Amsterdam – February 2015 Technical Quality Assurance of GA The aim of Technical Quality Assurance (TQA) was to assess the quality of the measurements conducted in the gait labs involved in the WP6:  KU Leuven (KUL);  VU Medisch Centrum (VUA);  Ospedale Pediatrico ‘Bambino Gesù’ (OPBG) TQA protocol is composed by two experimental sessions:  Low-level session: focused on the comparability of the measurement systems  High-level session: focused on the inter-laboratory/rater repeatability of gait analysis

Amsterdam – February 2015 Low-level session OS-validation  Wands were moved for 10 s inside the measurement volume  Distances and angles were evaluated and compared with the actual values  Repeatability of parameters was evaluated as RMSE Less accuracy of Bonita system

Amsterdam – February 2015 Low-level session S-synchro  Signal synchronization between the force platform and the EMG system  EMG Cometa system is provided with Foot-switch sensors (FS) which are acquired synchronously with EMG  Time delay between EMG and FP was individuated when both signals were zero Pointer Foot switch Force platform Values have to be considered in the activation time of muscles

Amsterdam – February 2015 Low-level session FP-validation  Pointer (LC-P) equipped with a 6-component load cell and 5 markers  Fully rotational ferrule allows LC-P to tilt in several directions  Application of arbitrary forces to each force platform with LC-P:  Oriented along the vertical axis  Tilted of approximately 30°around x and y  F and M were measured by LC-P and FP  F and M were projected on OS reference system  RMSEs were calculated for each pushed point

Amsterdam – February 2015 Low-level session FP-validation  Forces and Moments measured with FP were comparable with the same parameters evaluated by means of LC-P  Differences between curves were observable only at VUA for F x and F y

Amsterdam – February 2015 Low-level session FP-validation  F y and F x at OPBG and VUA got worse from the center to the edge of platform  Different behavior of platforms at each center was found  RMSE values were lower than the 1% of Full Scale of each platform  RMSEs have to be considered to estimate the contribution of FP to the overall uncertainty of the kinetic variables

Amsterdam – February 2015 The aim of high level is the evaluation of the repeatability of measurements conducted in different laboratories. The protocol includes the following features:  Two healthy children have been recruited;  Data were collected in OPBG, KUL and VUA;  Two therapists per center;  Five walking trials were collected. High level The considered variables are  Joint angles  Joint moments  Spatiotemporal parameters  Timing on EMG activation The repeatability indices used :  Coefficient of Multiple Correlation (CMC w ) within laboratory  Coefficient of Multiple Correlation (CMC b ) between laboratories

Amsterdam – February 2015 Within laboratory repeatability  In the sagittal plane the repeatability within laboratory was excellent  In the frontal and transverse plane the repeatability was lower than sagittal plane  CMC for hip rotation was the lowest value, it could be due to a different marker placement between therapists. Comparable values of CMC for second subjects were found (0.78, 0.83)

Amsterdam – February 2015 Between laboratories repeatability Two CMC between laboratories were calculated: 1.CMC between laboratories with PiG model (OPBG and KUL); 2.CMC between laboratories with different models (OPBG, KUL and VUA).  The repeatability between the three centers was lower than the repeatability between OPBG and KUL, however in the range of a good repeatability, with the exception of hip rotation between the three laboratories.

EMG Signal Amsterdam – February 2015 ANOVA Subject 1 Muscle attivation Time (%)OPBGKULVUAp-value Rectus Femoris13.9 (0.9)12.6 (1.5)12.8 (1.6)0.072 Medial Hamstring17.2 (1.4)16.9 (1.1)17.3 (1.7)0.814 Vastus Lateralis15.9 (1.2)15.8 (1.2)16.1(1.2)0.918 Biceps Femoris17.5 (3.3)16.1 (1.2)17.7 (2.1)0.271 Soleus18.3 (5.3)18.4 (4.9)17.6 (2.8)0.915 Gluteus Medius8.9 (1.9) VUA 7.7 (0.9) VUA 12.2 (3.2) OPBG-KUL <0.001  The activation time of 8 muscles (4 agonist and 4 antagonist) of one lower limb was calculated  T-tests were performed in order to find significant differences between the operators of the same laboratory  One-way ANOVA was conducted whether t-test was not significant  T-tests were significant only for Anterior Tibialis and Gastrocnemius  Statistical differences between laboratories were found for Gluteus Medius in both subjects  Therapists have to pay specific attention to place electrodes on Anterior Tibialis, Gastrocnemius and Gluteus Medius  Comparability of EMG data of these muscles among centers could be questionable

Future Works Amsterdam – February 2015  Analysis of comparability of the two models (PiG and HBM) on the same subject and the same gait  In order to do it, we need of:  Gait data from OPBG and KUL taking into account HBM model;  Gait data from VUA taking into account PiG model.