Cardiac biopotential analysis system and method

• US 5,020,540 A
• Filed: 08/22/1989
• Issued: 06/04/1991
• Est. Priority Date: 10/09/1987
• Status: Expired due to Term

• Current Assignee: Imperial Bank (Imperial Bancorp)
• Original Assignee: Aspect Medical Systems, Inc. (Medtronic Plc)
• Inventors: Chamoun, Nassib G.
• Primary Examiner(s): Jaworski, Francis J.
• Assistant Examiner(s): Manuel, George
• Application Number:
  US07/396,990
• Time in Patent Office: 651 Days
• Field of Search: 128/696, 128/, 128/697, 128/702, 128/703, 128/708, 364/413.05, 364/413.06
• US Class Current: 600/509
• CPC Class Codes: A61B 5/339 Displays specially adapted therefor
• CPC Class Codes: A61B 5/349 Detecting specific parameters of the electrocardiograph cycle
• CPC Class Codes: A61B 5/35 by template matching
• CPC Class Codes: A61B 5/363 Detecting tachycardia or bradycardia
• CPC Class Codes: A61B 5/413 Monitoring transplanted tissue or organ, e.g. for possible rejection reactions after a transplant
• CPC Class Codes: A61B 5/7257 using Fourier transforms
• CPC Class Codes: A61B 5/7275 Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor

Abstract

Disclosed is a cardiac biopotential analysis system and method for detecting and quantifying in a noninvasive manner nonlinear dynamic patterns of depolarization and repolarization on a beat to beat basis in real time using bispectral analysis. A suitable body surface electrode acquires the signal from a region of interest. The signal is then amplified, digitized, and transmitted to a host microcomputer where an arrhythmia-free QRST complex is chosen interactively. Using standard cross-correlation methods a suitable number of complexes are extracted from the record. The extracted complexes are used to compute bispectral parameters using a frequency domain or a parametric based approach. A reference clinical database is used to identify frequency pairs most sensitive to particular interventions or diagnostic states of interest. The values at these frequency pairs are then extracted from the patient'"'"'s bispectral arrays. The ensemble of values for the particular diagnostic determination under consideration is used to compute a single value index which serves as the diagnostic criterion by which the patient'"'"'s state is judged. In this way the quantification of nonlinear dynamic properties of the frequency structure of the QRST yields information about the presence and extent of coronary artery disease (CAD), myocardial ischemia, cardiac electrical stability, risk of malignant ventricular arrhydthmia, site(s) of origin of malignant arrhythmias, extent of malignancy of arrhythmias, degree of antiarrhythmic drug efficacy, neural and humoral inputs to the heart, pump function/ejection fraction, and ongoing organ rejection in cardiac transplant patients.

Claims

1. A method of noninvasively detecting cardiac phenomena comprising the steps of:

acquiring electrocardiographic signals from a body surface of a subject being analyzed through a surface electrode;

determining a waveform template representing a QRST waveform in said subject;

selecting a number of signals that match said waveform template and storing said signals for processing;

generating bispectral values from selected signals, comparing said bispectral values to reference bispectral values to analyze cardiac phenomena.

2. The method of noninvasively detecting cardiac phenomena of claim 1 wherein said generated bispectral values are autobispectral density values.

3. The method of noninvasively detecting cardiac phenomena of claim 1 wherein said generated bispectral values are autobicoherence values.

4. The method of noninvasively detecting cardiac phenomena of claim 1 wherein said generated bispectral values are autobiphase values.

5. The method of noninvasively detecting cardiac phenomena of claim 1 wherein said step of selecting a number of signals that match said waveform template comprises the step of selecting pairs of successive time series records that are used in the generation of bispectral values.

6. The method of noninvasively detecting cardiac phenomena of claim 5 wherein said generated bispectral values are crossbispectral density values.

7. The method of noninvasively detecting cardiac phenomena of claim 5 wherein said generated bispectral values are crossbispectral density values.

8. The method of noninvasively detecting cardiac phenomena of claim 5 wherein said generated bispectral values are crossbiphase values.

9. The method of noninvasively detecting cardiac phenomena of claim 1 further comprising the step of comparing said bispectral values to reference bispectral values to detect a presence and extent of coronary artery disease in said subject.

10. The method of noninvasively detecting cardiac phenomena of claim 1 further comprising the step of comparing said bispectral values to reference bispectral values to detect a presence and extent of myocardial ischemia in said subject.

11. The method of noninvasively detecting cardiac phenomena of claim 1 further comprising the step of comparing said bispectral values to reference bispectral values to analyze cardiac electrical stability in said subject.

12. The method of noninvasively detecting cardiac phenomena of claim 1 further comprising the step of comparing said bispectral values to reference bispectral values to analyze a risk of malignant ventricular arrhythmia in said subject.

13. The method of noninvasively detecting cardiac phenomena of claim 1 further comprising the step of comparing said bispectral values to reference bispectral values to locate a site of origin of malignant arrhythmias in said subject.

14. The method of noninvasively detecting cardiac phenomena of claim 1 further comprising the step of comparing said bispectral values to reference bispectral values to analyze an extent of malignancy of arrhythmias in said subject.

15. The method of noninvasively detecting cardiac phenomena of claim 1 further comprising the step of comparing said bispectral values to reference bispectral values to evaluate a degree of antiarrhythmic drug efficacy in said subject.

16. The method of noninvasively detecting cardiac phenomena of claim 1 further comprising the step of comparing said bispectral values to reference bispectral values to examine neutral and humoral inputs to the heart of said subject.

17. The method of noninvasively detecting cardiac phenomena of claim 1 further comprising the step of comparing said bispectral values to reference bispectral values to evaluate pump function or ejection fraction in the heart of said subject.

18. The method of noninvasively detecting cardiac phenomena of claim 1 further comprising the step of comparing said bispectral values to reference bispectral values to monitor ongoing organ rejection wherein said subject is a cardiac transplant patient.

19. The method of noninvasively detecting cardiac phenomena of claim 1 further comprising the step of distinguishing wide-complex supraventricular tachycardia from sustained ventricular tachycardia.

20. The method of noninvasively detecting cardiac phenomena of claim 1 wherein said step of generating bispectral values comprises:

assigning each selected signal matching said waveform template to a time series record;

computing an average complex triple product using Fourier transforms of said time series records;

computing biphase values as the inverse tangent of the quotient of an imaginary part of the average complex triple product divided by a real part of the average complex triple product.

21. The method of noninvasively detecting cardiac phenomena of claim 20 wherein said biphase values are crossbiphase values.

22. The method of noninvasively detecting cardiac phenomena of claim 20 wherein said biphase values are autobiphase values.

23. The method of noninvasively detecting cardiac phenomena of claim 1 wherein said step of generating bispectral values comprises:

assigning each selected signal matching said waveform template to a time series record;

determining Fourier transforms of said time series records;

estimating bispectrum of said time records by computing sequences of third-order moments, forming super matrices using said third-order moments, determining orders of autoregressive model filters of the bispectrum of the Fourier transforms of said time records from ranks of said super matrices, obtaining the skewness and autoregressive filter coefficients by;

solving a set of equations based on said third order moments and said order, and computing said bispectrum from said skewness and autoregressive filter coefficients.

24. The method of noninvasively detecting cardiac phenomena of claim 23 further comprising the step of estimating power spectra of said Fourier transforms, said step of estimating power spectra comprising the steps of:

computing two sequences of autocorrelations and performing a recursion function on each autocorrelation sequence in order to determine final prediction errors, minimums of said final prediction errors being order of the autoregression filters of said power spectra;

performing a recursion function on said autocorrelation sequences using said order of the autoregression filters in order to generate coefficient of said autoregression filters;

computing the power spectra as the final prediction error divided by a square of the magnitude of the Fourier transform of said coefficients.

25. The method of noninvasively detecting cardiac phenomena of claim 24 further comprising the steps of:

computing a real triple product as the product of three power spectra values;

computing an array of bicoherence values as the quotient of bispectral density divided by a square root of said real triple product.

26. The method of noninvasively detecting cardiac phenomena of claim 23 wherein said step of computing bispectrum of said Fourier transforms of said time series records comprises dividing the cubic root of the product of three skewnesses by a triple product of the Fourier transforms of the AR filter coefficient.

27. The method of noninvasively detecting cardiac phenomena of claim 23 further comprising the step of computing an array of bispectral densities by determining the absolute value for each bispectrum value.

28. The method of noninvasively detecting cardiac phenomena of claim 23 further comprising the step of computing an array of biphase values as the inverse tangent of the quotient of an imaginary part of the bispectrum divided by a real part of the bispectrum.

29. The method of noninvasively detecting cardiac phenomena of claim 1 wherein said bispectral values are crossbispectral values generated from consecutive signals from one surface electrode.

30. The method of noninvasively detecting cardiac phenomena of claim 1 wherein said bispectral values are crossbispectral values generated from signals obtained from two surface electrodes.

31. The method of noninvasively detecting cardiac phenomena of claim 1 wherein said template is a sinus rhythm template.

32. The method of noninvasively detecting cardiac phenomena of claim 1 wherein said template is an ectopic beat template.

33. The method of noninvasively detecting cardiac phenomena of claim 1 further comprising the steps of generating clinical reference arrays for use in identifying particular bispectral values that are sensitive to the cardiac phenomena being analyzed.

34. The method of noninvasively detecting cardiac phenomena of claim 33 further comprising the step of rank ordering said clinical reference arrays with respect to the diagnostic efficacy of the arrays for the cardiac phenomena being analyzed.

35. The method of noninvasively detecting cardiac phenomena of claim 33 wherein values in said clinical reference arrays that are not significant are suppressed to facilitate inspection of the most sensitive values.

36. The method of noninvasively detecting cardiac phenomena of claim 33 further comprising the steps of determining a diagnostic index from said clinical reference array, generated values above said index indicating a normal condition and below said index indicating an abnormal condition.

37. The method of noninvasively detecting cardiac phenomena of claim 36 wherein said diagnostic index is optimized to yield the best sensitivity and specificity with sensitivity being equal to true positives divided by positives and specificity being equal to true negatives divided by negatives.

38. The method of noninvasively detecting cardiac phenomena of claim 1 wherein said bispectral values are generated by computing the Fourier transform of the third order autocorrelation function of said filtering signals.

39. The method of noninvasively detecting cardiac phenomena of claim 1 wherein said bispectral values are generated by computing the Fourier transform of the third order crosscorrelation function of said filtered signals.

40. A system noninvasively detecting cardiac phenomena comprising:

means for acquiring electrocardiographics signals from a surface of a body of a subject being analyzed;

means for determining a waveform template representing a normal QRST waveform in said subject;

means for selecting a number of signals that match said waveform template;

means for storing said selected signals for processing;

means for generating bispectral values from said selected signals and for comparing said bispectral values to reference bispectral values to detect and analyze cardiac phenomena.

41. The system for noninvasively detecting cardiac phenomena of claim 40 wherein said bispectral values are crossbiphase values.

42. The system for noninvasively detecting cardiac phenomena of claim 40 wherein said bispectral values are autobiphase values.

43. The system for noninvasively detecting cardiac phenomena of claim 40 wherein said bispectral values are autobispectral density values.

44. The system for noninvasively detecting cardiac phenomena of claim 40 wherein said bispectral values are crossbispectral density values.

45. The system for noninvasively detecting cardiac phenomena of claim 40 wherein said bispectral values are autobicoherence values.

46. The system for noninvasively detecting cardiac phenomena of claim 40 wherein said bispectral values are crossbicoherence values.

47. The system for noninvasively detecting cardiac phenomena of claim 40 wherein said means for acquiring electrocardiographic signals is a plurality of surface electrodes and wherein said bispectral values are crossbispectral values generated from consecutive signals from at least one of said plurality of said surface electrodes.

48. The system for noninvasively detecting cardiac phenomena of claim 40 wherein said means for acquiring electrocardiographic signals is a plurality of surface electrodes and wherein said bispectral values are crossbispectral values generated from signals obtained from two of said surface electrodes.

49. The system for noninvasively detecting cardiac phenomena of claim 40 wherein said template is an ectopic beat template.

50. The system for noninvasively detecting cardiac phenomena of claim 41 further comprising means for generating clinical reference arrays for use in identifying particular bispectral values that are sensitive to the cardiac phenomena being detected and analyzed.

51. The system for noninvasively detecting cardiac phenomena of claim 41 further comprising means for automatically rank ordering clinical reference arrays of said reference bispectral values with respect to diagnostic efficacy of the arrays for the cardiac phenomena being detected and analyzed.

52. The system for noninvasively detecting cardiac phenomena of claim 41 further comprising means for suppressing values in clinical reference arrays of said reference bispectral values that are not significant in order to facilitate inspection of values most likely to indicate the presence of the cardiac phenomena.

53. The system for noninvasively detecting cardiac phenomena of claim 41 further comprising means for determining a diagnostic index from clinical reference arrays of said reference bispectral values, generated values above said index indicating a normal condition and values below said index indicating an abnormal condition.

54. The system for noninvasively detecting cardiac phenomena of claim 53 wherein said diagnostic index is optimized to yield the best sensitivity and specificity with sensitivity being equal to true positive tests divided by all true positives and specificity being equal to true negative tests divided by all true negatives.

55. The system for noninvasively detecting cardiac phenomena of claim 41 wherein said means for generating bispectral values comprises a means for computing the Fourier transform of the third order autocorrelation function of said filtered signals.

56. The system for noninvasively detecting cardiac phenomena of claim 41 wherein said means for generating bispectral values comprises a means for computing the Fourier transform of the third order crosscorrelation function of said filtered signals.