How R-R Data is Filtered
The Polar transmitter detects all heart beats. The recording may have some erroneous values due to measurement errors, or extra beats of the heart may appear as sudden peaks in the R-R data. An automatic filtering process in the Polar ProTrainer 5 software can be used to post process the accumulated R-R interval data.
Polar products recording R-R data do not store the ECG data in their memory. Therefore, it is usually impossible to know whether certain extraordinary data was caused by an extra beat or a measurement error. Normally this does not greatly affect the measurement results as all extraordinary data must be removed before further analysis. However, it is still very important to carefully wet the electrodes and tighten the transmitter sufficiently to ensure good skin contact and this way ensure accurate measurement results. In R-R recording it is advisable to attach the receiver for example on a waist belt to ensure perfect transmission of every heart beat.
Filtering process method in the software
First, the algorithm tries to find all possible measurement errors by using median and moving average based filtering methods. The criteria adapt to the R-R data to be checked: for example if your normal R-R data is very stable, relatively small peaks will be interpreted as errors. On the other hand, if your normal R-R data includes very large variability, the criteria will be looser. Very small errors are never corrected; you can determine this limit by changing the Minimum Protection Zone setting in the software.
In the correction phase, the algorithm computes several more matching 'candidates' to substitute the detected errors. Before drawing the preview curve, which shows the suggested corrections, the algorithm checks the difference between previous and the next accepted R-R interval, and then makes the series of corrected values (usually 2 - 4 intervals) to follow the same differential coefficient (derivative).
It is the special property of the error correction algorithm that it exactly maintains the total time of the recording, i.e. the sum of R-R intervals is exactly the same as the elapsed time. (This is not the case with conventional error correction procedures, such as linear interpolation.) This property has several benefits. Sometimes it could have one drawback, namely that supraventricular extra beats would not be removed. Instead these abnormal intervals would still be seen as sudden peaks after the correction. For this reason, by crossing a checkbox beside the scissors in the Error Correction box, you can force the algorithm to also correct these sudden peaks regardless of the small effects in the total time.
How to correct errors in R-R data heart rate file
1. Open the heart rate file saved in R-R mode.
2. Right-click on the heart rate curve view and select Error Correction. Alternatively you can select Error Correction from the Edit menu.
The preview of the suggested correction is displayed on the curve. By changing the options in the Error Correction box you can try to find the best possible result for the correction. When you have changed any option in the box click the Preview button to see how the changes effect in the HR curve.
For detailed information of correcting errors in heart rate curve see the software online help.