Non-invasive method and system for characterizing cardiovascular systems

• US 10,362,951 B2
• Filed: 04/24/2018
• Issued: 07/30/2019
• Est. Priority Date: 08/17/2012
• Status: Active Grant

• Current Assignee: Analytics For Life
• Original Assignee: Analytics For Life, Inc.
• Inventors: Gupta, Sunny, Yazdi, Mohsen Najafi, Burton, Timothy William Fawcett, Ramchandani, Shyamlal, Exner, Derek Vincent
• Primary Examiner(s): Mahmood, Nadia A
• Application Number:
  US15/961,213
• Publication Number:
  US 20180303360A1
• Time in Patent Office: 462 Days
• Field of Search: 600508, 600529, 600523, 600509
• CPC Class Codes: A61B 2576/023 for the heart
• CPC Class Codes: A61B 5/0205 Simultaneously evaluating both cardiovascular conditions and different types of body conditions, e.g. heart and respiratory condition
• CPC Class Codes: A61B 5/021 Measuring pressure in heart or blood vessels A61B5/0205 takes precedence
• CPC Class Codes: A61B 5/14551 for measuring blood gases
• CPC Class Codes: A61B 5/316 Modalities, i.e. specific diagnostic methods
• CPC Class Codes: A61B 5/318 Heart-related electrical modalities, e.g. electrocardiography [ECG]
• CPC Class Codes: A61B 5/339 Displays specially adapted therefor
• CPC Class Codes: A61B 5/341 Vectorcardiography [VCG]
• CPC Class Codes: A61B 5/7235 Details of waveform analysis A61B5/349 takes precedence
• CPC Class Codes: G16H 30/40 for processing medical images, e.g. editing

Abstract

The present disclosure uses physiological data, ECG signals as an example, to evaluate cardiac structure and function in mammals. Two approaches are presented, e.g., a model-based analysis and a space-time analysis. The first method uses a modified Matching Pursuit (MMP) algorithm to find a noiseless model of the ECG data that is sparse and does not assume periodicity of the signal. After the model is derived, various metrics and subspaces are extracted to image and characterize cardiovascular tissues using complex-sub-harmonic-frequencies (CSF) quasi-periodic and other mathematical methods. In the second method, space-time domain is divided into a number of regions, the density of the ECG signal is computed in each region and inputted into a learning algorithm to image and characterize the tissues.

Claims

1. A method of pre-processing data to extract variables for use in a machine learning operation to diagnose a pathology, the method comprising:

receiving a biopotential signal data set associated with a subject, said biopotential signal data set being associated with a biopotential signal collected from one or more electrical leads;

generating, via a processor, an estimated noiseless model of the received biopotential signal data set, wherein generation comprises iterative selection of member atoms of a pre-defined dictionary of member atoms to form a sparse approximation of the received biopotential signal data set;

extracting, via the processor, a plurality of features from a low-energy complex sub-harmonic subspace derived from the estimated noiseless model, wherein one or more of the plurality of extracted features includes one or more fractional derivative derived features of the low-energy complex sub-harmonic subspace; and

linking, via the processor, the one or more of the plurality of extracted features to a genetic algorithm to generate outputs that correlate with clinical parameters describing tissue architecture, structure and/or function.

2. The method of claim 1, wherein the biopotential signal data set comprises electrocardiographic (ECG) signals.

3. The method of claim 1, wherein the estimated noiseless model is generated based on a Matching Pursuit algorithm or a modified Matching Pursuit algorithm.

4. The method of claim 1, wherein the estimated noiseless model is generated based on a sparse approximation algorithm that does not assume periodicity of the signal.

5. The method of claim 1 further comprising:

removing, by the processor, baseline wander from the received biopotential signal data.

6. The method of claim 5 wherein the baseline wander is removed using a modified moving average filter.

7. The method of claim 1 wherein the estimated noiseless model is generated as a phase space representation.

8. The method of claim 7 further comprising:

performing, by the processor, a phase space transformation process, of the estimated noiseless model to generate the phase space representation.

9. The method of claim 8 further comprising:

synchronizing, by the processor, the phase space representation to a pre-defined dynamical behavior.

10. The method of claim 1, wherein the low-energy complex sub-harmonic subspace comprises about a last 20 percent of the selected member atoms.

11. The method of claim 1 comprising:

detecting, by the processor, a disease selected from the group consisting of hypertrophy, ischemia, scar, and abnormal electrical channel function.

12. The method of claim 1, wherein the one or more fractional derivative derived features comprise of the low-energy complex sub-harmonic subspace is based on an irrational fractional derivative of order α

.

13. The method of claim 2, wherein the outputs of the genetic algorithms are linked a 17-myocardial segment model of the ventricle.

14. The method of claim 13, further comprising:

identifying, by the processor, segments of the 17-myocardial segment model with altered architectural features and/or function.

15. The method of claim 14, wherein one or more segments of the 17-myocardial segment model are assigned a probability of a corresponding tissue having a pathophysiological abnormality.

16. The method of claim 15, wherein the pathophysiological abnormality is selected from the group consisting of hypertrophy, atrophy, scar, ischemia, edema, and fibrosis.

17. The method of claim 15, wherein the pathophysiological abnormality is ischemia.

18. The method of claim 2, wherein the one or more fractional derivative derived features comprise of the low-energy complex sub-harmonic subspace is based on an irrational fractional derivative of about −

1.5 or about −

2.5.

19. A system of pre-processing data to extract variables for use in a machine learning operation to diagnose a pathology, the system comprising:

a processor;

a memory having instructions stored thereon, wherein execution of the instructions by the processor causes the processor to;

receive a biopotential signal data associated with a subject, said biopotential signal data being associated with a biopotential signal collected from one or more electrical leads;

generate an estimated noiseless model of the received biopotential signal data, wherein generation comprises iterative selection of member atoms of a pre-defined dictionary of member atoms to form a sparse approximation of the received biopotential signal;

extract a plurality of features from a low-energy complex sub-harmonic subspace derived from the estimated noiseless model, wherein one or more of the plurality of extracted features includes one or more fractional derivative derived features of the low-energy complex sub-harmonic subspace; and

link the one or more of the plurality of extracted features to a genetic algorithm to generate outputs that correlate with clinical parameters describing tissue architecture, structure and/or function.

20. The system of claim 19, wherein the biopotential signal data set comprises electrocardiographic (ECG) signals.

21. The system of claim 19, wherein the estimated noiseless model is generated based on a Matching Pursuit algorithm or a modified Matching Pursuit algorithm.

22. The system of claim 19 wherein execution of the instructions by the processor further causes the processor to:

remove baseline wander from the received biopotential signal data.

23. The system of claim 22 wherein the baseline wander is removed using a modified moving average filter.

24. The system of claim 19 wherein the estimated noiseless model is generated as a phase space representation.

25. The system of claim 24 wherein execution of the instructions by the processor further causes the processor to:

perform a phase space transformation process, of the estimated noiseless model to generate the phase space representation.

26. The system of claim 25 wherein execution of the instructions by the processor further causes the processor to:

synchronize the phase space representation to a pre-defined dynamical behavior.

27. The system of claim 19, wherein the low-energy complex sub-harmonic subspace comprises about a last 20 percent of the selected member atoms.

28. The system of claim 19 wherein execution of the instructions by the processor further causes the processor to:

detect a disease selected from the group consisting of hypertrophy, ischemia, scar, and abnormal electrical channel function.

29. The system of claim 19, wherein the one or more fractional derivative derived features comprise of the low-energy complex sub-harmonic subspace is based on an irrational fractional derivative of order a.

30. The system of claim 19, wherein the outputs of the genetic algorithms are linked a 17-myocardial segment model of the ventricle.

31. The system of claim 19, wherein execution of the instructions by the processor further causes the processor to:

identify segments of the 17-myocardial segment model with altered architectural features and/or function.

32. The system of claim 31, wherein one or more segments of the 17-myocardial segment model are assigned a probability of a corresponding tissue having a pathophysiological abnormality.

33. The system of claim 32, wherein the pathophysiological abnormality is selected from the group consisting of hypertrophy, atrophy, scar, ischemia, edema, and fibrosis.

34. The system of claim 19, wherein the one or more fractional derivative derived features comprise of the low-energy complex sub-harmonic subspace is based on an irrational fractional derivative of about −

1.5 or about −

2.5.

35. A non-transitory computer readable medium having instructions stored thereon, wherein execution of the instructions by a processor cause the processor to:

receive a biopotential signal data associated with a subject, said biopotential signal data being associated with a biopotential signal collected from one or more electrical leads;

generate an estimated noiseless model of the received biopotential signal data, wherein generation comprises iterative selection of member atoms of a pre-deftned dictionary of member atoms to form a sparse approximation of the received biopotential signal;

extract a plurality of features from a low-energy complex sub-harmonic subspace derived from the estimated noiseless model, wherein one or more of the plurality of extracted features includes one or more fractional derivative derived features of the low-energy complex subharmonic subspace; and

link the one or more of the plurality of extracted features to a genetic algorithm to generate outputs that correlate with clinical parameters describing tissue architecture, structure and/or function.

36. The computer readable medium of claim 35, wherein the biopotential signal data set comprises electrocardiographic (ECG) signals.

37. The computer readable medium of claim 35, wherein the estimated noiseless model is generated based on a Matching Pursuit algorithm or a modified Matching Pursuit algorithm.

38. The computer readable medium of claim 35 wherein execution of the instructions by the processor further causes the processor to:

remove baseline wander from the received biopotential signal data.

39. The computer readable medium of claim 38 wherein the baseline wander is removed using a modified moving average filter.

40. The computer readable medium of claim 35 wherein the estimated noiseless model is generated as a phase space representation.

41. The computer readable medium of claim 40 wherein execution of the instructions by the processor further causes the processor to:

perform a phase space transformation process, of the estimated noiseless model to generate the phase space representation.

42. The computer readable medium of claim 41 wherein execution of the instructions by the processor further causes the processor to:

synchronize the phase space representation to a pre-deftned dynamical behavior.

43. The computer readable medium of claim 35, wherein the low-energy complex sub-harmonic subspace comprises about a last 20 percent of the selected member atoms.

44. The computer readable medium of claim 35 wherein execution of the instructions by the processor further causes the processor to:

detect a disease selected from the group consisting of hypertrophy, ischemia, scar, and abnormal electrical channel function.

45. The computer readable medium of claim 35, wherein the one or more fractional derivative derived features comprise of the low-energy complex sub-harmonic subspace is based on an irrational fractional derivative of order a.

46. The computer readable medium of claim 35, wherein the outputs of the genetic algorithms are linked a 17-myocardial segment model of the ventricle.

47. The computer readable medium of claim 35, wherein execution of the instructions by the processor further causes the processor to:

identify segments of the 17-myocardial segment model with altered architectural features and/or function.

48. The computer readable medium of claim 47 wherein one or more segments of the 17-myocardial segment model are assigned a probability of a corresponding tissue having a pathophysiological abnormality.

49. The computer readable medium of claim 48 wherein the pathophysiological abnormality is selected from the group consisting of hypertrophy, atrophy, scar, ischemia, edema, and fibrosis.

50. The computer readable medium of claim 35, wherein the one or more fractional derivative derived features comprise of the low-energy complex sub-harmonic subspace is based on an irrational fractional derivative of about −

1.5 or about −

2.5.