System and method for predicting ventricular tachycardia by adaptive high pass filter

• US 4,458,691 A
• Filed: 02/11/1982
• Issued: 07/10/1984
• Est. Priority Date: 02/11/1982
• Status: Expired due to Term

• Current Assignee: Arrhythmia Research Technology, Inc.
• Original Assignee: Arrhythmia Research Technology, Inc.
• Inventors: Netravali, Arun N.
• Primary Examiner(s): Kamm, William E.
• Application Number:
  US06/347,916
• Time in Patent Office: 880 Days
• Field of Search: 128/695, 128/696, 128/702-705, 128/708, 364/715, 364/717
• US Class Current: 600/518
• CPC Class Codes: A61B 5/366 Detecting abnormal QRS complex, e.g. widening

Abstract

A ventricular tachycardia (VT) prediction system for use with an ECG input having cyclical characteristics. The ECG input is periodically sampled during each of a plurality of ECG cycles, and signals representative of the ECG input for each corresponding sample time are stored. A high pass filter adaptively filters the signals for each ECG cycle, so as to provide a filter output systematically related to the high frequency content of the ECG input. The adaptive filter produces segments which are uninfluenced by other samples so as to negate ringing. A reference signal is provided for VT prediction. The filter output is compared to the reference signal, and the prediction of ventricular tachycardia is indicated responsive to the comparison. In one embodiment the samples are averaged for a plurality of ECG cycles prior to generating the filter output. In another embodiment, the system determines the duration of the QRS portion of the ECG cycles prior to generating the filter output. In another embodiment, the system determines the duration of the QRS portion of the ECG signal, provides a second reference signal for VT prediction, and compares the determined duration to the second reference signal. A prediction of ventricular tachycardia is indicated responsive to the second comparison.

Claims

1. A ventricular tachycardia prediction system for use with an ECG input having cyclical characteristics comprising:

means for periodically sampling the ECG input during each of a plurality of ECG cycles,

means for storing signals representative of the ECG input for each corresponding sample time;

filter means for adaptively high pass filtering the signals for each ECG cycle, so as to provide a filter output systematically related to the high frequency content of the ECG input;

means for providing a reference signal;

means for comparing the filter output to the reference signal; and

means for indicating the prediction of ventricular tachycardia responsive to said means for comparing.

2. The system as in claim 1 wherein:

said periodic sampling occurs N times per ECG cycle, and

said filter means is further comprised of

means for identifying a reference point P of the ECG input for each ECG cycle;

finite impulse response filter means, having a spread of samples about a center reference sample, comprising

means for deriving a filter output Q_n where n is a running index of the current sample number, and a_m are predefined filter constants selected to provide a finite impulse response such that ##EQU12## where (1) for filtering n>

p+l, then I_n+m =O for 0<

n+m, and for n+m>

N, and

(2) for filtering for p+l>

n>

p, then I_n+m =I_p for all n+m<

p; and

(3) for filtering for n<

p, then I_n+m =I_p for all n+m>

p.

3. The system as in claim 2

wherein the reference point P is the peak amplitude value of the ECG waveform for each cycle.

4. The system as in claim 1 wherein said filter means is further comprised of:

means for determining a peak of the ECG input waveform W,

means for sampling the waveform and storing N signals representative of N samples fo the waveform for each ECG cycle,

means for determining two reference samples Δ

1 and Δ

2,

means for determining a filter output Q_n where n is a running index of the samples, P is the center point of the ECG waveform,

2l+1 is the spread of filter samples, and a_m are predefined constants, according to;

##EQU13## (1) for filtering of n>

p+Δ

2, then W_n+m =W_p+Δ

2 for n+m<

p+Δ

2; and

(2) for filtering of p+Δ

2>

n>

p-Δ

1 then W_n+m =W_p+Δ

2 for n+m>

p+Δ

2; and

W_n+m =W_p-Δ

1 for n+m<

p-Δ

1; and

(3) for filtering of n<

p-Δ

1 then W_n+m =W_p-Δ

1, for n+m>

p-Δ

1.

5. The system as in claim 1 wherein said stored samples are filtered in the same time order as the ECG input was sampled.

6. The system as in claim 1 wherein said stored samples are filtered in the reverse time order as that which the ECG input was sampled.

7. The system as in claim 5 or 6 wherein said filter means is further comprising of:

means for averaging samples for a plurality of ECG cycles prior to deriving the filter output Q.

8. The system as in claim 1 further comprising:

means for determining the duration of the QRS portion of the ECG signal,

means for providing a second reference signal;

second means for comparing said determined duration to said second reference signal, and

means for indicating the prediction of ventricular tachycardia responsive to said second means for comparing.

9. The system as in claim 1 wherein said filter means adaptively filters the stored samples representative of the tail section of the QRS portion of the ECG to derive the filter output, Q.

10. The system as in claim 9 wherein said predefined value corresponds to a filter output indicative of a high frequency, low level, energy component in the tail of the ECG waveform.

11. A system, having X, Y, and Z ECG waveform inputs, for predicting ventricular tachycardia in a patient, said system comprising:

means for sampling and storing voltage inputs from each of the X, Y, and Z ECG waveforms at a fixed rate of N samples per predefined time interval,

means for identifying a peak value time point P for each of the X, Y, and Z ECG waveforms,

adaptive filter means responsive to a spread of ±

l sample points about any center sampled point, comprising

means for deriving a filter output Q_n responsive to the stored samples, where n is a running index, and A_m are predefined filter constants such that ##EQU14## where (1) for filtering of n<

p+l then I_n+m =0 for 0>

n+m and for n+m>

N;

(2) for filtering of P+l>

n>

p, then I_n+m =I_p for all n+m<

p; and

(3) for filtering of n<

p, then I_n+m =I_p for all n+m>

p; and

means for predicting the probable occurrence of ventricular tachycardia from the filter output.

12. A method for predicting ventricular tachycardia in a patient from cyclically recurring ECG inputs, comprising the steps of:

sampling and storing a plurality of points each ECG cycle at a fixed rate of N samples per cycle;

identifying a reference point p of the waveform,

filtering each sampled point with a spread of ±

l points about any center sampled point;

deriving a filter output Q_n responsive to the stored samples, where n is a running index and a_m are predefined filter constants, such that ##EQU15## (1) where for filtering of n>

p+l, then I_n+m =0 for 0>

n+m, and n+m>

N, and

(2) where for filtering of p+l>

n>

P, then I_n+m =I_p for all n+m<

p; and

(3) where for filtering of n<

p, then I_n+m =I_p for all n+m>

p;

determining a predefined reference signal,

comparing the filter output to said predefined reference signal and

indicating the prediction of potential ventricular tachycardia responsive to said comparison.

13. A method for predicting ventricular tachycardia from an ECG input waveform W comprising the steps of:

determining a peak value p of the waveform W,

sampling and storing the sampled values of the waveform W at a predefined rate;

determining two reference sample positions Δ

1 and Δ

2;

determining a filter output Q systematically related to the stored samples, where n is a running index of sample points, where 2l+1 is the spread of filter input samples analyzed, and where a_m are predetermined filter constants, such that ##EQU16## where (1) for filtering of n>

p+Δ

₂ then W_n+m =W_p+Δ

2 for n+m<

p+Δ

₂, and

(2) for filtering of p+Δ

₂ >

n>

p-Δ

₁ then W_n+m =W_p+Δ

2 for n+m>

p+Δ

2 and W_n+m =W_p-Δ

1 for n+m<

p-Δ

1; and

(3) for filtering of n<

p-Δ

1, then W_n+m =W_p-Δ

1 for n+m>

p-Δ

1;

determining a reference signal,

comparing the filter output to the reference signal, and

indicating the prediction of ventricular tachycardia responsive to said comparison.

14. A method for analyzing electrocardiograph signals to determine the presence or absence of a predetermined level of high frequency energy in the late QRS signal having X, Y, and Z parts, comprising the steps of:

sampling a series of analog QRS signals at spaced time segments,

converting each segment sample to a digital signal equivalent,

applying a portion of said digital signals to an adaptive high pass filter;

determining an arithmetic value of the amplitude of the output of said filter; and

comparing said value with said predetermined level.

15. The method of claim 14 further characterized in that said adaptive filter has a finite impulse response.

16. The method of claim 14 further comprising the step of generating an indication signal which is a function of said comparison.

17. The method of claim 14 further including the step of averaging the digital values of said series of QRS signals to obtain an averaged composite QRS signal before the application thereof to said adaptive high pass filter.

18. The method of claim 17 wherein said portion of said digital signals comprises signals representing at least the last 40 milliseconds of the averaged QRS signal.

19. The method of claim 18 wherein said arithmetic value is the root mean square.

20. The method of claim 18 wherein said arithmetic value is the sum of the absolute values of the filter output for the X, Y, and Z parts of the QRS portion of the ECG input.

21. The method of claim 17 further including the step of measuring the duration of said averaged QRS signal.

22. A system for analyzing electrocardiograph signals to determine the level of high frequency energy in the late QRS signal comprising:

means for converting X, Y, and Z lead electrocardiographic input signals to digital signals representative of times spaced samples;

means for examining said X, Y, and Z digital signals and selecting therefrom the QRS waveform portions thereof;

means for signal averaging a multiplicity of said selected QRS waveforms for each of said X, Y, and Z inputs and providing composite, digital signals representative of the averaged X, Y, and Z QRS waveforms;

high pass filter means having a finite impulse response;

means for applying to said filter means the digital signals representative of the late portion of each of said X, Y, and Z QRS waveforms;

means for providing a reference signal; and

means for comparing the output of said filter means with said reference signal to determine the presence of a high frequency, low level, energy component in the filter output predicative of ventricular tachycardia.

23. The system of claim 22 wherein said comparing means compares only a portion of said filter'"'"'s output.

24. The system of claim 23 wherein said portion of said filter'"'"'s output represents the filtered last 40 milliseconds of the QRS waveform.

25. The system of claim 24 wherein said comparing means combines selected portions of said filter'"'"'s output for each of the X, Y, and Z signal samples and determines an arithmetic value of said combined portions.

26. The system of claim 25 wherein said arithmetic value is the root mean square of said combined portions.

27. The system as in claim 25 wherein said arithmetic value is the sum of the absolute values of the filter outputs for each of said X, Y, and Z signals.

28. The system of claim 22 further comprising:

means for applying to said filter means the digital signals representative of the early portion of each of said X, Y, and Z QRS waveforms;

means for extracting from said filter'"'"'s output the time of commencement of said QRS waveform;

means responsive to said extraction means and said comparing means for determining the width of said QRS waveform,

means for providing a width reference signal,

means for indicating the prediction of probable ventricular tachycardia responsive to a comparison of the determined QRS width and the width reference signal.