WO1996028093A1

WO1996028093A1 - Apparatus and process for reducing the frequency and duration of apneic events

Info

Publication number: WO1996028093A1
Application number: PCT/US1996/003278
Authority: WO
Inventors: David Paydarfar
Original assignee: St. Elizabeth's Medical Center Of Boston, Inc.
Priority date: 1995-03-09
Filing date: 1996-03-07
Publication date: 1996-09-19
Also published as: AU5188696A

Abstract

Methods and apparatus for reducing the frequency and duration of apneic events in a subject are provided. The methods involve detecting the frequency and/or duration of apneic events and delivering a stimulus to the subject to reduce the frequency and/or duration of the apneic events. The methods and apparatus are useful for reducing the incidence of sudden infant death syndrome.

Description

APPARATUS AND PROCESS FOR REDUCING THE FREQUENCY AND DURATION OF APNEIC EVENTS

Government Support

The subject matter of this application was supported at least in part by grant number HL49848 from National Institutes of Health. The government has certain rights in this invention.

Field of the Invention

This invention relates to an apparatus and method for reducing the frequency and/or duration of apneic events. More particularly, the invention relates to an apparatus and method for preventing sudden infant death syndrome.

Background of the Invention

Sudden infant death syndrome (SIDS) refers to the sudden and unexplained cessation of respiration (apnea) in infants. SIDS is the leading cause of death in infants between two months and one year of age. Although the precise etiology of SIDS is unknown, it is generally accepted that repeated apneic events, as well as prolonged apneic events, are associated with an increased risk of irreversible brain damage and a reduced life expectancy. Apnea monitors have gained widespread acceptance in the medical community as the only practical solution for preventing SIDS. Such devices monitor infant respiration and, by activating an alarm upon detection of an apneic event, waken the infant or alert a nearby attendant or parent to come to the infant's aid. An exemplary apnea monitoring system is described in U.S. Patent No. 4,580,575, issued to Birnbaum et al. The Birnbaum device monitors respiration and heartbeat, converts these parameters into electrical signals and feeds the signal(s) into an appropriate circuitry which actuates an alarm upon detection of an apneic event.

Despite the widespread use of conventional apnea monitors, the incidence of SIDS has not been significantly reduced. In many instances, false alarms have desensitized or exhausted parents to the extent that parents have ignored or slept through real apnea-induced alarms and infants have died. As a result of such deaths, recent efforts have focussed on developing apnea monitoring systems which function independently of human intervention. Such monitoring systems include, for example, mechanisms for mechanically stimulating (wakening) an infant into a normal respiration rhythm and/or mechanisms for providing inhalation therapy to prevent sudden infant death syndrome.

Devices for mechanically stimulating a subject who has experienced an apneic event are reported in U.S. Patent Nos. 4,619,270, issued to Margolis et al. and 5,277,194, issued to Hosterman et al. The Margolis et al. patent describes a crib attached to a moveable frame that is operatively linked to an apnea monitor. Upon detection of an apneic event, the moveable frame is vigorously shaken until the infant wakens and normal breathing resumes. Loud noises and alarms supplement the shaking stimulation and are used to startle the child into a normal breathing rhythm. The Hosterman et al. patent describes a vest that is worn about the chest of the subject. The Hosterman et al. device is intended to stimulate the respiratory reflexes to resume normal breathing in the event of an apneic event. The Hosterman et al. device optionally further includes an alarm that is activated upon detection of an apneic event.

Devices employing mechanisms for inhalation therapy to stimulate normal breathing are reported in U.S. Patent Nos. 4,484,578 and 4,414,982, issued to Durkan; U.S. Patent No. 5,052,400, issued to Dietz; and U.S. Patent No. 4,813,427, issued to Schlaefke et al. The Durkan patents describe a device for delivering to a subject at least one dose of a respirating gas following detection of an apneic event. The device further includes a visual and/or audible alarm for alerting a nearby parent or attendant. The Dietz patent describes a device which delivers a prescribed dose of a therapeutic gas to the subject in response to the detection of an apneic event. Similarly, the Schlaefke patent describes an apnea monitoring device for administering a therapeutic gas to a subject in response to detection of an apneic event. The therapeutic gas is delivered in combination with an additional stimulus that is intended to restore normal respiration.

The above-cited patents disclose apnea monitoring systems for detecting apneic events having a preset duration. In general, the above-described monitoring systems provide mechanical stimulation or inhalation therapy, coupled with an alarm system, for waking the subject or alerting an attendant or parent. Such systems do not entrain the underdeveloped nervous system of an apnea-prone infant to develop a normal respiration cycle. Moreover, none of the cited patents disclose a monitoring system which is capable of detecting the frequency and/or duration of apneic events. Accordingly, there is still a need for methods and devices to detect apneic events in an infant and to stimulate, without waking, the infant into a normal respiration cycle. Such methods and devices would permit the entrainment of abnormal respiration to an environmental stimulus that is selected to induce normal respiration in the infant. Su marv of the Invention

The present invention provides an apparatus and method for reducing the frequency and/or duration of apneic events in a subject. The apparatus includes (a) at least one detector for detecting the frequency and/or duration of an apneic event in the subject; (b) at least one effector for delivering a stimulus to the subject, the stimulus having at least one adjustable stimulus parameter; and (c) a processor that is operatively coupled to the detector and the effector. The processor is actuated by the frequency and/or duration of the apneic event to perform a series of functions (described below) which result in delivering to the subject a stimulus that is selected to reduce the frequency and/or duration of apneic events in the subject. A normal respiratory period includes inspiratory and expiratory cycles and varies as a function of age. For example, the normal respiratory period for a premature infant is approximately one second (ranging from about 0.8 to about 1.3 seconds). The normal respiratory period for a sleeping infant between two months and six months of age is approximately 1.6 seconds. (See, e.g., Haddad G.G. et al., Maturation of ventilation and ventilatory pattern in normal sleeping infants, J. Appl. Physiol. 46:998-1002, 1979) for a discussion of the variation of normal respiratory cycle with age.) .An "apneic event" refers to the cessation of breathing for a period that exceeds the expected respiratory period. Obstruction of the upper airway is the most common cause of apnea in preterm infants under the age of thirty weeks. Older infants have more apnea of the non-obstructive variety. Both types of apnea are much more common during sleep and can result from failure of the neural control of respiratory movements. Clinically, apneic events present as an abnormally long expiratory cycle.

Apneic events can be divided into apneic pauses (cessation of airflow for between about 2 seconds and about 10 seconds) and apneic arrests (cessation of airflow for at least about 10 seconds). As used herein, the frequency of apneic events refers to the number of apneic pauses and apneic arrests that occur within a preselected time period. The duration of an apneic event refers to the length of time (typically measured in seconds) during which respiration is not detected. Apneic events having a duration of greater than about fifteen seconds are considered to be life-threatening in the absence of intervention.

The apparatus of the invention includes a detector for detecting the frequency and/or duration of apneic events. In general, the detector employs methods and/or devices that are known to those of ordinary skill in the art for gathering respiratory information (e.g., respiration rate, sleep state, cardiac rhythm, blood oxygen tension, blood carbon dioxide tension and temperature). Accordingly, the detectors of the invention measure one or more of the foregoing physiological parameters which, alone or in combination, are indicative of a subject's respiratory state.

Respiration rate typically is detected by measuring: (i) chest and/or abdominal wall movement (e.g., using pressure transducers or by measuring resistivity or inductance), (ii) airflow (e.g., by measuring expiration and/or inspiration), (iii) transdiaphragmatic pressure (i.e., the esophageal pressure differential across the diaphragm at the gastroesophageal junction, and/or (iv) blood oxygen and/or carbon dioxide tension. Other physiological parameters that can be measured using one or more detectors of the apparatus include, for example, sleep state (e.g., measure via surface electroencephalograms), cardiac rhythm (e.g., measure via electrocardiogram) and temperature.

A stimulus is administered to the subject to reduce the frequency and/or duration of apneic events experienced by the subject. As used herein, "stimulus" refers to an environmental signal that can elicit or evoke a response in an excitable tissue that results in an alteration in the respiratory pattern of the subject. Preferably, the stimulus evokes a response in a receptor (e.g., an auditory-, mechano- or pain-receptor) that is sufficient to transduce the environmental signal to the subjects' nervous system and induce normal respiration. Exemplary categories of stimuli that are useful to achieve the purposes of the invention include auditory, tactile (e.g., vibration or temperature changes to the skin), vestibular (i.e., rocking), lung-inflating (airflow) (e.g., continuous or intermittent positive airway pressure, "CPAP" or "IPAP"), laryngeal, pharyngeal, swallowing-inducing, light, and olfactory stimuli. Intermittent positive airway pressure can be achieved by applying a single puff of air, having a specified pressure, for a specified duration or by applying a positive airway pressure in a pulsatile fashion as described herein with respect to auditory stimuli. The preferred stimulus is one which combines auditory and tactile stimulation^ amounts that are sufficient to reduce the frequency and/or duration of apneic events but which are insufficient to arouse the infant from sleep.

The stimulus has at least one adjustable parameter having a dimension (value) that can be selected to deliver a stimulus having the desired attributes for reducing the frequency and/or duration of apneic events without waking the infant, i.e., a dimension is selected for each stimulus parameter to fall within a therapeutic window of dimensions. Thus, the therapeutic window comprises the range of stimulus parameter dimensions that is sufficient to reduce at least one of the frequency and duration of apneic events without awakening the subject.

The lower edge of the therapeutic window is defined as the lowest dimension of a stimulus parameter that results in a statistically significant reduction in the frequency and/or duration of apneic events. Further increases in the dimension of this parameter result in a further reduction in the frequency and/or duration of apneic events until the upper edge of the therapeutic window is reached. The upper edge of the therapeutic window, i.e., the dimension of the parameter which causes the infant to waken, is determined empirically (by observing the dimension of the stimulus amplitude which causes the infant to waken). The lower edge of the therapeutic window is determined by the processor or by the clinician. For example, the lower edge of the therapeutic window can be determined by programming the processor to increase the stimulus parameter dimension from zero and to record the lower edge as the smallest stimulus parameter dimension which resultsin a statistically significant reduction in the frequency and/or duration of apneic events. Alternatively, the clinician can determine the lower edge of the therapeutic window by observing the smallest stimulus parameter dimension that results in a statistically significant improvement in respiratory rhythm, i.e., which reduces the frequency and/or duration of apneic events. The processor uses the lower edge dimension (inputted by the clinician or determined by the processor) and the upper edge dimension to establish the therapeutic window. The three general categories of stimuli (continuous, pulsatile and trigger pulse) which define the stimulus protocol of the present invention are discussed in detail below. Each of these stimuli includes adjustable parameters of amplitude, frequency and duration. The stimulus frequency has a dimension that leads to activation of receptors, e.g., an 800 Hz tone or "white noise" with thousands of superimposed frequencies. This stimulus frequency is set, regardless of whether the stimulus is pulsatile or continuous. The period of the pulsatile signal is relevant only for a pulsatile stimulus, and is the frequency at which the pulse stimulus occurs. The period of the pulsatile signal is selected to be an integer ratio of the respiratory frequency. Thus, for example, an integer ratio of 1 : 1 means that a pulse is given for each breath, an integer ratio of 2: 1 means that there are two breaths for every pulse, and an integer ratio of 3:2 means there are three breaths for every two pulses.

The normal respiratory frequency of the subject is determined by the processor, using the respiratory information gathered by the above-described detector(s) (e.g., chest wall movement, airflow). Preferably, the normal respiratory frequency of the subject is established by monitoring the respiration of the subject for preselected periods of time (e.g., between about fifteen seconds to one minute; preferably, for about thirty seconds). The processor subtracts out apneic events which may have occurred during this observation period before calculating the normal respiratory frequency. Thereafter, the processor uses the subject's normal respiratory frequency during the observation period to select the stimulus frequency. For a continuous stimulus, for example, the processor selects a stimulus frequency that approximates that of the normal respiratory frequency. For a pulsatile stimulus, for example, the processor selects a stimulus frequency that is an integer ratio of the normal respiratory frequency.

In a particularly preferred embodiment, an auditory and/or tactile stimulus is delivered to the subject to reduce the frequency and/or duration of apneic events. Preferably, the frequency of the auditory/tactile stimulus is in the range of about 1 Hz to about 10 KHz. More preferably, the frequency of the auditory/tactile stimulus is in the range of about 20 Hz to about 1 KHz. Most preferably, the frequency of the auditory/tactile stimulus is in the range of about 40 Hz to about 800 Hz. The delivery of a continuous auditory/tactile stimulus of about 50 Hz to a premature infant resulted in a statistically significant reduction in the frequency and duration of apneic events, without waking the infant (see, e.g., Example 1 and Figures 5-7).

The processor also selects the dimension of the stimulus amplitude (or other parameter) which falls within the therapeutic window of the amplitude dimensions, i.e., the amplitude dimension ranges from a minimum value (lower edge dimension) that is sufficient to reduce the frequency and/or duration of apneic events to a maximum value (upper edge dimension) which is less that the amplitude dimension that would waken the infant. The amplitude dimension that would waken the infant is determined empirically for each subject and is entered into the processor prior to treatment. The minimum amplitude dimension can be determined by the processor or can be determined by the clinician as described above. An effector is used to administer (deliver) the stimulus to the subject. Effectors vary depending upon the nature of the stimulus being delivered. For example, effectors for delivering an auditory stimulus can include headphones or speakers placed near the ear of the subject or incorporated into the subject's mattress. Speakers that are incorporated into a mattress can be used to deliver an auditory/tactile stimulus to the subject, depending upon the frequency selected for stimulation (see, e.g., Example 1). Other exemplary effectors include: use of a mattress which can provide pure vibratory or percussive stimulation, e.g., using pneumatically driven devices (Dynamic Air Therapy, Hill-Rom, Inc., Charleston, SC), electromagnetic devices (electric motor or solenoid that vibrates with various adjustable frequencies), or waves created in water mattresses (Korner A.F. et al., Reduction of apnea and bradycardia in preterm infants on oscillating waterbeds: a controlled polygraphic study, Pediatrics 61:528-533, 1978); automatic rocking cribs (Sammon M.P. and R.A. Darnall, Entrainment of respiration to rocking in premature infants: coherence analysis, J. Appl. Physiol. 77:1548-1554, 1994); breathing bears (Ingersoll, E.W. and E.B. Tho an, The breathing bear: effects on respiration in premature infants, Physiology and Behavior 56: 855-859, 1994); swallowing-inducing stimulus and pharyngeal stimulus achieved, for example, by injecting small amounts of fluid in the hypopharynx; CPAP and IPAP achieved, for example, by using a device (an "INCA" unit) sold by Ackrad

Laboratories, Inc. Cranford, New Jersey; application of a puff of air into the nose without a seal, so that it is not under positive pressure (a "nasal cannula") (Salter Labs, Arvin); and direct electrical stimulation of skin receptors achieved, for example, using a TENS (transcutaneous electrical nerve stimulation) apparatus (EMPI, Inc. Minneapolis, Minnesota).

The processor, which is operatively coupled to the detector and the effector, is actuated by the frequency and/or duration of an apneic event to: (i) establish a therapeutic window of dimensions for one or more stimulus parameters; (ii) select at least one stimulus parameter having a dimension within the therapeutic window for delivery to the subject; (iii) instruct the effector to deliver the stimulus having the stimulus parameter of step (ii) to the subject; and (iv) repeat steps (i), (ii) and (iii) until at least one of the frequency and the duration of the apneic event in the subject is reduced. In a particularly preferred embodiment, the apparatus further includes an alarm mechanism that is operatively coupled to the processor. The processor activates the alarm mechanism to produce an alarm signal if an apneic event is not terminated after a predetermined period of time (generally on the order of about 15 seconds to about 30 seconds). The alarm provides a safety feature for alerting an attendant or parents to come to the infant's aid in the event that an apneic event is not terminated by application of the stimulus. Optionally, the alarm mechanism is configured to waken the infant in addition to alerting the attendant or parents.

According to yet another aspect of the invention, a method for reducing the frequency and/or duration of apneic events in a subject is provided. The method involves (a) detecting the frequency and/or duration of an apneic event in the subject, wherein detecting such events actuates the steps (b) through (d); (b) establishing a therapeutic window for a parameter of a stimulus; (c) selecting a stimulus parameter having a dimension within the therapeutic window; (d) delivering to the subject the stimulus having the stimulus parameter of the dimension selected in step (c); and (e) repeating steps (b), (c) and (d) until the frequency and/or duration of apneic events is reduced in the subject. These and other aspects of the invention, as well as various advantages and utilities will be more apparent with reference to the detailed description of the preferred embodiments and in the accompanying drawings.

Each of the references, patents and patent publications disclosed in this document are incorporated in their entirety herein by reference.

Brief Description of the Drawings

Figure 1 is a general block diagram of the system of the present invention; Figure 2 includes timing waveforms representing respiratory movement and the three categories of stimuli according to the stimulus protocol of the present invention;

Figure 3 is a graph illustrating the therapeutic window range of dimensions (values) for a stimulus parameter according to the invention;

Figure 4 is a table listing average respiratory periods during quiet sleep and REM sleep in 15 infants of different ages (estimated from Haddad G.G. et al., Maturation of ventilation and ventilatory pattern in normal sleeping infants, J. Appl. Physiol. 46:998-1002, 1979);

Figure 5 shows examples of the effects on respiratory rhythm of turning the stimulus on (Fig. 5 A) and stimulus off (Fig. 5B);

Figure 6 shows an example of the respiratory pattern seen with stimulus off (Fig. 6A) and the last 6 min 20 seconds of a 15 minute stimulation trial (Fig. 6B), showing continued regularity in rhythm; and

Figure 7 shows the effect of stimulation on respiratory period (Fig. 7 A) and number of breaths (Fig. 7B) for all trials, compared to the periods without stimulation.

Figure 8 show the critical effects of noise (n) on rhythmicity of the BvP model. Noise converts non-oscillatory system to one that exhibits spontaneous cycles (A, B). The Oscillatory system (C-G, H-J) exhibits increased dysrhythmia with certain levels of noise further increases above this critical level results in decreased dysrhythmia. Unit bars: y=l, t=15.

Figure 9 shows the histograms of T_b for different amounts of noise in the BvP model (z=-0.340). Noise levels (n): A. 0.05, B. 0.02, C. 0.003, D. 0.001, E. 0.0003, F. 0.00006. The noiseless cycle has period=13, T,=4, T_b=9.

Detailed Description of the Invention The system of the present invention monitors the respiratory rhythm of an infant and provides a stimulus to the infant during periods of predefined breathing irregularities in order to stimulate and reinstate a normal respiratory pattern. The system is a feedback system that continuously monitors the respiratory rhythm of the infant and periodically alters characteristics of the stimulus to normalize the infant's respiration. A preliminary clinical examination of the infant is made before use of the system to collect necessary data that preferably is input to the system, as described in greater detail below.

Figure 1 is a block diagram showing the system of the present invention as used on an infant 10. The system includes at least one detector 12 for collecting respiratory information from the infant. The respiratory information is provided along a line 14 to a central processor 16 which monitors the respiratory information. The system also includes at least one effector 20 for providing a stimulus or stimuli to the infant. The processor 16 causes the effector 20 to provide a stimulus by sending a stimulus control signal to effector 20 along line 18. Processor 16 also may provide an alarm control signal along line 22 to cause alarm 24 to provide an alarm signal to awaken the infant and/or parents or others in response to predefined respiratory conditions.

As used herein, a "detector" is a device for gathering respiratory information from an infant. The respiratory information preferably includes the respiratory rhythm (rate) of timing periods of inspiratory and expiratory respiratory cycles (explained below) as well as periods of apneic events. Many different types of detectors, described below, are available for gathering the respiratory information from the infant. Preferably, electrical signals (either analog or digital), representing the respiratory information, are provided by detector 12 along line 14 to processor 16 which monitors the respiratory information.

During periods of normal respiration, no stimulus is provided to the infant. A normal respiratory pattern is predefined for the particular infant based either on the age of the infant or on information gathered during a previously performed clinical examination of the infant, or both. Data representing a normal respiratory pattern for a particular infant can be input to the processor 16 by a clinician.

In response to either (1) a single apneic event of duration greater than a predefined duration (i.e., an apneic arrest), or (2) a frequency of apneic events (each of a predefined minimum duration such as an apneic pause) over a predefined time period that is greater than a predefined frequency, the processor provides a stimulus control signal to the effector to cause theeffector to provide a stimulus to the infant. As used herein, "stimulus" refers to an environmental signal that can elicit or evoke a response in an excitable tissue that results in an alteration in the respiratory pattern of the subject. Preferably, the stimulus evokes a response in a receptor (e.g., an auditory-, mechano- or pain-receptor) that is sufficient to transduce the environmental signal to the subjects' nervous system and induce normal respiration.

The duration of the single apneic event that would trigger the stimulus application preferably is great enough to be of some health concern to the infant (i.e., an apneic pause) but not too great to be imminently dangerous to the health of the infant (i.e., an apneic arrest). Such a duration range, as well as the predefined frequency and predefined time period (discussed above), preferably are determined before system use, based either on the age of the infant or on data gathered during clinical examination of the infant, or both. Information representing such parameters are input to the system by a clinician. The system is a feedback system that continuously monitors the respiratory information

(e.g., respiration rate), provided by the effector(s) and can periodically update the stimulus to achieve a normal or "best" respiratory pattern. At any time after the application of the stimulus to the infant, one or more parameters of the stimulus may be adjusted by the system depending on whether improvement in the respiratory pattern is detected. A "parameter", as used herein, is any adjustable characteristic of a stimulus such as, but not limited to, frequency, amplitude and duration.

As an example, assume that a stimulus is applied to the infant in response to a frequency of apneas (each of a minimum predefined duration) over a predefined time period that is greater than a predefined frequency. If, after a predefined time period subsequent to the provision of the stimulus, the frequency of apneic events is reduced, then a parameter of the stimulus may be adjusted, such as increasing the amplitude of the stimulus, to effect a further reduction in the frequency and/or duration of apneic events. If, after the predefined time period subsequent to the provision of the stimulus, however, the frequency and/or duration of apneic events increases, then the same or another parameter of the stimulus may be adjusted to reverse the apneic event frequency and/or duration increase. Alternatively, another type of stimulus altogether can be applied. This system is thus a "smart" feedback system that can periodically update the parameters of the stimulus or stimuli until a normal or "best" respiratory pattern or rhythm is achieved. If, at any time, a normal respiratory pattern occurs, then the stimulus application can be reduced decre entally (for the predefined time period) and as long as the respiratory pattern remains normal, the stimulus can continue to be reduced and eventually (provided that the respiratory pattern remains normal for at least about 30 to 60 minutes), can be turned off. In this manner, the stimulus parameter automatically is adjusted to the minimum level required to minimize the frequency and/or duration of apneic events without wakening the infant. If a single apneic events occurs of a predefined duration that is potentially dangerous to the health of the infant (i.e., an apneic arrest), then the processor 16 causes the alarm 24 to provide an alarm signal to awaken the infant and/or parents or others. The alarm signal preferably is an audio signal provided by one or more speakers at a volume sufficient to awaken sleeping individuals. The predefined duration of the apneic event that is potentially dangerous to the health of the infant may depend on the age of the infant or on data gathered during clinical examination of the infant, or both. Information representing this predefined duration preferably is input to the system by a clinician.

The need for the system of the present invention as well as its effective use can be shown and modeled using an attractor cycle mathematical model, as will be understood by those skilled in the art (see, e.g., U.S. Patent No. 5,167,228, issued to Czeisler et al ). For example, the Bonhoeffer-van der Pol (BvP) equation can be used to model the system of the present invention and to illustrate that the provision of a stimulus to an infant can help regulate the respiratory rhythm of the infant and that the increase of a parameter of the stimulus further aids in regulating the respiratory rhythm of the infant.

Figure 8 illustrates the remarkable influences of noise (n) on rhythmicity of the BvP model. Noise can convert a non-oscillatory system to one that exhibits spontaneous cycles. In run A, there are no spontaneous oscillations (z=-0.335) and a steady state is reached, reflecting a stable singular point. Run B shows that addition of noise (n=0.02) causes the system to exhibit spontaneous regenerative cycles with spontaneous episodes of dysrhythmia.

In the oscillatory BvP system, increasing noise leads to increases in the incidence and duration of dysrhythmia up to a critical level of noise. Further increases in noise above this critical level results in reduction in dysrhythmia. Runs C-G illustrate the critical effects of noise in the BvP oscillator with a stable singular point (z=-0.3400). A noise level of 0.001 results in permanent attenuation of rhythm; the tiny fluctuations reflect small orbits around the singular point. Runs H-J show similar effects of noise even for the BvP oscillator with an unstable singular point (z=-0.3450), although the dysrhythmias are much less prolonged than those associated with a stable singular point.

Figure 9 shows the effects of changing noise on the cycle periods in the BvP oscillator with a stable singular point (z=-0.340), using 6 different noise levels. The histograms represent the distribution of times (T_b) spent below an activity threshold (y,) midway between the maximum and minimum y value (transecting the attractor in half) For each noise level, computation proceeded for 75,000 time units, roughly equivalent to 5,800 cycles of the noiseless rhythm (cycle period = 13 time units) with the same parameters, and analogous to 6-8 hours of real breathing. The histograms show several features of the effects of noise on periodicity: (1) criticality, as illustrated in the shorter runs of Figure 8; (2) skew distribution of the variability in T_b for a given noise level; (3) asymmetry of noise effects on T_b below and above the critical noise level, i.e., below the critical noise level, a small increase in noise causes transition of the pattern from very regular to highly dysrhythmic, whereas above the critical noise level, reduction in dysrhythmia by increases in noise are gradual rather than abrupt. Histograms of the times above the activity threshold (TJ for the same set of runs show very little variability.

The BvP model, or other suitable equation(s) of an excitable system that exhibits attractor-cycle dynamics, can also be used to improve the accuracy and speed of the processor to arrive at the optimum therapeutic window. For example, for continuous stimulus protocols, the processor can generate surrogate respiratory waveforms for different surrogate stimulus parameters. Examples are shown in Figure 8. The histogram representing the distribution of various cycle periods is determined for each stimulus (Figure 9), and stored in the memory of the processing unit. The surrogate waveforms and stimulus parameters are in the non-physiological units of the BvP equation. However, in the real infant, stimulus parameters and distribution histograms of respiratory periods are acquired over time, and the surrogate data from the BvP equation can be "normalized" to each infant's response to stimulation. The potential advantage of this approach is to enable the processor to arrive at an optimum therapeutic stimulus more quickly than might be arrived without the use of the surrogate data. This approach is analogous to improving statistical power (i.e., by allowing the processor to determine whether and to what extent the stimulus should be increased or decreased) once the parameters of the population (distribution of respiratory periods vs. stimulus parameters) are known. By generating surrogate distributions, the ability of the processor to make decisions as to whether a respiratory pattern has changed are improved, thereby making the system faster and more accurate. This advantage is analogous to that known by statisticians, namely, that parametric statistics (statistics on distribution with known properties) are "more powerful" than non-parametric statistics (statistics on samples coming from populations that cannot be characterized or are not known).

The stimuli can take many forms, described below, of which most fall into three selectible categories according to a stimulus protocol of the present invention. The three categories of stimuli are shown as timing waveforms in Figure 2 at time lines B, C and D. Also shown at time line A is a timing waveform representing the respiratory rhythm 26 of a contrived infant. The waveform 26 is shown on a horizontal time axis wherein periods of normal respiration are illustrated as a somewhat distorted sine wave. Waveform 26 includes periods of respiratory inspiration T, and periods of respiratory expiration T_E, wherein the total respiratory period T_τoτ equals T, + T_E. Included also are a number of apneic events, the first occurring of the apneic events having a period T_A. Shown on Figure 2 at time lines B, C and D are waveforms 28, 30 and 32 respectfully representing the three general categories of stimuli which define the stimuli protocol of the present invention. The three signals are shown on the same time axis with one another and with the respiratory rhythm waveform for illustrative purposes. The relative frequencies and amplitudes are not drawn to scale and the illustration of one signal above another does not denote that one achieves a higher amplitude or magnitude.

Shown at time line B is a signal 28 representing a first category of stimuli that is "continuous". Waveform 28 is a continuous time signal having adjustable parameters of amplitude, frequency and duration (the time period during which the stimulus is delivered to the subject). The frequency of continuous stimulus signal 28 typically falls within an audible frequency range of 20Hz-lKHz depending on the type of stimulus but may have a frequency spectrum that is, at least in part, greater than lKHz. The continuous stimulus signal can be characterized by a single frequency or multiple frequencies within the audible range or with some frequencies in the purely tactile range of stimulation (<20 Hz). A broad spectrum of frequencies would be characterized as "white noise," as will be understood by those skill in the art. The second category of stimuli shown in Figure 2 at time line C is a pulsatile signal 30 which is a periodic signal, typically of varying amplitude, having adjustable parameters of amplitude, frequency, and duration. Like the continuous time stimulus, the pulsatile stimulus typically has a frequency spectrum within the audible range, depending on the type of stimulus. The pulsatile stimulus is aimed at entraining the respiration of the infant to obtain a normal respiratory pattern. The pulsatile stimulus therefore preferably has a period T_PULS that is equal to, or is an integer ratio of, the normal respiratory period T_τoτ of the infant calculated over a most recent predefined time period. For example, an integer ratio of 1 : 1 means that a pulse is given for each breath, an integer ratio of 2: 1 means that there are two breaths for every pulse, and an integer ratio of 3:2 means there are three breaths for every two pulses. Preferably, T_PULS = K x T_TOT, where K is an integer. After application of a pulsatile stimulus signal, the respiratory information is monitored by the system and the period of the pulsatile stimulus may be updated at each predetermined period of time to reflect any most recent change in the normal respiratory period of the infant's breathing. The normal respiratory period includes periods of inspiratory and expiratory cycles but excludes any apneic events.

For certain types of stimuli, the pulsatile category of stimulus has predefined exclusionary periods. For example, if the particular type of stimuli induces inspiration, then the pulsatile stimulus is prevented from being applied during an exclusionary time window between the inspiratory and expiratory cycles. Conversely, if a particular type of stimuli induces expiration, then the pulsatile stimulus is prevented from being applied during an exclusionary time window between the expiratory and inspiratory cycles. Data representing these exclusionary time windows preferably is input to the system by a clinician prior to use. The third category of stimuli of the stimulus protocol of the present invention, is shown by waveform 32 at time line D of Figure 2. Waveform 32 represents a trigger pulse stimulus of a short predefined duration used to "jump-start" a normal respiratory pattern. Like the other categories of stimuli, the trigger pulse has a frequency spectrum generally within the audible range, depending on the type of stimulus. The trigger pulse includes adjustable parameters of frequency, amplitude and duration.

The adjustable parameters of the stimuli preferably are limited by predefined minimum and maximum dimensions that together define a parameter range referred to herein as the "therapeutic window". Figure 3 is a graph showing the therapeutic window range of dimensions of a stimulus parameter. Typically, as the parameter dimension of a stimulus is increased, the respiratory rhythm of the infant improves (i.e., becomes more normal). Also, when the parameter dimension of the stimulus is increased beyond a certain value, the infant typically is awakened. Figure 3 shows the magnitude of the stimulus parameter on the horizontal axis and the frequency of apneic events or duration of a single apneic event on the vertical axis.

The therapeutic window includes the dimensions (values) of the stimulus parameter falling between the minimum (lower edge) and maximum (upper edge) dimensions. The minimum dimension of the stimulus parameter is the least dimension of the parameter of the stimulus that effectively reduces the frequency of apneic events and/or duration of a single apneic event to an acceptable level below a predefined threshold level The maximum dimension of the parameter is a peak dimension of the parameter of the stimulus that can be applied without awakening the infant from a sleeping state. It is desirable to provide a stimulus that will regulate the respiratory rhythm of the infant without awakening the infant. Clinical studies have shown that such restricted respiration regulation can be effectively achieved. The minimum and maximum dimensions of the parameter or parameters of the stimulus preferably are determined by clinical examination of each infant and data representing such dimensions are inputted to the system before use. Alternatively, the minimum dimension of the stimulus parameter can be determined by the processor as described above.

As mentioned, certain data is inputted to the system by a clinician before use. Such data includes that representing the minimum duration of a single apneic event that triggers the application of a stimulus, the minimum duration of a single apneic event that also triggers the provision of an alarm signal, the minimum frequency of apneic events and minimum time period of each apneic event that triggers the application of a stimulus. A clinician can also input to the system data representing the category of the stimulus protocol to be used for the particular infant, the particular type of stimulus to be used, and the beginning amplitude, frequency or frequency range, and duration of the stimulus. A particular category and/or type of stimulus having certain parameter ranges that are helpful in regulating the respiratory rhythm of the infant can be determined during clinical studies.

Also inputted to the system is data representing the exclusionary windows (described above), if any. Data representing multiple moving windows of different time periods during which apneic event frequencies are monitored also can be inputted. The inputted data may be determined through clinical examination or guided solely or partly by the age of the infant. Figure 4 includes a table listing average respiratory periods during quiet sleep and REM sleep in 15 infants of different ages (estimated from Haddad G.G. et al., Maturation of ventilation and ventilatory pattern in normal sleeping infants, J. Appl. Physiol. 46:998-1002, 1979).

Many different types of stimuli, preferably falling within the stimuli category protocol described above, can be provided to an infant. Some of the stimuli types and associated effectors for providing such stimuli are discussed below.

An audible stimulus signal can be provided to the infant through an effector including at least one speaker. The speaker may be placed near the infant, may be embedded within the mattress or surface upon which the infant rests, or may be part of a headpiece worn by the infant. The audible signal can be continuous, pulsatile or a trigger pulse. The signal preferably falls within the audible frequency range and can have a single constant frequency or multiple varying frequencies. As described above, all parameters of the stimulus signal (such as frequency, amplitude and duration) are adjustable and preferably fall within the therapeutic window range of dimensions. For example, the amplitude of the audible stimulus signal should be greater than a minimum amplitude that would reduce the frequency of apneic events and/or shorten the duration of a single apneic event to a satisfactory level and be less than the amplitude that would awaken the infant from a sleeping state. The processor conventionally provides an analog control signal to the speaker to provide such a stimulus.

Another type of stimulus includes rocking the infant. Automated rockers for rocking the infant in a swing or basket are available. Rockers typically include a manual potentiometer for controlling the rocking rate. Those skilled in the art will understand that the potentiometer varies the resistance within the power line to vary the power received by the swing rate control circuit. The processor of the present invention preferably provides an electrical signal directly to the swing rate control circuit to control the rocking rate and amplitude. The rocker preferably is operated at a rocking rate such that the period of rocking is equal to, or an integer ratio of, the period of normal respiration during a most recent predefined period. Thus, the period of rocking may be updated each predetermined time period to reflect a change in the recent normal period of respiration. The rocking rate preferably falls within the therapeutic window of dimensions. Yet another type of stimulus is a vibratory tactile stimulus that vibrates something in contact with the infant. The vibratory stimulus is applied to the infant by vibrating the surface (i.e., mattress) on which the infant rests. An effector including a solenoid, speaker or pneumatic motor, or the like, for example, is attached to the mattress or surface to cause the mattress or surface to vibrate. The effector conventionally vibrates in response to an analog control signal received from the processor. As will be appreciated by those skilled in the art, the magnitude of the vibration relates to the amplitude of the control signal and the frequency of vibration relates to the frequency of the control signal. The vibration preferably is one of continuous, periodic or of short duration, consistent with the stimulus category protocol of the present invention. Additionally, the parameters of the vibratory stimulus preferably fall within the therapeutic window range of parameters.

Providing air pressure to the throat of the infant is another form of readily-available stimulus. Continuous Positive Airway Pressure (CPAP), also referred to as a "pneumatic splint", is one type of effector for providing such air pressure. The CPAP includes an air-delivery mask which is sealed over the nose of the infant for applying air through the nostrils of the infant's nose to the throat of the infant at a certain pressure. A control signal provided from the processor to the air delivery effector controls the pressure and rate of the air applied. The application of the air pressure preferably is one of continuous, periodic or of short duration, consistent with the stimulus category protocol of the present invention. Additionally, the parameters of the air pressure application preferably fall within the therapeutic window range of dimensions. Another suitable stimulus is one applied by inflating a balloon on the distal end of an esophageal or stomach balloon catheter to trigger swallowing or other respiratory reflexes. Esophageal or stomach balloon catheters are well-known to those skilled in the art. The distal end of such a catheter carries an inflatable balloon. The distal end of the catheter is inserted orally through the mouth of the infant and conventionally advanced through the throat opening of the infant until the distal balloon is precisely located in the desired location within the esophagus or stomach. The balloon then is inflated conventionally. The inflation and deflation of the balloon preferably is rhythmic and controlled by the processor. The inflation/deflation rhythm of the balloon preferably is one of continuous, periodic or of short duration, consistent with the stimulus category protocol of the present invention.

Another suitable stimulus includes a fluid provision to the throat or esophageal area of the infant to trigger swallowing or other respiratory reflexes. The fluid is provided conventionally using a fluid provision tube that is orally inserted through the mouth of the infant. The fluid provision is controlled by the processor and the provision preferably is one of continuous, periodic or of short duration, consistent with the stimulus protocol of the present invention.

Yet another suitable stimulus includes the application of light to the eyes of the infant to stimulate respiration. The light can be applied by means of a conventional light source that is preferably controlled electrically by a control signal provided by the processor. The light provision preferably is one of continuous, intermittent or of short duration, consistent with the stimulus protocol of the invention. The parameters of the light application should fall within the therapeutic window range of dimensions. For example, the brightness parameter of the light should be sufficient to reduce the frequency of apneic events and/or the duration of a single apneic event to a satisfactory degree yet not too bright to awaken the infant from a sleeping state Still another suitable tactile stimulus includes the application of a temperature change (either warmer or cooler) to a certain area of the skin of the infant to effect respiration. The temperature change is provided conventionally by heating/cooling elements that are electrically powered and controlled by the processor. The protocol for the provision of a temperature change preferably is one of continuous, periodic or short duration. The temperature change preferably also includes parameters falling within the therapeutic window range of dimensions. For example, a cooler temperature change should be cool enough so that the frequency of apneic events and/or duration of a single apneic event is reduced to a satisfactory level yet warm enough so that the infant is not awakened from a sleeping state. The application of a unique odor to trigger the smelling sense of the infant is another suitable stimulus. The odor should be weak enough so that the infant is not awakened yet strong enough so that respiration is positively affected, such that the parameter of the odor has a dimension that is within the therapeutic window range of dimensions. The protocol for the odor provision preferably is one of continuous, periodic or short duration.

Electrical signals applied by electrodes to an area of the skin of the infant to stimulate skin nerve receptors of the infant is yet another type of suitable stimulus. Such signals can be applied conventionally by electrodes that are electrically driven by control signals provided by the processor. The protocol for such electrical stimulation preferably is one of continuous, periodic or short duration. Additionally, the amplitude, frequency and other parameters of the electrical stimulation preferably have dimensions that fall within the therapeutic window range of dimensions.

The application of an air stream to an area of the skin of the infant is another available type of tactile stimulus. The protocol for the air stream provision preferably is one of continuous, periodic or short duration and the parameters of such provision preferably have dimensions that fall within the therapeutic window range of dimensions.

The above discussion of available stimuli is not meant to be exhaustive. Similarly, the following discussion of available detectors for gathering respiratory information is non-exhaustive. Respiration rate typically is detected by measuring: (i) chest and/or abdominal wall movement (e.g., using pressure transducers or by measuring resistivity or inductance), (ii) airflow (e.g., by measuring expiration and/or inspiration), (iii) transdiaphragmatic pressure (i.e., the esophageal pressure differential across the diaphragm at the gastroesophageal junction, and or (iv) blood oxygen and/or carbon dioxide tension. Other physiological parameters that can be measured using one or more detectors of the apparatus include, for example, sleep state (e.g., measure via surface electroencephalograms), cardiac rhythm (e.g., measure via electrocardiogram) and temperature.

In general, three categories of detectors are available for measuring chest and/or abdominal wall movement: (1) devices which rely upon pressure transducers (e.g., the "Star Sync Abdominal Sensor" made by Infrasonics, Inc., San Diego, CA; (2) devices which rely upon measurement of electrical resistivity and/or inductance (e.g., the Pneumogram recorder system (respiration monitor) manufactured by Edentec, subsidiary Nellcor, Inc., Hayward, CA or the Respitrace™ system by Ambulatory Monitoring, Inc. Ardsley, NY); and (3) devices which rely upon measurement of non-electrical resistivity and/or inductance (e.g., devices that contain coilsthrough which oscillating signals are sent to measure resistivity, for example, the Respiband™ by Ambulatory Monitoring, Inc. Ardsley, NY). Such devices are well known to those of ordinary skill in the art and can be operatively linked to the processor of the invention using no more than routine skill.

At least two categories of detectors are available for measuring airflow: (1) devices that employ a "hot wire" to determine air flow by detecting airflow-induced cooling of one or more hot wires placed in the path of (expelled or inhaled) air in the vicinity of the nose or mouth of an infant; and (2) pneumotachometers (devices that measure the differential air pressure drop across a fine screen or membrane that is present in the sealed path of airflow from the subject).

Respiration can also be monitored by devices (disclosed herein) that measure transdiaphragmatic pressure (i.e., the esophageal pressure differential across the diaphragm at the gastroesophageal junction). Many of these detectors are well known to those of ordinary skill in the art and can be operatively linked to the processor of the invention using no more than routine skill.

The processor of the invention preferably includes a programmable microprocessor. The microprocessor preferably is programmed to implement the above-described functions of monitoring the respiratory information received from the detectors and for providing stimuli and alarm signals when certain criteria are met. The microprocessor also is programmable in the sense that data can be input by a clinician to reflect minimum and maximum parameter dimensions of the therapeutic window, types of stimuli to apply, specific stimuliparameters, as well as other useful information (described above).

As will be readily understood by those skilled in the art, the processor also may include digital and/or analog circuitry for implementing part or all of the processor functions. Such circuitry preferably would electrically communicate with the microprocessor over bus lines. An ASIIC may be employed as a silicon chip for storing the underlying software on which the system operates.

Input/output circuitry that couples the microprocessor to the detectors and effectors may include digital and/or analog circuitry for processing signals received from the detectors and providing suitable control signals to the effectors. Such circuitry should be readily available to those skilled in the art and would depend on the type of detectors and effectors utilized.

Conventional computer hardware input devices such as a computer terminal with associated keyboard and/or mouse or pen as well as a user-friendly menu driven display can be employed for inputting data to the system. Such input devices would communicate conventionally through digital bus lines to the processor.

Examples INTRODUCTION

This example describes the treatment of a preterm infant having a conceptional age of 34 weeks and a history of life-threatening apneic events with tactile and auditory stimuli to reduce the frequency and/or duration of apneic events. The infant exhibited a remarkable improvement in the regularity of respiratory rhythm with continuous tactile and auditory stimulation.

METHODS

1. Bedside Monitoring of Behavior and Physiological Variables

Electrocardiographic activity, respiratory movement, tissue oxygen saturation, and incubator air and skin temperatures weremonitored continuously in the neonatal intensive care unit at all times after birth including the duration of the study. Respiratory movements were monitored by two methods: inductance plethysmography used routinely in the newborn intensive care unit, and abdominal surface pressure transduction (Infrasonics, Inc., San Diego California, part no. 4403021). Tracings from the pressure transducer were used for the analysis because they were less prone to artifact related to body movements, and were a reliable index of respiratory rhythm. The incubator (Ohio* Care Plus, Ohmeda, Louisville, Colorado) contained a servo-controlled heater that maintained incubator air temperature at 86.4 (range + 0.1 °F).

During the study protocol the infant was observed continuously, and movements of the face, arms, legs, and eyes were noted. There were occasional instances of eye opening around the feeding period. After 30 minutes of feeding by nasogastric infusion, the infant was returned to the incubator until the next feeding period. Between feeding periods, the infant exhibit frequent (>50% of the time) spontaneous movements. The types of movement were quite varied, included stretching of the entire body, phasic locomotive movements of the extremities, grimacing, twitches of the extremities, generalized myoclonic movement similar to that seen with the Moro reflex, writhing movements of the hands, and chewing. Eyelids were closed but rapid irregular movements of the eyes were seen at least every few minutes throughout the study. On day 51 two-channel EEG monitoring (fronto-central, gain 5000, lowpass 70hz, highpass 1 hz) was recorded for 6 hours. The recordings showed continuous electrographic activity with slow (0.5 to 3 Hz) and much faster (6-20 Hz) irregular waves. There were no discontinuous or trace alternant recordings. These behavioral and electrographic findings are suggestive of active sleep or transitional sleep, the predominant forms of sleep in normal 34 week infants.

Tracings from the abdominal sensor with movement artifact were excluded from analysis. During these periods the infant exhibited stretching or generalized limb movements. Occasionally, during or immediately following these periods there was a period of no abdominal movement. These episodes have been shown to be associated with Valsalva maneuvers and obstructed inspiratory efforts. The time of onset of these apneic episodes was undefined because of obscuration by movement. Therefore, the duration of movement-associated apnealikewise was uncertain. However, this exclusion did not bias our analysis because it was applied consistently throughout all trials.

The analog respiratory signal was displayed and recorded digitally using a data acquisition board and computer software (DATAQ Instruments, Akron, Ohio) and a portable PC-AT 486 mHz computer (Toshiba T6400).

2. Tactile and auditory stimulation A crib mattress (64 x 32 x 6 cm) was constructed (TheraSound, Inc., Newton,

Massachusetts) to provide tactile and auditory stimulation. Two speakers, a midrange and a woofer, were mounted to a plywood board and embedded in foam. The entire unit was covered with vinyl and placed in the infant incubator. A sheet and blanket were placed between the infant and surface of the mattress. The speakers were powered by an amplifier (QSC model 1200, Costa Mesa, California) and a programmable sine wave oscillator (Optimization, Inc., North

Hollywood, California). The output voltage of the oscillator was recorded and displayed with the respiratory signal. The oscillator frequency was set at 50 Hz (38% amplitude modulation at 1.3 Hz) for the protocol because this stimulus appeared to cause no obvious arousal or EEG change when moderate levels of stimulation were used. In all trials, the peak output voltage to the mattress speakers was kept constant (1.25 ±

0.05 V). The magnitude of stimulation was measured by means of a calibrated sound level meter (B mode, type 2, model 886, Simpson Electric Co., Elgin, IL) and an accelerometer (Grass Instruments, Quincy, MA). At the surface of the mattress, this level of stimulation resulted in sound level of 77 ± 2 dB and vibratory peak acceleration of 0.1 ± 0.01 m-sec^*2. There was less than 2 dB attenuation by the bedding material. When the mattress speakers were turned off, the ambient sound level in the incubator was 60 ± 2 dB (due to the incubator motor), and surface vibration was <0.01 m-sec^"2. RESULTS

Figure 5 shows examples of the effects on respiratory rhythm of turning the stimulus on (Fig. 5 A) and stimulus off (Fig. 5B). Typically a change in respiratory rhythm was seen within 1-2 minutes after stimulus change, after which a different respiratory pattern emerged. Figure 5 A shows that prior to initiating the stimulus, the respiratory pattern was irregular, with conspicuous episodes of expiratory apnea. This pattern continued for 1 min after starting the stimulus, then the respiratory pattern became regular. Figure 5B shows regular respiratory rhythm during stimulation and for approximately 1 min after stimulation was discontinued, then the baseline irregular respiratory pattern with significant apneic pauses emerged. In all stimulation trials, the respiratory pattern became noticeably more regular within 2 minutes after stimulus onset. After this transitional period, respiratory rhythm remained highly rhythmical for the entire duration of stimulation. Figure 6 shows an example of the respiratory pattern seen with stimulus off (Fig. 6A) and the last 285 seconds of a 15 minute stimulation trial (Fig. 6B), showing continued regularity in rhythm. There was no entrainment between the respiratory rhythm and amplitude modulation of the stimulus. In Figure 7, the effect of stimulation (horizontal bars, Fig. 7B) on respiratory period is shown for all trials, compared to the periods without stimulation. The small gaps in data represent periods of movement-induced artifact on the respiratory tracings. During stimulation, the average respiratory period became shorter (1.05 seconds vs. 1.38 seconds, Fig. 7A), and the probability of observing a period greater than five seconds was markedly reduced. The finding was replicated in all trials on two separate days and at different times relative to the feeding cycle.

The incidence of expiratory apnea was quantified for all breaths unobscured by movement artifact. We analyzed breaths after 2 minutes of change in stimulation, to allow for near-steady state effects of the stimulus. Figure 7 shows the resulting histograms of the number of breaths containing expiratory periods of various durations. Compared to no stimulation (n=2002 breaths), stimulation (n=2360 breaths) was associated with significant reduction in the expiratory periods (P<0.0001, Kalmagorov-Smirnov one-tail test), and a 28-fold reduction in the probability of apnea greater than 5 sec (0.04% vs. 1.1%). Without stimulation there were 24 breaths with expiratory period greater than 5 sec. The two longest apneas (13.6 sec, 12.9 sec) were associated with oxygen desaturation to <80% and bradycardia to 60-70 beats/min. One episode resolved spontaneously, the other ended while the infant's back was rubbed as the first step in the resuscitation. During the stimulation trials, there was only one breath with an expiratory period greater than 5 sec. There were no episodes of bradycardia or oxygen saturation less than 95%. DISCUSSION

The salient finding is that low level tactile and auditory stimulation caused significant improvement in the regularity of respiratory rhythm and prevented apnea-induced hypoxia in a sleeping infant. We believe that the therapeutic response was due to stimulation of afferent neural pathways which affected the respiratory rhythm generator. The existence of such pathways for tactile (R. Shannon, "Reflexes from respiratory muscles and costovertebral joints," in Handbook of Physiology of the Respiratory System II, ed. N.S. Cherniack and J.G. Widdicombe (American Physiological Sociatiey, Bethesda , MD, 1986) pp.431-47) and auditory (Corbeille, C, and Baldes, E.J., "Respiratory responses to acoustic stimulation in intact and decerebrate animals", Am. J. Physiol. 1929, 88:481-97; Eldridge, F.L., et al., "Phase resetting of the respiratory cycle by auditory stimuli demonstrated in decerebrate cats." J. Physiol 1990, 423:28P; and Stewart, M.W., and Stewart, L.A., "Modification of sleep repiratory patterns by auditory stimulation: Indications of a technique for preventing sudden infant death syndreom ?", Sleep 1991, 14:241-48) stimuli has been demonstrated in animal models; these stimuli can cause perturbation of the respiratory oscillator's medullary and pontine circuits. Prorhythmic effects of low level continuous stimulation was anticipated based on the known generic properties of oscillatory control systems (Paydarfar, D. and D. Buerkel, 1995, Chaos, in press), together with our observations of respiratory dysrhythmias in anesthetized animals. Our stimulus frequency of 50 Hz is near the peak frequency response for cutaneous mechanoreceptors (Martin, J.H. and T.M. Jessell, (1991) Elsevier, pp. 346-347), and is within the audible frequency range, producing a low pitched tone. The relative contribution of tactile versus auditory stimulation in achieving the therapeutic response was not tested in our experiments. Nevertheless, the stimulus caused a transformation of the respiratory pattern, without awakening the infant. Higher pitched tones may cause more disruptive effects on the sleep cycle, a less desirable approach because repeated arousals can cause potentially harmful sleep deprivation. Although the infant was not awakened by the stimulation, it is possible that stimulation caused an alteration in the infant's sleep state.

It should be understood that the preceding is merely a detailed description of certain preferred embodiments. It therefore should be apparent to those of ordinary skill in the art that various modifications and equivalents can be made without departing from the spirit or scope of the invention.

Claims

CJaims

1. An apparatus for reducing at least one of the frequency and duration of apneic events in a subject, the apparatus comprising: (a) at least one detector for detecting at least one of the frequency and the duration of an apneic event in the subject;

(b) at least one effector for delivering a stimulus to the subject, the stimulus having at least one adjustable stimulus parameter; and

(c) a processor operatively coupled to the detector and the effector, wherein the processor is actuated by at least one of the frequency and the duration of the apneic event to:

(i) establish a therapeutic window of dimensions for the stimulus parameter, wherein the therapeutic window comprises a range of dimensions sufficient to reduce at least one of the frequency and duration of apneic events without awakening the subject;

(ii) select at least one stimulus parameter having a dimension within the therapeutic window for delivery to the subject;

(iii) instruct the effector to deliver the stimulus having the stimulus parameter of step (ii) to the subject; and

(iv) repeat steps (i), (ii) and (iii) until at least one of the frequency and the duration of the apneic event in the subject is reduced.

2. The apparatus of claim 1, wherein the detector detects at least one physiological parameter selected from the group consisting of respiration rate, sleep state, cardiac rhythm, blood oxygen tension, blood carbon dioxide tension and temperature.

3. The apparatus of claim 2, wherein the physiological parameter comprises respiration rate.

4. The apparatus of claim 3, wherein the detector detects movement of the chest or abdominal wall.

5. The apparatus of claim 3, wherein the detector detects airflow movement.

6. The apparatus of claim 2, wherein the detector detects at least one of blood oxygen tension and blood carbon dioxide tension.

7. The apparatus of claim 1, wherein the effector produces at least one stimulus selected from the group consisting of an auditory stimulus, a tactile stimulus, a vestibular stimulus, a laryngeal stimulus, a pharyngeal stimulus, a light stimulus, a swallowing-inducing stimulus, a lung-inflating stimulus, and an olfactory stimulus.

8. The apparatus of claim 1, wherein the subject has a normal respiratory frequency and wherein the effector produces a continuous stimulus having a frequency that is approximately equal to the normal respiratory frequency.

9. The apparatus of claim 1, wherein the subject has a normal respiratory frequency and wherein the effector produces a pulsating stimulus having a frequency that is an integer ratio of the normal respiratory frequency.

10. The apparatus of claim 1, wherein the effector produces at least one of an auditory stimulus and a tactile stimulus.

11. The apparatus of claim 10, wherein the parameters of the auditory or the tactile stimulus comprise frequency, amplitude and duration.

12. The apparatus of claim 11, wherein the frequency of the auditory or tactile stimulus ranges from about 1 Hz to about 10 KHz.

13. The apparatus of claim 1 1, wherein the auditory stimulus comprises a pulsating stimulus having a predetermined adjustable carrying frequency that is approximately equal to the normal respiratory frequency of the subject.

14. The apparatus of claim 11, wherein the amplitude of the auditory stimulus has a dimension ranging from a minimum dimension sufficient to reduce the frequency or duration of apneic events to a maximum dimension which is less than the amplitude dimension which would awaken the subject.

15. The apparatus of claim 1, further including an alarm mechanism, operatively coupled to the processor, wherein the processor activates the alarm mechanism to produce an alarm signal if an apneic event is not terminated after a predetermined period of time.

16. A method for reducing at least one of the frequency and the duration of apneic events in a subject, the method comprising:

(a) detecting at least one of the frequency and the duration of an apneic event in the subject, wherein detecting at least one of the frequency and duration of an apneic event actuates the following steps; (b) establishing a therapeutic window for a parameter of a stimulus, wherein the therapeutic window comprises a range of dimensions sufficient to reduce at least one of the frequency and duration of apneic events without awakening the subject;

(c) selecting a stimulus parameter having a dimension within the therapeutic window;

(d) delivering to the subject the stimulus having the stimulus parameter of the dimension selected in step (c); and

(e) repeating steps (b), (c) and (d) until at least one of the frequency and duration of apneic events is reduced in the subject.

17. The method of claim 16, wherein the stimulus is selected from the group consisting of an auditory stimulus, a tactile stimulus, a vestibular stimulus, a laryngeal stimulus, a pharyngeal stimulus, a light stimulus, a swallowing-inducing stimulus, a lung-inflating stimulus, and an olfactory stimulus.