For both squat and ankle rotation motions extracted from the ROM file, Table 4 shows the wide range of joint angle error caused by changing each subject measurement independently. The data suggest that the rotation z axis in each joint is the most prone to error in a majority of cases. However, Knee abduction y axis is also greatly affected by subject measurement errors. Ankle non-sagittal angles y and z axes show large error, but are not often considered in clinical analysis using the CGM. Importantly, during a squat motion the flexion x axis was not the largest source of error.
Additionally, the range of error between the minimum and maximum throughout the motion suggests that the error is not a fixed offset. The irregular effect of subject measurements on joint angles is due in part to both the calculation method S1 Fig and the hierarchal configuration. During motion such as a squat, the knee flexion axis is relatively unchanged, as this primary axis is defined by the lateral knee marker. Additionally, a change in the leg length measurement changes the hip joint center, which in turn changes the orientation of the knee rotation axis z.
Hence both the leg length Fig 9 and knee width Fig 10 measurements affect the knee joint angle in a different manner over the course of a motion. Graph of the right and left knee joint angles during a squat motion over frames 3,—3, of the ROM trial. The original angle is represented by the grey line and the leg length modifications are shown by the black line. Through the hierarchal system, the largest angle error of all combinations occurrs most often in the ankle joint Table 5.
The frame in which each joint incurs the largest error across all measurements is largely consistent within the individual joint, but varies across joints. The large difference in the most common joint axis angles between right and left sides shows that the initial frame orientation and definition greatly influences the resulting errors from subject measurements Table 6.
The code style of pyCGM was developed to be straight forward to understand and modify. Beyond the utility of being cross platform without modification there are no pointers or objects used in the calculation. Subject measurements are stored in a dictionary and are easily referred to throughout the code.
Improvement of the original CGM hip joint center location estimations [ 20 ] were implemented based on the work of Harrington et al. Fig 11 shows the minimal amount of code. The use of a keyword setting Harrington to True activates the if statement to use the Harrington Hip Regression Method. It is also possible to determine the knee joint center locations using the method proposed by Stief [ 22 ]. The midpoint between medial and lateral markers is used as the joint center with the vector defined as the midpoint to the medial marker.
Fig 12 shows the implementation in the pyCGM system. As the originally described method in [ 22 ] did not give exact details, the implementation here is approximate. However, modification of the method is simple. If a Medial marker is detected, the code will use it to determine the knee joint center and axis. This paper presents experimental data for high performance computing of the CGM used for calculating joint kinematics.
It demonstrates three methods; specifically, the use of HPC for a single large file of motion capture data by distributing frames to be calculated on separate cores and across nodes, and the distribution of frames across cores on a consumer grade desktop, was investigated. Second, large calculations were performed on a single dataset to derive valuable information about the model, such as the case of subject measurement errors, through the distribution of data across cores. As this implementation is based on direct kinematics and uses a single static calibration file, the computation time can easily be reduced through parallelization, an aspect that has been largely left out of discussions on joint kinematics models.
The choice to implement the CGM in Python was made primarily due to the portability between operating systems and the wide acceptance of Python for scripting purposes in the scientific community [ 38 ]. The interpreted language and easily understood syntax continues to promote the open source aspect of the work.
As much of the data reported in the literature is focused on the lower body due to a focus on gait , this paper focused on the lower body. While the upper body has been developed in pyCGM and is included in the computational time, the lack of data relating to marker placement and skin deformation of the upper body makes it difficult to compare the significance of subject measurement errors. Small differences observed may be due to numerical implementation differences or rounding differences.
Importantly, in the Vicon CGM, the method used to compute the knee and ankle joint center is unclear; however, here, an exact formulation based on the Rodrigues rotation formula [ 37 ] is presented. Moreover, the original work of Kadaba [ 12 ] specifies the use of the arcsin function which would result in erroneous joint angles during motions greater than 90 degrees. In the pyCGM code, the arctan function is used to solve this problem.
The flexibility of python and straightforward code of pyCGM allows researchers to easily view, modify, and expand the CGM. The non-optimized Direct Kinematic DK method allows for a frame-by-frame calculation for joint kinematics and, as such, provides an opportunity for easy HPC implementation to assist in the kinematic processing.
These calculation times are dramatically improved by moving from a classical desktop setup to an HPC, as detailed in the results section. As a stand-alone compiled program, comparison against Nexus also demonstrates the real speed benefit of parallelization as the interpreted python code is significantly faster than the compiled code of Nexus.
At the same time, under utilizing the parallelized methods by reducing the number of cores used has an adverse effect on computation time. While the HPC is orders of magnitude faster than Nexus when using multiple nodes, the use of only 1 core shows significantly slower processing times. This is attributed to the lower clock speed of processors containing large numbers of cores. However, challenges exist for the type of large databases of motion capture data that can fully take advantage of HPC.
The first is the need to improve the ability to track people in an efficient way. While devices such as the Microsoft Kinect may allow for cheaper 3D reconstructions, a more relevant advancement would be the replacement of reflective markers and infrared cameras with a regular RGB camera with tracking markers, such as QR tags, that can be easily implemented on-site in a variety of cases while taking advantage of the research already done on joint landmark kinematic calculations.
In the laboratory, more robust gap-filling and auto-labeling technology using the current motion capture systems would also greatly improve the efficiency and bring the use of HPC to a more common audience. The primary, or axis of most importance during an analysis, such as knee flexion during a squat motion, is central to understanding the importance of subject measurements. As seen in the experimental data, during a squat motion the flexion axis in both left and right knee axis were almost never the source of the largest error. As such, while subject measurements have an effect on joint angles, hip and knee flexion are the least affected and may still provide the most valid results.
The percentages of error displayed in Table 4 are based on the data found in [ 23 ] in which the maximum deviation from 7 laboratories in leg length measurements was 25 mm, knee diameter of 5 mm, and ankle diameter of 10 mm. While accurate subject measurements are required for the CGM, the importance of these measurements should be considered within the context of other possible sources of angle error. In [ 39 ], 10 mm of marker placement error resulted in 6. This results in a 2. While the kinematic model and subject motion from [ 39 ] slightly vary from those used in this research, this comparison should give researchers a basis for determining both which aspects of subject preparation are most vital and which aspects of the CGM new models should overcome.
In a more general sense, the results suggest that subject measurements can have a significant impact on joint angles. In practice, such a large error in leg length may not be likely for experienced users of the CGM c. Subject measurement errors within the range observed by [ 23 ] result in clinically negligible angle errors less than 2 degrees [ 40 ].
Past work has shown that rather than completely new models, modifications to existing models provide familiarity and improved accuracy. In [ 22 ], the knee joint center estimation was improved by using medial markers. This method was easily integrated into the pyCGM code by adding 18 lines of code directly in the knee joint calculation function with no removal or other modifications necessary as the code switches to this method when medial knee markers are detected.
Although this method requires additional markers, the overall marker set remains the same. Without additional markers, the hip joint center estimation can be improved with the Harrington method [ 41 ]. This method has been implemented through 9 lines of code in the hip axis calculation, 5 lines of code in the static calibration, and 1 line of code in the execution file which acts as the argument to switch between hip joint center methods. Likewise, [ 42 ] shows an improved knee joint center through the adjustment of the thigh offset. Given the open-source nature of pyCGM, interested users could implement this change in the future.
The numerous studies for validating the results during gait, the wide usage, and the deep understanding of the model [ 43 ] remain important standards for the CGM. In a large field in which multiple methods for calculating joint kinematics are possible, the use of standards for validated models is of great importance.
Since the creation of the original CGM, new methods for computing joint kinematics have been introduced, with the most relevant works modifying the CGM to be used with kinematic fitting and optimizations. C-motion Visual 3D offers a version of the CGM that uses optimization methods through inverse kinematics [ 44 ].
Likewise, methods such as the optimized lower-limb gait analysis OLGA use inverse kinematics by global optimization [ 3 ]. However, these methods often have implications for computational complexity, as in the case of OLGA, in which more than 50 frames of data are required to be considered for good convergence [ 4 ], making parallelization a more complex task and the overall computation time much larger. Furthermore, the use of inverse kinematics over direct kinematics is not a guarantee for more accurate results, as research has found that the anatomical model used in the study has a greater effect on kinematics than the different computational methods [ 5 ].
Likewise, the CGM is not without critics and drawbacks, however; as models aim to improve the accuracy of joint angle calculations over the CGM, the effect of subject measurements in errors must be understood. Beyond the work presented here, there is a need for new computational models to take into account spatial and time complexity, as these are directly related to the impact and implementation possibilities throughout the biomedical industry and other fields. Real-time algorithms for single subject analysis have been developed, which creates implications for clinician-patient interaction [ 45 ].
When considering least squares optimization and regression-based methods in the dynamic trial calculation, it is not possible to implement real-time calculations, something that may be useful in a clinical setting for doctor-patient interaction. At the same time, these methods may benefit from HPC when parallelization or distribution is possible.
Additionally, the number of markers being recorded has a significant impact on the storage required for these large datasets, hindering data sharing and distribution. However, further research into the actual computational time increase required by optimization methods would help define these limitations. Applications for computation of large databases of motion capture data extend from biomechanics and robotics to other, less obviously related fields, such as architecture.
In architecture, understanding human movement and movement abilities is important for design and necessary in order to move from prescriptive to performative design criteria [ 46 ]. In general, any field relying on interactions with humans and movement will at some point need to address the computational efficiency of large scale calculations for analysis. Finally, further work involving subject measurements in the CGM can shed light on how significant other measurements beyond leg length, knee width, and ankle width are.
This may also lead to the ability for subject measurements to be derived directly from marker locations during the static calibration, which would remove the need for storing patient specific data while maintaining a reasonable margin of error. For the knee joint center, the calculation is from the thigh marker a , hip joint center b , and knee marker c.
The intent is to find the plane in which all markers lay, with half the knee width kw being used in the calculation. The python code provides comparison between the mathematics and code that is easy to read and understand. The function receives three marker positions and half of the knee width measurement. The return value is the cartesian location of the calculated joint center. Ease of understanding the code is an important aspect of pyCGM, and as such, the steps are divided clearly so that users can both understand and modify the code to suit their needs.
Motion capture data of a sitting motion in which the Knee bends to 90 degrees. While this function was intended for use in gaits that would not commonly have a 90 degree flexion, the widespread use of the CGM includes researchers using it for purposes beyond typical gait. The Sum column is the sum of the times required for each step in the calculation process.
The Total column is the recorded start to finish time, with the difference shown in the last column. This difference includes communication time between cores and nodes, as can be seen from the increased difference when the calculation moved from 1 node to 2 nodes. The average, maximum, and minimum times for each core to complete the calculations are shown. Additionally, the longest time for any core to complete all calculations is shown. Loading data was all done on the initial core. Saving the results, dynamic trial calculation, and static trial calculation times are from every core.
The sum of these calculations and the total time recorded from the first node differ mostly due to data transfer between nodes. The author would like to thank Dr. Robert G. Van Wesep for his assistance in technical issues with the implementation on the high performance computer, Seungeun Yeon for her active development in pyCGM and assistance with early versions of the research, and the many interns that have provided contributions to the code throughout its development.
Additionally, the authors thank the University of Michigan Digital Media Commons for their resources and assistance in developing figures. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author s and do not necessarily reflect the views of the University of Tennessee, Oak Ridge National Laboratory, or the Joint Institute for Computational Sciences.
Browse Subject Areas? Click through the PLOS taxonomy to find articles in your field. Abstract The conventional gait model CGM is a widely used biomechanical model which has been validated over many years. Funding: The authors received no specific funding for this work. Download: PPT. Fig 2. Abstraction of averaged calculation vs. Fig 4. Abstraction diagram of a parallelized dynamic trial calculation on the HPC. Fig 5. Abstraction diagram of distributed computation of multiple files on the HPC. Table 1. Space requirement for saving lower body joint angles and axis to a.
Intervals of 0. An additional 9 3 combinations involving AT and IAD were used for the computational performance experiment, but not used for data analysis. Excluding duplicates from each range, such as all measurements being 0. Table 2. Fig 6. Fig 7. Computation time of a dynamic trial across multiple cores and nodes. Table 3. Computational performance of kinematic calculations for a dynamic trial on the HPC. Fig 8. Computation time of multiple variations of the ROM trial. Table 4. Joint angle error from variations in subject measurements. Fig 9. Joint angles of the knee during squat while changing the leg length.
Fig Joint angles of the knee during squat while changing the knee width. Table 6. Most frequent axis containing the largest angle error across all subject measurement variations. Supporting information. S1 Fig. Joint center calculation of the knee. S2 Fig. Sample code of pyCGM. S3 Fig. Flexion beyond 90 degrees using arcsin and arctan. S1 Equation. Equation of the knee joint calculation using Rodrigues rotation formula. S1 Table. Results of single dynamic trial scaled on HPC. S2 Table. Computational performance of kinematic calculations for multiple variations of a dynamic trial on the HPC.
S1 File. Output from the HPC. S2 File. Breakdown of subject measurement errors. Acknowledgments The author would like to thank Dr. References 1. Baker R. The history of gait analysis before the advent of modern computers. Gait Posture. View Article Google Scholar 2. A note on the description of articulating joint motion. Journal of Biomechanics.
View Article Google Scholar 3. Roren L, Tate P. A new lower body model using global optimisation techniques. View Article Google Scholar 4. Repeatability of an optimised lower body model. Gait and Posture. View Article Google Scholar 5. Joint kinematic calculation based on clinical direct kinematic versus inverse kinematic gait models Ms. View Article Google Scholar 6. Position and orietnation in space of bones during movement. Clin Biomech. View Article Google Scholar 7.
A six degrees-of-freedom marker set for gait analysis: repeatability and comparison with a modified Helen Hayes set. View Article Google Scholar 8. Repeatability of kinematic, kinetic, and electromyographic data in normal adult gait. Journal of Orthopaedic Research. View Article Google Scholar 9. Plug-in Gait manual v1; Accessed, Aug 22nd Nexus; Sutherland DH. The evolution of clinical gait analysis.
Part II Kinematics. View Article Google Scholar Measurement of lower extremity kinematics during level walking. Journal of orthopaedic research: official publication of the Orthopaedic Research Society. A joint coordinate system for the clinical description of three-dimensional motions: application to the knee. Journal of biomechanical engineering.
Assessment of the kinematic variability among 12 motion analysis laboratories. France L, Nester C. Effect of errors in the identification of anatomical landmarks on the accuracy of Q angle values. Clinical biomechanics Bristol, Avon. Bone position estimation from skin marker co-ordinates using global optimisation with joint constraints. Surface movement errors in shank kinematics and knee kinetics during gait. Human movement analysis using stereophotogrammetry Part 3.
Soft tissue artifact assessment and compensation; When used by an optimization, these actuators may represent forces and torques exerted by soft tissues and other non-muscular supportive mechanisms Rankin et al. Reserve actuator values were sensitive to TSL assumptions—shorter tendons placed muscle fibres on disadvantageous regions of their force-length curves, resulting in relatively larger reserve actuator torques to complete the StS dynamics. Both results suggest that these reserve actuators likely are acting as a direct compensation for the non-optimal operating ranges of muscle fibres during the StS movement.
Thus, tendon compliance appears to be a key component to successfully completing StS, in a potentially complex relationship with support from other passive tissues. Hence, it becomes clearer how critical tendons are to these simulations; future simulations using forward dynamics with compliant tendons are certain to obtain qualitatively different findings see also Rankin et al.ositough.henkel.buildingonline.com/beacon-lights-of-history-volume.php
Introduction and State of the Art - Kinematic Analysis of Human Movement - Wiley Online Library
Williams et al. Although the slow, near-static motions of StS would preclude the power modulation sensu Haldane et al. Nonetheless, until empirical measurements of what passive joint torques for given joint ranges of motion are available for greyhound hindlimbs such data do not yet exist to our knowledge , there is no right or wrong answer to this conundrum.
We found that greyhounds stand up through sequential proximal to distal within the hindlimbs; Figure 3 extension at the hip, then knee then ankle, consistent with past examination of StS in dogs Feeney et al. There are very few quantitative data on comparable StS motions. Nickel et al. Our greyhounds extended the hindlimbs first, for unclear reasons. Zannier-Tanner reported very diverse patterns of StS limb extension timing for mammalian herbivores, so general principles regarding these basic motions remain elusive.
Our results for maximal joint ranges of motion RoM also qualitatively match those for the only other StS study of dogs that we are aware of, by Feeney et al. Their RoMs tend to be greater for the forelimbs e. These differences are interesting in light of Bertram's et al. StS likewise seems to involve distinct mechanisms for these two breeds; corresponding in part to differences in limb muscle architecture Williams et al.
Joint RoMs during StS in greyhounds are larger for many degrees of freedom than published data indicate for locomotion in greyhounds or other canine breeds; much as Feeney et al. De Camp et al. We had expected substantial non-sagittal motions during StS as a way to circumvent the steep evolutionary constraints on sagittal motion imposed by walking and running.
Part of this contradiction may be by study design: we only examined trials where dogs began the motion from an adducted, bilaterally symmetrical posture on their bellies. This was a common behaviour used by almost all our subjects. However, many dogs chose to lie on their side while being positioned for this study, and would then frequently use non-sagittal rolling motions to stand up.
These motions were much more variable, often obscured motion capture markers, and excluded from this study. Rolling during StS would involve strongly asymmetrical limb function and likely require more non-sagittal movement; bolstering that aspect of our Hypothesis 2. Because of the large amount of uncertainty in our estimates of non-sagittal motions in our experimental data, we performed a sensitivity analysis to understand how changes in these joint angles may influence our simulations.
We examined two extreme scenarios no long-axis joint rotation vs. Greater long-axis rotations incurred greater muscle activity in the simulations Figure 9.
The initially selected representative trial had near-zero ankle internal rotations, which we assumed to reflect the limited rotations that are anatomically possible. Our study has provided the first dataset for limb kinetics during StS in dogs, revealing deeper biomechanical mechanisms used to accomplish the movement. As the GRFs required to complete StS are provided primarily by the hindlimbs, this reinforces our decision to simulate a hindlimb. GRFs are low during StS in greyhounds, compared with maximal speed locomotion.
For example, Bryant et al. For similar-sized subjects as ours, Williams et al. Our model's key extensor muscle moment arms for brevity, not plotted here are generally near their lowest values early in StS and qualitatively in agreement with cadaver data Williams et al. This reinforces that mechanical advantage is low when peak GRFs occur in StS, reflecting our Hypothesis 2 that muscle activations are high for greyhounds during StS. Such relatively high activation should apply in particular for comparisons to walking, in which peak GRFs are similar but joint moments are lower than in StS Wentink, We also observed a substantial craniad translation of the COP during StS in our greyhounds Figure S4 corresponding to a shift from a plantigrade to a digitigrade foot posture Figure 1 , Supplementary Movies S1 , S3 that kept joint moments lower than if the feet somehow remained in a digitigrade orientation as in standing.
Similar patterns of low GRFs but poor mechanical advantage leading to high demands placed on limb extensor muscles prevail for humans during StS. As previously described, it is widely accepted that knee extensor moments are large in StS, which imposes strength:weight ratio limitations on individuals with muscle weakness or other deficits Hughes et al.
Few computational models, let alone simulations, of dog hindlimbs exist but our active MTUs generally match those from a static optimization-based simulation of three-legged stance phase Shahar and Banks-Sills, The ExtDigLong is not typically active during stance phase in canine locomotion Wentink, , unlike our nominal simulation's estimated activation pattern that we judged questionable.
Investigating this result further, we found that, due to the extremely flexed posture at the ankle, the ankle moment arm of this muscle was extensor i. Altered kinematics resulted in the ExtDigLong not having an ankle extensor moment arm and thus was inactive Figure 9. Humans use homologous or analogous limb muscles to conduct StS. Increased activity occurs if the former muscles if load is added to subjects, whereas other muscles M.
Savelberg et al. Furthermore, StS in humans tends to involve marked co-activation of muscles with antagonist e. The common muscle coordination strategy in StS found in humans favours key antigravity muscles along with some coactivation of antagonist muscles and is similarly observed in our dog simulations, Our study therefore hints at a broader pattern that might prevail across mammals or even tetrapods—a subject worthy of further inquiry. Oddly however, our simulations did not estimate an activation sequence of muscles from proximal to distal: key antigravity muscles all became maximally active immediately with StS initiation Figures 6—8 , which is unlike in human StS and unlike our proximal-distal joint kinematics pattern, but could be an artefact of our static optimization criteria Pandy et al.
Regardless, distal limb muscle activations seemed to remain high for longer during StS relative to proximal limb muscles, and this was relatively insensitive to the input parameters we varied. While in vivo or modelling data on hindlimb muscle length changes are scarce for locomotor behaviours in greyhounds or other dog breeds or other species , prevailing evidence indicates more isometric patterns for most limb muscles, keeping muscles closer to their optima for force production e.
However, Goslow et al. Contrastingly, GMed and BF actively lengthened whereas VL actively shortened and GasM actively lengthened then shortened both to modest amounts during StS in our greyhound simulations Figures 6—8 — indicating clear differences in MTU work and power for locomotion vs. StS that future studies should pursue. Our actively shortening SM and actively lengthening VL in StS, however, qualitatively match direct sonomicrometry measurements in jumping and running dogs, as follows. Thus, at least for muscle fascicle length changes in two demanding behaviours and two muscles, MTU length changes in StS grossly match cf.
StS in humans and dogs shows some interesting similarities and differences in terms of MTU or fascicle length changes. Using experiments and a simple 2D model similar to Goslow's et al. These, especially VL and VM Figure 7 , roughly correspond to our patterns of active muscle length change except for the GSup and BF1 which actively lengthened Figure 6 ; our GasL and GasM results modest active stretch-shortening are more ambiguous comparisons Figure 8.
While tendons surely play an important role in StS see below , we predict little potential for elastic energy storage; unlike in locomotion or jumping e. This is because StS involves relatively slow quasi-static , vertical motions without a quick countermovement-style stretch that should disfavour rapid conversion of elastic strain energy into kinetic energy to raise the body's centre of mass or move it forwards. Similarly, Hughes et al. Our experimental data involved kinematic data from multiple markers placed on each segment to estimate 3D joint motions in StS from our greyhound subjects.
Small differences in marker placement have been observed to substantially alter measured joint angles Kadaba et al. Combined with the substantial individual variability in StS behaviour that was evident within and among individuals, this helps to explain the variation in our kinematic data e. Improved 3D kinematic data would augment our results by reducing uncertainties regarding the limb joint motions used in the simulations.
However, increasing the number of markers was prohibitive because of prolonged experimental setup time, and could alter the StS motion by creating discomfort or obstructing joint mobility. Regardless, the simplification we adopted was an appreciable step forward from the only other published study of StS in non-humans that we are aware of, involving 2D kinematics in Labrador retriever dogs Feeney et al.
Additionally, our model incorporated six muscles that were not in the Williams et al. While subjective investigator and measurement errors may have contributed, a large part of this difference may come from choice of subjects: we studied normal, domesticated, household greyhounds at a range of ages and fitness rather than using active athletes as were those studied by Williams et al.
A simulation of a more active athlete thus should have lower muscle activations due to larger muscle areas; if so, our Hypothesis 1 regarding length changes might remain unaffected but Hypotheses 2 and 3 muscle activations and passive tissue support might be weakened. By using an extensive experimental dataset for StS and anatomically realistic model, we simulated muscular mechanics in a greyhound in unprecedented detail while maximizing the rigor of the data involved. Tests of the validity of our model and simulations are less of a concern under these conditions, except in key areas as follows.
Our assumption that tendons were rigid was intentionally unrealistic, allowing us to tease apart how muscles alone may contribute to StS. It is interesting that muscles can successfully drive StS in our greyhound simulation; with some quantifiable passive support at more joints than others. We addressed this assumption of rigid tendons in more detail with our sensitivity analysis of tendon slack length and reserve actuator torques. Further sensitivity analyses of the input experimental data strengthened our hypothesis testing. However, some concerns remain.
Tendon slack length TSL values are a related limitation. This, however, might lead our model to be less able to simulate non-StS motions such as walking or running without re-tuning TSL values. Furthermore, experimental data e. We used static optimization to generate our simulations, which included an objective function that minimized the sum of muscle activations squared as did Actis et al. However, this approach likely generates results that do not exactly match the controls that greyhounds actually use in StS and may explain many disparities in our results vs.
Pandy et al. Indeed, we expect that their algorithm would give better results i. Later, Bobbert et al. Resolution of this issue awaits more study of what different species optimize in StS decisions see also Erdemir et al. Our models and simulations have considerable uncertainties and assumptions, yet even in light of these we contend that our combined experimental and computational analysis of StS dynamics in greyhounds supports our hypotheses that key antigravity hindlimb muscles operate close to their limits of length change, and even perhaps force and thus activation and mechanical work during StS.
We infer that this proximity to biomechanical limits requires substantial contributions from soft tissues including tendons and perhaps ligaments or other arthrological features in order to achieve StS. Although the limb forces in StS are less than in high-speed locomotion and more comparable to the forces experienced during walking, the unfavourable mechanical advantage of the limb joints requires substantial forces and activations from the limb muscles.
This requirement is only amplified by the increased length changes of the muscles required to produce the measured joint ranges of motion, moving the animals from a crouched, supine position to an erect, upright limb orientation and using most of the feasible ranges of motion of the joints themselves as well as the muscles. Relative to normal walking and running, non-sagittal motions are increased in the StS behaviour we focused on here and do impose some extra demands on muscles such as hip adductors.
Alternative StS strategies for quadrupeds include rolling behaviours that should exaggerate non-sagittal motions in return for poorly understood mechanical benefits, and limb coordination patterns that should involve greater force production and length changes from the limb muscles. Our study has shown how non-locomotor biomechanical demands deserve further consideration in the study of how the musculoskeletal system is adapted to, and constrained by, these demands vs. We have elucidated how greyhounds can sustain some aspects of the sit-to-stand transition with their muscles but need passive tissues, including but not necessarily limited to tendons, to fully achieve it because of their cursorial limb structure's emphasis on short-fibred distal muscles and adaptation to parasagittal, upright-limbed locomotion.
All authors contributed to the writing of the manuscript and gave final approval for publication. RE helped design the experiments and simulations and analysed the data. JR helped develop software, run simulations and assisted with data interpretation. JH conceived and guided the project and contributed to the experimental design and simulations and assisted with data interpretation. We thank the Whitaker International Program for its funding and support throughout this project. The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
We thank the dog owners who provided their animals for study, and staff of the Structure and Motion Laboratory for assistance with data collection and analysis. We thank Dr. Vivian Allen for advice on the usage of his segment mass and moment of inertia software. Additionally, RE thanks Dr. Jarrett Rushmore for the use of his lab space during writing. We thank the two reviewers for their helpful comments.
Actis, J. Lumbar loads and trunk kinematics in people with a transtibial amputation during sit-to-stand. Alexander, R. Winters and S. Google Scholar. Allen, V. Linking the evolution of body shape and locomotor biomechanics in bird-line archosaurs. Nature , — Amit, T. Segmental inertial properties in dogs determined by magnetic resonance imaging. Azizi, E. Variable gearing in pennate muscles. Bertram, J. Comparison of the trotting gaits of Labrador Retrievers and Greyhounds. Bieryla, K. Estimation of relative effort during sit-to-stand increase when accounting for variations in maximum voluntary torque with joint angle and angular velocity.
Biewener, A. Scaling body support in mammals: limb posture and muscle mechanics. Science , 45— Bobbert, M. Searching for strategies to reduce the mechanical demands of the sit-to-stand task with a muscle-actuated optimal control model. Bryant, J. Forces exerted on the ground by galloping dogs Canis familiaris. Buckley, T. Age impairs sit-to-walk motor performance. Burkholder, T. Sarcomere length operating range of vertebrate muscles during movement. PubMed Abstract Google Scholar. Colborne, G. Distribution of power across the hind limb joints in Labrador Retrievers and Greyhounds.
Cutts, A. Sarcomere length changes in muscles of the human thigh during walking. Dall, P. Frequency of the sit to stand task: an observational study of free-living adults. Davy, D. A dynamic optimization technique for predicting muscle forces in the swing phase of gait. De Camp, C. Kinematic gait analysis of the trot in healthy Greyhounds. The effect of the partially restricted sit-to-stand task on biomechanical variables in subjects with and without Parkinson's disease.
Deban, S. Activity of extrinsic limb muscles in dogs at walk, trot and gallop. Delp, S. OpenSim: open-source software to create and analyze dynamic simulations of movement. IEEE Transact. A graphics-based software system to develop and analyze models of musculoskeletal structures. Dimery, N. Muscle and sarcomere lengths in the hind limb of the rabbit Oryctolagus cuniculus during a galloping stride.
- The Dramatic Imagination: Reflections and Speculations on the Art of the Theatre (Theatre Arts Book)?
- Laurence Chèze.
- Shahnameh: The Persian Book of Kings!
- Keats and Embarrassment.
- Services on Demand;
Dries, B. Musculoskeletal modelling in dogs: challenges and future perspectives. Musculotendon excursion potential, tendon slack and muscle fibre length: the interaction of the canine gastrocnemius muscle and tendon. Erdemir, A. Model-based estimation of muscle forces exerted during movements. Feeney, L. Validation of two-dimensional kinematic analysis of walk and sit-to-stand motions in dogs. Fluit, R. Prediction of ground reaction forces and moments during various activities of daily living. Fujimoto, M. Dynamic balance control during sit-to-stand movement: an examination with the center of mass acceleration.
Gardner, R. Gillette, J. Simulation of functional neuromuscular stimulation assisted sit-to-stand movements. Goslow, G. Electrical activity and relative length changes of dog limb muscles as a function of speed and gait. Gregersen, C. Gear ratios at the limb joints of jumping dogs. External work and potential for elastic storage at the limb joints of running dogs. Haldane, D. Robotic vertical jumping agility via series-elastic power modulation. Robotics 1:eaag Helms, G.
Multi-body simulation of a canine hind limb: model development, experimental validation and calculation of ground reaction forces. Online Herbert, R. Change in length of relaxed muscle fascicles and tendons with knee and ankle movement in humans. Hicks, J. Is my model good enough?
Best practices for verification and validation of musculoskeletal models and simulations of movement. Hughes, M. The role of strength in rising from a chair in the functionally impaired elderly. Chair rise strategies in the elderly.
Kinematic analysis of human movement.
Janssen, W. Determinants of the sit-to-stand movement: a review. Jayes, A. Estimates of mechanical stresses in leg muscles of galloping Greyhounds Canis familiaris. Kadaba, M. Measurement of lower extremity kinematics during level walking. Khemlani, M. Muscle synergies and joint linkages in sit-to-stand under two initial foot positions. Lichtwark, G. A modified Hill muscle model that predicts muscle power output and efficiency during sinusoidal length changes. Lidfors, L. The use of getting up and lying down movements in the evaluation of cattle environments. Maganaris, C. Force—length characteristics of in vivo human skeletal muscle.
Acta Physiol. Manafzadeh, A. ROM mapping of ligamentous constraints on avian hip mobility: implications for extinct ornithodirans.
Manal, K. Subject-specific estimates of tendon slack length: a numerical method. McGowan, C. A phenomenological muscle model to assess history dependent effects in human movement. Millard, M. Flexing computational muscle: modeling and simulation of musculotendon dynamics. Newton, C. Textbook of Small Animal Orthopaedics.
Nickel, R. The Anatomy of the Domestic Animals. The Locomotor System of the Domestic Mammals. Berlin, Verlag Paul Parey. Nishikawa, K. A new twist on muscle contraction. B Biol. Pandy, M. Optimal control of non-ballistic muscular movements: a constraint-based performance criterion for rising from a chair. Payne, R. Functional specialisation of pelvic limb anatomy in horses Equus caballus. Pedersen, D. A model to predict canine pelvic limb musculoskeletal geometry. Cells Tissues Organs , — Rankin, J. Inferring muscle functional roles of the ostrich pelvic limb during walking and running using computer optimization.
Redl, C. Sensitivity of muscle force estimates to variations in muscle—tendon properties. Reisman, D. Coordination underlying the control of whole body momentum during sit-to-stand. Gait Posture 15, 45— Roberts, T. Muscular force in running turkeys: the economy of minimizing work. Science , — Robertson, G. Research Methods in Biomechanics. Rodosky, M. The influence of chair height on lower limb mechanics during rising. Roebroeck, M. Biomechanics and muscular activity during sit-to-stand transfer.
Rubenson, J. Running in ostriches Struthio camelus , three-dimensional joint axes alignment and joint kinematics. On the ascent: the soleus operating length is conserved to the ascending limb of the force—length curve across gait mechanics in humans. Sacks, R. Architecture of the hind limb muscles of cats: functional significance.
Savelberg, H. Scovil, C. Sensitivity of a Hill-based muscle model to perturbations in model parameters. Shahar, R. Biomechanical analysis of the canine hind limb: calculation of forces during three-legged stance. Morphometric and anatomic study of the hind limb of a dog. Shia, V. Stability basin estimates fall risk from observed kinematics, demonstrated on the Sit-to-Stand task.
Silva, A. Activation timing of soleus and tibialis anterior muscles during sit-to-stand and stand-to-sit in post-stroke vs. Motor Res.
Related Kinematic Analysis of Human Movement (Focus)
Copyright 2019 - All Right Reserved