Method and apparatus for analyzing and interpreting electrocardiograms using spectro-temporal mapping

• US 5,046,504 A
• Filed: 04/06/1990
• Issued: 09/10/1991
• Est. Priority Date: 02/01/1989
• Status: Expired due to Term

• Current Assignee: Arrhythmia Research Technology, Inc.
• Original Assignee: Corazonix Corporation (Micron Solutions, Inc.)
• Inventors: Albert, David E., Lander, Paul
• Primary Examiner(s): Jaworski, Francis
• Assistant Examiner(s): MANUEL, GEORGE C
• Application Number:
  US07/505,782
• Time in Patent Office: 522 Days
• Field of Search: 128/695, 128/696, 128/699, 128/700, 128/702, 128/703, 128/704, 128/705, 128/710, 364/413.05, 364/413.06
• US Class Current: 600/509
• CPC Class Codes: A61B 5/35 by template matching
• CPC Class Codes: A61B 5/7253 characterised by using transforms
• CPC Class Codes: A61B 5/7257 using Fourier transforms

Abstract

Method and apparatus for analyzing ECG signal data for diagnostic purposes by performing an incrementally moving, short window FFT analysis to produce a set of spectral templates representative of the spectral frequency content of the ECG at each window position. The spectral templates may be displayed as a representation of a three dimensional surface or mathematically analyzed. The changes in the shape of the templates or surface with time represent a change in the spectrum, which are shown to be an indication of abnormalities in the heart.

Claims

1. The method of evaluating electrocardiogram (ECG) signals of a patient which comprises:

digitizing a large number of successive, repetitive ECG waveforms, each waveform representing an amplitude with respect to time from the patient, at sampling points equal to a sampling rate greater than about 500 per second;

averaging the digital sample values at corresponding sample points on the successive ECG waveforms to produce a single digitized signal averaged, high resolution ECG (HRECG) waveform having significantly enhanced signal-to-noise ratio;

producing a set of digital spectrum values representative of an approximate power density spectrum at each of a large number of generally equally spaced sampling time intervals over the HRECG waveform, each spectrum value being a function of the power of the HRECG waveform in a finite time/amplitude window having a time deviation substantially less than the time period of the HRECG waveform which is of interest; and

determining the presence of late potentials from the set of power density spectrums.

3. The method of claim 1 wherein the onset of the QRS of the ECG is determined by locating the first power density spectrum approximation having a total power exceeding a predetermined value, the termination of the QRS is determined by locating the last power density spectrum approximations having a total power exceeding a predetermined value, and determining the time interval between the two PDS approximations as representative of the length of the QRS segment.

4. The method of claim 1 wherein late potentials in the RST time segment of the type indicative of ventricular tachycardia are detected by variations in the shape of the power density spectrum.

2. The method of claim 2 wherein the set of power density spectrum approximations are displayed as a three dimensional visual display in which the spectral values of the power density are represented as an apparent surface, one ordinate of the surface is time, the second ordinate is the frequency of the energy, and the third ordinate is power.

5. The method for evaluating an electrocardiogram (ECG) which comprises:

producing a set of digital values representative of a three dimensional surface defined by a three generally orthogonal axis coordinate system in which a first axis represents the time domain of the ECG, a second axis represents the frequency domain of the ECG and the third axis represents an approximation of the magnitude of the spectral frequency of the ECG; and

visually displaying the surface such that the three dimensional character of the surface can be observed and from which an evaluation of the ECG data can be made.

6. The method for producing and evaluating a high resolution electrocardiogram (HRECG) which comprises:

averaging corresponding digital sample points of successive waveforms of an ECG from a patient to produce digital data representative of a single HRECG waveform having an enhanced signal-to-noise ratio when compared to the original ECG data;

producing a set of digital values representative of a three dimensional surface defined by a three orthogonal axis coordinate system in which a first axis represents the time domain of the HRECG, a second axis represents the frequency domain of the HRECG and the third axis represents an approximation of the magnitude of the power of the HRECG; and

visually displaying the surface such that the three dimensional character of the surface can be observed and from which an evaluation of the ECG data can be made.

7. The method for evaluating an electrocardiogram (ECG) of a patient which comprises:

calculating from the ECG and storing digital data defining a three dimensional surface representative of the spectral frequency of the ECG in which the first axis is the time domain of the ECG, the second axis is the frequency domain, and the third axis is representative of the amplitude or power domain; and

evaluating the data representative of the surface to diagnose the condition of the patient'"'"'s heart.

8. The method of claim 7 further comprising visually inspecting a three dimensional visual representation of the surface data to diagnose the condition of the patient'"'"'s heart.

9. The method of claim 7 or 8 further comprising producing data representative of a set of spectral templates and evaluating the set of spectral templates to diagnose the condition of the patient'"'"'s heart.

10. The method of claim 9 further comprising producing a measure of the degree of correlation of the relative shapes of two templates spacially related in a selected manner to assist in the diagnosis.

11. The method of claim 10 wherein the degree of correlation is determined for adjacent templates of a selected set and presented as a visual image related to the time axis of the surface.

12. The method of claim 11 wherein the degree of correlation is a correlation factor comprising a discrete value plotted with respect to time.

13. The method of claim 11 wherein the degree of correlation is a residual area template for each template pair and the residual area templates are displayed as a three dimensional image.

14. The method of claim 7 or 8 further comprising producing a value with respect to the time domain representative of a selected attribute of each of a selected set of spectral templates and evaluating the data representative of the surface, based on the values produced, to diagnose the condition of the patient'"'"'s heart.

15. The method of claim 14 wherein the selected attribute is a spectral edge.

16. The method of claim 14 wherein the selected attribute is a mean or median frequency.

17. The method of claim 7 or 8 further comprising means for computing an area value representative of the total area of each of a set of spectral templates and evaluating the data representative of the surface, based on the area values computed, to diagnose the condition of the patient'"'"'s heart.

18. The method of claim 17 further comprising means for computing a volume value representative of the total volume defined by the surface and selected other limits of time and frequency and evaluating the data representative of the surface, based on the volume values computed, to diagnose the condition of the patient'"'"'s heart.

19. The method of claim 17 or 18 further comprising computing a series of the values and selecting the first occurring value about a predetermined value as a reference time point of the ECG.