a 4-target and 12-target abstract control task with both surface and intramuscular EMG recordings. The results suggest that intramuscular EMG is suitable for abstract control, and that performance could be increased in the future by stabilizing the amplitude of the processed intramuscular EMG signal.

Automatic Myoelectric Control Site Detection Using Candid Covariance-Free Incremental Principal Component Analysis.

The unknown composition of residual muscles surrounding the stump of an amputee makes optimal electrode placement challenging. This often causes the experimental set-up and calibration of upper-limb prostheses to be time consuming. In this work, we propose the use of existing dimensionality reduction techniques, typically used for muscle synergy analysis, to provide meaningful real-time functional information of the residual muscles during the calibration period. Two variations of principal component analysis (PCA) were applied to electromyography (EMG) data collected during a myoelectric task. Candid covariance-free incremental PCA (CCIPCA) detected task-specific muscle synergies with high accuracy using minimal amounts of data. Our findings offer a real-time solution towards optimizing calibration periods.

Myoelectric digit action decoding with multi-output, multi-class classification: an offline analysis.

The ultimate goal of machine learning-based myoelectric control is simultaneous and independent control of multiple degrees of freedom (DOFs), including wrist and digit artificial joints. For prosthetic finger control, regression-based methods are typically used to reconstruct position/velocity trajectories from surface electromyogram (EMG) signals. Unfortunately, such methods have thus far met with limited success. In this work, we propose action decoding, a paradigm-shifting approach for independent, multi-digit movement intent prediction based on multi-output, multi-class classification. At each moment in time, our algorithm decodes movement intent for each available DOF into one of three classes: open, close, or stall (i.e., no movement). Despite using a classifier as the decoder, arbitrary hand postures are possible with our approach. We analyse a public dataset previously recorded and published by us, comprising measurements from 10 able-bodied and two transradial amputee participants. We demonstrate the feasibility of using our proposed action decoding paradigm to predict movement action for all five digits as well as rotation of the thumb. We perform a systematic offline analysis by investigating the effect of various algorithmic parameters on decoding performance, such as feature selection and choice of classification algorithm and multi-output strategy. The outcomes of the offline analysis presented in this study will be used to inform the real-time implementation of our algorithm. In the future, we will further evaluate its efficacy with real-time control experiments involving upper-limb amputees.

Model-Based Control of Individual Finger Movements for Prosthetic Hand Function.

Prosthetic devices for hand difference have advanced considerably in recent years, to the point where the mechanical dexterity of a state-of-the-art prosthetic hand approaches that of the natural hand. Control options for users, however, have not kept pace, meaning that the new devices are not used to their full potential. Promising developments in control technology reported in the literature have met with limited commercial and clinical success. We have previously described a biomechanical model of the hand that could be used for prosthesis control. The goal of this study was to evaluate the feasibility of this approach in terms of kinematic fidelity of model-predicted finger movement and the computational performance of the model. We show the performance of the model in replicating recorded hand and finger kinematics and find average correlations of 0.89 between modelled and recorded motions; we show that the computational performance of the simulations is fast enough to achieve real-time control with a robotic hand in the loop; and we describe the use of the model for controlling object gripping. Despite some limitations in accessing sufficient driving signals, the model performance shows promise as a controller for prosthetic hands when driven with recorded EMG signals. User-in-the-loop testing with amputees is necessary in future work to evaluate the suitability of available driving signals, and to examine translation of offline results to online performance.

Multi-Grip Classification-Based Prosthesis Control With Two EMG-IMU Sensors.

In the field of upper-limb myoelectric prosthesis control, the use of statistical and machine learning methods has been long proposed as a means of enabling intuitive grip selection and actuation. Recently, this paradigm has found its way toward commercial adoption. Machine learning-based prosthesis control typically relies on the use of a large number of electrodes. Here, we propose an end-to-end strategy for multi-grip, classification-based prosthesis control using only two sensors, comprising electromyography (EMG) electrodes and inertial measurement units (IMUs). We emphasize the importance of accurately estimating posterior class probabilities and rejecting predictions made with low confidence, so as to minimize the rate of unintended prosthesis activations. To that end, we propose a confidence-based error rejection strategy using grip-specific thresholds. We evaluate the efficacy of the proposed system with real-time pick and place experiments using a commercial multi-articulated prosthetic hand and involving 12 able-bodied and two transradial (i.e., below-elbow) amputee participants. Results promise the potential for deploying intuitive, classification-based multi-grip control in existing upper-limb prosthetic systems subject to small modifications.

Effect of User Practice on Prosthetic Finger Control With an Intuitive Myoelectric Decoder.

Machine learning-based myoelectric control is regarded as an intuitive paradigm, because of the mapping it creates between muscle co-activation patterns and prosthesis movements that aims to simulate the physiological pathways found in the human arm. Despite that, there has been evidence that closed-loop interaction with a classification-based interface results in user adaptation, which leads to performance improvement with experience. Recently, there has been a focus shift toward continuous prosthesis control, yet little is known about whether and how user adaptation affects myoelectric control performance in dexterous, intuitive tasks. We investigate the effect of short-term adaptation with independent finger position control by conducting real-time experiments with 10 able-bodied and two transradial amputee subjects. We demonstrate that despite using an intuitive decoder, experience leads to significant improvements in performance. We argue that this is due to the lack of an utterly natural control scheme, which is mainly caused by differences in the anatomy of human and artificial hands, movement intent decoding inaccuracies, and lack of proprioception. Finally, we extend previous work in classification-based and wrist continuous control by verifying that offline analyses cannot reliably predict real-time performance, thereby reiterating the importance of validating myoelectric control algorithms with real-time experiments.

Continuous Versus Discrete Simultaneous Control of Prosthetic Fingers.

Modern, commercially available hand prostheses offer the potential of individual digit control. However, this feature is often not utilized due to the lack of a robust scheme for finger motion estimation from surface electromyographic (EMG) measurements. Regression methods have been proposed to achieve closed-loop finger position, velocity, or force control. In this paper, we propose an alternative approach, based on open-loop action-based control, which could be achieved through simultaneous finger motion classification. We compare the efficacy of continuous closed-loop and discrete open-loop control on the task of controlling the five degrees of actuation (DOAs) of a dexterous robotic hand. Eight normally-limbed subjects were instructed to teleoperate the hand using a data glove and the two control schemes under investigation in order to match target postures presented to them on a screen as closely as possible. Results indicate that, firstly, the performance of the two control methods is comparable and, secondly, that experience can lead to significant performance improvement over time, regardless of the method used. These results suggest that prosthetic finger control in a continuous space can be potentially achieved by means of myoelectric classification and discrete, action-based control and hence encourage further research in this direction.

Spatio-Temporal Inertial Measurements Feature Extraction Improves Hand Movement Pattern Recognition without Electromyography.

Recent studies indicate the limited clinical acceptance of myoelectric prostheses, as upper extremity amputees need improved functionality and more intuitive, effective, and coordinated control of their artificial limbs. Rather than exclusively classifying the electromyogram (EMG) signals, it has been shown that inertial measurements (IMs) can form an excellent complementary signal to the EMG signals to improve the prosthetic control robustness. We present an investigation into the possibility of replacing, rather than complementing, the EMG signals with IMs. We hypothesize that the enhancements achieved by the combined use of the EMG and IM signals may not be significantly different from that achieved by the use of Magnetometer (MAG) or Accelerometer (ACC) signals only, when the temporal and spatial information aspects are considered. A large dataset comprising recordings with 20 ablebodied and two amputee participants, executing 40 movements, was collected. A systematic performance comparison across a number of feature extraction methods was carried out to test our hypothesis. Results suggest that, individually, each of the ACC and MMG signals can form an excellent and potentially independent source of control signal for upper-limb prostheses, with an average classification accuracy of $\approx 93$% across all subjects. This study suggests the feasibility of moving from surface EMG to IM signals as a main source for upper-limb prosthetic control in real-life applications.

CMOS Magnetic Sensors for Wearable Magnetomyography.

Magnetomyography utilizes magnetic sensors to record small magnetic fields produced by the electrical activity of muscles, which also gives rise to the electromyogram (EMG) signal typically recorded with surface electrodes. Detection and recording of these small fields requires sensitive magnetic sensors possibly equipped with a CMOS readout system. This paper presents a highly sensitive Hall sensor fabricated in a standard $0.18~\mu \mathrm {m}$ CMOS technology for future low-field MMG applications. Compared with previous works, our experimental results show that the proposed Hall sensor achieves a higher current mode sensitivity of approximately 2400 V/A/mT. Further refinement is required to enable measurement of MMG signals from muscles.

Improved prosthetic hand control with concurrent use of myoelectric and inertial measurements.

BACKGROUND: Myoelectric pattern recognition systems can decode movement intention to drive upper-limb prostheses. Despite recent advances in academic research, the commercial adoption of such systems remains low. This limitation is mainly due to the lack of classification robustness and a simultaneous requirement for a large number of electromyogram (EMG) electrodes. We propose to address these two issues by using a multi-modal approach which combines surface electromyography (sEMG) with inertial measurements (IMs) and an appropriate training data collection paradigm. We demonstrate that this can significantly improve classification performance as compared to conventional techniques exclusively based on sEMG signals. METHODS: We collected and analyzed a large dataset comprising recordings with 20 able-bodied and two amputee participants executing 40 movements. Additionally, we conducted a novel real-time prosthetic hand control experiment with 11 able-bodied subjects and an amputee by using a state-of-the-art commercial prosthetic hand. A systematic performance comparison was carried out to investigate the potential benefit of incorporating IMs in prosthetic hand control. RESULTS: The inclusion of IM data improved performance significantly, by increasing classification accuracy (CA) in the offline analysis and improving completion rates (CRs) in the real-time experiment. Our findings were consistent across able-bodied and amputee subjects. Integrating the sEMG electrodes and IM sensors within a single sensor package enabled us to achieve high-level performance by using on average 4-6 sensors. CONCLUSIONS: The results from our experiments suggest that IMs can form an excellent complimentary source signal for upper-limb myoelectric prostheses. We trust that multi-modal control solutions have the potential of improving the usability of upper-extremity prostheses in real-life applications.

Towards low-dimensionsal proportional myoelectric control.

One way of enhancing the dexterity of powered myoelectric prostheses is via proportional and simultaneous control of multiple degrees-of-freedom (DOFs). Recently, it has been demonstrated that the reconstruction of finger movement is feasible by using features of the surface electromyogram (sEMG) signal. In such paradigms, the number of predictors and target variables is usually large, and strong correlations are present in both the input and output domains. Synergistic patterns in the sEMG space have been previously exploited to facilitate kinematics decoding. In this work, we propose a framework for simultaneous input-output dimensionality reduction based on the generalized eigenvalue problem formulation of multiple linear regression (MLR). We demonstrate that the proposed methodology outperforms simultaneous input-output dimensionality reduction based on principal component analysis (PCA), while the prediction accuracy of the full rank regression (FRR) method can be achieved by using only a few relevant dimensions.

Generalizability of EMG decoding using local field potentials.

Motor cortical local field potentials (LFPs) have been successfully used to decode both kinematics and kinetics of arm movement. For future clinically viable prostheses, however, brain activity decoders will have to generalize well under a wide spectrum of behavioral conditions. This property has not yet been demonstrated clearly. Here, we provide evidence for the first time, that an LFP-based electromyogram (EMG) decoder can generalize reasonably well across two different types of behavior. We implanted intracortical microelectrode arrays in the primary motor (M1) and ventral pre-motor (PMv) cortices of a rhesus macaque, and recorded LFP and EMG activity from arm and hand muscles of the contralateral forelimb during a two-dimensional (2-D) centre-out isometric wrist torque task (TT), and during free reach and grasp behavior (FB). Selected temporal and spectral features of the LFP signals were used to train EMG decoders using data from both types of behavior separately. We assessed the decoding performance for both within- and across-task cases. The average achieved generalization score was 65 ± 20%, while in many cases individual scores reached 100%.

Development of a real time sparse non-negative matrix factorization module for cochlear implants by using xPC target.

Cochlear implants (CIs) require efficient speech processing to maximize information transmission to the brain, especially in noise. A novel CI processing strategy was proposed in our previous studies, in which sparsity-constrained non-negative matrix factorization (NMF) was applied to the envelope matrix in order to improve the CI performance in noisy environments. It showed that the algorithm needs to be adaptive, rather than fixed, in order to adjust to acoustical conditions and individual characteristics. Here, we explore the benefit of a system that allows the user to adjust the signal processing in real time according to their individual listening needs and their individual hearing capabilities. In this system, which is based on MATLAB®, SIMULINK® and the xPC Target™ environment, the input/outupt (I/O) boards are interfaced between the SIMULINK blocks and the CI stimulation system, such that the output can be controlled successfully in the manner of a hardware-in-the-loop (HIL) simulation, hence offering a convenient way to implement a real time signal processing module that does not require any low level language. The sparsity constrained parameter of the algorithm was adapted online subjectively during an experiment with normal-hearing subjects and noise vocoded speech simulation. Results show that subjects chose different parameter values according to their own intelligibility preferences, indicating that adaptive real time algorithms are beneficial to fully explore subjective preferences. We conclude that the adaptive real time systems are beneficial for the experimental design, and such systems allow one to conduct psychophysical experiments with high ecological validity.