Atrial fibrillation detection method and apparatus

• US 20020143266A1
• Filed: 12/28/2000
• Published: 10/03/2002
• Est. Priority Date: 12/28/2000
• Status: Active Grant

• Current Assignee: GE Medical Systems Information Technologies Incorporated (General Electric Company)
• Original Assignee: GE Medical Systems Information Technologies Incorporated (General Electric Company)
• Inventors: Bock, Elizabeth Anne
• Granted Patent: US 6,490,479 B2
• Time in Patent Office: Days
• US Class Current: 600/518
• CPC Class Codes: A61B 5/02405 Determining heart rate variability
• CPC Class Codes: A61B 5/361 Detecting fibrillation
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems

Abstract

A method and apparatus to detect irregular heart activity based upon a ventricular activity analysis, P wave activity, similarities in R wave to R wave intervals, and a state evaluation. The invention includes a beat classification module that receives ECG information as an input. The beat classification module determines whether the heart beat being analyzed falls within classifications that are suitable for use in analyzing whether an irregular condition exists. If the beat falls within a class suitable for analysis, the ECG information is fed to an interval calculator. The interval calculator determines the interval between successive R waves. The information from the interval calculator is provided to a probability engine and to a contextual analysis module. The probability engine is designed to detect atrial fibrillation based upon beat classification and RR interval values from the interval calculator. The probability engine outputs a state variable that indicates a probability of whether an irregular condition is present. The contextual analysis module matches predefined maps to a running map of the current ECG information. The contextual analysis module also determines the similarity between consecutive RR intervals and checks for sequences of matching classes. ECG information is also supplied to a P wave detection module to detect the presence of P waves. Information from the probability analysis engine, the contextual analysis module, and the P wave detection module is then provided to a state evaluation module. The state evaluation module uses the outputs of the three modules to determine whether an irregular condition exists.

Claims

1. An arrhythmia detection apparatus comprising:

a beat classification module having an input to receive ECG information;

an interval calculator coupled to the beat classification module and operable to calculate an interval between recurring events;

a probability engine coupled to the interval calculator and operable to output a state variable;

a contextual analysis module coupled to the interval calculator and operable to match predefined maps to a current beat map and to generate a contextual output;

a P wave detection module coupled to the beat classification module and operable to produce a detection output; and

a state evaluation module coupled to the probability engine, the contextual analysis module, and the P wave detection module and operable to determine an arrhythmia condition based on the state variable, the contextual output, and the detection output.

2. An apparatus as claimed in claim 1, wherein the interval calculated by the interval calculator is an interval between successive R waves.

3. An apparatus as claimed in claim 1, wherein the probability engine implements an HMM analysis.

4. An apparatus as claimed in claim 3, wherein the probability engine is operable to classify intervals between recurring events using prematurity ratio thresholds.

5. An apparatus as claimed in claim 1, wherein the P wave detection module implements a sixth difference analysis.

6. An apparatus as claimed in claim 1, wherein the state evaluation module is operable to generate output in the form of alarm information.

7. An apparatus as claimed in claim 6, wherein the state evaluation module delays the output of alarm information until an alarm condition has existed for a predetermined amount of time.

8. An apparatus as claimed in claim 6, wherein the state evaluation module is operable to generate output in the form of trend information.

9. An apparatus as claimed in claim 1, wherein the contextual analysis module includes a buffer of interval values.

10. An apparatus as claimed in claim 9, wherein the contextual analysis module generates the contextual output based on a percentage of similarity of the intervals.

11. A method of detecting an arrhythmia from ECG information, the method comprising:

classifying the ECG information;

determining intervals between recurring events in the ECG information;

determining a probability that an irregular condition exists based on classifying the ECG information and determining an interval between recurring events;

generating a state variable based upon the determined probability;

generating a contextual output based on similarities in intervals between recurring events;

determining the presence of a P wave in the ECG information;

generating a detection output based on determining the presence of a P wave; and

determining the existence of the irregular condition based on the state variable, the contextual output, and the detection output.

12. A method as claimed in claim 11, wherein classifying ECG information includes detecting and correlating beats based on morphology.

13. A method as claimed in claim 11, wherein generating a state variable based upon the determined probability includes using a lookup table.

14. A method as claimed in claim 11, wherein determining a probability that an irregular condition exists includes maintaining a plurality of structures.

15. A method as claimed in claim 11, wherein generating a contextual output includes matching predefined maps to a current map of the ECG information.

16. A method as claimed in claim 11, wherein generating a contextual output includes performing a percent similarity test.

17. A method as claimed in claim 11, wherein determining the presence of a P wave includes calculating sixth differences of the ECG information.

19. An apparatus as claimed in claim 18, wherein the interval calculated by the interval calculator is an interval between successive waves.

20. An apparatus as claimed in claim 18, wherein the probability engine implements an HMM analysis.

21. An apparatus as claimed in claim 20, wherein the probability engine is operable to classify intervals between recurring events using prematurity ratio thresholds.

22. An apparatus as claimed in claim 18, wherein the wave detection module implements a sixth difference analysis.

23. An apparatus as claimed in claim 18, wherein the state evaluation module is operable to generate output in the form of alarm information.

24. An apparatus as claimed in claim 23, wherein the state evaluation module delays the output of alarm information until an alarm condition has existed for a predetermined amount of time.

25. An apparatus as claimed in claim 23, wherein the state evaluation module is operable to generate output in the form of trend information.

26. An apparatus as claimed in claim 18, wherein the contextual analysis module includes a buffer of interval values.

27. An apparatus as claimed in claim 26, wherein the contextual analysis module generates the contextual output based on a percentage of similarity of the intervals.

29. A method as claimed in claim 28, wherein classifying the physiological information includes detecting and correlating beats based on morphology.

30. A method as claimed in claim 28, wherein generating a state variable based upon the determined probability includes using a lookup table.

31. A method as claimed in claim 28, wherein determining a probability that an irregular condition exists includes maintaining a plurality of structures.

32. A method as claimed in claim 28, wherein generating a contextual output includes matching predefined maps to a current map of the physiological information.

33. A method as claimed in claim 28, wherein generating a contextual output includes performing a percent similarity test.

34. A method as claimed in claim 28, wherein determining the presence of a predetermined wave includes calculating sixth differences of the physiological information.

18. A physiological condition detection apparatus comprising:

a classification module having an input to receive physiological information;

an interval calculator coupled to the classification module and operable to calculate intervals between recurring events;

a probability engine coupled to the interval calculator and operable to output a state variable;

a contextual analysis module coupled to the interval calculator and operable to generate a contextual output based on the classification of the physiological information and the intervals between recurring events;

a wave detection module coupled to the classification module and operable to produce a detection output; and

a state evaluation module coupled to the probability engine, the contextual analysis module, and the wave detection module and operable to determine a physiological condition based on the state variable, the contextual output, and the detection output.

28. A method of analyzing physiological information, the method comprising:

classifying the physiological information;

determining intervals between recurring events in the physiological information;

determining a probability that an irregular condition exists based on classifying the physiological information and determining an interval between recurring events;

generating a state variable based upon the determined probability;

generating a contextual output based on similarities in intervals between recurring events;

determining the presence of a predetermined wave in the physiological information;

generating a detection output based on determining the presence of a predetermined wave; and

determining the existence of the irregular condition based on the state variable, the contextual output, and the detection output.