Medical device and method for detecting a ventricular arrhythmia event

1. A method for a medical device for detecting a ventricular arrhythmia event comprising:

receiving by input circuitry of the medical device an electrocardiogram (ECG) signal;

locating by processing circuitry of the medical device three or more QRS complexes within the ECG signal;

locating by the processing circuitry R peaks within the three or more QRS complexes;

determining by the processing circuitry a Q onset and a S offset;

determining by the processing circuitry an RR interval between consecutive R peaks;

determining by the processing circuitry search window boundaries between which a T wave and a P wave are located based upon the RR interval;

providing by the processing circuitry a calculated T wave amplitude threshold and a calculated P wave amplitude threshold;

determining by the processing circuitry a T wave location within the search window boundaries based upon the calculated T wave amplitude threshold;

determining by the processing circuitry a P wave location within the search window boundaries based upon the calculated P wave amplitude threshold;

extracting fiducial points by a fiducial point extraction block, wherein the fiducial points include at least a peak of the P wave location, the Q onset, and the S offset; and

determining by feature extraction circuitry coupled to the fiducial point extraction block interval variabilities between the fiducial points, wherein the interval variabilities include a combination of QP interval variability and SP interval variability.

2. The method for the medical device of claim 1 further including detecting by machine learning circuitry ventricular arrhythmia based upon the interval variabilities determined by the feature extraction circuitry.

3. The method for the medical device of claim 2 wherein the interval variabilities are further selected from the group consisting of RT interval variability, TR interval variability, PQ interval variability, and PS interval variability.

4. The method for the medical device of claim 1 wherein a T wave fiducial point is a peak of the T wave at the T wave location determined by a T wave delineation block.

5. The method for the medical device of claim 1 wherein the peak of the P wave at the P wave location is determined by a P wave delineation block.

6. The method for the medical device of claim 5 wherein other fiducial points are selected from the group consisting of P offset, R peak, and T offset.

7. The method for the medical device of claim 1 further comprises:

taking by the processing circuitry a derivative of a filtered ECG signal received by the input circuitry;

squaring by the processing circuitry the derivative of the filtered ECG signal; and

integrating by the processing circuitry a square of the derivative of the filtered ECG signal for a duration of at least two heartbeat cycles.

8. The method for the medical device of claim 1 further including removing, by the input circuitry, unwanted noise signals mixed with a received ECG signal.

9. A medical device for detecting a ventricular arrhythmia event comprising:

input circuitry configured to receive an electrocardiogram (ECG) signal;

processing circuitry coupled to the input circuitry and comprising:

a QRS complex demarcation block configured to locate three or more consecutive QRS complexes within the ECG signal;

an R peak detection block configured to locate R peaks within the located QRS complexes;

a Q onset and S offset block configured to determine a Q onset and an S offset;

an RR demarcation block configured to determine an RR interval between consecutive R peaks located by the R peak detection block;

a search windows calculator block configured to determine search window boundaries between which a T wave and a P wave are located based upon the RR interval;

a T and P wave thresholds calculator block configured to provide a calculated T wave amplitude threshold and a calculated P wave amplitude threshold;

a T wave delineation block configured to determine a T wave location within the search window boundaries based upon the calculated T wave amplitude threshold;

a P wave delineation block configured to determine a P wave location within the search window boundaries based upon the calculated P wave amplitude threshold;

a fiducial point extraction block configured to extract fiducial points within the search window boundaries, wherein the fiducial points include at least the P wave location, the Q onset and the S offset; and

feature extraction circuitry coupled to the fiducial point extraction block and configured to determine interval variabilities between the fiducial points, wherein the interval variabilities include a combination of QP interval variability and SP interval variability.

10. The medical device of claim 9 further including machine learning circuitry configured to detect ventricular arrhythmia based upon the interval variabilities determined by the feature extraction circuitry.

11. The medical device of claim 9 wherein the interval variabilities are further selected from the group consisting of RT interval variability, TR interval variability, PQ interval variability, and PS interval variability.

12. The medical device of claim 9 wherein a T wave fiducial point is a peak of the T wave at the T wave location determined by the T wave delineation block.

13. The medical device of claim 9 wherein a P wave fiducial point is a peak of the P wave at the P wave location determined by the P wave delineation block.

14. The medical device of claim 13 wherein other fiducial points are selected from the group consisting of P offset, Q onset, R peak, S offset, and T offset.

15. The medical device of claim 9 wherein the processing circuitry further comprises:

a differentiation block coupled to an output of the input circuitry and configured to take a derivative of a filtered ECG signal;

a squaring block coupled to an output of the differentiation block and configured to square the derivative of the filtered ECG signal; and

a moving window integral block coupled between an output of the squaring block and an input of the QRS complex demarcation block, wherein the moving window integral block is configured to integrate the square of the derivative of the filtered ECG signal for a duration of at least two heartbeat cycles.

16. The medical device of claim 9 wherein the input circuitry includes at least one filter configured to remove unwanted noise signals mixed with a received ECG signal.