Jörg Eschweiler

A survey on robotic devices for upper limb rehabilitation.

The existing shortage of therapists and caregivers assisting physically disabled individuals at home is expected to increase and become serious problem in the near future. The patient population needing physical rehabilitation of the upper extremity is also constantly increasing. Robotic devices have the potential to address this problem as noted by the results of recent research studies. However, the availability of these devices in clinical settings is limited, leaving plenty of room for improvement. The purpose of this paper is to document a review of robotic devices for upper limb rehabilitation including those in developing phase in order to provide a comprehensive reference about existing solutions and facilitate the development of new and improved devices. In particular the following issues are discussed: application field, target group, type of assistance, mechanical design, control strategy and clinical evaluation. This paper also includes a comprehensive, tabulated comparison of technical solutions implemented in various systems.

Surface-based determination of the pelvic coordinate system

In total hip replacement (THR) one technical factor influencing the risk of dislocation is cup orientation. Computer-assisted surgery systems allow for cup navigation in anatomy-based reference frames. The pelvic coordinate system most used for cup navigation in THR is based on the mid-sagittal plane (MSP) and the anterior pelvic plane (APP). From a geometrical point of view, the MSP can be considered as a mirror plane, whereas the APP can be considered as a tangent plane comprising the anterior superior iliac spines (ASIS) and the pubic tubercles. In most systems relying on the pelvic coordinate system, the most anterior points of the ASIS and the pubic tubercles are selected manually. As manual selection of landmark points is a tedious, time-consuming and error-prone task, a surface-based approach for combined MSP and APP computation is presented in this paper: Homologous points defining the MSP and the landmark points defining the APP are selected automatically from surface patches. It is investigated how MSP computation can benefit from APP computation and vice versa, and clinical perspectives of combined MSP and APP computation are discussed. Experimental results on computed tomography data show that the surface-based approach can improve accuracy.

Analysis of wrist bone motion before and after SL-ligament resection.

Abstract The analysis of the three-dimensional motion of wrist joint components in the physiological and injured wrist is of high clinical interest. Therefore, the purpose of this in vitro study was to compare the motion of scaphoid, lunate and triquetrum during physiological wrist motion in flexion and extension, and in radial- and ulnar-deviation, with those motion patterns after complete resection of the scapho-lunate-ligament. Eight fresh frozen cadaver wrists were carefully thawed and prepared for the investigation with an electromagnetic tracking system by implantation of measurement coils with 6 degrees of freedom. Electromagnetic tracking enabled the motion analysis of the scaphoid, lunate, and triquetrum bones with respect to the fixed radius in three planes of passive motion. After scapho-lunate-ligament injury changes in the translational and rotational motion pattern especially of the scaphoid bone occurred in dorsal-volar directions during flexion and extension, radial- and ulnar-deviation, and during rotation around the radio-ulnar- and longitudinal-axis of the wrist.

Development of a biomechanical model of the wrist joint for patient-specific model guided surgical therapy planning: Part 1

An enhanced musculoskeletal biomechanical model of the wrist joint is presented in this article. The developed computational model features the two forearm bones radius and ulna, the eight wrist bones, the five metacarpal bones, and a soft tissue apparatus. Validation of the model was based on information taken from the literature as well as own experimental passive in vitro motion analysis of eight cadaver specimens. The computational model is based on the multi-body simulation software AnyBody. A comprehensive ligamentous apparatus was implemented allowing the investigation of ligament function. The model can easily patient specific personalized on the basis of image information. The model enables simulation of individual wrist motion and predicts trends correctly in the case of changing kinematics. Therefore, patient-specific multi-body simulation models are potentially valuable tools for surgeons in pre- and intraoperative planning of implant placement and orientation.

Automatic extraction of the mid-sagittal plane using an ICP variant

Precise knowledge of the mid-sagittal plane is important for the assessment and correction of several deformities. Furthermore, the mid-sagittal plane can be used for the definition of standardized coordinate systems such as pelvis or skull coordinate systems. A popular approach for mid-sagittal plane computation is based on the selection of anatomical landmarks located either directly on the plane or symmetrically to it. However, the manual selection of landmarks is a tedious, time-consuming and error-prone task, which requires great care. In order to overcome this drawback, previously it was suggested to use the iterative closest point (ICP) algorithm: After an initial mirroring of the data points on a default mirror plane, the mirrored data points should be registered iteratively to the model points using rigid transforms. Finally, a reflection transform approximating the cumulative transform could be extracted. In this work, we present an ICP variant for the iterative optimization of the reflection parameters. It is based on a closed-form solution to the least-squares problem of matching data points to model points using a reflection. In experiments on CT pelvis and skull datasets our method showed a better ability to match homologous areas.

Impact of scapholunate dissociation on human wrist kinematics

Neither the complex motions of the scapholunate joint, nor the kinematic changes that occur as a result of injury to it, are fully understood. We used electromagnetic tracking within affected bones to evaluate the physiologic motions in the planes of flexion and extension, and of radial and ulnar deviation of human cadaver wrists, before and after complete transection of the scapholunate ligaments. Despite individual variance between each wrist, we were able to establish a pattern in the changes that occurred after scapholunate ligament injury. During the motions examined, the scaphoid showed an increase in translational deviation in almost all motion axes. In contrast, the movement of the lunate seemed to be impaired, especially in radial-ulnar deviation.

A biomechanical model of the wrist joint for patient-specific model guided surgical therapy: Part 2.

An enhanced musculoskeletal biomechanical model of the wrist joint is presented in this article. The computational model is based on the multi-body simulation software AnyBody. Multi body dynamic musculoskeletal models capable of predicting muscle forces and joint contact pressures simultaneously would be valuable for studying clinical issues related to wrist joint degeneration and restoration. In this study, the simulation model of the wrist joint was used for investigating deeper the biomechanical function of the wrist joint. In representative physiological scenarios, the joint behavior and muscle forces were computed. Furthermore, the load transmission of the proximal wrist joint was investigated. The model was able to calculate the parameters of interest that are not easily obtainable experimentally, such as muscle forces and proximal wrist joint forces. In the case of muscle force investigation, the computational model was able to accurately predict the computational outcome for flexion and extension motion. In the case of force distribution of the proximal wrist joint, the model was able to predict accurately the computational outcome for an axial load of 140 N. The presented model and approach of using a multi-body simulation model are anticipated to have value as a predictive clinical tool including effect of injuries or anatomical variations and initial outcome of surgical procedures for patient-specific planning and custom implant design. Therefore, patient-specific multi-body simulation models are potentially valuable tools for surgeons in pre- and intraoperative planning of implant placement and orientation.

Combined magnetic resonance imaging approach for the assessment of in vivo knee joint kinematics under full weight-bearing conditions

The development of detailed and specific knowledge on the biomechanical behavior of loaded knee structures has received increased attention in recent years. Stress magnetic resonance imaging techniques have been introduced in previous work to study knee kinematics under load conditions. Previous studies captured the knee movement either in atypical loading supine positions, or in upright positions with help of inclined supporting backrests being insufficient for movement capture under full-body weight-bearing conditions. In this work, we used a combined magnetic resonance imaging approach for measurement and assessment in knee kinematics under full-body weight-bearing in single legged stance. The proposed method is based on registration of high-resolution static magnetic resonance imaging data acquired in supine position with low-resolution data, quasi-static upright-magnetic resonance imaging data acquired in loaded positions for different degrees of knee flexion. The proposed method was applied for the measurement of tibiofemoral kinematics in 10 healthy volunteers. The combined magnetic resonance imaging approach allows the non-invasive measurement of knee kinematics in single legged stance and under physiological loading conditions. We believe that this method can provide enhanced understanding of the loaded knee kinematics.

Evaluation of Biomechanical Models for the Planning of Total Hip Arthroplasty.

Musculoskeletal loading plays an important role in the primary stability of THA. There are about 210,000 primary THA interventions p.a. in Germany. Consideration of biomechanical aspects during computer-assisted orthopaedic surgery is recommendable in order to obtain satisfactory long-term results. For this purpose simulation of the pre- and post-operative magnitude of the resultant hip joint force R and its orientation is of interest. By means of simple 2D-models (Pauwels, Debrunner, Blumentritt) or more complex 3D-models (Iglic), the magnitude and orientation of R can be computed patient-individually depending on their geometrical and anthropometrical parameters. In the context of developing a planning module for computer-assisted THA, the objective of this study was to evaluate the mathematical models. Therefore, mathematical model computations were directly compared to in-vivo measurements obtained from instrumented hip implants. With patient-specific parameters the magnitude and orientation of R were model-based computed for three patients (EBL, HSR, KWR) of the OrthoLoad-database. Their patient-specific parameters were acquired from the original patient X-rays. Subsequently, the computational results were compared with the corresponding in-vivo telemetric measurements published in the OrthoLoad-database. To obtain the maximum hip joint load, the static single-leg-stance was considered. A reference value for each patient for the maximum hip load under static conditions was calculated from OrthoLoad-data and related to the respective body weights (BW). On average there are large deviations of the results for the magnitude (O=147%) and orientation (O=14.35° too low) of R obtained by using Blumentritt9s model from the in-vivo results/measurements. The differences might be partly explained by the supplemental load of 20% BW within Blumentritt9s model which is added to the input parameter BW in order to consider dynamic gait influences. Such a dynamic supplemental load is not applied within the other static single-leg-stance models. Blumentritt9s model assumptions have to be carefully reviewed due to the deviations from the in-vivo measurement data. Iglic9s 3D-model calculates the magnitude (O17%) and the orientation (O49%) of R slightly too low. For the magnitude one explanation could be that his model considers nine individual 3D-sets of muscle origins and insertion points taken from literature. This is different from other mathematical models. The patient-individual muscle origin and insertion points should be used. Pauwels and Debrunner9s models showed the best results. They are in the same range compared to in-vivo data. Pauwels9s model calculates the magnitude (O5%) and the orientation (O28%) of R slightly higher. Debrunner9s model calculates the magnitude (O1%) and the orientation (O14%) of R slightly lower. In conclusion, for the orientation of R, all the computational results showed variations which tend to depend on the used model. There are limitations coming along with our study: as our previous studies showed, an unambiguous identification of most landmarks in an X-ray (2D) image is hardly possible. Among the study limitations there is the fact that the OrthoLoad-database currently offers only three datasets for direct comparison of static single leg stance with in-vivo measurement data of the same patient. Our ongoing work is focusing on further validation of the different mathematical models.

Application and evaluation of biomechanical models and scores for the planning of total hip arthroplasty

Intimate knowledge of the biomechanics of a given individual hip joint provides a potential advantage during the planning of total hip arthroplasty, and would thus have a positive influence over the outcome of such an intervention. In current clinical practise, the surgical planning is based solely on the status of the individual hip and its radiographic appearance. However, additional information could be gathered from the radiography to be used as input data for biomechanical models aimed at calculating the resultant force FR within the hip joint. An investigation of the biomechanical models by Pauwels, Debrunner and Iglič was performed, where the magnitude of FR calculated by the models showed a favourable comparison to the in-vivo data from instrumented prostheses by Bergmann. The Blumentritt model returned abnormally high results. The computational results showed large variations for FR orientation, which tends to depend more on the model used than on patient-specific parameters. Furthermore, a discrepancy was found between the data gathered from instrumented prostheses and the Standing Human Model within the ‘AnyBody Modeling System™’ software by AnyBody Tech. Additionally, the variations in inter-rater and intra-rater errors made while localizing radiographic landmarks were analysed with respect to their influence on Babisch-Layher-Blumentritt (BLB)-scoring using the Blumentritt hip model.