2. Artifact correction

1. Remove spikes: Clean up short period of noise. Figure 1. Once you have identified spikes then go to Transform→ Math Functions→ Connect Endpoints to remove them. Remember to select the point right before and right after the spike to connect endpoints.
2. Identify long period (>30s) of missing data in the focus area. (Refer to Helpful pointers for what is a focus area?)
3. For CNAP blood pressure BP channel #5, first, do Low pass filter at 50 Hz. Transform→Digital filter->FIR→Low Pass at 50 Hz (Figure 2). Then select Analysis→ Hemodynamic→ ABP classifier for the BP data. This will help identify the systole and diastole correctly (Figure 3). Remember to go Diastole to next Diastole (trough to trough for selecting systole when prompted) by the ABP classifier.

4. For Respiration signal channel# 2: (start respiration and ECG artifact correction simultaneously by doing step 4. b.)

   a. Do Low pass filtering at 60 Hz. (Transform→ Digital filters→ FIR→ Low Pass) for respiration channel.

   b. Go to Analysis→ HRV and RSA→ RSA time series. Make sure you select text only. If you select graph only as in A of Figure 10 make sure to delete the additional channels that will get created. It will still create a waveform called heart period, please remove it and you will notice that at the bottom journal section it outputs some RSA analysis values. For now, ignore the text. Y Heart Period is the R-R interval. 

   c. Once it starts putting inspire start at trough and expire start at peak. If there are errors then use Transform→ Math Functions→ Connect endpoints to remove error. If you need to add an inspire or expire start use flag and edit flag as discussed earlier in step 4 e.

5. 
   

   For ECG channel #9: 

a. Go to Analysis->find cycle , select maximum threshold in Cycles/Peaks dialog box at 0.5 mV (Figure 4) to select R peaks that are atleast 0.5 mV. 

c. Put the qrs peak on the maximum in the Output dialog box as shown in the Figure 5. Remember I have duplicated channel 9 of ECG to channel 7 as ECG duplicate. This analysis can be done on Channel 9 if all the work you do is in this ECG channel.

Make sure you select focus areas only in the Selection tab. In the end, you will get something like Figure 6.

d. Identify missed beat. Interpolate or add missed beat manually by going through the whole ECG time series manually.

e. Once you have the QRS peaks, you can easily see which wave missed a peak or which one has double picks. Example, misplaced QRS peak in Figure 7 can be corrected by using ALT- key and dragging the QRS marker from the other peak to the one that does not have it. Make sure that when you are dropping the QRK peak on an R peak, you are precise about the timing of the event. Moving it slightly ahead or before the time R peaked will change the marker position. Also make sure, you switch between select (B tool from helpful pointer #2 ) to edit flag to an hemodynamic event and flag tool to place a flag.

f. Also, you can use Flag and place it on the peak missing a QRS marker and then right click on the Flag go to Type in the popup window as in Figure 8 and change it to Hemodynamics→ ECG complex→ QRS peak.

g. On the original ECG channel # 9, not the duplicate ECG waveform channel do the following and then follow same process in e. and f. as above to ensure no misplaced  ECG marker. (Figure 9.) Following points out how to deal with 

Inverted heart beats

If there is only one heartbeat like this, I’m not sure how it happened, but you can flip it if you select the inverted section, go to Transform Waveform Math. Put the channel you’re working with on box 1, choose multiply, then K for box 2. In the box for K, write -1. This will flip that portion of the data. If the entire graph is this way, you likely connected the leads backwards.

Classify Heart Beats

• Analysis-> Detect and classify heart beats
• Go to the event palette button (+/-) and select “Event Palette”
• You see heart beats listed here as either “unclassifiable beat”, “Normal beat”, “Preventricular contractions”
• Select an event and close the event palette window and it will take you to that event (marked with a “Q”)
• This way, you can remove the unclassifiable beats in case they are noise
• If you select “Normal Heart Beat” from the dropdown menu in the Event Palette, “N”’s appear above every normal heart beat and you can perform analyses on those
• NOTE: once you perform the Heart Rate Variability analysis, the events will not show up as heart beats but as “QRS”, “P”, “O”, “e” and so on

https://www.youtube.com/watch?v=l4-v7euSFho&feature=youtu.be&list=UU7_iupIbHPbJPUC8Mgf3OGg

Figure 1. Example of noise in data evident from respiration and ECG data 

Figure 2. BP Low Frequency Filtering

Figure 3. Correct systole * and Diastole D. 

Figure 4. Use 0.5 mV threshold for ECG R peaks.

 Figure 5. To assign peaks in QRs complex for the focus area.

Figure 6. Sample QRS with R peaks assigned to all cycles.

Figure 7. Misplaced QRS peak.

Figure 8. How to fix a QRS marker on a missed beat.

Figure 9. How to classify QRS complex.

Figure 10. How to classify Respiration marker.