Skip to main content
. Author manuscript; available in PMC: 2016 Sep 15.
Published in final edited form as: Biosens Bioelectron. 2015 Apr 14;71:150–157. doi: 10.1016/j.bios.2015.04.027

Figure 3.

Figure 3

ECG Recording and Signal Processing

(a) Raw data of a representative ECG waveform in the presence of gill motion and low frequency baseline wandering. R waves were detected by using an in-house R wave detection algorithm. Specifically, all of the individual ECG signals were aligned with the peak R waves and were summed together to filter the unsynchronized breathing artifacts and wandering baselines. (b) A summed ECG waveform over 5 minutes was highlighted as the red ECG pattern while the black waveforms were the individual ECG signals. (c) The residual baseline was then fitted with a cubic spline curve (blue) and subtracted from the summed ECG waveform.