what kind of mechanical model can we use to represent the elbow joint?
Bioengineering (Basel). 2022 Jun; 5(ii): 31.
Musculoskeletal Model Development of the Elbow Joint with an Experimental Evaluation
Akin Cil
1Department of Civil and Mechanical Engineering, Academy of Missouri-Kansas City, 5110 Rockhill Road, Kansas City, MO 64110, United states of america; ude.ud@inaneRifirahS.neshoM (M.S.R.); gro.demcmt@liC.nikA (A.C.); ude.ckmu@auonailyts (A.P.S.)
iiSection of Orthopaedic Surgery, University of Missouri-Kansas Urban center, 2411 Holmes Street, Kansas Urban center, MO 64108, USA
3Department of Orthopaedics, Truman Medical Centers, 2301 Holmes Street, Kansas City, MO 64108, USA
Received 2022 Mar 30; Accepted 2022 April 18.
Abstract
A dynamic musculoskeletal model of the elbow joint in which muscle, ligament, and articular surface contact forces are predicted concurrently would exist an ideal tool for patient-specific preoperative planning, reckoner-aided surgery, and rehabilitation. Existing musculoskeletal elbow joint models have limited clinical applicability because of idealizing the elbow as a mechanical hinge joint or ignoring important soft tissue (e.g., cartilage) contributions. The purpose of this study was to develop a subject-specific anatomically correct musculoskeletal elbow joint model and evaluate it based on experimental kinematics and muscle electromyography measurements. The model included three-dimensional bone geometries, a joint constrained by multiple ligament bundles, deformable contacts, and the natural oblique wrapping of ligaments. The musculoskeletal model predicted the bone kinematics reasonably accurately in three different velocity atmospheric condition. The model predicted timing and number of muscle excitations, and the normalized muscle forces were as well in agreement with the experiment. The model was able to predict of import in vivo parameters that are not possible to measure experimentally, such equally musculus and ligament forces, and cartilage contact pressure. In improver, the developed musculoskeletal model was computationally efficient for body-level dynamic simulation. The maximum ciphering fourth dimension was less than thirty min for our 35 southward simulation. As a predictive clinical tool, the potential medical applications for this model and modeling arroyo are significant.
Keywords: musculoskeletal model, elbow joint, cartilage, ligaments, contact mechanics, kinematics, upper extremity
ane. Introduction
Equally an important joint of the upper extremity, the elbow joint serves as a fulcrum of the forearm lever that profoundly enhances the spatial positioning of the hand. Because of the axis of the elbow joint to the upper extremity system, loss of or diminished function of the elbow joint results in significant deficits in upper extremity function, and can jeopardize individual independence [1]. Therefore, biomechanical analysis of the elbow joint is extremely important in understanding elbow injuries, better execution of trauma direction, and prosthetic pattern.
Detailed knowledge of the in vivo loading of elbow structures is essential in understanding the biomechanical causes associated with both chronic (degenerative joint disease) and astute injuries, and for improving the design and implementation of therapeutic interventions. Since the direct measurement of the in vivo joint loads is not technologically viable, computational models need be implemented for predictions. A dynamic computational elbow joint model capable of concurrent prediction of muscle and ligament forces along with the cartilage contact mechanics would be an ideal tool in clinical practice. Once the model is validated, multiple disease states can be tested to better our clinical understanding of the elbow pathology, test the current elbow implants, and design a amend implant for partial or total elbow replacements. In addition, the model could provide subject-specific intervention strategies aimed at modifying upper extremity movement for targeted outcomes, such as reducing articular cartilage stress. Computational models tin can also mitigate the need for large sample sizes in clinical trials and can work at a fraction of the cost of cadaveric models.
Computational models of the elbow accept been employed to study the joint biomechanical behaviour and analyse musculoskeletal motion [2,3,4,v]. However, these models have express clinical applicability past bold a fixed joint axis of rotation (e.1000., hinge joint) rather than a true anatomical joint constrained by ligament forces and cartilage contacts. In some circumstances, such simplifications would exist helpful; however, the man elbow joint has significant laxity that should not be ignored [6,7]. In a recent study, Fisk and Wayne [viii] adult an anatomical model of the elbow joint where the articulation behaviour was dictated by ligament constraint and articular contact. Yet, in that location are nevertheless assumptions that have been made in this advanced model that cannot fully replicate the normal elbow physiology/clinical scenarios. The significant limitations of this model are that the model did not include articular cartilage, the ligaments were modelled as linear springs, the muscles were represented as abiding-magnitude forcefulness vectors, and the wrapping of ligaments and muscles effectually the bony structures was ignored. The need for dynamic anatomically correct computational elbow joint models that link muscle forces, movement, and articulation contact characteristics has long been recognized. To our knowledge, such an advanced musculoskeletal elbow articulation model does not exist in the literature.
The objective of this report was to develop an anatomically correct musculoskeletal elbow joint model and evaluate information technology based on experimental kinematics and muscle electromyography (EMG) measurement. The joint was constrained past multiple nonlinear ligament bundles and 3-dimensional deformable contacts. Moreover, the model included the natural oblique wrapping of muscle and ligaments around the joints. The developed elbow model is a step toward developing a complete musculoskeletal model of the upper extremity.
2. Materials and Methods
2.i. Experimental Measures
I healthy volunteer (24 years old, male) with no history of upper extremity problems was recruited for the project after providing written informed consent approved by the institutional human subject field review board. Ii localizers made of ABS plastic were fastened to the subject field'due south upper arm and the forearm; the localizers included two orthogonal tubes (Effigy anea). The tubes were filled with fluid (mustard) that appeared in the magnetic resonance images (MRIs). The localizers, forth with the tubes, were used to register the coordinate system of the bone, cartilage, ligament, and muscle geometries later in the modelling procedure. Loftier-resolution MRI (Siemens 1.5T automobile, TR: 14.viii, TE: half-dozen.xviii, slice thickness 0.five mm, imaging frequency 63.63 Hz, and group lengths 178) was acquired on the subject's dominant upper extremity. The localizers were not removed every bit the subject travelled to the UMKC Man Motility Lab. Two Optotrack move capture rigid bodies (Northern Digital Inc, Waterloo, Ontario, Canada) were attached to the subject's upper arm and two rigid bodies were attached to the subject's lower arm to collect the motility data. Each rigid trunk contains 3 infrared markers to capture 6 degrees of liberty of trunk segment (3 translations and 3 rotations).
Experimental measurement: (a) Two plastic localizers fastened to subject upper and lower arm; (b) Initial position drove using Optotrak probe; (c) The experimenter manipulating the arm for laxity examination; (d) The subject performing the experimental trial.
The subject field was positioned on a Biodex Multi-Joint Dynamometer organisation (Biodex Medical Systems, Shirley, NY, USA) where the ascendant arm was kept in a pronated residue position (Figure 1b). The initial position and orientation of the arm were determined past recording the coordinates of multiple points on the localizers, on the bony landmarks of the arm, and forth the arm surface by using an Optotrak probe tool. Following collection of the initial position, the localizers were removed, and an Orthopaedic surgeon performed a standard laxity test for the elbow (Figure 1c). This exam involved moving the elbow joint through its full range of motion (flexion–extension and varus–valgus) by applying minimal strength to the joint while motion data from the body segments were nerveless. This established the kinematic range of motion data from which the ligament nothing-load lengths (the lengths at which the ligament kickoff become taut) were extracted. The dynamic testing of the elbow was performed in the dynamometer (Figure aned). Along with the Optotrack markers, the subject was also outfitted with 2 surface EMG sensors (Delsys Myomonitor 4 wireless EMG system, Delsys Inc., Natick, MA, The states) on the middle of the muscle as suggested for bicep and tricep muscles to tape muscle action [9,10]. The subject was asked to perform three seated elbow flexion/extension trials in 3 different protocols: (1) isokinetic at 10 deg/s, (2) isokinetic at 60 deg/s, and (three) at self-selected angular velocity (as fast as possible by the bailiwick). For each trial, arm segment motions (at 100 Hz) and EMG activity (at 1400 Hz) were recorded.
2.2. Computational Model
Three-dimensional bone and cartilage geometries were generated from the MRI using 3D Slicer (world wide web.slicer.org). The geometries were imported to MeshLab for postprocessing that included removing the spikes, reducing the noise, and smoothing the surface [11]. The multibody model was created in ADAMS (MSC Software Corporation, Santa Ana, CA, USA) past aligning these geometries using the initial position and orientation obtained during the experimental study (Effigy 2a). The bone and the cartilage densities were ready at 1600 kg/g3 [12] and grand kg/chiliadthree [thirteen], respectively.
Multibody Model: (a) A subject-specific model representing bones and major muscles crossing the elbow articulation, (b) medial collateral ligament (MCL) complex, and (c) lateral collateral ligament complex in the model.
The ligaments and the interosseous membranes were modelled every bit a different number of bundles based on their structure and function. The model included three bundles for the medial collateral ligament (MCL) anterior function [xiv], three bundles for the MCL posterior office, three bundles for the lateral ulnar collateral ligament (LUCL) [15], three bundles for the radial collateral ligament (RCL) [8], and ii bundles for the annular ligament (Figure 2b,c). The ligaments were attached to the bone, according to the attachment sites identified in the MRI and in published studies [i,8,xvi,17]. Each ligament bundle was modelled as tension-merely nonlinear springs using a piecewise role that includes the "toe" region [eighteen,nineteen]. The toe region simulates the crimping result of the ligament which represents the parabolic transition between the zero strain and the linear region of the ligament.
The force-length relationship for each ligament bundle is described past Equations (1) and (2):
(i)
where k is the stiffness parameter, ε l is a spring parameter assumed to be 0.03 [xx], l is the length of the each packet, and l 0 is the cypher-load length [19,21]. The ligament stiffness parameter was obtained from the literature [8,22], and the zero-load length was calculated based on the laxity test. The zero-load length for each ligament bundle was determined past the maximum distance measured between ligament insertion and origin sites throughout the motion range during the laxity exam and then multiplied by a correction gene (eighty% of max length) [23]. This correction factor reduced the error unintentionally introduced by the experimenter when applying a minor amount of ligament strength. The ligaments were wrapped around the os to represent their anatomical physiology, and to prevent the penetration of the ligament into the os [24].
A custom macro was written in ADAMS to automatically split the humerus cartilage into discrete hexahedral elements. The macro performance was successfully tested in many of our previous lab studies [15,24,25,26,27,28,29]. Each cartilage element had an approximate three × 3 mm cross-sectional surface area. The macro also divers a deformable contact constraint with no friction using an ADAMS compliant contact model (Equation (3)) betwixt each humerus cartilage element with the radius and ulna cartilage geometry:
where 1000c is the contact stiffness, δ is the interpenetration of the geometries, and Bc(δ) is a damping coefficient. Optimization and pattern of the experiment approach were used to determine the contact parameters and the size of discretized cartilage elements from a cadaver study (Table 1). The optimization was washed in such a way that the maximum contact pressure and contact area errors were minimized between a multibody model and an identically loaded finite element model [27,30].
Table ane
Optimized contact parameter data.
| Parameters | Values |
|---|---|
| Contact blazon | Impact (deformable) |
| Contact chemical element size | 3 mm × 3 mm |
| Friction | No |
| Stiffness (gc ) | xl N/mm |
| Damping coefficient (Bc(δ)) | v Ns/mm |
| Exponent (north) | three.05 |
| Interpenetration of the geometries (δ) | 0.1 mm |
Since the model evaluation did not require the complete set of muscle, the model included three major muscles that cross the elbow joint: triceps (long, lateral, and medial), biceps (long, short), and brachialis. Muscle modelling mechanical parameters, insertions, origins, and via-points are based on published literature [four]. The musculoskeletal model simulation was executed in two phases. First, during the inverse kinematics, the experimental motion information were used to move the model as constrained by the joint contacts and ligaments. The shortening/lengthening pattern of each muscle element was recorded during this step. Side by side, the kinematic constraints were removed, and the muscles served as actuators during forward dynamics. The muscle forces were calculated via proportional–integral–derivative (PID) controllers implemented in Simulink (The MathWorks, Inc., Natick, MA, USA). During the forrad dynamics simulation, ADAMS and Simulink were linked in co-simulation. In the process, ADAMS sent the electric current muscle lengths to Simulink and Simulink then sent the musculus forces to ADAMS for the next step of calculation (Figure 3) [25]. The error betoken between the current forwards dynamics muscle length and the muscle length measured during the inverse kinematics simulation was minimized by the PID controllers. The output of the PID controller was the musculus forces for the forward dynamic simulation to track the inverse kinematics muscle length.
Feedback command scheme for calculating muscle strength.
The muscle strength was limited in a way that it tin only pull, not push. In improver, the PID parameters for each individual muscle were scaled based on the following equation:
(4)
where Pi is the proportional gain for musculus i. The PCSAi term is the physiological cantankerous-sectional expanse of each muscle and originates from the work by Holzbaur et al. [4]. The reference PCSA = 487 mm2 was calculated every bit the average of all muscles. Like equations were too applied for the integral and derivative gains. The global PID values for the muscle controller were P = 50, I = 5, and D = 0.0005 [25]. Muscles with a PCSA less than the reference PCSA will have smaller PID gains while larger muscles will take larger PID gains. Furthermore, the force generated by an individual muscle was express by its maximum forcefulness generating capacity.
Local coordinate systems for each bone segment were created to mensurate the ulna and radius motions relative to the humerus [27]. The translations were represented as medial–lateral (One thousand–L), anterior–posterior (A–P), and superior–junior (Southward–I) directions and the rotations were represented as flexion–extension (F–E), varus–valgus (VR–VL), and internal–external (I–E) rotation. The model was evaluated past comparing the model-predicted kinematics (forward dynamics) to the experimental measurements (inverse kinematics). The predicted musculus activation patterns were besides compared to the experimental EMG measurements. The experimental EMG signals were demeaned, rectified, so low-laissez passer filtered to eliminate measurement noise using a 2nd-order Butterworth low-pass filter with a cut-off frequency of six Hz. Then, the filtered EMG bespeak was normalized to the maximum value of the specific trial for each muscle. Nosotros calculated the root mean foursquare (RMS) fault and correlation coefficient to compare the model-predicted kinematics with experimental results. After model evaluation, ligament loads, articulation contact locations, contact area, and pressures were predicted from the musculoskeletal model simulation.
iii. Results
three.1. Model Evaluation
Kinematic comparisons between experimental results and model predictions are presented in each anatomical direction for the ulna and radius relative to the humerus (Figure four, Figure 5 and Figure vi). Overall, the model compared well with the experiment for all three velocity conditions of x deg/south, 60 deg/south, and free speed. The forearm bones rotated three° to vii° more internally in the model prediction than in the experiment (Tabular array 2). The model likewise predicted ii°–iii° greater valgus rotation and 2–iv mm lower lateral translation compared with the experimental measurement. All other model-predicted kinematics followed the experiment very well. Overall, we observed college RMS errors at lower speeds of forearm movement. On average, 8 out of 12 kinematics had a good correlation (greater than eighty%) between model and experiment.
Comparison of bone kinematics between experiment and model prediction for 10 deg/south trial. Increasing trend of the graph indicates more internal rotation for I–E, more than valgus rotation for VR–VL, and more extension for F–Eastward. Similarly, information technology indicates more superior translation for S–I, more inductive translation for A–P, and more medial translation for M–L. Decreasing tendency of the graph indicates the opposite.
Comparison of os kinematics between experiment and model prediction for 60 deg/s trial. Increasing trend of the graph indicates more internal rotation for I–Due east, more valgus rotation for VR–VL, and more extension for F–E. Similarly, information technology indicates more superior translation for Southward–I, more anterior translation for A–P, and more medial translation for Grand–L. Decreasing trend of the graph indicates the opposite.
Relative bone kinematics for experiment and model prediction for the free velocity trial. Increasing trend of the graph indicates more internal rotation for I–E, more valgus rotation for VR–VL, and more than extension for F–E. Similarly, it indicates more superior translation for S–I, more than anterior translation for A–P, and more medial translation for M–L. Decreasing trend of the graph indicates the reverse.
Table 2
RMS error (deg, mm) and correlation coefficients for ulna and radius kinematics (skilful correlations are in bold).
| Kinematics Description | ten deg/s | 60 deg/s | Costless Velocity | |||
|---|---|---|---|---|---|---|
| RMS Error | Correlation Coefficient | RMS Error | Correlation Coefficient | RMS Error | Correlation Coefficient | |
| Ulna I–E rotation | seven.4 | 0.48 | two.9 | 0.86 | 4.0 | 0.92 |
| Ulna VR–VL rotation | 2.eight | −0.63 | 2.ix | −0.13 | 2.seven | 0.57 |
| Ulna F–Due east rotation | 1.3 | 0.99 | ane.2 | 0.99 | ii.seven | 0.99 |
| Radius I–East rotation | v.5 | 0.44 | ii.iii | 0.55 | three.vi | 0.08 |
| Radius VR–VL rotation | ii.6 | −0.threescore | ii.8 | −0.fifteen | two.6 | 0.61 |
| Radius F–East rotation | i.1 | 0.99 | 1.i | 0.99 | 2.vii | 0.99 |
| Ulna Due south–I displacement | 0.4 | 0.07 | 0.two | 0.xc | 0.four | 0.97 |
| Ulna A–P deportation | 0.4 | 0.98 | 0.4 | 0.99 | 0.vi | 0.98 |
| Ulna One thousand–Fifty displacement | iii.4 | −0.45 | 2.0 | −0.62 | two.3 | −0.14 |
| Radius South–I deportation | ane.iii | 0.99 | 0.6 | 0.99 | ane.3 | 0.99 |
| Radius A–P deportation | 1.0 | 0.98 | 0.8 | 0.99 | 0.8 | 0.97 |
| Radius M–L displacement | 1.v | 0.86 | 0.half dozen | 0.91 | 0.9 | 0.91 |
For each velocity status, the timings when the muscle started to actuate in each loop were similar both in model prediction and experimental muscle excitation measured from EMG (Effigy 7). The total numbers of musculus activations were also consistent between the model and experiment. In addition, most of the normalized summit forces were in agreement between model and experiment. Still, in our model prediction, the maximum muscle contraction occurred faster than the experimental measurement; therefore, the musculus relaxation also started earlier. Although more pronounced, the model was able to reproduce the tendency very well.
Normalized experimental electromyography (EMG) and normalized muscle forces from model prediction. (a) 10 deg/s; (b) threescore deg/southward; (c) free velocity. Muscle forces are normalized to the maximum force produced by each musculus for the specific trial.
three.ii. Force Prediction
The core reward of the computation model is its ability to predict of import parameters that are very hard or impossible to measure out experimentally, such as ligament strength and cartilage contact pressure. In this written report, the model predicted noticeably increased peak ligament loads with increasing forearm velocity (Table 3). Contact pressure distributions on the humeral cartilage were also considerably different for diverse velocity conditions (Figure 8). Both contact area and contact pressure were noticeably increased with increasing velocity. Summit contact force per unit area (on medial cartilage) was 3.7 MPa for 10 deg/s, 4.two MPa for 60 deg/due south, and five.5 MPa for free velocity. Every bit expected, contact areas were much higher for the ulnohumeral joint (medial contact) than for the radiohumeral joint.
Contact pressure distribution on humeral cartilage for musculus-driven frontwards dynamic simulation. x- and y-axes units are in mm and the intensity of colour represents the intensity of contact pressure. The red boundary represents the unfolded 2D projection of the humerus cartilage.
Table 3
Ligament meridian load throughout the simulation flow for different velocity weather.
| Ligament | Bundles | Peak Ligament Load (N) | ||
|---|---|---|---|---|
| 10 deg/due south | 60 deg/southward | Complimentary Velocity | ||
| MCL anterior role | Inductive | 23 | 37 | 101 |
| Central | 45 | 61 | 123 | |
| Posterior | 61 | 76 | 112 | |
| MCL posterior part | Anterior | 35 | 40 | 55 |
| Central | 30 | 44 | 87 | |
| Posterior | 24 | 47 | 125 | |
| LUCL | Inductive | 43 | 58 | 80 |
| Cardinal | 23 | 33 | 44 | |
| Posterior | 26 | 25 | 50 | |
| RCL | Anterior | 21 | 44 | 131 |
| Central | eighteen | 26 | 55 | |
| Posterior | 12 | 31 | 88 | |
| Annular ligament | Proximal | 46 | 51 | 52 |
| Distal | 36 | 41 | 41 | |
Computation fourth dimension for our longest 35 s simulation (10 deg/s status) was less than 30 min (0.01 s step size, default ADAMS solver, desktop PC (Intel®Xenon®CPUE5-16070@3.00GHz with 32 GB of RAM). Simulation times for the other two cases were even smaller.
4. Discussion
The principal goal of this report was to develop an anatomically correct subject-specific musculoskeletal model of the elbow joint. The model was evaluated by comparing the bone kinematics from the model simulation (muscle-driven forward dynamics) to the experimental motion-captured kinematics. Equally an additional evaluation, the normalized muscle forces were likewise compared with the experimental normalized EMG data. The model included 3-dimensional bone and cartilage geometries. The elbow joint was constrained by nonlinear ligament bundles and three-dimensional deformable contacts, instead of an idealized mechanical joint. The model also included the natural oblique wrapping of muscle and ligaments. Articular contact force and contact expanse predictions were achieved by discretizing the humerus cartilage into multiple hexahedral elements. An optimization and the design of the experiment arroyo were used to determine the contact parameters and the size of discretized cartilage elements. The developed musculoskeletal model is capable of concurrently predicting muscle and ligament forces, along with cartilage contact mechanics during dynamic activities. Such concurrent prediction from an anatomical model has the potential to be a powerful predictive tool in Orthopaedics. The model is as well very computationally efficient for torso-level dynamic simulation. Such a simulation in finite element analysis would take an enormous corporeality of time to solve.
Previously, computational musculoskeletal models have been employed to predict the muscle activation and upper limb strength by constraining the elbow joint equally a single one-degree-of-freedom mechanical joint [31,32,33]. Withal, this assumption removes the influence of contact forces on muscle forces and musculus contributions to motion beyond the sagittal plane. Because of the interdependency between the articular contact and the musculus forces, those forces need to be computed concurrently. Furthermore, three-dimensional measurements of simulated active elbow motility by an electromagnetic tracking device revealed the amount of potential varus-valgus laxity that occurred during elbow flexion to be well-nigh 3 to four degrees [7]. Ignoring this laxity by placing a mechanical articulation can affect the muscle forcefulness predictions. In addition, the omission of this normal laxity into the implant design is the reason behind the failure of fully constrained elbow replacement implants, because it increases the stress transfer to the implant-cement-bone interfaces and results in aseptic loosening [34]. Therefore, an accurate elbow model should reverberate the intrinsic laxity of the elbow, particularly for clinical applications. In this study, the elbow joint was constrained by the ligaments and the ulnohumeral, radiohumeral, and radioulnar contact forces allowing 18 degrees of liberty.
The model-predicted kinematics compared well with the experimental measurements. However, some variations were observed in kinematic comparing, especially in the varus–valgus directions (Table ii). Nosotros did non include the joint capsule in our model which may be one of the major contributing factors to this error. Morrey and An [35] reported that the anterior and posterior capsule provided 32% varus and 33% valgus elbow stability, respectively. The mechanical properties of the elbow joint capsule were not available in the literature, which express its inclusion in our model. In addition, during a large range of flexion–extension, the Optotrak markers could slightly motility from their initial position due to relative musculus movement from bone and may introduce some measurement artefacts.
The activation patterns of each of the contributing major muscles were correctly identified for all velocity weather and favourably compared with experimental EMG measurements (Figure vii). The pocket-size departure of the muscle activation pattern may exist due in part to the experimental measurement of EMG voltage. We were able to measure two surface EMGs from 2 muscle groups (triceps and biceps) of the upper arm. Isolating the specific musculus group during the experiment was extremely difficult; this might induce some cross-signal errors in the EMG measurement. In addition, only iii muscles were powering the elbow flexion-extension in the model versus the total 24 muscles cantankerous the elbow articulation [1,36], which may influence the musculus to actuate faster than in the real subject experiment and may cause stage difference. To our understanding, incorporating more muscles in the model could better the muscle force prediction.
Our predicted joint contact pressure was increased by increasing the arm flexion–extension velocity. This appears reasonable since at higher velocity the musculus forces are expected to increase, which would lead to higher compression at the joint and increased contact force per unit area. This is also consequent with the increasing ligament load at a higher velocity of the arm. Although the experimental measurement of joint contact pressure is not feasible from a live subject, our external prediction of joint kinematics and muscle activation replicate the experiment, which gives some confidence of our internal joint contact pressure prediction. Furthermore, our predicted ulnohumeral contact and noncontact areas were consequent with the contact patterns reported by Eckstein et al. [37] and the maximum contact pressures were close in range to the values (0–5 MPa) reported by Make [38]. Overestimation of the pressure may exist explained by the choice of the ligament zip-load length. The correction factor used in our written report was constant, which may not exist ideal for all ligament bundles [23] and may innovate some ligament tightness.
A potential limitation of the present model was that information technology was adult based on a single-subject experiment and is therefore characteristic of a unmarried elbow. A larger sample size would allow more generalized conclusions to be made for clinical applications. However, the modelling techniques, efficiency, and accuracy were successfully tested by the research team for five cadaver studies [xiv,24] and 3 studies of live subject musculoskeletal modelling of a lower extremity [25,26,28]. Furthermore, the model included only three major muscles that crossed the elbow joint. Other muscles cantankerous either the shoulder or wrist articulation forth with the elbow articulation. Therefore, a complete musculoskeletal model of the upper extremity containing shoulder, elbow, and wrist joints will allow a complete testing of hypotheses that depict the forces responsible for elbow articulation loading. In addition, the muscle forces are computed using a feedback controller in this study. Ane problem with using a feedback controller is that information technology cannot predict muscle forces without irresolute the musculus length. Hybrid control methods where some EMG signals are used every bit input to the model may solve this issue.
The reward of the modelling arroyo and its applicability as a predicting tool greatly outweigh its limitation. To date, no other computational model has had the capability to predict the elbow joint kinematics, muscle and ligament forces, and cartilage contact force per unit area concurrently. Instead, past models accept assumed a particular joint degree of liberty and ignored the effect of ligaments, cartilage, or physiological properties of the joint. This study developed a subject-specific musculoskeletal elbow joint model in which the joint was constrained past multiple ligament bundles, 3-dimensional deformable contacts, and included the natural oblique wrapping of ligaments. The model has demonstrated a powerful power to predict of import parameters that are extremely hard or even impossible to measure experimentally, such as muscle and ligament forces, and cartilage contact pressure. This model and modelling approach are a step toward developing a complete musculoskeletal model of the upper extremity. This method is non intended to supervene upon existing models or modelling approaches, but rather to provide an additional supporting tool to understand the elbow joint functions and pathologies. Further refinements, such as more detailed subject-specific measurements and comeback of the musculature, can enhance its accurateness and clinical applicability. The potential medical awarding of this model and modelling approach is significant and is anticipated as a clinical tool for the development of patient-specific preoperative planning, computer-aided surgery, and figurer-aided rehabilitation.
Acknowledgments
This enquiry was supported in part by funds provided through the School of Graduate Studies Enquiry Grant program and School of Medicine of the University of Missouri-Kansas City (UMKC). The authors gratefully acknowledge the help of Jonathan Parman from Musculoskeletal Biomechanics Inquiry Laboratory, UMKC, in data collection.
Author Contributions
Thousand. Rahman, A.P. Stylianou, and A. Cil conceived and designed the project. M. Rahman, M.S. Renani, A. Cil, and A.P. Stylianou performed the experiment. G. Rahman developed the model. M. Rahman and M.S. Renani analyzed the data and performed statistical analyses. All authors discussed the results and contributed to the writing of the manuscript.
Conflicts of Interest
The authors declare no conflict of interest.
References
1. Morrey B.F. The Elbow and Its Disorders. W.B. Saunders; Philadelphia, PA, United states of america: 2000. [Google Scholar]
2. Garner B.A., Pandy M.Thou. Musculoskeletal model of the upper limb based on the visible human being male person dataset. Comput. Methods Biomech. Biomed. Eng. 2001;four:93–126. doi: 10.1080/10255840008908000. [PubMed] [CrossRef] [Google Scholar]
3. Gonzalez R.5., Hutchins E.50., Barr R.East., Abraham Fifty.D. Development and evaluation of a musculoskeletal model of the elbow joint circuitous. J. Biomech. Eng. 1996;118:32–40. doi: 10.1115/1.2795943. [PubMed] [CrossRef] [Google Scholar]
4. Holzbaur M.R., Murray Due west.M., Delp South.L. A model of the upper extremity for simulating musculoskeletal surgery and analyzing neuromuscular control. Ann. Biomed. Eng. 2005;33:829–840. doi: x.1007/s10439-005-3320-7. [PubMed] [CrossRef] [Google Scholar]
five. Willing R.T., Nishiwaki Thou., Johnson J.A., Rex Grand.J., Athwal G.Southward. Evaluation of a computational model to predict elbow range of motion. Comput. Aided Surg. 2014;19:57–63. doi: ten.3109/10929088.2014.886083. [PMC free commodity] [PubMed] [CrossRef] [Google Scholar]
6. Benham 1000.P., Wright D.Grand., Bibb R. Modelling soft tissue for kinematic analysis of multi-segment human body models. Biomed. Sci. Instrum. 2001;37:111–116. [PubMed] [Google Scholar]
seven. Tanaka Due south., An Thousand.-N., Morrey B.F. Kinematics and laxity of ulnohumeral articulation under valgus-varus stress. J. Musculoskelet. Res. 1998;ii:45–54. doi: 10.1142/S021895779800007X. [CrossRef] [Google Scholar]
8. Fisk J.P., Wayne J.Southward. Development and validation of a computational musculoskeletal model of the elbow and forearm. Ann. Biomed. Eng. 2009;37:803–812. doi: 10.1007/s10439-009-9637-x. [PubMed] [CrossRef] [Google Scholar]
ix. Asraf Ali M., Sundaraj K., Badlishah Ahmad R., Ahamed N.U., Islam A. Recent observations in surface electromyography recording of triceps brachii musculus in patients and athletes. Appl. Bionics Biomech. 2014;11:105–118. doi: 10.1155/2014/172815. [CrossRef] [Google Scholar]
10. Van Woensel W., Arwert H. Effects of external load and abduction angle on emg level of shoulder muscles during isometric activeness. Electromyogr. Clin. Neurophysiol. 1993;33:185–191. [PubMed] [Google Scholar]
eleven. Cignoni P., Callieri M., Corsini M., Dellepiane One thousand., Ganovelli F., Ranzuglia One thousand. Meshlab: An open up-source mesh processing tool; Proceedings of the Sixth Eurographics Italian Affiliate Conference; Salerno, Italy. ii–4 July 2008; pp. 129–136. [Google Scholar]
12. Donahue T.L., Hull M.L., Rashid M.K., Jacobs C.R. A finite element model of the homo genu articulation for the written report of tibio-femoral contact. J. Biomech. Eng. 2002;124:273–280. [PubMed] [Google Scholar]
13. Zielinska B., Donahue T.L. 3D finite element model of meniscectomy: Changes in articulation contact behavior. J. Biomech. Eng. 2006;128:115–123. doi: ten.1115/one.2132370. [PubMed] [CrossRef] [Google Scholar]
14. Rahman M., Cil A., Johnson Thousand., Lu Y., Gauge T.M. Development and validation of a computational multibody model of the elbow joint. Adv. Biomech. Appl. 2014;ane:169–185. doi: x.12989/aba.2014.1.three.169. [CrossRef] [Google Scholar]
15. Rahman M., Cil A., Bogener J.W., Stylianou A.P. Lateral collateral ligament deficiency of the elbow joint: A modeling approach. J. Orthop. Res. 2016;34:1645–1655. doi: x.1002/jor.23165. [PubMed] [CrossRef] [Google Scholar]
16. Miyake J., Moritomo H., Masatomi T., Kataoka T., Murase T., Yoshikawa H., Sugamoto K. In vivo and 3-dimensional functional anatomy of the anterior bundle of the medial collateral ligament of the elbow. J. Shoulder Elb. Surg. 2012;21:1006–1012. doi: 10.1016/j.jse.2011.07.018. [PubMed] [CrossRef] [Google Scholar]
17. Morrey B.F., An K.Due north. Functional anatomy of the ligaments of the elbow. Clin. Orthop. Relat. Res. 1985:84–90. doi: x.1097/00003086-198512000-00015. [PubMed] [CrossRef] [Google Scholar]
18. Blankevoort L., Huiskes R. Ligament-bone interaction in a three-dimensional model of the knee. J. Biomech. Eng. 1991;113:263–269. doi: 10.1115/1.2894883. [PubMed] [CrossRef] [Google Scholar]
19. Wismans J., Veldpaus F., Janssen J., Huson A., Struben P. A three-dimensional mathematical model of the genu-articulation. J. Biomech. 1980;xiii:677–685. doi: 10.1016/0021-9290(fourscore)90354-one. [PubMed] [CrossRef] [Google Scholar]
20. Li G., Gil J., Kanamori A., Woo Due south.L. A validated iii-dimensional computational model of a human being articulatio genus joint. J. Biomech. Eng. 1999;121:657–662. doi: x.1115/1.2800871. [PubMed] [CrossRef] [Google Scholar]
21. Blankevoort L., Kuiper J.H., Huiskes R., Grootenboer H.J. Articular contact in a three-dimensional model of the knee joint. J. Biomech. 1991;24:1019–1031. doi: 10.1016/0021-9290(91)90019-J. [PubMed] [CrossRef] [Google Scholar]
22. Regan W.D., Korinek Southward.Fifty., Morrey B.F., An K.Northward. Biomechanical study of ligaments around the elbow joint. Clin. Orthop. Relat. Res. 1991:170–179. doi: 10.1097/00003086-199110000-00023. [PubMed] [CrossRef] [Google Scholar]
23. Bloemker M.H., Judge T.Thou., Maletsky 50., Dodd M. Computational knee ligament modeling using experimentally determined goose egg-load lengths. Open up Biomed. Eng. J. 2012;half dozen:33–41. doi: 10.2174/1874120701206010033. [PMC free article] [PubMed] [CrossRef] [Google Scholar]
24. Rahman Grand., Cil A., Stylianou A.P. Prediction of elbow articulation contact mechanics in the multibody framework. Med. Eng. Phys. 2016;38:257–266. doi: 10.1016/j.medengphy.2015.12.012. [PubMed] [CrossRef] [Google Scholar]
25. Guess T.Chiliad., Stylianou A.P., Kia Thou. Concurrent prediction of muscle and tibiofemoral contact forces during treadmill gait. J. Biomech. Eng. 2014;136:021032. doi: 10.1115/1.4026359. [PMC gratuitous article] [PubMed] [CrossRef] [Google Scholar]
26. Kia K., Stylianou A.P., Approximate T.M. Evaluation of a musculoskeletal model with prosthetic knee through half dozen experimental gait trials. Med. Eng. Phys. 2014;36:335–344. doi: 10.1016/j.medengphy.2013.12.007. [PMC costless article] [PubMed] [CrossRef] [Google Scholar]
27. Renani M.S., Rahman M., Cil A., Stylianou A.P. Calibrating multibody ulno-humeral joint cartilage using a validated finite element model. Multibody Syst. Dyn. 2018 doi: 10.1007/s11044-018-9622-y. [CrossRef] [Google Scholar]
28. Stylianou A.P., Guess T.M., Melt J.L. Development and validation of a multi-body model of the canine stifle articulation. Comput. Methods Biomech. Biomed. Eng. 2014;17:370–377. doi: 10.1080/10255842.2012.684243. [PubMed] [CrossRef] [Google Scholar]
29. Stylianou A.P., Guess T.Thou., Kia Yard. Multibody muscle driven model of an instrumented prosthetic human knee during squat and toe rise motions. J. Biomech. Eng. 2013;135:041008. doi: 10.1115/one.4023982. [PMC free commodity] [PubMed] [CrossRef] [Google Scholar]
30. Renani M.S., Rahman K., Cil A., Stylianou A.P. Ulna-humerus contact mechanics: Finite element assay and experimental measurements using a tactile pressure level sensor. Med. Eng. Phys. 2017;50:22–28. doi: 10.1016/j.medengphy.2017.08.010. [PubMed] [CrossRef] [Google Scholar]
31. Carmichael Grand.K., Liu D. Upper limb strength estimation of physically impaired persons using a musculoskeletal model: A sensitivity assay. Conf. Proc. IEEE Eng. Med. Biol. Soc. 2015;2015:2438–2441. [PubMed] [Google Scholar]
32. Gonzalez R.V., Abraham L.D., Barr R.E., Buchanan T.S. Muscle action in rapid multi-caste-of-freedom elbow movements: Solutions from a musculoskeletal model. Biol. Cybern. 1999;80:357–367. doi: 10.1007/s004220050531. [PubMed] [CrossRef] [Google Scholar]
33. Gonzalez R.Five., Andritsos M.J., Barr R.Due east., Abraham L.D. Comparison of experimental and predicted muscle activation patterns in ballistic elbow articulation movements. Biomed. Sci. Instrum. 1993;29:9–16. [PubMed] [Google Scholar]
34. O'Driscoll S.W., An K.N., Korinek Southward., Morrey B.F. Kinematics of semi-constrained total elbow arthroplasty. J. Bone Jt. Surg. Br. 1992;74:297–299. doi: 10.1302/0301-620X.74B2.1544973. [PubMed] [CrossRef] [Google Scholar]
35. Morrey B.F., An K.N. Articular and ligamentous contributions to the stability of the elbow joint. Am. J. Sports Med. 1983;xi:315–319. doi: 10.1177/036354658301100506. [PubMed] [CrossRef] [Google Scholar]
36. Pigeon P., Yahia L., Feldman A.G. Moment arms and lengths of man upper limb muscles as functions of joint angles. J. Biomech. 1996;29:1365–1370. doi: 10.1016/0021-9290(96)00031-0. [PubMed] [CrossRef] [Google Scholar]
37. Eckstein F., Merz B., Muller-Gerbl K., Holzknecht N., Pleier M., Putz R. Morphomechanics of the humero-ulnar joint: 2. Concave incongruity determines the distribution of load and subchondral mineralization. Anat. Rec. 1995;243:327–335. doi: 10.1002/ar.1092430307. [PubMed] [CrossRef] [Google Scholar]
38. Make R.A. Articulation contact stress: A reasonable surrogate for biological processes? Iowa Orthop. J. 2005;25:82–94. [PMC gratis article] [PubMed] [Google Scholar]
Manufactures from Bioengineering are provided hither courtesy of Multidisciplinary Digital Publishing Establish (MDPI)
Source: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6027184/
0 Response to "what kind of mechanical model can we use to represent the elbow joint?"
Post a Comment