ABSTRACT

An electronic simulator of cardiorespiratory signals for testing and validating infant apnea monitors has been designed, built, tested, and applied. This device can simulate up to four signals that were as inputs to infant apnea monitors including: (1) transthoracic electrical impedance, (2) electrocardiogram, and (3) chest and abdominal inductance plethysmograph signals. Up to 17 minutes of four channels of actual monitored signals can be stored in ROM and played back in real time to simulate signals coming from sensors on patients.

INTRODUCTION

Cardio-respiratory (apnea) monitors are used in the home as well as the hospital in the care of infants. It is important that these devices function pmperly since they may be able to detect apparent life-thteatening events. Test equipment to evaluate the function of these monitors is, therefore, an important adjunct to the monitor. Electronic simulators are used to produce signals that can replace those coming from sensors on infants and can be used to detemiine whether a monitor is able to detect and alarm upon seeing specific cardiorespiratory pattems. Commercially available simulators produce artificial signals such as illustrated in Figure 1a. In this example of an apnea, a pause is seen in what is otherwise a regular, uniform pattern. Actual clinical data that is seen by the monitor looks more like the example shown in Figure 1b. Note that in this case breaths do not have regular amplitude or frequency, and noise often accompanies apneas. A monitor that can respond to the signal in Figure 1a may not necessarily respond appropriately to the signals seen in Figure 1b. Thus, the problem addressed by this work was to develop a simulator capable of producing biologically representative data to test apnea monitors.

Fig. 1. Simulated (a) and actual (b) breathing waveform in infant apnea.

METHODS

Our approach to this problem was to develop a simulator that stored actual physiologic data in memory and played this back rather than produce artificial, nonrepresentative wavefoms. The approach to the design of this device was to select representative patterns from actual signals recorded from infants and to store these signal segments in a read only memory (ROM). These signals were stored as a group of events. A microcontroller reads a sequence of these events according to its program. Individual events can be repeated as many times as desired depending on how the microcontroller is programmed. The respiration information read from memory by the microcontroller is passed through a digital-to-analog converter. The resulting analog signal thus produced is used to control an appropriate circuit for producing a change in the variable being measured by the monitor.

Respiration data from infants was collected using the sensors that were being simulated was mpled at a rate of 20 Hz with 8 bit resolution. A set of events were selected and written to the PROM.

RESULTS

A simulator capable of providing a transthoracic impedance signal, two inductance plethysmograph signals, and an electrocardiogram having an adjustable rate has been designed, fabricated, and evaluated. A block diagram of this circuit is shown in Figure 2. The main memory consists of a 512 KByte EPROM memory which stores the actual clinical data. An Intel 80196 microcontroller with its program stored in ROM reads the data stored in memory and produces an analog signal proportional to the variable stored in memory using a built in digital-to-analog converter. Three such signals canbe simultaneously produced. At present the simulator is configured such that three respiratory signals are stored in memory. One is the transthoracic electrical impedance wavefom. This signal is applied directly to a voltage-to-impedance converter circuit that provides the impedance signal to the monitor under test in place of the signal coming from the electrodes on the patient.

The other two channels provide chest and abdominal inductance plethysmograph signals. In this case the output from the digital-to-analog converter is fed into a circuit that converts this changing voltage into a changing inductance that corresponds to the inductance seen from the chest and abdominal belts on an infant.

The simulator also provides an electrocardiogram that is fed to the monitor under test to determine the heart rate. Because there was not enough space in memory to store the actual electrocardiogram, a different approach was taken. A typical QRS complex from an infant electrocardiogram was stored in memory at a sample rate of 50 Hz. A trigger signal was also stored in memory along with the respiration data such that it indicates when each heart beat occurs. The QRS complex stored in memory is then activated for each beat. Thus, the monitor sees an electrocardiogram with real QRS data that reproduces the heart rate information seen from the patient at the same time that the respiration data was recorded.

The simulator can be used as a two or four-channel signal source. In the former case only the transthoracic impedmce and the electrocardiogram channels are used, and up to 35 minutes of continuous data can be stored. In the four channel configuration, the chest and abdominal inductance plethysmograph signals are included with the other two signals. In this case the memory is capable of holding 17 minutes of continuous data.

Fig. 2. Block diagram of the simulator.

DISCUSSION

This work has shown that is possible to design a simulator for transthoracic impedance, inductance, and heart rate signals that presents physiologically realistic wavefoms to test infant cardio-respiratory monitors. The advantage of using such a system is that the monitors can be tested using actual physiologic waveforms rather than artificially produced approximations of these wavefoms. This makes it possible to test the monitor response to signals with varying respiration rates, cardiogenic artifact, motion artifact, and other types of noise frequently encountered when monitoring infants.

This simulator is currently being used to check monitors applied in the Collaborative Home Infant Monitor Evaluation to make sure that all of the monitors used in this nationwide study will respond to signals in the same way so that the data collected with different monitors can be pooled. By using signals recorded from actual infants, monitor function when connected to patients can be better simulated and evaluated.

This work was supported by the CHIME Cooperative Agreement number UIO-HD29071 of the National Institute of Child Health and Human Development, Bethesda, MD.

*Currently with Medtronic, Inc., Minneapolis, MN

**The C.H.I.M.E. Study group is a group of investigators at 6 academic facilities and the NICHD who are carrying out the Collaborative Home Infant Monitor Evaluation.