Adaptive frequency domain filtering for improved non-invasive blood pressure estimation

• US 8,556,821 B2
• Filed: 02/20/2008
• Issued: 10/15/2013
• Est. Priority Date: 02/20/2008
• Status: Active Grant

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Hersh, Lawrence T., Kolluri, Sai, Donehoo, Robert F.
• Primary Examiner(s): Yang, N. C.
• Assistant Examiner(s): MOERSCHELL, RICHARD P
• Application Number:
  US12/034,143
• Publication Number:
  US 20090209868A1
• Time in Patent Office: 2,064 Days
• Field of Search: None
• US Class Current: 600/490
• CPC Class Codes: A61B 5/02225 using the oscillometric method
• CPC Class Codes: A61B 5/0225 the pressure being controlled by electric signals, e.g. derived from Korotkoff sounds
• CPC Class Codes: A61B 5/024 Detecting, measuring or recording pulse rate or heart rate A61B5/0205, A61B5/021 take precedence
• CPC Class Codes: A61B 5/7257 using Fourier transforms

Abstract

A system for processing oscillometric data from a plurality of pressure steps to determine the blood pressure of a patient as disclosed herein. A heart rate monitor connected to the patient acquires the patient'"'"'s heart rate. A time to frequency domain converter receives oscillometric data and converts the oscillometric data into the frequency domain. A harmonic frequency calculator is connected to the heart rate monitor and derives at least the heart rate fundamental frequency. A filter connected to the time to frequency domain converter and the harmonic frequency calculator that produces a filter frequency domain oscillometric signal. A reconstruction calculator receives the filtered frequency domain oscillometric signal and reconstructs a time domain oscillometric signal. A method of computing an oscillometric envelope for use in determining the blood pressure of a patient is also disclosed herein.

Claims

1. A method of computing an oscillometric envelope for use in determining the blood, pressure of a patient, the method comprising:

receiving, with a processing unit, an oscillometric signal from a blood pressure cuff connected to a patient;

acquiring a heart rate from the patient with a heart rate monitor separate from the blood pressure cuff, the heart rate monitor being a different modality from the blood pressure cuff;

computing the fundamental frequency of the heart rate with a fundamental frequency computer operating on the processing unit;

computing at least one harmonic frequency of the heart rate with a harmonic frequency computer operating on the processing unit;

converting the oscillometric signal to a frequency domain oscillometric signal with a time to frequency converter;

removing acquisition artifacts from the received oscillometric signal by;

filtering, with the processing unit, the frequency domain oscillometric signal using a band pass filter centered on the fundamental frequency of the heart rate acquired from the heart rate monitor of a different modality to produce a first filtered frequency domain signal;

filtering, with the processing unit, the frequency domain oscillometric signal using a band pass filter centered on the at least one harmonic frequency of the heart rate acquired from the heart rate monitor of a different modality to produce a second filtered frequency domain signal; and

reconstructing the oscillometric signal with an oscillometric waveform reconstructor operating on the processing unit from the first and second filtered frequency domain signals to produce a reconstructed oscillometric signal;

computing at least one oscillometric envelope data point from the reconstructed oscillometric signal with an oscillation size calculator operating on the processing unit;

computing an oscillometric envelope from the computed at least one oscillometric data point; and

determining at least one of mean arterial pressure, systolic pressure, and diastolic pressure from the oscillometric envelope.

2. The method of claim 1, wherein the step of converting the oscillometric signal to the frequency domain is performed using a fast Fourier transform (FFT).

3. The method of claim 1, wherein the step of computing at least one harmonic frequency of the heart rate includes computing the first harmonic frequency and the second harmonic frequency.

4. The method of claim 3, further comprising the step of filtering the frequency domain oscillometric signal with a pass band centered at the first harmonic frequency to produce a second filtered frequency domain signal;

and

filtering the frequency domain oscillometric signal with a pass band centered at the second harmonic frequency to produce a third filtered frequency domain signal.

5. The method of claim 1 further comprising the steps of:

comparing the heart rate of the patient to a predetermined threshold; and

if the heart rate of the patient is below the predetermined threshold, filtering the frequency domain oscillometric signal with a single band pass filter that comprises the fundamental frequency and the at least one harmonic frequency.

6. The method of claim 1 wherein the steps of filtering the frequency domain oscillometric signal are performed using a band pass filter of a bandwidth less than or equal to the fundamental frequency centered at the fundamental frequency of the heart rate and a band pass filter of a bandwidth less than or equal to the fundamental frequency centered at least one harmonic frequency.

7. The method of claim 1, wherein the received oscillometric signal is from a single pressure step of the blood pressure cuff signal.

8. The method of claim 7, further comprising repeating the method at a plurality of blood pressure cuff pressure steps to produce a plurality of computed envelope data points.

9. The method of claim 8, wherein the heart rate acquired from the patient with the heart rate monitor is continuously acquired, and further comprising:

calculating at least one new harmonic frequency from the continuously acquired heart rate;

filtering the frequency domain oscillometric signal with band pass filters comprising the at least one new harmonic frequency.

10. The method of claim 6, wherein the step of computing an oscillometric envelope data point further includes removing the first and last reconstructed oscillations and determining an oscillometric envelope data point from the remaining reconstructed oscillations.

11. A system for processing an oscillometric signal from a plurality of pressure steps of a blood pressure cuff to determine the blood pressure of a patient, the oscillometric signal having at least one artifact due to the acquisition of the oscillometric signal with the blood pressure cuff, the system comprising:

a heart rate monitor connected to the patient, wherein the heart rate monitor is a separate modality from the blood pressure cuff, the heart rate monitor acquires the patient'"'"'s heart rate at each pressure step;

a time-to-frequency domain converter that receives the oscillometric signal acquired at each pressure step and converts the oscillometric data into the frequency domain;

a fundamental frequency computer connected to the heart rate monitor, the fundamental frequency computer deriving a fundamental frequency of the heart rate at each pressure step;

a fundamental frequency filter connected to the time-to-frequency domain converter and the fundamental frequency computer, the fundamental frequency filter comprising a pass band centered at the fundamental frequency of the heart rate at each pressure step, the fundamental frequency filter producing a harmonic frequency filter frequency domain oscillometric signal;

a harmonic frequency computer connected to the heart rate monitor, the harmonic frequency computer deriving at least a harmonic frequency of the heart rate at each pressure step;

a harmonic frequency filter connected to the time-to-frequency domain converter and the harmonic frequency computer, the harmonic frequency filter comprising at least one pass band centered at the harmonic frequency of the heart rate at each pressure step, the harmonic frequency filter producing a harmonic frequency filtered frequency domain oscillometric signal;

a reconstruction calculator that receives the fundamental frequency filtered frequency domain oscillometric signal and the harmonic frequency filtered frequency domain oscillometric signal and reconstructs a time domain oscillometric signal without the at least one artifact from the received filtered frequency domain oscillometric signals;

an oscillometric envelope calculator that processes the time domain oscillometric signal at each pressure step to produce an oscillometric envelope from envelope data points corresponding to each pressure step; and

a blood pressure determination calculator connected to the oscillometric envelope calculator, the blood pressure determination calculator analyzes the oscillometric envelope to determine the blood pressure of the patient.

12. The system of claim 11 wherein the harmonic frequency calculator further derives at least the heart rate first and second harmonic frequencies, and the filter further comprises the first and second harmonic frequencies.

13. The system of claim 12 wherein the filter further comprises pass bands centered at the bean rate fundamental frequency and at least the first and second harmonic frequencies.

14. The system of claim 13 wherein the pass bands centered at the heart rate fundamental frequency and at least the first and second harmonic frequencies are of a bandwidth less than the fundamental frequency.

15. The system of claim 11 further comprising a DC removal calculator, the calculator receiving the oscillometric data and removing the DC component from the oscillometric data.

16. A method of determining a blood pressure of a patient, the method comprising:

operating a blood pressure cuff through a plurality of pressure steps;

acquiring oscillometric signals at each of the a plurality of pressure steps, the oscillometric signals having at least one artifact due to the acquisition of the oscillometric signals with the blood pressure cuff;

acquiring a heart rate of the patient at each pressure step with a heart rate monitor of a separate modality from the blood pressure cuff, each acquired heart rate associated to an oscillometric signal acquired at a same pressure step;

computing the fundamental frequency of each of the acquired heart rates with a fundamental frequency computer operating on a processing unit;

computing at least one harmonic frequency of each of the acquired heart rates with a harmonic frequency computer operating on the processing unit;

converting each of the oscillometric signals to a frequency domain oscillometric signal with a time to frequency converter;

removing acquisition artifacts from each oscillometric signal by;

filtering, with the processing unit, the frequency domain oscillometric signal with a band pass filter centered on the fundamental frequency of the associated heart rate to produce a first filtered frequency domain signal consisting of frequency domain oscillometric data centered at the fundamental frequency of the associated heart rate;

filtering, with the processing unit, the frequency domain oscillometric signal with at least one band pass filter centered on at least one harmonic frequency of the associated heart rate to produce a second filtered frequency domain signal consisting of frequency domain oscillometric data centered at the at least one harmonic frequency of the associated bean rate; and

reconstructing the oscillometric signal in the time domain from the first and second filtered frequency domain signals with an oscillometric waveform reconstructor operating on the processing unit to produce a reconstructed oscillometric signal;

computing at least one data point from each of the reconstructed oscillometric signals;

computing an oscillometric envelope from the computed data points; and

determining at least one of mean arterial pressure, systolic pressure, and diastolic pressure from the oscillometric envelope.

17. The method of claim 16, wherein the band pass filter centered on the fundamental frequency and the at least one band pass filter centered on at least one harmonic frequency have bandwidths less than or equal to the fundamental frequency.

18. The method of claim 16, further comprising processing the oscillometry signals acquired al each pressure step to remove a dc portion of the oscillometric signal.