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.

Kinematic Analysis of a Drinking Task in Chronic Hemiparetic Patients Using Features Analysis and Statistical Parametric Mapping.

OBJECTIVE: To compare sitting posture and movement strategies between chronic hemiparetic and healthy subjects while performing a drinking task, using statistical parametric mapping (SPM) and feature analysis. DESIGN: Cross-sectional study. SETTING: A university physical therapy department. PARTICIPANTS: Participants (N=26) consisted of chronic hemiparetic (n=13) and healthy individuals (n=13) matched for sex and age. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: The drinking task was divided into phases: reaching, transporting the glass to mouth, transporting the glass to table, and returning to initial position. An SPM 2-sample t test was used to compare the entire kinematic waveforms of different joint angles (trunk, scapulothoracic, humerothoracic, elbow). Joint angles at the beginning and end of the motion, movement time, peak velocity timing, trajectory deviation, normalized integrated jerk, and range of motion were extracted from the motion data. Group differences for these parameters were analyzed using independent t tests. RESULTS: At the static posture and beginning of the reaching phase, patients showed a shoulder position more deviated from the midline and externally rotated with increased scapula protraction, medial rotation, anterior tilting, trunk anterior flexion and inclination to the paretic side. Altered spatiotemporal variables throughout the task were found in all phases, except for the returning phase. Patients returned to a similar posture as the task onset, except for the scapula, which was normalized after the reaching phase. CONCLUSIONS: Chronic hemiparetic subjects showed more deviations in the proximal joints during seated posture and reaching. However, the scapular movement drew nearer to the healthy individuals' patterns after the first phase, showing an interesting point to consider in rehabilitation programs.

Inter- and intrarater clinician agreement on joint motion patterns during gait in children with cerebral palsy.

AIM: This study aimed to quantify the inter- and intrarater clinician agreement on joint motion patterns in children with spastic cerebral palsy (CP), which were recently specified by a Delphi consensus study. It also examined whether experience with three-dimensional gait analysis (3DGA) is a prerequisite for using the patterns. METHOD: The experimental group consisted of 82 patients with CP (57 males, 25 females; uni-/bilateral CP [n=27/55]; Gross Motor Function Classification System levels I to III; mean age 9y 5mo [range 4y-18y]). Thirty-two clinicians were split into 'experienced' and 'inexperienced' rater groups. Each rater was asked to classify 3DGA reports of 27 or 28 patients twice. Inter- and intrarater agreement on 49 joint motion patterns was estimated using percentage of agreement and kappa statistics. RESULTS: Twenty-eight raters completed both classification rounds. Intrarater agreement was 'substantial' to 'almost perfect' for all joints (0.64<ĸ<0.91). Interrater agreement reached similar results (0.63<ĸ<0.86), except for the knee patterns during stance (ĸ=0.49, 'moderate agreement'). Experienced raters performed significantly better on patterns of the knee during stance and ankle during swing. INTERPRETATION: Apart from some specific knee patterns during stance and ankle patterns during swing, the results suggested that clinicians could use predefined joint motion patterns in CP with good confidence, even in case of limited experience with 3DGA.

A subject-specific method to measure dynamic spinal alignment in adult spinal deformity.

BACKGROUND CONTEXT: Two-dimensional static radiography currently forms the golden standard in spinal alignment measurement in adult spinal deformity (ASD). However, these static measurements offer no information on dynamic spinal behavior. To fully understand the functionality and compensation strategies of ASD patients, tools to assess dynamic spinal alignment are needed. PURPOSE: Therefore, the aim of this study was to introduce, validate and assess the reliability of a new kinematic model to measure dynamic spinal parameters in ASD based on a polynomial function, taking into account the subject-specific anatomy. STUDY DESIGN: Validation and reliability study OUTCOME MEASURES: Radiographic parameters, spinal kinematics and range of motion (ROM), Scoliosis Research Society Outcome Questionnaire (SRS-22), Core Outcome Measures Index (COMI). METHODS: Spinal alignment of 23 ASD patients and 18 controls was measured using both x-rays and motion capture. Marker positions were corrected to the underlying anatomy and a polynomial function was fitted through these corrected marker positions. By comparing the polynomial method to x-ray measurements concurrent validity was assessed. Test-retest, inter- and intrarater reliability during standing and sit-to-stand (STS) were assessed on a subsample of eight ASD patients and eight controls. RESULTS: The results showed good to excellent correlations (r>0.75) between almost all x-ray and anatomy-corrected polynomial parameters. Anatomy correction consistently led to better correlations than no correction. Intraclass correlation coefficients for the polynomial method were good to excellent (>0.75) between sessions and between and within raters and comparable or even better than radiographic measurements. Also, during STS reliability was excellent. Fair to moderate correlations were found between spinal ROM during STS and quality of life, measured with SRS-22 and COMI. CONCLUSIONS: The results of this study indicate the polynomial method, with subject-specific anatomy correction, can measure spinal alignment in a valid and reliable way using motion capture in both healthy and deformed spines. This method makes it possible to extend evaluation in ASD from mainly static, by means of x-ray measurements, to dynamic and functional assessments. CLINICAL SIGNIFICANCE: Eventually, this newly obtained dynamic spinal alignment information might lead to new insights in clinical decision-making and new treatment strategies, based and oriented on dynamic parameters and functionality.

Efficient image based method using water-filled balloons for improving probe spatial calibration in 3D freehand ultrasonography.

The ultrasound (US) probe spatial calibration is a key prerequisite for enabling the use of the 3D freehand US technique. Several methods have been proposed for achieving an accurate and precise calibration, although these methods still require specialised equipment. This equipment is often not available in research or clinical facilities. Therefore, the present investigation aimed to propose an efficient US probe calibration method that is accessible in terms of cost, easy to apply and capable of achieving results suitable for clinical applications. The data acquisition was carried out by performing two perpendicular US sweeps over water filled balloon phantoms. The data analysis was carried out by computing the similarity measures between 2D images from the first sweep and the corresponding images of the 3D reconstruction of the second sweep. These measures were maximized by using the Nelder-Mead algorithm, to find the optimal solution for the calibration parameters. The calibration results were evaluated in terms of accuracy and precision by comparing known phantom geometries with those extracted from the US images. The accuracy and the precision after applying the calibration method were improved. By using the parameters obtained from the plane phantom method as initialization of the calibration parameters, the accuracy and the precision in the best scenario was 0.4 mm and 1.5 mm, respectively. These results were in line with the methods requiring specialised equipment. However, the applied method was unable to consistently produce this level of accuracy and precision. The calibration parameters were also tested in a musculoskeletal application, revealing sufficient matching of the relevant anatomical features when multiple US sweeps are combined in a 3D reconstruction. To improve the current results and increase the reproducibility of this research, the developed software is made available.

An innovative solution to reduce muscle deformation during ultrasonography data collection.

BACKGROUND: 3D freehand ultrasound enables the creation of volumetric data. The acquisition of morphological features, such as muscle volume, is influenced by the variations in force applied to the skin with the ultrasound probe. To minimise the deformations, a concave-shaped plastic mount combined with a custom-shaped gel pad was developed for the ultrasound head, named Portico. This study analyses to what extent the Portico reduces muscle deformation and corresponding errors in estimating muscle volume. METHOD: Twenty medial gastrocnemius (MG) muscles were assessed (10 from typically developing children; 10 from children with spastic cerebral palsy). Two repetitions were acquired in each of the following approaches: (1) with the lower leg submerged in a water tank as a non-deformed reference; (2) probe-on-skin (PoS) as the conventional approach and (3) the newly introduced Portico. PoS and Portico data were registered with respect to the ones corresponding in a water tank. An in-house software package (Py3DFreeHandUS) was used to process the data and MG volume was estimated using MeVisLab. The minimal detectable change (MDC) was calculated. RESULTS: With respect to the PoS approach, the Portico reduced muscle deformation by 46%. For both the typically developing and spastic cerebral palsy cohorts, lower MDCs were found when using the Portico. DISCUSSION: Despite the improvements, the Portico did not yield statistically more reliable MG volume estimations than the traditional PoS approach. Further improvement can be attained by optimising the fit between the gel pad and the curvature of the limb, using a larger choice of Portico geometries.

Hip movement pathomechanics of patients with hip osteoarthritis aim at reducing hip joint loading on the osteoarthritic side.

This study aims at defining gait pathomechanics in patients with hip osteoarthritis (OA) and their effect on hip joint loading by combining analyses of hip kinematics, kinetics and contact forces during gait. Twenty patients with hip OA and 17 healthy volunteers matched for age and BMI performed three-dimensional gait analysis. Hip OA level was evaluated based on plane radiographs using the Tönnis classification. Hip joint kinematics, kinetics as well as hip contact forces were calculated. Waveforms were time normalized and compared between groups using statistical parametric mapping analysis. Patients walked with reduced hip adduction angle and reduced hip abduction and external rotation moments. The work generated by the hip abductors during the stance phase of gait was largely decreased. These changes resulted in a decrease and a more vertical and anterior orientation of the hip contact forces compared to healthy controls. This study documents alterations in hip kinematics and kinetics resulting in decreased hip loading in patients with hip OA. The results suggested that patients altered their gait to increase medio-lateral stability, thereby decreasing demand on the hip abductors. These findings support discharge of abductor muscles that may bear clinical relevance of tailored rehabilitation targeting hip abductor muscles strengthening and gait retraining.

Three-dimensional kinematics of the scapula and trunk, and associated scapular muscle timing in individuals with stroke.

Poor scapulothoracic control is a risk for developing shoulder pathology, but has received little attention so far in individuals with stroke (IwS). Trunk and scapular kinematics and surface muscle activity were measured in 15 healthy controls and 18 IwS during a low and high forward flexion (FF). Group-differences in trunk and scapular kinematics were assessed during low and high FF using a t-test (independent samples). Differences in muscle onset and offset time relative to movement start (both FF tasks) were determined using a mixed model taking into account the different groups and muscles. Recruitment patterns per group and task were described based on significant differences between muscles. In IwS, earlier lower trapezius and late infraspinatus offset were found during low FF, as well as a later onset and earlier offset of serratus anterior. For low FF, significantly more trunk axial rotation was found in IwS during both elevation and lowering. During high FF, IwS showed significantly less scapular posterior tilt during elevation and more scapular lateral rotation during lowering. IwS demonstrated adaptive muscle timing with earlier initiation and late inactivation of lower trapezius and infraspinatus, possibly to compensate for a late activation and early deactivation of the serratus anterior and to establish as such the correct pattern of scapulothoracic movement.

The reliability and validity of a clinical 3D freehand ultrasound system.

BACKGROUND AND OBJECTIVE: Acquiring large anatomical volumes in a feasible manner is useful for clinical decision-making. A relatively new technique called 3D freehand ultrasonography is capable of this by combining a conventional 2D ultrasonography system. Currently, a thorough analysis of this technique is lacking, as the analyses are dependent on the software implementation details and the choice of measurement systems. Therefore this study starts by making this implementation available under the form of an open-source software library to perform 3D freehand ultrasonography. Following that, reliability and validity analyses of extracting volumes and lengths will be carried out using two independent motion-tracking systems. METHODS: A PC-based ultrasonography device and two optical motion-tracking systems were used for data acquisition. An in-house software library called Py3DFreeHandUS was developed to reconstruct (off-line) the corresponding data into one 3D data set. Reliability and validity analyses of the entire experimental set-up were performed by estimating the volumes and lengths of ground truth objects. Ten water-filled balloons and six cross-wires were used. Repeat measurements were also performed by two experienced operators. RESULTS: The software library Py3DFreeHandUS is available online, along with the relevant documentation. The reliability analyses showed high intra- and inter-operator intra-class correlation coefficient results for both volumes and lengths. The accuracy analysis revealed a discrepancy in all cases of around 3%, which corresponded to 3 ml and 1 mm for volume and length measurements, respectively. Similar results were found for both of the motion-tracking systems. CONCLUSIONS: The undertaken analyses for estimating volume and lengths acquired with 3D freehand ultrasonography demonstrated reliable design measurements and suitable performance for applications that do not require sub-mm and -ml accuracy.

Evaluation of stair motion contributes to new insights into hip osteoarthritis-related motion pathomechanics.

Stair motion in the presence of hip osteoarthritis (OA) has received less attention than level walking. Its more strenuous aspect may shed the light on different locomotor strategies when compared to walking. We, therefore, aimed to define stair motion features associated to hip OA and to evaluate whether these specific features would differ from level walking and better characterize the hip pathological condition. Principal component and linear discriminant analyses were, respectively, used as data reduction and classification techniques. Our study highlighted that most of stair motion features associated to hip OA were similar to the ones of walking. Stair descent presented with the lowest misclassification error rate, ranging from 12% to 19% (estimated by cross-validation). But, features that may be considered as a mechanism to reduce demand on the hip abductors were found to be more important in the stair ascent condition. This was reflected by both, greater importance in the classification rule and variance compared with walking, that is, decreased hip internal rotation moment at mid-stance (72.50% vs. 57.63%) and increased trunk lateroflexion toward affected side (56.43% vs. 29.37%). This study emphasized the importance of investigating stair motion in hip osteoarthritic population by highlighting specific locomotor strategies.

Biomechanical gait features associated with hip osteoarthritis: Towards a better definition of clinical hallmarks.

Critical appraisal of the literature highlights that the discriminative power of gait-related features in patients with hip osteoarthritis (OA) has not been fully explored. We aimed to reduce the number of gait-related features and define the most discriminative ones comparing the three-dimensional gait analysis of 20 patients with hip osteoarthritis (OA) with those of 17 healthy peers. First, principal component analysis was used to reduce the high-dimensional gait data into a reduced set of interpretable variables for further analysis, including tests for group differences. These differences were indicative for the selection of the top 10 variables to be included into linear discriminant analysis models (LDA). Our findings demonstrated the successful data reduction of hip osteoarthritic-related gait features with a high discriminatory power. The combination of the top variables into LDA models clearly separated groups, with a maximum misclassification error rate of 19%, estimated by cross-validation. Decreased hip/knee extension, hip flexion and internal rotation moment were gait features with the highest discriminatory power. This study listed the most clinically relevant gait features characteristics of hip OA. Moreover, it will help clinicians and physiotherapists understand the movement pathomechanics related to hip OA useful in the management and design of rehabilitation intervention.

The relation between spasticity and muscle behavior during the swing phase of gait in children with cerebral palsy.

There is much debate about how spasticity contributes to the movement abnormalities seen in children with spastic cerebral palsy (CP). This study explored the relation between stretch reflex characteristics in passive muscles and markers of spasticity during gait. Twenty-four children with CP underwent 3D gait analysis at three walking velocity conditions (self-selected, faster and fastest). The gastrocnemius (GAS) and medial hamstrings (MEHs) were assessed at rest using an instrumented spasticity assessment that determined the stretch-reflex threshold, expressed in terms of muscle lengthening velocity. Muscle activation was quantified with root mean square electromyography (RMS-EMG) during passive muscle stretch and during the muscle lengthening periods in the swing phase of gait. Parameters from passive stretch were compared to those from gait analysis. In about half the children, GAS peak muscle lengthening velocity during the swing phase of gait did not exceed its stretch reflex threshold. In contrast, in the MEHs the threshold was always exceeded. In the GAS, stretch reflex thresholds were positively correlated to peak muscle lengthening velocity during the swing phase of gait at the faster (r = 0.46) and fastest (r = 0.54) walking conditions. In the MEHs, a similar relation was found, but only at the faster walking condition (r = 0.43). RMS-EMG during passive stretch showed moderate correlations to RMS-EMG during the swing phase of gait in the GAS (r = 0.46-0.56) and good correlations in the MEHs (r = 0.69-0.77) at all walking conditions. RMS-EMG during passive stretch showed no correlations to peak muscle lengthening velocity during gait. We conclude that a reduced stretch reflex threshold in the GAS and MEHs constrains peak muscle lengthening velocity during gait in children with CP. With increasing walking velocity, this constraint is more marked in the GAS, but not in the MEHs. Hyper-activation of stretch reflexes during passive stretch is related to muscle activation during the swing phase of gait, but has a limited contribution to reduced muscle lengthening velocity during swing. Larger studies are required to confirm these results, and to investigate the contribution of other impairments such as passive stiffness and weakness to reduced muscle lengthening velocity during the swing phase of gait.

Altered trunk movements during gait in children with spastic diplegia: compensatory or underlying trunk control deficit?

Altered trunk movements during gait in children with CP are considered compensatory due to lower limb impairments, although scientific evidence for this assumption has not yet been provided. This study aimed to study the functional relation between trunk and lower limb movement deficits during gait in children with spastic diplegia. Therefore, the relationship between trunk control in sitting, and trunk and lower limb movements during gait was explored in 20 children with spastic diplegia (age 9.2 ± 3 yrs; GMFCS level I n=10, level II n=10). Trunk control in sitting was assessed with the Trunk Control Measurement Scale (TCMS), a clinical measure that reflects the presence of an underlying trunk control deficit. Trunk movements during gait were measured with a recently developed trunk model including the pelvis, thorax, head, shoulder line and spine. Lower limb movements were assessed with the Plug-in-Gait model (Vicon(®)). Range of motion (ROM) of the different trunk segments was calculated, as well as the Trunk Profile Score (TPS) and Trunk Variable Scores (TVSs). Similarly, the Gait Profile Score (GPS) and Gait Variable Scores (GVSs) were calculated to describe altered lower limb movements during gait. Correlation analyses were performed between the presence of impaired trunk control in sitting (TCMS) and altered trunk movements during gait (ROM, TPS/TVSs) and between these altered trunk movements and lower limb movements (GPS/GVSs) during gait. A poorer performance on the TCMS correlated with increased ROM and TPS/TVSs, particularly for the thorax, indicating the presence of an underlying trunk control deficit. No significant correlation was found between the TPS and GPS, suggesting that overall trunk and lower limb movement deficits were not strongly associated. Only few correlations between specific lower limb deficits (GVSs for hip ab/adduction, knee flexion/extension and ankle flexion/extension) and TVSs for thorax lateral bending and rotation were found. This study provided first evidence that the altered trunk movements observed during gait should not be solely considered compensatory due to lower limb impairments, but that these may also partially reflect an underlying trunk control deficit. A better understanding of underlying trunk control deficits in children with CP may facilitate targeted therapy planning and ultimately can optimize a child's functionality.

Gaitography applied to prosthetic walking.

During walking on an instrumented treadmill with an embedded force platform or grid of pressure sensors, center-of-pressure (COP) trajectories exhibit a characteristic butterfly-like shape, reflecting the medio-lateral and anterior-posterior weight shifts associated with alternating steps. We define "gaitography" as the analysis of such COP trajectories during walking (the "gaitograms"). It is currently unknown, however, if gaitography can be employed to characterize pathological gait, such as lateralized gait impairments. We therefore registered gaitograms for a heterogeneous sample of persons with a trans-femoral and trans-tibial amputation during treadmill walking at a self-selected comfortable speed. We found that gaitograms directly visualize between-person differences in prosthetic gait in terms of step width and the relative duration of prosthetic and non-prosthetic single-support stance phases. We further demonstrated that one should not only focus on the gaitogram's shape but also on the time evolution along that shape, given that the COP evolves much slower in the single-support phase than in the double-support phase. Finally, commonly used temporal and spatial prosthetic gait characteristics were derived, revealing both individual and systematic differences in prosthetic and non-prosthetic step lengths, step times, swing times, and double-support durations. Because gaitograms can be rapidly collected in an unobtrusive and markerless manner over multiple gait cycles without constraining foot placement, clinical application of gaitography seems both expedient and appealing. Studies examining the repeatability of gaitograms and evaluating gaitography-based gait characteristics against a gold standard with known validity and reliability are required before gaitography can be clinically applied.