Fig. 5From: Kinect-based assessment of proximal arm non-use after a strokeTest-retest of PANU scores with the Kinect and CMS20s. Each panel compares the PANU scores in the test (R1) and retest (R2) sessions. Panels in the first row illustrate repeatability with the Kinect. Panels in the second row illustrate repeatability with the CMS20s. For each row, the left panel represents the Bland and Altman plot and the right panel represents the linear regression plot. The constant bias in the Bland and Altman plots (− 4.25 for Kinect; − 4.71 for CMS20s) indicates that the PANU scores decrease over repetitions, which was accurately determined by the Kinect and the CMS20sBack to article page