Method and Apparatus for Detecting Optical Pulses

This invention relates to non-invasive pulse oximetry and specifically to an improved method and apparatus for photoelectric determination of blood constituents. A 123 page computer program is appended as a part of this application.

Non-invasive photoelectric pulse oximetry has been previously described in U.S. Pat. No. 4,407,290, U.S. Pat. No. 4,266,554, U.S. Pat. No. 4,086,915, U.S. Pat. No. 3,998,550, U.S. Pat. No. 3,704,706, European patent application Ser. No. 102,816 published Mar. 13, 1984, European patent application No. 104,772 published Apr. 4, 1984, and European patent application No. 104,771 published Apr. 4, 1984. Pulse oximeters are commercially available from Nellcor Incorporated, Hayward, Calif., and are known as, for example, Pulse Oximeter Model N-100.

Pulse oximeters typically measure and display various blood flow characteristics including but not limited to blood oxygen saturation of hemoglobin in arterial blood, volume of individual blood pulsations supplying the flesh, and the rate of blood pulsations corresponding to each heartbeat of the patient. The oximeters pass light through human or animal body tissue where blood perfuses the tissue such as a finger, an ear, the nasal septum or the scalp, and photoelectrically sense the absorption of light in the tissue. The amount of light absorbed is then used to calculate the amount of blood constituent being measured.

The light passed through the tissue is selected to be of one or more wavelengths that is absorbed by the blood in an amount representative of the amount of the blood constituent present in the blood. The amount of transmitted light passed through the tissue will vary in accordance with the changing amount of blood constituent in the tissue and the related light absorption. For example, the Nellcor N-100 Pulse Oximeter measures oxygen saturation of hemoglobin using two light emitting diodes ("LED's"), one having a discrete frequency of about 660 nanometers in the red light range and the other having a discrete frequency of about 925 nanometers in the infrared range. The two LED's are illuminated alternately with a four-state clock so that the incident light will pass through a fingertip and the detected or transmitted light will be detected by a single photodetector. The clock uses a high strobing rate, e.g., two thousand cycles per second, to be easily distinguished from other light sources. The photodetector current changes in response to both red and infrared transmitted light, in sequence, and is then amplified and separated by a two-channel synchronous detector--one channel for processing the red light waveform and the other channel for processing the infrared light waveform. The separated signals are filtered to remove the strobing frequency, electrical noise, and ambient noise and then digitized by an analog to digital converter ("ADC"). As used herein, incident light or transmitted light refers to light generated by the LED or other light source, as distinguished from ambient or environmental light.

The light source intensity may be adjusted to accommodate variations among patients' skin color, flesh thickness, hair, blood, and other variants. The light transmitted is thus modulated by the variants, particularly the arterial blood pulse or pulsatile component, and is referred to as the optical signal. The digital representation of the optical signal is referred to as the digital optical signal. The portion of the digital optical signal that refers to the pulsatile component is labeled the optical pulse.

The digital optical signal is processed by the microprocessor of the Nellcor N-100 Pulse Oximeter in order to identify individual optical pulses and to compute the oxygen saturation from the ratio of maximum and minimum pulse levels as seen by the red wavelength compared to the pulse seen by the infrared wavelength.

Several alternate methods of processing and interpreting optical signal data have been disclosed in the patents and references cited above.

A problem with non-invasive pulse oximeters is that the optically derived pulse rate may be subject to irregular variants that interfere with the detection of the blood flow characteristics including but not limited to motion artifact. Motion artifact is caused by the patient's muscle movement proximate to the oximeter sensor, for example, the patient's finger, ear or other body part to which the oximeter sensor is attached, and may cause spurious pulses that are similar to pulses caused by arterial blood flow. These spurious pulses, in turn, may cause the oximeter to process the artifact waveform and provide erroneous data. This problem is particularly significant with infants, fetuses, or patients that do not remain still during monitoring.

A second problem exists in circumstances where the patient is in poor condition and the pulse strength is very weak. In continuously processing the optical data, it can be difficult to separate the true pulsatile component from artifact pulses and noise because of a low signal to noise ratio. Inability to reliably detect the pulsatile component in the optical signal may result in a lack of the information needed to calculate blood constituents.

It is well known that electrical heart activity occurs simultaneously with the heartbeat and can be monitored externally and characterized by the electrocardiogram ("EKG") waveform. The EKG waveform, as is known to one skilled in the art, comprises a complex waveform having several components that correspond to electrical heart activity. The QRS component relates to ventricular heart contraction. The R wave portion of the QRS component is typically the steepest wave therein, having the largest amplitude and slope, and may be used for indicating the onset of cardiovascular activity. The arterial blood pulse flows mechanically and its appearance in any part of the body typically follows the R wave of the electrical heart activity by a determinable period of time. See, e.g., Goodlin et al., "Systolic Time Intervals in the Fetus and Neonate", Obstetrics and Gynecology, Vol. 39, No. 2, February 1972, where it is shown that the scalp pulse of fetuses lag behind the EKG "R" wave by 0.03-0.04 second, and U.S. Pat. No. 3,734,086.

It is therefore an object of this invention to provide an improved method and apparatus for detecting the pulsatile component of the optical signal and measuring the amount of blood constituent and the pulse rate by incorporating the patient's heart activity, preferably detected electrically in the form of an EKG waveform, into the oximeter operation and thereby solve problems caused by motion artifact and low signal to noise ratio, as well as simplify and improve the operation of oximeters.

Another object of this invention is to have the oximeter analyze only those digital optical signals occurring during a period of time when the optical pulses are expected to be found and use information from that portion of the signal to calculate the amount of blood constituent. This increases the likelihood that the oximeter will process only optical waveforms that contain the pulsatile component of arterial blood, and will not process spurious pulses.

Another object of the invention is to provide for using pulse oximeters to monitor patients having irregular heartbeats by using the EKG information, particularly the R wave component, to determine when an arterial pulse is likely to occur and processing the digital optical signal waveform during that time period to make the desired measurement.

A further object of this invention is to cross correlate the pulse rate information determined by the oximeter from the digital optical signal with the heart rate determined from the EKG. The cross correlation function will allow measurement of the time relationship between the EKG and the optical pulse and is particularly advantageous when the optical signal may be weak and in the delivery room where fetal heart rate is an important and commonly monitored vital sign.

A further object of this invention is to provide for redundant measurement of the heart rate from both the optical signal and the EKG to continuously monitor the patient even if one of the signals were to be lost.

A further object of this invention is to provide a polarity compensation circuit for use with EKG detection so that the polarity of the EKG waveform can be made uniform, upgoing or downgoing, without having to adjust the leads.

This invention increases the accuracy and reliability of pulse oximeters used during surgery, life threatening medical situations, and childbirth, by measuring the patient's heart activity and correlating it with the patient's blood flow to more accurately calculate and measure vital information such as oxygen saturation and pulse rate. In one embodiment the correlation comprises using auto- and cross correlation techniques to enhance periodic information contained in each individual waveform as well as determine the time relationship of one waveform to another. In the preferred embodiment, the method comprises correlating the occurrence of cardiovascular activity with the detection of arterial pulses by measuring an EKG signal, detecting the occurrence of the R wave portion of the EKG signal, determining the time delay by which an optical pulse follows the R wave, and using the determined time delay between an R wave and the following optical blood pulse so as to evaluate arterial blood flow only when it is likely to present a true blood pulse for waveform analysis. The method also includes determining the heart rate of the patient based on the EKG signal, the optical pulse, or both.

In a preferred embodiment, the method and apparatus comprises an improvement in the use of a Model N-100 Pulse Oximeter (herein "N-100 oximeter") manufactured and sold by Nellcor Incorporated, Hayward, Calif. The improved method provides an oximeter with an additional parameter enabling the oximeter to better analyze the digital optical signal waveform of the patient. The apparatus comprises a heart activity detection device, the pulse oximeter functions of a Nellcor N-100 Pulse Oximeter, and a microprocessor system incorporating software and memory for controlling and processing the oximeter and heart activity information. Additional inputs to a multiplexer and a digital status input latch of the oximeter are provided to receive the inputs from the heart activity detection electronics. The improved oximeter processes the detected heart activity waveforms simultaneously with and independent of the optical signals, both waveforms having been converted to digital signals for signal processing by the signal processing components of the N-100 oximeter.

The heart activity parameter may be provided by conventional and nonconventional methods capable of detecting heart activity independent of peripheral arterial pulses, including but not limited to EKG signals, ultrasound, ballistocardiogram, accelerometers, nuclear magnetic resonators, electrical impedance techniques, and the like. The primary requirement of the heart activity parameter and the related circuitry is that it provide an identifiable and detectable signal in response to each heartbeat for use by the signal processing of the oximeter.

In the preferred embodiment, heart activity parameter is detected by electronic heart detection circuitry in the form of an EKG signal which is passed through an instrumentation amplifier electrically isolated from the oximeter, and system electronics to generate a variety of waveforms derived from the EKG signal. The amplifier differentially amplifies the raw EKG data, inverts and returns the common mode signal to the patient to null the patient's common mode voltage, amplifies and AC couples the signal to eliminate any DC (offset) voltage component, filters the signal to eliminate unwanted frequencies such as, for example, frequencies below 0.05 Hz, buffers, and then couples the EKG signal to the system electronics. Coupling may be effected, for example, by amplitude modulation of a carrier signal across a transformer having the appropriate circuitry, or by an optically coupled isolation barrier.

The system electronics demodulates the coupled signal, where necessary, amplifies the signal and passes it to an automatic gain control ("AGC") amplifier to maintain the EKG signal output within a desired range even though the actual EKG signal strength may vary from patient to patient or from lead location to lead location.

In the preferred embodiment, the output of the AGC amplifier is routed through a polarity compensation circuit that changes the polarity of the waveform to have a preselected upgoing or downgoing polarity, without having to switch the leads or manipulate the patient. This is advantageous in critical life threatening situations where an incorrect connection of EKG leads otherwise might not permit proper detection of a heart rate and correlation with an optical pulse, and with fetal patients where it is not desirable to apply and reapply leads.

The resultant signal, referred to as the diagnostic EKG, is an analog representation of the electrical heart activity and can be displayed on an analog device such as a cathode ray tube or a chart recorder. The diagnostic EKG is filtered to select for the R wave of the EKG waveform and AC coupled to remove the DC component. The resulting signal is the filtered EKG signal.

The filtered EKG is processed to detect when an R wave occurs so that a digital pulse may be generated and sent to the oximeter to indicate that an R wave has occurred.

The oximeter functions remain essentially unchanged, except as specified herein. The microprocessor provides a bipolar drive current for the two LED's so that a positive current pulse drives the infrared LED and a negative current pulse drives the red LED. The magnitude of the current is adjusted by the microprocessor to help account for the variants of the patient's tissue. The light emitted by the LED's is detected by a single photodetector, preferably a photodiode, which generates a current proportional to the amount of transmitted light detected. The photocurrent may be amplified by a current to voltage converter. The resulting voltage is processed by the system electronics under the control of the microprocessor, to analyze and detect arterial pulses and to develop a history as to pulse periodicity, pulse shape, and oxygen saturation. The oximeter decides whether or not to accept a detected pulse as corresponding to an arterial pulse by comparing the detected pulse against the pulse history. To be accepted, a detected pulse must meet certain predetermined criteria in accordance with a desired degree of confidence. The blood constituent measurement is then made on the basis of accepted pulses.

According to the improved method and apparatus, the EKG signals from the electronic heart detection circuitry are processed using the analog to digital conversion and digital processing circuitry of the N-100 Pulse Oximeter to determine polarity, rhythmicity, and amplitude of the EKG signals. During this determination, the microprocessor converts the diagnostic EKG, the filtered EKG signal, or both, into digital EKG signals, analyzes the digital EKG signals, determines the amplitude and the polarity of the EKG, and adjusts the AGC amplifier and the polarity compensation circuit accordingly.

In the preferred embodiment the microprocessor operates in an integrated mode in which it develops and compares information from an EKG waveform and the optical pulse signal. The microprocessor first separately measures the time period by which an optical pulse follows an R wave, averages it over several pulses, independently calculates the pulse rate for each waveform, and compares the optical and EKG pulse rates. This insures reliability of both the electrical heart and arterial blood flow waveform analyses.

Predetermined criteria for optical pulse signals may include, for example, the expected size of the pulse, when the pulse is expected to occur, and the expected ratio of the red light to infrared light of the detected optical pulse. The predetermined criteria may be preselected or established by creating a pulse history. The pulse history may comprise a number of most recent pulses, e.g., four, in a pushdown stack memory which may automatically store the data for the last four accepted detected optical pulses.

The improved oximeter uses the measured time delay between an R wave and an optical pulse to determine a time window when, following the occurrence of an R wave, the probability of finding an optical pulse corresponding to a true arterial pulse is high. The time window provides an additional criterion to be used in accepting or rejecting a detected pulse as an optical pulse. Any pulses detected that do not fall within the time window are rejected and not used to calculate the amount of blood constituent. Similarly these rejected pulses normally do not become a part of the pulse history. However, if there have been no acceptable pulses within the time window for approximately 3 pulse periods, pulses within the time window that normally would be rejected will be accepted. This may be accomplished, as discussed below, by changing the predetermined optical pulse criteria.

Adjustments may be made to the microprocessor so that, when the optical signals are of high quality and easily detected a relatively high correlation between a detected pulse and the pulse history can be required before a detected pulse is accepted as an optical pulse. This would provide measurements having a high confidence level. When the optical signals are of low quality, the degree of correlation necessary can be lessened, providing measurements having a lower confidence level. This confidence factor may be adjusted in accordance with the beat to beat variability of the optical signals or the relative strength of the optical pulse signal.

If, even with degraded criteria no acceptable optical pulse is detected within the window for a specified period, e.g., 10 seconds, the microprocessor will revert to the initialization procedure and re-establish a relationship between EKG R waves and acceptable optical pulses.

In its integrated mode, the improved oximeter can calculate the blood constituent amount from the digital optical signal detected only during the determined time window. The time window thus can be used to reduce the processing of any spurious pulses caused by motion artifact or noise so that integrating the EKG information establishes reliable measurement of oxygen saturation.

One advantage to the integrated measurement of heart activity and optical signals is that it indicates that the oximeter is detecting an optical pulse when it is expected to occur. One advantage to using EKG signals is the determinable relationship between an R wave and an arterial pulse, which can confirm the regularity or irregularity of the heart beat, and ensure, for example, that the oxygen saturation measurements are based on the pulsatile component of the blood flow and are accurate.

Another advantage is that if one of the EKG or optical signals were to fail, the oximeter can revert to a non-integrated mode, allowing independent processing of the EKG and optical signals. The non-failing signal would continue to provide certain vital information and, more importantly, indicate that the failure of the signal was not due to the patient's loss of bodily function, e.g., cardiac arrest. Thus, the improved oximeter provides for redundant measuring of the heart rate of the patient, and indicates that one of the EKG or optical signal detection devices is not working properly. In the event that the missing signal is restored, integrated operations would resume as described above.

A further advantage of the improved method and apparatus is that patients who do not have a regularly occurring heartbeat can now be reliably monitored. The improved oximeter of this invention has improved capability to deal with arrhythmias and can detect and analyze the period of time by which an optical pulse follows an R wave and determine an appropriate time window. Then, on the occurrence of successive R waves, including irregularly occurring R waves, the determined time window is used so that the oximeter digitally processes the digital optical signal detected during the time window, develops an optical pulse history, and calculates the amount of blood constituent present. A patient having an irregular heartbeat also can be monitored, and an amount of blood constituent measured based on the actual blood pulse.

FIG. 1 is a block diagram of the improved method and apparatus of this invention.

FIGS. 2a and 2b are a circuit schematic of the EKG detection circuitry and the system electronics of this invention.

FIG. 3 is a graphical representation of the outputs of FIG. 2.

FIGS. 4a and 4b are a detailed circuit schematic of the microprocessor status input of FIG. 1.

FIGS. 5a and 5b are a detailed schematic of the microprocessor analog multiplexors and digital to analog converter of FIG. 1.

FIGS. 6a, 6b, and 6c are flow charts for the EKG and optical pulse related microprocessor operation of this invention.

FIG. 7 is a graphical representation of the outputs of FIG. 1.

FIG. 8 is a graphical representation of the oximeter timing diagram.

FIGS. 9a and 9b are a detailed circuit schematic of the isolated front end printed circuit board of FIG. 1.

As shown in FIG. 1, an embodiment of this invention comprises EKG detection means 100 and pulse oximeter 200. EKG detection means 100 has positive lead 102, negative lead 104, and reference lead 106, each electrically connected to patient 108. Typically, positive lead 102 is connected to the right arm, negative lead 104 is connected to the left arm and reference lead 106 is connected to the right leg. When the patient is a fetus, positive lead 102 is connected to the fetus, negative lead 104 is connected to the material vaginal canal, and reference lead 106 is connected to the maternal right leg. An alternate perinatal oximeter probe combining optical detecting means and EKG detecting leads is described in co-pending and commonly assigned U.S. patent application Ser. No. 644,051, filed Aug. 24, 1984, which disclosure is incorporated herein by reference.

EKG detection means 100 also includes preamplifier 120, coupling circuit 138, automatic gain control ("AGC") amplifier 140, polarity switch 160, bandpass filter 170, DC level shifter 180, and R wave detector 190. In operation, EKG detector means 100 produces three outputs, diagnostic EKG waveform DEKG, filtered EKG waveform FEKG, and detected R wave DRW. These outputs are shown in FIG. 3.

Referring to FIGS. 2a and 2b preamplifier 120 is mounted on the EKG front end printed circuit board 164 ("board 164") which is electrically isolated to protect patient 108 from strong electrical signals used to analyze and process the EKG and the optical signal waveforms. Preamplifier 120 comprises instrumentation amplifier 122, preferably a Burr-Brown INA104HP Very High Accuracy Instrumentation Amplifier, protected against high voltage inputs by diodes 124 and isolation voltages .+-.V.sub.iso. It is designed to have a gain factor of about 100.

The signal generated at negative lead 104 is input to pin 17, the signal generated at positive lead 102 is input to pin 2, common mode signal CMS is returned to the patient by reference lead 106 thus lowering the null common mode voltage of the patient, increasing rejection of common mode signals. The other pins are connected as indicated in FIGS. 2a and 2b, as known to one skilled in the art. For ease of understanding and distinguishing what input or output of the particular solid state element is being discussed, the electronic circuit drawings may be referred to herein as element "A101-16", meaning element A101 at pin 16.

The output of instrumentation amplifier 122 is passed to isolation amplifier 500, preferably model 286J, manufactured by Analog Devices, Inc. Isolation amplifier 500 provides transformer coupling of the EKG signal from isolated preamplifier 120 to the system electronics. Isolation amplifier 500 also provides isolated power for instrumentation amplifier 122. Oscillation circuit 510 consists of hex Schmitt inverter 511, resistor 512, capacitor 513, and variable resistor 514. This circuit provides a 100 kHz signal for proper operation of isolation amplifier 500.

The EKG signal, once coupled to the system electronics, travels to two different circuits. The first circuit is LDOFF detector circuit 134. LDOFF circuit 134 indicates when the EKG leads have become disconnected or inoperative and comprises parallel comparators 135a and 135b arranged in a window comparator configuration so that when the output of isolated preamplifier 120 is within + or -3.8 volts, the voltage at node 136 will be at +5 volts, realized from the pullup resistor 137. Comparator 135a has a resistor-divider network with a filter capacitor connected to the inverting input of comparator 135a consisting of resistors 1134a, b and capacitor 1134d, and is provided with a +15 volt reference voltage. This voltage is divided down to +3.8 volts and is presented to the inverting input across resistor 1134c. Similarly, comparator 135b has resistor-divider network with a filter capacitor connected to the non-inverting input of comparator 135b consisting of resistors 1135a, b and capacitor 1135d and is provided a - 15 volt reference voltage. This reference voltage is divided down to -3.8 volts which is presented to the non-inverting input of comparator 135b across resistor 1135c. Comparator 135a has feedback resistor 1135e connected to the non-inverting input from the output to provide hysteresis.

The output from isolated preamplifier 120 is fed to both the inverting input of comparator 135b and the non-inverting input of comparator 135a across filter capacitor 1131 and resistor 1130. The voltage at node 136 will be at +5 volts when the leads 102 and 104 are properly connected to patient 108. If either lead 102 or 104 becomes disconnected or inoperative, the voltage at node 136 will be at 0 volts. This is the digital OVFLG which is presented to status latch 9G-13.

The EKG signal is also fed to the second circuit, a bandpass filter circuit 330 consisting of buffer amplifier 331, resistors 332-333 and capacitors 334-336 designed to selectively filter out frequencies below about 0.05 Hz and frequencies above about 100 Hz. The signal is then passed through notch filter 380 to eliminate selected signal components for example, 60 Hz or 50 Hz, primarily designed to eliminate any interference from noise sources such as from the power line. Notch filter 380 consists of amplifiers 381 and 382, resistors 383a-f, capacitors 384a-b, and variable resistors 385a and b, for tuning the filter to 60 Hz. The output of notch filter 380 is substantially identical in waveform to the output of instrumentation amplifier 122.

The output of notch filter 380, the EKG signal, is input to bandwidth limited inverting amplifier 142, and to AGC amplifier 140 which receives the bandwidth limited signal at pin 15, and an analog input, of digital to analog converter ("DAC") 144, located in the feedback loop of inverting amplifier 143. DAC 144 also receives digital input from latch 145. The digital word fed to DAC 144 is entered into latch 145 by microprocessor 16 of the oximeter. By changing the digital word fed to latch 145 in response to the amplitude of diagnostic EKG signal DEKG, microprocessor 16 can adjust the gain of AGC amplifier 140--DAC 144 is utilized as a variable resistor in the feedback loop.

Amplifier 147 provides a second level of gain to the signal, which is then fed to polarity switch 160. Polarity switch 160, preferably a DG201 Analog Switch, manufactured by Siliconix, is designed to maintain uniform polarity of the EKG signal as it is being processed by appropriately gating the signal to one of either an inverting or noninverting input of amplifier 162. Microprocessor 16 processes the filtered EKG waveform, detects polarity, and generates a voltage signal, for example, +5 volts, which is also inverted by inverter 161 to form a second voltage signal, e.g., 0.0 volts, which together to form a logical word (polarity, polarity). The voltage values of the logical word causes polarity switch 160 to gate the EKG signal being processed to the appropriate input of amplifier 162 accordingly. The output of amplifier 162 is diagnostic EKG signal DEKG which is buffered by amplifier 168 and sent to the analog to digital converter ("ADC") of the pulse oximeter for conversion by microprocessor 16.

The output of amplifier 162 is amplified by amplifier 166 and also fed to bandpass filter 170, for selectively passing frequencies from about 15 to about 40 Hz having a center frequency of about 20 Hz. The filtered signal is passed through capacitor 176 for eliminating any DC voltage components that may have been introduced during prior amplifications, and is then inputted to DC level-shifter 180. DC level-shifter 180 comprises an offset voltage V.sub.off, preferably +5 volts, being fed to the noninverting input of amplifier 182 and the filtered signal being fed to the inverting input of amplifier 182. V.sub.off is adjusted so that the output of amplifier 182 will be within the detectable range of the ADC of the pulse oximeter. In the preferred mode, the ADC of the pulse oximeter can only detect positive voltages, necessitating DC level-shifter 180. In circumstances where the ADC of the pulse oximeter can detect a bipolar signal having positive negative voltages, a DC level-shifter may not be required.

The output signal of amplifier 182 is filtered EKG signal FEKG, which is buffered by amplifier 186 and sent to the ADC of the pulse oximeter for waveform analysis. Output FEKG is also fed to R wave detector 190 comprising comparator 192, threshold voltage V.sub.th, preferably +5.5 volts, and digital pulse voltage V.sub.CL2, preferably a pull up voltage of +5 volts. When the amplitude of filtered EKG signal FEKG rises above the value of threshold voltage V.sub.th input to the inverting input of the comparator 192, comparator 192 generates as its output a digital pulse representing a logical 1, having an amplitude equal to V.sub.CL2, e.g., +5 volts. At all other times, comparator 192 has an output that is a logical 0, e.g., about 0 volts. V.sub.th may be adjustable by the operator of the pulse oximeter so that if no R wave pulses are generated, the threshold voltage (and the confidence level) can be reduced until R wave pulses occur. Alternately, V.sub.th can be adjusted by the microprocessor if no R wave pulses are detected and the LDOFF signal indicates there should be R wave pulses. The output signal of R wave detector 190 is detected R wave DRW, and each pulse (+5 volts) represents the occurrence of R waves (logical 1's) in the patient's EKG waveform, as shown in FIG. 3.

Referring to FIGS. 4a and 4b, detected R wave DRW is then fed to input 280-2 of flipflop 280. Flipflop 280 changes logical conditions from a logical 0 to a logical 1 output at output 280-1 when it detects the rising edge of an R wave pulse as the voltage of signal DRW rises from about 0 to V.sub.C12, and will maintain a logical 1 output until cleared by microprocessor 16 at clear input 280-3. When cleared, flipflop 280 has a logical 0 at output 280-1. The output of flipflop 280 is fed to status input latch 9G-19, where it is stored as a logical 1, representing an R wave flag. In this manner, the presence of an R wave pulse is indicated even though the instantaneous R wave signal DRW voltage has returned to logical 0.

Microprocessor 16 frequently checks the condition of each input of status input latch 9G for information relating to signal processing sequence control. As indicated in FIGS. 6(a-c), the presence of an R wave flag causes microprocessor 16 to (1) reset output 280-1 of flipflop 280 to a logical zero output, thereby clearing the R wave flag at status input latch 9G-19 so that flipflop 280 will return to a logical 1 output when it next detects an R wave pulse, and (2) either initiate non-integrated EKG waveform analysis to determine frequency and regularity of R waves to establish the period of delay between an R wave and an optical pulse (see FIG. 7) or initiate integrated searching for an optical pulse waveform during the established time period to analyze vital signs such as oxygen saturation, pulse flow, and pulse rate.

As shown in FIG. 3, diagnostic EKG waveform DEKG comprises an analog wave train of signals having components labeled P, Q, R, S, and T. The QRS portion is representative of ventricular contraction of the heart, the occurrence of the heartbeat. For normal patients, each heartbeat generates a similar PQRST pattern. Filtered EKG waveform FEKG is also an analog signal but contains substantially only the R portion of the diagnostic waveform as the other components are filtered out. The R portion is more distinctive than the other components having a significantly greater slope and amplitude. The R wave of filtered EKG waveform FEKG corresponds to the R wave portion of diagnostic EKG waveform DEKG, and detected R wave DRW contains a step or digital R pulse waveform that corresponds to the R wave portion of diagnotic EKG waveform DEKG.

The arterial blood pulse detection circuitry is the same as that found in the N-100 Pulse Oximeter manufactured and sold by Nellcor Incorporated, Hayward, Calif.

Referring to FIGS. 1, 4a, 4b, and 8, pulse oximetry occurs as follows. Clock 70 has a duty cycle of four segments .phi.1, .phi.2, .phi.3, .phi.4 that are sequential. Clock 70 is connected to microprocessor 16. Segment .phi.1 turns on LED 30, segment .phi.2 turns off LED 30, segment .phi.3 turns on LED 32 and segment .phi.4 turns off LED 32. The LED's are strobed in sequence so that only one LED is transmitting at a time. The LED's are turned off to allow the photoelectric detector to return to a quiescent condition to measure ambient environmental light levels. As clock 70 operates through its duty cycle, the light transmitted through the tissue of patient 108 is received by photodetector 38. Clock 70 has three signal outputs A, B and P. Outputs A and B are input to a conventional pulse width modulation circuit to couple the pulse widths from the system electronics to board 164, and to establish the desired LED intensities for the LEDs 30 and 32. The reference intensities are established by microprocessor 16 which generates intensity voltage V.sub.L1 for infrared LED 30 and intensity voltage V.sub.L2 for red LED 32. These reference voltages are adjusted as described elsewhere herein, and fornm a part of the output of hold circuit 200.

Referring to FIG. 1 and FIGS. 5a and 5b parallel pulse width modulation circuits 220 and 230 are shown. Circuit 220 has as inputs A, A, -15 volts, +15 volts, and V.sub.L1. Matching amplifiers 3E are used with the same resistor, gates, and capacitor networks shown as ramping generators to provide the waveform labeled "ramp" in FIG. 8. When A is a logical 1, gate 2E will open circuit, the -15 volt supply, otherwise connected to inverting input 3E-6 and the feedback loop comprising resistors R5 and R7 and capacitors C25 and C103, will be made conductive by the closing of gate 2E-9. This condition will cause the amplifier output to ramp from 0 volts to +15 volts as shown in FIG. 8 labeled ramp. When A is a logical 0, or at about 0 volts, gate 2E-16 is closed and a -15 volt supply is input to amplifier 3E at 3E-6, and feedback loop resistor R7 is open circuited by gate 2E-9 so that output 3E-7 will be and maintained at about 0 volts. During the time A is a logical 1, signal B is a logical 0. Because pulse width modulation circuit 230 works the same as circuit 220, ramping output 3E-1 will be at about 0 volts, except when signal B is a logical high when it will ramp from about 0 to +15 volts.

Outputs 3E-7 and 3E-1 are inputted to comparators 4F-6 and 4F-2, respectively. Intensity voltage inputs V.sub.L1 and V.sub.L2 Ll are fed to comparator inputs 4F-5 and 4F-3, respectively, so that when a ramping voltage exists it is compared to its respective intensity voltage. Thus, comparator output 4F-7 will reflect a logical 1 condition, about +5 volts from pull up voltage at resistor R72, for the time period when ramping voltage at 3E-7 is less than intensity voltage V.sub.L1. When ramping voltage is greater than V.sub.L1, output 4F-7 will change to a logical 0, creating a pulse having a width responsive to the intensity level. Similarly, comparator output 4F-1 will be a logical 1, about +5 volts, during the time period when ramp voltage 3E-1 is less than intensity voltage V.sub.L2 from pull up voltage +5 volts across resistor R73. Thus, the outputs of 4F-7 and 4F-1 are pulses having a width representing the desired voltage intensity for V.sub.L1 and V.sub.L2, respectively.

The outputs of comparators 4F-7 and 4F-1 are input to NAND gate inputs 3F-13 and 3F-1, respectively. Signal A is input to NAND gate input 3F-12 and signal V is input to NAND gate input 3F-2. The outputs of NAND gates 3F-11 and 3F-3 are input to NAND gate 3F-5 and 3F-4 respectively, where the signals are effectively combined to that output 3F-6 is a digital waveform slope shown in FIG. 8.

Output 3F-6 is also input to both inputs 2F-1 and 2F-2 of NAND gate 2F, and output 2F-3 is input to NAND gate input 2F-4. Signal LED*2 is input to NAND gate input 2F-5, and output 2F-6 is digital waveform zero, shown in FIG. 8.

Signals "Slope", "Zero", and "Phase", the latter being generated by clock 70 and shown in FIG. 8, are coupled to board 164 by optical couplers U4, U5, and U6, respectively (shown in FIGS. 9a and 9b). The signals are input to receiver circuit 800 for decoding of the pulse width information contained in signals Slope, Zero, and Phase and for generating voltages to be used for driving infrared LED 30 and red LEd 32 in accordance with the stated duty cycle.

Referring to FIGS. 8, 9a and 9b, signal Zero controls switching gate U8-16 which alters the gain of amplifier U11. Signal Slope controls gate U8-1 which controls whether there is input to integrator amplifier U11 and associated capacitor C6. The magnitude of output U11-1 is dependent upon the width of signal Slope. The greater the width is, the longer gate U8-1 will be closed. This directly relates to how long amplifier U11 and capacitor C6 will integrate the input signal, or the peak value at output U11-1. This in turn relates to a proportional current level in amplifier circuit 804 through the selected LED resulting in the selected intensity of emitted light. Signal Zero acts to turn off the gain of amplifier U11 at selected intervals so that the voltage will decay to about zero and the current in amplifier circuit 804 will decay to about zero, thus turning off whichever LED was on and allowing it to return to its quiescent state. Signal Zero also prevents leakage current from causing a progressive error in integrator amplifier U11 and associated capacitor C6. Signal Phase controls gates U8-8 and U8-9 which selects which voltage supply, +15 volts or -15 volts will be input to gate U8-2 for through-put to integrator amplifier U11 when gate U8-1 is closed by signal Slope.

The signal at amplifier output U11-1 thus provides the waveform (shown in FIG. 8) that controls LEDs 30 and 32. LEDs 30 and 32 are connected in parallel, anode to cathode and cathode to anode at the output ports J2-9 and J2-7 of amplifier circuit 804. Amplifier circuit 804 converts the voltage output U11-1 to the LED drive current using power amplifier U10 and current sensing resistor R25. Therefore, as output U11-1 varies from positive to negative, as converted to current by circuit 804, a positive current at port J2-9 turns on LED 30, LED 32 remaining off and open circuited by the current bias, and a negative current turns on LED 32 and open circuits LED 30. Between the positive and negative currents, the LED drive current has been turned off, turning off LEDs 30 and 32, due to the effect of signal Zero.

The light emitted by LEDs 30 and 32 is passed through tissue of patient 108, preferably through a finger. Alternate preferred tissue locations include the ear lobe, nasal septum, reflected light off the forehead and the like. In situations where reflected light is used, it is preferred to place an optical barrier (not shown) between the emitting LEDs and the detecting photodetector to prevent distortion of the light content transmitted through the tissue.

Referring to FIG. 1, photodetector 38 receives all light transmitted through the tissue of patient 108 so that photodetector 38 receives infrared plus ambient light and noise during clock output .phi.1, ambient light and noise during clock output .phi.2, red plus ambient light and noise during clock output .phi.3 and ambient light and noise during clock output .phi.4. This signal ("DLS") is passed through preamplifier 40, which converts the photodetected current into a voltage at a rate of about one volt per microamp, capacitor 41, and is then coupled by transformer 39 from electrically isolated board 164 to the system electronics. After coupling to system electronics, which may itself be electrically isolated, the signal is processed by parallel circuitry for separating the red and infrared signals, in order to adjust for the different gains required to process red and infrared signals. Clock 70 outputs .phi.1-.phi.4 control the synchronization detector gates 2H to divide composite signals DLS into infrared light signal IRLS and red light signal RLS and direct signals IRLS and RLS to parallel amplifiers 43 and 44. Parallel synchronous detector gates 2H and parallel amplifiers 43 and 44 also act to invert the pure ambient light and noise signals and, using a slow time constant filter, add them to the adjacent LED light plus ambient light and noise signals to subtract out the ambient light and noise signal components. These filtered signals then pass through parallel low pass filters 45 and 46 to eliminate the switching frequencies and noise. Signal outputs V.sub.a and V.sub.b are sent to the ADC of the oximeter for digitization, and also sent through parallel offset amplifiers 47 and 48 for subtracting out a portion of the DC bias and amplifying the remaining voltage signals. Offset amplifiers 47, 48 increase the resolution of the AC voltage signal component for digital conversion. Outputs V.sub.a' and V.sub.b' are also sent to the ADC of the oximeter for analog to digital conversion.

Referring to FIGS. 5a and 5b, the system electronics for processing signal DLS is described in more detail. Signal DLS is coupled through transformer 39 on front end printed circuit board 164 and enters the system electronics at pin 10 of connecting strip J1. Signal DLS consists of the time-sequenced response of photodetector 38 to infrared LED 30 and red LED 32 as shown in the timing diagram (see FIG. 8 under the label "Photocurrent"). Signal DLS is amplified by one-half of dual amplifier 1K.

The output of amplifier 1K is connected to four analog switches forming switching element 2H of parallel two-channel synchronous detectors 43 and 44 which separate the pulses of infrared and red light detected by photodetector 38 and eliminate low frequency noise and DC offset voltages. As clock 70 goes through its duty cycle, during state .phi.1 first analog switch 2H-10 closes, coupling preamplified signal DLS through detector channel 43 and amplifier 2K-1 having a gain of about -1 formed by resistors R33 and R34. During state .phi.2, first switch 2H-10 opens and second switch 2H-7 closes so that amplifier 2K-1 has a gain of about +1 wherein it acts essentially as a voltage follower with resistors R34 open circuited and resistor R33 forming the feedback loop. During states .phi.3 and .phi.4 red LED 32 turns on and off and a similar switching occurs for second detector channel 44 and amplifier 2K-7. The output of detector amplifiers 2K-1 and 2K-7 will thus be active with a duty cycle of 50 percent, half of that inverting and half non-inverting. Any DC or low-frequency voltage should be cancelled by the two adjacent pulses of opposite polarity, while photodetector signal DLS, which is present in only one of the two time states, will be amplified with an effective gain of about 0.25.

The outputs of amplifiers 2K-1 and 2K-7, signals IRLS and RLS, respectively, are fed to matched low-pass filters 45 and 46 for passing only frequencies below about 10.0 Hz, having amplifiers 3K-1 and 4K-1, and 3K-7 and 4K-7, for providing a respective gain of about 4 to signals IRLS and RLS. These filters remove the switching component which is about 2 kilohertz and filter out any high frequency noise.

The signal processing means for processing the EKG signals and the optical signals includes a programmed microprocessor such as the Intel 8085 A. The basic functions of the equipment will be described for comprehension, while the improvements forming a part of this invention will be described in detail.

Referring to FIG. 1, the signal processing means comprises microprocessor 16, data bus 17, RAM 19, ROM 18, latch 23, comparator 52, analog multiplexor 50, hold circuits 200, gate 24, latch select 21, latch digit 22, and display 20, each connected to bus 17 and thereby under the control of microprocessor 16. Data bus 17 shunts digital information into and out of microprocessor 16 and each of the components. Latch select 21, latch digit 22, and display 20 all relate to a preferred numerical display of the amount of blood constituent measured, e.g., optical pulse rate and oxygen saturation.

The function of the signal processing means is to convert the analog signals from the optical signal detector and from the EKG detector, independently, to digital signals, for subsequent waveform analysis. The waveform analysis is controlled by microprocessor 16, ROM 18, and RAM 19.

FIGS. 6(a-c) and the 123 page software appendix relate to the software used by the oximeter to control the signal processing of the optical signals and the EKG detection system. The microprocessor of the oximeter evaluates the optical signal to determine the oxygen saturation and pulse rate according to the following method.

A preferred embodiment of this invention incorporates into microprocessor 16 the means for processing the EKG signals and displaying the calculated EKG pulse rate, converting the analog diagnostic EKG signal DEKG and filtered EKG signal FEKG to digital EKG signals using the same analog to digital conversion circuits used for processing the optical signals. Referring to FIGS. 5a and 5b, analog multiplexer 50 is the input to the ADC of the pulse oximeter, and comprises two analog multiplexors 5G and 6G. Optical pulse signals V.sub.a', V.sub.b', V.sub.a, and V.sub.b are connected to pins 13, 14, 15 and 12 of multiplexer 5G. According to the improved apparatus, diagnostic EKG signal DEKG is connected to pin 15 of multiplexer 6G and filtered EKG signal FEKG is connected to pin 12 of multiplexor 6G.

In order to convert any of the analog inputs to digital signals, microprocessor 16 must address the proper channel of one of analog multiplexers 5G and 6G by inputting a three bit word through bus 17 to pins 9, 10, and 11 of both multiplexers 5G, 6G. The microprocessor program is configured to allow for analog to digital conversion of the EKG signals DEKG, FEKG in addition to V.sub.a, V.sub.b, V.sub.1', and V.sub.b', and appropriate storage of the digital signals in RAM 19.

Referring to FIGS. 4a, 4b, 5a and 5b, microprocesser 16 converts analog signals to digital signals by selecting which input is to be converted and loading a digital word into latches 8AH and 9K. Latches 8H and 9K store the digital word presented to the inputs of digital to analog converter ("DAC") 8K, which converts the digital word to analog signal DAC. Signal DAC is fed to pin 2 of comparator 5H. The other input to comparator 5H, at pin 3, is the analog signal from multiplexer 50 selected by microprocessor 16 for conversion. When the analog signal provided by DAC 8K exceeds the analog value presented by multiplexer 50, output DACMP of comparator 5H is at logical 1. The digital word which generates the analog voltage from DAC 8K that is less than the analog voltage present at multiplexer 50 will cause output DACMP of comparator 5H to change to logical 0. Output DACMP is inputted to status latch 9G-17 shown in FIG. 4b which is sampled by microprocessor 16 at a rate of about 57 cycles per second. When the microprocessor detects a logical 0, the word stored in latches 8H and 9K of FIG. 6 represents the digital value of the analog signal and is stored by microprocessor 16 into the accessed address of RAM 19 for later processing.

In connection with EKG signal processing, as shown in the software appendix, microprocessor 16 analyzes the stored digital words and calculates an amplitude for the EKG waveform. This amplitude is used to control AGC amplifier 140 by changing the digital word fed to DAC 144 so that outputs DEKG and FEKG will fall within and be compatible with the voltage range limits of the electronic circuitry used to process the signals, without losing any of the significant information contained therein. The start up or non-integrated condition includes independent and continuous signal processing of the optical pulse to calculate and display oxygen saturation and pulse rate, and simultaneously, continuously processing EKG waveforms DEKG, FEKG, and DRW. When certain conditions exist, flags are raised at status input latch 9G and internally within microprocessor 16, indicating what operation is to follow. Referring to FIGS. 4a, 4b, 5a and 5b, microprocessor 16 regularly searches status input latch 9G at a rate of about 57 cycles per second. According to this invention, output DRW is also input to status latch 9G through flipflop 280. Thus when detected R wave DRW is a logical 1, the microprocessor detects the 1 status and based upon that status selects the next operation. That operation can be one of the following events which occur in sequence. At startup conditions, upon detecting an R wave, microprocessor 16 clears output 280-1 of flipflop 280 to logical 0, clearing status input latch at input 9G-19 relating to EKG output DRW. At this first level, microprocessor 16 begins counting time intervals, using clock 70, from the detection of an R wave pulse DRW until the occurrence of the next logical 1 at status input latch 9G. Based upon this time interval, the improved pulse oximeter displays the pulse rate. After averaging several time intervals and establishing a regular EKG pulse rate, microprocessor 16 will change to the second level of processing.

With the detection of an R wave pulse, microprocessor 16 will begin to correlate the period of time by which an optical pulse, as separately determined by microprocessor 16 analyzing the digital optical signal, follows the detected R wave pulse to establish the time window during which the optical pulse is likely to occur. During this second level, the pulse oximeter is still calculating and displaying the time period or pulse rate between detected R wave DRW pulses.

The third level of processing is reached after a time window has been established. On detecting an R wave pulse, microprocessor 16 activates the time window so that only optical signals detected within the time window, following the occurrence of a R wave pulse, will be evaluated for acceptance or rejection for use in calculating and displaying vital measurements such as oxygen saturation, pulse flow, and pulse rate. The evaluation of a detected pulse is made in conjunction with a preselected confidence factor that is associated with the quality of the optical signals. The higher the optical signal quality, the better the correlation between the recorded pulse history and a detected pulse, and the higher the confidence level. The confidence level may be set automatically by the microprocessor, or it may be adjusted by the operator of the improved oximeter. Microprocessor 16 will reject any detected pulses occurring outside a time window. A typical time window for an adult male having a fingertip oximeter probe may be about 50 milliseconds, plus or minus 10 milliseconds, after the occurrence of an R wave.

The oximeter will also reject any additional pulses detected after an optical pulse is detected within the same time window, even though the time window has not expired.

However, if an optical pulse is not found within an opened time window, microprocessor 16 will continue to search for optical pulses using the degraded criteria during the time window period for a limited number of successive detected R wave DRW pulses, e.g., 3, after which it continues to search with degraded criteria. After a specific interval, e.g., 10 seconds, without detection of an optical pulse, microprocessor 16 will revert to independent or non-integrated processing of the optical signal and the EKG signals, returning the pulse oximeter to the start up condition. Therefore, if the oximeter cannot establish or maintain a reliable correlation between the R wave and the optical pulse, the waveforms will be processed independently. Preferably there is a display to indicate that the oximeter is integrating the EKG and optical signal data and so calculating the blood constituent amount. After attaining the third level of processing, losing either the EKG or optical pulse signals will activate an alarm and return the program to the start up condition.

Having described the overall operation, referring to FIGS. 6a, 6b, and 6c the flow chart for the software calculations is shown and described. In FIG. 6a, the R wave determination routine begins at 600 with electrical signals received from the EKG leads and calculating the period RRPER between the last detected R wave and the presented R wave at 601. The average period HISTORY from previous R waves and the present R wave is calculated at 602 and the determined period RRPER from 601 is compared to the average period HISTORY at 603. If RRPER does not correspond to HISTORY at 604, then the routine jumps to 613 where the R wave (or EKG flag) at flip flop 280 is reset and the routine is exited to await another R wave. If RRPER does correspond to HISTORY at 604, then a timer is activated at 605 to measure the interval from the occurrence of the R wave to the occurrence of the optical pulse. At 606, output HR (EKG heart rate) is calculated based on successive R waves. At 607, the system inquires whether a series of R-R periods have been synchronized (EKG synchronization). If not synchronized, then the system checks for alarms by comparing output HR at 609, to a preselected heart rate and generating an alarm if output HR is too low. If the EKG is synchronized but the optical pulse to optical pulse series is not synchronized at 608, then output HR is sent to display at 609 and then checked for alarms at 610. However, if the optical pulse is synchronized at 608, then the system just checks for alarms at 610. Only if the EKG is synchronized and the optical pulse is not synchronized, and if the R wave looks like a valid R wave by comparison with HISTORY, then HISTORY is updated using the new R wave at 611. After updating HISTORY, the system itself is updated (TIME OUT) to maintain synchronization at 612. If TIME OUT is not updated for a period of five seconds, then EKG synchronization is lost and must begin building a new history.

Referring to FIG. 6b the system routine for processing digital optical pulse information for optical pulses to send to LEVEL 3 (shown in FIG. 6c) is flow charted. The system begins by continuously evaluating the data from the detected digital optical signal at 644. The data is first evaluated for compatability with the signal processing at 645. If the data is over or undervalued electronically, i.e., beyond the voltage range of the circuitry, then the system exits the routine at 646, and the LED intensities are adjusted to correct the electrical values accordingly. When the data is compatible, it is next evaluated for a maximum signal. A relative maximum is determined and saved at 651. The next value is compared to the saved value, and if it is a new maximum it is saved at 651 instead. When the value found is now a new max, then a MAX FLAG is set at 650. Thereafter, the system evaluates the following data received, bypassing the maximum value section 648-652, to find the maximum slope at 653, again by successive comparisons. When the largest slope value is found it is saved at 658 and the SLOPE FLAG is raised at 656. Thereafter the following data is evaluated, bypassing the maximum and slope calculations, to find the minimum value corresponding to the end of the pulse at 659-662. When the smallest minimum is found, it is saved at 661 and the slope value saved at 658 is compared to a pre-established minimum threshold to determine whether it is large enough to be a possible optical pulse at 663. If it is not large enough, then the pulse is rejected at 664, the FLAGs raised at 659 and 656 are reset at 665 and the routine begins processing the next possible pulse at 644. If the slope is large enough then the pulse parameters, maximum, minimum, and slope, are saved in memory at 667 for use by LEVEL 3 processing in evaluating the possible pulse. Then, the time delay from the R wave to the possible pulse is calculated. Therafter, the DATA FLAG is set at 669, indicating to LEVEL 3 that there is a possible pulse to be evaluated, the MAX and SLOPE FLAGs are reset at 670, and the routine begins again to process the following data, looking for new maximum values corresponding to possible pulses.

Referring to FIG. 6c, LEVEL 3 of software for computing the saturation measurements is shown. The system starts by inputting a potential optical pulse at 615 after a DATA FLAG has been raised and inquiring whether there is EKG synchronization i.e., a regular EKG period has been established. If a DATA FLAG has not been raised, then the system exits the routine at 617. If there has not been EKG synchronization, then the microprocessor processes the optical pulse signals independent of the EKG, as would occur in the Nellcor N-100 oximeter without EKG capability, bypassing the inquiry into the presence of an R wave at 616.

If there is EKG synchronization, but no R wave has occurred, then the system exits at 617 and the pulse is not processed. If there is EKG synchronization and a R wave has occured, then the microprocessor processes the pulse as described below. The LED intensity is evaluated to see if adjustment is necessary at 618. The reset system gain, based on the minimum LED intensity required for adequate signal strength, is checked to see if adjustment is required. The optical pulse history is then computed at 620, based on the average historical period, amplitude, and ratio. The system then inquires whether the EKG apparatus is operating properly at 621. If it is, then the average time period between an R wave and the following optical pulse for the most recent four prior pulses is computed to give the TIME WINDOW at 622. Then the pulse waveform is analyzed to see if it is a dicrotic notch rather than a real optical pulse at 623. The downward slope of a dicrotic notch or other artifact can be misinterpreted as an optical pulse, but typically the pulse amplitude is less than half the amplitude of an actual pulse. If the pulse is determined to be a notch or artifact at 624, then the system exits at 625 and the next pulse presented will be processed. If not determined to be a notch, then it is analyzed to determine if it is a pulse at 626.

Assuming the EKG is synchronized, then the system determines if two criteria are met. The first is whether the time delay falls within the above-computed TIME WINDOW. If it does not, then the microprocessor rejects the pulse. The second criteria tested is whether or not the ratio is within acceptable limits. Only if the pulse statisfies both criteria, is the pulse accepted and a saturation calculation made.

If the EKG is not synchronized then the comparison must provide any two of three factors, (1) pulse period, (2) amplitude, and (3) ratio, as favorable for the pulse to pass as an accepted pulse at 627. E.g., pulse and period, period and amplitude, pulse and amplitude, or all three. If the pulse is accepted, then the oxygenation saturation is calculated at 628.

After the system is turned on (POWER UP) or after a TIME OUT alarm (a ten second period with no valid optical pulse found) a series of consistant pulses must be found to generate an optical pulse history before the oxygenation saturation will be sent to the display. Thus, if there is no optical pulse synchronization at 629, there will be no saturation display generated at 630. All optical pulses, those accepted and those not accepted, excluding pulses rejected as artifacts, enter the calculation routine section at 631-643. If the EKG is not synchronized then a pulse to pulse period and either an amplitude or a ratio must exist for the optical heart rate (OHR) calculation to be made at 632. If either the EKG or the optical pulse is synchronized, then the HR calculation made at 632 will be displayed at 634. If there is no synchronization, then the OHR is not displayed. At 635-643, the system is evaluating the status for pulse evaluation, i.e., whether signals should continue to be processed after a TIME WINDOW has been opened. If there was EKG synchronization and a good pulse was found, or the TIME WINDOW period has expired then TIME WINDOW is closed until opened by the detection of the next R wave.

In the preferred embodiment, the blood constituent measured is the oxygen saturation of the blood of a patient. The calculation of the oxygen saturation is made based on the ratio of the pulse seen by the red light compared to the pulse seen by the infrared light in accordance with the following equation: ##EQU1## wherein BO1 is the extinction coefficient for oxygenated hemoglobin at light wavelength 1 (Infrared)

BO2 is the extinction coefficient for oxygenated hemoglobin at light wavelength 2 (red)

BR1 is the extinction coefficient for reduced hemoglobin at light wavelength 1

BR2 is the extinction coefficient for reduced hemoglobin at light wavelength 2

light wavelength 1 is infrared light

light wavelength 2 is red light

and R is the ratio of the optical density of wavelength 2 to wavelength 1 and is calculated as: ##EQU2## wherein I.sub.max2 is the maximum light transmitted at light wavelength 2

I.sub.min2 is the minimum light transmitted at light wavelength 2 P0 I.sub.max1 is the maximum light transmitted at light wavelength 1

I.sub.min1 is the minimum light transmitted at light wavelength 1

The various extinction coefficients are determinable by empirical study and are set forth in the software appendix. For convenience of calculation, the natural lo of the ratios may be calculated by use of the Taylor expansion series for the natural log.

In an alternate embodiment, the microprocessor program can be adapted to utilize the relationship between the detected R wave DRW pulses and optical pulses without the need for first determining a pulse history. In this embodiment, microprocessor 16 searches status input latch 9G, and when detected R wave DRW is a logical 1, analyzes the optical signals that follow a detected R wave DRW pulse, regardless of the frequency of R wave pulses. By comparing the optical signals following a number of R wave pulses, microprocessor 16 correlates the detection of a pulse indicative of an optical pulse to the period of time by which such a detected pulse follows a detected R wave DRW pulse.

Referring to FIGS. 2a and 2b, the EKG front end printed circuit board schematic shows a respiratory monitor portion which may be used in conjunction with the EKG enhanced oximeter. The respiratory monitor is designed for use with a pressure sensitive transducer for detecting respiration or chest wall movement by measurement of pressure change. The sensor may be, for example, pneumatic-type sensor such as a Grasby Dynamics pressure capsule sensor, or a liquid mercury filled tube of silicon-like rubber secured across the patient's chest acting as a variable resistor. For a pneumatic type sensor, however it is designed, the small pressure change generated during respiratory chest movement is transmitted to a pressure transducer, for example, a Sensym model LX0503A bridge pressure transducer, for conversion to a voltage signal. The bridge output signal, or other voltage signal, is connected to differential amplifier 1A having a gain factor preferably over 100 nominally. The output of amplifier 1A is AC coupled to reject frequencies below about 0.07 Hz, and then buffered by amplifier 1B. The signal is then passed to low pass filter stages 1B and 1C which have a nominal cut off frequency of about 5 Hz. The output of the lowpass filters is further amplified and AC coupled to first buffer amplifier 1D. At that point the signal goes to second amplifier 1D for producing respiratory voltage V.sub.resp, an analog waveform of the chest wall movement. The signal output from first amplifier 1D is also passed to threshhold detector 1E which compares the amplitude of the respiratory waveform to a reference threshhold voltage, which may be fixed, or adjusted by the microprocessor. When the amplitude of the respiratory waveform is greater than the referenced threshhold, the output of the detector is driven to +5 volts. That +5 volts is inverted by invertor 2A to create digital pulse RSPTRG, corresponding to a respiratory breath. RSPTRG is then coupled to the system electronics of the oximeter at status latch 9G-11.

The respiratory activity is useful to monitor because, for example, many infants and small children have breathing problems and during their sleep have lapses in their breath. Monitoring chest wall breathing can check for such lapses or stoppage of breathing. In the preferred embodiment, the microprocessor creates a respiration history and establishes a regular pattern of breathing. Afterwords, if no breath is detected for a period of time, e.g., 15 seconds, then an alarm may be activated. Monitoring respiratory activity in conjunction with EKG enhanced oximetry can determine, when patient's blood flow characteristics drop during sleep, whether the reason for that drop was because of abnormal breathing, cessation of breathing, or some other cause. ##SPC1##