Description:
BACKGROUND OF THE INVENTION
The cathode ray tube was commercially introduced into medical monitoring in the nineteen fiftys. The display produced consisted in writing the electrocardiogram across the screen of the tube. In this type of display a narrow beam of electrons is directed onto a phosphor coated screen. Electrons striking the screen at a particular location cause the phosphor at that location to glow. Diversion of the electron beam from that location will cause the glowing phosphor to fade out. Thus in the conventional display a bright spot is seen to move across the screen at 25 mm/sec. moving vertically in direct proportion to the electrical signals from the patient's heart. This type of display has remained essentially unchanged. It suffers from the fact that while the phosphor does continue to glow slightly after the electron beam passes over it, in the ambient light levels of today's hospitals, the display is far from optimal in terms of readability and recognition of arrythmias.
Recently, a new type of display has appeared. This is the "infinite persistence" type display. Within this type of display, the electrocardiogram signal information is preserved at full brilliance on the screen for some predetermined amount of time (usually 1 to 10 seconds), giving the operator a clear view of arrythmias. A number of approaches have been made in this area. Among these have been (a) direct display of the data on a commercial storage cathode ray tube, (b) temporary storage of the data on a delay line and display on a television raster system, and (c) digitization of the data, temporary storage in a central computer, and conventional sweep display. Each of the above systems suffers from very definite drawbacks. Display of the data on a storage CRT suffers from two shortcomings in that the life of the storage tube itself is limited as compared with conventional CRT's, and the data itself is "printed" across the screen. This means that information appearing on the right side of the screen is erased as soon as it appears, and only information presented on the left side of the screen remains in view long enough for verification of possible arrythmias. These defects are eliminated by the raster display system. In this method the signal is passed through an acoustic delay line and stored there temporarily. As new data is fed into the line, old data (4-8 seconds) is removed. Thus the most recent history of the patient's ECG is retained. This memory is continually scanned and presented on a vertical raster display. This type of display uses a standard long-life CRT. The data itself is "pulled" across the screen. New data appears on the right side of the screen and moves slowly leftward across the screen, disappearing at the left edge of the screen. Thus, a given ECG complex remains visible on the screen for several seconds after occurrence.
The raster display system currently available also has certain disadvantages. The display itself is composed of a series of dots and has a discontinuity which is particularly noticeable in the QRS complex. The only available system is in the form of a central Nurses' station display. This system accepts conventional signals from bedside monitors, creates the nonfading display, and presents it to the central Nurses' display. Optionally, the persistent display may appear also at the bedside. This central processor concept allows for no failure margin. If the central delay line undergoes a major failure, the entire display system for up to four patients is affected.
The conventional swept display takes the data from the patient, converts it to a series of digital signals and stores it in a central computer. Like the raster display the data in memory is continually updated, and the data is "pulled" across the screen from right to left. Unlike the raster display, however, the data appears more continuous than segmented. The unit itself is once again central and a failure can affect up to eight patient displays both at the bedside and the Nurses' station, leaving the medical staff with no analog waveform display.
SUMMARY OF THE INVENTION
My invention embodies an apparatus and method for the storing of signals and eventual waveform display of the signals. The signals preferably are stored in at least two conditions, real time, and delay or alarm time. The real time signals are displayed in analog form on a recording medium. A waveform corresponding to the real time signal is analyzed to determine if it violates a preset condition or predetermined limits. If the limits are violated, then the stored alarm signals may then be displayed on the same or a different recording medium. Both the real time and alarm signals may be displayed on the same cathode ray tube, two different cathode ray tubes, on a cathode ray tube and an analog pen recorder, on two analog write-out devices, such as pen recorders, or in any other recording medium wherein the signals may be displayed in analog form, or may be sent in either analog or digital form to a computer for processing.
My invention is directed to an apparatus and method wherein the electrocardiogram signal (ECG) from the patient is stored in a memory and continually recycled and refreshed and displayed in analog form as a real time display. Preferably, new data appears on one side of a recording medium, such as a tube, and sweeps across the medium and is discharged a predetermined time later. All subsequent data follows the same pattern to provide a waveform which moves across the medium. The invention includes means to select an electrocardiogram signal, means to continuously display said signal, means to analyze the signal to determine if it falls within predetermined limits, means to store the signal as an alarm signal, and means to display the alarm signal.
More particularly, my invention is directed to a medical monitor for display of the electrocardiogram. Data from a patient is locally digitized, stored in a shift register, updated, and presented as a swept display. The patient is isolated from the monitor to eliminate fibrillation. A circuit analyzes a waveform corresponding to the ECG signal and determines if it violates certain limits. An alarm system is activated if the signal exceeds the established limits and the alarm condition is represented, visually and/or auditory. All of this is accomplished by a compact self-contained and independent monitor at the patient's bedside. In a group of eight (8) monitored patients, if a monitor fails only one patient's data is affected. The individual monitors can function without a central Nurses' station in the Emergency Ward, Operating Room, Cardiac Catheterization Laboratory, and as a mobile monitor in any part of the hospital. The invention includes the method of selecting an electrocardiogram signal, displaying continuously said signal for a predetermined time, analyzing said signal to determine if it falls within predetermined limits, storing the electrocardiogram signal as an alarm signal, and displaying said stored alarm signal as desired.
Briefly, certain features of the invention include a persistent or nonfading display, a totally floating input to assure patient isolation from the unit per se, a pattern recognition device for detection of the electrocardiogram signal to distinguish it from background noise, a digital display of the heartbeat updated on a beat-to-beat basis, and a time delay memory such as a five second auxiliary memory unit to freeze arrythmias or alarm conditions without disturbing the real time displays.
Signals from the patient are fed into an isolated front end wherein the selected electrocardiogram (ECG) signal is amplified and preferably filtered and optically coupled to the main frame of the instrument. The floating front end concept ensures that there is no electrical communication between the patient and the remainder of the system. This provides that in the event the patient is raised to some potential, such as by coming into contact with a source, there is no current path to ground through the patient and this prevents microshock. The front end is protected against defibrillation.
Once the signal has been transferred to the remainder of the system, it follows two paths.
To display the ECG signal, it is filtered to either a diagnostic or monitor band width depending upon the extent of analysis desired. After the signal has been filtered, it is amplified and then is fed into the display system. The signal passes to an analog-to-digital converter and is sampled a predetermined number of times, such as 200 times a second. Each time it is sampled a binary number, such as a seven digit binary number, is generated the value of which binary number corresponds to the amplitude of the sample. A plurality of these binary samples are stored in a real time shift register memory. Each time the ECG signal is sampled the binary number generated is entered into the memory displacing the number which represents the sample taken a predetermined sample time earlier. At the same time, the entire memory is being sampled at high speeds. The binary numbers are reconverted into the original signal and rapidly and repeatedly displayed on a cathode ray tube. In essence, the resultant display resembles an electronic analog pen recorder with the paper moving from right to left.
The second path which the signal follow is to a QRS waveform detector and then to a tachometer. The QRS detector itself is a pattern recognition device which detects the duration in time of the first excursion (be it plus or minus) taken in the QRS complex. The duration of this first excursion is approximately 10-40 milliseconds for both normal and infarcted hearts as nearly as has been determined. By allowing an appropriate window in time for the duration of this excursion, it can be extracted from background noise of similar amplitude. The tachometer includes a multiple beat rejection circuit which discards bogus beats caused by a patient motion. Such QRS appearing artifact is identified by the indication that it usually occurs in a salvo of rapidly successive peaks. This will not eliminate all artifactual beats but it is clearly an improvement over the present situation in commercially employed equipment. Once the beats are detected, the rate is obtained by inversion of the period between beats. This yields a continually updated rate. The rate is displayed digitally on the front panel of the instrument, and the rate voltage is sent to the alarm section of the unit.
In an alarm section of the unit, there are two comparators to compare the incoming rate voltage with high and low rate limits which are preset from the front panel. With the alarm enabled, violation of either limit will cause actuation of a signal, such as a buzzer, light, or both, each of which preferably correspond to the type of alarm detected; that is, exceedingly fast or slow heartbeat.
Another feature of the invention is an alarm memory in which data, preferably a predetermined time period old, such as five seconds, is continuously fed into the alarm memory. In the event of a violation of an alarm limit which is received on a delayed basis, this memory is sealed off. This captures the event causing the alarm, along with the data preceding and succeeding the alarm event. The information remains locked in this memory until the next alarm occurs unless cancelled or written out. The contents of the alarm memory may at any time be viewed by actuating a switch to display the alarm memory on the cathode ray tube which display preferably appears stationary and below the real time signal .
BREIF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a block diagram of the total invention;
FIG. 2 is a partially schematic and partially circuit diagram of the flotaing front end;
FIG. 3 is a block diagram of the filter;
FIG. 4 is a circuit diagram of the QRS wave detector;
FIG. 5 is a partially block and partially circuit diagram of the tachometer;
FIG. 6 is an illustration of the waveforms occurring in the tachometer;
FIG. 7 is a circuit diagram of the alarm system;
FIG. 8 is a block diagram of the audio oscillator circuit;
FIG. 9 is a block diagram of the real time memory;
FIG. 10 is a partially circuit and partially block diagram of the timing board employed with the real time memory;
FIG. 11 is a circuit diagram of the interconnection between the real time memory and the alarm memory including the timing for the alarm memory;
FIG. 12 is a block diagram of the alarm memory;
FIG. 13 is a circuit diagram of the multiplexer; and
FIG. 14 is a schematic illustration of the front panel of the instrument.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
The instrument is shown generally in FIG. 1 wherein a patient is shown at 10 and connected with the three-standard-limb leads or Eindhoven Triangle with additional sternal and/or auxiliary leads which communicate with a patient cable 12 which plugs into a patient cable jack 15 on the front panel of the instrument as shown in FIG. 14. The floating front end 14 is optically coupled to a filter-amplifier 26 shown in greater detail in FIG. 3. The signal is filtered and then is directed to a real time memory 200 shown in greater detail in FIG. 9, then to the multiplexer 340 shown in greater detail in FIG. 13, and then to the oscillograph of the cathod ray tube 230. The signal is also sent to the QRS detector 40 shown in FIG. 4, the tachometer 70 shown in FIG. 5, and the digital display 84. The signal from the tachometer is in communication with the alarm system 110 shown in FIG. 7; and if an alarm is actuated, it communicates with the alarm memory system 130 shown in FIG. 12.
FLOATING FRONT END
Referring to FIG. 2, the signal from the patient is selected by actuation of the multiple lead selection switch 16 which is also shown in FIG. 14 of the front panel of the instrument and may select any of various wet or dry electrode combinations from among the possible right arm (RA), left arm (LA), left leg (LL), and chest (C) combinations. The switch 16 as shown in FIG. 2 is in the 2 position to the RA-LL combination. The lead selection switch may also select a one millivolt calibration signal. Once the switch has selected the combinations of electrodes which will be chosen for display, the signal is amplified and filtered at 18 and optially communicated to the filter-amplifier of FIG. 3 by the optical coupler 24, such as a Monsanto Optical Isolator MCT-2.
The power supply 20 for the floating front end is a standard configuration zener regulated power supply. The grounds to the floating front end are identified as ##SPC1## . The supply 20 is coupled to a main power transformer (not shown) of the instrument which is actuated by the OFF-ON switch on the front panel. This communication is by a low capacitance 60 HZ power transformer 21, such as by a Stevens-Arnold isolation coupling No. IT-3408. To prevent a power surge a circuit 26 having neon tubes 28 (NE2H) is employed. When the voltage across them is less than 280 volts, the tubes form an open circuit. If voltage is 280 or greater, then the circuit 26 will act as a short circuit. With the tubes in the open circuit mode, there is no direct electrical communication between the patient and the main power ground of the instrument as shown.
Once the signal has been transferred to the remainder of the system, it will eventually go to the oscillograph of the cathode ray tube and also to the pattern recognition devices, the digital display, and the auxiliary alarm components as required.
FILTER-AMPLIFIER
The signal from the front end 14 of FIG. 2 is filtered as shown in FIG. 3. As mentioned in the discussion of FIG. 2, the signal through floating front end is initially filtered at the lower 3 DB point at 0.05 HZ. With the switch 36 on the diagnostic position, the signal is filtered at 29 with the upper 3 DB point at 100 HZ. The signal is amplified at gain adjust 30 and then goes to the real time memory of FIG. 9. Accordingly, when the switch 36 is on diagnostic, the signal will be filtered at the 3DB point with the upper limit at 100 HZ and the lower point at 0.05 HZ. After passing through filter 29, the signal then flows through gain adjust 30 wherein the signal may be amplified from 1 to 20 and then through switch 36.
As shown in FIG. 3, the signals to the QRS detector of FIG. 4 are always between the 1.0 and 40 HZ levels. It should be noted that the filter diagnostic settings correspond with the specifications of the American Heart Association for analog ECG data to be recorded.
QRS DETECTOR
Referring to FIG. 4, the QRS detector 40, a signal as shown enters the wave-shaping network shown at 42. The wave-shaping network is capable of producing two pulses whose waveforms are shown as +d/dt and -d/dt. The +d/dt and -d/dt pulses are then transmitted to circuit 44 which circuit produces a pulse corresponding in duration to ramp signals with slew rates in excess of +50 V per second or -50 V per second. Thus, the signal from the wave-shaping network 42 goes directly to gate 56 and into matching network 48.
The duration of the first linear rise or fall of the signal is approximately 10-40 milliseconds for both normal and infarcted hearts as nearly as has been determined. By allowing an appropriate window in time for the duration of this slope, such as from 10 to 40 milliseconds, it can be extracted from background noise of similar amplitude. Thus, a pattern recognition device for detection of the electrocardiographic signal is provided to distinguish it from background noise. When the pulse enters the compare network 48, an input pulse is entered into the network to determine whether or not the pulse received from the wave-shape network has a duration of at least 10 milliseconds. If the pulse received meets the minimum threshold level, 10 milliseconds, then a new pulse is generated by the network 48 which triggers the monostable multivibrator 52 which generates a new pulse of 30 milliseconds. This pulse in turn triggers the next monostable multivibrator 54 to generate a pulse of 0.1 milliseconds which is input to gate 56.
If the pulse from the wave-shaping network 42 which was transmitted directly to gate 56 ends prior to the pulse being received from the multivibrator 54, then in this condition gate 56 will be triggered to send a pulse to gate 60 when the pulse is received from the multivibrator. If the pulse from the multivibrator 54 is received by gate 56 prior to the end of the pulse directly from the waveforming network. then gate 56 will not be triggered and no pulse will be sent to gate 60. Therefore, if the pulse entering the compare network 48 is less than 10 milliseconds, no pulse will be received from the multivibrator 54 by gate 56 and thus gate 56 will not be actuated under any condition. If the pulse as compared is 10 milliseconds or more, then the multivibrator will transmit a pulse to gate 56. In order for the gate 56 to trigger a pulse to gate 60, the signal received by gate 56 directly from the pulse-shaping network must end prior to receiving the pulse from the multivibrator; that is, it must be 40 milliseconds or less. Thus, it is seen that if duration of the pulse initially received does not fall within the predetermined window or between 10 to 40 milliseconds, no further signal will be transmitted.
A similar pulse-shaping network, such as 44, to analyze and compare the -d/dt waveform has not been shown but functions in the same manner as the circuit 44 and feeds its input to gate 60. If the pulses from network 42 are between the limits set, then they will normally be received in quick succession by gate 60. In any event, even if they vary be a few milliseconds, any pulse received by gate 60 on either input will generate a pulse to trigger the nonretriggerable 100 millisecond delay 62.
Once the nonretriggerable delay has been actuated upon receipt of a pulse from gate 60, a pulse is generated shown as the QRS output to tachometer in FIG. 5. No new pulses will be generated for 100 milliseconds. This prevents bogus signals from being generated caused by possible overlapping of the signals from the gates.
The pulse is also sent to flasher 63 shown on FIGS. 1 and 12, and to the audio oscillator as will be descirbed in more detail in the description of FIG. 8. The flasher systole 63 is simply a light emitting diode, such as a Monsanto MV50, actuated upon receipt of a signal from the multivibrator 62. Any suitable visual indicator may be employed.
TACHOMETER
Referring to FIG. 5, the pulses from the QRS detector of FIG. 4 enter the tachometer; specifically, the pulse enters the "OR" gate 71 then to the sample and hold circuit 76 and the delay circuit 72, which triggers the ramp generator 74.
The pulse shapes are shown in FIG. 6 divided into three sections, 120 beats per minute (BPM), 60 BPM, and 20 BPM. Also shown in the 20 BPM section is the comparator output. Sample command from the gate 71 is shown as line a, the 100 microsecond delay pulse is shown as line b, the signal from the ramp generator is shown as line c, the sample and hold circuit output as line d, the divider output as line e, and smoothing circuit output as line f. As indicated, the sample and hold circuit selects that value of the slope from ramp generator 74 and provides and holds the waveform or voltage proportional to the period between pulses as shown in line 6d. Each time a pulse is received from the QRS detector, the ramp generator 74 is reset to zero by the 100 microsecond delay circuit 72. If no pulse is received from the QRS detector within the time period corresponding to the time it takes the ramp generator to reach a threshold value of 2.5 volts as determined by the comparator 74, the ramp generator will provide a signal to gate 71 to trigger a pulse corresponding in time to the 2.5 volt value, the sample and hold circuit thus generating a new waveform. The preset 2.5 volt value of the comparator corresponds to a heart rate of 20 beats/min. or approximately 3 seconds between beats. This is also shown in FIG. 6.
The pulse from the sample and hold circuit is received by the analog divider 78, wherein a one volt input at b provides the numerator for the division and the sample and hold output provides the denominator. The signal from the analog divider 78 is a voltage proportional to the frequency of the pulses. The output of the analog divider ranges from 0 V for 0 BPM to 5 V for 300 BPM. This signal is received by an amplifier and offset circuit 79 wherein the 0 to 300 BPM voltage range is changed to ±10 volts. From amplifier 79 the signal enters smoothing circuit 80 which is designed to prevent the transmission of any further signals to the analog to BCD converter unless they exceed certain limits. As shown, a short term change in BPM of ±11 or less, (±0.7 V) will not be reflected on the digital display 84. Thus, the smoothing circuit 80 prevents flicker. The one millisecond delay circuit 83 prevents entry of any signals into the converter 82 until the other components of this circuit have stabilized, particularly the sample and hold and ramp generator circuits. The converter 82 includes a gate which will transmit information to the display when signals are received from both the delay circuit 83 and the smoothing circuit 80.
ALARM
Referring now to FIG. 7, the alarm circuit is shown generally at 110 and receives a signal from the tachometer described in FIG. 5 which signal corresponds to the rate voltage sent to the digital display 84. Comparators 112 and 114 are preset through variable resistors 116 and 118 to determine the high and low limits of the rate voltage which should not be violated, such as a high limit of approximately 2 volts corresponding to 120 BPM and a low limit of approximately 0.6 volts corresponding to 40 BPM. The incoming signal from the tachometer is analyzed in the comparators. If neither condition is violated, no alarm signal is generated by gate 119. Violation of either limit will provide a signal and as shown will (a) actuate the front panel lamp, such as a MV-50, either tachycardia 120 or the brady-cardia 122 depending upon whether the upper or lower limit has been violated; (b) provide an audio signal, high or low, transmitted to the audio oscillator of FIG. 8; (c) transmit a further signal to the audio oscillator of FIG. 8; and (d) transmit a signal to the alarm memory of FIG. 12.
Two switches are shown, an alarm set-reset switch 128 and an alarm response switch 124, which includes a five second delay latch circuit 126. When in the normal position, switch 124 allows the signal to be transmitted instantaneously to the audio oscillator and the alarm memory. When actuated, the alarm response switch 124 will delay the alarm signal for five seconds by engaging the latch circuit 126.
Since it is possible for the unit to receive abberrant beats, this capability of selecting a five second delay to the alarm system is provided. If the alarm condition persists for five seconds, then the instrument will generate an alarm, but not before. The five second delay circuit may then be reset after the alarm condition has been checked through the switch 128 which also cancels the signal to the audio oscillator of FIG. 8. In instantaneous mode the alarm is reset automatically when the alarm condition ceases.
AUDIO OSCILLATOR
Referring to FIG. 8, a block diagram of the audio oscillator is shown wherein a QRS detector input line is shown receiving the signals from FIG. 4 which flow directly to an analog switch 91. From the switch they go to the mixer 92 where they are amplified through the speaker 95. The signals from the QRS detector provide the actual impetus for the sound corresponding to the heartbeat. A voltage controlled oscillator (VCO) systole sound is shown at 97 and receives a signal from the smoothing circuit of the tachometer of FIG. 5. This determines the frequency or the pitch at which the signal received from the QRS detector output will be heard. The alarm circuit pulses are received from low alarm, high alarm, and from the gate 119. A signal from high alarm will be transmitted to gate 94b. If a signal if received from gate 119 of FIG. 7, then the gate will actuate analog switch 96b which receives the characteristic sound from the phase shift oscillator 92b. The signal goes to the mixer 93 and then to the speaker 95 also shown on the front panel. The mixer is of the type which is adapted to receive and pass all signals simultaneously, as is well known in the art. It serves no other function than to allow the various signals to pass through a common point to the speaker 95. The VCO 97 may be adjusted to vary the volume of the systole sound to give an indication of the heartbeat, and the phase shift oscillators may also be adjusted to change the level of sound of the high and low alarms. These are shown on the front panel.
REAL TIME MEMORY
Referring to FIG. 9, the real time memory is shown generally at 200. The analog signal from FIG. 3 is received by the analog-to-digital converter where the analog is sampled and digitized each time a start pulse occurs, as will be explained later. As shown, this is 200 times per second. the a/d converter is of a standard downcounter design. The output of the a/d conversion is a 7 bit binary signal. Shift registers 204a-g, 206t-z, are each 512 bits long, and each have a binary input and a two phase clock input together with memory cells as with all commercially available shift registers, such as National Semiconductor MM5016. The clock input of the shift registers receives a continuous series of pulses at the rate of 204.8 KHZ/sec. from the clock 208. The digital signal appearing at the input of the registers will appear at the output of the registers after one-two-hundredth seconds or 1024 clock times. As shown, the output of the shift registers, for example shift registers 204a and 206t . . . 204g and 206z, are wired back to the inputs and data is continually passed back through the registers or recirculated. Switches SW-1 through SW-7 with continual clock pulses to the 1024 bit registers from the clock operate with the following cycles synchronized to the clock. For 1024 clock times, recirculation takes place, then for 1 clock time new data from the a/d converter 202 is entered into the register and the register and the cycle repeats. The result of this operation is that the registers are continually being updated with new information at the rate of 200 pieces of data per second. The output of the register is fed to a continuous d/a converter which produces an analog voltage from the 7 bit register 210.
The two phase clock 208, shown in greater detail in FIG. 10, provides two modes of operation during a clock time. The modes are respectively to read data and then shift data. On the read mode of the first clock pulse the data from the a/d converter is read or examined, and on the shift mode of the first clock pulse the data is entered into the first position of the register 204a. On the second clock time, the read pulse, new data is entered into the first position of the shift register while the old data shifts to the second position of the shift register. This continues on until completion of the 1024 clock time. On the 1025 clock time the original data is reentered into position 1. At the end of the 1025 time, the switch SW-1 is actuated to allow entry of new data. At the end of 1026 pulses or the first pulse of the second cycle, the original data is in position 2 and the new data is now in position 1.
Stated another way, referring specifically to registers 204a and 206t, which in series make up a 1024 bit register, a piece of data entering register 204a in the first position at time 0 will at time 1023 be in the 1024 position of shift register 206t. During clock time 1025 this bit of information will have been recirculated and entered shift register 204a in the second position of the 1024 positions available. At clock time 1025, switch SW-1 is actuated and new data is introduced into the first position of shift register 204a with the old data now in the second position of shift register 204a. With the 1025 clock times later the data in positions 1 and 2 will have recirculated through shift registers 204a and 206t and will have entered the second and third positions of shift register 204a. At clock time 2050 switch SW-1 is actuated and new data is entered into the first position of shift register 204a. From the above values it is seen that in approximately five seconds the first bit of information entered into the first position of register 204a will have recirculated approximately 1,024 times and be in the 1024th position of shift register 206t. The same sequence is followed for shift registers 204b-g and 206u-z.
Referring to FIG. 10, the two phase clock is shown generally at 208. The circuit generates complementary pulses, φ in and φ out. These voltages are alternately emitted at +5 (and when not at a voltage of +5, they are at a voltage of -12). As described above, these pulses are transmitted to the shift registers and correspond to the read pulse and shift pulse for each clock time. When the shift pulse goes to the shift registers, it also enters the level shifter shown at 240. This level shifter allows the 5 volt output to alternate between +5 and 0 volts. The output from the level shifter 240 goes to the a/d converter shown in FIG. 9, to the divide by 1025 circuit or register 242, and to gate 306b of FIG. 11. This circuit 242 is comprised of three divide by 16 counters 408a, b, and c. During clock time 1, the register reads zero. During clock time 2, the register reads one. During clock time 3, the register reads two, etc., until during clock time 1025, the register reads 1024. At the end of clock time 1025, the register reverts to the zero state. When the register reverts to zero, a signal is gated to pulse generator 244 which which actuates switches SW-1 to SW-7 to allow for the entry of new data into the shift registers. As shown the pulse generator has two outputs A and B. During clock times 1-1025 the output from A allows recirculation of data and the output from B inhibits entry of new data. When the signal is gated to the pulse generator 244, the outputs of A and B are reversed.
Referring to FIG. 11, the switching and command mechanism for the real time memory is shown in FIG. 11 for shift registers 204a and 206t, it being understood that similar command and switching mechanisms are duplicated for the remaining shift registers. The switching and command and the timing for the alarm memory are also shown. Gates 300a and 300b are synchronized with the clock times and upon completion the 1025 clock time, new data is entered into the first position of the 204a register and at the same time this new data is transmitted to the digital-to-analog converter 210 of FIG. 9. This condition occurs when the input to the 300a gate is 0 and the input to the 300b gate is 1. These inputs are 0 from A and 1 from B of the pulse generator 244. On the 1024 clock times, the input to the gate 300a is 1 and the input to the gate 300b is 0. This prevents the transmission of new data to the registers and allows recirculation of the present data to the shift registers. When the new data is accepted at the end of the 1025 clock times, the old data leaving the 1024 position of register 206t, goes to inverter 310, then to gate 302b and to D/A 210.
The alarm memory 320 comprises the same number of shift registers and of the same characteristics as of the real time memory 200. Thus, the data which is five seconds old is entered into the alarm memory 320. This five second old data is recycled in the same manner as the new data which enters the real time memory 200. New five second old data discharged from the real time memory 200 is entered into the alarm memory every 1025 clock times.
Shown in dotted lines is a two-phase clock 316 similar to clock 208 in that it provides θ in and θ out pulses for the alarm memory. The input to inverter 304 is from the level shifter 240 of FIG. 10. The two inputs to pulse generator 312 are from the alarm circuit and from the divide circuit 242 of FIG. 10. The input to pulse generator is from inverter 310 and from the pulse generator 244 of FIG. 10. The pulse generator provides two outputs A' and B' to gates 302a and 302b. For 1024 clock times recirculation in the alarm memory takes place and for 1025 clock times new data from the real time memory is entered.
When an alarm signal is entered into the gate 306b and inverter 310, the clock 316 provides the two-phase read and shift modes for the register of the alarm memory and 1024 clock pulses per cycle. Pulse generator 312 is inhibited and the values of A' and B' are held at +5 volts and 0 volts respectively whereby gate 302b is inhibited from allowing new data to enter the alarm memory and in essence isolates the alarm memory.
This condition remains until a view button is actuated at which time the information in the alarm memory will be displayed. The introduction of information from the alarm system does not affect the real time memory and that information is continuously displayed on the cathode ray tube.
Thus, it is seen that the data in the registers is read and displayed at each time (t) new data is entered. The data is displayed each time (t) while it moves up in position in the registers, the position in the registers corresponding to the position on the screen. The data will be displayed for a time t.n where n is the total times (t) it takes for the data to move from position 1 to 1024. This time n may vary depending upon the shift register used, i.e. the number of positions desired. That is, a time greater than or less than five seconds may be employed, say for example four seconds or six seconds.
In the preferred embodiment the five second old data from the real time memory is transferred to the alarm memory with an age of t.n recirculated n a times (where n a equals the total time (t) it takes for the data to move from position 1 to 1024 in the alarm memory) which in the example equals n and is discarded at time t.n 1 . Of course, the duration of recirculation in the alarm memory may vary and the transfer from the real time memory to the alarm memory may occur prior to the time t.n of the real time recirculation, at time t.n as in the example or subsequent to time t . n. Further, the total time to enter and be discarded from the register n a may also vary.
ALARM MEMORY
Referring to FIG. 12, the alarm memory is shown generally at 130 and includes a five second memory 132 into which memory data, which is five seconds old, is continuously loaded therein from the real time memory of FIG. 9; and when the data becomes ten seconds old, it is lost. The information is stored in the memory which is the same as shown for the real time memory 200. The timing pulses for the registers are derived from circuits as shown in FIG. 11; and the number and positions of registers, switches, etc., are the same, except, of course, there is no a/d converter since the five second old data is received in digitized form. Two discreet time delay circuits 134, a 7.5 second delay circuit, and 136, a 2.5 second delay circuit, are provided.
When the alarm response switch 124 of FIG. 7 is in instantaneous mode, the 7.5 second time delay circuit 134 is actuated. When in the instantanoous mode in the event of an alarm, the alarm signal to the delay circuit 134 does not enter memory 132 until 7.5 seconds subsequent to the receipt of the alarm signal. At the termination of the 7.5 seconds the memory 132 is sealed off as previously explained. This captures the event causing the alarm along with data 2.5 seconds earlier and later. It will remain locked in the memory 132 until the next alarm occurs or unless cancelled or written out, such as on the oscillograph of the cathode ray tube. When in the five second delay mode through switch 124 the 2.5 delay circuit is actuated to allow for the total 7.5 second delay in sealing the memory. Reset switch 124 resets the latch circuit 138.
MULTIPLEXER
Referring to FIG. 13, the multiplexer is shown generally at 340 and vertical input (FIG. 9) from the d/a converter main memory flows to amplifier 342 to field effect transistor switch 344 shown in greater detail in the dotted lines, then to the mixing, scaling and offset circuit 349 which simply comprises an operational amplifier configured as a weighted summing amplifier with dc inputs to bias the output from 358 or 360 and signal inputs for display from 344 or 348 as shown, and then to the vertical input to the CRT. Input from the alarm memory passes through amplifier 346 to field effect transistor switch 348 similar to FET switch 344 and then to the mixing, scaling, and offset circuit 349. Input from the clock 208 (see FIG. 9), specifically the level shifter 240, provides a blanking pulse to the JK flip-flop 350 whereby each time a blanking pulse is received this flip-flop changes state. NAND gate 352 receives inputs from the flip-flop 350 and view switch 354.
With the view switch 354 in the position shown, or OFF, it is grounded and the signal from the NAND gate 352 inhibits FET 348 and FET 360. The inverter provides a signal which turns on FET 344 and 358. When the view switch is OFF, the output from gate 352 is always a predetermined value, specifically in this instance 5 volts, which allows the input from the main memory to be displayed on the CRT in a conventional manner as shown in FIG. 14.
When the view switch 354 is actuated ON, the blanking input to JK flip-flop 350 goes to NAND gate 352. Thus, on each return sweep of the CRT which corresponds to the input of the blanking pulse FET's 344 and 358 and FET's 348 and 360 are alternately turned on and inhibited. With the view switch actuated ON and engaged with the +5 volt source when FET's 344 and 358 are on, circuit 362 is engaged whereby the signal to the scaling circuit is between +5 and +12 volts. When FET's 348 and 360 are on (344 and 358 being inhibited), circuit 364 is engaged whereby the signal transmitted to the mixing and scaling circuit varies between ground and -12. Therefore, when the view switch is actuated, the offset circuits 362 and 364 offset the displays of the real time and memory on the scope. Circuit 362 in combination with FET 358 offsets the real time display upwardly which is continuously displayed. Circuit 364 offsets the memory display downwardly. Since the information received from the alarm memory is the continuously recirculated data as previously described, no new data is introduced into the alarm memory until specifically cleared by actuating the set-reset button 124; and therefore, the memory data will appear as a permanent trace on the lower portion of the scope. This is also shown in FIG. 14 in the schematic illustration of the front panel.
The signals are displayed on the scope of a cathode ray tube, such as a Tektronix 560 Series, wherein the vertical input is from the mixing, scaling and offset circuit 349 (FIG. 13) and the horizontal sweep trigger input from the a/d converter 202 (FIG. 9). As shown in FIG. 14, the real time display during normal operation is across the center of the tube. When an alarm condition is displayed, it is frozen on the lower portion of the screen and the real time is displaced upwardly as shown in the dotted lines.
OPERATION
In the operation of our invention the power switch is actuated and the limit switches on the front panel of the device as shown on FIG. 14 are set to establish the upper and lower limits for the alarm circuit shown in FIG. 7. For example, the upper limit may be set at +2 volts corresponding to 120 BPM and the lower limit set at 0.6 volts corresponding to 40 BPM. The alarm system is set in the five second delay mode through actuation of the switch 124 which engages latch circuit 138 of FIG. 12, the band width filter for the electrocardiogram signal is set at diagnostic, the patient cable switch is set to lead 2 wherein the lead is attached to the right arm, left leg, and a reference lead to the right leg. The signal from the patient passes through the floating front end and is filtered at the diagnostic level and then passes to the real time memory of FIG. 9. As previously described, the signals from the patient are continually sampled, digitized, recirculated, passed to the multiplexer 340 of FIG. 13, and displayed on the scope of the cathode ray tube 230. As shown in FIG. 14, when the information in the alarm memory is not displayed, the real time signals appear as a moving trace from the right to left of the scope in the center of the scope as shown. A signal from the analog to digital converter 202 of FIG. 9 goes directly into the cathode ray tube as the horizontal sweep trigger input. The vertical input for the cathode ray tube is from the mixing, scaling, and offset circuit 349 of FIG. 13.
At the same time that the signal is transferred to the real time memory 200 and displayed on the scope of the cathode ray tube, it is filtered and passed to the QRS detector of FIG. 4. If the patient has a heartbeat of 80 BPM, the QRS detector will analyze the duration of the linear rise and fall of the signal and distinguish it from background noise. Assuming the heartbeat at this time is normal, the signal will be sent by the QRS detector to the flasher and sound system of FIG. 8 as previously described. The flashing light and audio signals are shown in FIG. 14 at 63 and 95. The signal entering the tachometer is converted to a digital representation of its value corresponding to beats per minute and is shown at 84 reflected as 80. The signal generated by the divider 78 then goes to the alarm circuit of FIG. 7. With 80 BPM the voltage would be approximately 1.3 volts. This signal is transferred to the comparators of the alarm circuit, which comparators were set at a high level of 2 volts corresponding to approximately 120 BPM and 0.6 volts corresponding to approximately 40 BPM. Since neither condition is violated, there is no signal generated by gate 119. Thus, while the heart rate in beats per minute is within the limits set by the comparators 112 and 114, no further signal is sent to the alarm memory and therefore, there is no change in the operation of the multiplexer and the signal is continuously displayed as described above.
Assuming that the patient now has a spell of sinus tachycardia, the signal to the filter and the real time memory 200 will continue as before and enter the multiplexer (FIG. 13) at 343. The horizontal sweep trigger pulse from the real time memory directly to the horizontal input of the cathode ray tube will also continue as before. With the heartbeat now 160 BPM the QRS detector circuit will examine the pulses to determine their slope. If they fall within the 10 to 40 millisecond window, then, of course, they are transmitted to the flasher and audio signal and to the tachometer. With a rate of 160 BPM entering the sample and hold circuit 76 (FIG. 5), a voltage output from divider 78 goes to amplifier and offset circuit 79, then to the converter 82 where a digital display reflecting the 160 BPM would be displayed. The voltage from the digital divider with a value of about 2.6 volts corresponding to the 160 BPM goes to the comparators 112 and 144 of the alarm circuit which have been set at 2 to 0.6 volts for the upper and lower limits. Thus, the 2.6 volt value would violate the upper limit of comparator 112 causing gate 119 to generate a signal. Also, the front panel lamp, in this instance the tachycardia 120, is actuated, a high alarm signal is transmitted to the oscillator of FIG. 8, and a further signal of the audio oscillator of FIG. 8 is transmitted, and the signal from gate 119 is transmitted to the alarm memory of FIG. 12. Since the switch 124 is in the five second delay mode, no signal will be received by the alarm memory from gate 119 until the condition has existed for five seconds. Therefore, as soon as the condition exists for five seconds, then, of course, the signal is transmitted to the alarm memory.
If no signal were received from the QRS circuit, then referring to FIG. 5, the sample and hold circuit would generate a 2.5 volt level which is approximately 3 seconds between heartbeats or a rate of 20 BPM or less. The ramp generator 74 of FIG. 5 when it reaches a threshold value of 2.5 volts as determined by the comparator 75 will provide a signal to gate 71 to trigger a pulse corresponding in time to the 2.5 volt value. This value, of course, of 0.4 volts to the alarm system would also exceed the preset limits and actuate the alarm system. Returning to the specific example, when in the five second delay mode the latch circuit 136 which provides a 2.5 second delay is actuated at the termination of the five seconds from when the alarm signal was first received as previously described. The data in the alarm memory is sealed off and remains in a recirculation mode or state. To view the display of the condition which caused the alarm the view switch of FIG. 13 is actuated and the real time display is displaced upwardly on the screen as shown in dotted lines in FIG. 14 and the display of the alarm condition appears on the lower half of the screen as a permanent display. After the data has been examined, then the information may be cleared by actuation of the set-reset button 124 of FIG. 7.