System and method for quantifying alternation in an electrocardiogram signal

• US 6,169,919 B1
• Filed: 06/18/1999
• Issued: 01/02/2001
• Est. Priority Date: 05/06/1999
• Status: Expired due to Term

• Current Assignee: Beth Israel Deaconess Medical Center, Inc. (Beth Israel Lahey Health, Inc.)
• Original Assignee: Beth Israel Deaconess Medical Center, Inc. (Beth Israel Lahey Health, Inc.)
• Inventors: Verrier, Richard L., Nearing, Bruce D.
• Primary Examiner(s): Getzow, Scott M.
• Application Number:
  US09/335,680
• Time in Patent Office: 564 Days
• Field of Search: 600/509, 600/510, 600/511, 600/512, 600/513, 600/515, 600/516, 600/517, 600/518
• US Class Current: 600/518
• CPC Class Codes: A61B 5/349 Detecting specific parameters of the electrocardiograph cycle
• CPC Class Codes: A61B 5/7203 for noise prevention, reduction or removal
• 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

Abstract

A system and method for quantifying alternation in the T-wave and ST segment of an ECG signal receives a digitized ECG signal (i.e., ECG data) for processing. The ECG data are used to calculate an odd median complex for the odd beats in the ECG data and an even median complex for the even beats in the ECG data. The odd median complex and the even median complex are then compared to obtain an estimate of the amplitude of beat-to-beat alternation in the ECG signal. Prior to calculation of the even and odd median complexes, the ECG data are filtered. Filtering of the ECG data involves low pass filtering the ECG data to remove high frequency noise, applying a baseline wander removal filter to the ECG data to remove low frequency artifacts, removing arrhythmic beats from the ECG data, and eliminating noisy beats from the ECG data. The filtered data are more suitable for calculation of an accurate estimate of alternation.

Claims

1. A method of quantifying alternation in an ECG signal having a plurality of beats, the method comprising the steps of:

(a) receiving digitized ECG data representing the ECG signal;

(b) calculating an odd median complex for at least one odd beat in said ECG data;

(c) calculating an even median complex for at least one even beat in said ECG data; and

(d) comparing said odd median complex with said even median complex to obtain an estimate of the amplitude of beat-to-beat alternation in said ECG signal.

2. The method of claim 1, further comprising the step before step (b) of:

filtering said ECG data.

3. The method of claim 2, wherein said filtering step comprises:

low pass filtering said ECG data; and

applying a baseline wander removal filter to said ECG data.

4. The method of claim 3, wherein said filtering step further comprises:

removing arrhythmic beats from said ECG data.

5. The method of claim 4, wherein said filtering step further comprises:

eliminating noisy beats from said ECG data.

6. The method of claim 5, wherein said eliminating step comprises:

calculating a mean value of all samples within a selected portion of a selected beat of said ECG data;

calculating a difference between said mean value and each sample within said selected portion of said selected beat;

calculating an average of the absolute value of said differences;

comparing said average to a threshold;

identifying said selected beat as a noisy beat based on said comparison of said average to said threshold; and

eliminating said noisy beat from said calculation of said odd and even median complexes.

7. The method of claim 3, wherein said applying step comprises:

(i) determining an isoelectric value at each of a first isoelectric point in a first beat, a second isoelectric point in a second beat, and a third isoelectric point in a third beat of said ECG data;

(ii) fitting a spline curve to said first three isoelectric values;

(iii) subtracting the values of said spline curve from the corresponding values of said ECG data between said first isoelectric point and said second isoelectric point;

(iv) subtracting the values of said spline curve from the corresponding values of said ECG data between said second isoelectric point and said third isoelectric point;

(v) determining an isoelectric value for a next isoelectric point in a next beat of said ECG data;

(vi) fitting a next spline curve to said next isoelectric value and isoelectric values corresponding to a previous two consecutive isoelectric points;

(vii) subtracting the values of said next spline curve from the corresponding values of said ECG data between said next isoelectric point and said previous isoelectric point; and

(viii) repeating steps (v) through (vii) until a desired plurality of beats in said ECG data have been processed to remove low frequency artifacts from said ECG data.

8. The method of claim 1, wherein step (d) comprises:

comparing a T-wave portion of said odd median complex with a T-wave portion of said even median complex to obtain an estimate of the amplitude of beat-to-beat alternation in said ECG signal.

9. The method of claim 1, wherein steps (a) through (d) are performed for a first plurality of beats of said ECG signal, and wherein said method further comprises:

repeating steps (a) through (d) for a second plurality of beats of said ECG signal to obtain a second estimate of the amplitude of beat-to-beat alternation in said ECG signal.

10. The method of claim 1, wherein step (b) comprises:

(i) initializing a first array with a plurality of odd median complex values, said first array representing said odd median complex;

(ii) initializing a second array with a plurality of even median complex values, said second array representing said even median complex;

(iii) comparing samples of an odd beat of said ECG data to corresponding values in said first array and adjusting said values of said first array as follows

A. if a sample of said odd beat exceeds said corresponding value of said first array by a first amount but by less than a second amount, then incrementing said corresponding value by said first amount,

B. if a sample of said odd beat exceeds said corresponding value of said first array by said second amount or by greater than said second amount, then incrementing said corresponding value by said second amount,

C. if a sample of said odd beat is less than said corresponding value of said first array by said first amount but not said second amount, then decrementing said corresponding value by said first amount, and

D. if a sample of said odd beat is less than said corresponding value of said first array by said second amount or by greater than said second amount, then decrementing said corresponding value by said second amount;

(iv) comparing samples of an even beat of said ECG data to corresponding values in said second array and adjusting said values of said second array as follows

A. if a sample of said even beat exceeds said corresponding value of said second array by a first amount but by less than a second amount, then incrementing said corresponding value by said first amount,

B. if a sample of said even beat exceeds said corresponding value of said second array by said second amount or by greater than said second amount, then incrementing said corresponding value by said second amount,

C. if a sample of said even beat is less than said corresponding value of said second array by said first amount but not said second amount, then decrementing said corresponding value by said first amount, and

D. if a sample of said even beat is less than said corresponding value of said second array by said second amount or by greater than said second amount, then decrementing said corresponding value by said second amount; and

(v) repeating steps (iii) and (iv) for a plurality of said odd and even beats of said ECG data.

11. The method of claim 10, wherein step (d) comprises:

calculating a difference between sample points of said odd median complex and sample points of said even median complex to obtain said estimate of the amplitude of beat-to-beat alternation in said ECG signal.

12. A system for quantifying alternation in an ECG signal having a plurality of beats, the method comprising the steps of:

means for receiving digitized ECG data representing the ECG signal;

means for calculating an odd median complex for at least one odd beat in said ECG data and for calculating an even median complex for at least one even beat in said ECG data; and

means for comparing said odd median complex with said even median complex to obtain an estimate of the amplitude of beat-to-beat alternation in said ECG signal.

13. The system of claim 12, further comprising:

means for filtering said ECG data.

14. The system of claim 12, wherein said comparing means comprises:

means for comparing a T-wave portion of said odd median complex with a T-wave portion of said even median complex to obtain an estimate of the amplitude of beat-to-beat alternation in said ECG signal.

15. The system of claim 12, wherein said calculating means comprises:

means for initializing a first array with a plurality of odd median complex values, said first array representing said odd median complex;

means for initializing a second array with a plurality of even median complex values, said second array representing said even median complex;

means for comparing samples of an odd beat of said ECG data to corresponding values in said first array and adjusting said values of said first array as follows

if a sample of said odd beat exceeds said corresponding value of said first array by a first amount but by less than a second amount, then incrementing said corresponding value by said first amount,

if a sample of said odd beat exceeds said corresponding value of said first array by said second amount or by greater than said second amount, then incrementing said corresponding value by said second amount,

if a sample of said odd beat is less than said corresponding value of said first array by said first amount but not said second amount, then decrementing said corresponding value by said first amount, and

if a sample of said odd beat is less than said corresponding value of said first array by said second amount or by greater than said second amount, then decrementing said corresponding value by said second amount; and

means for comparing samples of an even beat of said ECG data to corresponding values in said second array and adjusting said values of said second array as follows

if a sample of said even beat exceeds said corresponding value of said second array by a first amount but by less than a second amount, then incrementing said corresponding value by said first amount,

if a sample of said even beat exceeds said corresponding value of said second array by said second amount or by greater than said second amount, then incrementing said corresponding value by said second amount,

if a sample of said even beat is less than said corresponding value of said second array by said first amount but not said second amount, then decrementing said corresponding value by said first amount, and

if a sample of said even beat is less than said corresponding value of said second array by said second amount or by greater than said second amount, then decrementing said corresponding value by said second amount.

16. The system of claim 12, wherein said comparing means comprises:

means for calculating a difference between sample points of said odd median complex and sample points of said even median complex to obtain said estimate of the amplitude of beat-to-beat alternation in said ECG signal.

17. A computer program product comprising a computer useable medium having computer program logic for enabling at least one processor in a computer system to quantify alternation in an ECG signal, said computer program logic comprising:

first means for enabling the at least one processor to receive digitized ECG data representing the ECG signal;

second means for enabling the at least one processor to calculate an odd median complex for at least one odd beat in said ECG data and to calculate an even median complex for at least one even beat in said ECG data; and

third means for enabling the at least one processor to compare said odd median complex with said even median complex to obtain an estimate of the amplitude of beat-to-beat alternation in said ECG signal.