Detection and monitoring of atrial fibrillation
A real-time arrhythmia discrimination method is used in smartphones, which can discriminate between NSR, AF, PACs and PVCs using pulsatile time series collected from a smartphone's camera. To increase the sensitivity of AF detection and add the new capabilities of PVC and PAC identification, the arrhythmia discrimination method of these teachings combines Root Mean Square of Successive RR Differences (RMSSD), Shannon Entropy (ShE) and turning point ratio (TPR), with the Poincare plot, and utilizes the features of pulse rise/fall time and amplitude for arrhythmia discrimination.
1. A computer implemented method for discriminating between normal sinus rhythm without premature ventricular contractions (PVC) or premature atrial contractions (PAC) and atrial fibrillation and premature ventricular contractions (PVC) and premature atrial contractions (PACs), the method comprising:
obtaining root mean squared of successive differences, Shannon entropy and turning point ratio for peak-to-peak (PPI) interval data;
comparing the root mean square of successive differences to a first predetermined threshold; comparing the Shannon entropy to a second predetermined threshold;
comparing the turning point ratio to a third predetermined threshold; and
determining, when each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio is less than a corresponding predetermined threshold, a subject under test has normal sinus rhythm without PAC or PVC.
2. A computer implemented method for discriminating between atrial fibrillation and premature ventricular contractions (PVC) and premature atrial contractions (PACs) and normal sinus rhythm, the method comprising:
obtaining root mean squared of successive differences, Shannon entropy and turning point ratio for peak-to-peak (PPI) interval data;
comparing the root mean square of successive differences to a first predetermined threshold;
comparing the Shannon entropy to a second predetermined threshold;
comparing the tuning point ratio to a third predetermined threshold; wherein not each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio is less than a corresponding predetermined threshold;
demarcating boundaries in a Poincare plot space; the Poincare plot space being a space of time interval between consecutive pulses obtained by sensing variability in heart rate signal;
constructing a Poincare plot of time interval data from the PPI data; the time interval being a time interval between consecutive pulses obtained by sensing variability in heart rate signal from the subject under test;
identifying repeating patterns in the Poincare plot;
obtaining updated data by subtracting the repeating patterns from the time interval data from the subject under test;
obtaining root mean squared of successive differences, Shannon entropy and turning point ratio for the updated data;
comparing the root mean square of successive differences for the updated data to the first predetermined threshold;
comparing the Shannon entropy for the updated data to a second predetermined threshold;
comparing the turning point ratio for the updated data to the third predetermined threshold; and
determining, when not each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio for the updated data is less than a corresponding predetermined threshold, that the subject under test has AF.
3. A computer implemented method for discriminating between normal sinus rhythm (NSR) with premature ventricular contractions (PVC) or normal sinus rhythm (NSR) with premature atrial contractions (PACs) and normal sinus rhythm without PVC or PAC and atrial fibrillation, the method comprising:
obtaining root mean squared of successive differences, Shannon entropy and turning point ratio for peak-to-peak (PPI) interval data;
comparing the root mean square of successive differences to a first predetermine threshold;
comparing the Shannon entropy to a second predetermined threshold;
comparing the turning point ratio to a third predetermined threshold; wherein not each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio is less than a corresponding predetermined threshold;
demarcating boundaries in a Poincare plot space; the Poincare plot space being a space of time interval between consecutive pulses obtained by sensing variability in heart rate signal;
constructing a Poincare plot of time interval data from the PPI data; the time interval being a time interval between consecutive pulses obtained by sensing variability in heart rate signal from the subject under test;
identifying repeating patterns in the Poincare plot;
obtaining updated data by subtracting the repeating patterns from the time interval data from the subject under test;
obtaining root mean squared of successive differences, Shannon entropy and turning point ratio for the updated data;
comparing the root mean square of successive differences for the updated data to the first predetermined threshold;
comparing the Shannon entropy for the updated data to a second predetermined threshold;
comparing the turning point ratio for the updated data to the third predetermined threshold; wherein each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio for the updated data is less than a corresponding predetermined threshold;
determining pulse rise times and false full-time for a pulse time series from which the PPI data was derived;
determining a difference between pulse rise times of consecutive pulses;
determining a difference between pulse fall times of consecutive pulses;
comparing the difference between pulse rise times of consecutive pulses to a fourth predetermined threshold;
comparing the difference between pulse fall times of consecutive pulses to a fifth predetermined threshold; and
determining from comparison of the difference between pulse rise times of consecutive pulses to the fourth predetermined threshold and comparison of the difference between pulse fall times of consecutive pulses to the fifth predetermined threshold, whether the subject under test has one of normal sinus rhythm (NSR) with PAC or NSR with PVC.
4. A system for discriminating between normal sinus rhythm without premature ventricular contractions or premature atrial contractions and atrial fibrillation and premature ventricular contractions (PVC) and premature atrial contractions (PACs), the system comprising:
one or more processors; and
computer usable media having computer readable code embodied there in that, when executed in the one or more processors, causes the one or more processors to:
obtain root mean squared of successive differences, Shannon entropy and turning point ratio for peak-to-peak (PPI) interval data;
compare the root mean square of successive differences to a first predetermined threshold;
compare the Shannon entropy to a second predetermined threshold;
compare the turning point ratio to a third predetermined threshold; and
determine, when each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio is less than a corresponding predetermined threshold, a subject under test has normal sinus rhythm without PAC or PVC.
5. A system for discriminating between atrial fibrillation and premature ventricular contractions (PVC) and premature atria contractions (PACs) and normal sinus rhythm, the system comprising:
one or more processors; and
computer usable media having computer readable code embodied there in that, when executed in the one or more processors, causes the one or more processors to:
obtain root mean squared of successive differences, Shannon entropy and turning point ratio for peak-to-peak (PPI) interval data:
compare the root mean square of successive differences to a first predetermined threshold;
compare the Shannon entropy to a second predetermined threshold;
compare the turning point ratio to a third predetermined threshold;
determine, whether each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio is less than a corresponding predetermined threshold; wherein not each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio is less than a corresponding predetermined threshold;
demarcate boundaries in a Poincare plot space; the Poincare plot space being a space of time interval between consecutive pulses obtained by sensing variability in heart rate signal;
construct a Poincare plot of time interval data from the PPI data; the time interval being a time interval between consecutive pulses obtained by sensing variability in heart rate signal from the subject under test;
identify repeating patterns in the Poincare plot;
obtain updated data by subtracting the repeating patterns from the time interval data from the subject under test;
obtain root mean squared of successive differences, Shannon entropy and turning point ratio for the updated data;
compare the root mean square of successive differences for the updated data to the first predetermined threshold;
compare the Shannon entropy for the updated data to a second predetermined threshold;
compare the turning point ratio for the updated data to the third predetermined threshold; and
determine, when each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio for the updated data is not less than a corresponding predetermined threshold, the subject under test has AF.
6. A system for discriminating between normal sinus rhythm (NSR) with premature ventricular contractions (PVC) or normal sinus rhythm (NSR) with premature atrial contractions (PACs) and normal sinus rhythm without PVC or PAC and atrial fibrillation, the system comprising:
one or more processors; and
computer usable media having computer readable code embodied there in that,
when executed in the one or more processors, causes the one or more processors to:
obtain root mean squared of successive differences, Shannon entropy and turning point ratio for peak-to-peak (PPI) interval data;
compare the root mean square of successive differences to a first predetermine threshold;
compare the Shannon entropy to a second predetermined threshold;
compare the turning point ratio to a third predetermined threshold;
determine, whether each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio is less than a corresponding predetermined threshold; wherein not each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio is less than a corresponding predetermined threshold;
demarcate boundaries in a Poincare plot space; the Poincare plot space being a space of time interval between consecutive pulses obtained by sensing variability in heart rate signal;
construct a Poincare plot of time interval data from the PPI data; the time interval being a time interval between consecutive pulses obtained by sensing variability in heart rate signal from the subject under test;
identify repeating patterns in the Poincare plot;
obtain updated data by subtracting the repeating patters from the time interval data from the subject under test;
obtain root mean squared of successive differences, Shannon entropy and turning point ratio for the updated data:
compare the turning point ratio for the updated data to the third predetermined threshold;
wherein each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio for the updated data is less than a corresponding predetermined threshold;
determine pulse rise times and pulse fall-time for a pulse time series from which the PPI data was derived;
determine a difference between pulse rise times of consecutive pulses;
determine a difference between pulse fall times of consecutive pulses;
compare the difference between pulse rise times of consecutive pulses to a fourth predetermined threshold;
compare the difference between pulse fall times of consecutive pulses to a fifth predetermined threshold; and
determine from comparison of the difference between pulse rise times of consecutive pulses to the fourth predetermined threshold and comparison of the difference between pulse fall times of consecutive pulses to the fifth predetermined threshold, whether the subject under test has one of normal sinus rhythm (NSR) with PAC or NSR with PVC.
7. A non-transitory computer usable medium having computer readable code embodied therein, the computer readable code, when executed in one or more processors, causing the one or more processors to:
obtain root mean squared of successive differences, Shannon entropy and turning point ratio for peak-to-peak (PPI) interval data;
compare the root mean square of successive differences to a first predetermined threshold;
compare the Shannon entropy to a second predetermined threshold;
compare the turning point ratio to a third predetermined threshold; and
determine, when each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio is less than a corresponding predetermined threshold, a subject under test has normal sinus rhythm without PAC or PVC.
8. A non-transitory computer readable medium having computer readable code embodied therein, the computer readable code, when executed in one or more processors, causing the one or more processors to:
obtain root mean squared of successive differences, Shannon entropy and turning point ratio for peak-to-peak (PPI) interval data;
compare the root mean square of successive differences to a first predetermined threshold;
compare the Shannon entropy to a second predetermined threshold;
compare the turning point ratio to a third predetermined threshold;
determine, whether each of the root mean square of successive differences, the Shannon entropy, and the turning pint ratio is less than a corresponding predetermined threshold; wherein not each of the root mean square of successive differences the Shannon entropy, and the turning point ratio is less than a corresponding predetermined threshold;
demarcate boundaries in a Poincare plot space; the Poincare plot space being a space of time interval between consecutive pulses obtained by sensing variability in heart rate signal;
construct a Poincare plot of time interval data from the PPI data; the time interval being a time interval between consecutive pulses obtained by sensing variability in heart rate signal from the subject under test;
identify repeating patterns in the Poincare plot;
obtain updated data by subtracting the data in the patterns from the time interval data from the subject under test;
obtain root mean squared of successive differences, Shannon entropy and turning point ratio for the updated data;
compare the root mean square of successive differences for the updated data to the first predetermined threshold;
compare the Shannon entropy for the updated data to a second predetermined threshold;
compare the turning point ratio for the updated data to the third predetermined threshold; and
determine, when each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio for the updated data is not less than a corresponding predetermined threshold, the subject under test has AF.
9. A non-transitory computer readable medium having computer readable code embodied therein, the computer readable code, when executed in one or more processors, causing the one or more processors to:
obtain root mean squared of successive differences, Shannon entropy and turning point ratio for peak-to-peak (PPI) interval data;
compare the root mean square of successive differences to a first predetermined threshold;
compare the Shannon entropy to a second predetermined threshold;
compare the turning point ratio to a third predetermined threshold;
determine, whether each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio is less than a corresponding predetermined threshold wherein not each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio is less than a corresponding predetermined threshold;
demarcate boundaries in a Poincare plot space; the Poincare plot space being a space of time interval between consecutive pulses obtained by sensing variability in heart rate signal;
construct a Poincare plot of time interval data from the PPI data; the time interval being a time interval between consecutive pulses obtained by sensing variability in heart rate signal from the subject under test;
identify repeating patterns in the Poincare plot;
obtain updated data by subtracting the data in the patterns from the time interval data from the subject under test;
obtain root mean squared of successive differences, Shannon entropy and turning point ratio for the updated data;
compare the root mean square of successive differences for the updated data to the first predetermined threshold;
compare the Shannon entropy for the updated data to a second predetermined threshold;
compare the turning point ratio for the updated data to the third predetermined threshold;
wherein each of the root mean square of successive differences, the Shannon entropy, and the turning point ratio for the updated data is less than a corresponding predetermined threshold;
determine pulse rise times and pulse fall-times for a pulse time series from which the PPI data was derived;
determine a difference between pulse rise times of consecutive pulses;
determine a difference between pulse fall times of consecutive pulses;
compare the difference between pulse rise times of consecutive pulses to a fourth predetermined threshold;
compare the difference between pulse fall times of consecutive pulses to a fifth predetermined threshold; and
determine from comparison of the difference between pulse rise times of consecutive pulses to the fourth predetermined threshold and comparison of the difference between pulse fall times of consecutive pulses to the fifth predetermined threshold, whether the subject under test has one of normal sinus rhythm (NSR) with PAC or NSR with PVC.
10. The method of claim 3 further comprising:
determining, when the difference between pulse rise times of consecutive pulses is less than the fourth predetermined threshold and the difference between pulse fall times of consecutive pulses is less than the fifth predetermined threshold, that the subject under test has NSR with PAC.
11. The method of claim 3 further comprising:
determining, when the difference between pulse rise times of consecutive pulses is not less than the fourth predetermined threshold and the difference between pulse fall times of consecutive pulses is not less than the fifth predetermined threshold, that the subject under test has NSR with PVC.
12. The system of claim 6 wherein the computer readable code, when executed in the one or more processors, further causes the one or more processors to:
determine, when the difference between pulse rise times of consecutive pulses is less than the fourth predetermined threshold and the difference between pulse fall times of consecutive pulses is less than the fifth predetermined threshold, that the subject under test has NSR with PAC.
13. The system of claim 6 wherein the computer readable code, when executed in the one or more processors, further causes the one or more processors to:
determine, when the difference between pulse rise times of consecutive pulses is not less than the fourth predetermined threshold and the difference between pulse fall times of consecutive pulses is not less than the fifth predetermined threshold, that the subject under test has NSR with PVC.
14. The non-transitory computer readable medium of claim 8 wherein the computer readable code, when executed in the one or more processors, further causes the one or more processors to:
determine, when the difference between pulse rise times of consecutive pulses is less than the fourth predetermined threshold and the difference between pulse fall times of consecutive pulses is less than the fifth predetermined threshold, that the subject under test has NSR with PAC.
15. The non-transitory computer readable medium of claim 8 wherein the computer readable code, when executed in the one or more processors, further causes the one or more processors to:
determine, when the difference between pulse rise times of consecutive pulses is not less than the fourth predetermined threshold and the difference between pulse fall times of consecutive pulses is not less than the fifth predetermined threshold, that the subject under test has NSR with PVC.