United States Patent [191
`Swedlow et al.
`
`111111111111111111111111111111111111111111111111111111111111111111111111111
`US005226417 A
`[11] Patent Number:
`[45] Date of Patent:
`
`5,226,417
`Jul. 13, 1993
`
`Introducing the Nellcor N-200 with ECG synchroniza(cid:173)
`tion".
`Nellcor pamphlet "N-200. Nellcor N-200 pulse oxime(cid:173)
`ter with C-LOCK ECG synchronization".
`"C-LOCK ECG Synchronization Principles of Opera(cid:173)
`tion", Pulse Oximetry Note Number 6, Reference Note,
`Nellcor Inc., 1988.
`
`Primary Examiner-Lee S. Cohen
`Assistant Examiner-Kevin Pontius
`Attorney, Agent, or Firm-Townsend and Townsend
`Khourie and Crew
`
`[57]
`ABSTRACI'
`An apparatus for detecting movement in patients cou(cid:173)
`pled to pulse oximeters and a method for using the
`signal generated by the apparatus to filter out the effects
`of motion from the test results generated by the pulse
`oximeter are disclosed. In a preferred embodiment, a
`piezoelectric fllm located in close proximity to the pulse
`oximeter's sensor provides a voltage signal whenever
`movement occurs near the sensor. This voltage signal is
`processed and the resulting signal is used to correct the
`oximeter's measurements. In addition to piezoelectric
`film, accelerometers and strain gauges are also usable to
`provide a signal indicative of motion.
`
`6 Claims, 3 Drawing Sheets
`
`[75]
`
`[54] APPARATUS FOR THE DETECTION OF
`MOTION TRANSIENTS
`Inventors: David B. Swedlow, Foster City;
`Robert L. Clark, Hayward; Adnan I.
`Merchant, Fremont; Deborah A.
`Briggs, San Ramon; Jessica A.
`Warring, Millbrae, all of Calif.
`[73] Assignee: Nellcor, Inc., Hayward, Calif.
`[21] Appl. No.: 667,152
`Mar. 11, 1991
`[22] Filed:
`[51]
`Int. Cl.s ................................................ A61B S/00
`[52] U.S. Cl ........................................ 128/633; 356/41
`[58] Field of Search ............... 128/633, 664, 665, 670,
`128/677,682; 356/40,41
`References Cited
`U.S. PATENT DOCUMENTS
`B1 4,653,498 4/1989 New, Jr ..
`4,802,486 2/1989 Goodman .
`4,830,014 5/1989 Goodman .
`4,869,254 9/1989 Stone .
`4,911,167 3/1990 Corenman .
`5,025,791 6/1991 Niwa .
`5,099,702 3/1992 French ............................. 73/862.68
`
`[56]
`
`OTHER PUBLICATIONS
`Nellcor pamphlet "Nellcor redefines pulse oximetry.
`
`19
`
`Apple Inc.
`APL1006
`U.S. Patent No. 8,989,830
`
`0001
`
`

`
`U.S. Patent
`
`July 13, 1993
`
`Sheet 1 of 3
`
`5,226,417
`
`19
`
`FIG.
`
`I.
`
`,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
`f* ' -21
`
`FIG. 2.
`
`0002
`
`

`
`-l
`~· .....
`...
`0'\
`N
`N
`...
`f.ll
`
`CN
`0 .....,
`N
`(1) a
`00 =(cid:173)
`
`CN
`~
`.....
`JN
`~ .....
`~
`
`~ a ~ = ~
`c •
`
`•
`rJJ.
`
`OUT
`MOTION
`
`r
`: 100
`
`I
`I
`I
`I
`~
`ISO
`I
`I
`
`'
`:
`.---------,
`
`.01 Jlf
`
`I
`I
`I
`
`I
`I
`
`ADJUST FOR 2.1 V OFFSET
`
`FIG. 3.
`•
`25K
`
`L------__ ....J +15V--'\IW---15V
`I
`
`V
`. •'T
`....... .V2LF 412
`
`22
`
`1)01
`
`FROM KYNAR
`
`FILM
`
`,
`
`:
`1
`I
`I
`I
`-----------~-----.
`
`,~
`
`22pf
`
`-15V
`
`i .IJ&f
`
`499K
`-15V
`
`•15V
`-15V +.IJif
`
`+15V
`
`/
`
`101
`
`.499K
`
`22pf
`
`151""'
`
`ANODE
`DETECTOR
`
`CAT HOD
`DETECTOR
`
`.47 Jlf
`
`1 5oK r
`
`,----. .J
`I
`:
`I
`
`0003
`
`

`
`""""'
`......:1
`~
`,.
`0\
`~
`~
`,.
`Ul
`
`~
`0 ......,
`~
`~
`(T)
`1:1"
`00
`
`'""'
`'""' ~
`~
`
`~
`\C
`\C
`
`~ = f""to. n> =
`
`f""to.
`
`•
`rJJ.
`~ •
`
`FIG. 4.
`
`X • X+l
`
`HISTORY
`-ADD TO
`RESET TIMER
`
`221
`
`YES
`
`---1
`
`YES
`
`213
`
`I YES
`
`211
`
`RESET TINER
`
`HISTORY
`-AOO TO
`
`PROCESSING
`.NO MOTION.
`
`SELECT
`
`NO
`
`YES
`
`207
`
`PROCESSING
`•MOTION•
`SELECT
`
`SIGNALS
`READ NEXT
`
`r------L---.._203
`
`,...--------.-201
`
`X •. I
`
`0004
`
`

`
`1
`
`5,226,417
`
`BACKGROUND OF THE INVENTION
`
`APPARATUS FOR THE DETECTION OF MOTION
`TRANSIENTS
`
`2
`digital optical signal that refers to the pulsatile compo(cid:173)
`nent is called the optical pulse.
`The detected digital optical signal is processed by the
`microprocessor of the N-200 to analyze and identify
`5 arterial pulses and to develop saturation. The micro(cid:173)
`processor decides whether or not to accept a detected
`This invention relates generally to non-invasive pulse
`pulse as corresponding to an arterial pulse by compar-
`monitors such as pulse oximeters. In particular, it relates
`ing the detected pulse against the pulse history. To be
`accepted, a detected pule must meet certain predeter-
`to the detection of motion transients and the flltering of
`these transients from the blood oxygen signals sent to 10 mined criteria, including the expected size of the pulse,
`when the pulse is expected to occur, and the expected
`the pulse oximeter.
`Phot~lectric pulse oximet~y is kno~n. Pulse oxime-
`ratio of the red light to infrared light in the detected
`ters typ•c:Ul.Y ~easur~ and display vanous blood_ flow
`optical pulse. Identified individual optical pulses ac-
`characten~tu::s mclu?mg the blood oxygen sat_ur~tl<;m of
`cepted for processing are used to compute the oxygen
`hemoglobm ~n arterial b~ood, the volume of mdlV1dual 15 saturation from the ratio of maximum and minimum
`blood pulsa~1ons supplymg _the flesh, and the rate of
`pulse levels as seen by the infrared wavelength.
`bl~ pulsat10n~ correspond~g to each heartbea! oft~e
`A problem with pulse oximeters is that the plethys-
`patlent: The OX1meters pass light thro~gh ~y t1ssue m
`mograph signal and the optically derived pulse rate may
`a locatiOn where blood pe~uses the t1ssue (1.e. a ~mger
`be subject to irregular variants in the blood flow that
`or an ear) and photoelectncally sense the absorpt10n of 20 interfere with the detection of the blood flow character-
`light in the tissue. The amount of light absorbed i~ then
`istics. For example, when a patient moves, inertia may
`u~d to calculate the amount of the blood constituent
`cause a slight change in the venous blood volume at the
`bemg measu~ed.
`.
`.
`sensor site. This, in tum, alters the amount oflight trans-
`Several d1fferent _wavelengths of hgh~ are srmulta-
`mitted through the blood and the resetting optical pulse
`neously or. nearly s1multaneously transm1tted through 25 signal. These spurious pulses, called motion artifacts,
`the bo~y t1ssue. !hese wavelengths are selected b~ed may cause the oximeter to process the artifact wave-
`on the1r absorpt10n by the blood _comp~nents be~ng
`form and provide erroneous data.
`measured. The amount of transm1tted light passmg
`It is well known that electrical heart activity occurs
`through the tissue will vary in accordance with the
`simultaneously with the heartbeat and can be monitored
`changing amount of blood c?nstitue~t in the tissue:
`30 externally and characterized by an electrocardiogram
`A:n example of a commerc1ally avallable. pulse OXlme-
`('ECG') waveform. The ECG waveform comprises a
`ter 1s the Nellcor Incorporated Pulse Oxrmeter model
`complex waveform having several components that
`N-200 (herein "N-200"). The N-200 is a microprocessor
`correspond to electrical heart activity. A QRS compo-
`controlled device that measures oxygen saturation of
`nent relates to ventricular heart contraction. The R
`hemoglobin using light from two light emitting diodes 35 wave portion of the QRS component is typically the
`steepest wave therein, having the largest amplitude and
`("LEOs"), one having a discrete frequency of about ~60
`nanometers in the red light range and the other ~avmg
`slope, and may be used for indicating the onset of cardi-
`a discrete frequency of about 925 nanometers m the
`ovascular activity. The arterial blood pulse flows me-
`infrared range. The N-200's microprocessor uses a four-
`chanically and its appearance in any part of the body
`state clock to provide a bipolar drive current for the 40 typically follows the R wave of the electrical heart
`activity by a determinable period of time that remains
`two LEOs so that a positive current pulse drives the
`infrared LED and a negative current pulse drives the
`essentially constant for a given patient.
`red LED. This illuminates the two LEOs alternately so
`One method to reduce or eliminate the effects of
`that the transmitted light can be detected by a single
`motion artifacts is to synchronize the ECG signal and
`photodetector. The clock uses a high strobing rate, 45 the optical pulse signal and process the two signals to
`roughly 1,500 Hz, and is consequently easy to distin-
`form a composite signal. This composite signal is then
`guish from other light sources. The photodetector cur-
`used to measure the level of oxygen saturation. This
`rent changes in response to the red and infrared light
`method is called ECG synchronization.
`transmitted and is converted to a voltage signal, ampli-
`In the first stage of synchronization, the optical pulse
`tied and separated by a two-channel synchronous detec- so signal is flltered to minimize the effects of electronic
`tor-one channel for processing the red light wave
`high frequency noise, using a low pass fllter. Next, the
`form and the other channel for processing the infrared
`oximeter positions the newly acquired optical pulse in
`light waveform. The separated signals are flltered to
`memory, using the QRS complex as a reference point
`for aligning sequential signals. In other words, when the
`remove the strobing frequency, electrical noise and
`ambient noise and then digitized by an analog to digital 55 QRS complex occurs, the oximeter begins processing
`the optical pulse data.
`converter ("ADC"). As used herein, incident light and
`transmitted light refers to light generated by the LEOs
`In the third stage, the new optical pulse signal is
`or other light sources, as distinguished from ambient or
`combined with the composite of the signals that were
`previously stored in the memory. Signals are combined
`environmental light.
`The light source intensity can be adjusted to accom- 60 using an adjustable weighted algorithm wherein, when
`the new composite signal is calculated, the existing
`modate variations in patients' skin color, flesh thickness,
`hair, blood, and other variants. The light transmitted is
`memory contents are weighted more heavily than the
`thus modulated by the absorption of light in the blood
`new optical signal pulse.
`pulse, particularly the arterial blood pulse or pulsatile
`Finally, t~e o~ygen sat_uration 17vel is ~easured fr~m
`component. The modulated light signal is referred to as 65 the compos1te s1gnal. This determmaton 1s on the rat1os
`the plethysmograph waveform, or the optical signal.
`of the maximum and minimum transmission of red and
`The digital representation of the optical signal is re-
`infrared light. As each sequential QRS complex and
`ferred to as the digital optical signal. The portion of the
`optical pulse signal are acquired, the process of filtering,
`
`0005
`
`

`
`5,226,417
`
`4
`3
`chrony. If the new optical pulse is identical to the com-
`positioning, combining and measuring saturation is re-
`posite pulse then the updated result is a composite opti-
`peated. As aperiodic signals such as motion artifacts
`cal pulse having the same magnitude. If the magnitudes
`will not occur synchronously on the ECG and the de-
`tected optical pulse, the effect of these aperiodic signals
`differ, the additive result will differ according to the
`5 relative weights.
`is rapidly attenuated.
`Another method to detect and reduce the effect of
`As a result of the collected, synchronized additive
`process, any aperiodic information that may be present
`motion artifacts involves correlating the occurrence of
`cardiovascular activity with the detection of arterial
`in the portions of the detected optical signals are also
`weighted and added to the weighted composite portion
`pulses by measuring the ECG signal, detecting the oc-
`currence of the R-wave portion of the ECG signal, 10 waveform. However, because aperiodic signals differ in
`pulse shape, duration, height, and relative time of oc-
`determining the time delay by which an optical pulse in
`the detected optical signal follows the R-wave, and
`currence within each portion, and are not synchronous
`using the determined time delay between the R-wave
`with heart activity, they do not add in phase. Rather,
`and the following optical pulse to evaluate arterial
`they add .in a cancelling manner whereby their
`blood flow only when it is likely to represent a true 15 weighted sum is spread across the relative time frame of
`blood pulse. The measured time delay is used to deter-
`the composite portion.
`mine a time window when, following the occurrence of
`By processing portions including the periodic infor-
`an R-wave, the probability of finding an optical pulse
`mation collectively, aperiodic information is attenuated
`corresponding to a true arterial pulse is high. The time
`by the absence of any corresponding historical aperi-
`window provides an additional criterion to be used in 20 odic signal in the prior composite portion or any subse-
`quent aperiodic signal at that relative time following
`accepting or rejecting a detected pulse as an optical
`pulse. Any spurious pulses caused by motion artifacts or
`heart activity. As the new information can be given a
`noise occurring outside of the correct time window are
`small weight compared to the absolute weight given the
`typically rejected and are not used to calculate the
`amount of blood constituent. Correlating the ECG with 25 prior composite, new aperiodic information is quickly
`and effectively attenuated and filtered out of the resul-
`the detached optical pulses thus provides for more reli-
`able measurement of oxygen saturation.
`tant additive portions.
`~though all of the. descri!'ed methods improve the
`Other methods to detect and eliminate the effects of
`~ualtty of the pulse o~eter s ~easurements by re~uc-
`patient motion have been developed. A time-measure of
`the detected optical signal waveform containing a plu- 30 ~g the effects of I?otton t~anstents and .ot~er ~punous
`stgn~s, they proVIde no mdependent mdtcat~on t~at
`rality of periodic information corresponding to arterial
`motto~ has oc~u~ed. Such an mdepe~dent verificatt~n
`pulses caused by the patient's heartbeat and periodic
`ofpatt~n~ motton IS ~seful for pulse o~etry. I.n certain
`information unrelated to pulsatile flow is collected, and
`casc:s, tt IS also posstble that ~ EC~ stgnal will not. be
`the collected time measure of information is processed
`to obtain enhanced periodic information that is closely 35 availa~le. In th~ cases, haVIng an ~dependent mot~on
`related to the most recent arterial pulsatile blood flow.
`de~ectton capabiltty would be essential to detect motton
`The time-measure may comprise a continuous portion
`artifacts.
`of detected optical signals including a plurality of peri-
`SUMMARY OF THE INVENTION
`odic information from successive heartbeats, or a plural(cid:173)
`A preferred embodiment of the present invention
`ity of discrete portions of detected optical signals ih- 40
`comprises a method and apparatus for minimizing the
`eluding a corresponding plurality of periodic informa(cid:173)
`effect of motion artifacts in pulse oximetry. Unlike
`tion.
`known methods, the present invention derives a motion
`By updating the time-measure of information to in(cid:173)
`detection signal independently of the pulse signal. Al(cid:173)
`clude the most recently detected aperiodic information,
`though the present invention will be described relative
`and processing the updated measure collectively, an 45
`to its use in pulse oximetry, its usefulness is not limited
`updated enhanced periodic information is obtained (in(cid:173)
`to that area alone.
`cluding the new and historical data) from which aperi(cid:173)
`A preferred embodiment of the present invention will
`odic information (including any new aperiodic informa(cid:173)
`be described in connection with an adhesive fmger
`tion) is attenuated. In some embodiments, the updating
`sensor for use with a pulse oximeter. Other sensors may
`process includes subtracting detected signals older than 50
`be used, however, without departing from the scope of
`a certain relative time from the collected time-measure.
`the invention.
`By collectively processing a time-measure including
`In an adhesive finger sensor for a pulse oximeter, a
`successive periodic information to obtain the enhanced
`strip of piezoelectric film has been incorporated. The
`periodic information, and using the enhanced periodic
`film covers the nearest movable joint to the sensor; in
`information as the basis for making oxygen saturation 55
`this example, the joint on the fmger to which the sensor
`calculations, the accuracy and reliability of oxygen
`is attached. The change of strain on the motion sensing
`saturation determinations can be significantly increased.
`element caused by moving the finger to which the sen(cid:173)
`The time-means may be collectively processed in either
`sor is attached generates a charge within the element, as
`the time domain or the frequency domain.
`in a capacitor. A gain resistor mounted across the mo(cid:173)
`By synchronizing the occurrence of successive R- 60
`tion sensing element bleeds off the charge, thereby cre(cid:173)
`waves, it becomes possible to add the corresponding
`ating a voltage signal that is proportional to the rate of
`successive portions of the detected optical signal to(cid:173)
`bending.
`gether so that the periodic information (optical pulses)
`By properly processing this voltage signal, motion
`corresponding to the arterial pulse in each portion will
`artifacts can be detected and their effect on the calcula(cid:173)
`add in phase. The weighted magnitude of the new peri- 65
`odic information is reinforced by the existence of the
`tion of blood oxygen compensated for.
`The invention will now be described in detail, with
`weighted enhanced periodic information at the same
`time location in accordance with the degree of syn-
`reference to the figures tested and described below.
`
`0006
`
`

`
`1;
`
`FIG. 3 is a schematic of the preamplifier used in the
`presentinvention; and
`FIG. 4 is a flow chart showing how the present in(cid:173)
`vention processes motion transient signals.
`
`5,226,417
`
`5
`BRIEF DESCRIPTION OF THE DRAWINGS
`FIG. 1 shows an assembled sensor according to the
`present invention;
`FIG. 2 is a cross-section of the sensor shown in FIG.
`
`6
`sensing element is made from KYNAR fllm, a product
`of Atochem, Inc. The motion sensing element extends
`across the sensor head from one butterfly wing end 18
`through and beyond the other end 11, into a tab 27 that,
`5 in the preferred embodiment, is disposed over the first
`joint of a fmger, on the dorsal side, when the sensor is
`applied to the patient.
`The change of strain on the motion sensing element
`(such as by bending the mm strip) generates a charge
`10 within the element, as in a capacitor. A gain resistor
`mounted across the motion sensing element bleeds off
`DESCRIPTION OF THE SPECIFIC
`the charge, thereby creating a voltage signal that is
`EMBODIMENT(S)
`proportional to the rate of bending. The size of the gain
`A preferred embodiment of the motion detection
`resistor may be varied to permit differently dimensioned
`sensor of the present invention is shown in FIGS. 1 and 15 oximeter sensors (with differently dimensioned motion
`2. Sensor package 10, which includes light-emitting
`detection elements) to be used with the same oximeter.
`diodes 13 and 16 and photodetector IS, is for transillu-
`In the preferred embodiment, the gain resistor is
`mination of a blood perfused portion of flesh to measure
`mounted in the sensor connector.
`light extinction during transillumination. The sensor is
`preferably mounted on a fmgertip but any digit or other 20 cu An :lec~ri~ cableh 23todprtovtideslSthe dLEDt. driving
`rren an re urns P 0
`e ec or
`an mo ton sens-
`blood perfused tissue will work. The sensor conforms to
`ing element 19 signals to the oximeter. In the preferred
`and with the cutaneous layer of the blood perfused
`portion of flesh upon which the sensor is placed. A first
`embodiment, the cable contains three shielded, twisted
`end 11 of sensor 10 is disposed on one side of the flesh
`pairs of conductors, one pair each for the detector, the
`to be transilluminated and a second end 18 is disposed 25 emitters and the motion sensing element. The cable's
`on the opposite and opposed side of the flesh to be
`inner shield is coupled to the photodetector's Faraday
`transilluminated.
`shield. Both the outer and inner cable shields are tied to
`When the sensor is adhesively fastened, the effect of
`analog ground. All wires are terminated in the sensor
`the light source and photodetector being integrated into
`connector.
`.
`.
`.
`.
`the adhesive fastener is that they become, in effect, a 30
`I~ t~e prefe~ed embodtment, an emttter ~mg rests-
`tor ts _mclu~ed m the sensor connector. As 1s more fully
`part of the skin. The resulting device is resistant to
`accidental removal and avoids constriction of blood
`explamed m U.S. Pat. No. 4,621,643, the value of the
`vessels. Most importantly, the low mass of the sensor
`coding res.istor is relate~ to the operating wavl!lengt~s
`of ~he emttters. ~e o~eter read~ the value ~f thts
`itself and its conformance to the skin prevents motion
`and the possible resulting contact interruption between 35 reststo~ to dete~e which coeffictents to use m the
`the light source, photodetector and flesh.
`saturauon calculatton.
`In the present invention, as illustrated in FIGS. 1 and
`In the preferred embodiment, the sensor connector is
`2, the dimensions of the butterfly-shaped bandage con-
`plugged into. the front end of a custom preamplifier.
`taining the sensor are such that the butterfly "wings"
`!fle preamplifi7r ~ay be :xtemal to the oximeter or
`(ends 11 and 18) do not extend beyond the first joint of 40 mcorporated wtthin the OXtmeter.
`the patient's fmger when the sensor is attached to a
`As shown in FIG. 3 preamplifier 100 comprises a first
`patient. Bandage layer 21 is preferably an adhesive
`section 101 to amplify the photodetector signal used to
`cotton elastic material "which completely covers opaque
`compute oxygen saturation and a second section lSI to
`white polypropylene layer 14. Holes are formed in
`condition the motion detector's output.
`opaque, adhesive coated polypropylene layer 14 for the 45
`First section 101 comprises a differential input ampli-
`optical components. A clear, double-coated 0.003 thick
`fier with an approximate gain of 1 million. This requires
`polyethylene layer 12 covers these holes.
`the sensor to be configured in a differential mode with
`The LEDs 13 and 16, as well as photodetector IS are
`shielded twisted pair conductors. No offset voltage is
`placed beneath layer 12. Photodetector IS is mounted
`provided for dynamic range improvement but could be
`on lead frame package 26 and is surrounded by Faraday so added. The output of the differential amplifier is trans-
`mitted to the pulse oximeter in known fashion.
`shield 17. The LEDs, photodetector and Faraday shield
`are all coupled to the pulse oximeter by means of leads
`As stated previously, KYNAR piezoelectric film
`running through cable 23. LEOs 13/16 are commer-
`element 19 can be modeled as a capacitor. When a strain
`cially available and are mounted in a lead frame pack-
`is placed on the film, a charge is produced. The output
`age 29. The red wavelength LED generates at least 0.85 55 of the film is proportional to the rate of change of the
`milliwatts and the I.R. LED generates 1.45 milliwatts of
`strain and it is A. C. coupled. To use this charge, a resis-
`power. In an alternate embodiment, the lead frame
`tor 1S2 needs to be coupled in parallel with the fllm.
`packages 26 and 29, photodetector 1S and LEDs 13 and
`The value of this resistor affects the voltage sensitivity
`16 are mounted on a flexible substrate 25. In the pre-
`of the fllm, which simply means that different sensor
`ferred embodiment, opaque layer 14 and clear layer 12 60 geometries need to be tuned with different resistors.
`The voltage signal from the fllm/resistor combina-
`are peanut-shaped to provide adequate coverage of the
`optical components, wires and motion sensor. The pea-
`tion is then passed through a unity gain, second order
`Butterworth fllter ISS with a cut-off frequency of 10 Hz
`nut shape also provides sufficient surface area to adhere
`to the butterfly without subsequent delamination and
`to reject line noise pickup. The band-limited signal is
`minimizes assembly time.
`6S then amplified in amplifier 160 by a factor of 33,000
`Motion sensing element 19 is a strip of piezoelectric
`along with an inserted (adjustable) offset of 2.1 volts.
`ftlm placed between the optical components and the
`The selection of the gain is arbitrary, based on obtaining
`bandage layer 21. In the preferred embodiment, the
`"reasonable" output for typical motions. The offset was
`
`0007
`
`

`
`5,226,417
`
`25
`
`40
`
`55
`
`7
`8
`filter. The final answer becomes the new N. Note that as
`added to place the A. C. coupled output approximately
`N rises, the effective filtering decreases.
`in the middle of a 5 volt ADC input range.
`When the oximeter enters the "motion present" state
`In the preferred embodiment, the N-200 is modified
`(step 205), pulses which do not conform to the "his-
`to receive the conditioned motion signal through an
`unused channel of an ADC. The optical pulse signal is 5 tory" accumulated prior to entering the "motion pres-
`ent" state are not accepted. This prevents the oximeter
`sampled at 57 hz, the ECG signal at 200 hz, and the
`motion signal at about 57 hz. The N-200 software is
`from mistakenly accepting false pulses caused by roo-
`modified to read this additional ADC channel and pro-
`tion artifacts which pass other criteria! checks em-
`cess it along with the optical and ECG information.
`ployed by the oximeter after the first four bad pulses
`Collection of the optical and ECG signals is not 10 have gone by. Additionally, it prevents the oximeter
`from building up a history consisting of false pulses
`changed.
`Referring now to FIG. 4, the oximeter detects the
`caused by motion artifact which would then prevent the
`presence of motion by subtracting from the baseline
`N-200 from accepting good pulses once the motion
`signal of the motion signal at step 203, after it has been
`artifact ceases.
`conditioned to remove background noise, taking the 15 Also, the oximeter uses a higher "N" value in the
`absolute value of the result and entering a "motion
`flltered ratio calculation for accepted pulses (step 207).
`This change permits the oximeter to use tighter ftltering
`present" state at step 207 whenever the processed signal
`passes a fixed threshold as determined at step 205. In the
`on data during the motion present state, while allowing
`the instrument to return its normal response time when
`preferred embodiment, the optimum threshold was de-
`termined empirically to be 1.22 millivolts. The oximeter 20 motion is not present. Finally, the oximeter employs a
`45 second pulse time-out period (step 207), as compared
`leaves the "motion present" state 1.5 seconds after the
`to the 15-20 second time out used when motion is not
`processed signal falls below the threshold.
`Entering a "motion present" state at step 207 changes
`present (step 213) before triggering an alarm indicative
`of loss of pulse in the patient.
`the way the optical signals are processed and, therefore,
`the way blood oxygen saturation is calculated. Outside
`The foregoing description provides a full and com-
`of the "motion present" state (step 213), the oximeter
`plete disclosure of the preferred embodiments of the
`calculates blood oxygen saturation in any known appro-
`invention. Various modifications, alternate construc-
`priate manner. In the preferred embodiment, the oxime-
`tions, and equivalents may be employed without depart-
`ter maintains a history (step 221) consisting of the mean 30 ing from the true spirit and scope of the invention. For
`values over four consecutive pulses of three parameters
`example, although only the use of a piezoelectric film to
`as part of the saturation calculation algorithm: the per-
`provide motion detection has been described herein,
`iod between successive optical minima, the IR optical
`other motion detection means such as accelerometers,
`pulse amplitudes, and the "ratio-of-ratios". The period
`or stain gauges could be substituted without changing
`and amplitude information is displayed by the oximeter. 35 the substance of this application. Therefore, the above
`"Ratio of ratios" is used in the saturation calculation
`description and illustrations should not be construed as
`limiting the scope of the invention which is defmed by
`and is defined as follows:
`the appended claims.
`What is claimed is:
`1. A sensor for attaching to a patient for electro-opti-
`cal measurement of at least one blood characteristic,
`comprising:
`optical signal means for generating a first electrical
`signal indicativ~ of the at least one characteristic of
`the blood in a portion of the patient's tissue;
`a piezoelectric fllm;
`signal processing means, coupled to said piezoelectric
`mm, for generating a second electrical signal indic(cid:173)
`ative of movement in and of the portion of the
`patient's tissue; and
`means for transmitting the first and second electrical
`signal to an instrument for determining the blood
`characteristics.
`2. The sensor of claim 1 wherein the signal processing
`means comprises an electrical impedance means cou(cid:173)
`pled to the piezoelectric fllm.
`3. The sensor of claim 1 wherein the signal processing
`means further comprises an electrical impedance means
`coupled to the piezoelectric film, the value of the elec(cid:173)
`trical impedance means indicating the geometry of the
`piezoelectric film.
`4. A system for measuring a blood characteristic of a
`patient comprising:
`a sensor comprising:
`optical means for generating a first electrical signal
`indicative of a characteristic of the blood in a
`portion of the patient's tissue;
`a piezoelectric film;
`
`Incoming pulses are checked against the history, and
`pulses are rejected if they are outside the permitted
`limits of variation (step 215). The first four pulses re- 45
`jected for variation excess are not placed into the pulse
`histories (step 217 and 219). Once four pulses are re(cid:173)
`jected for this reason, subsequent pulses are placed into
`the history at step 221 to permit the history to reflect
`changing physiological conditions. If the pulse is ac- 50
`cepted, a time-out clock is reset. The time-out clock
`normally sounds an alarm if no qualified pulse is de(cid:173)
`tected within 15-20 seconds.
`Before using the ratio-of-ratios in the saturation cal-
`culation, it is ftltered as follows:
`
`L Red max
`0 Red min
`IR max
`L
`0
`IR min
`
`Filtered Ratio=unfiltered ratio •(N/256)+filtered
`ratio •(256- N)/256,
`
`where 1 ~N~255 and N varies according to pulse rate 60
`and amplitude. For the first 5 pulses after locking onto
`the optical pulse, use N=255. After the first 5 pulses,
`calculate N for each pulse. The initial N varies depend(cid:173)
`ing upon the type and physiology of the patient. If
`locked on ECG, multiply the result by 3. If the rate is 65
`greater than 100, divide the result by 2. If the average
`IR amplitude is small, divide the result by 2. Filter the
`result against the previous result using a i old, i new
`
`0008
`
`

`
`5,226,417
`
`9
`signal processing means, coupled to said piezoelec(cid:173)
`tric film, for generating a second electrical signal
`indicative of movement in and of the portion of
`the patient's tissue; and
`means for transmitting the first and second electri- 5
`cal signals to an instrument for determining a
`blood characteristic;
`means for receiving the first and second electrical
`signals from the sensor;
`first processing means for operating on the second
`electrical signal for generating a signal indicative of
`motion; and
`
`10
`second processing means for operating on the first
`electrical signal and the signal generated by the
`first processing means for determining a blood
`characteristic.
`5. The system of claim 4 wherein the signal process(cid:173)
`ing means comprises an electrical impedance means
`coupled to the piezoelectric film.
`6. The system of claim 4 wherein the signal process(cid:173)
`ing means further comprises an electrical impedance
`10 means coupled to the piezoelectric film, the value of the
`electrical impedance means indicating the geometry of
`the piezoelectric film.
`·
`• • • • •
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`0009