JP2003079732A - System for assisting respirating, swallowing and coughing function - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a system which can comfortably be worn by a patient having consciousness for a long time and can integrally and efficiently support a respiring, coughing or swallowing function damaged by various diseases or injuries. SOLUTION: This system is provided with a respiration assisting device such as electrodes 1a and 1b, a solenoid valve 3 for opening/closing a bronchi incising tube 2, an intra-pharynx pressure sensor 4, a signal input button 5 for inputting the intention indication of cough reflection by a person to be assisted and a controller 6 composed of at least one selected out of analog circuit, digital circuit and computer program for generating an output waveform approximate to biological respiratory center output ordinarily, changing the output waveform corresponding to the detecting signal of the sensor and/or input signal of the respiration assisting device and controlling the inhalation or exhalation of the respiration assisting device and the opening/closing degree of an opening/closing mechanism.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system for physiologically and consistently controlling a breathing assistance device represented by a ventilator and a functional electrical stimulator.
More specifically, the present invention relates to a technique for controlling a breathing / swallowing / coughing function assisting device using an artificial neuron network.

[0002]

2. Description of the Related Art As a means for assisting respiration, a positive pressure type ventilator which is widely used at present, a negative pressure type ventilator used for a relatively limited number of patients and pathological conditions, a patient with high spinal cord injury, etc. There is a diaphragm pacer (functional electrical stimulator) used in. In the positive pressure type ventilator, the controlled object is the airway pressure or the airway flow velocity (and its integrated value, the ventilation volume), and in the negative pressure type ventilator, the negative pressure chamber internal pressure, and in the diaphragm pacer, it is stimulated. It is a pulse. At present, tracheostomy and positive pressure artificial respiration are generally used to secure the respiration of patients with neurodegenerative diseases such as alveolar hypoventilation such as cerebral infarction, cerebral hemorrhage, postoperative brain tumor surgery, and amyotrophic lateral sclerosis. There is.

The respiratory center output is not a simple rectangular wave or triangular wave, but a non-linear pulse adapted to the mechanical characteristics of the lung and thorax. Moreover, the inspiratory time, expiratory time, and waveform pattern flexibly change due to non-respiratory exercise such as coughing and swallowing, and the timing of entry.
For example, physiologically, when swallowing water or food, the activity of the tongue muscle pushes the water or food into the pharynx and suppresses inspiratory respiratory neurons to prevent aspiration.

[0004]

However, in any of the above conventional control objects, most of the current control waveforms are rectangular waves, and in other cases, they are triangular waves or sine waves and are emitted from the respiratory center. It is far from the physiological nerve impulses. In addition, the conventional respiratory assistance device does not have a control mechanism that suppresses inspiratory respiratory neurons during swallowing, and is not subject to control of the cough reflex at all, and discharges sputum and aspirated foreign matter. I couldn't. For this reason, medical staff have been performing airway cleaning with frequent suction using a suction tube or bronchial fiberscope (or electronic scope), which is extremely painful for patients and even prevents disease progression. Even if medical treatments related to recovery of various physiological functions such as exercise, defecation and urination were advanced, the range of action was still limited.

Therefore, a first object of the present invention is to provide a breathing / swallowing / coughing function assisting system which a conscious patient can wear comfortably for a long period of time. The second challenge is to establish a system that can assist the respiratory, cough, and swallowing functions impaired by various diseases and injuries in an integrated and effective manner without losing consistency. .

[0006]

In order to solve the problem, a respiratory / swallow / cough function assisting system of the present invention detects a respiratory assist device, an opening / closing mechanism for opening and closing a tracheostomy tube, and a pharyngeal pressure. A sensor and a signal input device that inputs an indication of the cough reflex by the person being assisted, and an output waveform that is close to a physiological respiratory center output at all times is generated, and is used as a detection signal of the sensor and / or an input signal of the signal input device. A controller comprising at least one selected from an analog circuit, a digital circuit, and a computer program, which changes the output waveform in accordance with the operation, and controls the inspiratory operation or expiratory operation of the respiratory assistance device and the degree of opening / closing of the opening / closing mechanism. Is characterized by.

In the present invention, the breathing assistance device is not particularly limited, and may be any of electrical stimulation devices such as positive pressure type respirator, negative pressure type respirator, and diaphragm pacer. The opening / closing mechanism is, for example, a known solenoid valve or shutter. The controller is an implementation of an artificial neuron network circuit (hereinafter "network circuit").
1) neurons form a network.
No matter how many neurons there are in one neuron group, there are hundreds or thousands of neurons, and only the properties of the entire group are specified, and it behaves like a real complex neuron network.
(Presentation Y. Oku and TE Dick, NeuroReport 6: 379-38
3, 1995).

According to the system of the present invention, the controller always controls the breathing operation of the breathing assistance device and the degree of opening / closing of the opening / closing mechanism based on the output waveform close to the physiological respiratory center output to assist the breathing of the living body. . When the sensor detects an increase in the pharyngeal pressure, it is determined to be swallowing, the signal is transmitted to the controller, and the controller changes the output waveform to stop the inspiration operation. Also, when the person who wants to cough, he / she inputs the intention indication to the signal input device.

[0009]

BEST MODE FOR CARRYING OUT THE INVENTION [Configuration of Network Circuit] It is assumed that there are n neuron groups, and they are mutually excitatory or inhibitory. The average membrane potential as a group of neuron groups j = 1, ..., N is u _j , the output as a group (average firing rate) is x _j , the decay time constant is τ (= 80 ms), and the i-th group Let w _{ji be} the connection weight when the output from the neuron group is connected to the j-th group neuron group, s _{j be} the sum of input signals received from the outside of the network by the j-th group neuron group, and h _j be the excitation threshold. , The average membrane potential of the jth neuron group is determined according to τdu _j (t) / dt = -u _j + Σw _ji x _j (t) + s _j (t) -h _j , and the output is x _j ( t) = f (u _j (t)). f is a non-linear saturation type sigmoid function which is 0 until it exceeds the threshold value h _j , and an output starts to be output when it exceeds h _j , which monotonically increases and saturates at a certain value. Here, to simplify the implementation, h _j = 0 and s _j (t) = constant value T _j . f can be defined as follows, for example.

F (x) = max (g (x), 0) g (x) = 4 / (1 + e ^{−1.8 (x−2)} ) −0.106 Furthermore, a certain (I-DEC and E-DEC) accelmodification property that the membrane potential gradually returns to the resting membrane potential even when the external stimulus is unchanged.
Suppose that. In this modeling of the acclimation characteristic, the first neuron group having the acclimation characteristic has a connection weight of 1 and is connected to a second virtual neuron with a relatively slow decay time constant (τ = 1.6s). This is realized by assuming that the two neuron groups are feedback-coupled to the first neuron group by the coupling weight.

The network circuit used for control is, as shown in FIG. 1, I-DEC / E-DEC / E-AU which forms a respiratory rhythm.
It consists of three brainstem proper neuron groups of G, a primary motor neuron group I-AUG which transmits respiratory motion command to the diaphragm, and three neuron groups of cough interneuron (n = 7 in the above formula). In the connection part between neuron groups, ● represents an inhibitory connection and ○ represents an excitatory connection.

The I-DEC / E-DEC / E-AUG and I-AUG neuron groups correspond to neurons existing in the brain stem of an actual mammal, and their names are derived from the firing pattern. . For example, I-DEC is a group of neurons that emits (Inspiratory) tapering (DECrementing) during inspiration, and E-AUG is
It is a group of neurons that emits expiratory (AUGmenting) firing. It is known that the three neuronal groups of I-DEC, E-DEC, and E-AUG are inhibitoryly connected to each other by a known electro-neurophysiological method, and that they are major elements of respiratory pattern formation. In addition, the I-AUG neuron group is in inhibitory connection with these three neuron groups. The model respiratory neuron group shown in Fig. 1 (I-DEC, E-DEC, E-AUG
-I-AUG) binding is also based on neurophysiological findings.

The primary motor neurons that control the abdominal muscle group exhibit a firing pattern similar to that of E-AUG neurons, and the activity of the glottal obturator muscle (M. Thyroarytenoid) is similar to that of E-DEC neurons. I know that.

A method for quantitatively measuring the coupling strength between neuron groups has not been established so far.
If two neuron groups have inhibitory cyclic connections, that is, A suppresses B, B suppresses C, and C suppresses A, if the connection strength is appropriately determined, It is theoretically known that this network circuit can be a system that returns to a stable oscillation state, that is, a limit cycle oscillator, no matter what timing the disturbance enters.

The three neuron groups of I-DEC, E-DEC, and E-AUG are
In fact, it has a suppressive ring structure, and regarding the coupling between neuron groups based on the findings of electroneurological physiology, when a simulation is performed by setting an appropriate coupling load, it becomes a limit cycle oscillator. Furthermore, assuming that the I-DEC and E-DEC have acclimation properties, assuming that the I-DEC neuron group has a self-excitatory connection that feeds back to itself, the change in the membrane potential of the neuron is further enhanced by that of the actual neuron. You can get closer. When the limit cycle oscillator receives an external stimulus, the vibration cycle changes. For example, in the case of respiratory movement, if the timing of the stimulation is inspiration due to stimulation of the superior laryngeal nerve, inspiration ends and transitions to exhalation, or it is temporarily suppressed and inspiration is extended, and stimulation is exhaled. Exhalation may be extended at times (hereinafter, this characteristic is referred to as a phase-resetting characteristic). In the network circuit modeled in this way, the behavior with respect to the external stimulus is consistent with the actual behavior of the mammalian respiratory center. The external stimulus includes at least the detection signal of the pharyngeal pressure sensor and the input signal of the signal input device.

It is not clear what kind of nerve connection the cough center has with the respiratory center, but there is a report on the behavior of the network circuit when coughing (announcement:
OkuY, Tanaka I, and Ezure K. Activity of bulbar re
spiratory neurons duringfictive coughing and swall
owing in the decerebrate cat. J. Physiol. Lond. 48
0: 309-324, 1994). Now, the input neuron group (not considering the membrane property) that triggers the cough first is CIN-
1. C is the neuron group that forms a cough pattern from the stimulation
IN-2 and CIN-3. Since CIN-2 and CIN-3 receive the same membrane characteristics and input from the same respiratory neuron, they have exactly the same membrane potential changes, but the difference between them is that CIN-2 is a group of inhibitory neurons and respiratory neurons. CIN-3 is an excitatory neuron group that acts as an inhibitor of coughing, whereas
It is a group of secondary motor neurons. By adding these three cough interneurons to the network circuit, the behavior of the network circuit with respect to the cough stimulus can be matched to some extent with the actual behavior of the mammalian respiratory center.

The swallowing reflex is said to occur when food or the like reaches the pharyngeal cavity and stimulates the superior laryngeal nerve, and the excitement is transmitted to the swallowing center of the brain stem. The electrical stimulation of the superior laryngeal nerve either temporarily suppresses the inspiration during inspiration, or completely terminates it and shifts it to exhalation. In addition, the electric stimulation during exhalation prolongs exhalation. By that,
It is considered that aspiration is effectively prevented when swallowing or when food etc. is inside the pharyngeal cavity. Such a response that changes depending on the timing of stimulus (phase response)
Can be reproduced, assuming that the superior laryngeal nerve excites E-DEC neurons in the artificial neural network described above, as shown in FIG.

In order to extract a signal from such a network circuit in real time, a method in which the network is configured by an analog or digital circuit as hardware and a method in which it is configured as software by describing it in a computer program and numerical differentiation is performed Although there is a method of extracting the signal, the latter is more flexible in terms of easy modification. Here, the above network circuit is described in C language, and the fourth-order Runge-Kut is used.
The output obtained by real-time numerical analysis by the ta method is shown. Table 1 shows the parameter values used in the program. FIG. 3 shows the behavior of the network circuit when the intention input is input from the signal input device.

[0019]

[Table 1]

[Implementation of network circuit on system]
A schematic diagram of the entire system is shown in FIG. This system
The phrenic nerve electrode 1a attached to the diaphragm and the abdominal muscle electrode 1b also attached to the abdominal muscle as the breathing assistance device 1.
A solenoid valve 3 attached to the tracheostomy tube 2 to open and close the tube, a sensor 4 for detecting the pressure inside the pharynx, a signal input button 5 for inputting an indication of intention to cough, and an FES controller embedded in the body. 6 and a power supply device 7 such as a battery held by the patient outside the body. Among them, the phrenic nerve electrode 1a and the FES controller 6 have a function corresponding to a diaphragm pacer.

The FES controller 6 normally generates an output waveform close to the physiological respiratory center output, changes the output waveform according to the detection signal of the sensor 4 and / or the input signal of the signal input button 5, and On the basis of the output waveform, a network circuit is built to energize the phrenic nerve electrode 1a and the abdominal muscle electrode 1b or control the degree of opening / closing of the solenoid valve.

The output from the I-AUG neuron group is used to control the phrenic nerve electrode 1a. Since this output reflects phrenic nerve activity, the pulse waveform with the signal waveform obtained by DA conversion of the output from the I-AUG neuron group as an envelope is used as a control signal to energize the phrenic nerve electrode 1a to cause the diaphragm to pass through. Electrically stimulate. This assists the intake. In the case of other respiratory assistance devices, the change in the ventilation volume of the respiratory assistance device with time may be controlled so as to match the I-AUG waveform. When further exhalation assistance is required such as during exercise, the abdominal muscle electrode 1b is used as a control signal with a pulse waveform having a signal waveform obtained by DA conversion of the output from the E-AUG neuron group as an envelope.
To stimulate the abdominal muscles. However, when the present system is used also during exercise, it is preferable to combine a sensor for detecting exercise, such as an acceleration sensor, with the system. As an aid for coughing, the output from the CIN-3 neuron group is
There is a method of energizing the abdominal muscle electrode 1b to electrically stimulate the abdominal muscle using a pulse waveform having an A-converted signal waveform as an envelope as a control signal. This electrode may be placed on the lumbar spinal cord instead of on the abdominal muscle.

[Glottic Closure] In order to cause an effective cough reflex that expels sputum and foreign substances, in addition to controlling the diaphragm muscles and abdominal muscles, the glottis is closed at the beginning of the cough calling phase to reduce airway pressure. Needs to be raised. The cough reflex occurs because the airway pressure once increased is rapidly released to atmospheric pressure. This glottic closure technique has not previously been established by functional electrical stimulation or other methods. In the present invention, the electromagnetic valve 3 is attached to the end of the tracheostomy tube 2, the signal input button 5 is pressed to excite the CIN-1 neuron group, and a command is issued to act on the entire network circuit and output to the CIN-3 neuron group. , Using its output 10
By closing the solenoid valve 3 for 0-150 ms, the same effect as the glottal closure in the early stage of coughing can be obtained.

Respiratory control to prevent aspiration during swallowing
(Suppression) can be performed by converting a change in pharyngeal pressure into an electric signal by the sensor 4, detecting swallowing activity, and causing the E-DEC neuron to perform excitatory feedback.
The overall effect on the network of excitatory inputs to E-DEC neurons is inspiratory depression. That is, if an input is entered in the inspiratory phase, it has the effect of ending the inhalation and shifting to the expiratory breath as shown in FIG. 5, and if entering the expiratory phase, it has the effect of prolonging the exhaled breath as shown in FIG. This method can reduce the risk of inhalation of food or the like due to inhalation during swallowing.

[0025]

EXAMPLE FIG. 7 shows the result of electrical stimulation of the phrenic nerve and the abdominal muscle by the output from the artificial neuron network circuit using a rabbit. From the top, the phrenic nerve stimulation signal (Phr; actually stimulated by a pulse wave with this waveform as the envelope), abdominal muscle stimulation signal (Abd), and airway flow (Flow) are shown. The expiratory phase is shown above the reference line of the airway flow rate, and the inspiratory phase is shown below. The cough stimulation signal input in the figure corresponds to when the input button 5 is pressed. It can be seen from FIG. 7 that the breathing pattern and the cough pattern are generated by the output from the artificial neuron network circuit, and thereby the electrical stimulation of the phrenic nerve and the abdominal muscle nerve can assist the breathing and the cough.

[0026]

INDUSTRIAL APPLICABILITY The present invention provides a means for assisting a coughing / swallowing function which has never been found in a conventional ventilator.
Furthermore, by combining it with a diaphragm pacer, which has conventionally only been very narrowly adapted, it becomes possible to develop a respiratory assist device that is much smaller and more portable than conventional ventilators.

[Brief description of drawings]

FIG. 1 is a diagram showing an artificial network circuit used in a system of an embodiment.

FIG. 2 is a diagram showing an input position of a swallowing stimulus in the artificial network circuit.

FIG. 3 is a graph showing the behavior of the artificial network circuit during cough stimulation.

FIG. 4 is a schematic diagram showing a state in which the system of the embodiment is attached to a living body.

FIG. 5 is a graph showing the behavior of the artificial network circuit when a swallowing stimulus is input in the inspiratory phase.

FIG. 6 is a graph showing the behavior of the artificial network circuit when a swallowing stimulus is input in the expiratory phase.

FIG. 7 is a graph showing the airway flow rate of rabbits when electrical stimulation is performed by the output from the artificial network circuit.

[Explanation of symbols]

1a Phrenic nerve electrode (respiratory assist device) 1b Abdominal muscle electrode (respiratory assist device) 2 tracheostomy tube 3 Solenoid valve 4 pharyngeal pressure sensor 5 Signal input button 6 FES controller 7 Power supply device

Claims (3)

[Claims] 1. A breathing assistance device, an opening / closing mechanism for opening and closing a tracheostomy tube, a sensor for detecting intrapharyngeal pressure, a signal input device for inputting an indication of intention of a cough reflex by a person being assisted, and a physiological input at all times. Generate an output waveform close to the respiratory center output,
An analog circuit, a digital circuit, and a computer program that change the output waveform according to the detection signal of the sensor and / or the input signal of the signal input device to control the inspiratory operation or expiratory operation of the respiratory assistance device and the degree of opening / closing of the opening / closing mechanism A breathing, swallowing, and coughing function assisting system, which is provided with a controller consisting of one or more kinds selected from among them. 2. The controller comprises three I-DEC neuron groups that gradually reduce firing during inspiration, an E-DEC neuron group that gradually reduces firing during exhalation, and an E-AUG neuron group that gradually fires during expiration. Brainstem proper neuron group, primary motor neuron group I-AUG that transmits respiratory movement commands to the diaphragm, CIN-1 neuron group that triggers cough reflex based on the input signal of the signal input device, and the cough pattern is formed by the trigger In addition, an artificial neuron network consisting of three cough interneurons, a CIN-2 neuron group acting inhibitory to the brain stem proper neuron group and a CIN-3 neuron group acting excitatory, is provided. -DEC and E-AUG neuron groups are inhibitoryly connected to each other, I-AUG neuron group is inhibitoryly connected to these three neuron groups, and I-DEC neuron group Acclimation characteristics and have a self-excitatory coupling system of claim 1 E-DEC neuron group with acclimatization characteristics. 3. The system of claim 1, wherein the breathing assistance device is an electrical stimulator such as a diaphragm pacer.