Method, apparatus and system for removing motion artifacts from measurements of bodily parameters

• US 6,393,311 B1
• Filed: 10/01/1999
• Issued: 05/21/2002
• Est. Priority Date: 10/15/1998
• Status: Expired due to Fees

• Current Assignee: Philips Electronics North America Corporation (Koninklijke Philips N.V.)
• Original Assignee: NTC Technology, Inc. (Dixtal Medical, Inc.)
• Inventors: DelFavero, John R., Jaffe, Michael B., Allo, August J. Jr., Edgar, Reuben W. Jr., Martin, Jesus D.
• Primary Examiner(s): Winakur, Eric F.
• Assistant Examiner(s): KREMER, MATTHEW J
• Application Number:
  US09/410,991
• Time in Patent Office: 963 Days
• Field of Search: 600/309-311, 600/322-326, 600/330-331, 600/336, 600/473, 600/476, 356/39-41, 369/60.01
• US Class Current: 600/323
• CPC Class Codes: A61B 5/14551 for measuring blood gases
• CPC Class Codes: A61B 5/7207 of noise induced by motion artifacts

Abstract

A method for removing motion artifacts from devices for sensing bodily parameters and apparatus and system for effecting same. The method includes analyzing segments of measured data representing bodily parameters and possibly noise from motion artifacts. Each segment of measured data may correspond to a single light signal transmitted and detected after transmission or reflection through bodily tissue. Each data segment is frequency analyzed to determine dominant frequency components. The frequency component which represents at least one bodily parameter of interest is selected for further processing. The segment of data is subdivided into subsegments, each subsegment representing one heartbeat. The subsegments are used to calculate a modified average pulse as a candidate output pulse. The candidate output pulse is analyzed to determine whether it is a valid bodily parameter and, if yes, it is output for use in calculating the at least one bodily parameter of interest without any substantial noise degradation. The above method may be applied to red and infrared pulse oximetry signals prior to calculating pulsatile blood oxygen concentration. Apparatus and systems disclosed incorporate methods disclosed according to the invention.

Claims

1. A method of removing motion-induced noise artifacts from a single electrical signal representative of a pulse oximetry light signal, comprising:

receiving a segment of raw data spanning a plurality of heartbeats from said single electrical signal;

analyzing said segment of raw data for candidate frequencies, one of which said candidate frequencies may be representative of a valid plethysmographic pulse;

analyzing each of said candidate frequencies to determine a best frequency including narrow bandpass filtering said segment of raw data at each of said candidate frequencies;

outputting an average pulse signal computed from said segment of raw data and said best frequency; and

repeating the above steps with a new segment of raw data.

2. The method of claim 1, wherein said analyzing said segment of raw data for candidate frequencies includes computing a power spectrum of said segment of raw data and identifying n candidate frequencies associated with the n largest amplitude peaks in power spectrum which are not harmonics of each other.

3. The method of claim 2, wherein n equals two.

4. The method of claim 3, further including comparing said two candidate frequencies with a previous known valid frequency and replacing the smaller amplitude candidate frequency with said previous known valid frequency if neither of said two candidate frequencies is within about ±

0.1 Hertz of said previous known valid frequency.

5. The method of claim 2, further including filtering said segment of raw data to reduce spectral leakage prior to computing said power spectrum.

6. The method of claim 2, wherein said analyzing each of said candidate frequencies to determine a best frequency by selecting n candidate frequencies associated with the n largest amplitude peaks in power spectrum which are not harmonics of each other further comprises:

subsegmenting said bandpass filtered segment of raw data into individual heartbeat subsegments;

computing an average heartbeat pulse from said individual heartbeat subsegments;

evaluating said average heartbeat pulse against known plethysmographic pulse characteristics to obtain a quality measure for said average heartbeat pulse;

repeating the above steps for each of said n candidate frequencies; and

selecting a candidate frequency and average heartbeat pulse with highest quality measure to be said best frequency.

7. The method of claim 6, wherein said bandpass filter is a finite impulse response (FIR) filter.

8. The method of claim 7, wherein the coefficients of said FIR filter are adjusted to place the center frequency of said FIR filter at said candidate frequency.

9. The method of claim 6, wherein said computing an average heartbeat pulse from said individual heartbeat subsegments comprises:

calculating an average for each individual heartbeat subsegment;

aligning corresponding individual heartbeat subsegment points from beginning to end for each of said individual heartbeat subsegments; and

calculating a modified average pulse by eliminating all data points outside x standard deviations from said average for each individual heartbeat subsegment to obtain said average heartbeat pulse.

10. The method of claim 9, wherein x equals about two.

11. The method of claim 6, further comprising minimizing dispersion of saturation values, wherein said minimizing dispersion of saturation values includes adjusting the phase relationship of red and IR pulses and calculating a minimum dispersion of saturation values.

12. The method of claim 11, wherein adjusting the phase relationship of red and IR pulses includes calculating saturation values for a plurality of delays between said red and IR signals, wherein said plurality of delays are sequentially swept from about −

20 ms to about +20 ms in 1 ms increments.

13. The method of claim 11, wherein minimizing dispersion of saturation values includes:

generating a histogram of saturation values with bin sizes of about y percent calculated with delays;

locating a peak bin and its immediately surrounding z bins and calculating a measure of dispersion based on the percentage of saturation values which fall within said peak bin and its immediately surrounding z bins; and

selecting an optimum delay corresponding to measure of dispersion with highest percentage.

14. The method of claim 13, wherein y equals 0.25.

15. The method of claim 13, wherein z equals two.

16. A method of removing motion-induced noise artifacts from a measured pulse oximetry signal comprising:

acquiring a segment of pulse oximetry data from the measured pulse oximetry signal;

filtering said segment of pulse oximetry data;

performing frequency analysis on said filtered segment of pulse oximetry data to determine frequency components of said filtered segment of pulse oximetry data;

selecting a frequency component with a largest power as an initial candidate heart rate response;

subsegmenting said filtered segment of pulse oximetry data into a plurality of subsegments, wherein each subsegment represents an individual heartbeat;

averaging data values for each subsegment at identical points along said length of each subsegment to form an average subsegment;

determining whether said average subsegment represents a valid pulse oximetry signal;

outputting valid average subsegments;

updating said segment of pulse oximetry data;

measuring dispersion of saturation of latest pulse subsegment to determine whether said averaging step may be skipped; and

repeating the above steps.

17. A method of removing motion-induced noise artifacts from measured pulse oximetry signals prior to calculating functional pulsatile blood oxygen concentration, SpO₂, comprising:

acquiring a red data segment spanning a plurality of heartbeats from oldest to newest;

frequency analyzing said red data segment for candidate red frequency components, one of which may be representative of a valid plethysmographic pulse;

analyzing each of said candidate red frequency components to determine a best red frequency component including narrow bandpass filtering said segment of red data at each of said candidate red frequency components;

outputting an average red pulse signal computed from said red data segment and said best red frequency component;

repeating the above steps with a new red data segment; and

substantially simultaneously performing the above steps with respect to an infrared data segment.

18. A method of removing motion-induced noise artifacts from a single electrical signal representative of a pulse oximetry light signal comprising:

(a) acquiring a segment of raw data spanning a plurality of heartbeats from said single electrical signal;

(b) frequency analyzing said segment of raw data to determine two candidate frequencies with greatest amplitude, F₁ and F₂, which are not harmonics of each other;

(c) removing F₂ from said segment of raw data to obtain F₂ filtered data;

(d) subsegmenting said F₂ filtered data into a plurality of subsegments each containing a single heartbeat pulse;

(e) calculating a first average pulse signal based on averaging said plurality of F₂ filtered subsegments;

(f) removing F₁ from said segment of raw data to obtain F₁ filtered data;

(g) subsegmenting said F₁ filtered data into a plurality of subsegments each containing a single heartbeat pulse;

(h) calculating a second average pulse signal based on averaging said plurality of F₁ filtered subsegments;

(i) determining which of said first average pulse signal and said second average pulse signal represents a valid plethysmographic pulse;

(j) outputting said first average pulse signal or said second average pulse signal which represents a valid plethysmographic pulse; and

(k) repeating steps (b) through (j) with a new segment of raw data.

19. The method of claim 18, wherein after step (b), frequencies F₁ and F₂ are compared with a valid frequency, F₀, from a previous signal processing cycle and F₀ is substituted for F₂ if neither F₁ or F₂ are ±

0.1 Hertz of F₀.

20. The method of claim 18, wherein removing F₂ from said segment of raw data is accomplished by filtering said segment of raw data with a bandstop filter of a width about 0.5 Hertz centered at F₂.

21. A method of removing motion-induced noise artifacts from a measured pulse oximetry signal comprising:

acquiring a segment of pulse oximetry data from the measured pulse oximetry signal;

filtering said segment of pulse oximetry data;

performing frequency analysis on said filtered segment of pulse oximetry data to determine frequency components of said filtered segment of pulse oximetry data;

selecting a frequency component with a largest power as an initial candidate heart rate response;

narrow bandpass filtering at said selected frequency component;

subsegmenting said filtered segment of pulse oximetry data into a plurality of subsegments, wherein each subsegment represents an individual heartbeat;

averaging data values for each subsegment at identical points along said length of each subsegment to form an average subsegment;

determining whether said average subsegment represents a valid pulse oximetry signal;

selecting a next largest power frequency component, if one exists, and repeating said narrow bandpass filtering step, said subsegmenting step, said averaging step and said determining step;

outputting valid average subsegments;

updating said segment of pulse oximetry data; and

repeating the above steps.

22. The method of claim 21, further comprising measuring dispersion of saturation of latest pulse subsegment to determine whether said averaging step may be skipped.

23. The method of claim 21, further comprising minimizing dispersion of saturation values prior to said outputting valid average subsegments.

24. A circuit card for use in a pulse oximetry system to remove motion-induced noise artifacts from a measured pulse oximetry signal, said circuit card comprising:

a circuit board for mounting electronic circuitry and interfacing with the pulse oximetry system;

a processor mounted on said circuit board for processing at least one input signal according to instructions; and

memory for storing a computer program, wherein said memory is operably coupled to said processor, and wherein said computer program includes instructions for implementing a method of removing motion artifacts from said measured pulse oximetry signal, said method comprising;

receiving a segment of raw data spanning a plurality of heartbeats from said measured pulse oximetry signal;

analyzing said segment of raw data for candidate frequencies one of which said candidate frequencies may be representative of a valid plethysmographic pulse including narrow bandpass filtering said segment of raw data at each of said candidate frequencies;

analyzing each of said candidate frequencies to determine a best frequency based on highest quality measure;

outputting an average pulse signal computed from said segment of raw data and said best frequency; and

repeating the above steps with a new segment of raw data.

25. The circuit card of claim 24, wherein said processor is a digital signal processor.

26. A pulse oximetry system for removing motion-induced noise artifacts from measured pulse oximetry signals comprising an input device, an output device, and a motion artifact circuit card, wherein said motion artifact circuit card comprises:

a circuit board configured for coupling to said pulse oximetry system and communicating with other electronics within said pulse oximetry system;

a processor mounted on said circuit board for processing at least one input signal according to instructions; and

memory operably coupled to said processor for storing a computer program, wherein said computer program includes instructions for implementing a method of removing motion artifacts from said measured pulse oximetry signal, wherein said method comprises;

receiving a segment of raw data spanning a plurality of heartbeats from said measured pulse oximetry signal;

analyzing said segment of raw data for candidate frequencies one of which said candidate frequencies may be representative of a valid plethysmographic pulse including narrow bandpass filtering said segment of raw data at each of said candidate frequencies;

analyzing each of said candidate frequencies to determine a best frequency based on highest quality measure;

outputting an average pulse signal computed from said segment of raw data and said best frequency; and

repeating the above steps with a new segment of raw data.

27. The pulse oximetry system of claim 26, wherein said processor is a digital signal processor.

28. A pulse oximetry system for removing motion-induced noise artifacts from measured pulse oximetry signals comprising an input device, an output device, and a motion artifact circuitry, wherein said motion artifact circuitry includes:

a processor for processing at least one input signal according to instructions; and

memory operably coupled to said processor for storing a computer program, wherein said computer program includes instructions for implementing a method of removing motion artifacts from said measured pulse oximetry signal, wherein said method comprises;

receiving a segment of raw data spanning a plurality of heartbeats from said signal;

analyzing said segment of raw data for candidate frequencies one of which said candidate frequencies may be representative of a valid plethysmographic pulse including narrow bandpass filtering said segment of raw data at each of said candidate frequencies;

analyzing each of said candidate frequencies to determine a best frequency based on highest quality measure;

outputting an average pulse signal computed from said segment of raw data and said best frequency; and

repeating the above steps with a new segment of raw data.

29. The pulse oximetry system of claim 28, wherein said processor is a digital signal processor.

30. A method of removing motion-induced noise artifacts from a single electrical signal representative of a pulse oximetry light signal, comprising:

receiving a segment of raw data spanning a plurality of heartbeats from said single electrical signal;

analyzing said segment of raw data for candidate frequencies, one of which said candidate frequencies may be representative of a valid plethysmographic pulse, said analyzing including computing a power spectrum of said segment of raw data and identifying n candidate frequencies associated with n largest amplitude peaks in power spectrum which are not harmonics of each other, further comprising;

bandpass filtering said segment of raw data;

subsegmenting said bandpass filtered segment of raw data into heartbeat subsegments, comprising;

beginning each heartbeat subsegment at one quarter pulse width before a diastolic peak; and

ending each heartbeat subsegment at one quarter pulse width after an immediately following diastolic peak, wherein each heartbeat subsegment overlaps with both a preceding heartbeat subsegment and a succeeding heartbeat subsegment, and each plethysmographic pulse is centered within each heartbeat subsegment;

computing an average heartbeat pulse from said heartbeat subsegments;

evaluating said average heartbeat pulse against known plethysmographic pulse characteristics to obtain a quality measure for said average heartbeat pulse;

repeating the above steps for each of said n candidate frequencies; and

selecting a candidate frequency and average heartbeat pulse with highest quality measure to be said best frequency;

analyzing each of said candidate frequencies to determine a best frequency;

outputting an average pulse signal computed from said segment of raw data and said best frequency; and

repeating the above steps with a new segment of raw data.

31. A method of removing motion-induced noise artifacts from a single electrical signal representative of a pulse oximetry light signal, comprising:

receiving a segment of raw data spanning a plurality of heartbeats from said single electrical signal;

analyzing said segment of raw data for candidate frequencies, one of which said candidate frequencies may be representative of a valid plethysmographic pulse, said analyzing including computing a power spectrum of said segment of raw data and identifying n candidate frequencies associated with n largest amplitude peaks in power spectrum which are not harmonics of each other, further comprising;

bandpass filtering said segment of raw data;

subsegmenting said bandpass filtered segment of raw data into individual heartbeat subsegments;

computing an average heartbeat pulse from said individual heartbeat subsegments;

evaluating said average heartbeat pulse against known plethysmographic pulse characteristics to obtain a quality measure for said average heartbeat pulse;

repeating the above steps for each of said n candidate frequencies;

selecting a candidate frequency and average heartbeat pulse with highest quality measure to be said best frequency; and

minimizing dispersion of saturation values including adjusting the phase relationship of red and IR pulses including calculating saturation values for a plurality of delays between said red and IR signals, wherein said plurality of delays are sequentially swept from about −

20 ms to about +20 ms in 1 ms increments and calculating a minimum dispersion of saturation values;

analyzing each of said candidate frequencies to determine a best frequency;

outputting an average pulse signal computed from said segment of raw data and said best frequency; and

repeating the above steps with a new segment of raw data.

32. A method of removing motion-induced noise artifacts from a single electrical signal representative of a pulse oximetry light signal, comprising:

receiving a segment of raw data spanning a plurality of heartbeats from said single electrical signal;

analyzing said segment of raw data for candidate frequencies, one of which said candidate frequencies may be representative of a valid plethysmographic pulse, said analyzing including computing a power spectrum of said segment of raw data and identifying n candidate frequencies associated with n largest amplitude peaks in power spectrum which are not harmonics of each other, further comprising;

bandpass filtering said segment of raw data;

subsegmenting said bandpass filtered segment of raw data into individual heartbeat subsegments;

computing an average heartbeat pulse from said individual heartbeat subsegments;

evaluating said average heartbeat pulse against known plethysmographic pulse characteristics to obtain a quality measure for said average heartbeat pulse;

repeating the above steps for each of said n candidate frequencies;

selecting a candidate frequency and average heartbeat pulse with highest quality measure to be said best frequency; and

minimizing dispersion of saturation values including adjusting the phase relationship of red and IR pulses and calculating a minimum dispersion of saturation values including;

generating a histogram of saturation values with bin sizes of about y percent calculated with delays;

locating a peak bin and its immediately surrounding z bins and calculating a measure of dispersion based on the percentage of saturation values which fall within said peak bin and it immediately surrounding z bins; and

selecting an optimum delay corresponding to measure of dispersion with highest percentage;

analyzing each of said candidate frequencies to determine a best frequency;

outputting an average pulse signal computed from said segment of raw data and said best frequency; and

repeating the above steps with a new segment of raw data.

33. The method of claim 32, wherein y equals 0.25.

34. The method of claim 32, wherein z equals two.