Respiratory state adjustment method, device and computer equipment

Abstract

The application relates to a respiratory state adjustment method, device and computer equipment. The method comprises the following steps: acquiring a physiological signal of a target object in a preset historical time period when being in a continuous positive airway pressure mode, and determining a characteristic value of the physiological signal; predicting the characteristic value through a pre-constructed respiratory state prediction model to obtain a respiratory prediction state of the target object at a next moment, wherein the respiratory state prediction model is used for predicting multiple respiratory states, the multiple respiratory states comprise an easy-to-occur respiratory event state, a stable sleep state and a micro-awakening state; predicting a respiratory state change condition of the target object according to the respiratory prediction state, and adjusting an operation pressure of the respiratory equipment according to the respiratory state change condition. The method can reduce the occurrence of respiratory events, reduce the harm caused to the user's body due to various respiratory events, and improve the rationality of pressure adjustment of the breathing machine.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical devices, in particular to a respiration state adjustment method and device and a computer device. BACKGROUND

[0002] The ventilator has an APAP (Auto-adjusting Positive Airway Pressure) mode. In the APAP mode, the ventilator monitors the user's respiration in real time. When the airway is blocked or not unobstructed, the ventilator actively increases the pressure to support the collapsed or falling soft tissue, so as to achieve airway unobstructed. When there is no apnea, the pressure of the ventilator is automatically reduced.

[0003] In the related art, the APAP mode of the ventilator is to adjust the pressure after the occurrence of a respiratory event, and the blood oxygen has decreased and caused damage to the human body. The pressure adjustment method is unreasonable. SUMMARY

[0004] Therefore, it is necessary to provide a respiration state adjustment method, device, computer device and storage medium capable of reducing the occurrence of respiratory events, reducing the damage to the user's body caused by various respiratory events, and improving the rationality of the pressure adjustment of the ventilator.

[0005] In a first aspect, the present application provides a respiration state adjustment method, comprising:

[0006] When in a continuous positive airway pressure mode, obtaining a physiological signal of a target object in a preset historical time period, and determining a characteristic value of the physiological signal;

[0007] predicting the characteristic value by a pre-constructed respiration state prediction model to obtain a respiration prediction state of the target object at a next time; the multiple respiration states include an easy-to-occur respiratory event state, a stable sleep state and a micro-awakening state;

[0008] According to the respiration prediction state, predicting the respiration state change of the target object, and adjusting the operating pressure of the respiration device according to the respiration state change.

[0009] In one of the embodiments, the method further comprises:

[0010] When an unpredicted respiratory event occurs in the stable sleep state, storing the characteristic value before the respiratory event;

[0011] After the continuous positive airway pressure mode ends, training the pre-constructed respiration state prediction model according to the stored characteristic value to obtain an optimized respiration state prediction model.

[0012] updating the optimized respiratory state prediction model to the pre-constructed respiratory state prediction model.

[0013] In one of the embodiments, the adjusting the operation pressure of the respiratory equipment according to the respiratory state change of the target object predicted based on the respiratory prediction state comprises:

[0014] when the respiratory prediction state is the respiratory event-prone state and the respiratory state change of the target object predicted is from the stable sleep state to the respiratory event-prone state, increasing the operation pressure of the respiratory equipment at a first change rate until reaching the normal operation pressure corresponding to the respiratory event-prone state;

[0015] after the target object changes from the stable sleep state to the respiratory event-prone state, if the respiratory state change of the target object predicted is from the respiratory event-prone state to the stable sleep state, decreasing the operation pressure of the respiratory equipment at a second change rate.

[0016] In one of the embodiments, the adjusting the operation pressure of the respiratory equipment according to the respiratory state change of the target object predicted based on the respiratory prediction state comprises:

[0017] when the respiratory prediction state is the respiratory event-prone state and the respiratory state change of the target object predicted is from the stable sleep state to the respiratory event-prone state, increasing the operation pressure of the respiratory equipment at a first change rate;

[0018] when the respiratory event occurs during the operation pressure increasing process, determining a to-be-increased pressure value of the respiratory equipment according to the severity of the respiratory event, after the respiratory event ends, increasing the operation pressure of the respiratory equipment according to the to-be-increased pressure value, and continuing to increase the operation pressure of the respiratory equipment to obtain a first target pressure of the respiratory equipment, the first target pressure being greater than the normal operation pressure corresponding to the respiratory event-prone state;

[0019] after the target object changes from the stable sleep state to the respiratory event-prone state, if the respiratory state change of the target object predicted is to keep the respiratory event-prone state, controlling the respiratory equipment to keep the first target pressure;

[0020] when the respiratory state change of the target object predicted is from the respiratory event-prone state to the stable sleep state, decreasing the operation pressure of the respiratory equipment according to the to-be-increased pressure value to obtain a second target pressure of the respiratory equipment, the second target pressure being greater than the minimum operation pressure.

[0021] In one of the embodiments, the method further comprises: predicting a respiratory state change of the target object according to the respiratory prediction state; and adjusting the operating pressure of the respiratory device according to the respiratory state change, wherein the adjusting the operating pressure of the respiratory device according to the respiratory state change comprises:

[0022] When the respiratory prediction state is the respiratory event-prone state, and the respiratory state change of the target object is predicted to be a transition from the stable sleep state to the respiratory event-prone state, the operating pressure of the respiratory device is increased at the first change rate until the normal operating pressure corresponding to the respiratory event-prone state is reached.

[0023] After the target object transitions from the stable sleep state to the respiratory event-prone state, if the target object has a respiratory event in the respiratory event-prone state, a pressure value to be increased of the respiratory device is determined according to a severity of the respiratory event; after the respiratory event ends, the operating pressure of the respiratory device is increased according to the pressure value to be increased, to obtain a first target pressure of the respiratory device.

[0024] When the respiratory state change of the target object is predicted to be a maintenance of the respiratory event-prone state, the respiratory device is controlled to maintain the first target pressure.

[0025] When the respiratory state change of the target object is predicted to be a transition from the respiratory event-prone state to the stable sleep state, the operating pressure of the respiratory device is decreased according to the pressure value to be increased, to obtain a second target pressure of the respiratory device, the second target pressure being greater than the minimum operating pressure.

[0026] In one of the embodiments, the method further comprises: predicting a respiratory state change of the target object according to the respiratory prediction state; and adjusting the operating pressure of the respiratory device according to the respiratory state change, wherein the adjusting the operating pressure of the respiratory device according to the respiratory state change comprises:

[0027] When the respiratory prediction state is the stable sleep state, and the respiratory state change of the target object is predicted to be a maintenance of the stable sleep state, the operating pressure of the respiratory device is decreased.

[0028] When an unexpected respiratory event occurs in the stable sleep state, a pressure value to be increased of the respiratory device is determined according to a severity of the respiratory event.

[0029] After the respiratory event ends, the operating pressure of the respiratory device is increased according to the pressure value to be increased, and the operating pressure of the respiratory device is continuously decreased until the minimum operating pressure is reached.

[0030] In one of the embodiments, the method further comprises: predicting a respiratory state change of the target object according to the respiratory prediction state; and adjusting the operating pressure of the respiratory device according to the respiratory state change, wherein the adjusting the operating pressure of the respiratory device according to the respiratory state change comprises:

[0031] When the breathing prediction state is the micro-awakening state, and the breathing state change of the target object is predicted to be from the stable sleep state to the micro-awakening state, the operation pressure of the breathing device is reduced in the exhalation phase of the target object, and the current operation pressure is maintained in the inhalation phase of the target object.

[0032] In one of the embodiments, the breathing state change of the target object is predicted according to the breathing prediction state, and the operation pressure of the breathing device is adjusted according to the breathing state change, including:

[0033] When the breathing prediction state is the micro-awakening state, and the breathing state change of the target object is predicted to be from the state prone to respiratory events to the micro-awakening state, after the transition to the micro-awakening state, the operation pressure of the breathing device is reduced in the exhalation phase of the target object, and the current operation pressure is maintained in the inhalation phase of the target object.

[0034] In a second aspect, the present application also provides a breathing state adjustment device, including:

[0035] The feature value acquisition module is configured to acquire the physiological signal of the target object in a preset historical time period when the target object is in the continuous positive airway pressure mode, and determine the feature value of the physiological signal.

[0036] The state prediction module is configured to predict the feature value by a pre-constructed breathing state prediction model to obtain the breathing prediction state of the target object at the next moment, and the plurality of breathing states include the state prone to respiratory events, the stable sleep state, and the micro-awakening state.

[0037] The pressure adjustment module is configured to predict the breathing state change of the target object according to the breathing prediction state, and adjust the operation pressure of the breathing device according to the breathing state change.

[0038] In a third aspect, the present application also provides a computer device, including a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0039] The feature value of the physiological signal of the target object in a preset historical time period is acquired when the target object is in the continuous positive airway pressure mode.

[0040] The feature value is predicted by a pre-constructed breathing state prediction model to obtain the breathing prediction state of the target object at the next moment, and the plurality of breathing states include the state prone to respiratory events, the stable sleep state, and the micro-awakening state.

[0041] The breathing state change of the target object is predicted according to the breathing prediction state, and the operation pressure of the breathing device is adjusted according to the breathing state change.

[0042] In a fourth aspect, the present application also provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the following steps:

[0043] acquiring physiological signals of the target object in a preset historical time period when in the continuous positive airway pressure mode, and determining characteristic values of the physiological signals;

[0044] predicting the characteristic values by a pre-constructed respiratory state prediction model to obtain a respiratory prediction state of the target object at a next moment; the multiple respiratory states include a state prone to respiratory events, a stable sleep state, and a micro-awakening state;

[0045] predicting a respiratory state change of the target object according to the respiratory prediction state, and adjusting an operating pressure of the respiratory device according to the respiratory state change.

[0046] In a fifth aspect, the present application also provides a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement the following steps:

[0047] acquiring physiological signals of the target object in a preset historical time period when in the continuous positive airway pressure mode, and determining characteristic values of the physiological signals;

[0048] predicting the characteristic values by a pre-constructed respiratory state prediction model to obtain a respiratory prediction state of the target object at a next moment; the multiple respiratory states include a state prone to respiratory events, a stable sleep state, and a micro-awakening state;

[0049] predicting a respiratory state change of the target object according to the respiratory prediction state, and adjusting an operating pressure of the respiratory device according to the respiratory state change.

[0050] The respiratory state adjustment method, device, computer device, storage medium, and computer program product described above, when in the continuous positive airway pressure mode, predict the characteristic values of the physiological signals of the target object by a pre-constructed respiratory state prediction model to obtain a respiratory prediction state of the target object at a next moment, which not only can predict multiple states of a sleep stage, but also can predict a state prone to respiratory events. According to the respiratory prediction state, a respiratory state change of the target object is predicted, and the operating pressure of the respiratory device is adjusted in advance according to the respiratory state change, which can reduce the occurrence of respiratory events, reduce the harm to the target object caused by the respiratory events, and reasonably adjust the operating pressure of the respiratory device. At the same time, when the micro-awakening state is predicted, the operating pressure of the respiratory device is adjusted, so that the operating pressure is more in line with the human respiratory law. BRIEF DESCRIPTION OF DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0052] Figure 1 A diagram showing an application environment of a respiratory state adjustment method according to an embodiment;

[0053] Figure 2 Schematic diagram of a flow chart of a respiratory state adjustment method according to an embodiment;

[0054] Figure 3 FIG1 is a flow chart of a model optimization step after the continuous positive airway pressure mode ends in one embodiment;

[0055] Figure 4 Schematic diagram of the structure of the forward propagation stage of the BP neural network in one embodiment;

[0056] Figure 5 is a schematic diagram of a neuron computing process in one embodiment;

[0057] Figure 6 A schematic diagram of adjusting treatment pressure in a respiratory event prone state with no respiratory event occurring in one embodiment;

[0058] Figure 7 A schematic diagram of treatment pressure adjustment under continuous state changes in one embodiment;

[0059] Figure 8 A schematic diagram of adjusting treatment pressure when a respiratory event occurs during a state transition in one embodiment;

[0060] Figure 9 is a schematic diagram of adjusting treatment pressure when a respiratory event occurs in a state prone to respiratory events in one embodiment;

[0061] Figure 10 A schematic diagram of adjusting the therapeutic pressure of a respiratory device during a continuous stable sleep state according to one embodiment;

[0062] Figure 11 is a schematic diagram of adjusting the therapeutic pressure of a respiratory device when a respiratory event occurs during a stable sleep state in one embodiment;

[0063] Figure 12 A schematic diagram of adjusting treatment pressure when a micro-arousal state occurs in one embodiment;

[0064] Figure 13 A schematic diagram of adjusting treatment pressure when a state prone to respiratory events transitions to a state of minimal arousal in one embodiment;

[0065] Figure 14 FIG. 4 is a structural block diagram of a respiratory state adjustment device in one embodiment. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0067] In the related art, the APAP mode of the ventilator adjusts the pressure after a respiratory event occurs, when the blood oxygen level has dropped and damage has been caused to the human body. This has had a negative impact on the user, the response is relatively delayed, and most respiratory events are caused by increased airway resistance. At the same time, the pressure adjustment method in the related art is unreasonable.

[0068] Based on the above problems, a respiratory state adjustment method is proposed, which can predict a state prone to respiratory events due to increased airway resistance, adjust pressure in advance, reduce the occurrence of respiratory events, reduce the impact of adverse consequences caused by various respiratory events, and reasonably adjust pressure.

[0069] The respiratory state adjustment method provided in the embodiment of the present application can be applied to Figure 1The respiratory device shown in the figure includes a respiratory pressure support device, a sensor, a control system, a display screen, and a storage device. The user represents a target subject. The display screen can be used to display parameters of the continuous positive airway pressure mode, such as the minimum treatment pressure, maximum treatment pressure, pressure adjustment sensitivity, pressure reduction level for minimal arousal states, and pressure increase level for respiratory event prone states. The continuous positive airway pressure mode in this application can adjust the treatment pressure based on the predicted respiratory state. The sensor is used to acquire the target subject's physiological signals within a preset historical time period while in continuous positive airway pressure mode and transmit them to the control system. The control system is used to determine the characteristic values ​​of the physiological signals and predict various respiratory states based on the characteristic values ​​using a pre-established respiratory state prediction model to obtain the predicted respiratory state of the target subject at the next moment; the various respiratory states include a respiratory event prone state, a stable sleep state, and a minimal arousal state. Based on the predicted respiratory state, the target subject's respiratory state changes are predicted, and the operating pressure provided by the respiratory pressure support device, i.e., the treatment pressure, is adjusted accordingly. The storage device is used to store usage data during the respiratory adjustment process and the characteristic values ​​used for model training. The respiratory device in this embodiment may be, but is not limited to, a sleep apnea machine.

[0070] In an exemplary embodiment, Figure 2 As shown, a breathing state adjustment method is provided, which is applied to Figure 1 Taking the respiratory device in FIG. 1 as an example, the method includes the following steps 202 to 206. Among them:

[0071] Step 202 : When in continuous positive airway pressure mode, obtain physiological signals of the target subject within a preset historical time period and determine characteristic values ​​of the physiological signals.

[0072] The target subject refers to a user of a respiratory device, and the physiological signal refers to a parameter related to the respiratory state of the target subject.

[0073] Optionally, when the respiratory device is in continuous positive airway pressure mode, it collects physiological signals of the target object within a preset historical time period through sensors, and transmits the physiological signals of the target object within the preset historical time period to the control system. The preset historical time period is a fixed-length time window, that is, a sliding time window. For example, the amplitudes of the pressure and flow signals in the last minute are collected and stored as two one-dimensional arrays of 1000 sampling points, and the characteristic values ​​of these two arrays are calculated. Physiological signals may include flow changes and pressure changes during sleep stages. Furthermore, physiological signals may also include blood oxygen, pulse, etc. The control system determines the characteristic values ​​of the physiological signals. The characteristic values ​​include but are not limited to the mean value, variance, difference between adjacent peaks, zero crossing point, deviation mean value, peak-to-valley difference, autocorrelation coefficient, etc.

[0074] For example, the calculation process of the eigenvalue is as follows:

[0075] average value:

[0076]

[0077] in, is the number of sampling points in the sliding time window, For the sampling point.

[0078] variance:

[0079]

[0080] in, is the number of sampling points in the sliding time window, is the sampling point, is the average value of the sampling points in the sliding time window.

[0081] Difference between adjacent peaks:

[0082]

[0083] in, is the number of peaks in the sliding time window, is the amplitude of the i-th peak, is the amplitude of the i+1th peak.

[0084] Zero crossing:

[0085] ZCP

[0086] in, is the i-th sampling point, is the i+1th sampling point.

[0087] Deviation from the mean:

[0088]

[0089] in, is the number of sampling points in the sliding time window, is the sampling point, is the average value of the sampling points in the sliding time window.

[0090] Peak-to-valley difference:

[0091]

[0092] in, is the number of peaks and valleys in the sliding time window, For the crest, For the trough.

[0093] Autocorrelation coefficient:

[0094] R = ∑ i=1 N-k [ x i × x i+k ] / [ ∑ i=1 N-k x i 2 × x i+k 2 ]

[0095] Where k is the time delay, is the i-th sampling point, is the i+kth sampling point.

[0096] Optionally, the sensor may include a flow sensor and a pressure sensor, which are respectively used to collect flow changes and pressure changes of the target object within a preset historical time period.

[0097] In step 204, a pre-built respiratory state prediction model is used to predict multiple respiratory states of the characteristic values ​​to obtain the predicted respiratory state of the target object at the next moment; the multiple respiratory states include a state prone to respiratory events, a stable sleep state, and a slightly aroused state.

[0098] This application divides the predicted states of the target subject when using a respiratory device into: a state prone to respiratory events, a stable sleep state, and a micro-arousal state. Among them, the state prone to respiratory events is due to the prediction that the target subject's airway resistance is very high during sleep, and it is very easy to have respiratory events caused by excessive airway resistance, such as obstructive sleep apnea events, hypopnea events, airflow limitation events, etc. The stable sleep state is a stage in which breathing is relatively stable during sleep, with normal blood oxygen concentration and stable breathing. The micro-arousal state is a brief, partial brain awakening of the target subject during sleep, but it is not enough to cause full wakefulness. However, if the treatment pressure is too high when using a respiratory device, it may also cause the target subject to wake up.

[0099] Optionally, the control system inputs the characteristic values ​​of the physiological signal into a pre-built respiratory state prediction model, performs predictions of multiple respiratory states, and obtains the predicted respiratory state of the target object at the next moment. The pre-built respiratory state prediction model may be, but is not limited to, a BP neural network (Back Propagation Neural Network), a convolutional neural network, a recurrent neural network, etc. Taking the BP neural network as an example, the BP neural network may include an input layer, a hidden layer, and an output layer. The input layer is used to receive the characteristic values ​​of the physiological signal. The hidden layer is used to extract and transform the characteristic values. The output layer is used to output the prediction result of the network, that is, the predicted respiratory state of the target object at the next moment. The layers are fully connected, that is, each neuron in the input layer is connected to each neuron in the hidden layer, and each neuron in the hidden layer is connected to each neuron in the output layer.

[0100] Optionally, when setting the hidden layer, three hidden layers are set by comprehensively considering factors such as performance and accuracy. The hidden layer contains a maximum of 30 neurons, which generate output signals to the output layer after nonlinear transformation, thereby achieving the accuracy requirements of respiratory state prediction with a smaller number of layers and performance occupancy.

[0101] Step 206 : Predict changes in the target subject's respiratory state based on the predicted respiratory state, and adjust the operating pressure of the respiratory device based on the changes in the respiratory state.

[0102] Optionally, after obtaining the predicted respiratory state, the subject's respiratory state change can be determined based on the subject's current respiratory state. Different respiratory state changes may require different operating pressure adjustment methods. The operating pressure of the respiratory device is adjusted accordingly based on the different respiratory state changes. Operating pressure refers to the therapeutic pressure.

[0103] Optionally, when the predicted breathing state indicates a state prone to respiratory events, the operating pressure of the respiratory device is increased. When the predicted breathing state indicates a stable sleep state, the operating pressure of the respiratory device is decreased. When the predicted breathing state indicates a state of minimal arousal, the operating pressure is decreased during the subject's exhalation phase.

[0104] In the above-mentioned respiratory state adjustment method, when in continuous positive airway pressure mode, the characteristic values ​​of the target object's physiological signals are predicted by a pre-built respiratory state prediction model to predict the state prone to respiratory events, stable sleep state, and micro-arousal state, and the target object's respiratory prediction state at the next moment is obtained. This method can not only predict multiple states of the sleep stage, but also predict the state prone to respiratory events. Predicting the changes in the target object's respiratory state based on the respiratory prediction state, and adjusting the operating pressure of the respiratory equipment in advance based on the changes in the respiratory state can reduce the occurrence of respiratory events, reduce the damage caused to the target object by respiratory events, and reasonably adjust the operating pressure of the respiratory equipment. At the same time, when a micro-arousal state is predicted, the operating pressure of the respiratory equipment is adjusted to make the operating pressure more consistent with the human body's breathing rules.

[0105] In an exemplary embodiment, Figure 3 As shown, the above method further includes: a model optimization step after the continuous positive airway pressure mode ends, which includes steps 302 to 306.

[0106] Step 302: When an unpredicted respiratory event occurs in a stable sleep state, the feature value before the respiratory event is stored.

[0107] Step 304 : After the continuous positive airway pressure mode ends, the pre-built respiratory state prediction model is trained according to the stored characteristic values ​​to obtain an optimized respiratory state prediction model.

[0108] Step 306, update the optimized respiratory state prediction model to the pre-constructed respiratory state prediction model.

[0109] Optionally, if an unexpected respiratory event occurs in the stationary sleep state, the feature values before the respiratory event are stored for model training. After the treatment ends, i.e., the continuous positive airway pressure mode ends, the respiratory state prediction model is trained by the stored feature values to obtain a respiratory prediction result. The respiratory prediction result is compared with the true result through the output layer of the respiratory state prediction model to obtain the error between the respiratory prediction result and the true result. When the training end condition is not met, the model parameters are updated through the back propagation algorithm so that the model can better fit the feature values. After multiple iterations of training, when the training end condition is met, the optimized respiratory state prediction model is obtained. The optimized respiratory state prediction model is updated to the pre-constructed respiratory state prediction model for respiratory state prediction.

[0110] Exemplarily, when the pre-constructed respiratory state prediction model is a bp neural network, the working process of the bp neural network mainly includes two stages: forward propagation and back propagation. The input data in the forward propagation stage passes through the input layer, the hidden layer, and finally reaches the output layer, as shown in Figure 4 wherein x includes x1, x2, …, x m , y includes y1, y2, and y3, and y is the output respiratory prediction result. The neurons in each layer perform weighted summation on the input data and generate output through an activation function, as shown in Figure 5 The activation function can be represented as:

[0111]

[0112] wherein x i is the input data, is the weight, is the activation function, is the output respiratory prediction result. In this application, the activation function uses ReLU (Rectified Linear Unit) and softmax function (normalized exponential function).

[0113] In the back propagation stage, the error between the respiratory prediction result output by the respiratory state prediction model and the true result is calculated, the error is propagated in the reverse direction along the network connection in the respiratory state prediction model, and the weights and thresholds of the respiratory state prediction model are adjusted according to the error. The adjustment of the weights can adopt the gradient descent method, so that the total error of the respiratory state prediction model gradually decreases. The error between the output of the respiratory state prediction model and the true result is calculated by the following loss function:

[0114]

[0115] wherein Loss is a loss function, is a true result, is an output respiratory prediction result in forward propagation.

[0116] In order to minimize the error between the respiratory prediction result output by the respiratory state prediction model and the true result, the minimum value of the loss function is to be found. The iterative method for finding the minimum value of the loss function is to find the numerical solution by gradient descent and iteration. The iterative method generally needs to be repeated several times, because the finding of the minimum value of the loss function may need to be repeated several times, and in each iteration, the weights between nodes of each layer will be updated iteratively. The iterative process of the weight is as follows:

[0117]

[0118] wherein, denotes the weight of the next iteration, denotes the weight of the current iteration, is a learning rate, denotes the gradient of the loss function Loss with respect to the parameter ω.

[0119] The present application proposes an intelligent adjustment of the continuous positive airway pressure mode, realizes the respiratory state prediction model only on the breathing device, continuously optimizes the model, saves data during treatment, trains the respiratory state prediction model after the end of treatment, and as the target object continues to use the breathing machine device, the model will be more adapted to the target object used, and more accurate respiratory state prediction is provided.

[0120] Among the easy-to-occur respiratory event state, the stable sleep state and the micro-awakening state, the stable sleep state is the basic state and is maintained for the longest time. In the treatment process, the treatment pressure will be adjusted according to different future states.

[0121] For the easy-to-occur respiratory event state, because the airway resistance of the target object will continue to increase in this state, it is easy to cause the occurrence of respiratory events, so the treatment pressure needs to be increased on the basis of the current treatment pressure to offset the airway resistance and maintain the smoothness of the respiratory airway. In the stable sleep state, the respiratory state prediction model predicts that the respiratory state of the target object changes, and the state changes from the stable sleep state to the easy-to-occur respiratory event state. At this time, the treatment pressure of the breathing device is adjusted, and there are the following situations: no respiratory event occurs, continuous state change, respiratory event occurs in the state transition process, and respiratory event occurs in the easy-to-occur respiratory event state stage.

[0122] In an exemplary embodiment, predicting changes in the respiratory state of a target object based on a respiratory prediction state, and controlling the operating pressure of the respiratory apparatus based on the changes in the respiratory state include: when the respiratory prediction state is a state prone to respiratory events, and the target object's respiratory state change is predicted to be a transition from a stable sleep state to a state prone to respiratory events, increasing the operating pressure of the respiratory apparatus at a first rate of change until reaching the normal operating pressure corresponding to the state prone to respiratory events; after the target object transitions from a stable sleep state to a state prone to respiratory events, if it is predicted that the target object's respiratory state change is to maintain the state prone to respiratory events, controlling the respiratory apparatus to maintain the normal operating pressure corresponding to the state prone to respiratory events; when the target object's respiratory state change is predicted to be a transition from a state prone to respiratory events to a stable sleep state, reducing the operating pressure of the respiratory apparatus at a second rate of change until reaching the minimum operating pressure.

[0123] The first rate of change refers to the rate of increase in treatment pressure when transitioning from a stable sleep state to a state prone to respiratory events. The normal operating pressure corresponding to the state prone to respiratory events refers to the treatment pressure when no respiratory events occur. The second rate of change refers to the rate of decrease in treatment pressure when transitioning from a state prone to respiratory events to a stable sleep state.

[0124] This embodiment is to adjust the treatment pressure for the state prone to respiratory events without respiratory events, such as Figure 6 As shown in the figure, the sliding window is used to calculate the feature value and input it into the input layer of the respiratory state prediction model. When the respiratory state is predicted to be prone to respiratory events, and the target subject's respiratory state is predicted to change from a stable sleep state to a state prone to respiratory events (②→①), the treatment pressure is gradually increased in advance according to the first change rate. This allows the airway to remain open due to the higher treatment pressure, thereby eliminating the occurrence of respiratory events. Respiratory events in this case may be caused by sleeping posture or different sleep stages.

[0125] After the target subject changes from a stable sleep state to a state prone to respiratory events, if it is predicted that the target subject's respiratory state changes to remain in the state prone to respiratory events, the respiratory device is controlled to maintain a normal treatment pressure corresponding to the state prone to respiratory events.

[0126] After the target subject ends the state prone to respiratory events and returns to a stable sleep state (①→②), the treatment pressure is slowly reduced according to the second change rate to return to the treatment pressure before the increase.

[0127] In an exemplary embodiment, predicting changes in the respiratory state of a target object based on a respiratory prediction state, and adjusting the operating pressure of the respiratory apparatus based on the changes in the respiratory state include: when the respiratory prediction state is a state prone to respiratory events, and the target object's respiratory state change is predicted to be from a stable sleep state to a state prone to respiratory events, increasing the operating pressure of the respiratory apparatus at a first change rate until reaching the normal operating pressure corresponding to the state prone to respiratory events; after the target object changes from a stable sleep state to a state prone to respiratory events, if it is predicted that the target object's respiratory change is from a state prone to respiratory events to a stable sleep state, reducing the operating pressure of the respiratory apparatus at a second change rate.

[0128] This embodiment adjusts the treatment pressure according to the continuous state change, such as Figure 7 As shown in the figure, when the predicted respiratory state indicates a high risk of respiratory events, and the target subject's respiratory state is predicted to change from a stable sleep state to a state prone to respiratory events (②→①), the respiratory device's treatment pressure is increased at a first rate of change. After the stable sleep state transitions to a state prone to respiratory events, it briefly transitions back to a stable sleep state (①→②). At this point, the treatment pressure is also decreased at a second rate of change based on the state transition. The treatment pressure is adjusted slowly between different states, and the entire process of increasing or decreasing the treatment pressure can last several minutes.

[0129] In an exemplary embodiment, predicting a change in the respiratory state of a target subject based on a predicted respiratory state, and adjusting the operating pressure of a respiratory device based on the change in the respiratory state includes: when the predicted respiratory state is a state prone to respiratory events, and the predicted change in the respiratory state of the target subject is a transition from a stable sleep state to a state prone to respiratory events, increasing the operating pressure of the respiratory device at a first rate of change; when a respiratory event occurs during the operating pressure increase process, determining a pressure value to be increased for the respiratory device based on the severity of the respiratory event; after the respiratory event ends, increasing the operating pressure of the respiratory device based on the pressure value to be increased, and continuing to increase the operating pressure of the respiratory device to obtain a first target pressure of the respiratory device, the first target pressure being greater than the normal operating pressure corresponding to the state prone to respiratory events; after the target subject transitions from a stable sleep state to a state prone to respiratory events, if the predicted change in the respiratory state of the target subject is to remain in the state prone to respiratory events, controlling the respiratory device to maintain the first target pressure; and when the predicted change in the respiratory state of the target subject is a transition from a state prone to respiratory events to a stable sleep state, reducing the operating pressure of the respiratory device based on the pressure value to be increased to obtain a second target pressure of the respiratory device, the second target pressure being greater than the minimum operating pressure.

[0130] The first target pressure refers to the treatment pressure required for the respiratory device to transition to a state prone to respiratory events if a respiratory event occurs during the state transition. The second target pressure refers to the treatment pressure required to return to a stable sleep state if a respiratory event occurs during the state transition.

[0131] This embodiment is to adjust the treatment pressure when a respiratory event occurs during the state transition, such as Figure 8 As shown, when the predicted respiratory state is a state prone to respiratory events, and the target subject's respiratory state is predicted to change from a stable sleep state to a state prone to respiratory events (②→①), the treatment pressure of the respiratory device is increased according to the first change rate.

[0132] When a respiratory event occurs during the process of slowly increasing the treatment pressure, the respiratory device determines the pressure value to be increased based on the severity of the respiratory event. Figure 8 After an event occurs, the treatment pressure is increased to the pressure value to be increased, and the slow increase in pressure is not affected, but the first target pressure reached will be higher than the normal treatment pressure when no respiratory event occurs.

[0133] After transitioning from a state prone to respiratory events back to a stable sleep state (①→②), the slowly reduced and stabilized treatment pressure will also increase by a certain amount based on the previous treatment pressure value. Figure 8 Taking a 0.5 cm water column increase as an example, in the previous stable sleep state, the treatment pressure is 4 cm water column. After the intermediate state transition and respiratory events, the treatment pressure maintained in the stable sleep state is 4.5 cm water column.

[0134] In an exemplary embodiment, predicting a change in the respiratory state of a target subject based on a predicted respiratory state, and adjusting the operating pressure of a respiratory apparatus based on the change in the respiratory state includes: when the predicted respiratory state is a state prone to respiratory events, and the predicted change in the respiratory state of the target subject is a transition from a stable sleep state to a state prone to respiratory events, increasing the operating pressure of the respiratory apparatus at a first rate of change until reaching a normal operating pressure corresponding to the state prone to respiratory events; after the target subject transitions from a stable sleep state to a state prone to respiratory events, if the target subject experiences a respiratory event in the state prone to respiratory events, determining a pressure value to be increased for the respiratory apparatus based on the severity of the respiratory event; after the respiratory event ends, increasing the operating pressure of the respiratory apparatus based on the pressure value to be increased to obtain a first target pressure of the respiratory apparatus; when the predicted change in the respiratory state of the target subject is a state prone to respiratory events, controlling the respiratory apparatus to maintain the first target pressure; and when the predicted change in the respiratory state of the target subject is a transition from a state prone to respiratory events to a stable sleep state, reducing the operating pressure of the respiratory apparatus based on the pressure value to be increased to obtain a second target pressure of the respiratory apparatus, wherein the second target pressure is greater than the minimum operating pressure.

[0135] This embodiment is to adjust the treatment pressure when a respiratory event occurs in a state prone to respiratory events, such as Figure 9 As shown, when the predicted respiratory state is a state prone to respiratory events, and the target subject's respiratory state is predicted to change from a stable sleep state to a state prone to respiratory events (②→①), the treatment pressure of the respiratory device is increased according to the first change rate until it reaches the normal treatment pressure corresponding to the state prone to respiratory events.

[0136] When a respiratory event occurs during the process of slowly increasing the treatment pressure, the respiratory device determines the pressure value to be increased based on the severity of the respiratory event. Figure 9 After an event occurs in the treatment, the treatment pressure is increased to the pressure value to be increased.

[0137] After transitioning from a state prone to respiratory events back to a stable sleep state (①→②), the slowly decreasing, stabilized treatment pressure will be increased by a certain amount based on the previous value. Furthermore, the feature values ​​used to predict respiratory state before the respiratory event are stored in a storage device. After treatment, the model is optimized based on this stored data to make the model prediction more accurate.

[0138] In the above embodiment, adopting corresponding treatment pressure adjustment strategies for different situations prone to respiratory events can better reduce the AHI index (Apnea Hypopnea Index, the average number of apneas plus hypopneas per hour of sleep), reduce the number of respiratory events during sleep, and thus reduce the damage to the target subject's body caused by respiratory events.

[0139] During a stable sleep state, the target patient's airway remains unobstructed and respiratory events are unlikely to occur. The current treatment pressure only needs to be sufficient to maintain airway patency. During a stable sleep state, the following situations may occur: sustained stable sleep and respiratory events occurring during a stable sleep state.

[0140] In an exemplary embodiment, predicting changes in the respiratory state of a target object based on a respiratory prediction state, and controlling the operating pressure of the respiratory apparatus based on the changes in the respiratory state include: when the respiratory prediction state is a stable sleep state, and when it is predicted that the changes in the respiratory state of the target object are to maintain a stable sleep state, reducing the operating pressure of the respiratory apparatus until a minimum operating pressure is reached.

[0141] This embodiment is to adjust the therapeutic pressure of the respiratory device in a continuous and stable sleep state. Figure 10 As shown, if the subject remains in a stable sleep state for an extended period, the respiratory device's treatment pressure will decrease slowly until it reaches the minimum treatment pressure. Some of the feature values ​​used to predict respiratory states during the stable sleep state are also stored in the storage device and used to optimize the model for more accurate predictions of other states.

[0142] In an exemplary embodiment, predicting changes in the respiratory state of a target object based on a respiratory prediction state, and adjusting the operating pressure of the respiratory apparatus based on the changes in the respiratory state include: when the respiratory prediction state is a stable sleep state, and the target object's respiratory state changes are predicted to maintain a stable sleep state, reducing the operating pressure of the respiratory apparatus; when an unpredicted respiratory event occurs in the stable sleep state, determining a pressure value to be increased for the respiratory apparatus based on the severity of the respiratory event; after the respiratory event ends, increasing the operating pressure of the respiratory apparatus based on the pressure value to be increased, and continuing to reduce the operating pressure of the respiratory apparatus until the minimum operating pressure is reached.

[0143] This embodiment is to adjust the treatment pressure of the respiratory device when a respiratory event occurs in a stable sleep state. Figure 11 As shown, when the predicted respiratory state is a stable sleep state and the target subject's respiratory state change is predicted to maintain a stable sleep state, the treatment pressure of the respiratory device is reduced.

[0144] When an unforeseen respiratory event occurs during a stable sleep state, the pressure level to be increased is determined based on the severity of the event. After the event ends, the treatment pressure is increased to the desired level, while the ventilator's treatment pressure is continuously reduced until the minimum treatment pressure is reached. The feature values ​​used to predict respiratory status before the event are stored in a storage device. After treatment ends, the model is optimized based on this stored data.

[0145] In the above embodiment, corresponding treatment pressure adjustment strategies are adopted according to different situations of the stable sleep state, so that the treatment pressure can be adjusted more reasonably.

[0146] A microarousal state refers to a brief, partial awakening of the brain during sleep, but not enough to cause full wakefulness. The following situations may occur during a microarousal state: a microarousal state occurs and a state prone to respiratory events transitions to a microarousal state.

[0147] In an exemplary embodiment, predicting changes in the respiratory state of a target subject based on a respiratory prediction state, and adjusting the operating pressure of the respiratory device based on the changes in the respiratory state include: when the respiratory prediction state is a micro-arousal state, and the target subject's respiratory state is predicted to change from a stable sleep state to a micro-arousal state, controlling the respiratory device to reduce the operating pressure during the target subject's exhalation phase, and maintaining the current operating pressure during the target subject's inhalation phase.

[0148] This embodiment adjusts the treatment pressure when a micro-awakening state occurs. Figure 12 As shown, the respiratory prediction state model predicts that the target subject is about to enter a state of micro-arousal. In the state of micro-arousal, it is easy for the target subject's comfort to be reduced due to high treatment pressure, and there is a possibility that the target subject will wake up. Therefore, after the micro-arousal state is predicted (②→③), the treatment pressure will be adjusted according to the target subject's inhalation and exhalation. The current treatment pressure will be maintained during the inhalation stage, and the treatment pressure will be reduced to a certain extent during the exhalation stage. When it is predicted that the target subject will transition from a micro-arousal state to a stable sleep state, the treatment pressure will be restored to the inhalation stage. Figure 12 For example, the inhalation phase is 5 cmH2O and the exhalation phase is 4 cmH2O. Lowering the treatment pressure during the exhalation phase will make the target subject's exhalation smoother, improve the target subject's comfort, and reduce the possibility of the target subject being awakened by high treatment pressure.

[0149] In an example embodiment, the adjusting the operation pressure of the respiratory device according to the respiratory state change of the target object predicted according to the respiratory prediction state comprises: when the respiratory prediction state is the micro-awakening state, and the respiratory state change of the target object predicted is a transition from the easy-to-respiratory-event state to the micro-awakening state, after the transition to the micro-awakening state, controlling the respiratory device to reduce the operation pressure in the exhalation phase of the target object, and maintaining the current operation pressure in the inhalation phase of the target object.

[0150] The embodiment is to adjust the treatment pressure when the easy-to-respiratory-event state transitions to the micro-awakening state. As shown in FIG. 5, if the target object transitions from the easy-to-respiratory-event state to the micro-awakening state (①→③), the treatment pressure will start to be adjusted according to the inhalation and exhalation of the target object in the process of slowly decreasing after the transition to the micro-awakening state. The treatment pressure is maintained in the inhalation phase, and the treatment pressure is reduced in the exhalation phase. Figure 13

[0151] In the embodiment, when it is predicted that the target object enters the micro-awakening state, the treatment pressure is adjusted according to the switching of the respiratory phase. The current treatment pressure is maintained in the inhalation phase of each breath, and the treatment pressure is appropriately reduced in the exhalation phase. This can improve the comfort of the target object when using the respiratory device, and enable the respiratory device to accurately adjust the treatment pressure according to the human respiratory rhythm.

[0152] It should be understood that, although each step in the flowchart involved in each of the above embodiments is displayed in sequence according to the arrow, these steps are not necessarily executed in sequence according to the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other sequences. Moreover, at least part of the steps in the flowchart involved in each of the above embodiments can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or stages.

[0153] Based on the same inventive concept, the embodiments of the present application also provide a respiratory state adjustment device for implementing the above-mentioned respiratory state adjustment method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, and therefore the specific limitations in one or more respiratory state adjustment device embodiments provided below can refer to the limitations of the respiratory state adjustment method described above, which will not be described here again.

[0154] In an example embodiment, as shown in FIG. 6, the method for adjusting the respiratory state of the target object comprises the following steps. Figure 14 ​As shown, a respiratory state adjustment apparatus is provided, comprising: a characteristic value acquisition module 1402, a state prediction module 1404, and a pressure adjustment module 1406, wherein:

[0155] The characteristic value acquisition module 1402 is configured to acquire a physiological signal of a target object in a predetermined historical time period when in a continuous positive airway pressure mode, and determine a characteristic value of the physiological signal.

[0156] The state prediction module 1404 is configured to predict the characteristic value by a pre-constructed respiratory state prediction model to obtain a respiratory prediction state of the target object at a next time point, wherein the respiratory prediction state comprises an easy-to-occur respiratory event state, a stable sleep state, and a micro-awakening state.

[0157] The pressure adjustment module 1406 is configured to predict a respiratory state change of the target object according to the respiratory prediction state, and adjust an operating pressure of a respiratory device according to the respiratory state change.

[0158] In an exemplary embodiment, the apparatus further comprises:

[0159] A model optimization module is configured to store the characteristic value before a respiratory event when an unexpected respiratory event occurs in the stable sleep state, train the pre-constructed respiratory state prediction model according to the stored characteristic value after the continuous positive airway pressure mode ends, obtain an optimized respiratory state prediction model, and update the optimized respiratory state prediction model as the pre-constructed respiratory state prediction model.

[0160] In an exemplary embodiment, the pressure adjustment module 1406 is further configured to, when the respiratory prediction state is the easy-to-occur respiratory event state and the respiratory state change of the target object is predicted to be from the stable sleep state to the easy-to-occur respiratory event state, increase the operating pressure of the respiratory device at a first change rate until a normal operating pressure corresponding to the easy-to-occur respiratory event state is reached, and after the target object changes from the stable sleep state to the easy-to-occur respiratory event state, if the respiratory change of the target object is predicted to be from the easy-to-occur respiratory event state to the stable sleep state, decrease the operating pressure of the respiratory device at a second change rate.

[0161] In an exemplary embodiment, the pressure adjustment module 1406 is further configured to, when the predicted respiratory state is a state prone to respiratory events and the target subject's respiratory state is predicted to change from a stable sleep state to a state prone to respiratory events, increase the operating pressure of the respiratory apparatus at a first rate of change; when a respiratory event occurs during the operating pressure increase process, determine a pressure value to be increased for the respiratory apparatus based on the severity of the respiratory event; after the respiratory event ends, increase the operating pressure of the respiratory apparatus based on the pressure value to be increased, and continue to increase the operating pressure of the respiratory apparatus to obtain a first target pressure of the respiratory apparatus, wherein the first target pressure is greater than the normal operating pressure corresponding to the state prone to respiratory events; after the target subject changes from a stable sleep state to a state prone to respiratory events, if the target subject's respiratory state is predicted to remain in the state prone to respiratory events, control the respiratory apparatus to maintain the first target pressure; when the target subject's respiratory state is predicted to change from a state prone to respiratory events to a stable sleep state, reduce the operating pressure of the respiratory apparatus based on the pressure value to be increased to obtain a second target pressure of the respiratory apparatus, wherein the second target pressure is greater than the minimum operating pressure.

[0162] In an exemplary embodiment, the pressure adjustment module 1406 is further configured to, when the predicted respiratory state is a state prone to respiratory events and the target subject's respiratory state is predicted to change from a stable sleep state to a state prone to respiratory events, increase the operating pressure of the respiratory apparatus at a first change rate until the normal operating pressure corresponding to the state prone to respiratory events is reached; after the target subject changes from a stable sleep state to a state prone to respiratory events, if the target subject has a respiratory event in the state prone to respiratory events, determine a pressure value to be increased for the respiratory apparatus based on the severity of the respiratory event; after the respiratory event ends, increase the operating pressure of the respiratory apparatus based on the pressure value to be increased to obtain a first target pressure of the respiratory apparatus; when it is predicted that the target subject's respiratory state will remain in the state prone to respiratory events, control the respiratory apparatus to maintain the first target pressure; when it is predicted that the target subject's respiratory state will change from a state prone to respiratory events to a stable sleep state, reduce the operating pressure of the respiratory apparatus based on the pressure value to be increased to obtain a second target pressure of the respiratory apparatus, and the second target pressure is greater than the minimum operating pressure.

[0163] In an exemplary embodiment, the pressure adjustment module 1406 is also used to reduce the operating pressure of the respiratory device when the predicted respiratory state is a stable sleep state and the target subject's respiratory state change is predicted to maintain a stable sleep state; when an unpredicted respiratory event occurs in the stable sleep state, determine the pressure value to be increased of the respiratory device according to the severity of the respiratory event; after the respiratory event ends, increase the operating pressure of the respiratory device according to the pressure value to be increased, and continue to reduce the operating pressure of the respiratory device until the minimum operating pressure is reached.

[0164] In an exemplary embodiment, the pressure adjustment module 1406 is further configured to control the respiratory device to reduce the operating pressure during the target subject's exhalation phase and to maintain the current operating pressure during the target subject's inhalation phase when the predicted respiratory state is a micro-arousal state and the target subject's respiratory state is predicted to change from a stable sleep state to a micro-arousal state.

[0165] In an exemplary embodiment, the pressure adjustment module 1406 is also used to control the respiratory device to reduce the operating pressure during the target subject's exhalation phase and maintain the current operating pressure during the target subject's inhalation phase when the predicted respiratory state is a micro-arousal state and the target subject's respiratory state is predicted to change from a state prone to respiratory events to a micro-arousal state after the transition to the micro-arousal state.

[0166] Each module in the respiratory state adjustment device can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in the computer device in the form of software, so that the processor can call and execute the corresponding operations of each module.

[0167] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0168] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0169] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0170] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0171] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), data processing logic devices based on quantum computing, and the like.

[0172] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0173] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the following steps are implemented: When in continuous positive airway pressure mode, obtaining a physiological signal of the target subject within a preset historical time period and determining a characteristic value of the physiological signal; The characteristic value is used to predict multiple respiratory states using a pre-built respiratory state prediction model to obtain the predicted respiratory state of the target subject at the next moment; the multiple respiratory states include a state prone to respiratory events, a stable sleep state, and a slightly aroused state; Predicting changes in the respiratory state of the target object according to the respiratory prediction state, and adjusting the operating pressure of the respiratory equipment according to the changes in the respiratory state, including: when the respiratory prediction state is a micro-arousal state, and the target object's respiratory state is predicted to change from a stable sleep state to a micro-arousal state, controlling the respiratory equipment to reduce the operating pressure during the target object's exhalation phase, and maintaining the current operating pressure during the target object's inhalation phase; when the respiratory prediction state is a micro-arousal state, and the target object's respiratory state is predicted to change from a state prone to respiratory events to a micro-arousal state, after the target object is transformed into the micro-arousal state, controlling the respiratory equipment to reduce the operating pressure during the target object's exhalation phase, and maintaining the current operating pressure during the target object's inhalation phase.

2. The computer-readable storage medium according to claim 1, wherein When the computer program is executed by a processor, the following steps are implemented: When an unpredicted respiratory event occurs in a stable sleep state, the feature value before the respiratory event is stored; After the continuous positive airway pressure mode ends, the pre-built respiratory state prediction model is trained according to the stored characteristic values ​​to obtain an optimized respiratory state prediction model; The optimized respiratory state prediction model is updated to the pre-built respiratory state prediction model.

3. The computer-readable storage medium according to claim 1, wherein When the computer program is executed by a processor, the following steps are implemented: When the predicted respiratory state is a state prone to respiratory events, and the target subject's respiratory state is predicted to change from a stable sleep state to a state prone to respiratory events, increasing the operating pressure of the respiratory device at a first rate of change until reaching a normal operating pressure corresponding to the state prone to respiratory events; After the target subject transitions from a stable sleep state to a state prone to respiratory events, if it is predicted that the target subject's respiratory change is transitioning from a state prone to respiratory events to a stable sleep state, the operating pressure of the respiratory device is reduced according to a second change rate.

4. The computer-readable storage medium according to claim 1, wherein When the computer program is executed by a processor, the following steps are implemented: When the predicted respiratory state is a state prone to respiratory events, and the target subject's respiratory state is predicted to change from a stable sleep state to a state prone to respiratory events, increasing the operating pressure of the respiratory device at a first change rate; When a respiratory event occurs during the operating pressure increase process, determining a pressure value to be increased for the respiratory device based on the severity of the respiratory event; after the respiratory event ends, increasing the operating pressure of the respiratory device based on the pressure value to be increased, and continuing to increase the operating pressure of the respiratory device to obtain a first target pressure of the respiratory device, wherein the first target pressure is greater than the normal operating pressure corresponding to the state in which a respiratory event is likely to occur; After the target subject changes from a stable sleep state to a state prone to respiratory events, if it is predicted that the change in the target subject's respiratory state is to remain in a state prone to respiratory events, controlling the respiratory device to maintain the first target pressure; When it is predicted that the respiratory state of the target object changes from a state prone to respiratory events to a stable sleep state, the operating pressure of the respiratory device is reduced according to the pressure value to be increased to obtain a second target pressure of the respiratory device, and the second target pressure is greater than the minimum operating pressure.

5. The computer-readable storage medium according to claim 1, wherein When the computer program is executed by a processor, the following steps are implemented: When the predicted respiratory state is a state prone to respiratory events, and the target subject's respiratory state is predicted to change from a stable sleep state to a state prone to respiratory events, increasing the operating pressure of the respiratory device at a first rate of change until reaching a normal operating pressure corresponding to the state prone to respiratory events; After the target subject transitions from a stable sleep state to a state prone to respiratory events, if the target subject experiences a respiratory event in the state prone to respiratory events, determining a pressure value to be increased for the respiratory device based on the severity of the respiratory event, and after the respiratory event ends, increasing the operating pressure of the respiratory device based on the pressure value to be increased to obtain a first target pressure of the respiratory device; When it is predicted that the change in the respiratory state of the target subject is to maintain a state prone to respiratory events, controlling the respiratory device to maintain the first target pressure; When it is predicted that the respiratory state of the target object changes from a state prone to respiratory events to a stable sleep state, the operating pressure of the respiratory device is reduced according to the pressure value to be increased to obtain a second target pressure of the respiratory device, and the second target pressure is greater than the minimum operating pressure.

6. The computer-readable storage medium according to claim 1, wherein When the computer program is executed by a processor, the following steps are implemented: When the predicted respiratory state is a stable sleep state, and the target subject's respiratory state change is predicted to maintain a stable sleep state, reducing the operating pressure of the respiratory device; When an unpredicted respiratory event occurs in a stable sleep state, determining a pressure value to be increased for the respiratory device according to the severity of the respiratory event; After the respiratory event ends, the operating pressure of the respiratory device is increased according to the pressure value to be increased, and the operating pressure of the respiratory device is continuously decreased until the minimum operating pressure is reached.

7. A breathing state adjustment device, characterized in that: The device comprises: a characteristic value acquisition module, configured to acquire physiological signals of a target subject within a preset historical time period and determine characteristic values ​​of the physiological signals when the subject is in a continuous positive airway pressure mode; A state prediction module is used to predict multiple respiratory states of the characteristic value using a pre-built respiratory state prediction model to obtain the predicted respiratory state of the target subject at the next moment; the multiple respiratory states include a state prone to respiratory events, a stable sleep state, and a slightly aroused state; a pressure adjustment module, configured to predict a change in the breathing state of the target subject according to the predicted breathing state, and adjust the operating pressure of the respiratory device according to the change in the breathing state; The pressure adjustment module is further used to control the respiratory device to reduce the operating pressure during the exhalation phase of the target object and to maintain the current operating pressure during the inhalation phase of the target object when the predicted respiratory state is a micro-arousal state and the target object's respiratory state is predicted to change from a stable sleep state to a micro-arousal state; when the predicted respiratory state is a micro-arousal state and the target object's respiratory state is predicted to change from a state prone to respiratory events to a micro-arousal state, after the target object changes to the micro-arousal state, control the respiratory device to reduce the operating pressure during the exhalation phase of the target object and to maintain the current operating pressure during the inhalation phase of the target object.

8. The device according to claim 7, characterized in that The device further comprises: The model optimization module is used to store the characteristic values ​​before the respiratory event when an unpredicted respiratory event occurs in a stable sleep state; after the continuous positive airway pressure ventilation mode ends, train the pre-built respiratory state prediction model according to the stored characteristic values ​​to obtain an optimized respiratory state prediction model; and update the optimized respiratory state prediction model to the pre-built respiratory state prediction model.

9. The device according to claim 7, characterized in that The pressure adjustment module is also used to increase the operating pressure of the respiratory equipment according to a first change rate until the normal operating pressure corresponding to the state prone to respiratory events is reached when the predicted respiratory state is a state prone to respiratory events and the target object's respiratory state is predicted to change from a stable sleep state to a state prone to respiratory events; after the target object changes from a stable sleep state to a state prone to respiratory events, if it is predicted that the target object's respiratory change is from a state prone to respiratory events to a stable sleep state, the operating pressure of the respiratory equipment is reduced according to a second change rate.

10. A computer device comprising the computer-readable storage medium and a processor according to any one of claims 1 to 6, wherein: The processor is configured to execute the computer program.

Images