Systems and methods for personalized and real-time adaptation for spontaneous breathing onset detection

Abstract

This invention relates to a method comprising receiving a pressure signal from a pressure sensor (17) and / or a flow signal from a pressure sensor and a flow sensor (16), and receiving signals from one or more sensors (56, 58, 60, 62) measuring different physiological parameters from the pressure sensor (17) and the flow sensor (16). The method includes detecting the onset of spontaneous breathing in a patient based on the pressure signal and / or the flow signal, and synchronizing the provision of respiratory support to the patient with the onset of spontaneous breathing detected using the pressure signal and / or the flow signal. The method includes calibrating parameters and thresholds to be used for detecting the onset of spontaneous breathing based on these signals. The method includes, after calibration, switching to: detecting the onset of spontaneous breathing in the patient based on these signals, and synchronizing the provision of respiratory support to the patient with the onset of spontaneous breathing detected using these signals.

Description

Background Technology

[0001] The topic disclosed in this article relates to the control of ventilators.

[0002] Basic modes of ventilator operation can be subcategorized based on whether breathing is initiated by the ventilator or the patient. Patient-initiated breathing is called spontaneous breathing. When ventilation is triggered by a spontaneous breathing event, it is crucial that the ventilator begins the inspiratory process as soon as the patient attempts to breathe. In other words, there should be a timing synchronization between the patient's effort and the ventilator's action. It is estimated that patient-ventilator asynchrony exists in approximately 50% of breaths. Any patient-ventilator asynchrony can lead to patient discomfort, sleep disturbances, and delays in weaning. Currently, there is a delay (e.g., several hundred milliseconds) between the onset of diaphragmatic muscle activation and the time it causes changes in pressure and flow in the facial area. Therefore, earlier detection of the patient's inspiratory effort and improvement of patient-ventilator synchronization during spontaneous breathing are needed. Summary of the Invention

[0003] The following outlines certain embodiments commensurate with the scope of the initially claimed subject matter. These embodiments are not intended to limit the scope of the claimed subject matter, but rather are intended only to provide a brief overview of possible implementations. In practice, the invention may include various forms that are similar to or different from the embodiments described below.

[0004] In one embodiment, a computer-implemented method is provided for detecting the onset of spontaneous breathing in a patient coupled to a ventilation system. The method includes receiving, at a processor, a pressure signal from a pressure sensor and / or a flow signal from a flow sensor coupled to the patient. The method also includes receiving, at the processor, signals from one or more sensors coupled to the patient, the one or more sensors measuring different physiological parameters from the pressure sensor and the flow sensor. The method further includes detecting the onset of spontaneous breathing in the patient based on the pressure signal and / or the flow signal via the processor. The method also includes synchronizing, via the processor, the provision of respiratory support to the patient via the ventilation system with the onset of spontaneous breathing detected using the pressure signal and / or the flow signal. The method further includes calibrating, via the processor, parameters and thresholds to be used for detecting the onset of spontaneous breathing when synchronizing the provision of respiratory support using the pressure signal and / or the flow signal. The method further includes, after calibration, switching to: detecting the onset of spontaneous breathing in the patient based on the signals from the one or more sensors via the processor, and synchronizing, via the processor, the provision of respiratory support to the patient via the ventilation system with the onset of spontaneous breathing detected using the signals from the one or more sensors.

[0005] In another embodiment, a ventilation system is provided. The ventilation system includes a plurality of sensors configured to be coupled to a patient and generate signals related to the patient's respiratory function, wherein the plurality of sensors include a flow sensor, a pressure sensor, and at least one electromyography (EMG) sensor. The ventilation system includes a memory encoding processor-executable routines. The ventilation system also includes a processor configured to access the memory and execute the processor-executable routines, wherein the routines, when executed by the processor, cause the processor to perform actions. These actions include extracting respiratory features from corresponding signals from each of the plurality of sensors. These actions also include estimating or measuring the presence of an electrocardiogram (ECG) in the EMG signals received from the at least one EMG sensor. These actions also include detecting the onset of spontaneous breathing in the patient based on the respiratory features extracted from the corresponding signals from the plurality of sensors and the estimated or measured ECG presence.

[0006] In another embodiment, a computer-implemented method is provided for real-time calibration of parameters used to detect the onset of spontaneous breathing in a patient coupled to a ventilation system. The method includes receiving, at a processor, flow signals, pressure signals, and EMG signals, respectively, from a flow sensor, a pressure sensor, and an electromyography (EMG) sensor coupled to the patient. The method also includes detecting the onset of spontaneous breathing via the EMG signal using an initial value range for at least one parameter of a start detection algorithm. The method further includes detecting the onset of spontaneous breathing via the processor using at least one of the flow signal and the pressure signal. The method also includes determining, via the processor, a time difference for detecting the onset of spontaneous breathing in the EMG signal and at least one of the flow signal and the pressure signal. The method further includes selecting, via the processor, at least one updated parameter of the start detection algorithm used to detect the onset of spontaneous breathing in the EMG signal, based at least on the time difference. Attached Figure Description

[0007] These and other features, aspects, and advantages of the invention will be better understood when reading the following detailed description with reference to the accompanying drawings, in which like reference numerals denote like parts throughout the drawings, wherein:

[0008] Figure 1 It is a schematic diagram of a mechanical ventilation system based on various aspects of this disclosure;

[0009] Figure 2 This is a schematic diagram of processing sensor signals according to various aspects of this disclosure to determine the onset of spontaneous breathing in a patient coupled to a ventilation system;

[0010] Figure 3This is a flowchart of a method for patient personalization and real-time adaptation for spontaneous breathing initiation detection according to various aspects of this disclosure;

[0011] Figure 4 This is a flowchart of a method for estimating the occurrence of ECG by utilizing respiratory characteristics at the onset of spontaneous breathing in a patient, according to various aspects of this disclosure;

[0012] Figure 5 This is a schematic diagram of different processes for extracting respiratory features from signals according to various aspects of this disclosure;

[0013] Figure 6 It is a schematic diagram of a process for detecting the presence and abnormality of ECG (e.g., reverse ECG) according to various aspects of this disclosure;

[0014] Figure 7 It is the reference ECG mode in the signal;

[0015] Figure 8 It is a graph showing the occurrence of ECG within an EMG signal according to various aspects of this disclosure;

[0016] Figure 9 It is a graph showing ECG anomalies within an EMG signal according to various aspects of this disclosure;

[0017] Figure 10 This is a schematic diagram of a process for detecting the onset of spontaneous breathing in a signal (e.g., an EMG signal) using ECG information, according to various aspects of this disclosure.

[0018] Figure 11 This is a graph showing the use of multiple thresholds and ECG information to detect the onset of spontaneous breathing according to various aspects of this disclosure;

[0019] Figure 12 This is a graph showing the detection of the onset of spontaneous breathing in the flow signal compared to the EMG signal, according to various aspects of this disclosure;

[0020] Figure 13 This is a graph showing the detection of the onset of spontaneous breathing in the flow signal compared to the EMG signal, according to various aspects of this disclosure;

[0021] Figure 14 It is a graph of EMG and flow data collected from various aspects of this disclosure for calibration dataset;

[0022] Figure 15 It is a time-normalized curve of the moving average for respiration based on various aspects of this disclosure;

[0023] Figure 16It is a graph of the maximum likelihood function for various respiratory actions derived from EMG data used for calibration, based on various aspects of this disclosure;

[0024] Figure 17 It is a graph showing the use of a modified dynamic threshold in automatic threshold selection on EMG calibration data according to various aspects of this disclosure;

[0025] Figure 18 This is a schematic diagram of an adaptive method for selecting parameters of an EMG-based initiation detection algorithm for an individual patient, according to various aspects of this disclosure.

[0026] Figure 19 It is a graph of the quality factor of respiration calculated based on initial data collected within a range of values ​​of parameters (e.g., lower threshold) according to various aspects of this disclosure;

[0027] Figure 20 It is a graph of the average time difference of the breath start detection calculated based on initial data collected within a range of values ​​of parameters (e.g., lower threshold) according to various aspects of this disclosure;

[0028] Figure 21 It is a graph of erroneous breathing detection calculated based on initial data collected within a range of values ​​for parameters (e.g., lower threshold) according to various aspects of this disclosure;

[0029] Figure 22 It is a graph showing the relationship between erroneous breath detection and window size within a range of parameter (e.g., lower threshold) values, based on various aspects of this disclosure; and

[0030] Figure 23 This is a flowchart of a method for real-time calibration of parameters used to detect the onset of spontaneous breathing in a patient coupled to a ventilation system, according to various aspects of this disclosure. Detailed Implementation

[0031] One or more specific implementations will be described below. To provide a concise description of these implementations, not all features of an actual implementation will be described in this specification. It should be understood that in the development of any such actual implementation, as in any engineering or design project, many implementation-specific decisions must be made to achieve the developer's specific objectives, such as complying with system-related and business-related constraints that may differ from implementation to implementation. Furthermore, it should be understood that such development efforts may be complex and time-consuming, but remain routine tasks of design, fabrication, and manufacturing for those skilled in the art who benefit from this disclosure.

[0032] When describing the elements of various embodiments of the subject matter of this invention, the articles “a,” “an,” “the,” and “the” are intended to indicate the presence of one or more of the stated elements. The terms “comprising,” “including,” and “having” are intended to be inclusive and mean that additional elements may be present in addition to the listed elements. Furthermore, any numerical examples in the following discussion are intended to be non-limiting, and therefore the additional numerical values, ranges, and percentages are within the scope of the disclosed embodiments.

[0033] As will be understood, specific embodiments of this disclosure may be embodied as systems, methods, apparatus, or computer program products. Therefore, aspects of this disclosure may take the form of entirely hardware implementations, entirely software implementations (including firmware, resident software, microcode, etc.), or implementations combining software and hardware aspects, all of which may generally be referred to herein as “circuit,” “module,” or “system.” Furthermore, aspects of this disclosure may take the form of computer program products embodied in one or more computer-readable media having computer-readable program code embodied thereon.

[0034] This disclosure provides a system and method for detecting the onset of spontaneous breathing in a patient using a personalized onset detection algorithm adapted in real-time to the patient's connection to a ventilation system (e.g., a mechanical ventilator). The disclosed implementation initially (and during recalibration) utilizes signals from a pressure / flow sensor to detect the onset of spontaneous breathing and synchronously provide respiratory support to the patient. Subsequently, after calibration (or recalibration), signals from other sensors (e.g., an EMG sensor) can be used to detect the onset of spontaneous breathing (e.g., using parameters determined during calibration / recalibration) and synchronously provide respiratory support to the patient. Sometimes, both EMG signals and pressure / flow signals can be used to determine the onset of spontaneous breathing. Detecting the onset of spontaneous breathing in one or more of these signals can trigger (e.g., via onset triggering) the provision (and synchronization) of respiratory support to the patient. The disclosed implementation can provide faster and more accurate detection of a patient's respiratory effort. Furthermore, the disclosed implementation can improve synchronization between the ventilator and the patient.

[0035] Figure 1An example of a mechanical ventilation system 10 is schematically shown. The ventilation system 10 provides a pneumatic circuit that delivers breathing gas to the patient 11 and exhales air from the patient to assist the patient 11 in breathing. As described in more detail below, the ventilation system is configured to detect the onset of spontaneous breathing in the patient 11 in a faster and more accurate manner than a typical ventilation system. The ventilation system 10 includes a ventilator 12, a breathing circuit 14, and sensors 16, 17. The ventilator 12 can operate in multiple modes. These modes include a controlled mode (breathing initiated by the ventilator 12), a spontaneous mode (breathing initiated by the patient), and a supported mode (breathing initiated by both the ventilator and the patient). Both the spontaneous and supported modes require synchronization with the patient's effort.

[0036] The ventilator 12 supplies gas (e.g., air or air containing anesthetics, drugs, etc.) to the patient 11 via a breathing circuit 14, and receives exhaled air via the breathing circuit 14. In the illustrated example, the ventilator 12 receives air from an air source 16 via a conduit 18 and oxygen (O2) from an oxygen source 20 (e.g., a compressed oxygen container) via a conduit 22. The ventilator 12 includes valves 24 and 26, sensors 28 and 30, valve 32 and sensor 34, and a controller or processing unit 36. Valves 24 and 26 control the supply of air and oxygen (a mixture thereof) to the breathing circuit 14 via conduit 38, respectively. Sensors 28 and 30 sense or detect the supply of air and oxygen, respectively, and transmit signals representing such sensed values ​​to the processing unit 36.

[0037] Valve 32 includes a valve mechanism connected to breathing circuit 14 via conduit 40 to control the flow of exhaled air received from breathing circuit 14 to discharge conduit 42. Sensor 34 includes means for sensing the flow of exhaled air to discharge port 42. This sensed value of the exhaled air is further transmitted to controller or processing unit 36.

[0038] Controller 36 generates control signals for the operation of control valves 24, 26, and 28. Controller 36 includes one or more processors 44 and memory 46. The one or more processors 44 execute instructions contained in memory 46. The execution of these instructions causes controller 36 to perform steps such as generating control signals. These instructions may be loaded from read-only memory (ROM), mass storage devices, or some other persistent storage device into random access memory (RAM) for execution by the processing unit. In other embodiments, hard-wired circuitry may be used in place of or in combination with software instructions to implement the described functions. For example, processor 44 may be embodied as part of one or more application-specific integrated circuits (ASICs). Unless otherwise specifically indicated, controller 36 is not limited to any particular combination of hardware circuitry and software, nor to any particular source of instructions executed by the processing unit. In other specific embodiments...

[0039] Breathing circuit 14 delivers breathing gases (air, oxygen, and possibly other additives such as anesthetics, drugs, etc.) from ventilator 12 to patient 11, while also guiding exhaled air from patient 11 to system 10 and ventilator 12. Breathing circuit 14 includes an inspiratory segment or segment 48, an expiratory segment 50, a Y-connector 52, and a patient segment 54. Inspiratory segment 48 extends from and is pneumatically connected to catheter 38 at one end, and extends from and is pneumatically connected to catheter 52 at the other end. During forced, assisted, or spontaneous inspiration by patient 11, segment 48 delivers gas from catheter 38 to patient segment 54. Expiratory segment 50 delivers exhaled gas (exhaled air) from patient segment 54 to catheter 40. Y-connector 52 connects both segments 48 and 50 to patient segment 54. Patient segment 54 extends from Y-connector 52 to patient 11. The patient segment 54 may include means for pneumatic connection to the patient 11, for example, through the patient 11’s nose, mouth or trachea.

[0040] During inhalation (inspiratory) breathing, breathing air is delivered through patient segment 54 and enters the lungs of patient 11. During exhalation or expiration, exhaled or escaping breathing air leaves the lungs of patient 11 and is received in patient segment 54. Exhaled breathing air is delivered or transferred through patient segment 54, via Y-connector 52, to expiratory segment 50. Although not shown, in other embodiments, ventilation system 10 may include additional devices or systems. For example, in one embodiment, system 10 may also include a nebulizer located between ventilator 12 and inspiratory segment 48 to introduce drugs or anesthetics into the patient's breathing air. In other embodiments, breathing circuit 14 may include components such as a humidifier for humidifying breathing air, a heater for heating breathing air, or a water / vapor trap for removing excess moisture from a specific segment or section of ventilation system 10.

[0041] In some embodiments, the ventilator 12 may also include a carbon dioxide scavenger that removes carbon dioxide from the exhaled air and returns or recirculates the air by guiding this recirculated air to the conduit 18. In one embodiment, the ventilator 12 utilizes a bellows to pressurize the air supplied to the conduit 48. For example, in one embodiment, the ventilator 12 selectively supplies pressurized air to and from the outside of the bellows assembly. During inspiration, the ventilator 12 supplies gas or air to the outside of the bellows, causing the bellows to contract to force the gas within the bellows through the carbon dioxide scavenger and to the breathing circuit 14 and the patient's lungs. During expiration, the gas expelled from the patient's lungs fills the bellows through valve 32. In other embodiments, the carbon dioxide scavenger and the bellows may be omitted.

[0042] Ventilation system 10 includes multiple non-invasive sensors for determining the onset of breathing (e.g., spontaneous breathing) in patient 11. In some embodiments, some sensors may be invasive (e.g., EMG sensors on a tube passing through the mouth). Sensor 16 includes a flow sensor to measure the flow rate and direction of gas or air within a channel (e.g., patient segment 54). Sensor 17 includes a pressure sensor to measure the pressure from the flow through the channel (e.g., patient segment 54). In some embodiments, sensors 16, 17 are part of a device configured to sense or detect the pressure and / or flow rate (with or without additives) of air corresponding to forced, spontaneous, or assisted inhalation and exhalation by patient 11. In the illustrated example, sensor 16 is located within patient segment 54. In other embodiments, sensor 16 may be located in other locations. For example, in other embodiments, sensor 16 may be positioned as part of the patient 11's inhalation and exhalation interfaces. Sensors 16 and 17 provide signals (e.g., flow and pressure signals) to provide feedback to controller 36.

[0043] Ventilation system 10 includes an EMG sensor 56 (e.g., an EMG patch or surface sensor) disposed on the skin of patient 11 adjacent to the upper airway muscles (e.g., adjacent to the posterior cricoarytenoid muscle group in the back of the neck). EMG sensor 56 measures the action potentials of the respiratory muscles. EMG sensor 56 provides an EMG signal to controller 36. Ventilation system 10 also includes an EMG sensor 58 (e.g., an EMG patch or surface sensor) disposed on the skin of patient 11 adjacent to the intercostal spaces. EMG sensor 58 measures the movement of the thoracic / abdominal cavity and / or diaphragm. EMG sensor 58 provides an EMG signal to controller 58. In some embodiments, the EMG sensor may be invasive (e.g., via a catheter through the mouth). In some embodiments, ventilation system 10 includes one or more ECG sensors 60 disposed on the patient's skin (e.g., in the chest and / or intercostal spaces). ECG sensor 60 measures cardiac electrical activity. ECG sensor 60 provides an ECG signal to controller 58. In some embodiments, the ventilation system 10 includes one or more additional sensors 62 (e.g., ultrasonic, piezoelectric, and / or inductive sensors) disposed on the patient's skin in the thoracic cavity or intercostal space. The sensors 62 measure the movement of the thoracic / abdominal cavity and / or diaphragm. The sensors 62 provide signals to the controller 36.

[0044] As discussed in more detail below, each sensor may be associated with a different start-triggered algorithm for determining the onset of breathing in patient 11. In some embodiments, data collected from the sensors may be fused together to determine the onset of breathing in patient 11. The sensors may be connected to controller 36 via a wired or wireless connection. Measurements from the sensors are synchronized.

[0045] Figure 2 It is to process signals to determine coupling to the ventilation system (e.g., Figure 1 A schematic diagram of the onset of spontaneous breathing in a patient using a ventilation system 10. Signal processing used to determine the onset of spontaneous breathing in a patient can be performed via several modules (e.g., stored in a processing or computing device, e.g., stored in a system such as...). Figure 1 The events occur either within and executed by the controller 36 in a processing or computing device, or stored in and executed by a remote processing or computing device coupled to the controller 36. These modules include a data acquisition module 64, a signal preprocessing module 66, a calibration module 68, a calibration scheduler module 70, and a start-trigger detection module 72. In some embodiments, these modules may further include a respiratory feature detection module 74 operating between the signal preprocessing module 66 and the start-trigger detection module 72.

[0046] Data acquisition module 64 (via signal) collects or receives data from various sensors coupled to the patient (who is coupled to the ventilation system). Figure 1 The data is from a sensor (in the sensor module). The sensor is coupled to the data acquisition module 64 via a wired or wireless connection. The data acquisition module 64 is configured to synchronize the measurement results from the sensor.

[0047] Signal preprocessing module 66 receives signals from data acquisition module 64. Signal preprocessing module 66 is configured to reduce noise and / or remove motion artifacts from the signal to improve the signal-to-noise ratio. In some embodiments, signal preprocessing module 66 is configured to estimate or measure the occurrence of ECG activity in the signal (e.g., signals from surface EMG sensors and / or ECG sensors). The estimated or measured ECG occurrence can be utilized in the start-trigger detection module to reduce or avoid erroneous triggering of spontaneous breathing start detection due to ECG spikes. Machine learning algorithms, peak detection algorithms, and / or QRS composite algorithms can be used to estimate or measure ECG occurrence. In some embodiments, algorithms utilizing windowed fast Fourier transform, ECG rate, or wavelet algorithms can be used to estimate or measure ECG occurrence. In some embodiments, the estimation or measurement of ECG occurrence can occur in start-trigger detection module 72 (e.g., when respiratory feature detection module 74 is being used).

[0048] Calibration module 68 is configured to calibrate or determine parameters for a start detection algorithm using signals from various sensors to detect the onset of spontaneous breathing in a patient. Examples of parameters include thresholds, window sizes, and variations in EMG baseline entropy. Calibration takes into account patient-to-patient variability and changes in patient status (e.g., sleep, respiratory rate variations, etc.). Calibration via calibration module 68 may include initially (and during recalibration) using data from pressure and flow sensors as the truth or baseline for a given time period, while determining or selecting parameters for a start detection algorithm using sensors other than pressure and flow sensors (e.g., EMG sensors). In some embodiments, the initial parameters may be used during calibration for start detection algorithms of other sensors (e.g., EMG sensors), which may be used to select updated parameters for these start detection algorithms. Optimization, dynamic thresholding, and / or maximum likelihood are some of the techniques used during calibration. The aim of calibration is to minimize the following: average response time (e.g., for providing respiratory support in response to the detection of the onset of spontaneous breathing), response time variations, the number of false triggers, and the number of missed detections.

[0049] The calibration scheduler module 70 is configured to command the calibration module 68 to recalibrate / calculate the parameters of the start detection algorithm of the start trigger detection module 72. Calibration scheduling can be time-based. For example, calibration can occur after a set time (e.g., 2 hours). Calibration scheduling can also be event-triggered. For example, a significant change in one or more physiological parameters, a change in receiver settings (e.g., from medical personnel), or a trigger parameter (e.g., based on the difference between a threshold obtained using a dynamic thresholding method and the current threshold).

[0050] The start-trigger detection module 72 is configured to detect the onset of spontaneous breathing (e.g., inspiration) in signals from each sensor. Different algorithms can be used to detect the onset of spontaneous breathing in signals from different sensor types. For example, for EMG signals, a combination of moving average, fixed sample entropy, or the Hodges-Teager-Kaiser Energy (TKE) operator with thresholding (e.g., single or multiple) can be used for start-on detection. For signals from pressure and flow sensors, thresholding (e.g., single or multiple) can be used for start-on detection. For signals from piezoelectric or inductive sensors, moving average, rate of change, and thresholding can be used for start-on detection.

[0051] In some embodiments, each sensor-based trigger for each sensor is associated with its own confidence score based on multiple factors (e.g., noise level in the underlying data, threshold, and rate of change of the signal). The start trigger detection module 72 is configured to determine the onset of spontaneous breathing and provide a signal (e.g., a final start trigger signal) to the ventilation system (e.g., to a controller) to provide respiratory support, thereby synchronizing the respiratory support with the onset of spontaneous breathing. In some embodiments, the start trigger detection module 72 may provide a start trigger signal when the start trigger for one of the sensors reaches and / or exceeds a minimum confidence score.

[0052] In some implementations, to improve accuracy and robustness, the start-trigger detection module is configured to combine or fuse information from various sensors and utilize a trial-and-error approach when determining to provide a final start-trigger signal to the ventilation system. For example, a majority-based approach can be used, where the start-trigger detection module 72 provides a final start-trigger signal once a majority of signals from the sensors detect the onset of spontaneous breathing. The trial-and-error approach can reduce false triggering. Examples of trial-and-error approaches may include skipping samples, analyzing slopes, ignoring inspiratory phases from the ventilator, or considering any health-specific abnormalities in the patient.

[0053] In some implementations, to improve accuracy and robustness, the start-trigger detection module is configured to multiply the signals from two different sensors before utilizing the start-trigger detection algorithm. For example, after signal conditioning, the flow signal from the flow sensor can be multiplied with the EMG envelope signal before searching for the start-trigger.

[0054] These methods reduce false triggering by decreasing robustness to measurement disturbances. Additionally, they are able to utilize lower thresholds for faster start-up detection. Although the lower threshold reduces confidence, the fusion of information from the sensors generally improves confidence.

[0055] The respiratory feature detection module 74 uses sensor-specific algorithms to extract useful respiratory features from signals from each sensor. In EMG signals, examples of respiratory features include average signal strength, energy, entropy, randomness, or peak frequency. For signals from flow and pressure sensors, examples of respiratory features include scaled signal amplitude, rate of change, or trend estimation. Examples of algorithms used for extracting respiratory features include moving average (for average signal strength), fixed sample entropy, Hodges-TKEO (for signal energy), and numerical methods such as the finite difference method for calculating the rate of change of the signal. Different algorithms can be used for different sensors when extracting respiratory features. For example, for an EMG signal from a first EMG sensor, two different algorithms can be applied to the EMG signal (e.g., moving average and sample entropy), while average entropy is applied to the EMG signal from a second EMG sensor, and trend estimation is applied to the flow signal from a flow sensor. Extracting respiratory features involves detecting the EMG envelope. Parameters of the respiratory feature extraction algorithm (e.g., moving average window size) can be predetermined or obtained in real-time from the calibration module 68.

[0056] As described above, in an embodiment utilizing the respiratory feature detection module 74, the start detection module 72 is configured to estimate the timing of ECG occurrence from EMG signals from an EMG sensor. The start detection module 72 combines (e.g., fused information) multiple respiratory features extracted from the sensor signals and ECG occurrence information upon detecting the start of spontaneous breathing. Examples of usable fusion logic may include combining a majority-based approach using thresholding (i.e., where the start of spontaneous breathing is detected by a majority of signals from the sensor). Another example of usable fusion logic includes thresholding and weighted averaging. Another example of usable fusion logic includes maximum likelihood or another confidence measure. Yet another example of fusion logic includes multiplying EMG signals from multiple locations with a pressure / flow rate signal (or the rate of change of pressure / flow rate). The advantages of using information fusion include reduced false triggering and the ability to detect spontaneous breathing events more quickly.

[0057] Figure 3 This is a flowchart of a method 76 for patient personalization and real-time adaptation for spontaneous breathing initiation detection (e.g., for patients coupled to a ventilation system and monitored using non-invasive sensors). Method 76 may be processed or computed by a ventilation system processing or computing system (e.g., Figure 1The method 76 is executed by a remote processing or computing system of the controller 36 or a controller coupled to the ventilation system. Method 76 may initially be performed for a set time period (e.g., 20 minutes) when the patient is coupled to the ventilation system and various sensors (e.g., pressure sensors, flow sensors, EMG sensors, etc.), or during recalibration in response to reaching a predetermined recalibration time (e.g., 2 hours), a change in physiological parameters (e.g., a significant change exceeding a specific threshold), or a received external input. Method 76 includes receiving pressure and flow signals, respectively, from the pressure and flow sensors coupled to the patient (box 78). Method 76 also includes receiving signals from one or more sensors coupled to the patient (one or more EMG sensors, inductive sensors, piezoelectric sensors, etc.), which measure different physiological parameters from the pressure and flow sensors (box 80). Method 76 further includes detecting the onset of spontaneous breathing (e.g., a respiratory event) of the patient based on the pressure and flow signals (box 82). Method 76 further includes synchronizing the provision of respiratory support to the patient via the ventilation system with the onset of spontaneous breathing detected using the pressure signal and the flow signal (box 84). Method 76 also includes calibrating or determining parameters and thresholds to be used to detect the onset of spontaneous breathing based on signals from the one or more sensors while synchronizing the provision of respiratory support using the pressure signal and the flow signal (box 86). After calibration or recalibration, method 76 further includes switching after calibration to: detecting the onset of spontaneous breathing (e.g., a respiratory event) of the patient based on signals from the one or more sensors (box 88), and synchronizing the provision of respiratory support to the patient via the ventilation system with the onset of spontaneous breathing detected using signals from the one or more sensors (box 90). As described above, the triggering of the provision of respiratory support may be in response to a single signal from a single sensor detecting the onset event or signals from multiple sensors, each detecting the onset event.

[0058] Figure 4 This is a flowchart of method 92 for using respiratory characteristics and estimating ECG occurrence at the onset of spontaneous breathing in a patient. Method 76 can be processed or calculated by a ventilation system (e.g., Figure 1The method is executed by a remote processing or computing system of the controller 36 in the ventilation system or a controller coupled to the ventilation system. Method 92 includes receiving corresponding signals from a flow sensor and / or pressure sensor coupled to the patient (which is coupled to the ventilation system (e.g., a ventilator)) and at least one sensor (e.g., an EMG sensor) coupled to the patient and providing physiological signals different from the flow and pressure signals (box 94). Method 92 also includes extracting respiratory features from the corresponding signals of each of the plurality of sensors (box 96). For example, a respiratory signal can be extracted from each of the pressure / flow signal and the EMG signal. Method 92 also includes estimating the presence of ECG in the EMG signal received from the at least one EMG sensor (box 98). Method 92 also includes detecting the onset of spontaneous breathing in the patient based on the respiratory features extracted from the corresponding signals of the plurality of sensors and the estimated presence of ECG (box 100). Method 92 even includes patient respiratory synchronization based on the detected onset of spontaneous breathing to provide respiratory support to the patient via the ventilation system (box 102).

[0059] Figure 5 This is a schematic diagram of the different processes used to extract respiratory features from signals. These processes can be handled by the ventilation system or by a computational system (e.g., Figure 1 The signal can be executed by a controller 36 in the system or a remote processing or computing system coupled to the controller of the ventilation system. For example, the signal can be an EMG signal, and the respiratory characteristics can be an EMG envelope. The signal can be processed by a signal preprocessing module (e.g., Figure 2 The signal preprocessing module 68) in the middle undergoes a high-pass filter 104, and the filtered signal is then passed to the respiratory feature detection module (e.g., Figure 2 The respiratory feature detection module 74 is used to extract respiratory features. In an energy-based method (e.g., Hodges TKE) 105, the filtered signal is sequentially rectified (box 106), TKE preprocessing (box 108), rectified (box 110), and processed by the Hodges function (box 112). In an entropy-based method (e.g., fixed sample entropy), the filtered signal is processed using a fixed sample entropy algorithm (box 114), from which the rate of change can be utilized (box 116). The respiratory features are then passed to the start-trigger detection module, as indicated by arrow 118.

[0060] As described above, in the implementation scheme for extracting respiratory features from signals, the triggering of the detection module can determine the presence of ECG. Figure 6 This is a schematic diagram of a process used to detect the presence and abnormalities of ECG (e.g., reverse ECG). This process can be handled by the ventilation system's processing or a computing system (e.g., Figure 1The controller 36 in the system or a remote processing or computing system coupled to the controller of the ventilation system is executed. If a peak (i.e., a P-complex wave) is detected in the signal (see...), the controller 36 in the system or a remote processing or computing system coupled to the controller of the ventilation system is executed. Figure 7 (Box 120), then the process includes determining whether a trough value (i.e., Q in the QRS complex, see...) is received. Figure 7 (Box 122). If no valley value is detected, then no ECG is detected (Box 124). If a valley value is received, the process includes determining whether the valley value is less than Q. 阈值 (Box 126). If the trough is low, the process includes determining whether a peak (i.e., the R value of the QRS complex, see...) is detected. Figure 7 (Box 128). If a peak is detected, an anomaly is detected (reverse ECG) (Box 130). If the valley value is not less than Q 阈值 The process then includes determining whether a peak (i.e., R in the QRS complex) is detected (box 132). If a peak is detected, then an ECG is detected (box 134). Figure 8 In the figure, graph 136 shows the ECG occurrence 138 detected in the EMG signal 140. Figure 9 In the figure, graph 142 shows the occurrence of abnormal ECG 144 in EMG signal 146.

[0061] As mentioned above, ECG can be used to minimize false triggers when detecting the onset of a patient's spontaneous breathing. Figure 10 This is a schematic diagram of a process used to detect the onset of spontaneous breathing in a signal (e.g., an EMG signal) using ECG information. This process can be handled by a ventilation system or a computing system (e.g., Figure 1 The controller 36) or a remote processing or computing system coupled to the controller of the ventilation system performs the operation. For example... Figure 10 As shown, multiple thresholds (h1 and h2) can be used at the start of spontaneous breathing detection, where threshold h1 is smaller than threshold h2. A lower threshold h1 improves the start detection time. Figure 10 On the right side, if the signal is greater than h2 (box 148), then the start of spontaneous breathing has definitely been detected (box 150). Figure 10 On the left side, if the signal is greater than threshold h1 but less than h2 (box 152), ECG information is used to determine whether an ECG spike exists in the original signal around that point in the signal (box 154). The presence of ECG allows it to be determined whether the detected start is due to breathing or ECG activity. If an ECG spike is present at that point, no start of breathing is detected (box 156). If no ECG spike is present at that point, it is determined whether the slope is greater than 0 (box 158). If the slope is greater than 0, the start of spontaneous breathing is detected (box 160).

[0062] Figure 11Graph 162 illustrates the detection of the onset of spontaneous breathing using multiple thresholds 166 (h1) and 164 (h2) along with ECG information. Curve 168 represents the normalized signal from a spirometer coupled to the patient. Curve 170 represents the raw EMG signal (multiplied by 3) from an EMG sensor coupled to the patient. Curve 172 represents the respiratory features (fixed sample entropy) extracted from the EMG signal. Curves 174 and 176 represent the onset of spontaneous breathing detected using the respiratory features (fixed sample entropy) detected using thresholds 164 and 166, and the signal from the spirometer, respectively. As shown in the figure... Figure 10 The use of the ECG information minimizes any error triggering because Figure 11 There are no error triggers in the curve 162.

[0063] Figure 12 Graphs 178 and 180 illustrate the detection of the onset of spontaneous breathing in the flow signal compared to the EMG signal. Curve 182 represents the flow signal over time in graph 178 (e.g., from a spirometer coupled to the patient). Line 184 represents the onset of spontaneous breathing in the flow signal. Curve 186 represents the filtered EMG signal over time in graph 180 (e.g., obtained from a surface EMG sensor coupled to the patient). Curve 188 represents the fixed sample entropy of the EMG signal over time in graph 180. Line 190 represents the onset of spontaneous breathing in the EMG signal. The onset of spontaneous breathing was detected earlier in the EMG signal than in the flow signal. Figure 12 In this process, the EMG signal is filtered during preprocessing. For ECG detection in the EMG signal, a windowed Fast Fourier Transform (FFT) is used. For envelope detection in the EMG signal, a fixed sample entropy is used. For thresholding of the EMG signal, multiple thresholds are used. A trial-and-error method is used to trigger the start of respiration in the EMG signal.

[0064] Figure 13 Graph 192 shows the detection of the onset of spontaneous breathing in the flow signal compared to the EMG signal. Curve 194 represents the flow signal over time (e.g., from a spirometer coupled to the patient). Line 196 represents the onset of spontaneous breathing in the flow signal. Curve 198 represents the fixed sample entropy derived over time from the EMG signal (e.g., from a surface EMG sensor coupled to the patient). Line 200 represents the onset of spontaneous breathing in the fixed sample entropy. Curve 202 represents the onset of spontaneous breathing based on the rate of change of entropy (e.g., slope) in the fixed sample entropy. Because the slope of noise and other interference is relatively low, the onset of spontaneous breathing can be detected by setting a threshold for the slope, further reducing the number of false detections. Figure 13In this process, the EMG signal was filtered during preprocessing. For envelope detection in the EMG signal, a fixed sample entropy rate was used. For ECG detection in the EMG signal, pattern recognition was used. For thresholding of the EMG signal, a single threshold was used. A heuristic method was used to trigger the onset of respiration in the EMG signal.

[0065] Many methods or techniques are used to calibrate past breathing thresholds for algorithms that determine the onset of spontaneous breathing. For example, statistical methods can be employed. In statistical methods, EMG and pressure or flow data can be collected over a period of time (e.g., 15 minutes) to calibrate the dataset. Figure 14 Includes graph 204, which has curves 206 and 208 representing the envelope of the raw EMG signal and the EMG signal collected for calibration data over time. Figure 14 It also includes a graph 210, which has a curve 212 representing the flow signal collected for calibration data as a function of time. An EMG processing algorithm (e.g., moving average) with conservative values ​​for algorithm parameters can be used for time normalization of the EMG data and the respirations performed. Figure 15 Examples of time-normalized curves for moving averages of respiration (collected, for example, for different values ​​of algorithm parameters) are shown in graphs 214 and 216. Figure 16 As shown in graph 218, the statistical inter-respiratory variation (e.g., the 25th percentile of EMG processed during respiration) can be calculated. If the variation at the 25th percentile is within tolerance, the average of the 25th percentiles can be used as a threshold. If the 25th percentile is not within tolerance, then the inter-respiratory variation can be calculated using a lower percentile.

[0066] In addition to statistical methods, a modified dynamic thresholding method can be used to calibrate thresholds from past breaths. Multiple breaths can be used to calculate the modified dynamic threshold. In some implementations, more recent breaths can be given more weight when calculating the modified dynamic threshold. In some implementations, the modified dynamic threshold can be used to calibrate thresholds. As an example of using a multi-threshold method along with dynamic thresholds to calibrate thresholds, the average of the modified dynamic thresholds from past breaths can be used as the lower threshold, and the maximum value of the modified dynamic threshold can be used as the upper threshold. In some implementations, a dynamic thresholding method can be used to calibrate a single threshold of a single thresholding method. When the onset of spontaneous breathing is detected in real time on EMG data (e.g., using previously calculated modified dynamic thresholds), the modified dynamic threshold can be calculated. If the difference between the newly calculated dynamic threshold and the previously calculated modified dynamic threshold exceeds a given value, the scheduler (e.g., ...) is calibrated. Figure 2 The calibration scheduler module 70 in the middle) triggers a calibration event, and the calibration module (e.g., Figure 2The calibration module 68) can be used to update or recalibrate the threshold that will be used to start detection.

[0067] Figure 17 This is graph 220 illustrating the use of a modified dynamic threshold in automatic threshold selection on EMG calibration data. Curve 222 represents the fixed sample entropy of the EMG signal. The threshold is calibrated using the envelope of the first 30% of respirations. The upper threshold 224 is set to be greater than the ECG peak amplitude. The lower threshold 226 is set using the amplitude distribution during the expiratory phase. Point 228 represents respiratory offset. Point 230 represents the start of respiration. Point 232 represents other peaks. In some implementations, manual threshold selection can be used (with similar performance). Thresholds can be automatically set for different objects and algorithm parameters.

[0068] When detecting spontaneous breathing, an adaptive approach can be used in selecting parameters (including parameters) of an EMG-based algorithm for an individual patient. Figure 18 This is a schematic diagram of an adaptive method for selecting parameters of an EMG-based initiation detection algorithm for an individual patient. This adaptive method can be implemented by a ventilation system's processing or computational system (e.g., Figure 1 The adaptive method is executed by a controller 36 in the system or a remote processing or computing system coupled to the controller of the ventilation system. The first step of the adaptive method includes generating a result 234. For example, initial data (e.g., calibration data) can be collected by running an EMG-based start-detection algorithm, where parameters vary within a range on the EMG data collected from the patient. Examples of important parameters include window size (W) and threshold (t). Simultaneously, calibration data can be collected using a pressure / flow sensor. The adaptive selection of parameters is based on the patient's initial data. Results can be generated from the results of the algorithm execution. For example, the start-detection time difference (Δt) between the EMG data from the EMG-based start-detection algorithm and the start-detection in the pressure / flow data.

[0069] The next step in the adaptive method involves calculating factor 236 that governs the final parameter selection. One factor calculated is the false breath detection (F). False breath detection is the ratio of false breath detections to true breath detections. Another factor calculated is the mean Δt (avg.Δt), which is the average time difference between the start detection times of all breaths of the patient (between the EMG-based start detection algorithm and the start detection times in the pressure / flow data). Yet another factor calculated is the breath quality factor (Q). The breath quality factor is the quality of a breath relative to other breaths in the same dataset. For the breath quality factor, normalization of one breath relative to other breaths in the same dataset also occurs. Figure 19 The graph 237 shows Q calculated based on initial data collected within a range of parameter (e.g., lower threshold) values. Figure 20The curve 239 shows the avg.Δt calculated based on initial data collected within a range of parameter (e.g., lower threshold) values. Figure 21 The graph 241 shows Q calculated based on initial data collected within a range of parameter (e.g., lower threshold) values.

[0070] After calculating the factors, the adaptive method includes decision 238. During decision-making, the factors calculated above are substituted into the following decision equation:

[0071] Result = min(F*(k1*Q+k2*avg.Δt), (1)

[0072] Where F equals the false detection (W,t) / true detection (W,t) ratio, avg.Δt equals the average (Δt(i)) of all i (i being respirations), and Q equals the average (normalized (Δt(i))) of all i. Finally, the adaptive method involves determining or selecting parameters for patient 240 based on the lowest values ​​of parameters (e.g., W and t) that satisfy the decision outcome (i.e., Equation 1). Figure 22 Graph 242 shows the relationship between F and W within a range of parameter values ​​(e.g., the lower threshold). Δt increases with increasing W. Figure 22 As shown, the number of false detections decreases exponentially with increasing W. Additionally, as... Figure 22 As shown, the level of error detection reaches point 244, where increasing W does not decrease the error detection rate.

[0073] Figure 23 This is a flowchart of method 246 for real-time calibration of parameters used to detect the onset of spontaneous breathing in a patient coupled to a ventilation system. Method 246 may be generated by a ventilation system processing or computing system (e.g., Figure 1The method is executed by a remote processing or computing system of a controller 36 in the system or a controller coupled to the ventilation system. Method 246 includes receiving corresponding signals from a flow sensor, a pressure sensor, and at least one EMG sensor coupled to the patient (who is coupled to the ventilation system (e.g., a ventilator)) (box 248). Method 246 also includes using the EMG signal to detect the onset of spontaneous breathing using an initial range of values ​​for at least one parameter (e.g., a threshold) for initiating the detection algorithm (box 250). Method 246 also includes using at least one of the flow signal and the pressure signal to detect the onset of spontaneous breathing (box 252). Method 246 further includes determining the time difference (Δt) relative to the initial value of at least one parameter at the time of detecting the onset of spontaneous breathing in the EMG signal and at least one of the flow signal and the pressure signal (box 254). Method 246 also includes determining the respiratory quality (Q) from the EMG signal relative to the initial value of at least one parameter (box 256). Method 246 also includes using EMG data and pressure / flow data to determine the false detection rate (F) in the EMG signal relative to the initial value of at least one parameter (box 258). Method 246 even includes at least one updated parameter (e.g., window size) of the start detection algorithm, which will be used to detect the start of spontaneous breathing in the EMG signal, based at least on Δt. The updated parameter may also be based on F and / or Q.

[0074] It should be noted that although various techniques have been discussed for signals from EMG signals, the same techniques can be applied to other physiological signals from other sensors (e.g., piezoelectric sensors).

[0075] The technical advantages of the disclosed implementation include the use of a real-time adaptive and personalized initiation detection algorithm tailored to the patient coupled to the ventilation system (e.g., a mechanical ventilator) to detect the onset of spontaneous breathing in the patient. The disclosed implementation can provide faster and more accurate detection of the patient's respiratory effort. Furthermore, the disclosed implementation can improve synchronization between the ventilator and the patient.

[0076] Referring to the technology presented herein and protected by the claims, and applying it to physical objects and concrete examples of practical nature, which explicitly improves the present art, it is therefore not abstract, intangible, or purely theoretical. Furthermore, if any claim appended to the end of this specification contains one or more elements designated as “means for [performing]…” or “steps for [performing]…”, such elements are intended to be interpreted pursuant to Section 35, Section 112(f) of the USC. However, for any claim containing elements designated in any other manner, such elements are not intended to be interpreted pursuant to Section 35, Section 112(f) of the USC.

[0077] This written description uses examples to disclose the subject matter, including best practices, and also enables those skilled in the art to practice the invention, including making and using any apparatus or system and performing any included methods. The patent scope of this subject matter is defined by the claims and may include other examples that would occur to those skilled in the art. Such other examples are intended to fall within the scope of the claims if they have structural elements that are not indistinguishable from the literal language of the claims, or if they include equivalent structural elements that have minor differences from the literal language of the claims.

Claims

1. A ventilation system, the ventilation system comprising a processor, the processor being configured to: Receive pressure signals from a pressure sensor coupled to the patient and / or flow signals from a flow sensor, respectively; The system receives signals from one or more sensors coupled to the patient, which measure different physiological parameters from the pressure sensor and the flow sensor. The onset of spontaneous breathing in the patient is detected based on the pressure signal and / or the flow signal; The provision of respiratory support to the patient via the ventilation system is synchronized with the initiation of spontaneous breathing detected using the pressure signal and / or the flow signal; When providing respiratory support is synchronized using the pressure signal and the flow signal, the parameters and thresholds used to detect the onset of spontaneous breathing are calibrated based on the signals from the one or more sensors. After calibration, switch to: The initiation of spontaneous breathing in the patient is detected based on the signals from the one or more sensors; as well as The provision of respiratory support to the patient via the ventilation system is synchronized with the initiation of spontaneous breathing detected using signals from the one or more sensors. The one or more sensors mentioned above include at least one EMG (electromyography) sensor. The processor is configured to estimate or measure the presence of ECG (electrocardiogram) in the EMG signal received from the at least one EMG sensor. Detecting the onset of spontaneous breathing in the patient based on signals from the one or more sensors includes: using an estimated ECG occurrence to determine whether the detected onset is due to breathing or due to the ECG occurrence.

2. The ventilation system according to claim 1, wherein the processor is further configured to: In response to changes in a patient's physiological parameters exceeding a certain threshold, switch to: The initiation of the patient's spontaneous breathing is detected based on the pressure signal and / or the flow signal; Synchronize the provision of respiratory support to the patient via the ventilation system with the initiation of spontaneous breathing detected using the pressure signal and / or the flow signal; and When using the pressure signal and / or the flow signal to synchronize the provision of respiratory support, the parameters and the thresholds are recalibrated.

3. The ventilation system according to claim 1, wherein the processor is further configured to: After the set time period, switch to: The initiation of the patient's spontaneous breathing is detected based on the pressure signal and the flow signal; Synchronize the provision of respiratory support to the patient via the ventilation system with the initiation of spontaneous breathing detected using the pressure signal and / or the flow signal; and When using the pressure signal and / or the flow signal to synchronize the provision of respiratory support, the parameters and the thresholds are recalibrated.

4. The ventilation system of claim 1, wherein the one or more sensors comprise a plurality of EMG (electromyography) sensors.

5. The ventilation system of claim 4, wherein the plurality of EMG sensors include a first surface EMG sensor disposed on the patient adjacent to the upper airway muscles and a second surface EMG sensor disposed on the patient adjacent to the intercostal spaces.

6. The ventilation system of claim 1, wherein the one or more sensors include one or more of a piezoelectric sensor, an electrocardiogram sensor, and an ultrasound sensor.

7. The ventilation system of claim 1, wherein the one or more sensors comprise a plurality of sensors, and wherein detecting the initiation of the patient's spontaneous breathing based on signals from the one or more sensors comprises: The initiation of spontaneous breathing is detected in the corresponding signals from the plurality of sensors.

8. The ventilation system of claim 7, wherein detecting the start of the spontaneous breathing in corresponding signals comprises: The start of the spontaneous breathing is determined using a different start detection algorithm for each corresponding signal.

9. The ventilation system of claim 7, wherein respiratory support is provided when the start of spontaneous breathing is detected in at least two signals from the plurality of sensors.

10. A ventilation system, the ventilation system comprising: Multiple sensors configured to be coupled to a patient and generate signals related to the patient's respiratory function, wherein the multiple sensors include a flow sensor, a pressure sensor, and at least one EMG (electromyography) sensor; A memory that encodes processor-executable routines; A processor configured to access the memory and execute processor-executable routines, wherein the routines, when executed by the processor, cause the processor to: Respiratory features are extracted from the corresponding signals of each of the plurality of sensors; Estimate or measure the presence of ECG (electrocardiogram) in the EMG signal received from the at least one EMG sensor; as well as The onset of spontaneous breathing in the patient is detected based on the respiratory features extracted from the corresponding signals of the multiple sensors and the estimated or measured ECG findings. The detection of the onset of spontaneous breathing in the patient based on the respiratory features extracted from the corresponding signals of the plurality of sensors and the estimated ECG occurrence includes: using the estimated ECG occurrence to determine whether the onset detected in one or more of the respiratory features is due to breathing or due to the ECG occurrence.

11. The ventilation system of claim 10, wherein the routine, when executed by the processor, causes the processor to provide respiratory support to the patient via the ventilation system upon detecting the start of the spontaneous breathing.

12. The ventilation system of claim 11, wherein respiratory support is provided to the patient when the start of spontaneous breathing is detected in at least two respiratory features of the corresponding signals from the plurality of sensors.

13. The ventilation system of claim 10, wherein detecting the initiation of the patient's spontaneous breathing based on the respiratory features extracted from the corresponding signal comprises: The respiratory characteristics are compared with at least one threshold to determine the onset of spontaneous breathing.

14. The ventilation system of claim 13, wherein detecting the initiation of the patient's spontaneous breathing based on the respiratory features extracted from the corresponding signal comprises: Determine whether the respiratory feature exceeds a first threshold; When the respiratory feature exceeds the first threshold, the estimated ECG occurrence is used to determine whether the start detected in the respiratory feature is due to breathing or ECG occurrence. And when no ECG is present when the respiratory characteristics exceed the first threshold, it is determined that the initiation of spontaneous breathing is occurring.

Images