Separating motion from cardiac signals using second order derivative of the photo-plethysmogram and fast fourier transforms

• US 20030225337A1
• Filed: 01/30/2003
• Published: 12/04/2003
• Est. Priority Date: 01/31/2002
• Status: Active Grant

• Current Assignee: Spacelabs Healthcare LLC (OSI Systems Incorporated)
• Original Assignee: Dolphin Medical, Inc. (OSI Systems Incorporated)
• Inventors: Scharf, John E., Shah, Bhavin B.
• Granted Patent: US 7,020,507 B2
• Time in Patent Office: Days
• US Class Current: 600/508
• CPC Class Codes: A61B 5/14551 for measuring blood gases
• CPC Class Codes: A61B 5/7207 of noise induced by motion artifacts
• CPC Class Codes: A61B 5/7239 using differentiation including higher order derivatives
• CPC Class Codes: A61B 5/7253 characterised by using transforms
• CPC Class Codes: A61B 5/7257 using Fourier transforms
• CPC Class Codes: A61B 5/726 using Wavelet transforms

Abstract

The present invention is directed toward a pulse oximetry system for the determination of a physiological parameter capable of removing motion artifacts from physiological signals comprises a hardware subsystem and a software subsystem. The software subsystem is used in conjunction with the hardware subsystem to perform a method for removing a plurality of motion artifacts from the photo-plethysmographic data and for obtaining a measure of at least one physiological parameter from the data. The method comprises acquiring the raw photo-plethysmographic data, transforming the data into the frequency domain, analyzing the transformed data to locate a series of candidate cardiac spectral peaks (primary plus harmonics), reconstructing a photo-plethysmographic signal in the time domain with only the candidate cardiac spectral peaks (primary plus harmonics), computing the second order derivative of the reconstructed photo-plethysmographic signal, analyzing the candidate second order derivative photo-plethysmographic signal to determine the absence or presence of cardiac physiologic signal characteristics, and finally selecting the best physiologic candidate from the series of potential cardiac spectral peaks (primary plus harmonics) based upon a second derivative scoring system. This scoring system is preferentially based upon second derivative processing analysis, but can be equally applied using the first, third, fourth or other similar derivative processing analysis.

Claims

1. A method for use in a system comprising a sensor for acquiring photo-plethysmographic data of a patient and a signal processing system for removing a plurality of motion artifacts from said photo-plethysmographic data and for obtaining a measure of at least one physiological parameter from said data, the method comprising:

(a) acquiring said photo-plethysmographic data;

(b) transforming said data;

(c) analyzing said transformed data to locate candidate spectral peaks;

(d) reconstructing a photo-plethysmographic signal;

(e) computing the second order derivative of said reconstructed photo-plethysmographic signal;

(f) analyzing said second order derivative photo-plethysmographic signal to determine a plurality of parameters;

(g) identifying signal representations of a plurality of motion artifacts; and

(h) selecting a candidate spectral peak from a plurality of potential cardiac spectral peaks.

2. The method of claim 1 further comprising the step of conditioning the photo-plethysmographic data prior to transforming the data.

3. The method of claim 2, wherein conditioning said photo-plethysmographic data reduces spectral leakage from subsequent signal processing and frequency analysis.

4. The method of claim 3, wherein said conditioning includes filtering said photo-plethysmographic data.

5. The method of claim 4, wherein said filtering is performed with a Hanning window.

6. The method of claim 1 wherein transforming said data comprises the step of transforming the data from a time domain to a frequency domain.

7. The method of claim 6, wherein transforming said data into the frequency domain is performed with a Fast Fourier Transform.

8. The method of claim 7, wherein said transforming is performed with a Fast Fourier Transform of a window size of approximately 1024 data points.

9. The method of claim 8, wherein said window size of approximately 1024 data points is obtained by sampling a 37 second segment of said data with a rate of approximately 27.5 data points per second.

10. The method of claim 6, wherein transforming said data into the frequency domain is performed with a technique selected from the group comprising a periodogram, a correlogram, autoregressive methods, Prony'"'"'s method, minimum variance methods, maximum likelihood methods, a discrete cosine transform, a wavelet transform, a discrete Hartley transform, and a Gabor transform.

11. The method of claim 1, wherein analyzing said transformed data to locate candidate spectral peaks comprises the steps of assigning a largest power amplitude from said data as a primary candidate spectral peak and assigning a next largest power amplitude that is not a harmonic of said primary candidate spectral peak as a secondary candidate spectral peak.

12. The method of claim 11, wherein said analyzing transformed data to locate candidate spectral peaks is performed using the first derivative criteria, wherein said candidate spectral peak is a frequency bin having a slope value equal to approximately zero.

13. The method of claim 12, wherein said first derivative criteria comprises a first criterion for detecting extrema, having a plurality of slopes, and a second criterion for verifying maxima or minima, wherein said extrema are considered maxima or minima based upon whether a second derivative of the slopes of extrema is lesser than or greater than zero.

14. The method of claim 1, further comprising the step of classifying said candidate spectral peaks detected into primary peaks (P_j) and harmonics (H_j) to said primary peaks, where j=1, . . . , m and wherein a maximum of m primary frequencies are identified as candidate spectral peaks.

15. The method of claim 1 wherein reconstructing a photo-plethysmographic signal comprises the step of transforming data from a frequency domain data into a time domain.

16. The method of claim 15, wherein transforming the frequency domain data into the time domain is performed using an inverse fast Fourier transform.

17. The method of claim 15 wherein the time domain photo-plethysmographic data is reconstructed by including only a plurality of primary candidate spectral peaks and a plurality of corresponding harmonic spectral peaks.

18. The method of claim 1 wherein the second order derivative photo-plethysmographic signal is analyzed to determine the absence or presence of cardiac physiologic signal characteristics.

19. The method of claim 18 wherein analyzing said second order derivative photo-plethysmographic signal comprises the step of performing a morphological analysis of the second order derivative photo-plethysmographic signal to determine a unique cardiac morphology, wherein said cardiac morphology consists of a plurality of systolic and a plurality of diastolic waves, and wherein said analysis accounts for a plurality of wave transitions.

20. The method of claim 19 wherein analyzing said second order derivative photo-plethysmographic signal further comprises the step of performing a morphological analysis of the second order derivative photo-plethysmographic signal to determine at least one height of said wave transitions.

21. The method of claim 1 wherein analyzing said second order derivative photo-plethysmographic signal comprises the steps of:

determining a plurality of wave transitions, calculating a plurality of heights of said wave transitions, and applying a criterion for detecting a predefined type of wave transition.

22. The method of claim 21 wherein the criterion for detecting a predefined type of wave transition among a first wave, a second wave, a third wave, a fourth wave, and a fifth wave comprises identifying the largest transition between a first wave and a second wave and determining whether a set of first, second, third, fourth, and fifth waves adhere to a formula:

height (first wave to second wave)>

{height (second wave to third wave), height (third wave to fourth wave), height (fourth wave to fifth wave)}.

23. The method of claim 21 wherein the criterion for detecting a predefined type of wave transition among a first wave, a second wave, a third wave, a fourth wave, a fifth wave, and a sixth wave comprises identifying the largest transition between a first wave and a second wave and determining whether a set of first, second, third, fourth, fifth, and sixth waves adhere to a formula:

height (first wave to second wave)>

{height (second wave to third wave), height (third wave to fourth wave), height (fourth wave to fifth wave), height (fifth wave to sixth wave)}.

24. The method of claim 1, wherein the removal of motion artifacts from the photo-plethysmographic signal comprises classifying a first primary-harmonic set with a first highest score_j as a likely cardiac geometry and classifying a second primary-harmonic set with a highest score_k as a likely motion geometry.

25. The method of claim 1, wherein the removal of motion artifacts from the photo-plethysmographic signal comprises classifying a first primary-harmonic set with a first highest score_j as a likely cardiac geometry and classifying a second primary-harmonic set with a highest score_k as a likely motion geometry based upon a wavelet analysis.

26. The method of claim 25, wherein the wavelet analysis is conducted by calculating correlation coefficients.

27. The method of claim 26, wherein the correlation coefficients are calculated by applying a wavelet correlation coefficient scoring system.

28. The method of claim 27, wherein the wavelet correlation coefficient scoring system comprises at least one of Symlets #7, Daubechies #9 or Morlet wavelets applied to a second derivative of candidate cardiac spectral peaks.

29. The method of claim 1 further comprising the step of removing a plurality of motion artifacts from the reconstructed photo-plethysmographic signal to create a motion artifact independent photo-plethysmographic signal.

30. The method of claim 1 further comprising the step of computing a physiological parameter from the motion artifact independent photo-plethysmographic signal.

31. The method of claim 1 wherein the candidate spectral peak is selected from a set of potential cardiac spectral peaks based upon a second order derivative scoring system.

32. The method of claim 1 wherein the candidate spectral peak is selected from a set of potential cardiac spectral peaks based upon a derivative scoring system.

33. A pulse oximetry system for the determination of a physiological parameter capable of removing motion artifacts from physiological signals comprising a hardware subsystem and a software subsystem,

wherein said hardware subsystem comprises an optical sensor unit for providing said pulse oximetry system with a patient'"'"'s photo-plethysmographic data, wherein said sensor comprises a light emitting diode for emitting light, a photo-detector for detecting said light, wherein said detected light contains information on the photo-plethysmographic data, a processor for processing the photo-plethysmographic data, and a memory unit, wherein said memory unit is operably coupled to said processor, and

wherein the software subsystem performs a method comprising the steps of:

(a) acquiring said photo-plethysmographic data from the sensor;

(b) conditioning said photo-plethysmographic data for signal processing;

(c) transforming the conditioned data into a frequency domain;

(d) analyzing the frequency domain data to locate candidate spectral peaks;

(e) transforming the frequency domain data into a time domain

(f) reconstructing the photo-plethysmographic signal;

(g) computing the second order derivative of the time domain photo-plethysmographic signal;

(h) analyzing the second order derivative time domain photo-plethysmographic signal to identify parameters;

(i) measuring said parameters;

(j) removing a motion artifact from the photo-plethysmographic signal; and

(k) computing a physiological parameter for the motion artifact independent photo-plethysmographic signal.

34. The system of claim 33 wherein the method further comprising the step of conditioning the photo-plethysmographic data prior to transforming the data.

35. The system of claim 34 wherein conditioning said photo-plethysmographic data reduces spectral leakage from subsequent signal processing and frequency analysis.

36. The system of claim 33 wherein the step of transforming said data comprises the step of transforming the data from a time domain to a frequency domain.

37. The system of claim 33 wherein analyzing said transformed data to locate candidate spectral peaks comprises the steps of assigning a largest power amplitude from said data as a primary candidate spectral peak and assigning a next largest power amplitude that is not a harmonic of said primary candidate spectral peak as a secondary candidate spectral peak.

38. The system of claim 33 wherein said analyzing transformed data to locate candidate spectral peaks is performed using the first derivative criteria, wherein said candidate spectral peak is a frequency bin having a slope value equal to approximately zero.

39. The system of claim 33 wherein said first derivative criteria comprises a first criterion for detecting extrema, having a plurality of slopes, and a second criterion for verifying maxima or minima, wherein said extrema are considered maxima or minima based upon whether a second derivative of the slopes of extrema is lesser than or greater than zero.

40. The system of claim 33 further comprising the step of classifying said candidate spectral peaks detected into primary peaks (P_j) and harmonics (H_j) to said primary peaks, where j=1, . . . , m and wherein a maximum of m primary frequencies are identified as candidate spectral peaks.

41. The system of claim 33 wherein reconstructing a photo-plethysmographic signal comprises the step of transforming data from a frequency domain data into a time domain.

42. The system of claim 41 wherein the time domain photo-plethysmographic data is reconstructed by including only a plurality of primary candidate spectral peaks and a plurality of corresponding harmonic spectral peaks.

43. The system of claim 33 wherein the second order derivative photo-plethysmographic signal is analyzed to determine the absence or presence of cardiac physiologic signal characteristics.

44. The system of claim 43 wherein analyzing said second order derivative photo-plethysmographic signal comprises the step of performing a morphological analysis of the second order derivative photo-plethysmographic signal to determine a unique cardiac morphology, wherein said cardiac morphology consists of a plurality of systolic and a plurality of diastolic waves, and wherein said analysis accounts for a plurality of wave transitions.

45. The system of claim 44 wherein analyzing said second order derivative photo-plethysmographic signal further comprises the step of performing a morphological analysis of the second order derivative photo-plethysmographic signal to determine at least one height of said wave transitions.

46. The system of claim 33 wherein analyzing said second order derivative photo-plethysmographic signal comprises the steps of:

determining a plurality of wave transitions, calculating a plurality of heights of said wave transitions, and applying a criterion for detecting a predefined type of wave transition.

47. The system of claim 46 wherein the criterion for detecting a predefined type of wave transition among a first wave, a second wave, a third wave, a fourth wave, and a fifth wave comprises identifying the largest transition between a first wave and a second wave and determining whether a set of first, second, third, fourth, and fifth waves adhere to a formula:

height (first wave to second wave)>

{height (second wave to third wave), height (third wave to fourth wave), height (fourth wave to fifth wave)}.

48. The system of claim 46 wherein the criterion for detecting a predefined type of wave transition among a first wave, a second wave, a third wave, a fourth wave, a fifth wave, and a sixth wave comprises identifying the largest transition between a first wave and a second wave and determining whether a set of first, second, third, fourth, fifth, and sixth waves adhere to a formula:

height (first wave to second wave)>

{height (second wave to third wave), height (third wave to fourth wave), height (fourth wave to fifth wave), height (fifth wave to sixth wave)}.

49. The system of claim 33 wherein the removal of motion artifacts from the photo-plethysmographic signal comprises classifying a first primary-harmonic set with a first highest score_j as a likely cardiac geometry and classifying a second primary-harmonic set with a highest score_k as a likely motion geometry.

50. The system of claim 33 wherein the removal of motion artifacts from the photo-plethysmographic signal comprises classifying a first primary-harmonic set with a first highest score_j as a likely cardiac geometry and classifying a second primary-harmonic set with a highest score_k as a likely motion geometry based upon a wavelet analysis.