Monitoring physiological activity using partial state space reconstruction

• US 7,996,075 B2
• Filed: 03/15/2005
• Issued: 08/09/2011
• Est. Priority Date: 10/20/2004
• Status: Active Grant

• Current Assignee: Braemar Incorporated
• Original Assignee: Cardionet Incorporated
• Inventors: Korzinov, Lev, Kremliovsky, Michael
• Primary Examiner(s): Getzow; Scott M
• Application Number:
  US11/081,401
• Publication Number:
  US 20060084881A1
• Time in Patent Office: 2,338 Days
• Field of Search: 600508-523, 607 5- 30
• US Class Current: 600/515
• CPC Class Codes: A61B 5/366 Detecting abnormal QRS complex, e.g. widening
• CPC Class Codes: G06F 2218/18 by plotting the signal segments against each other, e.g. analysing scattergrams

Abstract

Systems and techniques relating to monitoring physiological activity using partial state space reconstruction. In general, in one aspect, a partial state space is produced using an orthogonal, frequency-independent transform, such as Hilbert transform. The partial state space can then be analyzed using state space techniques to identify physiological information for the biological system. The described techniques can be implemented in a distributed cardiac activity monitoring system, including a cardiac monitoring apparatus, and a QRS detector thereof.

Claims

1. A machine-implemented method comprising:

obtaining a physiological signal of a biological system of an organism;

generating a transformed signal that is mathematically orthogonal to the physiological signal by applying a frequency-independent transform to the physiological signal;

producing, from the physiological signal and the transformed signal, a partial state space representing dynamics of the biological system, the partial state space being a partial reconstruction of a state space representing system dynamics for the biological system; and

identifying physiological information concerning the organism based on an analysis of signal morphology in the partial state space;

wherein the producing and the identifying are performed by one or more data processing apparatus.

2. The method of claim 1, wherein obtaining the physiological signal comprises receiving a sensed cardiac signal.

3. The method of claim 2, wherein identifying the physiological information comprises classifying heart beats in the sensed cardiac signal.

4. The method of claim 2, wherein identifying the physiological information comprises characterizing a physiological condition of the organism.

5. The method of claim 1, wherein identifying the physiological information comprises detecting a physiological occurrence for the biological system based on a dynamical quantity comprising a value derived from the physiological signal and the transformed signal.

6. The method of claim 5, wherein obtaining the physiological signal comprises receiving an electrically-sensed time series x(t), generating the transformed signal comprises applying Hilbert (H) transform to the time series x(t) to obtain H(x(t)), and producing the partial state space comprises considering x(t) and H(x(t)) together as components of a state vector.

7. The method of claim 6, further comprising generating multiple dynamical quantities from the physiological signal and the transformed signal, and wherein identifying the physiological information comprises assessing the multiple dynamical quantities with respect to one or more predefined physiological aspects of the biological system.

8. The method of claim 7, wherein the multiple dynamical quantities comprise nonlinear transformations of x(t) and H(x(t)) in state space, excepting simple linear combinations of amplitude and phase.

9. The method of claim 5, wherein obtaining the physiological signal comprises receiving a real-time, electrocardiogram time series for an actively monitored human heart, generating the analytic signal comprises applying Hilbert transform directly to the received electrocardiogram time series, and detecting the physiological occurrence comprises assessing the dynamical quantity with respect to one or more predefined physiological aspects of the human heart.

10. The method of claim 1, wherein the frequency-independent transform comprises a nonlinear transform.

11. The method of claim 10, wherein the frequency-independent transform comprises a linear transform and a nonlinear transform.

12. A software product tangibly embodied in a machine-readable medium, the software product comprising instructions operable to cause data processing apparatus to perform operations comprising:

obtaining a physiological signal of a biological system of an organism;

generating a transformed signal that is mathematically orthogonal to the physiological signal by applying a frequency-independent transform to the physiological signal;

producing, from the physiological signal and the transformed signal, a partial state space representing dynamics of the biological system, the partial state space being a partial reconstruction of a potential full state space for the biological system; and

identifying physiological information concerning the organism based on an analysis of signal morphology in the partial state space.

13. The software product of claim 12, wherein obtaining the physiological signal comprises receiving a sensed cardiac signal.

14. The software product of claim 13, wherein identifying the physiological information comprises classifying heart beats in the sensed cardiac signal.

15. The software product of claim 13, wherein identifying the physiological information comprises characterizing a physiological condition of the organism.

16. The software product of claim 12, wherein identifying the physiological information comprises detecting a physiological occurrence for the biological system based on a dynamical quantity comprising a value derived from the physiological signal and the transformed signal.

17. The software product of claim 16, wherein obtaining the physiological signal comprises receiving an electrically-sensed time series x(t), generating the transformed signal comprises applying Hilbert (H) transform to the time series x(t) to obtain H(x(t)), and producing the partial state space comprises considering x(t) and H(x(t)) together as components of a state vector.

18. The software product of claim 17, the operations further comprising generating multiple dynamical quantities from the physiological signal and the transformed signal, and wherein identifying the physiological information comprises assessing the multiple dynamical quantities with respect to one or more predefined physiological aspects of the biological system.

19. The software product of claim 18, wherein the multiple dynamical quantities comprise nonlinear transformations of x(t) and H(x(t)) in state space, excepting simple linear combinations of amplitude and phase.

20. The software product of claim 16, wherein obtaining the physiological signal comprises receiving a real-time, electrocardiogram time series for an actively monitored human heart, generating the analytic signal comprises applying Hilbert transform directly to the received electrocardiogram time series, and detecting the physiological occurrence comprises assessing the dynamical quantity with respect to one or more predefined physiological aspects of the human heart.

21. The software product of claim 12, wherein the frequency-independent transform comprises a nonlinear transform.

22. The software product of claim 21, wherein the frequency-independent transform comprises a linear transform and a nonlinear transform.

23. A cardiac monitoring apparatus comprising:

an ECG input element;

a state space transformation component responsive to an output of the ECG input element and configured to generate a transformed signal that is mathematically orthogonal to a physiological signal, for a biological system of an organism, by applying a frequency-independent transform to the physiological signal; and

a QRS identification component responsive to an output of the state space transformation component, the QRS identification component comprising one or more dynamical quantity calculators that each produce a value derived from a combination of the physiological signal and the transformed signal in a partial state space representing dynamics of the biological system, the partial state space being a partial reconstruction of a state space representing system dynamics for the biological system.

24. A cardiac monitoring apparatus comprising:

an ECG input element;

a state space transformation component responsive to an output of the ECG input element; and

a QRS identification component responsive to an output of the state space transformation component, the QRS identification component comprising one or more dynamical quantity calculators;

the apparatus further comprising a pQRST parameter averaging component coupled with the QRS identification component, a noise estimator responsive to the output of the ECG input element, and final QRS decision logic coupled with the QRS identification component, the pQRST parameter averaging component and the noise estimator.

25. The apparatus of claim 24, further comprising a filter coupled between the ECG input element and both the state space transformation component and the noise estimator.

26. The apparatus of claim 25, wherein the output of the ECG input element comprises a split output, the apparatus further comprising an arrhythmia identification component coupled with the split output and with an input of the final QRS decision logic.

27. The apparatus of claim 26, wherein the arrhythmia identification component comprises a ventricular fibrillation detector.

28. A system comprising:

an ECG input element;

a state space transformation component responsive to an output of the ECG input element and configured to generate a transformed signal that is mathematically orthogonal to a physiological signal, for a biological system of an organism, by applying a frequency-independent transform to the physiological signal;

a QRS identification component responsive to an output of the state space transformation component, the QRS identification component comprising one or more dynamical quantity calculators that each produce a value derived from a combination of the physiological signal and the transformed signal in a partial state space representing dynamics of the biological system, the partial state space being a partial reconstruction of a state space representing system dynamics for the biological system; and

an antenna coupled with the QRS identification component and configured to wirelessly transmit physiological information.

29. The system of claim 28, further comprising a monitoring station configured to receive the transmitted physiological information.

30. A system comprising:

an ECG input element;

a state space transformation component responsive to an output of the ECG input element;

a QRS identification component responsive to an output of the state space transformation component, the QRS identification component comprising one or more dynamical quantity calculators;

an antenna coupled with the QRS identification component and configured to wirelessly transmit physiological information; and

a monitoring station configured to receive the transmitted physiological information,

the system further comprising a pQRST parameter averaging component coupled with the QRS identification component, a noise estimator responsive to the output of the ECG input element, and final QRS decision logic coupled with the QRS identification component, the pQRST parameter averaging component and the noise estimator.

31. The system of claim 30, further comprising a filter coupled between the ECG input element and both the state-space transformation component and the noise estimator.

32. The system of claim 31, wherein the output of the ECG input element comprises a split output, the apparatus further comprising an arrhythmia identification component coupled with the split output and with an input of the final QRS decision logic.

33. The system of claim 32, wherein the arrhythmia identification component comprises a ventricular fibrillation detector and an asystole detector.

34. A machine-implemented method comprising:

obtaining a cardiac signal of a heart;

translating the cardiac signal into an embedding space that represents coarse-grained dynamics of the heart, said translating comprising applying Hilbert transform to the cardiac signal; and

employing state space analysis techniques to extract physiological information for the heart from the embedding space;

wherein the translating and the employing are performed by one or more data processing apparatus.

35. The method of claim 34, wherein obtaining the cardiac signal comprises obtaining multiple cardiac signals from independent leads, and translating the cardiac signal comprises applying the Hilbert transform directly to the multiple cardiac signals to form the embedding space having an embedding dimension greater than or equal to twice a number of the independent leads.

36. The method of claim 35, wherein the number of the independent leads is two, and the embedding space has four spatial dimensions.

37. The method of claim 35, wherein obtaining the multiple cardiac signals comprises retrieving the multiple cardiac signals from a database.

38. The method of claim 34, wherein translating the cardiac signal comprises calculating a nonlinear combination of the cardiac signal and the Hilbert transform of the cardiac signal.

39. The method of claim 34, wherein employing state space analysis techniques comprises deriving multiple dynamical quantities from the embedding space.

40. The method of claim 39, wherein the multiple dynamical quantities comprise speed of trajectory in state space, length of trajectory in state space, area integral of a speed vector, and threshold crossings in state space.

41. The method of claim 34, further comprising detecting abnormal heart beats based on the extracted physiological information.

42. The method of claim 41, further comprising estimating a physiological condition based on the detected abnormal heart beats.

43. A software product tangibly embodied in a machine-readable medium, the software product comprising instructions operable to cause one or more data processing apparatus to perform operations comprising:

obtaining a cardiac signal of a heart;

translating the cardiac signal into an embedding space that represents coarse-grained dynamics of the heart, said translating comprising applying Hilbert transform to the cardiac signal; and

employing state space analysis techniques to extract physiological information for the heart from the embedding space.

44. The software product of claim 43, wherein obtaining the cardiac signal comprises obtaining multiple cardiac signals from independent leads, and translating the cardiac signal comprises applying the Hilbert transform directly to the multiple cardiac signals to form the embedding space having an embedding dimension greater than or equal to twice a number of the independent leads.

45. The software product of claim 44, wherein the number of the independent leads is two, and the embedding space has four spatial dimensions.

46. The software product of claim 44, wherein obtaining the multiple cardiac signals comprises retrieving the multiple cardiac signals from a database.

47. The software product of claim 43, wherein translating the cardiac signal comprises calculating a nonlinear combination of the cardiac signal and the Hilbert transform of the cardiac signal.

48. The software product of claim 43, wherein employing state space analysis techniques comprises deriving multiple dynamical quantities from the embedding space.

49. The software product of claim 48, wherein the multiple dynamical quantities comprise speed of trajectory in state space, length of trajectory in state space, area integral of a speed vector, and threshold crossings in state space.

50. The software product of claim 43, the operations further comprising detecting abnormal heart beats based on the extracted physiological information.

51. The software product of claim 50, the operations further comprising estimating a physiological condition based on the detected abnormal heart beats.