**#OPTIHEALTH** This project aims to compare reaching movement in young healthy participants recorded with the Kinect V2 against the Vicon motion capture system as the gold standard. This project is linked to the article : **Faity G, Mottet D, Froger J. Validity and Reliability of Kinect v2 for Quantifying Upper Body Kinematics during Seated Reaching. Sensors. 2022; 22(7):2735. https://doi.org/10.3390/s22072735** Preprint available here: https://doi.org/10.1101/2022.01.18.476737 ######### Repertory **DAT** contains all data. Repertory **DAT/000-raw** contains all raw data, in text format (txt or csv). -> Files like **001, LapBri - 07-02-2019_15-09-57 - TAR_00.txt** are Kinect files. -> Files like **001, LapBri - V - TAR_00.csv** are corresponding Vicon files. TAR files are TARGET files: one reach with the loaded arm, and one reach with the unloaded arm. SAU files are SPONTANEOUS ARM USE files: 5 reaches with the loaded arm. MAU files are MAXIMAL ARM USE files: 5 reaches with the loaded arm, while minimising trunk use. See below for the logic of the TAR/SAU/MAU system. Additional information contains sex (M/F), birthdate (DD/MM/YYYY), height (m), weight (kg), maximum voluntary force of shoulder flexion (MVC) (kg), DumblMass (kg), Nb of exercises (Spontaneous and Maximal), Beg and End of each exercise (timeframe). Repertory **DAT/001-cleaned** contains all cleaned data (resampled, filtered) in scilab format (.sod). Repertory **DAT/002-mergeBySession** are cleaned data combined by participant and session (1 big scilab data file by session and by participant). Repertory **DAT/003-summary** are results from scilab analysis. -> Repertory **DAT/003-summary/all_csv_concat** contains summary of kinematic analysis (big table). -> Repertory **DAT/003-summary/all_csv_comparison_concat** contains summary of raw signal comparison (Kinect VS Vicon). -> Repertory **DAT/003-summary/ViconVSKinect_...** contains detailed results for each participants (with figures). Repertory **DAT/004-statistics_ViconVSKinect** contains statistics results (from comparison of Kinematic summary between kinect and vicon. -> Repertory **DAT/004-statistics_ViconVSKinect/LOADED ARM** contains statistics when the arm is loaded. -> Repertory **DAT/004-statistics_ViconVSKinect/UNLOADED ARM** contains statistics when the arm is not loaded. ######### # PANU_analysis Analyse motion capture data from Proximal Arm Non Use (PANU) test. Step 1 (scilab) : Preprocessing. Primary analysis from raw data (txt) is provided by modified MacokiV6 program : https://github.com/DenisMot/MACOKIv6 Primary analysis is : read from txt (/DAT/000-raw), tidy in a scilab struct, data processing, export scilab struct in sod file(/DAT/001-cleaned and /DAT/002-mergeBySession). Step 2 (scilab) : Analysis. Secondary analysis from sod file : recompute PANU (/DAT/002-mergeBySession), compute number of velocity peaks, elbow shoulder and trunk angles, torques... Export these results in a summary table in .csv + graphics in png (in /DAT/003-summary). Another analysis is done (see Main.sce) : for comparison Kinect VS Vicon files. Step 3 (R) : Statistics. Tertiary analysis from summary table (/DAT/003-summary/all_csv_concat/resAllSessions.csv) is statistics (between groups) then export results in /DAT/004-statistics_ViconVSKinect. #HOW IT WORKS Method (manual) : Step 1 : run step 1 (/SRC/step1-preprocessing/CheckPANU.sce) with scilab 6.0.2+ Step 2 : run step 2 (/SRC/step2-analysis/Main.sce) with scilab 6.0.2+ Step 3 : run step 3 (/SRC/step3-statistics_ViconVSKinect/Main.R) with R 3.6.1+ #CAUTION - UPDATE 2020-06-30 15h36 Step 2 from Axel Martinez thesis have been updated (new way to compute torques, more precise). In consequence, step 3 has been updated too. Old program from Axel Martinez thesis is still available at this adress : https://www.euromov.fr/d/d8f01c63/ Changement : in the old version, when the subject was leaning back (back on the backrest), the weight of the trunk was removed in the computation of hipsCenter torque, whatever the weight of arm and dumbbel. So when we computed hipsCenterTorque, we could have a postiv torque but without taking into account trunk weight. So hips center torques and sum of torque could be overestimated. In the new version, when the subject is leaning back (back on the armrest), we compute the total torque on hipsCenter (trunk + arm + dumbbel) and then, if the sagital torque tilt the trunk back, we put a 0 (and keep frontal torque integer). #GLOBAL DESCRIPTION - PANU project Task : seated reaching tasks from PANU protocol (Proximal Arm Non Use - seated reaching tasks). Theses tasks are recorded with a 3D motion system. Related to "Proximal Arm Non Use when reaching after a stroke" (Bakhti et al., 2017) Kinect One (Microsoft) is validated for PANU score in Bakhti et al., 2018. The kinematic analysis of the dataset (only vicon) is in Faity et al., 2021 (https://www.nature.com/articles/s41598-021-02338-4). 3 Blocks : TAR = TARGET : definition of target. A reach by paretic hand (held 5 sec) then a reach by control hand (held 5 sec). SAU = Spontaneous Arm Use : 5 reach by paretic hand (held 1 sec) then 5 reach by control hand (held 5 sec), in spontaneous condition. MAU = Maximal Arm Use : 5 reach by paretic hand (held 1 sec) then 5 reach by control hand (held 5 sec), in maximal condition (trunk restraint). For files from directory Faity 2019 : R0 = a block (control condition) at time 1 R1 = a block (loaded condition) at time 1 R2 = a block (control condition) at time 2 R3 = a block (loaded condition) at time 3 In control condition, the "paretic" hand is not loaded (the "non paretic" hand is not loaded neither). In loaded condition, the "paretic" hand is loaded in all reachs. Load = 85% CMV in shoulder flexion (90°). The "non paretic" hand is not loaded. AUTHORS Germain FAITY <germain.faity@umontpellier.fr> Denis MOTTET <denis.mottet@umontpellier.fr> STRUCTURE Each step is organized as: - WRK - PRG = all program files - DAT = all input data - RES = all results of processing here NOTE : subdirectory are possible, and certainly useful if many files... DATAS Exemples are given in DAT directory. These exemples may be already computed (RES path). Don't hesitate to delete the contents of RES directory to compute them by yourself. Additionnal data are openly avalaible at : https://osf.io/r3xcu/files/ ########################################################################################################### STEP DESCRIPTION SEE subdirectories' README WARNING ! SEE subdirectories' README PROGRAM ORDER SEE subdirectories' README HOW TO LAUNCH THE CODE SEE subdirectories' README ########################################################################################################### ## Repertory **PANU_preprocessing** contains code needed to preprocessed data from txt to sod. The repertory archives the code at the time of article submission. See README.md in the repertory. The **PANU_preprocessing** code is still being updated for further analyses, see https://github.com/DenisMot/MACOKIv6 ######### Analysis conducted with SciLab 6.0.2 and R 3.6.1.