Data processing and analysis

The data were then exported into Microsoft Excel for processing. To begin offline HRV analysis, the processed data were imported into Kubios Premium (V. 3.2, The Biomedical Signals Analysis and Medical Imaging Group, University of Kuopio, Finland). As a primary outcome measure, only the time domain HRV measure of RMSSD (root mean square of successive differences) was calculated in this study given its established reliability and validity in HRV_UST analysis [4–6].

All arterial waveform data were visually inspected to locate any ectopic, missed, or misplaced beats, and were manually corrected when needed. For data processing in Kubios Premium, the smoothness priors (500), and interpolation (cubic spline: 4 Hz with 50ms R-R threshold) were chosen as previously used [24]. The maximal acceleration point in the upslope in the arterial waveform was the default selection within Kubios to determine the IBI as previously validated [24]. The automatic correction and noise levels were set to none and medium, respectively throughout the analysis. Since the minimum analysable length of the signal in Kubios is 10 seconds, the signals shorter than 10 seconds were interpolated using the merge analysis type. Otherwise, the analysis type was set to single.