Search

One of the most important factors in successful upper limb prostheses is the socket design. Sockets must be individually fabricated to arrive at a geometry that suits the user's morphology and appropriately distributes the pressures associated with prosthetic use across the residual limb. In higher levels of amputation, such as transhumeral, this challenge is amplified as prosthetic weight and the physical demands placed on the residual limb are heightened. Yet, in the upper limb, socket fabrication is largely driven by heuristic practices. An analytical understanding of the interactions between the socket and residual limb is absent in literature. This work describes techniques, adapted from lower limb prosthetic research, to empirically characterize the pressure distribution occurring between the residual limb and well-fit transhumeral prosthetic sockets. A case series analyzing the result of four participants with transhumeral amputation is presented. A Tekscan VersaTek pressure measurement system and FaroArm Edge coordinate measurement machine were employed to capture socket-residual limb interface pressures and geometrically register these values to the anatomy of participants. Participants performed two static poses with their prosthesis under two separate loading conditions. Surface pressure maps were constructed from the data, highlighting pressure distribution patterns, anatomical locations bearing maximum pressure, and the relative pressure magnitudes. Pressure distribution patterns demonstrated unique characteristics across the four participants that could be traced to individual socket design considerations. This work presents a technique that implements commercially available tools to quantitatively characterize upper limb socket-residual limb interactions. This is a fundamental first step toward improved socket designs developed through informed, analytically-based design tools.

Position-aware myoelectric prosthesis controllers require long, data-intensive training routines. Transfer Learning (TL) might reduce training burden. A TL model can be pre-trained using forearm muscle signal data from many individuals to become the starting point for a new user. A recurrent convolutional neural network (RCNN)-based classifier has already been shown to benefit from TL in offline analysis (95% accuracy). The present real-time study tested whether an RCNN-based classification controller with TL (RCNN-TL) could reduce training burden, offer improved device control (per functional task performance metrics), and mitigate what is known as the "limb position effect". 27 participants without amputation were recruited. 19 participants performed wrist/hand movements across multiple limb positions, with resulting forearm muscle signal data used to pre-train RCNN-TL. 8 other participants donned a simulated prosthesis, retrained (calibrated) and tested RCNN-TL, plus trained and tested a conventional linear discriminant analysis classification controller (LDA-Baseline). Results confirmed that TL reduces user training burden. RCNN-TL yielded improved task performance durations over LDA-Baseline (in specific Grasp and Release phases), yet other metrics worsened. Overall, this work contributes training condition factors necessary for TL success, identifies metrics needed for comprehensive control analysis, and contributes insights towards improved position-aware control.

To mitigate the "limb position effect" that hinders myoelectric upper limb prosthesis control, pattern recognition-based models must accurately predict user-intended movements across a multitude of limb positions. Such models can use electromyography (EMG) and inertial measurement units to capture necessary multi-position data. However, this data capture solution requires lengthy user-performed model training routines, with movements in many limb positions, plus retraining thereafter due to inherent signal variations over time. While a general-purpose control model (trained with a dataset that represents numerous device users) eliminates the user-training requirement altogether, it yields low movement predictive accuracy. Conversely, a user-specific control model (trained with a smaller dataset from an individual) yields high predictive accuracy, but requires retraining over time. This study capitalizes on the benefits offered by both such control options, and contributes an alternative control solution-a novel recurrent convolutional neural network (RCNN)-based Composite Model that combines the representation portion of a general-purpose model, with the decision portion of a user-specific model. The resulting Composite Model offers moderate movement predictive accuracy across various limb positions and a reduction in user training routine requirements, suggesting a new research direction to help mitigate the limb position effect along with model training burden.

Competing Interests: Competing interest statement: C.L.S. is employed through Torreyana Corp, San Diego, CA; Blackhawk Genomics, Concord, CA; and Advagenix, Rockville, MD. Volunteer activities of C.L.S. include those for College of American Pathologists, Northfield, IL, and Clinical Laboratory Standards Institute, Wayne, PA. T.J.O. is a Member, Scientific Advisory Committee, MioDx and Integrated Nano-Technologies. A.W.I.L.’s laboratory receives sponsorships from AstraZeneca (HK) Ltd. and MSD (HK) Ltd. for providing selected companion diagnostic tests free to public patients. N.C. is an employee at Quest Diagnostics, Inc. F.A.M. is an employee and stock option holder at Castle Biosciences, Inc. A.B.C. is paid teaching faculty for the American Medical Informatics Association Clinical Informatics Board Review Course and receives small honoraria as well as travel reimbursement to speak at multiple scientific and professional medical society meetings. T.L.P. has been consulted (compensated) for Bio-Rad, Inc. and is Director of Pathology Strategies for the Sturge Weber Foundation (compensated). H.E.W. has employment through Viapath, a majority National Health Service-owned independent pathology service provider; has been a paid faculty member at Kingston University, accepted paid accommodation and subsistence as an invited speaker to Cytocell User Group meeting for the United Kingdom and Ireland, and accepted paid event registration as an invited speaker to Digital Pathology/Global Engage meeting. H.E.W.’s laboratory received scholarship funds from The International Council for Standardization in Haematology for providing JAK2 testing. The other authors (K.F., E.A.B., P.M., N.R.P., K.P.P., B.S., M.S., M.L.G., S.M.M., H.F., and S.O.) do not have any affiliations, memberships, funding, or financial holdings that might be perceived as affecting the objectivity of this article.

Studies that investigate myoelectric prosthesis control commonly use non-disabled participants fitted with a simulated prosthetic device. This approach improves participant recruitment numbers but assumes that simulated movements represent those of actual prosthesis users. If this assumption is valid, then movement performance differences between simulated prosthesis users and normative populations should be similar to differences between actual prosthesis users and normative populations. As a first step in testing this assumption, the objective of this study was to quantify movement performance differences between simulated transradial myoelectric prosthesis hand function and normative hand function. Motion capture technology was used to obtain hand kinematics for 12 non-disabled simulated prosthesis participants who performed a functional object-manipulation task. Performance metrics, end effector movement, and grip aperture results were compared to 20 nondisabled participants who used their own hand during task execution. Simulated prosthesis users were expected to perform the functional task more slowly, with multiple peaks in end effector velocity profiles, and a plateau in grip aperture when reaching to pick up objects, when compared to non-disabled participants. This study confirmed these expectations and recommends that subsequent research be undertaken to quantify differences in actual myoelectric prosthesis hand function versus normative hand function.