Fig 1.
ECG system configuration for simultaneous measurement (adapted from [34]).
Table 1.
General information about the measurement devices.
See S1 Fig for the table with device images.
Table 2.
Special technical information for scientific studies.
Fig 2.
Illustration of the morphological signal quality process in a sliding window of an ECG (number of cardiac cycles k = 8).
Fig 3.
Examples of ECG segments with different signal quality measured by quantifying the similarity of ECG templates.
Very clean segments have a morphSQ near to zero, noisy segments have values larger than 0.1.
Fig 4.
Illustration of a Smith-Waterman-like algorithm for RR interval sequences to identify incorrect RR intervals in a query sequence.
Table 3.
Morphological Signal Quality Index (morphSQI) expressed as mean over all participants (average SD) and proportion of sufficient morphological signal quality (morphSQ < 10%).
Table 4.
QRS detection performance expressed by false positive (FP) and false negative (FN) counts and positive predictive values (PPV) and false negative rates (FNR) from bandpass-filtered (3–20 Hz) ECGs (heartbeats annotated by eplimited).
Fig 5.
QRS detection performance expressed by F1 scores of the used devices per experimental phase with individual curve progression.
Table 5.
QRS detection performance through RR interval alignment using the modified Smith-Waterman algorithm for numericals.
Errors are expressed by false positive counts (FP), false negative counts (FN), and the number of misplaced annotations from unsynchronized data (50 ms tolerance).