Nontraditional Electrocardiogram and Algorithms for Inconspicuous In-Home Monitoring: Comparative Study

Figures

• 

  Figure 1. Stainless-steel electrodes are integrated into the seat and are connected directly to the integrated ECG analog front-end. The output of each front-end is low-impedance, allowing for a low-noise connection to the differential ECG amplifier inside of the BIOPAC MP150 system (left). The resulting full-scale ECG signal is acquired on a NI CompactRIO DAQ. For the HF subject study, all instrumentation and data acquisition circuitry are integrated into the toilet seat (right). DAQ: data acquisition system; ECG: electrocardiogram; bECG: buttocks ECG; NI: National Instruments; HF: heart failure.

• 

  Figure 2. Signal quality is calculated from the bandpass filtered (1-45 Hz) and notch filtered (60 Hz) bECG (top) using a moving kurtosis (middle) and moving SD (bottom). The first signal quality index (moving kurtosis) rejects regions with in-band noise that have a kurtosis below a threshold of 3.6. The second signal quality index (moving SD) rejects noisy spikes that have a moving SD greater than 1.25 times the overall SD. R-peak delineations are shown as red circles to indicate where each beat is located within regions that have passed the signal quality test. bECG: buttocks electrocardiogram.

• 

  Figure 3. The modified PT algorithm processes the raw ECG signal (top) to isolate QRS complexes for delineation using a bandpass filter and derivative (middle). Dual thresholds (dashed and dotted lines) are calculated using a moving window average and are continually updated (bottom), rather than updating only when a delineated feature is found. The resultant delineations, shown as red circles, are often shifted from the R-peak location, necessitating a subsequent refinement stage. ECG: electrocardiogram; PT: Pan–Tompkins.

• 

  Figure 4. Ensemble averaging aligns multiple beats based on the R-wave peak location to calculate an average beat. Individual beats that passed all the rejection criteria are shown as thin light-colored lines, and the resulting ensemble averaged beat is shown as a thick dark line.

• 

  Figure 5. Box plots across normative recordings (N=140) show the correlation between the bECG and the standard ECG leads. Leads aVR, V1, aVL, and V2 have been inverted to facilitate visual comparison. Leads are organized from left to right by highest to lowest correlation. The top 3 correlated leads were Lead II, aVR, and aVF with median correlations of .904, .899, and .893, respectively. Statistical significance was tested using a paired Student t-test (*P<.05, **P<.01, and ***P<.001). bECG: buttocks electrocardiogram, ECG: electrocardiogram.

• 

  Figure 6. The signal quality of the bECG can vary across subjects and measurements because it is captured using active, dry electrodes. Three examples of signal quality are shown for the bECG (top) compared with a time synchronized Lead II ECG (bottom). The best signal quality example (a) shows the best signal quality achieved from the bECG study. The average example quality (b) shows the typical signal quality of the bECG. The poor example (c) shows a very noisy waveform that passed the signal quality check before analysis. bECG: buttocks electrocardiogram; ECG: electrocardiogram.

• 

  Figure 7. An example waveform from a heart failure subject during a period of arrhythmia, demonstrates that the bECG has sufficient quality to perform single-lead based rhythm analysis for those with disease states. bECG: buttocks electrocardiogram.

• 

  Figure 8. Heart rate (left) and heart rate variability (right) extracted from the bECG signal for normative rest (red x), normative post-stress (blue o), and heart failure (green +) are closely aligned with those extracted from the gold standard Lead II ECG. The dashed line shows the mean error. The dotted lines show 1.96 times the SD, corresponding to a 95% limits of agreement. bECG: buttocks electrocardiogram; ECG: electrocardiogram.

• 

  Figure 9. QRS duration (left) and QTc interval (right) extracted from the bECG signal for normative rest (red x), normative poststress (blue o), and heart failure (green +) are closely aligned with those extracted from the gold-standard Lead II ECG. The dashed line shows the mean error. The dotted lines show 1.96 times the SD, corresponding to a 95% limits of agreement. bECG: buttocks electrocardiogram; ECG: electrocardiogram; QTc: corrected QT.