Corrigendum to “Connecting the wrist to the hand: A simulation study exploring changes in thumb-tip endpoint force following wrist surgery” [J. Biomech. 58 (2017) 97–104] (Journal of Biomechanics (2017) 58 (97–104), (S0021929017302257), (10.1016/j.jbiomech.2017.04.024))

Daniel C. McFarland, Jennifer A. Nichols, Michael S. Bednar, Sarah J. Wohlman, Wendy M. Murray*

*Corresponding author for this work

Research output: Contribution to journalComment/debatepeer-review


The published paper (Nichols et al., 2017) used three musculoskeletal models of the wrist and thumb: a model of the nonimpaired system, a model of proximal row carpectomy (PRC), and a model of scaphoid excision four corner fusion (SE4CF). All three of these models incorporated the intrinsic thumb muscle paths from Wohlman and Murray (2013). As described in a parallel addendum (McFarland et al., submitted for publication), while completing a thorough verification process of the biomechanical models in order to release the thumb model to the biomechanics community for use in OpenSim (Delp et al., 2007; Seth et al., 2018), we discovered a computational error in the original MATLAB code that required correction and resulted in the need to re-optimize the intrinsic thumb muscle paths. We have described the error and the effects of updating the intrinsic thumb muscle paths in the biomechanical model on all simulations originally reported in Wohlman and Murray (2013) in detail elsewhere McFarland et al. (submitted for publication). For completeness, we have also re-simulated the work presented in Nichols et al. (2017). Here, we detail the results of these re-simulations. In this corrigendum, we repeated the simulations described in Section 2.3 (“Simulations of muscle control strategy”) and Section 2.4 (“Simulations of muscle force transmission”) with the updated intrinsic thumb muscle paths. All simulations were performed in OpenSim v3.3, using the methods described in the original publication. The simulations described in Section 2.2 (“Simulations of joint torques”) were not repeated because these simulations do not include muscles and are therefore not sensitive to the new muscle paths. In these simulations, computed muscle control is used to calculate the muscle activations that are required to produce a target pinch force while simultaneously maintaining a specified lateral pinch and wrist posture. Seven simulations, with target pinch forces that varied from 0 to 60 N were performed, where target pinch force was varied in 10 N increments. Muscle activations were computed for the primary wrist, extrinsic thumb, and intrinsic thumb muscles. As described and implemented in the published paper, additional torque generators were included at the wrist to account for wrist flexion–extension and radial-ulnar deviation torques produced by the extrinsic muscles that actuate the four fingers, which were not included in these models. In these simulations, forward-dynamic simulations of lateral pinch force were performed with each of the three models (nonimpaired, PRC, and SE4CF). Identical muscle activations served as inputs to each model. Because muscle activations were held constant across the three models, any differences in the simulated pinch force vectors arise from changes to wrist joint kinematics and musculoskeletal geometry imposed by the surgeries. The muscle activation inputs were determined using computed muscle control and the nonimpaired model, without any reserve actuators at the wrist, so that the lateral pinch forces produced only required inputs for the 14 muscles in each model. The resulting lateral pinch force and equilibrium posture were analyzed. Overall, with the new intrinsic muscle paths, the results for the muscle control strategy (Section 2.3) and the muscle force transmission (Section 2.4) simulations support the conclusions of the original publication. In particular, the simulations that incorporate the new intrinsic thumb muscle paths demonstrated that different control strategies are needed for each of the three biomechanical models to generate a well-directed pinch force while maintaining wrist posture. As with the original publication, the computed activation patterns differed for the wrist, intrinsic thumb, extrinsic thumb and extrinsic finger (represented by reserve torques) muscles across the three models. The primary effect of the new intrinsic thumb muscle paths is that all models are approximately 10 N weaker compared to the original publication. As described in the original publication, the strength limit for each model was defined by the largest magnitude of target pinch force for which the computed muscle control algorithm could successfully solve for activation patterns. Unlike in the original publication, the updates to the muscle paths resulted in the SE4CF model producing equivalent lateral pinch forces to the nonimpaired model during the forward dynamic simulations when actuated with the nonimpaired control strategy. However, this difference did not change the conclusion that the SE4CF was unable to maintain the initial wrist posture during force production. Below we provide detailed results of the simulations with the new intrinsic muscle paths (Sections 2.3 and 2.4). The computed muscle control algorithm succeeded at solving for muscle activation patterns that produced target lateral pinch forces up to a maximum of 50 N of lateral pinch force for the nonimpaired and SE4CF models (i.e., the simulations we ran for the 60 N target force failed for these two models). For the PRC model, successful simulations were limited to a maximum of 40 N. In each case (nonimpaired, SE4CF, and PRC) the maximum target force was smaller than in the original publication by a 10 N increment (recall that simulations were repeated in 10 N increments only). As in the original results, the models representing the different wrist conditions required different control strategies to produce equivalent, well-directed thumb-tip endpoint forces while maintaining the specified posture (Fig. 1). Differences were evident for each group of muscles (primary wrist, extrinsic thumb, and intrinsic thumb). Additionally, activations of the reserve torque generators, which represent the wrist torques generated by the extrinsic finger muscles, suggest that different control strategies are required for these muscles following these surgical salvage procedures. For example, the PRC model requires more total reserve wrist torque than the other wrist conditions (Fig. 2). Overall, the reserve torque magnitudes are similar to the original publication. However, the SE4CF model that incorporates the new intrinsic thumb muscle paths now requires reserve torques in the distal direction, whereas in the original publication, the SE4CF model required reserve torque in the proximal direction. This difference from the original results occurred even though the only changes implemented in the models were the paths of the intrinsic muscles, which do not cross the wrist. In general, due to the new intrinsic muscle paths, different activation patterns were selected to produce the same well-directed lateral pinch forces. Ultimately, the intrinsic activation patterns affected the required activations for the primary wrist and extrinsic thumb muscles, changing the demand at the wrist. In forward simulations with identical muscle activation inputs, the PRC model produced thumb-tip endpoint forces that are approximately 10% weaker than the nonimpaired model (Table 1 ). The resultant force of the PRC model is misdirected in the palmar and proximal direction (Fig. 3); these results are consistent with the original findings. In contrast, the SE4CF model with the new intrinsic muscle paths produces endpoint forces that match the desired pinch force (Table 1) and are well-directed (Fig. 3). In the original publication, the SE4CF model produced endpoint forces approximately 8% weaker than the nonimpaired model and more misdirected. As in the original results, each model produced a different equilibrium posture. The equilibrium posture for the nonimpaired model was closest to the initial posture (Table 2). In contrast, for both the SE4CF and PRC models, the wrist moved by nearly 10˚ from neutral in the radial-extension direction. Thus, in the re-simulations, the nonimpaired model is the only model that can successfully maintain the desired posture while producing well-directed endpoint forces using the same activation pattern. This result supports our original conclusion that following these surgical salvage procedures, adapting muscle coordination patterns may be required to best improve post-operative hand function. Using “simulation-only” methods, the original publication demonstrated how drastic surgical changes to the musculoskeletal design of the wrist influenced hand function. In particular, the original publication showed that surgically altering wrist kinematics and wrist muscle moment arms substantially influences muscle control strategies and muscle force transmission associated with the production of lateral pinch force. In this corrigendum, we have verified these major findings of the original publication, despite implementation of new intrinsic thumb muscle paths. The results with the new muscle paths continue to support the original conclusions that the wrist's torque-generating capacity and adaptability of muscle coordination patterns are key areas to focus on to improve post-operative hand function. The nonimpaired, PRC, and SE4CF models that can replicate these lateral pinch simulations are now available for download on SimTK for use in OpenSim. Daniel C. McFarland: Methodology, Software, Validation, Formal analysis, Writing – original draft. Jennifer A. Nichols: Conceptualization, Methodology, Software, Formal analysis, Writing – reviewing and editing. Michael S. Bednar: Conceptualization, Supervision, Writing – reviewing and editing. Sarah J. Wohlman: Methodology, Software, Writing – reviewing and editing. Wendy M. Murray: Conceptualization, Methodology, Supervision, Writing – reviewing and editing.

Original languageEnglish (US)
Article number110859
JournalJournal of Biomechanics
StateAccepted/In press - 2021


  • Computer simulation
  • Proximal row carpectomy
  • Scaphoid-excision four
  • Thumb
  • Wrist

ASJC Scopus subject areas

  • Biophysics
  • Biomedical Engineering
  • Orthopedics and Sports Medicine
  • Rehabilitation


Dive into the research topics of 'Corrigendum to “Connecting the wrist to the hand: A simulation study exploring changes in thumb-tip endpoint force following wrist surgery” [J. Biomech. 58 (2017) 97–104] (Journal of Biomechanics (2017) 58 (97–104), (S0021929017302257), (10.1016/j.jbiomech.2017.04.024))'. Together they form a unique fingerprint.

Cite this