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

Automated Analysis of Ultrasound Images to Measure Muscle-Tendon Junction Excursions by Using the Multiple Feature Point Tracking Algorithm

Ultrasound imaging is used to measure the muscle-tendon junction (MTJ) to investigate the mechanical properties of the tendon and the interaction of the muscle-tendon unit in vivo. Although the MTJ can be observed clearly in the resting state, accurate tracking of the MTJ is difficult during muscle contractions due to changes in its morphology. We devised a novel method using an algorithm that extracts and tracks multiple feature points in ultrasound images to automatically measure the MTJ that moves during muscle contraction. Instead of using a single reference image, multiple feature points are used to improve the tracking performance during the deformation of the MTJ. Subsequently, we experimentally evaluated the usefulness of this method. Tests were conducted on 20 healthy participants performing isometric maximal contractions, and ultrasound echo images of the medial gastrocnemius and Achilles tendon junctions were recorded. MTJ excursion was calculated using the developed multiple feature point algorithm and two conventional methods-multi-updating template-matching and modified Lucas-Kanade (LK)-based on automatic and manual analyses. The root mean square error (RMSE) was used to compare the results. The intraclass correlation coefficient (ICC) was used to evaluate the repeatability among examiners. RMSE was 1.57 ± 0.62 for the proposed algorithm and 2.18 ± 0.89 and 1.84 ± 1.13 for the conventional methods. The Bland-Altman plot showed that the proposed method exhibited a lower 95% confidence interval than the two conventional methods. Thus, the proposed algorithm had the smallest error. Furthermore, the ICC values were 0.96, 0.40, and 0.86 for the proposed algorithm, multi-updating template-matching, and the modified LK method, respectively. When tracking an MTJ excursion that flexibly changes its shape, the use of multiple feature points provides robust results and achieves tracking that approximates the manual analysis results.

Reduced Achilles tendon stiffness in aging associates with higher metabolic cost of walking

The mechanisms responsible for increased metabolic cost of walking in older adults are poorly understood. We recently proposed a theoretical premise by which age-related reductions in Achilles tendon stiffness (kAT) can disrupt the neuromechanics of calf muscle force production and contribute to faster rates of oxygen consumption during walking. The purpose of this study was to objectively evaluate this premise. We quantified kAT at a range of matched relative activations prescribed using electromyographic biofeedback and walking metabolic cost and ankle joint biomechanics in a group of 15 younger (age: 23 ± 4 yr) and 15 older (age: 72 ± 5 yr) adults. Older adults averaged 44% lower kAT than younger adults at matched triceps surae activations during isokinetic dorsiflexion tasks on a dynamometer (P = 0.046). Older adults also walked with a 17% higher net metabolic power (P = 0.017) but indistinguishable peak Achilles tendon forces than younger adults. Thus, data implicate altered tendon length-tension relations with age more than differences in the operating region of those length-tension relations between younger and older adults. In addition, we discovered empirical evidence that lesser kAT-likely due to the shorter muscle lengths and thus higher relative activations it imposes-was positively correlated with higher net metabolic power during walking (r = -0.365, P = 0.048). These results pave the way for interventions focused on restoring ankle muscle-tendon unit structural stiffness to improve walking energetics in aging.NEW & NOTEWORTHY This study provides the first empirical evidence to our knowledge that age-related decreases in kAT exact a potentially significant metabolic penalty during walking. These results pave the way for interventions focused on restoring ankle muscle-tendon unit structural stiffness to improve walking energetics in aging.

Reduced Achilles tendon stiffness in aging persists at matched activations and associates with higher metabolic cost of walking

The mechanisms responsible for increased walking metabolic cost among older adults are poorly understood. We recently proposed a theoretical premise by which age-related reductions in Achilles tendon stiffness (k AT ) can disrupt the neuromechanics of calf muscle behavior and contribute to faster rates of oxygen consumption during walking. The purpose of this study was to objectively evaluate this premise. We quantified k AT at a range of matched activations prescribed using electromyographic biofeedback and walking metabolic cost in a group of 15 younger (age: 23±4 yrs) and 15 older adults (age: 72±5 yrs). Older adults averaged 44% less k AT than younger adults at matched triceps surae activations (p=0.046). This effect appeared to arise not only from altered tendon length-tension relations with age, but also from differences in the operating region of those length-tension relations between younger and older adults. Older adults also walked with a 17% higher net metabolic power than younger adults (p=0.017). In addition, we discovered empirical evidence that lesser k AT exacts a metabolic penalty and was positively correlated with higher net metabolic power during walking (r=-0.365, p=0.048). These results pave the way for interventions focused on restoring ankle muscle-tendon unit structural stiffness to improve walking energetics in aging.

A comparison of point-tracking algorithms in ultrasound videos from the upper limb

Tracking points in ultrasound (US) videos can be especially useful to characterize tissues in motion. Tracking algorithms that analyze successive video frames, such as variations of Optical Flow and Lucas-Kanade (LK), exploit frame-to-frame temporal information to track regions of interest. In contrast, convolutional neural-network (CNN) models process each video frame independently of neighboring frames. In this paper, we show that frame-to-frame trackers accumulate error over time. We propose three interpolation-like methods to combat error accumulation and show that all three methods reduce tracking errors in frame-to-frame trackers. On the neural-network end, we show that a CNN-based tracker, DeepLabCut (DLC), outperforms all four frame-to-frame trackers when tracking tissues in motion. DLC is more accurate than the frame-to-frame trackers and less sensitive to variations in types of tissue movement. The only caveat found with DLC comes from its non-temporal tracking strategy, leading to jitter between consecutive frames. Overall, when tracking points in videos of moving tissue, we recommend using DLC when prioritizing accuracy and robustness across movements in videos, and using LK with the proposed error-correction methods for small movements when tracking jitter is unacceptable.

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.

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/.

A sound approach to improving exoskeletons and exosuits

[Figure: see text].