HAL Id: hal-02109095
https://hal.archives-ouvertes.fr/hal-02109095
Submitted on 9 May 2019
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
A simple geometrical method to minimise crosstalk during clinical gait analysis
Alexandre Naaim, Alice Bonnefoy-Mazure, Raphaël Dumas, Stéphane Armand
To cite this version:
Alexandre Naaim, Alice Bonnefoy-Mazure, Raphaël Dumas, Stéphane Armand. A simple geometrical
method to minimise crosstalk during clinical gait analysis. 15th International Symposium on 3-D
Analysis of Human Movement, Jul 2018, Manchester, France. 2 p. �hal-02109095�
22 XV International Symposium on 3-D Analysis of Human Movement
Wednesday July 4
th2018
O01
A simple geometrical method to minimise crosstalk during clinical gait analysis Alexandre Naaim1, Alice Bonnefoy-Mazure2, Raphaël Dumas1, Stephane Armand2
1Univ Lyon, Université Claude Bernard Lyon 1, IFSTTAR, LBMC UMR_T9406, F69622, Lyon, France. 2Willy Taillard Laboratory of Kinesiology, Geneva University Hospitals and Geneva University, Geneva, Switzerland
Introduction
Stereophotogrammetry has been widely used for measuring gait. Conventional gait analysis systems tend to be based on the Davis marker-set [1]. This marker-set is dependent on the position of a wand on the thigh to define medial-lateral thigh axis. This wand is supposed to lie in the frontal plane of the thigh and can be difficult to position properly. A poor
positioning leads to axis misorientation which can cause errors known as crosstalk. Crosstalk corresponds to an unphysiological knee Abduction Adduction (AA) amplitude correlated with knee Flexion-Extension (FE) [2]. This
phenomenon also affects hip Internal External Rotation (IER) which is a parameter of high interest in clinical gait analysis (CGA). Several methods have been proposed previously to better define the medial-lateral axis of the thigh and improve the quality of both hip and knee kinematics. However, the corrections were not always satisfying or difficult to implement [2–3].
Research Question
This study aimed to propose an a-posteriori simple geometrical method allowing re-orientating the medial-lateral thigh axis to correct, at the same time, knee AA and hip IER.
Methods
To test this new geometrical method a database of 118 patients was used. These patients had a CGA, at the Willy Taillard Laboratory of Kinesiology (Geneva University Hospitals), before and after a knee joint replacement [4].
Protocol: CGA were performed using a twelve-camera system using the lower limb Davis marker-set. Patients were asked to walk on a 10-m walkway at their self-selected speed until at least 10 gait cycles were recorded.
Method: In the Davis marker-set model, the thigh anterior-posterior axis was initially constructed thanks to the normal to the plane defined by the hip joint centre, the knee joint centre and the wand marker position [1]. The medial-lateral thigh axis lies in this normal plane, orthogonal to the longitudinal axis defined from knee to hip joint centre. The geometrical method introduced in this study aimed to correct this axis. For this, the sagittal plane of the lower limb was defined by the hip, knee and ankle joint centres. In order to obtain a correct orientation of the normal to this sagittal plane, the mean normal was only calculated when the knee flexion was superior to 20° during the cycle. Then, the thigh anterior-posterior axis was replaced by the cross product between this mean normal and the longitudinal axis, while the medial-lateral thigh axis was replaced by another cross product.
Data analysis: Only data with at least 20° of knee flexion were selected. The remaining data were then divided into two groups in function of the knee AA. If the knee AA amplitude was inferior to 10°, trials were considered as reference data, otherwise the geometrical method was applied. The knee AA in function of the knee FE for each point of gait cycle was plotted. Then, the Pearson coefficient was calculated between both angles for non-corrected and corrected data. The root mean square difference (RMSD) was also calculated between the reference and the data before and after correction for knee AA and hip IER.
Results
On the 9964 gait cycles available from the database, 326 were discarded (maximal knee flexion < 20°), 2372 gait cycles were used as a reference (AA amplitude < 10°) and 7266 gait cycles (AA amplitude > 10°) were corrected using the geometrical method tested.
XV International Symposium on 3-D Analysis of Human Movement 23
Wednesday July 4
th2018
Figure 1: Mean and standard deviation of knee AA and hip IER during a gait cycle for the reference, non-corrected and corrected data with the RMSD relative to reference
Figure 2: Knee AA in function of knee FE for all trials with an amplitude of knee AA > 10° with and without correction.
Pearson coefficients between both angles are reported (ρ) Discussion
Person’s coefficient between the FE and the AA dropped from 0.45 to -0.02 after the use of the correction method. In addition, the knee AA kinematics pattern after correction was flatter. The geometrical method seemed to reduce the crosstalk phenomenon. Considering the hip IER, the geometrical method allowed to decrease the RMSD relative to reference from 6.63° to 3.00°. However, it must be highlighted that the reference used cannot be considered as a gold standard. It represents only data for which the knee AA was considered physiological.
Consequently, it seems that the geometrical method introduced allows to correct the thigh axes orientation and reduce errors due to crosstalk. The proposed method has strong advantages to be simple and implemented a-posteriori. It can be performed on retrospective data or on new data without any modification of existing clinical protocols (no additional markers or calibration movements…). Moreover, it is not specific to the Davis marker-set.
References
1. Davis, Human movement science 1991;10: 575-587.
2. Baker, Human Movement Science 1999;18:655-667.
3. Baudet, Plos One 2014;9:1-7.
4. Bonnefoy-Mazure, The Journal of Arthroplasty 2017;32:793-800.