Manual and semi-automatic determination of elbow angle-independent parameters for a model of the biceps brachii distal tendon based on ultrasonic imaging

Tendons consist of passive soft tissue with non linear material properties. They play a key role in force transmission from muscle to skeletal structure. The properties of tendons have been extensively examined in vitro. In this work, a non linear model of the distal biceps brachii tendon was parameterized based on measurements of myotendinous junction displacements in vivo at different load forces and elbow angles. The myotendinous junction displacement was extracted from ultrasound B-mode images within an experimental setup which also allowed for the retrieval of the exerted load forces as well as the elbow joint angles. To quantify the myotendinous junction movement based on visual features from ultrasound images, a manual and an automatic method were developed. The performance of both methods was compared. By means of exemplary data from three subjects, reliable fits of the tendon model were achieved. Further, different aspects of the non linear tendon model generated in this way could be reconciled with individual experiments from literature.

Analysis on synergistic cocontraction of extrinsic finger flexors and extensors during flexion movements: A finite element digital human hand model

Fine hand movements require the synergistic contraction of intrinsic and extrinsic muscles to achieve them. In this paper, a Finite Element Digital Human Hand Model (FE-DHHM) containing solid tendons and ligaments and driven by the Muscle-Tendon Junction (MTJ) displacements of FDS, FDP and ED measured by ultrasound imaging was developed. The synergistic contraction of these muscles during the finger flexion movements was analyzed by simulating five sets of finger flexion movements. The results showed that the FDS and FDP contracted together to provide power during the flexion movements, while the ED acted as an antagonist. The peak stresses of the FDS, FDP and ED were all at the joints. In the flexion without resistance, the FDS provided the main driving force, and the FDS and FDP alternated in a "plateau" of muscle force. In the flexion with resistance, the muscle forces of FDS, FDP, and ED were all positively correlated with fingertip forces. The FDS still provided the main driving force, but the stress maxima occurred in the FDP at the DIP joint.

Reliability and Accuracy of a Time-Efficient Method for the Assessment of Achilles Tendon Mechanical Properties by Ultrasonography

The assessment of the force-length relationship under mechanical loading is widely used to evaluate the mechanical properties of tendons and to gain information about their adaptation, function, and injury. This study aimed to provide a time-efficient ultrasound method for assessing Achilles tendon mechanical properties. On two days, eleven healthy young non-active adults performed eight maximal voluntary isometric ankle plantarflexion contractions on a dynamometer with simultaneous ultrasonographic recording. Maximal tendon elongation was assessed by digitizing ultrasound images at rest and at maximal tendon force. Achilles tendon stiffness index was calculated from the ratio of tendon force-to-strain. No within- and between-day differences were detected between the proposed method and manual frame by frame tracking in Achilles tendon maximal force, maximal elongation, maximal strain, and stiffness index. The overall coefficient of variation between trials ranged from 3.4% to 10.3% and average difference in tendon tracking between methods was less than 0.6% strain. Furthermore, an additional assessment demonstrated significant differences between elite athletes, healthy young, and older adults in Achilles tendon force and stiffness index. Hence, the analysis has the potential to reliably and accurately monitor changes in Achilles tendon mechanical properties due to aging and altered mechanical loading in a time-efficient manner.

A Human-Centered Machine-Learning Approach for Muscle-Tendon Junction Tracking in Ultrasound Images

Biomechanical and clinical gait research observes muscles and tendons in limbs to study their functions and behaviour. Therefore, movements of distinct anatomical landmarks, such as muscle-tendon junctions, are frequently measured. We propose a reliable and time efficient machine-learning approach to track these junctions in ultrasound videos and support clinical biomechanists in gait analysis. In order to facilitate this process, a method based on deep-learning was introduced. We gathered an extensive data set, covering 3 functional movements, 2 muscles, collected on 123 healthy and 38 impaired subjects with 3 different ultrasound systems, and providing a total of 66864 annotated ultrasound images in our network training. Furthermore, we used data collected across independent laboratories and curated by researchers with varying levels of experience. For the evaluation of our method a diverse test-set was selected that is independently verified by four specialists. We show that our model achieves similar performance scores to the four human specialists in identifying the muscle-tendon junction position. Our method provides time-efficient tracking of muscle-tendon junctions, with prediction times of up to 0.078 seconds per frame (approx. 100 times faster than manual labeling). All our codes, trained models and test-set were made publicly available and our model is provided as a free-to-use online service on https://deepmtj.org/.

Automated analysis of medial gastrocnemius muscle-tendon junction displacements in heathy young adults during isolated contractions and walking using deep neural networks

BACKGROUND AND OBJECTIVE

Direct measurement of muscle-tendon junction (MTJ) position is important for understanding dynamic tendon behavior and muscle-tendon interaction in healthy and pathological populations. Traditionally, obtaining MTJ position during functional activities is accomplished by manually tracking the position of the MTJ in cine B-mode ultrasound images - a laborious and time-consuming process. Recent advances in deep learning have facilitated the availability of user-friendly open-source software packages for automated tracking. However, these software packages were originally intended for animal pose estimation and have not been widely tested on ultrasound images. Therefore, the purpose of this paper was to evaluate the efficacy of deep neural networks to accurately track medial gastrocnemius MTJ positions in cine B-mode ultrasound images across tasks spanning controlled loading during isolated contractions to physiological loading during treadmill walking.

METHODS

Cine B-mode ultrasound images of the medial gastrocnemius MTJ were collected from 15 subjects (6M/9F, 23 yr, 71.9 kg, 1.8 m) during treadmill walking at 1.25 m/s and during maximal voluntary isometric plantarflexor contractions (MVICs). Five deep neural networks were trained using 480 manually-labeled images collected during walking, defined as the ground truth, and were then used to predict MTJ position in images from novel subjects: 1) during walking (novel-subject) and 2) during MVICs (novel-condition).

RESULTS

We found an average mean absolute error of 1.26±1.30 mm and 2.61±3.31 mm between the ground truth and predicted MTJ positions in the novel-subject and novel-condition evaluations, respectively.

CONCLUSIONS

Our results provide support for the use of open-source software for creating deep neural networks to reliably track MTJ positions in B-mode ultrasound images. We believe this approach to MTJ position tracking is an accessible and time-saving solution, with broad applications for many fields, such as rehabilitation or clinical diagnostics.

Automatic Tracking of the Muscle Tendon Junction in Healthy and Impaired Subjects using Deep Learning

Recording muscle tendon junction displacements during movement, allows separate investigation of the muscle and tendon behaviour, respectively. In order to provide a fully-automatic tracking method, we employ a novel deep learning approach to detect the position of the muscle tendon junction in ultrasound images. We utilize the attention mechanism to enable the network to focus on relevant regions and to obtain a better interpretation of the results. Our data set consists of a large cohort of 79 healthy subjects and 28 subjects with movement limitations performing passive full range of motion and maximum contraction movements. Our trained network shows robust detection of the muscle tendon junction on a diverse data set of varying quality with a mean absolute error of 2.55 ± 1 mm. We show that our approach can be applied for various subjects and can be operated in real-time. The complete software package is available for open-source use.

Semi-automatic methods for tracking the medial gastrocnemius muscle-tendon junction using ultrasound: a validation study

NEW FINDINGS

What is the central question of this study? Is the proposed semi-automatic algorithm suitable for tracking the medial gastrocnemius muscle-tendon junction in ultrasound images collected during passive and active conditions? What is the main finding and its importance? The validation of a method allowing efficient tracking of the muscle-tendon junction in both passive and active conditions, in healthy as well as in pathological conditions. This method was tested in common acquisition conditions and the developed software made freely available.

ABSTRACT

Clinically relevant information can be extracted from ultrasound (US) images by tracking the displacement of the junction between muscle and tendon. This paper validated automatic methods for tracking the location of muscle-tendon junction (MTJ) between the medial gastrocnemius and the Achilles tendon during passive slow and fast stretches, and active ankle rotations while walking on a treadmill. First, an automatic algorithm based on an optical flow approach was applied on collected US images. Second, results of the automatic algorithm were evaluated and corrected using a quality measure that indicated which critical images need to be manually corrected. US images from 12 typically developed (TD) children, 12 children with spastic cerebral palsy (SCP) and eight healthy adults were analysed. Automatic and semi-automatic tracking methods were compared to manual tracking using root mean square errors (RMSE). For the automatic tracking, RMSE was less than 3.1 mm for the slow stretch and 5.2 mm for the fast stretch, the worst case being for SCP. The tracking results in the fast stretch condition were improved (especially in SCP) by using the semi-automatic approach, with an RMSE reduction of about 30%. During walking, the semi-automatic method also reduced errors, with a final RMSE of 3.6 mm. In all cases, data processing was considerably shorter using the semi-automatic method (2 min) compared to manual tracking (20 min). A quick manual correction considerably improves tracking of the MTJ during gait and allows to achieve results suitable for further analyses. The proposed algorithm is freely available.

Reliability of ultrasound speckle tracking with singular value decomposition for quantifying displacement in the carpal tunnel

Inhibited movement patterns of carpal tunnel structures have been found in carpal tunnel syndrome (CTS) patients. Motion analysis on ultrasound images allows us to non-invasively study the (relative) movement of carpal tunnel structures and recently a speckle tracking method using singular value decomposition (SVD) has been proposed to optimize this tracking. This study aims to assess the reliability of longitudinal speckle tracking with SVD in both healthy volunteers and patients with CTS. Images from sixteen healthy volunteers and twenty-two CTS patients were used. Ultrasound clips of the third superficial flexor tendon and surrounding subsynovial connective tissue (SSCT) were acquired during finger flexion-extension. A custom made tracking algorithm was used for the analysis. Intra-class correlation coefficients (ICCs) were calculated using a single measure, two-way random model with absolute agreement and Bland-Altman plots were added for graphical representation. ICC values varied between 0.73 and 0.95 in the control group and 0.66-0.98 in the CTS patients, with the majority of the results classified as good to excellent. Tendon tracking showed higher reliability values compared to the SSCT, but values between the control and CTS groups were comparable. Speckle tracking with SVD can reliably be used to analyze longitudinal movement of anatomical structures with different sizes and compositions within the context of the carpal tunnel in both a healthy as well as a pathological state. Based on these results, this technique also holds relevant potential for areas where ultrasound based dynamic imaging requires quantification of motion.

Comparison of the Achilles tendon moment arms determined using the tendon excursion and three-dimensional methods

The moment arm of muscle‐tendon force is a key parameter for calculating muscle and tendon properties. The tendon excursion method was used for determining the Achilles tendon moment arm (ATMA). However, the accuracy of this method remains unclear. This study aimed to investigate the magnitude of error introduced in determining the ATMA using the tendon excursion method by comparing it with the reference three‐dimensional (3D) method. The tendon excursion method determined the ATMA as the ratio between the Achilles tendon displacement during foot rotation from 15° of dorsiflexion to 15° of plantarflexion and the joint rotation angle. A series of foot images was obtained at 15° of dorsiflexion, the neutral position, and 15° of plantarflexion. The 3D value of the ATMA was determined as the shortest distance between the talocrural joint axis and the line of action of the Achilles tendon force. The ATMA determined by the tendon excursion method was smaller by 3.8 mm than that determined using the 3D method. This error may be explained mainly by the length change in the Achilles tendon due to the change in the force applied to it, as passive plantarflexion torque was different by 11 Nm between 15° of dorsiflexion and 15° of plantarflexion. Furthermore, the ATMAs determined using the 3D and tendon excursion methods were significantly correlated but the coefficient of determination was not large (R² = 0.352). This result suggests that the tendon excursion method may not be feasible to evaluate the individual variability of the ATMA.

Use of a Lucas-Kanade-Based Template Tracking Algorithm to Examine In Vivo Tendon Excursion during Voluntary Contraction Using Ultrasonography

Ultrasound imaging can be used to study tendon movement during muscle contraction to estimate the tendon force-length relationship in vivo. Traditionally, such tendon displacement measurements are made manually (time consuming and subjective). Here we evaluated a Lucas-Kanade-based tracking algorithm with an optic flow extension that accounts for tendon movement characteristics between consecutive frames of an ultrasound image sequence. Eleven subjects performed 12 voluntary isometric plantar flexion contractions on a dynamometer. Simultaneously, the gastrocnemius medialis tendon was visualized via ultrasonography. Tendon displacement was estimated manually and by using two different automatic tracking algorithms. Maximal tendon elongation (manual: 17.9 ± 0.3 mm, automatic: 17.0 ± 0.3 mm) and tendon stiffness (209 ± 4 N/mm, 218 ± 5 N/mm) generated by the developed algorithm correlated with those obtained with the manual method (0.87 ≤ R ≤ 0.91), with no differences between methods. Our results suggest that optical flow methods can potentially be used for automatic estimation of tendon movement during contraction in ultrasound images, which is further improved by adding a penalty function.

The use of normalized cross-correlation analysis for automatic tendon excursion measurement in dynamic ultrasound imaging

The work describes an automated method of tracking dynamic ultrasound images using a normalized cross-correlation algorithm, applied to the patellar and gastrocnemius tendon. Displacement was examined during active and passive tendon excursions using B-mode ultrasonography. In the passive test where two regions of interest (2-ROI) were tracked, the automated tracking algorithm showed insignificant deviations from relative zero displacement for the knee (0.01 ± 0.04 mm) and ankle (-0.02 ± 0.04 mm) (P > .05). Similarly, when tracking 1-ROI the passive tests showed no significant differences (P > .05) between automatic and manual methods, 7.50 ± 0.60 vs 7.66 ± 0.63 mm for the patellar and 11.28 ± 1.36 vs 11.17 ± 1.35 mm for the gastrocnemius tests. The active tests gave no significant differences (P > .05) between automatic and manual methods with differences of 0.29 ± 0.04 mm for the patellar and 0.26 ± 0.01 mm for the gastrocnemius. This study showed that automatic tracking of in vivo displacement of tendon during dynamic excursion under load is possible and valid when compared with the standardized method. This approach will save time during analysis and enable discrete areas of the tendon to be examined.

In vivo strain analysis of the intact supraspinatus tendon by ultrasound speckles tracking imaging

Limited data exist regarding why delaminated tears occur within the rotator cuff tendon, and no general agreement exists on how to handle this complicated tear. To analyze in vivo intratendinous strain of the supraspinatus tendon, the superficial, middle, and deep regions were marked with speckles using 2D speckle tracking echocardiography (2D STE) of 15 shoulders. The displacement and the strain of each speckle during isotonic and isometric shoulder motion were evaluated. Significantly different displacement and strains in the tendon were found between isometric and isotonic shoulder motions. In isometric motion, the average longitudinal displacement of the speckle at the superficial region (1.66 mm) was larger than at the deep region (0.61 mm), and the average peak strain at the superficial region (17.03%) was also higher than that at the deep region (3.42%). Conversely, in isotonic motion, the average longitudinal displacement of the speckle at the superficial region (0.70 mm) was less than that at the deep region (1.61 mm), and the average peak strain at the superficial region (4.73%) was also lower than that at the deep region (15.69%). A different strain was found between the superficial and deep regions within the intact live supraspinatus tendon. The strain and displacement patterns vary according to isometric versus isotonic shoulder motions. On the basis of our observations, we suggest that the delaminated tear of the rotator cuff tendon must be repaired separately layer by layer to resist the inhomogeneous strain after the repair.

Automated regional analysis of B-mode ultrasound images of skeletal muscle movement

To understand the functional significance of skeletal muscle anatomy, a method of quantifying local shape changes in different tissue structures during dynamic tasks is required. Taking advantage of the good spatial and temporal resolution of B-mode ultrasound imaging, we describe a method of automatically segmenting images into fascicle and aponeurosis regions and tracking movement of features, independently, in localized portions of each tissue. Ultrasound images (25 Hz) of the medial gastrocnemius muscle were collected from eight participants during ankle joint rotation (2° and 20°), isometric contractions (1, 5, and 50 Nm), and deep knee bends. A Kanade-Lucas-Tomasi feature tracker was used to identify and track any distinctive and persistent features within the image sequences. A velocity field representation of local movement was then found and subdivided between fascicle and aponeurosis regions using segmentations from a multiresolution active shape model (ASM). Movement in each region was quantified by interpolating the effect of the fields on a set of probes. ASM segmentation results were compared with hand-labeled data, while aponeurosis and fascicle movement were compared with results from a previously documented cross-correlation approach. ASM provided good image segmentations (<1 mm average error), with fully automatic initialization possible in sequences from seven participants. Feature tracking provided similar length change results to the cross-correlation approach for small movements, while outperforming it in larger movements. The proposed method provides the potential to distinguish between active and passive changes in muscle shape and model strain distributions during different movements/conditions and quantify nonhomogeneous strain along aponeuroses.

Ultrasound assessment of the motion patterns of human flexor digitorum superficialis and profundus tendons with speckle tracking

The purposes of our study were to correlate ultrasonographically measured and joint angle estimated excursions of the flexor digitorum superficialis (FDS) and flexor digitorum profundus (FDP) tendons of the hand and to estimate the relative motion of FDS and FDP while gripping cylinders of standard diameter in normal human subjects. Thirty wrists from 15 human subjects were imaged with an ultrasound scanner. Speckle tracking was used to measure the excursion of the FDS and FDP tendons. The tendon excursions necessary to grip three differently sized acrylic tubes were measured and correlated with the corresponding finger joint angles. The FDP/FDS excursion ratio was calculated. The Pearson's correlation coefficient between the FDS excursion and MP + PIP joint angle was 0.61. The Pearson's correlation coefficient between the FDP + FDS excursion and the DIP + PIP + MP joint angle was 0.67. The FDP/FDS excursion ratio was smaller for larger excursions (gripping a smaller diameter tube) and larger for small excursions (gripping a larger diameter tube, P < 0.01). These data suggest that speckle tracking may be a useful method to discriminate the relative motion of flexor tendons, which in turn may be relevant in evaluating tendon function, for example after tendon injury.

Direct comparison of in vivo Achilles tendon moment arms obtained from ultrasound and MR scans

Accurate and reliable estimation of muscle moment arms is a prerequisite for the development of musculoskeletal models. Numerous techniques are available to estimate the Achilles tendon moment arm in vivo. The purposes of this study were 1) to compare in vivo Achilles tendon moment arms obtained using the center of rotation (COR) and tendon excursion (TE) methods and 2) to assess the reliability of each method. For the COR method, magnetic resonance (MR) images from nine participants were obtained at ankle angles of −15°, 0°, and +15° and analyzed using Reuleaux' method. For the TE method, the movement of the gastrocnemius medialis-Achilles tendon junction was recorded using ultrasonography as the ankle was passively rotated through its range of motion. The Achilles tendon moment arm was obtained by differentiation of tendon displacement with respect to ankle angular excursion using seven different differentiation techniques. Moment arms obtained using the COR method were significantly greater than those obtained using the TE method ( P < 0.01), but results from both methods were well correlated. The coefficient of determination between moment arms derived from the COR and TE methods was highest when tendon displacement was linearly differentiated over a ±10° interval ( R2 = 0.94). The between-measurement coefficient of variation was 3.9% for the COR method and 4.5–9.7% for the TE method, depending on the differentiation technique. The high reliabilities and strong relationship between methods demonstrate that both methods are robust against their limitations. The large absolute between-method differences (∼25–30%) in moment arms have significant implications for their use in musculoskeletal models.

Ultrasonographic assessment of long finger tendon excursion in zone v during passive and active tendon gliding exercises

PURPOSE

Cadaver and in vivo studies report variable results for tendon excursion during active and passive hand movements. The purpose of this study was to measure long finger flexor digitorum profundus (FDP) tendon excursion during active and passive movement using high-resolution ultrasound images.

METHODS

The FDP tendon excursion was measured at the wrist level in 10 healthy subjects during full tip-to-palm active and passive flexion of the fingers. Passive movement was performed 2 ways: (1) straight to full fist: passive flexion starting at the metacarpophalangeal joint, followed by proximal interphalangeal and distal interphalangeal joint flexion; and (2) hook to full fist: passive flexion starting at the distal interphalangeal joint, followed by proximal interphalangeal and metacarpophalangeal joint flexion. Tendon excursion was measured using an in-house-developed, frame-to-frame analysis of high-resolution ultrasound images.

RESULTS

Median FDP excursion was 24.3 mm, 14.0 mm, and 13.6 mm for active fist, straight to full fist, and hook to full fist movements, respectively. Tendon excursions during active movements was significantly larger than excursions during passive movements (p = .005). The adjusted median tendon excursion was 12.7 mm/100 degrees , 7.5 mm/100 degrees , and 7.4 mm/100 degrees for active fist, straight to full fist, and hook to full fist movements, respectively. Adjusted tendon excursions during active movement were significantly larger than those achieved during passive straight to full fist movement). Adjusted tendon excursions during straight to full fist movements were significantly larger than those achieved during passive hook to full fist movement.

CONCLUSIONS

Active motion produced 74% and 79% increases in excursions compared to both passive motions in healthy controls. The study results can serve as a reference for evaluating excursions in patients with tendon pathology, including those who have had tendon repair and reconstruction.

Architectural properties of the first dorsal interosseous muscle

Muscle architecture is considered to reflect the function of muscle in vivo, and is important for example to clinicians in designing tendon-transfer and tendon-lengthening surgeries. The purpose of this study was to quantify the architectural properties of the FDI muscle. It is hypothesized that there will be consistency, that is low variability, in the architectural parameters used to describe the first dorsal interosseous muscle because of its clear functional role in index finger motion. The important architectural parameters identified were those required to characterize a muscle adequately by modeling. Specifically the mass, cross-sectional area, and length of the tendon and muscle were measured in cadavers along with the muscle fiber optimum length and pennation angle, and the moment arm of the first dorsal interosseous at the metacarpophalangeal joint. These parameters provide a characterization of the architecture of the first dorsal interosseous, and were used to indicate the inherent variability between samples. The results demonstrated a large amount of variability for all architectural parameters measured; leading to a rejection of the hypothesis. Ratios designed to describe the functioning of the muscles in vivo, for example the ratio of tendon to fiber optimum lengths, also demonstrated a large variability. The results suggest that function cannot be deduced from form for the first dorsal interosseous, and that subject-specific architectural parameters may be necessary for the formulation of accurate musculoskeletal models or making clinical decisions.

Built for speed: musculoskeletal structure and sprinting ability

The musculoskeletal structure of the foot and ankle has the potential to influence human sprinting performance in complex ways. A large Achilles'tendon moment arm improves the mechanical advantage of the triceps surae but also produces larger shortening velocity during rapid plantarflexion, which detracts from the force-generating capacity of the plantarflexors. The lever arm of the ground reaction force that resists the muscular plantarflexor moment during propulsive push-off is constrained in part by the skeletal structure of the foot. In this study, we measured the plantarflexion moment arms of the Achilles' tendon, lateral gastrocnemius fascicle lengths and pennation angles, and anthropometric characteristics of the foot and lower leg in collegiate sprinters and height-matched non-sprinters. The Achilles' tendon moment arms of the sprinters were 25% smaller on average in sprinters than in non-sprinters (P&lt;0.001) whereas the sprinters' fascicles were 11%longer on average (P=0.024). The ratio of fascicle length to moment arm was 50% larger in sprinters (P&lt;0.001). Sprinters were found to have longer toes (P=0.032) and shorter lower legs (P=0.026)than non sprinters. A simple computer simulation of the sprint push-off demonstrated that shorter plantarflexor moment arms and longer toes, like those measured in sprinters, permit greater generation of forward impulse. Simulated propulsion was enhanced in both cases by increasing the `gear ratio'of the foot, thus maintaining plantarflexor fibre length and reducing peak fibre shortening velocity. Longer toes especially prolonged the time of contact, giving greater time for forward acceleration by propulsive ground reaction force.