Arnold Huurnink

Chapter 8.2 8.2 120 Dear editor in chief We thank Pagnacco and colleagues for their interest in our work, but we have to disagree with their conclusions [1]. Below we will argue that the rhetorical question in the title of their letter can be answered in the affirmative. Pagnacco and colleagues raise concerns about the methodology of our study, and even claim that the conclusions offered are misleading and unsupported. We agree that the two devices (laboratory grade force plate (FP) and Wii balance Board (WBB)) cannot be considered interchangeable, but this is not what we state in our paper. In fact, the limitations of the WBB are extensively presented and discussed in the paper. We did, however, conclude that the WBB is sufficiently accurate to be used for sway measures of single-leg stance, as used in the field of sports medicine and sports rehabilitation. Although not exactly equal, the center of pressure (CoP) measurements with the WBB were shown to provide very good proxies for parameters derived from more accurately measured CoP data. For more information about the specifications and durability of the WBB, we would like to advise the recent work of Bartlett et al. (2014) [2]. To support our conclusion, we have presented comprehensive results on the agreement between the WBB and FP in 420 trials. In our work, we have performed two analyses to evaluate the agreement between WBB and FP: 1) a comparison of CoP trajectories (time series), and 2) a comparison of commonly used parameters calculated from these trajectories (CoP path velocity, mean CoP sway). 1) To evaluate the similarity of CoP trajectories, both time series data sets (WBB and FP) were compared using the Pearson’s correlation coefficient (r), and the root-mean-square (RMS) technique. In contrast to Lee et al. (1989) (cited in the letter by Pagnacco et al.) [3] , the study by Derrick et al. (1994) (cited in our paper) [4] comprehensively investigated the pros and cons of the use of r to assess similarity between two time series. They showed that r is sensitive to differences in both amplitude and timing. Advantages are that it is easy to use and can be used to evaluate the entire curve. Derrick et al. (1994) concluded that a very high correlation is always indicative of temporal similarity, however, a low correlation does not guarantee a lack of temporal similarity. Obviously, a high r does not guarantee absence of amplitude differences. This is why we also calculated the root- mean-squared differences between CoP trajectories of FP and WBB. Since our Pearson’s correlation coefficients were consistently very high (mean r > 0.996), and our root-mean- square differences were consistently very low (mean RMS < 0.74 mm), we are confident in concluding that the CoP trajectory of FP and WBB possess similar characteristics. 2) The CoP parameters derived from the CoP trajectories were compared between FP and WBB by means of Pearson’s correlation coefficient (linearity), the mean difference (systematic bias), and the standard deviation of the difference (consistency). We dispute