Method and apparatus for spectral analysis of electrocardiographic signals
First Claim
1. A method for detecting abnormalities in physiological functions of an organism by performing a frequency analysis of a single cycle of a time varying bioelectric signal, said method comprising:
- (a) partitioning said single cycle of said time varying signal into a sequence of time-displaced, overlapping segments or slices,(b) performing a frequency analysis of each of said slices, thereby resolving each of said slices into a sequence of discrete spectral components of varying amplitude,(c) storing a number proportional to the magnitude of each of said spectral components in a table in which frequency is plotted along a first axis, and time is plotted along a second axis,(d) assessing the degree of difference between the spectra of different slices of said table, and (e) comparing said spectral differences with corresponding differences obtained from representative samples of normal and abnormal test organisms.
1 Assignment
0 Petitions
Accused Products
Abstract
A frequency domain signal processing and analysis method and apparatus displays, plots and makes measurements upon electrocardiographic signals (ECG) recorded from the body surface. Both the graphic plots and the numeric parameters measured reveal abnormalities of electrical conduction within the heart thought to be important in the identification of patients at risk of serious disturbances of heart rhythm including sudden death. The invention employs Fourier analysis of short overlapping segments of ECG signal to create a three dimensional map ("spectrocardiogram") whose axes are time, fequency and power thus disclosing changes in the frequency content of the ECG signal over short intervals of time. In additoin to the Fourier analysis itself, the method and apparatus provides a wide choice of user-selectable signal pre-processing, post-processing, display and plotting options as well as performing custom mathematical computations upon the Fourier spectra to distinguish and quantify visually apparent differences between normal and abnormal spectrocardiograms.
208 Citations
59 Claims
-
1. A method for detecting abnormalities in physiological functions of an organism by performing a frequency analysis of a single cycle of a time varying bioelectric signal, said method comprising:
-
(a) partitioning said single cycle of said time varying signal into a sequence of time-displaced, overlapping segments or slices, (b) performing a frequency analysis of each of said slices, thereby resolving each of said slices into a sequence of discrete spectral components of varying amplitude, (c) storing a number proportional to the magnitude of each of said spectral components in a table in which frequency is plotted along a first axis, and time is plotted along a second axis, (d) assessing the degree of difference between the spectra of different slices of said table, and (e) comparing said spectral differences with corresponding differences obtained from representative samples of normal and abnormal test organisms. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9)
-
-
10. A method for analyzing in the frequency domain a single cycle of an ECG signal, which single cycle may be an ensemble average of a plurality of cycles, said method comprising:
-
a. locating within said ECG signal, a time interval of interest, typically the QRS region of said ECG signal; b. dividing said time interval of interest into a sequence of overlapping segments, or slices; c. multiplying each said slice by a windowing function to form a windowed slice; d. performing a discrete fast Fourier transform (FFT) on each said windowed slice, in which FFT the power spectral density (PSD) at a sequence of frequencies is calculated; e. producing a numerical matrix or table of PSD'"'"'s in which discrete frequencies head each column in said table and discrete times head each row of said table, thereby resulting in a matrix of PSDs for a vertically-stacked array of time-slices, and, f. comparing said matrix of PSD s with corresponding matrices obtained from representative samples of normal and abnormal test subjects. - View Dependent Claims (11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33)
-
-
34. An apparatus for analyzing ECG signals to detect abnormalities of electrical conduction within the heart comprising:
-
(a) means for selectably storing a waveform representative of the amplitude-versus-time variation of a single cycle of said ECG signal, or an ensemble average of said cycles; (b) means for locating within said waveform a time interval of interest; (c) means for dividing said time interval into a sequence of overlapping segments, or time slices; (d) means for performing a frequency analysis of each of said overlapping segments; (e) means for producing a numerical matrix table of the amplitude of each frequency component obtained by said frequency analysis means, said matrix table having in a first line of headings discrete frequency values, and in a second orthogonal line of headings, discrete time intervals corresponding to the beginning of each slice, and (f) means for comparing said table with corresponding tables obtained from normal and abnormal test subjects. - View Dependent Claims (35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59)
-
Specification