Advertisement

Variations of Marker Sets and Models for Standard Gait Analysis

  • Felix Stief
Reference work entry

Abstract

A variety of different approaches is used in 3D clinical gait analysis. This chapter provides an overview of common terms, different marker sets, underlying anatomical models, as well as a fundamental understanding of measurement techniques commonly used in clinical gait analysis and the consideration of possible errors associated with these different techniques. Besides the different marker sets, two main approaches can be used to quantify marker-based joint angles: a prediction approach based on regression equations and a functional approach. The prediction approach uses anatomical assumptions and anthropometric reference data to define the locations of joint centers/axes relative to specific anatomical landmarks. In the functional approach, joint centers are determined via optimization of marker movement. The accuracy of determining skeletal kinematics is limited by ambiguity in landmark identification and soft-tissue artifacts. When the intersubject variability of control data becomes greater than the expected change due to pathology, the clinical usefulness of the data becomes doubtful. To allow a practical interpretation of a comparison of approaches, differences and the measurement error should be quantified in the unit of interest (i.e., degree or percent). The highest reliability indices occurred in the hip and knee in the sagittal plane, with lowest reliability and highest errors for hip and knee rotation in the transverse plane. In addition, knowledge about sources of errors should be known before the approach is applied in practice.

Keywords

Marker sets Anatomical markers Technical markers Clusters Modeling Segment definition Prediction approach Functional approach Regression equations Conventional Gait Model Measurement error Soft-tissue artifacts Reliability Accuracy 

Introduction

This chapter provides an overview of common terms, different marker sets, underlying anatomical models, as well as a fundamental understanding of measurement techniques commonly used in clinical gait analysis and the consideration of possible errors associated with these different techniques.

It is possible for a clinician or physician to subjectively study gait; however, the value and repeatability of this type of assessment is questionable due to poor inter- and intra-tester reliability. For instance, it is impossible for one individual to study, by observation alone, the movement pattern of all the main joints involved during an activity like walking simultaneously. Therefore, skeletal movements in three dimensions during gait are typically recorded using markers placed on the surface of the skin on various anatomical landmarks to represent body segments. The marker-based analysis of human movement helps to better understand normal and pathological function and results in a detailed and objective clinical assessment of therapeutic and surgical interventions.

A variety of different anatomical models and marker sets were used for clinical gait analysis. While a certain amount of standardization could be established in recent years for the marker placement on anatomical points and the definition for most of the rigid body segments (pelvic, thigh, shank, foot), protocols differ in the underlying biomechanical model, the definition of joint centers and axes, and the number of markers used. These differences have an effect on the outcome measures (e.g., joint angles and moments). The main focus of this chapter is to demonstrate the impact of marker sets and joint angle calculations on gait analysis results.

State of the Art

In Standard Gait Analysis either passive or active markers are being used. Passive markers for camera-based systems are generally made of a retroreflective material. This material is used to reflect light emitted from around the camera back to the camera lens. Some camera-based systems use a stroboscopic light, while others use light from synchronized infrared light-emitting diodes mounted around the camera lens.

In contrast, active markers produce light at a given frequency, so these systems do not require illumination, and, as such, the markers are more easily identified and tracked (Chiari et al. 2005). These light-emitting diodes (LED) are attached to a body segment in the same way as passive markers, but with the addition of a power source and a control unit for each LED. Active markers can have their own specific frequency which allows them to be automatically detected. This leads to very stable real-time three-dimensional motion tracking as no markers can be misidentified as adjacent markers. Regardless of whether passive or active, the use of markers should not significantly modify the movement pattern being measured.

Anatomical and Technical Markers

Anatomical markers are used to set up the segment reference frame. This is generally done during a static trial with the subject standing still. Anatomical markers may be attached on bony landmarks directly to the skin or fixed to a pointer. These markers are not required for the dynamic trials as long as at least three fixed points are available on each segment.

Technical markers have no specific location and are chosen purely to meet the other requirements above. Additional technical markers can be used to create a technical coordinate system from data collected in a static calibration trial during which both anatomical and technical markers are present. In subsequent dynamic trials, absent anatomical markers can be expressed in relation to the technical coordinate system. Technical markers can also be used to avoid areas of adipose tissue in obese patients, to accommodate walking aids, or to replace markers that are obscured dynamically. Two approaches are commonly used. Technical markers may be used to replace only those anatomical markers that cannot be used dynamically. In this case, the majority of anatomical markers remain in place for the walking trials. Alternatively, clusters of technical markers attached to a plate (see “Marker Clusters” below) may be used to provide all the dynamic information needed. Anatomical markers are then only used for the static trial to allow segment reconstruction.

Marker Clusters

Other techniques for minimizing soft-tissue artifacts and in order to reduce intersubject variability are marker clusters (an array of markers) (Cappozzo et al. 1995). They must be in place during the static anatomical calibration. The exact placement of the clusters is less reliant as this technique uses the relative positions to the anatomical landmarks used in the static calibration. The purpose is to define the plane of each segment with 3–5 markers and then track its movement through the basic reference planes. Clusters can be directly attached to the skin or mounted on rigid fixtures (Fig. 1), which are dependent upon the anatomy, the activity, and the nature of the analysis.
Fig. 1

Rigid marker cluster with four retroreflective markers

In a rigid body or cluster, the distance between any two points within the body or cluster does not change. In general, tracking of marker clusters helps to reduce noise within the motion signal and improve accuracy of kinematic data. When the markers are fixed to rigid plates, the markers never move independently with deformation of the skin. It has been shown that the absolute and relative variance in out-of-sagittal plane rotations tended to be higher for the Conventional Gait Model ( “The Conventional Gait Model - Success and Limitations”) compared with a cluster technique (Duffell et al. 2014) and that a cluster marker set overcomes a number of theoretical limitations compared to the conventional set (Collins et al. 2009) when both models were compared simultaneously. Much work has been carried out determining the optimal configuration of marker clusters, and it is now widely accepted that a rigid shell with a cluster of four markers is a good practical solution (Cappozzo et al. 1997; Manal et al. 2000). However, when the cluster markers were fixed to a rigid plate, these methods were not able to address absolute errors and can still result in inaccurate identification of joint centers (Holden and Stanhope 1998). Although an extended version of this method has reported improvements in estimation of the position of the underlying bones (Alexander and Andriacchi 2001), it can only model skin deformations and has limited use in clinical applications due to the number of additional markers required.

The Definition of a Segment

In general, three markers are needed to fix a rigid body in space. When using motion capture to define the pelvic segment ( “The Conventional Gait Model - Success and Limitations”) and measure pelvic motion, the International Society of Biomechanics (ISB) recommends the pelvic anatomical coordinate system be defined by surface markers placed on the right and left anterior superior iliac spines (ASISs) and on the right and left posterior superior iliac spines (PSISs). The pelvic anatomical coordinate system can be described as the origin at the midpoint between the right ASIS and the left ASIS; the Z-axis points from the origin to the right ASIS; the X-axis lies in the plane defined by the right ASIS, left ASIS, and the midpoint of the right PSIS and left PSIS markers and points ventrally orthogonal to the Z-axis; and the Y-axis is orthogonal to these two axes (Wu et al. 2002). These markers would ideally be used to track the pelvis during gait or clinical assessment protocols that involve movement. However, situations in which the ASIS or PSIS markers are obscured from view require that alternative technical marker sets are used. Occlusion of the ASIS markers could be as a result of soft tissue around the anterior abdomen (a common issue in overweight and obese subjects), arm movement, or activities that require high degrees of hip and trunk flexion, such as running, stair climbing, or level walking. It has been shown that pelvic models that include markers placed on the ASISs and the iliac crests (ICs), and PSISs and ICs, are suitable alternatives to the standard pelvic model (ASISs and PSISs) for tracking pelvic motion during gait (Bruno and Barden 2015). Alternatively, the use of a rigid cluster of three orthogonal markers as technical markers attached to the sacrum can be used (Borhani et al. 2013). Using the calibrated anatomical system technique (Benedetti et al. 1998; Cappello et al. 2005) allows the position of ASIS defined relative to the cluster in a static trial, and then during dynamic trial, the position of the ASIS is linked to the cluster and thus affected by the same skin movement artifact that affects the cluster. Another alternative to solve skin artifacts is to use the right and left hip joint centers described in the technical coordinate system of the right and left thighs, together with the right PSIS and left PSIS markers, as technical markers for tracking the pelvis movement (Kisho Fukuchi et al. 2010).

Prediction Approach or the Conventional Gait Model

Besides the different marker sets, two main approaches can be used to quantify joint angles: a prediction approach based on regression equations and a functional approach. The prediction approach uses anatomical assumptions and anthropometric reference data to define the locations of joint centers/axes relative to specific anatomical landmarks (Isman and Inman 1969; Weidow et al. 2006). In the functional approach, joint centers are determined via optimization of marker movement. The advantages and disadvantages of both approaches were described below in detail.

Most biomechanical analysis systems use regression equations based on predictive methods to calculate joint centers. Kadaba et al. (1989), Davis III et al. (1991), and Vaughan et al. (1992) provided detailed descriptions of a marker-based system to calculate joint centers in the lower extremities. This marker setup has become one of the most commonly used models in gait analysis. It is referred to as Helen Hayes Hospital marker setup, and the regression equations are referred to as the Plug-in-Gait (PiG) model or the Conventional Gait Model ( “The Conventional Gait Model - Success and Limitations”).

Functional Approach

In general, technical marker sets require data capture in a static standing trial to determine rotation values (offsets) to place these markers into the anatomical coordinate system. If a marker does, for instance, not accurately represent the position of the hip during standing data capture, the technical markers will not be placed into the correct anatomical plane for the dynamic trial. This is particularly problematic if the static and dynamic positions of the hip vary from one another. It has been shown that static standing posture greatly affected the dynamic hip rotation kinematics when using a thigh wand in the typical clinical gait analysis process for the Conventional Gait Model (McMulkin and Gordon 2009). Therefore, if a thigh wand is to be used in clinical practice, it is necessary that patients stand in a hip rotation posture that is equivalent to hip rotation position used in gait. This can be very difficult because it requires clinicians to have a priori knowledge of the gait hip rotation before testing. Also, patients may use different strategies in static standing than with walking posture. One way of addressing this issue is to use functional joint center techniques (Ehrig et al. 2006; Leardini et al. 1999; Schwartz and Rozumalski 2005). This functional approach is considered functional due to the calculation of subject-specific joint centers/axes by using specific movement data of adjacent segments derived from basic motion tasks. With a focus on assessing motion patterns in a subject-specific manner, functional methods rely on the relative motion between the marker clusters of neighboring segments to identify joint centers and axes (Cappozzo et al. 1997; Ehrig et al. 2006). Previously developed functional methods have been demonstrated to be precise (Ehrig et al. 2006; Kornaropoulos et al. 2010; Kratzenstein et al. 2012) as well as rapid and robust (Schwartz and Rozumalski 2005) in estimating joint centers. Nevertheless, in many patient groups, functional calibration has been reported to be difficult (Sangeux et al. 2011) due to the fact that the range of motion (ROM) of affected joints is restricted. In addition, functional methods have not been able to demonstrate consistent advantages over more traditional regression-based approaches (Assi et al. 2016; Bell et al. 1990; Davis III et al. 1991; Harrington et al. 2007), possibly due to issues of marker placement and the nonlinear distribution of soft-tissue artifacts across a segment (Gao and Zheng 2008; Stagni et al. 2005). Kratzenstein et al. (2012) present an approach for understanding the contribution of different regions of marker attachment on the thigh toward the precise determination of the hip joint center. This working group used a combination of established approaches (Taylor et al. 2010) to reduce skin marker artifacts (Taylor et al. 2005), determine joint centers of rotation (Ehrig et al. 2006), and quantify the weighting of each of a large number of markers (Heller et al. 2011) attached to the thigh. Consequently, markers that are suboptimally located and therefore strongly affected by soft-tissue artifacts are assigned a lower weighting compared to markers that follow spherical trajectories around the joint. Based on these methods, six regions of high importance were determined that produced a symmetrical center of rotation estimation (Ehrig et al. 2011) almost as low as using a marker set that covered the entire thigh. Such approaches could be used to optimize marker sets for targeting more accurate and robust motion capture for aiding in clinical diagnosis and improving the reliability of longitudinal studies.

Impact of Marker Set and Joint Angle Calculation on Gait Analysis Results

Errors Involved with Marker Placement and Soft-Tissue Artifacts

The accuracy of determining skeletal kinematics is limited by ambiguity in landmark identification and soft-tissue artifacts that is the motion of markers over the underlying bones due to skin elasticity, muscle contraction, or synchronous shifting of the soft tissues (Leardini et al. 2005; Taylor et al. 2005). Generally, two types of errors are referred to soft-tissue artifacts. Relative errors are defined as the relative movement between two or more markers that define a rigid segment. Absolute errors are defined as the movement of a marker with respect to the bony landmark it is representing (Richards 2008). Relative and absolute errors are often caused by movement of the soft tissue on which the markers are placed (Cappozzo et al. 1996). The magnitude of these errors has been studied by using pins secured directly into the bone and comparing the data collected from skin-mounted markers to markers attached to bone pins. These data give a direct measure of soft-tissue movement with respect to the skeletal system (Cappozzo 1991; Cappozzo et al. 1996; Reinschmidt et al. 1997a, b). However, the applicability of this method is limited due to their invasive nature. The amount and the effects of soft-tissue artifacts from skin markers are discussed controversial with relative skin to bone marker movements in the range of 3 mm up to 40 mm, dependent upon the specific body segment and soft-tissue coverage (Cappozzo et al. 1996; Holden et al. 1997; Manal et al. 2000, 2003; Reinschmidt et al. 1997b). Differences can be accounted for by variation in marker placement and configuration, differences in techniques, intersubject differences, and differences in the task performed (Leardini et al. 2005).

Inaccuracies in lower limb motion and in particular knee kinematics are present mainly because of soft-tissue artifacts at the thigh segment (Alexander and Andriacchi 2001; Cappello et al. 1997; Fuller et al. 1997; Leardini et al. 2005; Lucchetti et al. 1998). Conversely, soft-tissue movement on the shank has only a small effect on three-dimensional kinematics and moments at the knee (Holden et al. 1997; Manal et al. 2002). In addition, mainly in the frontal and transverse planes, substantial angular variabilities were noted (Ferrari et al. 2008; Miana et al. 2009) due to the small ROM in these planes compared to sagittal plane movements. This reasoning agrees with the results of Leardini et al. (2005) who assert that angles out-of-sagittal planes should be regarded with much more caution as the soft-tissue artifact produces spurious effects with magnitudes comparable to the amount of motion actually occurring in the joints. In addition, an increase in velocity (for instance, during running) produces an increased variability of the joint centers’ distances and increases the maximum differences between the joint angles when using different protocols (Miana et al. 2009).

Errors Associated with the Regression Equations

Besides soft-tissue artifacts and variability of the marker placement, errors associated with the regression equations used to calculate the joint center locations are also considerable (Harrington et al. 2007; Leardini et al. 1999; Sangeux et al. 2011). Clinically, the definition of the joint center is generally achieved by using palpable anatomical landmarks to define the medial-lateral axis of the joint. From these anatomical landmarks, the center of rotation is generally calculated in one of two ways: through the use of regression equations based on standard radiographic evidence or simply calculated as a percentage offset from the anatomical marker based on some kind of anatomical landmarks (Bell et al. 1990; Cappozzo et al. 1995; Davis III et al. 1991; Kadaba et al. 1989).

The issue of hip joint center (HJC) identification is one that has been covered in much depth, and there are still many debates around this area. The location of this joint center is one of the most difficult anatomic reference points to define. The center of the femoral head is the center of the hip joint and located within the acetabulum on the obliquely aligned and tilted lateral side of the pelvis. Therefore, common approaches have used landmarks on the pelvis as the anatomical reference (Perry and Burnfield 2010). The regression equations in the Conventional Gait Model are based on the HJC regression equations by Davis et al. (1991) and chord functions to predict the knee and the ankle joint centers. The HJC regression equation was based on 25 male subjects and has been validated in later studies (Harrington et al. 2007; Leardini et al. 1999; Sangeux et al. 2011) showing significant errors, which were corrected with new regression equations (Sandau et al. 2015).

In the chord function , the HJC, thigh wand marker, and the epicondyle marker were used to define a plane. The knee joint center (KJC) was then found so that the epicondyle marker was at a half knee diameter distance from the KJC in a direction perpendicular to the line from the HJC to KJC. The ankle joint center (AJC) was predicted in the same way as the knee, where the chord function was used to predict the joint center based on the KJC, the calf wand marker, and the malleolus marker. The chord functions predict the KJC and the AJC with the assumption that the joint centers are lying on the transepicondylar axis and the transmalleolar axis in the frontal plane, respectively. This assumption seems reasonable for the knee (Asano et al. 2005; Most et al. 2004), but to a lesser extent regarding the ankle joint (Lundberg et al. 1989). The exact position of the joint centers influences the joint angles as well as joint angular velocity and acceleration which are part of inverse dynamics. Likewise, the location of segmental center of mass will influence the inverse dynamics calculations via the moment arms acting together with both proximal and distal joint reaction forces.

How to Address the Measurement Error and What is the Extent of This Error?

In general, when addressing the measurement error in marker-based movement analysis, it is helpful to provide an absolute measure of reliability , for instance, the root mean square error or standard error of measurement (SEM). It is thus possible to express the variability in a manner that can be directly related to the measurement itself, in the same measurement units (e.g., degrees). Furthermore, with the transformation of the absolute error into relative error, one can obtain the error expressed as percentage corresponding to the total ROM of the variable to be analyzed. This is of particular importance for the between-plane comparison of the measurement error with different amplitude of the kinematic and kinetic parameters (Stief et al. 2013).

In contrast, the commonly reported intraclass correlation coefficient or coefficient of variation and coefficient of multiple correlations allow limited information, as high coefficient values can result from a low mean value of the variable of interest and thus could hide measurement errors of clinical importance (Luiz and Szklo 2005). Furthermore, expressing data variability as a coefficient results in units that are difficult to interpret clinically (Leardini et al. 2007).

Regarding the literature, kinematic measurement errors of less than 4° and 6° were reported for the intertrial and intersession variability, respectively (Stief et al. 2013). A systematic review from McGinley et al. (2009) identifies that the highest reliability indices occurred in the hip and knee in the sagittal plane, with lowest reliability and highest errors for hip and knee rotation in the transverse plane. Most studies included in this review article providing estimates of data error reported values of less than 5°, with the exception of hip and knee rotation. Fukaya et al. (2013) investigated the interrater reliability of knee movement analyses during the stance phase using a rigid marker set with three attached markers affixed to the thigh and shank. Each of three testers independently attached the infrared reflective markers to four subjects. The SEM values for reliability ranged from 0.68° to 1.13° for flexion-extension, 0.78°–1.60° for external-internal rotation, and 1.43°–3.33° for abduction-adduction. In general, the measurement errors between testers are considered to be greater than the measurement errors between sessions and within testers (Schwartz et al. 2004).

Accuracy for Marker-Based Gait Analysis

The accuracy of body protocols can hardly be assessed in clinical routine since invasive methods such as radiographic imaging (Garling et al. 2007) or bone pins (Taylor et al. 2005) are required in order to provide sufficient access to the skeletal anatomy but are generally not available. Ultrasound assessment of the joint provides one noninvasive opportunity (Sangeux et al. 2011), but assessment of the images can be somewhat subjective. According to Schwartz and Rozumalski (2005), the following indirect indicators of accuracy can be computed instead:
  1. 1.

    Knee varus/valgus ROM during gait: An accurate knee flexion axis alignment minimizes the varus/valgus ROM resulting from cross-talk, that is, one joint rotation (e.g., flexion) being interpreted as another (e.g., adduction or varus) due to axis malalignment (Piazza and Cavanagh 2000).

     
  2. 2.

    Knee flexion/extension ROM during gait: An accurate knee flexion axis alignment maximizes knee flexion/extension ROM by reducing cross-talk.

     

In general, the knee varus/valgus curve can be evaluated for signs of marker misplacement or Knee Alignment Device misalignment. Moreover, it has been shown that for the stable knee joint, the physiological ROM of knee varus/valgus only varies between 5° and 10° (Reinschmidt et al. 1997a). Minimization of the knee joint angle cross talk can therefore be considered to be a valid criterion to evaluate the relative merits of different protocols and marker sets.

Comparison of Marker Sets and Models for Standard Gait Analysis

There is still a variety of different approaches being used in clinical gait analysis. Protocols differ in the underlying biomechanical model, associated marker sets, and data recording and processing. The former defines properties of the modeled joints, the number of involved segments, the definitions of joint centers and axes, the used anatomical and technical reference frames, and the angular decomposition technique to calculate joint angles. Despite apparent differences of the outcome measures derived from different gait protocols (Ferrari et al. 2008; Gorton et al. 2009), specifically for out-of-sagittal plane rotations (Ferrari et al. 2008), data of different studies are compared and interpreted.

Any protocol for movement analysis will only prove useful if it displays adequate reliability (Cappozzo 1984). Moreover, and as stated before, the placement of the markers has considerable influence on the accuracy of gait studies (Gorton et al. 2009). One of the first protocols proposed by Davis et al. (1991), and known as Conventional Gait Model or PiG model, is still used by a vast majority of gait laboratories (Schwartz and Rozumalski 2005). Although the protocol is practicable and has been established over the years, some main disadvantages exist. It has been shown that intersession and interexaminer reliability are low for this protocol, especially at the hip and knee joint in the frontal and transverse plane (McGinley et al. 2009). The errors in the PiG protocol, for example, knee varus/valgus ROM up to 35° (Ferrari et al. 2008), are very likely caused by inconsistent anatomical landmark identification and marker positioning by the examiner. This leads to well-documented errors of skin movement (Leardini et al. 2005) and kinematic cross talk. Moreover, accurate placement of the wand markers on the shank and the thigh is difficult (Karlsson and Tranberg 1999). Wands on the lateral aspect of the thighs and shanks are also likely to enlarge skin motion artifact effects (Manal et al. 2000) and variability of the gait results (Gorton et al. 2009).

One way of addressing these errors is the usage of additional medial malleolus and medial femoral condyle markers to determine joint centers. This eliminates the reliance on the difficult, subjective palpation of the thigh and tibia wand markers necessary for the PiG model, which has been shown to have large variability (Gorton et al. 2009) between laboratories and to enlarge skin motion artifact effects (Manal et al. 2002), especially when placed proximally where the greatest soft-tissue artifact of any lower-limb segment is found (Stagni et al. 2005). Besides that, it has been shown that thigh wand markers capture approximately half of actual femoral axial rotation (Schache et al. 2008; Schulz and Kimmel 2010; Wren et al. 2008). The reason for this may be that substantial proportions of hip external-internal rotations were being detected as knee motions by the marker sets using thigh markers (Schulz and Kimmel 2010). Wren et al. (2008) have suggested using a patella marker (placed in the center of the patella), which was reported to detect 98% of the actual hip rotation ROM. And indeed, dynamic hip rotation during gait when utilizing a patella marker in lieu of a thigh wand was not effected by static hip posture (McMulkin and Gordon 2009).

In a comparative study, the reliability and accuracy of the PiG model and an advanced protocol (MA) with additional medial malleolus and medial femoral condyle markers were estimated (Stief et al. 2013) (Fig. 2).
Fig. 2

Marker set of both lower body protocols. The markers indicated by circles are part of the standard Plug-in-Gait (PiG) marker set (Conventional Gait Model); those indicated by triangles are the additional markers used in the custom made protocol (MA)

For the MA, neither anthropometric measurements nor joint alignment devices are necessary. Knowledge of anatomical landmarks spatial location enables automatic calculation of anthropometric measurements necessary for joint center determination. In both protocols, the center of the hip joint was calculated using a geometrical prediction method (Davis III et al. 1991). The PiG model derived the rotational axis of the knee joint from the position of the pelvic, knee, and thigh markers and the rotational axis of the ankle joint from the position of the knee, ankle, and tibia markers. In contrast to the PiG model, the centers of the knee and ankle joints using the MA were statically defined as the midpoint between the medial and lateral femoral condyle and malleolus markers. The anatomical medial malleolus and femoral condyle markers can then be removed for the dynamic trials.

The results of this comparative study (PiG model vs. MA) show for both protocols and healthy subjects a good intersession reliability for all ankle, knee, and hip joint angles in the sagittal plane. Nevertheless, the lower intersession errors for the MA compared to the PiG model regarding frontal plane knee angles and moments and transverse plane motion in the knee and hip joint suggest that the error in repeated palpation of the landmarks is lower using the MA. Moreover, the MA significantly reduced the knee axis cross talk phenomenon, suggesting improved accuracy of knee axis alignment compared to the PiG model. These results are comparable to those reported by Schwartz and Rozumalski (2005) using a functional approach in comparison with the PiG model. The MA eliminates the reliance on the subjective palpation of the thigh and tibia wand markers and the application of a Knee Alignment Device method (Davis and DeLuca 1996), which is difficult to handle and less reliable within or between therapists than manual palpation, especially in non-experienced investigators (Serfling et al. 2009). Nevertheless, a correct marker placement based on the exact identification of the characteristic anthropological points of the body (bony landmarks) is required. Especially the position of the knee markers is very important, because it influences not only knee joint kinematics, but also hip and ankle joints. It has been shown that simultaneous knee hyperextension, internal hip rotation, and external ankle rotation can be caused by back lateral knee marker misplacement, and simultaneous knee overflexion, external hip rotation, and internal ankle rotation may be influenced by forward knee marker misplacement (Szczerbik and Kalinowska 2011). Therefore, if such phenomena are represented by kinematic graphs, their presence should be confirmed by video registration prior to the formulation of clinical conclusions.

Future Directions

There are many anatomical models and marker sets reported in the literature. The increase in complexity in the models relates not only to the ability of movement analysis systems to track more and more markers but also in the increase in the knowledge of modeling human movement.

The Importance of Repeatability Studies

Some of the theoretical aspects of marker placement have been presented in this chapter. The practical implications are best explored in the gait laboratory by repeat marker placement. Repeated testing of a single subject will give some insight into the variability for a single person placing markers (intrasubject reliability) and between different people placing markers (intersubject reliability). Intersubject variability would additionally be affected by differences in each subject’s walking style and between-subject differences in marker placement and motion relative to bony landmarks. When the intersubject variability of control data becomes greater than the expected change due to pathology, the clinical usefulness of the data becomes doubtful. To allow a practical interpretation of a comparison of approaches, differences and their variability should be quantified in the unit of interest (i.e., degree or percent).

Removing the Effects of Marker Misplacement

The placement of markers is not easy, and there is a limit to the accuracy we can realistically achieve. Even if the markers are in the right place, the effects of skin movement and oscillation will introduce errors once the subject is walking. One possibility is that the marker placement is “corrected” as part of the data processing. Complex algorithms are now becoming available for performing such corrections. A simpler approach has been used for some time to increase the accuracy in joint center determination. In addition to the Conventional Gait Model, at least the use of medial malleolus and medial femoral condyle markers is recommended when analyzing frontal and transverse plane gait data. This should lead to lower measurement errors for most of the gait variables and to a more accurate determination of the knee joint axis. Nevertheless, gait variables in the transverse plane are poorly reproducible (Ferber et al. 2002; Krauss et al. 2012), and their variability associated with the underlying biomechanical protocol is substantial (Ferrari et al. 2008; Krauss et al. 2012; Noonan et al. 2003).

In future, approaches that combine key characteristics of proven methods (functional and/or predictive methods) for the assessment of skeletal kinematics could be used to optimize marker sets for targeting more accurate and robust motion capture for aiding in clinical diagnosis and improving the reliability of longitudinal studies. On the other hand, procedural distress should be minimized. Especially children cannot always stand still for a long time, walk wearing a large number of markers, and perform additional motion trials. The marker set and possible associated anatomical landmark calibration or anthropometric measurement procedures, therefore, must be minimized to contain the time taken for subject preparation and data collection (Leardini et al. 2007).

Conclusion

When comparing movement data, it is worth noting that care must be taken where different marker sets have been used. Whatever approach is used, the problem is separating patterns produced by errors from those produced by pathology. Till this day, it is, for instance, not clear how different marker configurations impact hip rotation for the typical clinical gait analysis process. For this reason, the “true” values for rotation often remain unknown. Therefore, gait protocols have to be described precisely, and comparison with other studies should be done critically. In addition, knowledge about sources of errors should be known before the approach is applied in practice. Learning and training of the examiners, which is considered to be a critical issue (Gorton et al. 2009), is important to ensure exact anatomical landmark locations which may also reduce intra- and inter-examiner variability. Moreover, the graphs from instrumented gait analysis should be confirmed by video registration prior to the formulation of clinical conclusions.

References

  1. Alexander EJ, Andriacchi TP (2001) Correcting for deformation in skin-based marker systems. J Biomech 34(3):355–361CrossRefGoogle Scholar
  2. Asano T, Akagi M, Nakamura T (2005) The functional flexion-extension axis of the knee corresponds to the surgical epicondylar axis: in vivo analysis using a biplanar image-matching technique. J Arthroplast 20(8):1060–1067.  https://doi.org/10.1016/j.arth.2004.08.005CrossRefGoogle Scholar
  3. Assi A, Sauret C, Massaad A, Bakouny Z, Pillet H, Skalli W, Ghanem I (2016) Validation of hip joint center localization methods during gait analysis using 3D EOS imaging in typically developing and cerebral palsy children. Gait Posture 48:30–35.  https://doi.org/10.1016/j.gaitpost.2016.04.028CrossRefGoogle Scholar
  4. Bell AL, Pedersen DR, Brand RA (1990) A comparison of the accuracy of several hip center location prediction methods. J Biomech 23(6):617–621CrossRefGoogle Scholar
  5. Benedetti MG, Catani F, Leardini A, Pignotti E, Giannini S (1998) Data management in gait analysis for clinical applications. Clin Biomech (Bristol, Avon) 13(3):204–215CrossRefGoogle Scholar
  6. Borhani M, McGregor AH, Bull AM (2013) An alternative technical marker set for the pelvis is more repeatable than the standard pelvic marker set. Gait Posture 38(4):1032–1037.  https://doi.org/10.1016/j.gaitpost.2013.05.019CrossRefGoogle Scholar
  7. Bruno P, Barden J (2015) Comparison of two alternative technical marker sets for measuring 3D pelvic motion during gait. J Biomech 48(14):3876–3882.  https://doi.org/10.1016/j.jbiomech.2015.09.031CrossRefGoogle Scholar
  8. Cappello A, Cappozzo A, La Palombara PF, Lucchetti L, Leardini A (1997) Multiple anatomical landmark calibration for optimal bone pose estimation. Hum Mov Sci 16(2–3):259–274.  https://doi.org/10.1016/S0167-9457(96)00055-3CrossRefGoogle Scholar
  9. Cappello A, Stagni R, Fantozzi S, Leardini A (2005) Soft tissue artifact compensation in knee kinematics by double anatomical landmark calibration: performance of a novel method during selected motor tasks. IEEE Trans Biomed Eng 52(6):992–998.  https://doi.org/10.1109/tbme.2005.846728CrossRefGoogle Scholar
  10. Cappozzo A (1984) Gait analysis methodology. Hum Mov Sci 3(1–2):27–50.  https://doi.org/10.1016/0167-9457(84)90004-6CrossRefGoogle Scholar
  11. Cappozzo A (1991) Three-dimensional analysis of human walking: experimental methods and associated artifacts. Hum Mov Sci 10(5):589–602.  https://doi.org/10.1016/0167-9457(91)90047-2CrossRefGoogle Scholar
  12. Cappozzo A, Catani F, Croce UD, Leardini A (1995) Position and orientation in space of bones during movement: anatomical frame definition and determination. Clin Biomech (Bristol, Avon) 10(4):171–178CrossRefGoogle Scholar
  13. Cappozzo A, Catani F, Leardini A, Benedetti MG, Croce UD (1996) Position and orientation in space of bones during movement: experimental artefacts. Clin Biomech (Bristol, Avon) 11(2):90–100CrossRefGoogle Scholar
  14. Cappozzo A, Cappello A, Della Croce U, Pensalfini F (1997) Surface-marker cluster design criteria for 3-D bone movement reconstruction. IEEE Trans Biomed Eng 44(12):1165–1174.  https://doi.org/10.1109/10.649988CrossRefGoogle Scholar
  15. Chiari L, Della Croce U, Leardini A, Cappozzo A (2005) Human movement analysis using stereophotogrammetry. Part 2: instrumental errors. Gait Posture 21(2):197–211.  https://doi.org/10.1016/j.gaitpost.2004.04.004CrossRefGoogle Scholar
  16. Collins TD, Ghoussayni SN, Ewins DJ, Kent JA (2009) A six degrees-of-freedom marker set for gait analysis: repeatability and comparison with a modified Helen Hayes set. Gait Posture 30(2):173–180.  https://doi.org/10.1016/j.gaitpost.2009.04.004CrossRefGoogle Scholar
  17. Davis RB, DeLuca PA (1996) Clinical gait analysis: current methods and future directions. In: Harris GF, Smith PA (eds) Human motion analysis: current applications and future directions. The Institute of Electrical and Electronic Engineers Press, New York, pp 17–42Google Scholar
  18. Davis RB III, Õunpuu S, Tyburski D, Gage JR (1991) A gait analysis data collection and reduction technique. Hum Mov Sci 10(5):575–587.  https://doi.org/10.1016/0167-9457(91)90046-ZCrossRefGoogle Scholar
  19. Duffell LD, Hope N, McGregor AH (2014) Comparison of kinematic and kinetic parameters calculated using a cluster-based model and Vicon’s plug-in gait. Proc Inst Mech Eng H 228(2):206–210.  https://doi.org/10.1177/0954411913518747CrossRefGoogle Scholar
  20. Ehrig RM, Taylor WR, Duda GN, Heller MO (2006) A survey of formal methods for determining the centre of rotation of ball joints. J Biomech 39(15):2798–2809.  https://doi.org/10.1016/j.jbiomech.2005.10.002CrossRefGoogle Scholar
  21. Ehrig RM, Heller MO, Kratzenstein S, Duda GN, Trepczynski A, Taylor WR (2011) The SCoRE residual: a quality index to assess the accuracy of joint estimations. J Biomech 44(7):1400–1404.  https://doi.org/10.1016/j.jbiomech.2010.12.009CrossRefGoogle Scholar
  22. Ferber R, McClay Davis I, Williams DS 3rd, Laughton C (2002) A comparison of within- and between-day reliability of discrete 3D lower extremity variables in runners. J Orthop Res 20(6):1139–1145.  https://doi.org/10.1016/s0736-0266(02)00077-3CrossRefGoogle Scholar
  23. Ferrari A, Benedetti MG, Pavan E, Frigo C, Bettinelli D, Rabuffetti M, Crenna P, Leardini A (2008) Quantitative comparison of five current protocols in gait analysis. Gait Posture 28(2):207–216.  https://doi.org/10.1016/j.gaitpost.2007.11.009CrossRefGoogle Scholar
  24. Fukaya T, Mutsuzaki H, Wadano Y (2013) Interrater reproducibility of knee movement analyses during the stance phase: use of anatomical landmark calibration with a rigid marker set. Rehabil Res Pract 2013:692624.  https://doi.org/10.1155/2013/692624Google Scholar
  25. Fuller J, Liu LJ, Murphy MC, Mann RW (1997) A comparison of lower-extremity skeletal kinematics measured using skin- and pin-mounted markers. Hum Mov Sci 16(2–3):219–242.  https://doi.org/10.1016/S0167-9457(96)00053-XCrossRefGoogle Scholar
  26. Gao B, Zheng NN (2008) Investigation of soft tissue movement during level walking: translations and rotations of skin markers. J Biomech 41(15):3189–3195.  https://doi.org/10.1016/j.jbiomech.2008.08.028CrossRefGoogle Scholar
  27. Garling EH, Kaptein BL, Mertens B, Barendregt W, Veeger HE, Nelissen RG, Valstar ER (2007) Soft-tissue artefact assessment during step-up using fluoroscopy and skin-mounted markers. J Biomech 40(Suppl 1):S18–S24.  https://doi.org/10.1016/j.jbiomech.2007.03.003CrossRefGoogle Scholar
  28. Gorton GE 3rd, Hebert DA, Gannotti ME (2009) Assessment of the kinematic variability among 12 motion analysis laboratories. Gait Posture 29(3):398–402.  https://doi.org/10.1016/j.gaitpost.2008.10.060CrossRefGoogle Scholar
  29. Harrington ME, Zavatsky AB, Lawson SE, Yuan Z, Theologis TN (2007) Prediction of the hip joint centre in adults, children, and patients with cerebral palsy based on magnetic resonance imaging. J Biomech 40(3):595–602.  https://doi.org/10.1016/j.jbiomech.2006.02.003CrossRefGoogle Scholar
  30. Heller MO, Kratzenstein S, Ehrig RM, Wassilew G, Duda GN, Taylor WR (2011) The weighted optimal common shape technique improves identification of the hip joint center of rotation in vivo. J Orthop Res 29(10):1470–1475.  https://doi.org/10.1002/jor.21426CrossRefGoogle Scholar
  31. Holden JP, Stanhope SJ (1998) The effect of variation in knee center location estimates on net knee joint moments. Gait Posture 7(1):1–6CrossRefGoogle Scholar
  32. Holden JP, Orsini JA, Siegel KL, Kepple TM, Gerber LH, Stanhope SJ (1997) Surface movement errors in shank kinematics and knee kinetics during gait. Gait & Posture 5(3):217–227.  https://doi.org/10.1016/S0966-6362(96)01088-0CrossRefGoogle Scholar
  33. Isman RE, Inman VT (1969) Anthropometric studies of the human foot and ankle. Bull Prosthet Res 10(11):97–219Google Scholar
  34. Kadaba MP, Ramakrishnan HK, Wootten ME, Gainey J, Gorton G, Cochran GV (1989) Repeatability of kinematic, kinetic, and electromyographic data in normal adult gait. J Orthop Res 7(6):849–860.  https://doi.org/10.1002/jor.1100070611CrossRefGoogle Scholar
  35. Karlsson D, Tranberg R (1999) On skin movement artefact-resonant frequencies of skin markers attached to the leg. Hum Mov Sci 18(5):627–635.  https://doi.org/10.1016/S0167-9457(99)00025-1CrossRefGoogle Scholar
  36. Kisho Fukuchi R, Arakaki C, Veras Orselli MI, Duarte M (2010) Evaluation of alternative technical markers for the pelvic coordinate system. J Biomech 43(3):592–594.  https://doi.org/10.1016/j.jbiomech.2009.09.050CrossRefGoogle Scholar
  37. Kornaropoulos EI, Taylor WR, Duda GN, Ehrig RM, Matziolis G, Muller M, Wassilew G, Asbach P, Perka C, Heller MO (2010) Frontal plane alignment: an imageless method to predict the mechanical femoral-tibial angle (mFTA) based on functional determination of joint centres and axes. Gait Posture 31(2):204–208.  https://doi.org/10.1016/j.gaitpost.2009.10.006CrossRefGoogle Scholar
  38. Kratzenstein S, Kornaropoulos EI, Ehrig RM, Heller MO, Popplau BM, Taylor WR (2012) Effective marker placement for functional identification of the centre of rotation at the hip. Gait Posture 36(3):482–486.  https://doi.org/10.1016/j.gaitpost.2012.04.011CrossRefGoogle Scholar
  39. Krauss I, List R, Janssen P, Grau S, Horstmann T, Stacoff A (2012) Comparison of distinctive gait variables using two different biomechanical models for knee joint kinematics in subjects with knee osteoarthritis and healthy controls. Clin Biomech (Bristol, Avon) 27(3):281–286.  https://doi.org/10.1016/j.clinbiomech.2011.09.013CrossRefGoogle Scholar
  40. Leardini A, Cappozzo A, Catani F, Toksvig-Larsen S, Petitto A, Sforza V, Cassanelli G, Giannini S (1999) Validation of a functional method for the estimation of hip joint centre location. J Biomech 32(1):99–103CrossRefGoogle Scholar
  41. Leardini A, Chiari L, Della Croce U, Cappozzo A (2005) Human movement analysis using stereophotogrammetry. Part 3. Soft tissue artifact assessment and compensation. Gait Posture 21(2):212–225.  https://doi.org/10.1016/j.gaitpost.2004.05.002CrossRefGoogle Scholar
  42. Leardini A, Sawacha Z, Paolini G, Ingrosso S, Nativo R, Benedetti MG (2007) A new anatomically based protocol for gait analysis in children. Gait Posture 26(4):560–571.  https://doi.org/10.1016/j.gaitpost.2006.12.018CrossRefGoogle Scholar
  43. Lucchetti L, Cappozzo A, Cappello A, Della Croce U (1998) Skin movement artefact assessment and compensation in the estimation of knee-joint kinematics. J Biomech 31(11):977–984CrossRefGoogle Scholar
  44. Luiz RR, Szklo M (2005) More than one statistical strategy to assess agreement of quantitative measurements may usefully be reported. J Clin Epidemiol 58(3):215–216.  https://doi.org/10.1016/j.jclinepi.2004.07.007CrossRefGoogle Scholar
  45. Lundberg A, Svensson OK, Nemeth G, Selvik G (1989) The axis of rotation of the ankle joint. J Bone Joint Surg (Br) 71(1):94–99CrossRefGoogle Scholar
  46. Manal K, McClay I, Stanhope S, Richards J, Galinat B (2000) Comparison of surface mounted markers and attachment methods in estimating tibial rotations during walking: an in vivo study. Gait Posture 11(1):38–45CrossRefGoogle Scholar
  47. Manal K, McClay I, Richards J, Galinat B, Stanhope S (2002) Knee moment profiles during walking: errors due to soft tissue movement of the shank and the influence of the reference coordinate system. Gait Posture 15(1):10–17CrossRefGoogle Scholar
  48. Manal K, McClay Davis I, Galinat B, Stanhope S (2003) The accuracy of estimating proximal tibial translation during natural cadence walking: bone vs. skin mounted targets. Clin Biomech (Bristol, Avon) 18(2):126–131CrossRefGoogle Scholar
  49. McGinley JL, Baker R, Wolfe R, Morris ME (2009) The reliability of three-dimensional kinematic gait measurements: a systematic review. Gait Posture 29(3):360–369.  https://doi.org/10.1016/j.gaitpost.2008.09.003CrossRefGoogle Scholar
  50. McMulkin ML, Gordon AB (2009) The effect of static standing posture on dynamic walking kinematics: comparison of a thigh wand versus a patella marker. Gait Posture 30(3):375–378.  https://doi.org/10.1016/j.gaitpost.2009.06.010CrossRefGoogle Scholar
  51. Miana AN, Prudencio MV, Barros RM (2009) Comparison of protocols for walking and running kinematics based on skin surface markers and rigid clusters of markers. Int J Sports Med 30(11):827–833.  https://doi.org/10.1055/s-0029-1234054CrossRefGoogle Scholar
  52. Most E, Axe J, Rubash H, Li G (2004) Sensitivity of the knee joint kinematics calculation to selection of flexion axes. J Biomech 37(11):1743–1748.  https://doi.org/10.1016/j.jbiomech.2004.01.025CrossRefGoogle Scholar
  53. Noonan KJ, Halliday S, Browne R, O’Brien S, Kayes K, Feinberg J (2003) Interobserver variability of gait analysis in patients with cerebral palsy. J Pediatr Orthop 23(3):279–287 discussion 288-291Google Scholar
  54. Perry J, Burnfield JM (2010) Gait Analysis. Normal and pathological function, 2nd edn. SLACK Incorporated, ThorofareGoogle Scholar
  55. Piazza SJ, Cavanagh PR (2000) Measurement of the screw-home motion of the knee is sensitive to errors in axis alignment. J Biomech 33(8):1029–1034CrossRefGoogle Scholar
  56. Reinschmidt C, van den Bogert AJ, Lundberg A, Nigg BM, Murphy N, Stacoff A, Stano A (1997a) Tibiofemoral and tibiocalcaneal motion during walking: external vs. skeletal markers. Gait Posture 6(2):98–109.  https://doi.org/10.1016/S0966-6362(97)01110-7CrossRefGoogle Scholar
  57. Reinschmidt C, van den Bogert AJ, Nigg BM, Lundberg A, Murphy N (1997b) Effect of skin movement on the analysis of skeletal knee joint motion during running. J Biomech 30(7):729–732CrossRefGoogle Scholar
  58. Richards J (2008) Biomechanics in clinic and research. Elsevier, PhiladelphiaGoogle Scholar
  59. Sandau M, Heimburger RV, Villa C, Jensen KE, Moeslund TB, Aanaes H, Alkjaer T, Simonsen EB (2015) New equations to calculate 3D joint centres in the lower extremities. Med Eng Phys 37(10):948–955.  https://doi.org/10.1016/j.medengphy.2015.07.001CrossRefGoogle Scholar
  60. Sangeux M, Peters A, Baker R (2011) Hip joint centre localization: evaluation on normal subjects in the context of gait analysis. Gait Posture 34(3):324–328.  https://doi.org/10.1016/j.gaitpost.2011.05.019CrossRefGoogle Scholar
  61. Schache AG, Baker R, Lamoreux LW (2008) Influence of thigh cluster configuration on the estimation of hip axial rotation. Gait Posture 27(1):60–69.  https://doi.org/10.1016/j.gaitpost.2007.01.002CrossRefGoogle Scholar
  62. Schulz BW, Kimmel WL (2010) Can hip and knee kinematics be improved by eliminating thigh markers? Clin Biomech (Bristol, Avon) 25(7):687–692.  https://doi.org/10.1016/j.clinbiomech.2010.04.002CrossRefGoogle Scholar
  63. Schwartz MH, Rozumalski A (2005) A new method for estimating joint parameters from motion data. J Biomech 38(1):107–116.  https://doi.org/10.1016/j.jbiomech.2004.03.009CrossRefGoogle Scholar
  64. Schwartz MH, Trost JP, Wervey RA (2004) Measurement and management of errors in quantitative gait data. Gait Posture 20(2):196–203.  https://doi.org/10.1016/j.gaitpost.2003.09.011CrossRefGoogle Scholar
  65. Serfling DM, Hooke AW, Bernhardt KA, Kaufman KR, 2009 Comparison of techniques for finding the knee joint center. In: Proceedings of the gait and clinical movement analysis society. p 43Google Scholar
  66. Stagni R, Fantozzi S, Cappello A, Leardini A (2005) Quantification of soft tissue artefact in motion analysis by combining 3D fluoroscopy and stereophotogrammetry: a study on two subjects. Clin Biomech (Bristol, Avon) 20(3):320–329.  https://doi.org/10.1016/j.clinbiomech.2004.11.012CrossRefGoogle Scholar
  67. Stief F, Bohm H, Michel K, Schwirtz A, Doderlein L (2013) Reliability and accuracy in three-dimensional gait analysis: a comparison of two lower body protocols. J Appl Biomech 29(1):105–111CrossRefGoogle Scholar
  68. Szczerbik E, Kalinowska M (2011) The influence of knee marker placement error on evaluation of gait kinematic parameters. Acta Bioeng Biomech 13(3):43–46Google Scholar
  69. Taylor WR, Ehrig RM, Duda GN, Schell H, Seebeck P, Heller MO (2005) On the influence of soft tissue coverage in the determination of bone kinematics using skin markers. J Orthop Res 23(4):726–734.  https://doi.org/10.1016/j.orthres.2005.02.006CrossRefGoogle Scholar
  70. Taylor WR, Kornaropoulos EI, Duda GN, Kratzenstein S, Ehrig RM, Arampatzis A, Heller MO (2010) Repeatability and reproducibility of OSSCA, a functional approach for assessing the kinematics of the lower limb. Gait Posture 32(2):231–236.  https://doi.org/10.1016/j.gaitpost.2010.05.005CrossRefGoogle Scholar
  71. Vaughan CL, Davis BL, O’Conner JC (1992) Dynamics of human gait. Human Kinetics Publishers, ChampaignGoogle Scholar
  72. Weidow J, Tranberg R, Saari T, Karrholm J (2006) Hip and knee joint rotations differ between patients with medial and lateral knee osteoarthritis: gait analysis of 30 patients and 15 controls. J Orthop Res 24(9):1890–1899.  https://doi.org/10.1002/jor.20194CrossRefGoogle Scholar
  73. Wren TA, Do KP, Hara R, Rethlefsen SA (2008) Use of a patella marker to improve tracking of dynamic hip rotation range of motion. Gait Posture 27(3):530–534.  https://doi.org/10.1016/j.gaitpost.2007.07.006CrossRefGoogle Scholar
  74. Wu G, Siegler S, Allard P, Kirtley C, Leardini A, Rosenbaum D, Whittle M, D’Lima DD, Cristofolini L, Witte H, Schmid O, Stokes I (2002) ISB recommendation on definitions of joint coordinate system of various joints for the reporting of human joint motion – part I: ankle, hip, and spine. International Society of Biomechanics. J Biomech 35(4):543–548CrossRefGoogle Scholar

Copyright information

© Springer International Publishing AG, part of Springer Nature 2018

Authors and Affiliations

  1. 1.Movement Analysis LabOrthopedic University Hospital Friedrichsheim gGmbHFrankfurt/MainGermany

Section editors and affiliations

  • Sebastian I. Wolf
    • 1
  1. 1.Movement Analysis LaboratoryClinic for Orthopedics and Trauma Surgery; Center for Orthopedics, Trauma Surgery and Spinal Cord Injury;Heidelberg University HospitalHeidelbergGermany

Personalised recommendations