Method and system for detecting a respiratory signal

• US 9,351,665 B2
• Filed: 01/14/2010
• Issued: 05/31/2016
• Est. Priority Date: 01/14/2009
• Status: Active Grant

• Current Assignee: Widemed Technologies Ltd.
• Original Assignee: Widemed Technologies Ltd.
• Inventors: Amos, Yariv Avraham, Kaminski, Gil, Barak-Shinar, Deganit
• Primary Examiner(s): Kahelin, Michael
• Assistant Examiner(s): Alter, Mitchell E
• Application Number:
  US13/144,502
• Publication Number:
  US 20110275910A1
• Time in Patent Office: 2,329 Days
• US Class Current: 1/1
• CPC Class Codes: A61B 5/02416 using photoplethysmograph signals, e.g. generated by infrared radiation A61B5/14552 takes precedence
• CPC Class Codes: A61B 5/0816 Measuring devices for examining respiratory frequency measuring frequency of electric signals G01R23/00
• CPC Class Codes: A61B 5/1455 using optical sensors, e.g. spectral photometrical oximeters

Abstract

A method of analyzing a plethysmograph signal is disclosed. The method comprises: extracting from the signal a plurality of features, thereby constructing a feature space. The method further comprises employing a path selection procedure to the feature space for determining at least one sequence of respiration frequencies and reconstructing a respiratory signal using the sequence.

Claims

1. A method of analyzing a photoplethysmograph signal, comprising:

segmenting the photoplethysmograph signal into at least a plurality of temporal segments;

extracting from each temporal segment of said plurality of temporal segments a plurality of features including at least respiration parameters, to construct a multidimensional feature space;

employing a path selection procedure to said multidimensional feature space to assign a respiration frequency or a range of respiration frequencies separately for each segment of said plurality of temporal segments, thereby determining at least one sequence of respiration frequencies, wherein each respiration frequency or range of respiration frequencies corresponds to a state of said multidimensional feature space and wherein said path selection procedure calculates probabilities of transitions between states;

reconstructing a respiratory signal using said sequence; and

displaying said reconstructed respiratory signal and adjusting a therapy based on said displayed reconstructed respiratory signal.

2. The method according to claim 1, further comprising using said sequence of respiration frequencies for selecting a corresponding sequence of respiration tidal volumes.

3. The method according to claim 1, wherein said plurality of features further comprises spectral parameters, and the method further comprises analyzing the spectral parameters for detecting temporal segments corresponding to a physiological event, wherein said path selection procedure is employed separately for different types of physiological events.

4. The method according to claim 3, wherein said physiological event is selected from the group consisting of atrial fibrillation, periodic limb movements and apnea.

5. The method according to claim 3, further comprising updating said feature space following identification of a segment corresponding to a physiological event.

6. The method according to claim 5, further comprising calculating a transfer function filter for each segment other than segments corresponding to physiological events, wherein said updating of said feature space comprises filtering a respective segment corresponding to a respective physiological event using a transfer function filter of a preceding segment.

7. The method according to claim 1, wherein said path selection procedure comprises an expectation maximization algorithm.

8. The method according to claim 1, wherein said path selection procedure features a hidden Markov model.

18. A method of assessing respiration of a subject, comprising connecting a photoplethysmograph sensor to the subject and analyzing a photoplethysmograph signal received from said sensor using the method of claim 1.

19. The method of claim 1, wherein said plurality of features comprises at least one additional feature selected from the group consisting of spectral parameters and signal envelope parameters.

9. Apparatus for analyzing a photoplethysmograph signal, comprising:

an input unit for receiving the photoplethysmograph signal; and

a data processor configured for segmenting the photoplethysmograph signal into at least a plurality of temporal segments;

extract from each temporal segment of said plurality of temporal segments a plurality of features including at least respiration parameter, thereby constructing a multidimensional feature space;

employing a path selection procedure to said multidimensional feature space to assign a respiration frequency or a range of respiration frequencies separately for each segment of said plurality of temporal segments, thereby determining at least one sequence of respiration frequencies;

reconstructing a respiratory signal using said sequence, wherein each respiration frequency or range of respiration frequencies corresponds to a state of said feature space and wherein said path selection procedure calculates probabilities of transition between states, and displaying said reconstructed respiratory signal.

10. The apparatus according to claim 9, wherein said data processor is configured for using said sequence of respiration frequencies for selecting a corresponding sequence of respiration tidal volumes.

11. The apparatus according to claim 9, wherein said plurality of features further comprises spectral parameters, and wherein said data processor is configured for analyzing the spectral parameters for temporal segments corresponding to a physiological event, wherein said path selection procedure is employed separately for different types of physiological events.

12. The apparatus according to claim 11, wherein said physiological event is selected from the group consisting of atrial fibrillation, periodic limb movements and apnea.

13. The apparatus according to claim 11, wherein said data processor is configured for updating said feature space following identification of segment corresponding to a physiological event.

14. The apparatus according to claim 13, wherein said data processor is configured for calculating a transfer function filter for each segment other than segments corresponding to physiological events, wherein said updating of said feature space comprises filtering a respective segment corresponding to a respective physiological event using a transfer function filter of a preceding segment.

15. The apparatus according to claim 9, wherein said path selection procedure comprises an expectation maximization algorithm.

16. The apparatus according to claim 9, wherein said path selection procedure features a hidden Markov model.

20. The apparatus of claim 9, wherein said plurality of features comprises at least one additional feature selected from the group consisting of spectral parameters and signal envelope parameters.

17. A photoplethysmograph system, comprising;

a photoplethysmograph sensor configured for generating a photoplethysmograph signal and an analysis apparatus which comprises an input unit for receiving the photoplethysmograph signal; and

a data processor configured for segmenting the photoplethysmograph signal into at least a plurality of temporal segments;

extracting from each temporal segment of said plurality of temporal segments a plurality of features including at least respiration parameters, thereby constructing a multidimensional feature space;

employing a path selection procedures to said multidimensional feature space to assign a respiration frequency or a range of respiration frequencies separately for each segment of said plurality of temporal segments, determining at least one sequence of respiration frequencies;

reconstructing a respiratory signal using said sequence, wherein each respiration frequency or range of respiration frequencies corresponds to a state of said feature space and wherein said path selection procedure calculates probabilities of transitions between states, and displaying said reconstructed respiratory signal.