Myoelectric Upper-limb Prostheses Research Articles

Perspectives on the comparative benefits of body-powered and myoelectric upper limb prostheses

BackgroundPatient access to body-powered and myoelectric upper limb prostheses in the United States is often restricted by a healthcare system that prioritizes prosthesis prescription based on cost and perceived value. Although this system operates on an underlying assumption that design differences between these prostheses leads to relative advantages and disadvantages of each device, there is limited empirical evidence to support this view.Main textThis commentary article will review a series of studies conducted by our research team with the goal of differentiating how prosthesis design might impact user performance on a variety of interrelated domains. Our central hypothesis is that the design and actuation method of body-powered and myoelectric prostheses might affect users’ ability to access sensory feedback and account for device properties when planning movements. Accordingly, other domains that depend on these abilities may also be affected. While our work demonstrated some differences in availability of sensory feedback based on prosthesis design, this did not result in consistent differences in prosthesis embodiment, movement accuracy, movement quality, and overall kinematic patterns.ConclusionCollectively, our findings suggest that performance may not necessarily depend on prosthesis design, allowing users to be successful with either device type depending on the circumstances. Prescription practices should rely more on individual needs and preferences than cost or prosthesis design. However, we acknowledge that there remains a dearth of evidence to inform decision-making and that an expanded research focus in this area will be beneficial.

The Utility of Synthetic Reflexes and Haptic Feedback for Upper-Limb Prostheses in a Dexterous Task Without Direct Vision.

Individuals who use myoelectric upper-limb prostheses often rely heavily on vision to complete their daily activities. They thus struggle in situations where vision is overloaded, such as multitasking, or unavailable, such as poor lighting conditions. Able-bodied individuals can easily accomplish such tasks due to tactile reflexes and haptic sensation guiding their upper-limb motor coordination. Based on these principles, we developed and tested two novel prosthesis systems that incorporate autonomous controllers and provide the user with touch-location feedback through either vibration or distributed pressure. These capabilities were made possible by installing a custom contact-location sensor on the fingers of a commercial prosthetic hand, along with a custom pressure sensor on the thumb. We compared the performance of the two systems against a standard myoelectric prosthesis and a myoelectric prosthesis with only autonomous controllers in a difficult reach-to-pick-and-place task conducted without direct vision. Results from 40 able-bodied participants in this between-subjects study indicated that vibrotactile feedback combined with synthetic reflexes proved significantly more advantageous than the standard prosthesis in several of the task milestones. In addition, vibrotactile feedback and synthetic reflexes improved grasp placement compared to only synthetic reflexes or pressure feedback combined with synthetic reflexes. These results indicate that autonomous controllers and haptic feedback together facilitate success in dexterous tasks without vision, and that the type of haptic display matters.

An Exploration of the Optimal Feature-Classifier Combinations for Transradial Prosthesis Control.

Within state-of-the-art gesture-based upper-limb myoelectric prosthesis control, gesture recognition commonly relies on the classification of features extracted from electromyographic (EMG) data gathered from the amputee's residual forearm musculature. Despite best efforts in broadly maximizing gesture recognition accuracy, there does not yet exist a feature-classifier combination accepted as best-practice. In turn, this work hypothesizes that no single feature-classifier combination can consistently maximize accuracy across subjects, positing instead that control schemes should be personalized to the individual. To investigate this hypothesis, the study employed the 40-subject, 49-gesture Ninapro Database 2 (DB2) to compare the performance of 7 different historic, more recent, and state-of-the-art feature sets when classified by 5 machine learning algorithms commonly seen within EMG-based pattern recognition literature. The results demonstrate the ability of Linear Discriminant Analysis (LDA) to marginally exceed other more computationally intensive classifiers in terms of mean accuracy, while the feature set which maximized the highest proportion of individuals' accuracies was shown to vary with both classifier choice and gesture count.

A Haptic Sleeve as a Method of Mechanotactile Feedback Restoration for Myoelectric Hand Prosthesis Users.

Current myoelectric upper limb prostheses do not restore sensory feedback, impairing fine motor control. Mechanotactile feedback restoration with a haptic sleeve may rectify this problem. This randomised crossover within-participant controlled study aimed to assess a prototype haptic sleeve's effect on routine grasping tasks performed by eight able-bodied participants. Each participant completed 15 repetitions of the three tasks: Task 1—normal grasp, Task 2—strong grasp and Task 3—weak grasp, using visual, haptic, or combined feedback All data were collected in April 2021 in the Scottish Microelectronics Centre, Edinburgh, UK. Combined feedback correlated with significantly higher grasp success rates compared to the vision alone in Task 1 (p < 0.0001), Task 2 (p = 0.0057), and Task 3 (p = 0.0170). Similarly, haptic feedback was associated with significantly higher grasp success rates compared to vision in Task 1 (p < 0.0001) and Task 2 (p = 0.0015). Combined feedback correlated with significantly lower energy expenditure compared to visual feedback in Task 1 (p < 0.0001) and Task 3 (p = 0.0003). Likewise, haptic feedback was associated with significantly lower energy expenditure compared to the visual feedback in Task 1 (p < 0.0001), Task 2 (p < 0.0001), and Task 3 (p < 0.0001). These results suggest that mechanotactile feedback provided by the haptic sleeve effectively augments grasping and reduces its energy expenditure.

Analysis of Myoelectric Signals to Prosthesis Applications

This work presents a myoelectric system measurement incorporated in an arm band device, particularly designed for application in upper-limb myoelectric prosthesis with pattern recognition-based sEMG control. For this particular use, specifications such as low cost, high sampling frequency, size and weight, may be observed. Moreover, it is introduced an acquisition protocol to evaluation of the myoelectric signal in daily activities.

Addressing unpredictability may be the key to improving performance with current clinically prescribed myoelectric prostheses

The efferent control chain for an upper-limb myoelectric prosthesis can be separated into 3 key areas: signal generation, signal acquisition, and device response. Data were collected from twenty trans-radial myoelectric prosthesis users using their own clinically prescribed devices, to establish the relative impact of these potential control factors on user performance (user functionality and everyday prosthesis usage). By identifying the key factor(s), we can guide future developments to ensure clinical impact. Skill in generating muscle signals was assessed via reaction times and signal tracking. To assess the predictability of signal acquisition, we inspected reaction time spread and undesired hand activations. As a measure of device response, we recorded the electromechanical delay between electrode stimulation and the onset of hand movement. Results suggest abstract measures of skill in controlling muscle signals are poorly correlated with performance. Undesired activations of the hand or incorrect responses were correlated with almost all kinematics and gaze measures suggesting unpredictability is a key factor. Significant correlations were also found between several measures of performance and the electromechanical delay; however, unexpectedly, longer electromechanical delays correlated with better performance. Future research should focus on exploring causes of unpredictability, their relative impacts on performance and interventions to address this.

Robotic hand design with linear actuators based on Toronto development

In this work, the design of a robotic hand with 7 degrees of freedom is presented that allows greater flexibility, achieving the usual actions performed by a normal hand. The work consists of a prototype designed with linear actuators and myoelectric sensor, following the mechanism of the University of Toronto for the management of functional phalanges. The design, construction description, components and recommendations for the elaboration of a flexible and useful robotic hand for amputee patients with a residual limb for the socket are presented.&#x0D; Keywords: Robotic hand, Degree of freedom, Toronto´s Mechanism, lineal actuator.&#x0D; References&#x0D; [1]W. Diane, J. Braza and M. Yacub, Essentials of Physical Medicine and Rehabilitation, 4th ed. Philadelphia: Walter R. Frontera and Julie K. Silver and Thomas D. Rizzo, 2020, pp. 651 - 657.&#x0D; [2]A. Heerschop, C. Van Der Sluis, E. Otten, &amp; R.M. Bongers, Looking beyond proportional control: The relevance of mode switching in learning to operate multi-articulating myoelectric upper-limb prostheses, . Biomedical Signal Processing and Control, 2020, doi:10.1016/j.bspc.2019.101647.&#x0D; [3]L. Heisnam, B. Suthar, 20 DOF robotic hand for tele-operation: — Design, simulation, control and accuracy test with leap motion. 2016 International Conference on Robotics and Automation for Humanitarian Applications (RAHA), 2016, doi:10.1109/raha.2016.7931886.&#x0D; [4]Y. Mishima, R. Ozawa, Design of a robotic finger using series gear chain mechanisms. 2014 IEEE/RSJ International Conference on Intelligent Robots and Systems, 2014, doi:10.1109/iros.2014.6942961.&#x0D; [5]N. Dechev, W. Cleghorn, S. Naumann, Multi-segmented finger design of an experimental prosthetic hand,Proceedings of the Sixth National Applied Mechanisms &amp; Robotics Conference, december 1999.&#x0D; [6]O. Flor, “Building a mobile robot,” Education for the future. Accessed on: December 29, 2019. [Online] Available: https://omarflor2014.wixsite.com/misitio.&#x0D; [7]Vargas, O., Flor,O., Suarez, F., Design of a robotic prototype of the hand and right forearm for prostheses, Universidad, Ciencia y Tecnología, 2019.&#x0D; [8]O. Vargas, O. Flor, F. Suarez, C. Chimbo, Construction and functional tests of a robotic prototype for human prostheses, Revista espirales, 2020.&#x0D; [9]P. PonPriya, E. Priya, Design and control of prosthetic hand using myoelectric signal. International Conference on Computing and Communications Technologies (ICCCT), 2017, doi:10.1109/iccct2.2017.7972314.&#x0D; [10]N. Bajaj, A. Spiers, A. Dollar, State of the Art in Artificial Wrists: A Review of Prosthetic and Robotic Wrist Design. IEEE Transactions on Robotics, 2019, doi:10.1109/tro.2018.2865890.

Online Grasp Force Estimation From the Transient EMG.

Myoelectric upper limb prostheses are controlled using information from the electrical activity of residual muscles (i.e. the electromyogram, EMG). EMG patterns at the onset of a contraction (transient phase) have shown predictive information about upcoming grasps. However, decoding this information for the estimation of the grasp force was so far overlooked. In a previous offline study, we proved that the transient phase of the EMG indeed contains information about the grasp force and determined the best algorithm to extract this information. Here we translated those findings into an online platform to be tested with both non-amputees and amputees. The platform was tested during a pick and lift task (tri-digital grasp) with light objects (200 g - 1 kg), for which fine control of the grasp force is more important. Results show that, during this task, it is possible to estimate the target grasp force with an absolute error of 2.06 (1.32) % and 2.04 (0.49) % the maximum voluntary force for non-amputee and amputees, respectively, using information from the transient phase of the EMG. This approach would allow for a biomimetic regulation of the grasp force of a prosthetic hand. Indeed, the users could contract their muscles only once before the grasp begins with no need to modulate the grasp force for the whole duration of the grasp, as required with continuous classifiers. These results pave the way to fast, intuitive and robust myoelectric controllers of limb prostheses.

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.

Upper limb activity of twenty myoelectric prosthesis users and twenty healthy anatomically intact adults

The upper limb activity of twenty unilateral upper limb myoelectric prosthesis users and twenty anatomically intact adults were recorded over a 7-day period using two wrist worn accelerometers (Actigraph, LLC). This dataset reflects the real-world activities of the participants during their normal day-to-day routines. Participants included students, working adults, and retirees recruited from across the United Kingdom. This dataset offers a potential wealth of knowledge into a poorly understood cohort. The raw unprocessed data files and the activity count data exported from the Actilife software are provided. We also provide a non-wear algorithm developed for the removal of prosthesis non-wear periods and resulting activity count data corresponding to prothesis wear periods. Finally, we have included the transposed activity diaries provided by the participants. Analysis to date has primarily involved assessment of the symmetry of upper limb activity, however, there is potential to undertake additional analysis such as understanding the differences in the way a prosthesis is used compared to an anatomical arm.

Users' and therapists' perceptions of myoelectric multi-function upper limb prostheses with conventional and pattern recognition control.

ObjectiveTo describe users’ and therapists’ opinions on multi-function myoelectric upper limb prostheses with conventional control and pattern recognition control.DesignQualitative interview study.SettingsTwo rehabilitation institutions in the Netherlands and one in Austria.SubjectsThe study cohort consisted of 15 prosthesis users (13 males, mean age: 43.7 years, average experience with multi-function prosthesis: 3.15 years) and seven therapists (one male, mean age: 44.1 years, average experience with multi-function prostheses: 6.6 years). Four of these users and one therapist had experience with pattern recognition control.MethodThis study consisted of semi-structured interviews. The participants were interviewed at their rehabilitation centres or at home by telephone. The thematic framework approach was used for analysis.ResultsThe themes emerging from prosthesis users and therapists were largely congruent and resulted in one thematic framework with three main themes: control, prosthesis, and activities. The participants mostly addressed (dis-) satisfaction with the control type and the prosthesis itself and described the way they used their prostheses in daily tasks.ConclusionProsthesis users and therapists described multi-function upper limb prostheses as more functional devices than conventional one-degree-of-freedom prostheses. Nonetheless, the prostheses were seldom used to actively grasp and manipulate objects. Moreover, the participants clearly expressed their dissatisfaction with the mechanical robustness of the devices and with the process of switching prosthesis function under conventional control. Pattern recognition was appreciated as an intuitive control that facilitated fast switching between prosthesis functions, but was reported to be too unreliable for daily use and require extensive training.

Looking beyond proportional control: The relevance of mode switching in learning to operate multi-articulating myoelectric upper-limb prostheses

ObjectiveOperating a multi-articulating myoelectric prosthetic hand requires both proportional control of surface electromyography (sEMG) signals and accurately producing triggers to switch between grip types. Proportional control, used for hand opening/closing, has been studied extensively but the grip-switching control far less so. Can switching be trained and if so, how early on can improvement be seen? And is the ability to produce precise switches related to proficiency in proportional control? We examined individual learning differences, monitored improvements in switching control, and assessed whether skilful proportional control predicts proficient switching. MethodsTwenty-three unimpaired participants fitted with sEMG electrodes on the wrist flexors/extensors performed two 10-minute tasks: an adaptive virtual catching task (proportional control) and an sEMG- feedback-based switching task, with multiple performance measures being recorded and compared. ResultsMost outcome measures revealed large individual differences, with 47–87% of the participants showing improvement in switching and 14% deteriorating. As hypothesized, no correlations were found between mode-switching and proportional-control skills. ConclusionLearning to generate accurate grip-switching control proved difficult and might not be feasible for all individuals with upper-limb amputations. Unaffected participants showing proficient proportional control were not automatically good at producing well-tuned mode switches. SignificanceIn rehabilitation, clients with good proportional control are often offered a multi-articulating hand prosthesis. Since we found no clear associations between the proficiency in proportional movements and mode switches, this rationale may warrant reconsideration. Rather, when selecting the most optimal prosthesis, an individual’s skills in both control types should be examined separately.

An Insulated Flexible Sensor for Stable Electromyography Detection: Applicationto Prosthesis Control.

Electromyography (EMG), the measurement of electrical muscle activity, is used in a variety of applications, including myoelectric upper-limb prostheses, which help amputees to regain independence and a higher quality of life. The state-of-the-art sensors in prostheses have a conductive connection to the skin and are therefore sensitive to sweat and require preparation of the skin. They are applied with some pressure to ensure a conductive connection, which may result in pressure marks and can be problematic for patients with circulatory disorders, who constitute a major group of amputees. Due to their insulating layer between skin and sensor area, capacitive sensors are insensitive to the skin condition, they require neither conductive connection to the skin nor electrolytic paste or skin preparation. Here, we describe a highly stable, low-power capacitive EMG measurement set-up that is suitable for real-world application. Various flexible multi-layer sensor set-ups made of copper and insulating foils, flex print and textiles were compared. These flexible sensor set-ups adapt to the anatomy of the human forearm, therefore they provide high wearing comfort and ensure stability against motion artifacts. The influence of the materials used in the sensor set-up on the magnitude of the coupled signal was demonstrated based on both theoretical analysis and measurement.The amplifier circuit was optimized for high signal quality, low power consumption and mobile application. Different shielding and guarding concepts were compared, leading to high SNR.

Robust feature sets for contraction level invariant control of upper limb myoelectric prosthesis

In spite of the tremendous progress of upper limb myoelectric prosthetic control in the field of rehabilitation engineering, there still exist several real world challenges to be met, before realizing it as a good substitute for a natural arm. Incompetence of the system to accommodate variations in contraction levels of muscle movements has been identified as one of the significant challenges, as these variations have a subsequent impact on the performance of pattern recognition based myoelectric control. Non-linear techniques are more suited to characterize myoelectric signals since one of their major properties is nonlinearity. Based on this we propose two feature combinations which can lead to a reliable control scheme that is robust against contraction level variations. The performance of our proposed features when tested on nine transradial amputees for six motion classes at three different force levels outweighed other established feature extraction methods meant for contraction variation independent control. Significant improvement of around 8% in average classification performance was achieved across all subjects and force levels, subjected to training, both with all force levels and with unseen force levels. Moreover, these features achieved superior performance in classifying flexion as well as grip movements.

上肢切断者に対する能動義手，筋電義手の現状と課題 —当院6年間の義手処方と処方後調査より—

上肢切断者に対する能動義手，筋電義手の現状と課題 —当院6年間の義手処方と処方後調査より—

SoftHand at the CYBATHLON: a user\u2019s experience

BackgroundRoughly one-quarter of upper limb prosthesis users reject their prosthesis. Reasons for rejection range from comfort, to cost, aesthetics, function, and more. This paper follows a single user from training with and testing of a novel upper-limb myoelectric prosthesis (the SoftHand Pro) for participation in the CYBATHLON rehearsal to training for and competing in the CYBATHLON 2016 with a figure-of-nine harness controlled powered prosthesis (SoftHand Pro-H) to explore the feasibility and usability of a flexible anthropomorphic prosthetic hand.MethodsThe CYBATHLON pilot took part in multiple in-lab training sessions with the SoftHand Pro and SoftHand Pro-H; these sessions focused on basic control and use of the prosthetic devices and direct training of the tasks in the CYBATHLON. He used these devices in competition in the Powered Arm Prosthesis Race in the CYBATHLON rehearsal and 2016 events.ResultsIn training for the CYBATHLON rehearsal, the subject was able to quickly improve performance with the myoelectric SHP despite typically using a body-powered prosthetic hook. The subject improved further with additional training using the figure-of-nine harness-controlled SHPH in preparation for the CYBATHLON. The Pilot placed 3rd (out of 4) in the rehearsal. In the CYBATHLON, he placed 5th (out of 12) and was one of only two pilots who successfully completed all tasks in the competition, having the second-highest score overall.ConclusionsResults with the SoftHand Pro and Pro-H suggest it to be a viable alternative to existing anthropomorphic hands and show that the unique flexibility of the hand is easily learned and exploited.

Skill assessment in upper limb myoelectric prosthesis users: Validation of a clinically feasible method for characterising upper limb temporal and amplitude variability during the performance of functional tasks

Upper limb myoelectric prostheses remain challenging to use and are often abandoned. A proficient user must be able to plan/execute arm movements while activating the residual muscle(s), accounting for delays and unpredictability in prosthesis response. There is no validated, low cost measure of skill in performing such actions. Trial-trial variability of joint angle trajectories measured during functional task performance, linearly normalised by time, shows promise. However, linear normalisation of time introduces errors, and expensive camera systems are required for joint angle measurements.This study investigated whether trial-trial variability, assessed using dynamic time warping (DTW) of limb segment acceleration measured during functional task performance, is a valid measure of user skill. Temporal and amplitude variability of forearm accelerations were determined in (1) seven myoelectric prosthesis users and six anatomically-intact controls and (2) seven anatomically-intact subjects learning to use a prosthesis simulator over repeated sessions.(1): temporal variability showed clear group differences (p<0.05). (2): temporal variability considerably increased on first use of a prosthesis simulator, then declined with training (both p<0.05). Amplitude variability showed less obvious differences. Analysing forearm accelerations using DTW appears to be a valid low-cost method for quantifying movement quality of upper limb prosthesis use during goal-oriented task performance.

Visualisation of upper limb activity using spirals: A new approach to the assessment of daily prosthesis usage.

Background:Current outcome measures used in upper limb myoelectric prosthesis studies include clinical tests of function and self-report questionnaires on real-world prosthesis use. Research in other cohorts has questioned both the validity of self-report as an activity assessment tool and the relationship between clinical functionality and real-world upper limb activity. Previously,1 we reported the first results of monitoring upper limb prosthesis use. However, the data visualisation technique used was limited in scope.Study Design:Methodology development.Objectives:To introduce two new methods for the analysis and display of upper limb activity monitoring data and to demonstrate the potential value of the approach with example real-world data.Methods:Upper limb activity monitors, worn on each wrist, recorded data on two anatomically intact participants and two prosthesis users over 1 week. Participants also filled in a diary to record upper limb activity. Data visualisation was carried out using histograms, and Archimedean spirals to illustrate temporal patterns of upper limb activity.Results:Anatomically intact participants’ activity was largely bilateral in nature, interspersed with frequent bursts of unilateral activity of each arm. At times when the prosthesis was worn prosthesis users showed very little unilateral use of the prosthesis (≈20–40 min/week compared to ≈350 min/week unilateral activity on each arm for anatomically intact participants), with consistent bias towards the intact arm throughout. The Archimedean spiral plots illustrated participant-specific patterns of non-use in prosthesis users.Conclusion:The data visualisation techniques allow detailed and objective assessment of temporal patterns in the upper limb activity of prosthesis users.Clinical relevanceActivity monitoring offers an objective method for the assessment of upper limb prosthesis users’ (PUs) activity outside of the clinic. By plotting data using Archimedean spirals, it is possible to visualise, in detail, the temporal patterns of upper limb activity. Further work is needed to explore the relationship between traditional functional outcome measures and real-world prosthesis activity.

Targeted Muscle Reinnervation for the Upper and Lower Extremity.

Myoelectric devices are controlled by electromyographic signals generated by contraction of residual muscles, which thus serve as biological amplifiers of neural control signals. Although nerves severed by amputation continue to carry motor control information intended for the missing limb, loss of muscle effectors due to amputation prevents access to this important control information. Targeted Muscle Reinnervation (TMR) was developed as a novel strategy to improve control of myoelectric upper limb prostheses. Severed motor nerves are surgically transferred to the motor points of denervated target muscles, which, after reinnervation, contract in response to neural control signals for the missing limb. TMR creates additional control sites, eliminating the need to switch the prosthesis between different control modes. In addition, contraction of target muscles, and operation of the prosthesis, occurs in reponse to attempts to move the missing limb, making control easier and more intuitive. TMR has been performed extensively in individuals with high-level upper limb amputations and has been shown to improve functional prosthesis control. The benefits of TMR are being studied in individuals with transradial amputations and lower limb amputations. TMR is also being investigated in an ongoing clinical trial as a method to prevent or treat painful amputation neuromas.

Resolving the effect of wrist position on myoelectric pattern recognition control

BackgroundThe use of pattern recognition-based methods to control myoelectric upper-limb prostheses has been well studied in individuals with high-level amputations but few studies have demonstrated that it is suitable for partial-hand amputees, who often possess a functional wrist. This study’s objective was to evaluate strategies that allow partial-hand amputees to control a prosthetic hand while allowing retain wrist function.MethodsEMG data was recorded from the extrinsic and intrinsic hand muscles of six non-amputees and two partial-hand amputees while they performed 4 hand motions in 13 different wrist positions. The performance of 4 classification schemes using EMG data alone and EMG data combined with wrist positional information was evaluated. Using recorded wrist positional data, the relationship between EMG features and wrist position was modeled and used to develop a wrist position-independent classification scheme.ResultsA multi-layer perceptron artificial neural network classifier was better able to discriminate four hand motion classes in 13 wrist positions than a linear discriminant analysis classifier (p = 0.006), quadratic discriminant analysis classifier (p < 0.0001) and a linear perceptron artificial neural network classifier (p = 0.04). The addition of wrist position data to EMG data significantly improved performance (p < 0.001). Training the classifier with the combination of extrinsic and intrinsic muscle EMG data performed significantly better than using intrinsic (p < 0.0001) or extrinsic muscle EMG data alone (p < 0.0001), and training with intrinsic muscle EMG data performed significantly better than extrinsic muscle EMG data alone (p < 0.001). The same trends were observed for amputees, except training with intrinsic muscle EMG data, on average, performed worse than the extrinsic muscle EMG data. We propose a wrist position–independent controller that simulates data from multiple wrist positions and is able to significantly improve performance by 48–74% (p < 0.05) for non-amputees and by 45–66% for partial-hand amputees, compared to a classifier trained only with data from a neutral wrist position and tested with data from multiple positions.ConclusionsSensor fusion (using EMG and wrist position information), non-linear artificial neural networks, combining EMG data across multiple muscle sources, and simulating data from different wrist positions are effective strategies for mitigating the wrist position effect and improving classification performance.