Method and apparatus for ECG rhythm analysis

• US 4,589,420 A
• Filed: 07/13/1984
• Issued: 05/20/1986
• Est. Priority Date: 07/13/1984
• Status: Expired due to Fees

• Current Assignee: Spacelabs Incorporated A Corporation OFCA
• Original Assignee: Spacelabs, Inc. (OSI Systems Incorporated)
• Inventors: Carlson, James M., Yeshaya, Amalia, Adams, Isabelle M., Rooks, James B.
• Primary Examiner(s): Kamm, William E.
• Assistant Examiner(s): NOT, DEFINED
• Application Number:
  US06/630,782
• Time in Patent Office: 676 Days
• Field of Search: 128/696, 128/702-706, 128/708, 128/710, 364/415, 364/417
• US Class Current: 600/515
• CPC Class Codes: A61B 5/35 by template matching
• 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: A61B 5/7267 involving training the classification device
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems

Abstract

A method and apparatus for automatic rhythm analysis of digitized ECG signals. Candidate heartbeats are detected and the noise of the ECG signal portion containing the candidate is evaluated to determine whether the candidate is a valid nominal or noisy beat or an invalid noisy saturated signal. The timing of detected heartbeats, both noisy and nominal, are determined relative to prior validated heartbeats. Features of the detected heartbeats are automatically extracted including onset, offset and R wave locations; QRS complex width; R-R interval based on the onset points; instantaneous and averaged heartrates; and others. A 16 point template of the heartbeat starting at the onset is also determined. Beats are then classified in a two step process; first, by template comparison with a previously determined dominant template, and second, by a finite state machine process based on the results of the template comparison and relative timing of the beat. As a result of the two step classification beats are classified as belonging to a plurality of beat types including both single and sequence beat types. A class record for each beat type is created and stored, any portion of which is accessible for display upon demand.

Claims

1. An apparatus for automatic heartbeat rhythm analysis of ECG signals comprising:

means for digitizing said ECG signals;

means responsive to said digitized ECG signals for automatically detecting heartbeats including means for analyzing said detected heartbeats;

means for assessing the relative timing of said detected heartbeats relative to prior heartbeats;

means responsive to said analysis for classifying each of said detected heartbeats as one of a plurality of beat types, said plurality of beat types including a plurality of sequences of beat types, said classifying means further comprising;

first stage means responsive to said timing means for automatically classifying each of said detected heartbeats as one of a first plurality of single heartbeat types; and

state machine means responsive to said first stage classification for automatically determining when said detected heartbeats are part of one of said plurality of sequences of heartbeat types and for classifying said detected heartbeats so determined as belonging to one of said sequence type plurality, said apparatus further comprising;

means responsive to said classifying means for automatically forming and storing class records, at least one for each of said plurality of beat types including said plurality of sequences of beat types; and

means for displaying all or a portion of preselected ones of said class records.

2. The apparatus of claim 1 wherein said analysis by said analyzing means comprises forming a template including a plurality of digital signals representative of a portion of an associated detected heartbeat, and

said classifying means further comprises forming a dominant heartbeat type by template comparison of detected heartbeats;

and

said first stage means further comprises means for forming a plurality of abnormal single heartbeat types by template comparison of said detected heartbeats with said dominant heartbeat template.

3. The apparatus of claim 2 wherein said means for comparing said templates comprises:

means for determining the absolute values of the corresponding point by point differences between two of said templates;

means for comparing the absolute values of said differences with a first predetermined number for said differences obtained over a first interval of said templates and with a second predetermined number for a second interval of said templates; and

means for adding the number of times that said absolute values exceed said first and second predetermined numbers; and

means for determining whether said incoming and dominant templates match in response to said number of times.

4. The apparatus of claim 1 wherein said relative timing means further comprises:

means for automatically determining the R-R interval of said detected heartbeats;

means for determining an instantaneous and average heartbeat in response to said R-R interval; and

means for comparing said instantaneous heartrate with said average heartrate in a predetermined manner to determine said relative timing.

5. The apparatus of claim 1 wherein said second plurality of sequences of heartbeat types comprises:

couplets, triplets, and runs.

6. The apparatus of claim 1 wherein said detection means further comprises:

means for automatically detecting candidate heartbeats;

means for evaluating the noise content of said ECG signals;

said evaluating means further comprising;

means for identifying said candidate beats as valid, noisy or invalid.

7. The apparatus of claim 6 wherein said noise evaluating means further comprises:

means for automatically detecting local maximum and minimum peaks within a predetermined interval including said detected heartbeat;

means for determining the number of said local peaks exceeding a first predetermined maximum or minimum signal value throughout said predetermined interval; and

means for determining the number of said local peaks exceeding a second predetermined maximum or minimum signal value throughout a preselected portion of said predetermined interval.

8. The apparatus of claim 7 wherein said preselected portion comprises the ST portion of said ECG signal associated with said detected heartbeat.

9. The apparatus of claim 1 wherein said means for automatically analyzing detected heartbeats further comprises:

peak detection means for automatically locating the peak of the QRS complex of said ECG signals;

means for automatically locating the onset of said QRS complex relative to said peak locations; and

means for automatically determining the R-R interval between the current and prior detected heartbeats in response to said onset points.

10. The apparatus of claim 9 wherein said means for automatically analyzing heartbeats further comprises;

means for determining instantaneous and averaged heart rates in response to said R-R interval determination; and

timing means for determining the relative timing of said heartbeat in response to said instantaneous and averaged heart rates.

11. The apparatus of claim 9 wherein said analysis means further comprises:

means for forming a template representative of said detected heartbeat, said template beginning at said onset point.

12. The apparatus of claim 11 wherein each of said class records comprises:

the number of detected heartbeats classified in the class;

a trend record;

the template of at least one of said detected heartbeats classified in said class; and

a representative ECG waveform sample over a predetermined time interval including at least one of said detected heartbeat types of said class record.

13. A method of automatic rhythm analysis of digitized ECG signals comprising:

detecting said heartbeats from said digitized ECG signals;

analyzing said detected heartbeats including assessing the timing of said detected heartbeats relative to prior detected heartbeats;

automatically classifying each of said detected heartbeats in response to said analysis as one of a plurality of heartbeat types including a plurality of sequences of heartbeat types, said classifying step including;

identifying a dominant heartbeat type;

comparing said detected heartbeats with said dominant heartbeat and classifying said detected heartbeats as dominant or abnormal;

determining whether a detected heartbeat is part of one of said plurality of sequences of heartbeat types from said relative timing and said classification of said heartbeat as dominant or abnormal; and

classifying said detected heartbeat as one of said sequence plurality when said detected heartbeat is determined to be part of one of said sequence plurality, said method further comprising;

automatically forming and storing class records for each of said plurality of heartbeat types including said plurality of sequences of heartbeat types; and

displaying all or a portion of selected ones of said class records.

14. The method of claim 13 wherein said method further comprises the step of forming a template for each detected heartbeat;

and

wherein the step of comparing said detected heartbeats comprises a point by point template matching comparison.

15. The method of claim 13 wherein the step of assessing the relative timing of said heartbeats comprises:

determining the R-R interval between successive detected heartbeats;

determining instantaneous and averaged heartrates in response to said R-R intervals; and

comparing said instantaneous heartrate with said averaged heartrate in a predetermined manner whereby said detected heartbeats are determined to be early, regular or late.

16. The method of claim 13 wherein the step of detecting said heartbeats comprises:

detecting candidate heartbeats;

evaluating the noise of a portion of said ECG signal including said candidate heartbeat;

determining whether said candidate beat is a valid, noisy or invalid heartbeat in response to said noise evaluation.