Method of analysis of medical signals

• US 7,171,269 B1
• Filed: 05/02/2000
• Issued: 01/30/2007
• Est. Priority Date: 05/01/1999
• Status: Expired due to Fees

• Current Assignee: CardioDigital Ltd.
• Original Assignee: CardioDigital Ltd.
• Inventors: Addison, Paul Stanley, Watson, James Nicholas
• Primary Examiner(s): Casler; Brian L.
• Assistant Examiner(s): JUNG, WILLIAM C
• Application Number:
  US09/980,770
• Time in Patent Office: 2,464 Days
• Field of Search: 600/508, 600/504, 600/513
• US Class Current: 607/7
• CPC Class Codes: A61B 5/349 Detecting specific parameters of the electrocardiograph cycle
• CPC Class Codes: A61B 5/361 Detecting fibrillation
• CPC Class Codes: A61B 5/726 using Wavelet transforms
• 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: Y10S 128/901 Suppression of noise in electric signal
• CPC Class Codes: Y10S 128/902 Biological signal amplifier

Abstract

A method of analysis of medical signals which uses wavelet transform analysis to decompose cardiac signals. Apparatus for carrying out the method, and cardiac apparatus adapted to employ the method are also described.

Claims

1. A method of decomposition of waveforms in a cardiac signal comprising the steps of:

a) connecting electrodes to a patient whose heart is in Ventricular Fibrillation (VF);

b) deriving analogue input signals from the electrodes;

c) sampling said analogue input signals to derive the cardiac signal (EKG);

d) digitising said EKG signal;

e) employing wavelet transform analysis to process said digitised EKG signal;

f) extracting key features from the wavelet transform representation to predict the outcome of a specific interim therapeutic intervention during the Ventricular Fibrillation; and

g) guiding a resuscitation protocol based on the prediction, said guidance comprising the steps of;

h) using an analytical method to determine the likely outcome of a defibrillation shock; and

i) determining whether to provide at least one interim therapeutic intervention from a group comprising defibrillatory shock, CPR and pharmaceutical, before shocking.

2. The method of claim 1, wherein the analytical method is characterised by learning vector quantisation (LVQ) methods, for example Kohonen Networks.

3. The method of claim 1, where the analytical method is characterised by statistical, stochastic methods, for example Baysian Methods.

4. The method of claim 1, where the analytical method is characterised by multi-layered neural network methods, for example Radial Basis Neural Networks.

5. A method of decomposition of waveforms in a cardiac signal comprising the steps of:

a) connecting electrodes to a presenting patient whose heart is in Ventricular Fibrillation (VF) after the commencement of Cardio-Pulmonary Resuscitation (CPR);

b) deriving analogue input signals from the electrodes;

c) sampling the analogue input signals to derive the cardiac signal (EKG);

d) digitising said EKG signal;

e) employing wavelet transform analysis to process said digitised EKG signal; and

f) extracting key features from the wavelet transform representation to predict the outcome of a specific interim therapeutic intervention during the Ventricular Fibrillation.

6. The method of claim 5, further comprising the steps of:

a) filtering said cardiac signal such that the CPR component is disassociated/separated from the heart signal;

b) producing an energy wavelet scalogram; and

c) temporally filtering the scalogram using ridge following techniques.

7. The method of claim 6, wherein said ridge following techniques are characterised by modulus maxima techniques.

8. The method of claim 6, further comprising the steps for guiding resuscitation protocol of:

a) using an analytical method for determining the likely outcome of a defibrillation shock; and

b) determining whether to provide at least one interim therapeutic intervention from a group comprising immediate defibrillatory shock and CPR, before shocking.

9. The method of claim 8, wherein said analytical method is characterised by learning vector quantisation (LVQ) methods, for example Kohonen Networks.

10. The method of claim 8, wherein said analytical method is characterised by statistical, stochastic methods, for example Baysian Methods.

11. The method of claim 8, wherein said analytical method is characterised by multi-layered neural network methods, for example Radial Basis Neural Networks.

12. A method of decomposition of waveforms in a cardiac signal comprising the steps of:

a) connecting electrodes to a presenting patient whose heart is in Atrial Fibrillation (AF);

b) deriving analogue signals from said electrodes;

c) sampling the analogue input signals to derive the cardiac signal (EKG);

d) digitising said EKG signal; and

e) employing wavelet transform analysis to process said digitised EKG signal; and

f) extracting key features from the wavelet transform representation to predict the outcome of a specific interim therapeutic intervention during the Atrial Fibrillation.

13. The method of claim 12, further comprising the step of filtering said cardiac signal such that the QRS complex and T components are disassociated/separated from the heart signal, comprising:

a) producing an energy wavelet scalogram; and

b) temporally filtering the scalogram using ridge following techniques.

14. The method of claim 13, wherein said ridge following techniques are characterised by modulus maxima techniques.

15. The method of claim 13, further comprising the step for guiding the course of therapeutic intervention taken, comprising:

a) using an analytical method for determining the likely outcome of a cardioversion shock; and

b) determining whether to at least one therapeutic intervention from a group comprising cardioversion shock, and drug therapy.

16. The method of claim 15, wherein said analytical method is characterised by learning vector quantisation (LVQ) methods, for example Kohonen Networks.

17. The method of claim 15, wherein said analytical method is characterised by statistical, stochastic methods, for example Baysian Methods.

18. The method of claim 15, wherein said analytical method is characterised by multi-layered neural network methods, for example Radial Basis Neural Networks.