 |
Description  |
|
|
FIELD OF THE INVENTION
This invention relates to apparatus and methods for measuring the
constituents of blood and, more particularly, apparatus for noninvasive
determination of constituent concentrations utilizing light wave
absorption measurements and methods for processing signals generated by
such measurements.
BACKGROUND OF THE INVENTION
One known blood constituent measuring device is a blood oximeter which is a
photoelectric photometer utilized for measurement of the fraction of
hemoglobin in blood which is in the form of oxygenated Hb, that fraction
is normally expressed in a percentage referred to as saturated oxygen
content. The basic principles of blood oximetry based on these techniques
are discussed in an article entitled .cent.Oximetry" by Earl H. Wood, et.
al., appearing at pages 416-455, Medical Physics, vol. 2, O. Glasser, Ed.,
Year Book Medical, Chicago, Ill (1960).
A number of oximetry devices and methods utilizing light in the red and
infrared regions are shown in U.S. Pat. Nos. 3,638,640; 3,647,299;
3,704,706; 3,804,539; 3,998,550; 4,086,915 and 4,167,331.
Many previous oximetry devices use complex logarithmic functions to
determine oxygen saturation of the blood as is shown, for example, in U.S.
Pat. Nos. 3,638,140; 3,804,539; 3,998,550; and 4,167,331. The apparatus
disclosed in U.S. Pat. No. 4,167,331 employs a digital microprocessor in
conjunction with analog logarithmic amplifiers to calculate blood oxygen
saturation. Other devices have utilized three frequencies, with
synchronous detection, peak detection and ratio circuits, for example,
that shown in U.S. Pat. No. 3,647,299.
Finally, there are blood oxygen saturation measurement devices employing
digital techniques including computer microprocessors for calculating
pulse rate, oxygen saturation and confidence factors for those values (see
European Patent Application Publication No. 0104771) and one such device
further includes normalization of the DC signal components from each light
source and an AC modulated test signal for testing the device (see U.S.
Pat. No. 4,407,290).
Nevertheless, even those devices employing digital techniques have not
proven to be completely satisfactory in delivering reliable, blood
constituent measurements particularly in the operating room environment.
Therefore, a need exists for a compact, highly noise resistant, accurate
and reliable blood constituent measurement apparatus and method.
SUMMARY OF THE INVENTION
This invention provides an apparatus and method for measuring variations in
the opacity of blood due to blood volume changes related to Traube-Herring
waves, respiration, heart pulsation, and hydrostatic pressure. Data
relating to opacity variations is obtained by passing light of known
wavelengths through blood containing tissue such as a finger or earlobe.
For the purpose of determining the saturated oxygen content of the blood,
the cyclic variations due only to the heart pulsations are extracted from
the measurements and are corrected for the effects of the Traube-Herring
waves. Calculation of heart rate and saturated oxygen content is performed
using a programmed microprocessor. The same microprocessor also
dynamically controls the data measurement function in a manner that
optimizes the measurements under a variety of changing conditions.
It is therefore a principle object of the invention to provide a blood
constituent measurement method and apparatus that is dynamically adaptive
to the measurement environment.
It is a further object of the invention to continuously monitor and
evaluate the quality of the measured values and display the result of such
evaluation as a signal confidence level. It is also an object of the
invention to utilize the signal quality evaluation to adjust the
measurement parameters and signal analysis so as to obtain and maintain
the highest signal confidence level.
It is a further object of the invention to measure and evaluate the
presence of non-measurement signal noise and to adapt the measurement
sequence so as to minimize interference of data acquisition by such noise.
It is still a further object of the invention to provide pulse rate data
and blood oxygen saturation data along with a continuous indication of the
confidence level in the accuracy of that data.
BRIEF DESCRIPTION OF THE DRAWINGS
These and other objects and advantages of the present invention are
apparent from a consideration of the entire specification and the
following drawings in which:
FIG. 1 is a Schematic Block Diagram of an apparatus of the invention.
FIGS. 2A, B and C are a more detailed schematic diagram of components of
the LED driver circuits shown in FIG. 1.
FIGS. 3A, B and C are a more detailed schematic diagram of components of
the channel separation and compensation for additive noise circuits shown
in FIG. 1.
FIGS. 4A, B and C are a more detailed schematic diagram of the offset
voltage circuitry shown in FIG. 1. FIGS. 5A, B, C and D are a more
detailed schematic diagram of the sequence control circuitry shown in FIG.
1.
FIG. 6 is a timing diagram of a data collection sequence generated by the
apparatus of FIG. 1.
FIG. 7 is a timing diagram of the noise detection sequence caused by the
noise detection and tracking control modules shown in FIG. 1.
DESCRIPTION
Referring now to FIG. 1, there is shown in functional block diagram form an
apparatus of the invention. The apparatus performs two functions
simultaneously under the control of a programmed microprocessor (not
shown); (1) control of signal generation; tracking the desired signal and
measurement of the data provided by the signal (2) calculation of
saturated blood oxygen level and pulse rate from the measured data and
housekeeping functions associated with data display and operator
interaction.
Broadly common to both of these functions is the circuitry to measure the
raw data. This consists of the sequence control 47, LED control circuit
43, LED's 4, 5, a sample area causing pulsating changes in attenuation of
the light 6, a signal detection method 44, channel separation and
compensation for additive noise circuitry 11, analog signal processing
circuits 45, and circuits to convert the signals to digital form 46.
The programmed microprocessor stores the data in memories 30 and 32,
provides additional filtering to eliminate the noise associated with
analog to digital conversion, 31 and 33, and detects inflection points to
determine the minimum and maximum values of the plethysmographic waveform
29.
The microprocessor is also programmed to provide the tracking control 24
which is used to analyze the data to determine if the plethysmographic
waveform is of sufficient amplitude and if the mean level of the signal at
the input of the A/D convertor 19 is within an acceptable range. The
tracking control program is also responsible for maintaining an acceptable
amplitude for the pulse information by controlling the intensity of the
LED's and adjusting the offset, 22 and 23, to maintain the signals within
the acceptable range of the subsequent digital conversion circuits. These
tasks are performed separately on each channel. The data necessary to make
these decisions comes from the filtered digital signals from 31 and 33 as
well as a separate determination of the amplitude of the plethysmographic
waveform accomplished by the MIN/MAX detector 29. Tracking control and
associated data acquisition is performed as a foreground task by the
microprocessor.
The second function of the microprocessor is principally calculation and
display of pulse rate 28, 37, 40 and calculation and display of saturated
blood oxygen level (SaO.sub.2) 36, 39, 42. A key part of the calculations
process is determination and display of signal quality 38, 41. This
function and the associated operator interaction are performed by the
microprocessor as background tasks.
With the foregoing as an overview, the function of each block in FIG. 1
will be briefly described with reference to specific devices shown in
FIGS. 2 and 3.
The light emitting diode (LED) control 43, controls the intensities of the
individual light emitting diodes 4, 5, in response to signals generated by
the Tracking Control 24. Each LED intensity is adjusted independently
based on the data analyzed and received from that channel.
Referring to FIG. 1 there is shown intensity controls 1, 2 for LED's 4, 5
consisting mainly of a 12 bit digital to analog convertors, 100, 104,
shown in FIG. 2A, which use a reference voltage 110 and 114 that has been
temperature compensated to match the characteristics of the LED current
drivers 3, particularly at low currents, and a circuit 50 for
communicating with the microprocessor.
Returning to FIG. 1, the LED driver circuitry 3 assures predictable linear
control over the maximum LED intensity range and minimizes the number of
wires connecting the LED's to the instrument. A bipolar current driver,
shown in FIG. 2B, for each data channel 200, 205, 300, 305 is used with
associated components of a watchdog timer 118, 122 and 126 to insure that
the LED's cannot remain on for extended time intervals at high currents,
even in the event of a tracking or sequence control module failure. This
is necessary to guarantee that the patient will not be burned and that the
LED's will not be destroyed due to a momentary failure in the controlling
software.
The light emitting diodes 4, 5, are composed of two RED (660 nm) dies wired
in series and two INFRARED (940 nm) dies wired in series. The two diode
pairs are mounted and wired in parallel but with opposite polarities on
the same substrate and covered with an optically clear epoxy.
The received signals from the LED's as well as from any interfering sources
are received by the photo detector 7, which is a standard PIN diode
selected to be sensitive to light over the wavelengths from 600 nm to 1000
nm. The signal is amplified by a current to voltage convertor 8 and
transmitted over a length of cable 9 to be received by an amplifier 10.
The current to voltage convertor 8, is located in the cable assembly near
the photo detector and is constructed as a hybrid to minimize the size and
susceptibility to electromagnetic interference. The configuration is
optimized to the frequency and gain requirements of the overall apparatus.
The buffer amplifier 10, is used to amplify the signal by a factor of 20.
The configuration of this amplifier is also optimized for frequency and
noise requirements of the overall apparatus.
Channel separation 11 is performed on the amplified signal. The signals
received by the amplifier 10 consist of a sequence of pulses determined by
sequence selection and control circuitry, 47. The data collection sequence
is ambient, ambient+RED, and ambient+INFRARED. This serially multiplexed
data is separated by analog switches shown in FIGS. 3A and C, 130 and 134
and stored respectively in sample and hold elements 400, 210 and 310.
These sample and hold circuits are unique in that the time constant for
the circuit is many times greater than the sample period. In operation,
the sampling circuit causes the signal to charge a capacitor through a
resistive element for a predetermined short time period. The time constant
of the R-C combination is many times longer than the sample period,
therefor, many samples are required to charge the capacitor to a level
representative of the original signal. The effect is to provide a high
frequency filter near the sampled data frequency without distorting the
desired data. Asymmetrical high frequency noise results in a constant
residual value which will appear on all of the sampled and held signals
and thus be cancelled in the subsequent subtraction operation. This unique
method allows the three sets of data to be sampled nearly simultaneously
while also filtering the data over a longer interval to remove any high
frequency interference received by the photosensor. From the sample and
hold elements the ambient data is subtracted from the INFRARED+ambient
data by 215 and from the RED+ambient data by 315 both shown in FIG. 3B.
Returning to FIG. 1, after the data has been separated into two channels
and compensated for additive noise at 11, it is further processed by
elements designated analog signal processing circuits 45. Each channel is
individually amplified by a factor of 5 by the difference amplifiers 12,
13. The RED channel data and INFRARED channel data are filtered by analog
low pass filters 14, 15 to remove additional noise. These filters are of
standard design and have been tuned for a cutoff frequency of
approximately 10 Hz. This frequency was selected since the maximum heart
rate expected to be measured will be 4 Hz.
Offset voltages are generated for each channel in offset voltage sources
22, 23. Referring to FIG. 4 A, a precisely known voltage is generated by
the digital to analog convertors 220 and 320 for each of the data
channels. As shown in FIGS. 4A, B and C, the references for these voltages
are developed by 138 and 225 and 325 for stability and to minimize
crosstalk and noise between channels. The offset voltages are buffered by
230 and 330 and are made available to the difference amplifiers 235 and
335 as well as the multiplexer 18.
Returning to FIG. 1, the tracking control program in the microprocessor
relatively corrects for the voltage present on each channel when no data
is expected (both LED sources 4 and 5 and the offset voltages from 22 and
23 are turned off) by measuring it through common amplifiers 16, 17 and a
common analog to digital convertor 19.
Up to this point, all of the signal processing has been accomplished by
analog circuit techniques. Analog to digital conversion 46 is accomplished
by directing the data to a common analog to digital convertor 19 by means
of a multiplexer 18. Analog to digital conversion permits all subsequent
data processing to be performed by the microprocessor.
The RED channel data and INFRARED channel data are directed by Multiplexer
18 to a common analog to digital convertor 19. In addition, the direct
offset voltages are preset relative to Ref. 1 and Ref. 2 in the same A/D
convertor, which permits calibration of the offset voltages relative to
the same references used to measure the data.
The operations of sequentially turning on the multiple diodes, 4, 5;
receiving signals from the light emitted by the diodes and the ambient
light; separating, sampling and holding this data 11; multiplexing the
data 18; and converting the data by a common A/D convertor 19 for
subsequent digital processing are all controlled by a sequence control
circuit 20. Up to 4 unique sequences are provided to optimize the
instrument for the particular measurement environment. Referring to FIG.
5C the sequence to be run is selected by sequence selection 21 from those
stored in a Read Only Memory, 52. Upon initialization of the instrument,
the counter 54 will increment through 1024 states for each sequence at a
rate controlled by a clock synchronous with the microprocessor. At each
counter increment the ROM 52 will yield a new pattern on 8 digital lines
to be latched by 56. The outputs from latch 56 are used by sequence
control 20 to control the operations of the LED drivers 3, sample and hold
circuits 11, multiplexer 18, and A/D convertor 19. The end of the sequence
is decoded by decoder 58 shown in FIG. 5D and an appropriate signal is
sent to the microprocessor to indicate that the operation has been
completed. Referring to FIGS. 5B and C, the selection of the desired
control sequence is accomplished by selecting the appropriate starting
address in the Read Only Memory 54 from the digital latch 60. The
conditions for sequence selection and the purpose of the various sequences
are discussed hereinbelow.
Returning to FIG. 1, the data from the A/D convertor 19 is separated and
stored in memory buffers 30 and 32. This data is additionally filtered by
digital lowpass filters 31, 32 to remove any noise created in the A/D
conversion process.
At the onset of alarm conditions and at any time when the tracking control
module determines that the instrument cannot track and obtain useful data
from the subject, a test for the existence of outside interference is made
by the Noise Detection module 27. As is explained more fully hereinbelow,
this module selects a sequence in sequence selector 21 to cause the
instrument to measure the amount of interference being received.
The MIN/MAX detector 29 samples the data points which are compared with
data stored previously to determine if the data in question is a maximum
value or a minimum value for that pulse cycle. When the MAX or MIN values
are determined, they are stored for each channel in 34. The actual
determination of MIN and MAX values is made on the INFRARED channel data
only and the corresponding RED channel data points are stored. This
operation assures that the subsequent calculations are made on related
data even if they are not exactly at the maximum and minimum levels of the
data received on the RED channel.
The oscillator 25 used to drive and synchronize the system is based on a
crystal controlled clock that is an integral part of the computer.
The sample rate 26 of the data collection is based on the system oscillator
and can be adjusted by the sequence control module of the microprocessor
program.
The pulse rate module 28 analyses the information from the MIN/MAX detector
29 and scales it by the current sample rate to develop a continuously
updated value representative of the subjects pulse rate.
Correction for motion artifact module 35 determines if an interference has
occurred due to motion of the subject. This interference usually occurs as
low frequency noise which is noise added to the data at a frequency lower
than the desired data and is defined herein as less than 20 Hz. Errors
caused by Traube-Herring waves are also in this frequency range and are
also corrected for by this module.
Suppression of this type of noise is accomplished by comparing the current
maximum value of the photoplethysmographic waveform for the infrared
incident radiation with that of the proceeding waveform. If the relative
level of the photoplethysmographic waveform maximum values are not with in
a predetermined value, which is a predetermined percentage of the peak to
peak amplitude of the first waveform, than the data calculated from the
second waveform is not used.
The minimum and maximum values, including the offset value 22 and 23
subtracted from the data are used to calculate the value of the saturated
oxygen in the subjects blood by the calculate SaO.sub.2 module 36. The
equation used is as follows:
RR=[log(REDmax/REDmin)]/[log(INFRAREDmax/INFRAREDmin)]
SaO.sub.2 A +B(RR)+C(RR).sup.2
Where the constants "A" , "B" & "C" are determined from experimental data
and are dependent upon the actual wavelengths of the LED's. The values of
these constants are stored in Read Only Memory and are accessed by the
microprocessor in performing these calculations.
The value of the SaO.sub.2 resulting from calculation module 36 is filtered
by a lowpass adaptive filter 39 with a cutoff frequency selected to be
slightly greater than the most rapid change expected in the saturated
oxygen of a human subject. Similarly, the value of the pulse rate from 28
is filtered by pulse adaptive filter 37 with a cutoff frequency slightly
higher than the most rapid change expected in the pulse rate of a human
subject. In addition each new data point is compared with the filtered
value to determine if the data is within reason. If a data point (Dn) is
determined to be outside of predetermined allowable limits, relative to
previous data, the filtered value is not modified by the new data,
however, the new data (Dn) is not discarded but is stored with a "flag" ,
to indicate the direction of change for future reference. The next data
point (Dn+1) is compared to the filtered data and again a determination is
made. If the data (Dn+1) is within acceptable limits then the previously
rejected data (Dn) is thrown out and the process continues. However, if
the second data (Dn+1) is also outside the allowable limits it is compared
to the previously questionable data (Dn). If the data are similar, i.e.,
within the same predetermined limits of each other, then the second data
is stored. The process continues for a third time and if at that time the
data is in the same direction as the two previous data points (Dn and
Dn+1) it is assumed that the last three data points are to be considered
as good data and the filter is biased to adjust its final value to be
equal to the average of the last three data points.
The final value for SaO.sub.2 is made available on a digital display 42
located on the front panel of the instrument.
Based on data from adaptive filters 37, 39, determinations are made by
signal quality analysis module 38 to form a judgment on the quality, or
accuracy, of the displayed data. Both the pulse rate and the calculated
SaO.sub.2 are examined for consistency and how those values deviate from
the predetermined norm causes a quality value to be assigned.
The results of the quality analysis module 38 are displayed by signal
quality indicator 41 as either a .-+.*" , "x" or a "+" depending on the
assigned worth of the displayed data This quality indicator is based on
the evaluated consistency of the pulse rate and SaO.sub.2 values. If the
SaO.sub.2 value changes by more than 3% from the average a predetermined
number of times or the pulse changes by more than 50 sample units a
predetermined number of times, the quality indicator will drop
accordingly. In each case, the change causes a variable called the change
count to receive 1 point. If the values stay within range, the change
count loses 1 point. If the value is out of range but is in the same
direction as the previous value, no change is made to the count. Normally,
both the pulse and SaO.sub.2 change counters start with 4 points. The two
counters are subtracted from the value 9 to arrive at a confidence value.
For example, on start up the confidence is 9-4-4=1. The lowest confidence
value possible is 0 (values less than 0 are set to 0), while the highest
is 9. The confidence levels are indicated on the meter display by changing
the appearance of a display "needle" . A "+" is used to indicate
confidence levels of 0 to 3, an "x" to indicate levels of 4 to 7, and an
"*" to show high confidence of 8 or 9. The displayed symbol is caused to
move horizontally as a representation of the raw data from 31 and 33. In
this manner the heart beat of the subject is displayed as a rhythmic
motion from side to side of an indicator that is representative of the
quality of the displayed data.
Referring again to FIG. 1, it can be seen that all of the modules discussed
above are directly or indirectly controlled by tracking control module 24.
It is responsible for maintaining an acceptable amplitude for the pulse
information by controlling the intensity of the LED's and adjusting the
offset, 22 and 23, to maintain the signals within the acceptable range of
the subsequent circuits. These tasks are continuously performed separately
on each channel. The data necessary to make these decisions comes from the
filtered signals 31 and 33 as well as a separate determination of the
amplitude of the plethysmographic waveform accomplished by the MIN/MAX
detector 29. Specifically, the tracking control monitors the signals of
both light sources independently to optimize the signal at each incident
wavelength. This tracking of the signals is composed of two primary tasks:
(1) to maintain a photoplethysmographic amplitude within predetermined
limits, and (2) to adjust for signal offsets 16, 17, 22, 23 due to the
variations in the transparency of the tissue sample and to adjust the
amount of signal offset required to maintain the signals within the range
of A/D converter 19.
Most significantly, the control flexibility provided by the tracking
control module 24 interacting with the noise detection module 27 provides
the apparatus of the invention with its unique capability of obtaining
reliable data in the presence of the normally interfering noise found in
the environments where such instruments are typically operated.
Initially, the apparatus is operated with the LED's 4 and 5 turned off. The
signal received is measured noise that is processed in the same manner as
measured data therefore a determination can be made as to the effect such
noise would have on the data. If the noise is of such a magnitude or if it
exceeds preset limits the apparatus turns off the data display and audibly
and visually indicates that data cannot be gathered due to interfering
noise.
The foregoing noise detection process is also performed periodically during
the normal data collection process to indicate the presence of excessive
noise and assure that reliable data is being measured. In addition,
whenever the tracking control module 24 detects an abrupt drop in the
results of the quality indicator calculation or any rapid change in
measured physiological parameters a noise detection sequence is initiated.
As can be readily seen from the foregoing discussion the control, and data
analysis functions are performed by a programmed microprocessor such as
the 8086 microprocessor manufactured by the Intel Corporation. A series of
program modules written for use with a microprocessor, such as the Intel
8086, running under a real time operating system, such as the RMX
operating system sold by the Intel Corporation, provide all of the
microprocessor controlled functions described in respect of FIG. 1. In
particular it is the unique control of data generation by using analysis
of the acquired data that provides the instrument of the invention with
its surprisingly superior performance in a wide variety of measurement
environments.
Referring now to FIG. 6 there is shown an example of the Data Collection
Sequence waveforms generated by the microprocessor. As can be seen a
sample sequence time t.sub.s (456.5 .mu.sec) occurs within an overall
cycle time T (3665 .mu.sec). During data collection the sequence
controller causes multiple cycles of the "ambient-red-infrared" and
ambient-infrared-red" sequence with the result that the time interval
between measurement of ambient conditions and red or infrared light
measurements are the same, on average, as stored in the averaging sample
and hold circuits of FIG. 1 block 11.
Referring to FIG. 7 there is shown an example of the Noise Detection
Sequence waveforms generated by the microprocessor. The sequence selected
for measuring the amount of interfering noise is identical to that used in
collecting the data, however, the LED's are never turned on. If there is
no interfering noise the data available to the A/D converter will be close
to zero in value.
* * * * *
|
| |
