Phonopneumograph system

• US 6,261,238 B1
• Filed: 03/29/1999
• Issued: 07/17/2001
• Est. Priority Date: 10/04/1996
• Status: Expired due to Term

• Current Assignee: Isonea Limited
• Original Assignee: Karmel Medical Acoustic Technologies Ltd. (Noam Gavriely, MD)
• Inventors: Gavriely, Noam
• Primary Examiner(s): Nasser, Robert L.
• Application Number:
  US09/269,513
• Time in Patent Office: 841 Days
• Field of Search: 600/525, 600/532-4, 600/538, 600/586
• US Class Current: 600/532
• CPC Class Codes: A61B 5/0823 Detecting or evaluating cough events
• CPC Class Codes: A61B 5/087 Measuring breath flow
• CPC Class Codes: A61B 7/003 Detecting lung or respiration noise

Abstract

A method of analyzing breath sounds produced by a respiratory system, the method comprising: measuring breath sounds produced by the respiratory system; tentatively identifying a signal as being caused by a breath sound of a given type if it meets a first criteria characteristic of the breath sound of the given type; and confirming said identification if a tentatively identified signal meets a second criteria characteristic of the breath sound of the given type.

Claims

1. A method of analyzing breath sounds produced by a respiratory system, the method comprising:

measuring breath sounds produced by the respiratory system;

tentatively identifying a signal as being caused by a breath sound of a given type if it meets a first criteria that tentatively identifies only the breath sound of the given type; and

confirming said identification if a tentatively identified signal meets a second criteria characteristic of the breath sound of the given type.

2. A method according to claim 1 comprising segmenting said breath sound data into segments and wherein tentatively identifying and confirming are based on time segments of breath sound data.

3. A method according to claim 1 wherein confirming said identification comprises comparing breath sounds tentatively identified as being of a given type to at least one of a plurality of second criteria each of which is characteristic of the tentatively identified given type and confirming that the breath sound is of the tentatively identified given type if it meets the second criteria characteristic of the given type.

4. A method according to claim 1 wherein tentatively identifying comprises comparing the breath sound to a plurality of first criteria, each said criteria being tentatively identify a breath sound as being of a given type, wherein said breath sound is tentatively identified as being only of the type for which it meets the first criteria.

5. A method according to claim 1 or claim 4 wherein the breath sound comprises a wheeze.

6. A method according to claim 5 wherein tentatively identifying the breath sound as a wheeze comprises detecting narrow peaks within the spectrum of the breath sound and determining if the narrow peaks are located within a narrow frequency range over a number of consecutive time periods.

7. A method according to claim 6 wherein confirming comprises determining if the narrow peaks of said tentatively identified wheeze have less than three harmonics each.

8. A method according to claim 6 wherein said consecutive time periods span at least 150 ms.

9. A method according to claim 6 wherein a breath sound is confirmed as a squeak type wheeze when said consecutive time periods span between 80 and 150 ms.

10. A method according to claim 6 wherein a breath sound is classified as low frequency wheeze when the frequency of the narrow peak is less than 400 Hz.

11. A method according to claim 6 wherein a breath sound is classified as a high frequency wheeze when the frequency of the narrow peak is between 400 Hz and 1600 Hz.

12. A method according to claim 6 wherein a breath sound is classified as an ultra-high frequency wheeze if the frequency of the narrow beak is above 1600 Hz.

13. A method according to claim 5 wherein said narrow frequency range is not greater then 64 Hz among any two consecutive time periods.

14. A method of according to claim 1 or claim 4 wherein the given breath type comprises a rhonchus.

15. A method according to claim 14 wherein tentatively identifying comprises detecting narrow peaks within a spectrum of said segment and determining if said narrow peaks are located within a predetermined small frequency range across during consecutive time periods.

16. A method according to claim 15 wherein confirming the identification of a breath sound as a rhonchus comprises, if there are more than two harmonics in each of the consecutive time periods:

generating a transfer function in the frequency domain between said breath sound data and measured ambient sound in the space surrounding the patient;

determining a coherence graph of said transfer function;

confirming each narrow peak as a rhonchus if the frequency range of high coherence of said coherence graph does not correspond to the frequency range of said narrow peaks.

17. A method according to claim 1 or claim 4 wherein the breath sound comprises a cough.

18. A method according to claim 17 wherein tentatively identifying comprises:

coincidentally detecting sudden loud ambient noise, a sudden loud breath sound, and a sudden chest motion.

19. A method according to claim 18 wherein confirming that a breath sound is a cough comprises:

generating an envelope of said breath sound data and determining the duration of said envelope;

determining that the sound is a cough if it fulfills all of the following conditions;

a) the breath sound takes place during expiration;

b) the breath sound peaks generally coincide with the ambient noise peaks;

c) the envelope of the breath sound data has a double hump shape;

d) the duration of the envelope is within a predetermined time range; and

e) determining that the frequency spectra of the sound are broad band with at least a predetermined high coherence level between the ambient and the breath sounds.

20. A method according to claim 19 and wherein said predetermined time range is 0.2-3.0 seconds.

21. A method according to claim 19 or claim 20 wherein said predetermined coherence level is 0.7 or greater.

22. A method according to claim 19 wherein the cough is classified as a productive cough if the second hump has a variance greater than 40 msec and is skewed to later times.

23. A method according to claim 19 wherein the cough is classified as a dry cough if it has a variance of less than 40 msec and is not substantially skewed.

24. A method according to claim 19 wherein the cough is classified as a barking cough if it is has a duration of between 200 and 350 msec and has a second hump of 10%-25% above the value between the humps.

25. A method according to claim 1 or claim 4 wherein the breath sound comprises a crackle.

26. A method according to claim 25 wherein tentatively identifying a breath sound as a crackle comprises:

finding the locations of abrupt changes in said breath sound data.

27. A method according to claim 26 wherein confirming that a breath sound is a crackle comprises:

matching breath sound data following the said abrupt changes to the following curve;

 

y=A*B(t)*C(t)

where y is said breath sound data starting at the abrupt change at which t begins, A is an amplitude parameter, B(t) is an envelope function and C(t) is an oscillatory function.

28. A method according to claim 27 wherein:

$B<br><br>\left(t\right)=\frac{t^n}{1+{\left(\mathrm{mt}\right)}^k}$$\mathrm{and}$$C<br><br>\left(t\right)=\sin <br><br>\left(\frac{2<br><br>\mathrm{\pi <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>f_0<br><br>t}{1+\mathrm{Ct}}\right)$

where ∫

₀, C, n, m and k are parameters.

29. A method according to claim 28 including identifying the type of crackle from the values of said parameters.

30. A method according to claim 29 wherein the cackle is identified as a fine cackle if ∫

₀ is above 600 Hz.

31. A method according to claim 29 wherein the cackle is defined as a coarse cackle if the if C, the rate of change of the cackles internal frequency is greater than 100.

32. A method according to claim 26 and including identifying the portion of the breathing cycle during which the crackle occurs.

33. A method according to claim 1 or claim 4 wherein the breath sound comprises a snore.

34. A method according to claim 33 wherein tentatively identifying a breath sound as a snore comprises:

determining that the breath sound occurs during an inspiratory period; and

determining that the breath sound is highly correlated with ambient noise in the region surrounding the patient.

35. A method according to claim 34 wherein confirming that the tentatively identified breath sound is a snore comprising:

identifying peaks in the inspiratory period and determining the average peak-to-peak time delta_t;

identifying at least three peaks in the power spectrum of the inspiratory period which are significantly large and determining the average peak-to-peak frequency delta_f;

generating a coherence graph for a transfer function in the frequency domain between said breath sound data and measured ambient noise of the space where said breath sound data was gathered;

identifying a snore if delta_t is close to the inverse of delta_f and if the frequency range of high coherence of said coherence graph corresponds to the frequency range of said narrow peaks.

36. A method according to claim 35 ad wherein identifying peaks in the inspiratory period comprises:

calculating the inverse Fourier transform of said power spectrum of the inspiratory period said generating thereby cleaned data;

searching for a first peak in said cleaned data beginning between 5 to 25 ms from the start of said segment; and

searching for peaks following said first peak.

37. A method according to claim 35 wherein identifying three peaks in the power spectrum comprises:

calculating the histogram of the power spectrum,

determining the variance of the histogram; and

defining as peaks those points having values which are at k variance or higher, where k is at least three.

38. A method according to claim 35 wherein the snore is classified as a simple snore if the spectrum contains 3 or 4 resonance peaks.

39. A method according to claim 35 wherein the snore is classified as a complex snore if the spectrum contains 5 or more resonance peaks.

40. A method according to claim 1 or claim 4 wherein the breath sound comprises a chest wall breath sound.

41. A method according to claim 40 wherein the breath sound is identified as a chest wall breath sound if it fits a low pass filter shaped curve having the parameters of amplitude, corner frequency and slope.

42. A method according to claim 41 wherein the breath sound is confirmed as a bronchial chest wall breath sound if the corner frequency is higher than a normal frequency for the age of the patient.

43. A Phonopneumograph for analyzing breath sounds produced by a respiratory system, the phonopneumograph comprising:

at least one auditory breath sensor which receives sound related to breath sounds and produces signals in response to the received sounds;

at least one first sound signal analyzer which analyzes said signals to initially screen said breath sounds said initial screening being operative to produce a tentative identification of the signals as being caused by a breath sound of a given type if it meets a first criteria; and

at least one second sound signal analyzer which further analyzes those of said signals which have been tentatively identified to confirm if the signals are caused by a breath sound of the given type.

44. A Phonopneumograph according to claim 43 wherein the at least one first sound analyzer comprises a plurality of first sound analyzers, each of which is operative to initially screen said breath sounds, each of said first sound analyzers producing a tentative identification of a breath sound of a different type.

45. A phonopneumograph according to claim 43 wherein the at least one breath sensor comprises a plurality of breath sensors placed at different positions with respect to the respiratory system.

46. A phonopneumograph according to claim 43 or claim 44 wherein the at least one second sound analyzer comprises a plurality of second sound analyzers, each of which is operative to further screen breath sounds of the tentatively identified given type and to produce a confirmation of the tentative identification if the breath sound additional criteria characteristic of the second type.

47. A phonopneumograph according to claim 43 or claim 44 wherein the breath sound of a given type comprises a crackle.

48. A phonopneumograph according to claim 43 or claim 44 wherein the breath sound of a given type comprises a cough.

49. A phonopneumograph according to claim 43 or claim 44 wherein the breath sound of a given type comprises a snore.

50. A phonopneumograph according to claim 43 or claim 44 wherein the breath sound of a given type comprises a rhonchus.

51. A phonopneumograph according to claim 43 or claim 44 wherein the breath sound of a given type comprises a wheeze.

52. A phonopneumograph according to claim 43 or claim 44 wherein the breath sound of a given type comprises a breath.

53. A phonopneumograph according to claim 43 or claim 44 wherein the first sound analyzer divides the sound into time segments and analyzes the sound on a segment by segment basis.

54. A method of analysis of breath sound data comprising:

determining the presence of a first breath sound in the breath sound data;

adjusting the breath sound data by reducing the effect of the first breath sound in the breath sound data; and

determining the presence of a second breath sound in the adjusted breath sound data.

55. A method according to claim 54 wherein the first breath sound is a wheeze.

56. A method according to claim 54 or claim 55 wherein the second breath sound is a wheeze.

57. A method according to claim 54 wherein the first breath sound is a crackle.

58. A method according to claim 54 or 57 wherein the second breath sound is a crackle.

59. A method according to claim 54 wherein the first breath sound and the second breath sound are of the same type.

60. A method according to claim 54 wherein the first breath sound and the second breath sound are adventitious breath sounds.

61. A method according to claim 54 wherein reducing the effect of the first breath sound comprises substantially removing the effect of the first breath sound.

62. A method according to claim 54 wherein determining the presence of a breath sound comprises segmenting the breath sound data segments and making the determination based on time segments of breath sound data.

63. A method according to any of claim 54, 55, 61, 59 or 60 wherein determining comprises:

tentatively identifying a signal as being caused by a breath sound of a given type if it meets a first criteria that tentatively identifies only the breath sound of a given type; and

confirming said identification if a tentatively identified signal meets a second criteria characteristic of the breath sounds of the given type.

64. A method according to claim 63 wherein confirming said identification comprises comparing breath sounds tentatively identified as being of a given type to at least one of a plurality of second criteria each of which is characteristic of the tentatively identified given type and confirming that the breath sound is of the tentatively identified given type if it meets the second criteria characteristic of the given type.

65. A method according to claim 63 wherein tentatively identifying comprises comparing the breath sound to a plurality of fir criteria, each said criteria being used to tentatively identity a breath sound as being of a given type, wherein said breath sound is tentatively identified as being only of the type for which it meets the first criteria.