Thanks to the observations of a user, we have discovered that a warning message displayed when loading data to the KINARM_add_Jerk.m function is incorrect. The warning message indicates that unfiltered data must be used when applying this function. Our user discovered a variance between different Kinarm systems which could be explained by whether the Kinarm Lab had a (older) PMAC or (new) EtherCAT based control card. We have now determined that data must be filtered with a 10 Hz filter before using the Jerk function not after as indicated by our warning message.
Who is impacted: users with a PN 10288 (Kinarm End-Point Lab w PMAC) or PN 10755 (Kinarm Classic w PMAC) could experience significant differences between pre-filtered jerk analysis and post-filtered jerk analysis. Why: PMAC systems collect position data at 1.13 kHz and then re-sampled, which introduces noise.
The fix: only load data that has been filtered with a 10 Hz filter into the KINARM_add_Jerk.m function.
The correction: The filtering needs to be applied BEFORE calculating jerk, in order to avoid possible edge effects at the beginning (or end) of a trial. The warning and documentation will be updated in a future release.
Please contact us at support@kinarm.com if you have any concerns about the impact of this change. Your feedback will be used to help us determine the urgency of releasing a new version of KAS.