Multi-modal heart and cerebral hemodynamics monitoring method and device

Abstract

The invention relates to a multi-mode heart and cerebral hemodynamics monitoring method and device, and belongs to the technical field of medical equipment. According to the method, brain absorbance information and pulse wave information at the middle artery of the brain are obtained based on a near infrared spectrum technology by utilizing the relevance among the cardiac output, blood pressure, intracranial pressure and cerebral perfusion pressure of a human body, cardiac output information and cuff-free blood pressure information are obtained based on a thoracic impedance method, and physiological signals are obtained based on the two methods; and carrying out fusion processing to obtain a multi-modal signal, and further constructing a non-invasive monitoring model of non-invasive intracranial pressure, cardiac output, blood pressure and cerebral perfusion pressure, so as to realize continuous non-invasive monitoring reflecting human heart and cerebral hemodynamic parameters. According to the invention, through fusion application of multi-mode signals, the mechanism of cardio-cerebral hemodynamics interaction can be effectively disclosed, and a more effective, safer and noninvasive monitoring means is provided for clinic.

Description

Technical Field

[0001] The present invention belongs to the technical field of medical equipment and relates to a multimodal cardiocerebral hemodynamics monitoring method and device. Background Art

[0002] Cerebral perfusion pressure (CPP) is a key indicator for measuring the state of cerebral hemodynamics. It represents the effective pressure that promotes cerebral blood flow. Its calculation formula is CPP = MAP-ICP (mean arterial pressure minus intracranial pressure). Under normal physiological conditions, CPP needs to be maintained in an appropriate range to ensure that the brain tissue obtains sufficient blood supply and oxygen delivery, thereby maintaining the normal function of brain cells. The subtle changes in this pressure balance reflect the complex interaction between the cardiovascular system and the cerebrovascular system. After craniocerebral injury, the increase in ICP leads to increased cerebrovascular resistance, affecting cerebral perfusion and reduced cerebral blood flow. The blood flow velocity may change to a certain extent, which is manifested as changes in parameters related to cerebral hemodynamics. Hemodynamics is the science of studying the characteristics and regularity of the movement of blood and its components in the body. Its main monitoring indicators include blood pressure, central venous pressure, oxygen delivery, heart rate and cardiac output. Based on the laws of physics and combined with the concepts of physiology and pathophysiology, the regularity of blood movement in the circulatory system is quantitatively, dynamically and continuously measured and analyzed in order to understand the development of the disease and guide clinical treatment. Through hemodynamic monitoring, medical professionals can obtain real-time information on key physiological parameters such as cardiac pumping function, vascular resistance, blood volume status, etc., which is critical for the diagnosis and management of cardiovascular disease, shock, heart failure, and postoperative recovery.

[0003] In the human physiological system, the relationship between cardiopulmonary blood flow and middle cerebral artery blood flow, as well as the interaction between cardiopulmonary vascular and cerebral vascular, constitute a highly coordinated hemodynamic network. Understanding these relationships is of great significance for the development of cardiocerebral hemodynamic monitoring and evaluation methods based on multimodal physiological signals.

[0004] The heart is the core of the circulatory system. It pumps blood through contraction and relaxation to provide oxygen and nutrients to tissues and organs throughout the body, while taking away metabolic waste. Cardiac output (CO) is the amount of blood pumped out by the heart per minute and is a key indicator of cardiac function and hemodynamic status. The size of cardiac output is affected by heart rate and stroke volume (SV), among which stroke volume depends on the contractility, preload and afterload of the myocardium.

[0005] After blood is pumped out of the heart, it enters the systemic circulatory system through the aorta. The aorta is the main channel for cardiac output, and blood enters the head and upper limbs through the aortic arch. The blood supply to the brain mainly comes from two pairs of arterial systems: the internal carotid artery and the vertebral artery. The two form the cerebral artery ring at the base of the brain to ensure that the brain can still get blood supply when different arteries are damaged. After entering the skull, the internal carotid artery is divided into the anterior cerebral artery, the middle cerebral artery, the anterior choroid plexus artery, and the posterior communicating artery. Among them, the middle cerebral artery (MCA) is an important blood vessel that supplies the front and middle parts of the brain. Its blood flow is crucial to maintaining brain function. It is one of the main arteries for brain blood supply and is responsible for supplying most areas on the side of the brain, including the frontal lobe, parietal lobe, and part of the temporal lobe. The effect of increased ICP on cerebral blood flow velocity is most obvious in the middle cerebral artery (MCA). The MCA is the direct continuation of the internal carotid artery and is the largest artery supplying the cerebral hemisphere, supplying about 80% of the blood flow required by the cerebral hemisphere.

[0006] In general, cardiopulmonary function directly affects the blood oxygen supply to the brain. Any cardiopulmonary dysfunction may lead to insufficient blood supply or oxygenation to the brain, which in turn affects brain function. The connection between cardiopulmonary blood flow and middle cerebral artery blood flow, as well as the connection between cardiopulmonary blood vessels and cerebral blood vessels, is a highly coordinated system. By monitoring physiological parameters such as cardiac output, blood pressure, and cerebral blood flow, the state of cardiocerebral hemodynamics can be better understood and evaluated. However, there is currently no method or device for integrated monitoring and evaluation of cardiocerebral hemodynamics. Summary of the invention

[0007] In view of this, the purpose of the present invention is to provide a multimodal cardiocerebral hemodynamic monitoring method and device, which regards the cardiovascular and cerebrovascular system as a whole, and realizes comprehensive monitoring of cardiopulmonary function and brain function through non-invasive means (it can simultaneously realize non-invasive and continuous monitoring of indicators such as cardiac output, blood pressure, intracranial pressure and cerebral perfusion pressure), providing important physiological information support for clinical diagnosis and treatment.

[0008] In order to achieve the above object, the present invention provides the following technical solutions:

[0009] Scenario 1:

[0010] A multimodal cardiocerebral hemodynamic monitoring method specifically comprises the following steps:

[0011] S1: Acquire brain absorbance information and brain pulse wave information based on near-infrared light method, acquire cardiac output information and cuffless blood pressure information based on thoracic impedance method, and then preprocess and extract features of the acquired information;

[0012] S2: Based on the absorbance information and brain pulse wave characteristics, a non-invasive intracranial pressure prediction model M1 is constructed to obtain preliminary non-invasive prediction results of intracranial pressure; based on the chest impedance signal characteristic information and the ECG signal characteristic information, a non-invasive prediction model M2 of cardiac output and cuff-free blood pressure is constructed to obtain preliminary non-invasive prediction results of cardiac output and blood pressure;

[0013] S3: The physiological information features obtained in step S1 based on the near-infrared light method and the thoracic impedance method are fused and processed to obtain multimodal physiological information; based on the preliminary non-invasive prediction results of intracranial pressure, cardiac output and blood pressure obtained in step S2 and the multimodal physiological information, a non-invasive comprehensive monitoring model M3 of intracranial pressure and cerebral perfusion pressure is constructed to obtain accurate prediction results of intracranial pressure and blood pressure, and then obtain non-invasive prediction results of cerebral perfusion pressure, thereby realizing continuous non-invasive monitoring that synchronously reflects the human cardiocerebral hemodynamic state.

[0014] Furthermore, in step S1, the near-infrared light method utilizes a multi-wavelength sensor monitoring probe to monitor the absorbance changes of the frontal lobe of the brain and the pulse wave signal of the middle cerebral artery.

[0015] Furthermore, in step S1, the thoracic impedance method utilizes a human heart hemodynamic parameter monitoring method to obtain a physiological signal reflecting changes in human heart hemodynamics.

[0016] Further, in step S2, a non-invasive intracranial pressure prediction model M1 is constructed, which specifically includes the following steps:

[0017] S201: Using a multi-wavelength sensing monitoring probe, obtain n wavelengths λ1, λ2, ..., λ n Absorbance information at the frontal lobe of the brain under the light source: I = {i λ1 ,i λ2 ,...,i λ n};

[0018] S202: Using the sensor monitoring probe to obtain the pulse wave time domain signal PPG of the middle cerebral artery, and after preprocessing it, obtain the pulse wave characteristic parameters related to the changes in intracranial pressure and hemodynamics of the brain: P = {p1, p2, ..., p m}, where m represents the number of preprocessed pulse wave features;

[0019] S203: Construct a non-invasive intracranial pressure prediction model M1 to obtain non-invasive prediction results of intracranial pressure and mean cerebral arterial pressure: M1(ICP0, MAP0) = f(I, P, PPG), where ICP0 represents the preliminary predicted intracranial pressure and MAP0 represents the preliminary predicted mean cerebral arterial pressure.

[0020] Further, in step S2, a non-invasive prediction model M2 of cardiac output and blood pressure is constructed, which specifically includes the following steps:

[0021] S211: obtaining physiological signals reflecting changes in human cardiac hemodynamics, including thoracic impedance change signals, thoracic impedance change differential signals, basic impedance signals, electrocardiogram signals and pulse wave signals, and preprocessing the signals;

[0022] S212: Extract characteristic parameters related to cardiac output and hemodynamic changes from the preprocessed signals and construct a non-invasive prediction model for cardiac output M21(CO)=f(p1,p2,p3,...,p k ), wherein k represents the number of characteristic parameters related to cardiac output and hemodynamic changes, and CO represents the predicted cardiac output; the characteristic parameters related to cardiac output and hemodynamic changes include periodic information, strength information, electro-mechanical conduction information and energy information of thoracic impedance changes;

[0023] S213: extracting the characteristic transmission time information of the heart activity reflected in the transition from the electrocardiogram signal to the chest impedance change signal, from the electrocardiogram signal to the pulse wave signal, and from the chest impedance signal to the pulse wave signal, and the time, amplitude and energy information of the characteristic changes of the chest impedance change signal and the pulse wave signal from the preprocessed signal, and forming a parameter set of time T, value A and energy P. Based on these three parameter sets, construct a prediction model M22(SBP,DBP)=f(T,A,P) for cuffless blood pressure, where SBP represents systolic pressure and DBP represents diastolic pressure.

[0024] S214: The non-invasive prediction models M21 and M22 obtained by steps S212 and S213 together constitute a non-invasive prediction model M2=(M21, M22) based on the thoracic impedance method.

[0025] Further, in step S3, a non-invasive comprehensive monitoring model M3 for cardiocerebral hemodynamics is constructed, which specifically includes the following steps:

[0026] S301: Fusing the physiological information features obtained based on the near infrared light method and the chest impedance method to obtain multimodal physiological information I fusion ;

[0027] S302: Using the correlation between changes in cardiocerebral hemodynamics, intracranial pressure, blood pressure and cardiac output, multimodal physiological information is added to the non-invasive comprehensive monitoring model M3 for cardiocerebral hemodynamics based on the intracranial pressure values, cerebral hemodynamic characteristics, cardiac output, blood pressure and cardiac hemodynamic characteristics obtained by the prediction models M1 and M2. fusion , construct a non-invasive intracranial pressure prediction model based on multimodal information M31(MAP,ICP)=f(I fusion,ICP0,MAP0,CO,SBP,DBP), so that the outputs of models M1 and M2 affect each other, and improve the output accuracy of cardiac output, blood pressure and intracranial pressure; where MAP represents the mean cerebral arterial pressure predicted by model M31; ICP represents the intracranial pressure predicted by model M31;

[0028] S303: In the non-invasive comprehensive monitoring model M3 for cardiocerebral hemodynamics, based on the result of step S302, a non-invasive prediction model M32 (CPP) = f(CO, MAP, ICP, I fusion ).

[0029] S304: The non-invasive prediction models M31 and M32 obtained from steps S302 and S303 together constitute a non-invasive comprehensive monitoring model M3 = (M31, M32) of intracranial pressure and cerebral perfusion pressure.

[0030] Further, in step S301, the fusion processing is specifically based on the correlation of cardiocerebral hemodynamic parameters, and utilizes a time synchronization mechanism and a weighted fusion algorithm to combine near-infrared spectral signals and chest impedance signals to form multimodal physiological information that can comprehensively reflect the cardiocerebral hemodynamic state.

[0031] Solution 2: A multimodal cardiocerebral hemodynamic monitoring device: including a signal acquisition module, a signal processing module, a data fusion module, a non-invasive monitoring model module and a display and alarm module; the signal acquisition module includes a near-infrared spectral sensor and a chest impedance electrode.

[0032] The near-infrared spectral sensor and chest impedance electrode are used to be connected to the patient's body surface and are responsible for acquiring multimodal physiological signals of the human body; the signal processing module performs denoising, filtering and feature extraction on the acquired signals; the data fusion module fuses the physiological data of different signal sources; the non-invasive monitoring model module integrates the multimodal cardiocerebral hemodynamic monitoring method described in Scheme 1, and generates non-invasive monitoring results based on the fused data; the display and alarm module is used to present the non-invasive monitoring results, i.e., the values ​​of cardiac output, blood pressure, intracranial pressure and cerebral perfusion pressure, in real time, and automatically trigger an alarm when the parameters exceed the safe range.

[0033] First, the near-infrared spectral sensor and chest impedance electrode are correctly connected to the patient's body surface; after starting the equipment, the signal acquisition module begins to acquire physiological signals, and the signal processing unit preprocesses and extracts features of the acquired signals; then, the data fusion module fuses the processed signals, and finally the non-invasive monitoring model module generates non-invasive monitoring results; the monitoring model module integrates the multimodal cardiocerebral hemodynamic monitoring method in Scheme 1; the display and alarm module presents the numerical values ​​and change trend waveforms of cardiac output, blood pressure, intracranial pressure and cerebral perfusion pressure in real time, and automatically triggers an alarm when the parameters exceed the safe range.

[0034] The beneficial effects of the present invention are:

[0035] (1) The present invention fully considers the individual difference information of the subjects and the correlation and integrity of the cardiovascular and cerebrovascular systems, and obtains multimodal physiological signals reflecting the hemodynamic state of the human cardiovascular and cerebrovascular system through the comprehensive application of near-infrared spectroscopy technology and thoracic impedance technology.

[0036] (2) Based on the obtained multimodal physiological signals, the present invention constructs a multi-level non-invasive prediction model for different hemodynamic indicators, which helps to improve the accuracy of non-invasive monitoring of different cardiocerebral hemodynamic indicators while achieving an overall evaluation of cardiocerebral hemodynamic indicators, thus helping clinicians to make more accurate diagnosis and treatment decisions.

[0037] (3) The present invention improves the accuracy and real-time performance of monitoring: by fusing multimodal physiological signals, the device of the present invention can accurately reflect the changes in cardiovascular and cerebral hemodynamic parameters and provide key physiological data to medical staff in real time. Compared with the traditional single signal monitoring method, the multimodal fusion method significantly improves the accuracy of monitoring and reduces the monitoring error caused by interference or loss of a single signal.

[0038] (4) The present invention reduces the pain and monitoring risks of patients: Traditional invasive monitoring methods often require the insertion of catheters or sensors, which not only increases the risk of infection and complications, but also causes discomfort and pain to patients. The present invention monitors in a non-invasive manner, completely avoiding the risks of invasive operations, providing patients with a more comfortable and safe monitoring experience, especially for critically ill patients who require long-term continuous monitoring.

[0039] (5) The present invention provides more quantifiable physiological parameters: The device of the present invention can not only monitor cardiovascular hemodynamic parameters such as cardiac output and blood pressure, but also provide important cerebral hemodynamic indicators such as mean cerebral arterial pressure, intracranial pressure and cerebral perfusion pressure. The provision of these parameters enables doctors to conduct a more comprehensive assessment and diagnosis of the patient's cardiovascular and cerebral blood flow status, thereby formulating a more accurate treatment plan.

[0040] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below in conjunction with the accompanying drawings, wherein:

[0042] Figure 1 This is a flow chart of multimodal cardiocerebral hemodynamic monitoring of the present invention;

[0043] Figure 2 Hardware implementation for cardiocerebral hemodynamic signal acquisition;

[0044] Figure 3 It is the software implementation for multi-modal physiological signal fusion;

[0045] Figure 4 It is a non-invasive prediction model for multi-level cardiocerebral hemodynamic parameters;

[0046] Figure 5 This is a block diagram of a non-invasive monitoring device for cardiovascular and cerebral hemodynamics. DETAILED DESCRIPTION

[0047] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner, and the following embodiments and the features in the embodiments can be combined with each other without conflict.

[0048] See also Figure 1 to Figure 5 The present invention designs a multimodal cardiocerebral hemodynamic monitoring method and device, the purpose of which is to achieve continuous monitoring of cardiocerebral hemodynamic parameters such as cardiac output, blood pressure, intracranial pressure and cerebral perfusion pressure through non-invasive means.

[0049] The method of the present invention utilizes the correlation between human cardiac output, blood pressure, intracranial pressure and cerebral perfusion pressure, obtains brain absorbance information and pulse wave information at the middle cerebral artery based on near-infrared spectroscopy technology, obtains cardiac output information and cuffless blood pressure information based on the thoracic impedance method, and performs fusion processing on the physiological signals obtained based on the two methods to obtain multimodal signals, thereby constructing a noninvasive monitoring model of noninvasive intracranial pressure, cardiac output, blood pressure and cerebral perfusion pressure, thereby realizing continuous noninvasive monitoring of human cardiocerebral hemodynamic parameters.

[0050] like Figure 1 As shown, the device of the present invention obtains human physiological signals by combining near-infrared spectroscopy technology and chest impedance technology, and constructs a multimodal non-invasive monitoring model based on these signals, thereby providing an effective clinical monitoring method.

[0051] In the implementation process, the device of the present invention is mainly composed of a signal acquisition module, a signal processing module, a data fusion module, a non-invasive monitoring model module, and a display and alarm module. Figure 5 As shown in the figure, the signal acquisition module is responsible for acquiring the multimodal physiological signals of the human body. The signal processing unit performs denoising, filtering and feature extraction on the acquired signals. The data fusion module fuses the physiological data of different signal sources. The non-invasive monitoring model module generates non-invasive monitoring results based on the fused data. The display and alarm module is used to display the monitoring results in real time and issue an alarm when necessary.

[0052] 1. Signal acquisition (such as Figure 2 (shown)

[0053] 1) Collection of near-infrared spectral signals

[0054] The present invention adopts near-infrared spectroscopy technology and obtains the absorbance changes of the frontal lobe of the brain and the pulse wave information of the middle cerebral artery through a multi-wavelength sensing monitoring probe. The probe is equipped with multiple near-infrared light sources and is made of flexible materials so as to fit the scalp closely and reduce motion artifacts. The probe is also integrated with a highly sensitive photodiode for receiving the reflected near-infrared light signal. The sensor probes are respectively arranged in the patient's frontal lobe and temporal window, and the multi-wavelength light beam penetrates the scalp. Part of the light is absorbed by the brain tissue, and the unabsorbed light is received by the probe and converted into an electrical signal. After the photoelectric signal is amplified by the preamplifier circuit, it enters the analog-to-digital converter (ADC) for digitization. The digitized signal is input into the embedded processor for subsequent signal processing and feature extraction to obtain absorbance data and pulse wave signals reflecting the hemodynamic state of the brain.

[0055] 2) Collection of chest impedance signals

[0056] The chest impedance measurement adopts a four-electrode configuration, in which two pairs of electrodes are used to apply a constant weak AC current and measure the voltage difference. By measuring the voltage difference and the known current value, the impedance change of the chest cavity can be calculated. The module also includes a high-precision amplifier and filter to amplify and filter out the noise in the measurement signal. After the signal is processed, it also enters the ADC for digital processing. Since the impedance changes of blood and other tissues during cardiac contraction and relaxation are regular, cardiac output and cuffless blood pressure information can be extracted by analyzing these changes. The signal processing unit filters and extracts features from these collected impedance signals to ensure the accuracy and effectiveness of the signal.

[0057] 2. Signal processing and data fusion (such as Figure 3 (shown)

[0058] 1) Signal preprocessing and feature extraction

[0059] The signal processing module preprocesses the signals collected from the near-infrared spectrum and chest impedance, mainly including denoising, baseline correction and feature extraction. For the near-infrared spectrum signal, the waveform characteristics of the absorbance and pulse wave are mainly extracted; for the chest impedance signal, the characteristic signals related to cardiac output and blood pressure are mainly extracted. And relevant physiological parameter features are extracted from the preprocessed signals. For example, the pulse wave amplitude, rise time, and absorbance change in the near-infrared spectrum signal; the ECG R wave spacing and impedance change in the chest impedance signal.

[0060] 2) Multimodal data fusion

[0061] In the data fusion module, signals from different physiological sources are fused. Specifically, the near-infrared spectral signal and the chest impedance signal characteristics are combined using the time synchronization mechanism and the weighted fusion algorithm to form multimodal physiological data that can fully reflect the cardiocerebral hemodynamic state. The fusion process mainly includes the following steps:

[0062] (1) Time-align the signals to ensure that physiological data from different signal sources are processed on the same timeline.

[0063] (2) Based on the correlation between cardiocerebral hemodynamic parameters, a weighted fusion algorithm based on machine learning is applied to combine the feature weights of different signals to generate a unified physiological data output.

[0064] 3. Construction of non-invasive monitoring model (such as Figure 4 (shown)

[0065] 1) Non-invasive intracranial pressure prediction model (M1)

[0066] Based on the absorbance information and brain pulse wave characteristics obtained from near-infrared spectroscopy, the present invention constructs a non-invasive intracranial pressure prediction model, called the M1 model. The original intention of the design of this model is to achieve non-invasive real-time monitoring of intracranial pressure, thereby avoiding the risks and discomfort caused by traditional invasive detection methods. The core of the M1 model is to use machine learning algorithms to deeply analyze and process the collected absorbance information and pulse wave characteristics. These physiological signals are non-invasively obtained through near-infrared spectroscopy technology, reflecting the changes in brain hemodynamics. Through feature extraction, the key information in the signal is extracted as the input of the M1 model. In the training stage of the model, the M1 model uses a large amount of historical data from the clinic, which covers the intracranial pressure measurements of different patient groups and the corresponding near-infrared spectra and pulse wave signal characteristics. The trained M1 model can make real-time predictions for unknown patients in a new clinical environment. By inputting the absorbance information and brain pulse wave characteristics collected in real time, the model can quickly output non-invasive measurements of intracranial pressure. This predicted value can not only reflect the patient's current intracranial pressure status, but also help clinicians obtain important physiological information without performing invasive operations, thereby providing a basis for the patient's treatment and management.

[0067] 2) Non-invasive cardiac output and blood pressure prediction model (M2)

[0068] The construction of the non-invasive cardiac output and blood pressure prediction model (M2) is based on the cardiac hemodynamic signals obtained by thoracic impedance technology. These signals contain rich physiological information, and through signal processing technology, characteristic data closely related to cardiac output and blood pressure can be extracted from them. Specifically, the thoracic impedance signal can reflect the hemodynamic changes of each heart beat, including the impedance changes during ventricular contraction and relaxation. This information is an important basis for predicting cardiac output and blood pressure.

[0069] In order to accurately predict cardiac output and blood pressure, the M2 model uses machine learning algorithms to conduct in-depth learning and training of these feature data, and calibrates and optimizes them through a large amount of clinical data to improve the accuracy and reliability of the prediction. The trained M2 model can achieve real-time prediction of cardiac output and cuff-free blood pressure, and these predicted values ​​are non-invasive and can be obtained without directly measuring blood or arterial pressure, thereby reducing patient discomfort and medical risks.

[0070] 3) Comprehensive cardiocerebral hemodynamic monitoring model (M3)

[0071] The prediction results of the M1 and M2 models are further input into the comprehensive cardiocerebral hemodynamic monitoring model (M3). The comprehensive cardiocerebral hemodynamic monitoring model (M3) is an advanced prediction model further developed on the basis of the M1 and M2 models. The original intention of the design of the M3 model is to provide more comprehensive and accurate cardiocerebral hemodynamic monitoring to meet the clinical needs for comprehensive evaluation of multiple physiological parameters. Specifically, the M1 model is responsible for predicting intracranial pressure, while the M2 model predicts cardiac output and blood pressure. The M3 model deeply integrates the prediction results of these two models to generate more accurate and comprehensive cardiocerebral hemodynamic parameters.

[0072] The uniqueness of the M3 model is that it does not simply integrate the outputs of M1 and M2, but takes into account the mutual influence and correlation between intracranial pressure, cardiac output and blood pressure through a multimodal signal fusion algorithm. For example, cerebral perfusion pressure (CPP) is determined by intracranial pressure and mean arterial pressure. The M3 model can provide more accurate CPP prediction results than a single model through a comprehensive analysis of these parameters.

[0073] To ensure the accuracy and applicability of the model, the M3 model is also trained and validated using extensive clinical data. The ultimate goal of the M3 model is to provide a more comprehensive and accurate non-invasive monitoring system by fusing multimodal signals to help clinicians better understand and manage patients' cardiocerebral hemodynamic status. Its application can not only improve the treatment effect of patients, but also reduce the risks and discomfort caused by the monitoring process, thereby greatly improving the overall quality of care for patients.

[0074] 4. Monitoring device and operation

[0075] 1) Structure and function of the device

[0076] like Figure 5 As shown, the monitoring device of the present invention includes a signal acquisition module, a signal processing module, a data fusion module, a non-invasive monitoring model module and a display and alarm module. These modules are integrated to jointly complete the real-time monitoring of cardiocerebral hemodynamic parameters.

[0077] 2) Operation process

[0078] When operating the device, first correctly connect the near-infrared spectroscopy sensor and chest impedance electrode to the patient's body surface. After starting the device, the signal acquisition module begins to acquire physiological signals, and the signal processing unit preprocesses and extracts features from these signals. Subsequently, the data fusion module fuses the processed signals, and finally the non-invasive monitoring model module generates non-invasive monitoring results. The display and alarm module presents the values ​​of cardiac output, blood pressure, intracranial pressure, and cerebral perfusion pressure in real time, and automatically triggers an alarm when the parameters exceed the safe range.

[0079] 3) Clinical application

[0080] The device of the present invention is particularly suitable for scenarios that require real-time monitoring of cardiocerebral hemodynamics, such as intensive care, cardiovascular and cerebrovascular surgery, and acute brain injury. Non-invasive, continuous monitoring methods can significantly reduce the pain of patients and improve the safety and effectiveness of monitoring.

[0081] 5. Experimental verification and results

[0082] 1) Experimental verification

[0083] In clinical experiments, the monitoring accuracy and reliability of the device of the present invention were verified by applying it to intensive care patients or healthy subjects. The monitoring device of the present invention was compared with standard monitoring equipment (such as brain puncture intracranial pressure invasive monitoring equipment, flash evoked visual potential measurement intracranial pressure system, and cardiac hemodynamics detector based on thoracic impedance method), focusing on the monitoring accuracy of the device for key parameters such as cardiac output, intracranial pressure, blood pressure and cerebral perfusion pressure.

[0084] 2) Implementation effect

[0085] The monitoring device of the present invention realizes non-invasive and continuous monitoring of cardiocerebral hemodynamic parameters by fusing multimodal signals. Compared with the prior art, the present invention has the following advantages:

[0086] (1) Improved monitoring accuracy and real-time performance: By fusing multimodal physiological signals, the device of the present invention can accurately reflect changes in cardiovascular and cerebral hemodynamic parameters and provide key physiological data to medical staff in real time. Compared with the traditional single signal monitoring method, the multimodal fusion method significantly improves the monitoring accuracy and reduces the monitoring error caused by single signal interference or loss.

[0087] (2) Reduces the pain and monitoring risks of patients: Traditional invasive monitoring methods often require the insertion of catheters or sensors, which not only increases the risk of infection and complications, but also causes discomfort and pain to patients. The present invention monitors in a non-invasive manner, completely avoiding the risks of invasive operations, providing patients with a more comfortable and safe monitoring experience, especially for critically ill patients who require long-term continuous monitoring.

[0088] (3) Providing more quantifiable physiological parameters: The device of the present invention can not only monitor cardiac hemodynamic parameters such as cardiac output and blood pressure, but also provide important cerebral hemodynamic indicators such as intracranial pressure and cerebral perfusion pressure. The provision of these parameters enables doctors to conduct a more comprehensive assessment and diagnosis of the patient's cardiocerebral blood flow status, thereby formulating a more accurate treatment plan.

[0089] Through the implementation of the present invention, non-invasive and continuous monitoring of human cardiocerebral hemodynamics can be achieved, providing a safer and more effective monitoring method with wide clinical application value.

[0090] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution, which should be included in the scope of the claims of the present invention.

Claims

1. A multimodal cardiocerebral hemodynamic monitoring method, characterized in that: The method specifically comprises the following steps: S1: Acquire brain absorbance information and brain pulse wave information based on near-infrared light method, acquire cardiac output information and cuffless blood pressure information based on thoracic impedance method, and then preprocess and extract features of the acquired information; S2: Based on the absorbance information and brain pulse wave characteristics, a non-invasive intracranial pressure prediction model M1 is constructed to obtain preliminary non-invasive prediction results of intracranial pressure; based on the chest impedance signal characteristic information and the electrocardiogram signal characteristic information, a non-invasive prediction model M2 based on the chest impedance method is constructed to obtain preliminary non-invasive prediction results of cardiac output and blood pressure; S3: The physiological information features obtained in step S1 based on the near-infrared light method and the thoracic impedance method are fused and processed to obtain multimodal physiological information; based on the preliminary non-invasive prediction results of intracranial pressure, cardiac output and blood pressure obtained in step S2 and the multimodal physiological information, a non-invasive comprehensive monitoring model M3 of intracranial pressure and cerebral perfusion pressure is constructed to obtain accurate prediction results of intracranial pressure and blood pressure, and then obtain non-invasive prediction results of cerebral perfusion pressure, thereby realizing continuous non-invasive monitoring that synchronously reflects the human cardiocerebral hemodynamic state.

2. The multimodal cardiocerebral hemodynamic monitoring method according to claim 1, characterized in that: In step S1, the near-infrared light method uses a multi-wavelength sensor monitoring probe to monitor the absorbance changes of the frontal lobe of the brain and the pulse wave signal of the middle cerebral artery.

3. The multimodal cardiocerebral hemodynamic monitoring method according to claim 1, characterized in that: In step S1, the thoracic impedance method utilizes the human heart hemodynamic parameter monitoring means to obtain a physiological signal reflecting the human heart hemodynamic change.

4. The multimodal cardiocerebral hemodynamic monitoring method according to claim 2, characterized in that: In step S2, a non-invasive intracranial pressure prediction model M1 is constructed, which specifically includes the following steps: S201: Using a multi-wavelength sensing monitoring probe, obtain n wavelengths λ1, λ2, ..., λ n Absorbance information at the frontal lobe of the brain under the light source: I = {i λ1 ,i λ2 ,...,i λ n }; S202: Using the sensor monitoring probe to obtain the pulse wave time domain signal PPG of the middle cerebral artery, and after preprocessing it, obtain the pulse wave characteristic parameters related to the changes in intracranial pressure and hemodynamics of the brain: P = {p1, p2, ..., p m }, where m represents the number of preprocessed pulse wave features; S203: Construct a non-invasive intracranial pressure prediction model M1 to obtain non-invasive prediction results of intracranial pressure and mean cerebral arterial pressure: M1(ICP0, MAP0) = f(I, P, PPG), where ICP0 represents the preliminary predicted intracranial pressure and MAP0 represents the preliminary predicted mean cerebral arterial pressure.

5. The multimodal cardiocerebral hemodynamic monitoring method according to claim 3, characterized in that: In step S2, a non-invasive prediction model M2 of cardiac output and blood pressure is constructed, which specifically includes the following steps: S211: obtaining physiological signals reflecting changes in human cardiac hemodynamics, including thoracic impedance change signals, thoracic impedance change differential signals, basic impedance signals, electrocardiogram signals and pulse wave signals, and preprocessing the signals; S212: Extract characteristic parameters related to cardiac output and hemodynamic changes from the preprocessed signals and construct a non-invasive prediction model for cardiac output M21(CO)=f(p1,p2,p3,...,p k ), wherein k represents the number of characteristic parameters related to cardiac output and hemodynamic changes, and CO represents the predicted cardiac output; the characteristic parameters related to cardiac output and hemodynamic changes include periodic information, strength information, electro-mechanical conduction information and energy information of thoracic impedance changes; S213: extracting the characteristic transmission time information of the heart activity reflected in the transition from the electrocardiogram signal to the chest impedance change signal, from the electrocardiogram signal to the pulse wave signal, and from the chest impedance signal to the pulse wave signal, and the time, amplitude and energy information of the characteristic changes of the chest impedance change signal and the pulse wave signal from the preprocessed signal, and forming a parameter set of time T, value A and energy P. Based on these three parameter sets, construct a non-invasive prediction model M22(SBP,DBP)=f(T,A,P) for cuffless blood pressure, where SBP represents systolic pressure and DBP represents diastolic pressure. S214: The non-invasive prediction models M21 and M22 obtained by steps S212 and S213 together constitute a non-invasive prediction model M2=(M21, M22) based on the thoracic impedance method.

6. The multimodal cardiocerebral hemodynamic monitoring method according to claim 1, characterized in that: In step S3, a non-invasive comprehensive monitoring model M3 for cardiocerebral hemodynamics is constructed, which specifically includes the following steps: S301: Fusing the physiological information features obtained based on the near infrared light method and the chest impedance method to obtain multimodal physiological information I fusion ; S302: Using the correlation between changes in cardiocerebral hemodynamics, intracranial pressure, blood pressure and cardiac output, multimodal physiological information is added to the non-invasive comprehensive monitoring model M3 for cardiocerebral hemodynamics based on the intracranial pressure values, cerebral hemodynamic characteristics, cardiac output, blood pressure and cardiac hemodynamic characteristics obtained by the prediction models M1 and M2. fusion , construct a non-invasive intracranial pressure prediction model based on multimodal information M31(MAP,ICP)=f(I fusion ,ICP0,MAP0,CO,SBP,DBP), so that the outputs of models M1 and M2 affect each other, and improve the output accuracy of cardiac output, blood pressure and intracranial pressure; where MAP represents the mean cerebral arterial pressure predicted by model M31; ICP represents the intracranial pressure predicted by model M31; S303: In the non-invasive comprehensive monitoring model M3 for cardiocerebral hemodynamics, based on the result of step S302, a non-invasive prediction model M32 of cerebral perfusion pressure CPP (CPP) = f(CO, MAP, ICP, I fusion ); S304: The non-invasive prediction models M31 and M32 obtained from steps S302 and S303 together constitute a non-invasive comprehensive monitoring model M3 = (M31, M32) of intracranial pressure and cerebral perfusion pressure.

7. The multimodal cardiocerebral hemodynamic monitoring method according to claim 6, characterized in that: In step S301, the fusion processing is specifically based on the correlation of cardiocerebral hemodynamic parameters, using a time synchronization mechanism and a weighted fusion algorithm to combine near-infrared spectral signals and chest impedance signals to form multimodal physiological information that can comprehensively reflect the cardiocerebral hemodynamic state.

8. A multimodal cardiocerebral hemodynamic monitoring device, characterized in that: It includes a signal acquisition module, a signal processing module, a data fusion module, a non-invasive monitoring model module and a display and alarm module; the signal acquisition module includes a near-infrared spectrum sensor and a chest impedance electrode; The near-infrared spectral sensor and chest impedance electrode are used to be connected to the patient's body surface and are responsible for acquiring multimodal physiological signals of the human body; the signal processing module performs denoising, filtering and feature extraction on the acquired signals; the data fusion module fuses the physiological data of different signal sources; the non-invasive monitoring model module integrates the multimodal cardiocerebral hemodynamic monitoring method described in any one of claims 1 to 7, and generates non-invasive monitoring results based on the fused data; the display and alarm module is used to present the non-invasive monitoring results, namely the numerical values ​​and change trend waveforms of cardiac output, blood pressure, intracranial pressure and cerebral perfusion pressure in real time, and automatically trigger an alarm when the parameters exceed the safety range.

Images