Data compression of cardiac electrical signals using scanning correlation and temporal data compression

• US 5,215,098 A
• Filed: 08/12/1991
• Issued: 06/01/1993
• Est. Priority Date: 08/12/1991
• Status: Expired due to Term

• Current Assignee: Pacesetter Incorporated (Abbott Laboratories)
• Original Assignee: Telectronics Pacing Systems, Inc (Abbott Laboratories)
• Inventors: Steinhaus, Bruce M., Wells, Randy T.
• Primary Examiner(s): Howell, Kyle L.
• Assistant Examiner(s): Getzow, Scott M.
• Application Number:
  US07/744,466
• Time in Patent Office: 659 Days
• Field of Search: 128/696, 128/702
• US Class Current: 600/515
• CPC Class Codes: A61B 5/30 Input circuits therefor
• CPC Class Codes: A61B 5/308 for electrocardiography [ECG]
• CPC Class Codes: A61B 5/7232 involving compression of the physiological signal, e.g. to extend the signal recording period
• 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: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems

Abstract

A method and apparatus for storing a representation of a cardiac signal by compressing the data using scan correlation and temporal data compression techniques. The method and apparatus sense cardiac signals when the heart is functioning in a known cardiac state, then characterize this known state by storing a temporally compressed template of time sequence samples. The method and apparatus may perform testing for multiple different cardiac states and store templates associated with each state. Subsequently when the heart is functioning in an unknown cardiac state, the method and apparatus monitor cardiac signals by temporally compressing samples and scan correlating these samples with the previously stored templates to derive correlation coefficients. These correlation coefficients are a basis for identifying and classifying cardiac signal waveforms. For waveforms which correlate highly with a particular template, analysis of the timing of the maximum correlation coefficient provides a fiducial time which designates the time relationship of waveforms within different cardiac cycles. The method and apparatus store information in the form of templates, fiducial timing markers, and waveform occurrence counts. This data provides the information necessary to subsequently reproduce a long-term signal record.

Claims

1. A method of data compressing and storing cardiac electrical signals which are sensed from a patient'"'"'s heart, comprising the steps of:

sensing cardiac electrical signals when the heart is functioning in at least one known cardiac state;

characterizing each of said at least one known cardiac states in a time sequence of template samples;

temporally compressing and storing each of said time sequences of template samples;

monitoring a time sequence of cardiac electrical signal samples when the heart is functioning in an unknown cardiac state by sampling and temporally compressing said unknown cardiac state samples, scan correlating said compressed unknown cardiac state samples with at least one of said stored compressed template sample sequences to derive correlation coefficients associated with each of said temporally compressed unknown cardiac state samples and each of said temporally compressed template samples in said at least one stored temporally compressed template sample sequence, and mutually comparing said derived correlation coefficients associated with each of said unknown cardiac state samples and each of said template samples;

classifying each cf said unknown cardiac state time sequences of cardiac electrical signal samples as included or excluded from each of said known cardiac states on the basis of the results of said mutually comparing step and codifying the results of said classifying step;

determining a fiducial time for each of said unknown cardiac state sequences on the basis of the results of said mutually comparing step; and

storing a record of said codifications and associated fiducial times in a memory.

2. A method according to claim 1, further comprising the step of highpass filtering said sensed cardiac electrical signals when the heart is operating in each of said at least one known cardiac states and said unknown cardiac state to substantially eliminate the mean amplitude of each of said temporally compressed template samples and said temporally compressed unknown cardiac state samples.

3. A method according to claim 2, wherein said scan correlating sub-step of said monitoring step disregards the mean amplitude of each of said temporally compressed template samples and each of said temporally compressed unknown cardiac state samples when deriving said correlation coefficient.

4. A method according to claim 3, wherein each of said temporally compressing steps comprises the sub-steps of:

subtracting each of a predetermined number of consecutive noncompressed samples of said sensed cardiac electrical signal from the most recently determined compressed sample, wherein said predetermined number is a compression ratio;

storing each of said noncompressed samples and its associated absolute difference value from each subtracting step result;

mutually comparing each of said stored absolute difference values to determine the largest absolute difference; and

setting the current compressed sample to the value of the stored noncompressed sample associated with the largest absolute difference.

9. A method according to claim 1, wherein each of said temporally compressing steps comprises the sub-steps of:

determining the differences between the most recent sample of said temporally compressed time sequence and each of a predetermined number of consecutive noncompressed samples of said sensed cardiac electrical signal sequence, wherein said predetermined number is a compressions ratio;

mutually comparing said differences to identify the sample in said sensed cardiac electrical signal sequence associated with the largest difference; and

setting the current compressed sample to the value of the identified sample.

10. A method according to claim 1, wherein said step of characterizing said at least one known cardiac state in a time sequence of template samples comprises the substeps of:

sampling a first sequence of cardiac electrical signals sensed when the heart is functioning in said at least one known cardiac state;

storing said first sequence samples in a template memory;

sampling and storing subsequent cardiac electrical signals;

scan correlating said subsequent cardiac electrical signal samples with said template samples to derive a template correlation coefficient;

comparing said template correlation coefficient with a template threshold value;

if said template correlation coefficient is greater than said threshold value, aligning said subsequent cardiac electrical signal samples in time;

averaging said aligned subsequent cardiac electrical signal samples into said template; and

repeating said sampling and storing, scan correlating, comparing, aligning, and averaging steps for a predetermined number of iterations.

11. A method according to claim 10, further comprising the steps of:

monitoring said cardiac electrical signals to determine an intrinsic heart rate and its associated intrinsic cycle length; and

setting a template length limit restricting the size of the template to a predetermined percentage of said intrinsic cycle length.

13. A method in accordance with claim 11, further comprising the steps of:

locating an R wave in said sequence of cardiac electrical signals; and

aligning said first sequence samples in the template memory so that the R wave sample is stored in a predetermined sample location.

5. A method according to claim 5, wherein said step of characterizing said known cardiac state in a time sequence of template samples comprises the substeps of:

sampling a first sequence of cardiac electrical signals sensed when the heart is functioning in a known cardiac state;

storing said first sequence samples in a template memory;

sampling and storing subsequent cardiac electrical signals;

scan correlating said subsequent cardiac electrical signal samples with said template samples to derive a template correlation coefficient;

comparing said template correlation coefficient with a template threshold value;

if said template correlation coefficient is greater than said threshold value, aligning said subsequent cardiac electrical signal samples in time;

averaging said aligned subsequent cardiac electrical signal samples into said template; and

repeating said sampling and storing, scan correlating, comparing, aligning, and averaging steps for a predetermined number of iterations.

6. A method according to claim 5, further comprising the steps of:

monitoring said cardiac electrical signals to determine an intrinsic heart rate and its associated intrinsic cycle length; and

setting a template length limit restricting the size of the template to a predetermined percentage of said intrinsic cycle length.

7. A method in accordance with claim 5, further comprising the steps of:

locating an R wave in said sequence of cardiac electrical signals; and

aligning said first sequence samples in the template memory so that the R wave sample is stored in a predetermined sample location.

8. A method according to claim 6, further comprising the steps of:

locating an R wave in said sequence of cardiac electrical signals; and

aligning said first sequence samples in the template memory so that the R wave sample is stored in a predetermined sample location.

12. A method according to claim further comprising the steps of:

locating an R wave in said sequence of cardiac electrical signals; and

aligning said first sequence samples in the template memory to that the R wave sample is stored in a predetermined sample location.

14. An apparatus for data compressing and storing cardiac electrical signals which are sensed from a patient'"'"'s heart, comprising:

means for sensing cardiac electrical signals when the heart is functioning in at least one known cardiac state;

means for characterizing each of said at least one known cardiac states in a time sequence of template samples;

means for temporally compressing and storing each of said time sequences of template samples;

means for monitoring a time sequence of cardiac electrical signal samples when the heart is functioning in an unknown cardiac state, said monitoring means including means for sampling and temporally compressing said unknown cardiac state samples, means for scan correlating said compressed unknown cardiac state samples with at least one of said stored compressed template sample sequences to derive correlation coefficients associated with each of said temporally compressed unknown cardiac state samples and each of said temporally compressed template samples in said at least one stored temporally compressed template sample sequence, and means for mutually comparing said derived correlation coefficients associated with each of said unknown cardiac state samples and each of said template samples;

means for classifying each of said unknown cardiac state time sequences of cardiac electrical signal samples as included or excluded from each of said known cardiac states on the basis of the results of the comparison made by said mutually comparing means, said classifying means including means for codifying the results of the classification made by said classifying means;

means for determining a fiducial time for each of said unknown cardiac state sequences on the basis of the results of the comparison made by said mutually comparing means; and

means for storing a record of said codifications and associated fiducial times in a memory.

15. An apparatus according to claim 14, further comprising:

means for highpass filtering said sensed cardiac electrical signals when the heart is operating in each of said known cardiac states and said unknown cardiac state to substantially eliminate the mean amplitude of each of said temporally compressed template samples and said temporally compressed unknown cardiac state samples.

16. An apparatus according to claim 15, wherein said scan correlating means of said monitoring means disregards the mean amplitude of each of said temporally compressed template samples and each of said temporally compressed unknown cardiac state samples when deriving said correlation coefficient.

17. An apparatus according to claim 16, wherein each of said temporally compressing means further comprises:

means for subtracting each of a predetermined number of consecutive noncompressed ;

samples of said sensed cardiac electrical signal from the most recently determined compressed sample, wherein said predetermined number is a compression ratio;

means for storing each of said noncompressed samples and its associated absolute difference value from each subtraction result;

means for mutually comparing each of said stored absolute difference values to determine the largest absolute difference; and

means for setting the current compressed sample to the value of the stored noncompressed sample associated with the largest absolute difference.

18. An apparatus according to claim 17, further comprising:

means for sampling a first sequence of cardiac electrical signals sensed when the heart is functioning in a known cardiac state;

means for storing said first sequence samples in a template memory;

means for sampling and storing subsequent cardiac electrical signals;

means for scan correlating said subsequent cardiac electrical signal samples with said template samples to derive a template correlation coefficient;

means for comparing said template correlation coefficient with a template threshold value;

means, operative when said template correlation coefficient is greater than said threshold value, for aligning said subsequent cardiac electrical signal samples in time;

means for averaging said aligned subsequent cardiac electrical signal samples into said template; and

means for repeating said sampling, storing, scan correlating, comparing, aligning, and averaging means for a predetermined number of iterations.

19. An apparatus according to claim 18, further comprising:

means for monitoring said cardiac electrical signals to determine an intrinsic heart rate and its associated intrinsic cycle length; and

means for setting a template length limit restricting the size of the template to a predetermined percentage of said intrinsic cycle length.

20. An apparatus according to claim 19, further comprising:

means for locating an R wave in said sequence of cardiac electrical signals; and

means for aligning said first sequence samples in the template memory so that the R wave sample is stored in a predetermined sample location.

21. An apparatus according to claim 20, wherein said known cardiac states include ventricular tachycardia and normal sinus rhythm.

22. An apparatus according to claim 14, wherein each of said temporally compressing means further comprises:

means for determining the differences between the most recent sample of said temporally compressed time sequence and each of a predetermined number of consecutive noncompressed samples of said sensed cardiac electrical signal sequence, wherein said predetermined number is a compression ratio;

means for mutually comparing said differences to identify the sample in said sensed cardiac electrical signal sequence associated with the largest difference; and

means for setting the current compressed sample to the value of the identified sample.

23. An apparatus according to claim 14, wherein said means for characterizing said known cardiac state in a time sequence of template samples further comprises:

means for sampling a first sequence of cardiac electrical signals sensed when the heart is functioning in a known cardiac state;

means for storing said first sequence samples in a template memory;

means for sampling and storing subsequent cardiac electrical signals;

means for scan correlating said subsequent cardiac electrical signal samples with said template samples to derive a template correlation coefficient;

means for comparing said template correlation coefficient with a template threshold value;

means, operative when said template correlation coefficient is greater than said threshold value, for aligning said subsequent cardiac electrical signal samples in time;

means for averaging said aligned subsequent cardiac electrical signal samples into said template; and

means for sequentially actuating said sampling, storing, scan correlating, comparing, aligning, and averaging means for a predetermined number of iterations.

24. An apparatus according to claim 23, further comprising:

means for monitoring said cardiac electrical signals to determine an intrinsic heart rate and its associated intrinsic cycle length; and

means for setting a template length limit restricting the size of the template to a predetermined percentage of said intrinsic cycle length.

25. An apparatus according to claim 23, further comprising:

means for locating an R wave in said sequence of cardiac electrical signals; and

means for aligning said first sequence samples in the template memory so that the R wave sample is stored in a predetermined sample location.

26. An apparatus according to claim 14, further comprising:

means for locating an R wave in said sequence of cardiac electrical signals; and

means for aligning said first sequence samples in the template memory so that the R wave sample is stored in a predetermined sample location.