Anesthesia depth monitoring method based on multi-modal dynamic response index (MDRI)

Abstract

The invention relates to the technical field of anesthesia depth monitoring, in particular to an anesthesia depth monitoring method based on a multi-modal dynamic response index (MDRI), which comprises the following steps of: acquiring physiological signals of a patient, and applying external stimulation to obtain a physiological signal result set # imgabs0 #; performing data preprocessing on the physiological signal result set # imgabs 1 # to obtain a preprocessing result set # imgabs 2 #; and extracting the characteristics of the preprocessing result set # imgabs3 # to obtain a multi-modal dynamic response index (MDRI). According to the anesthesia depth monitoring method based on the MDRI, by integrating different physiological signals from a patient and combining the influence of external stimulation on the signals, the anesthesia state of the patient can be evaluated more accurately, and further, anesthesia depth evaluation is obtained by accurately monitoring the anesthesia depth; an anesthetist can be helped to better adjust the dosage of anesthetic medicine, it is ensured that a patient keeps a proper anesthesia state in the operation period, and the risk that anesthesia is too shallow or too deep is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of anesthesia depth monitoring, and in particular to an anesthesia depth monitoring method based on a multimodal dynamic response index MDRI. Background Art

[0002] The depth of anesthesia refers to the level of anesthesia that the patient is in during general anesthesia. The ideal depth of anesthesia should ensure that the patient has no pain sensation, no conscious activity, stable hemodynamics, and can wake up smoothly after surgery without adverse reactions such as awareness during the operation.

[0003] Although traditional monitoring methods such as the bispectral index (BIS) and the auditory evoked potential (AEP) index can reflect the anesthesia state to a certain extent, they have limitations and cannot accurately detect the depth of anesthesia. Therefore, an anesthesia depth monitoring method based on the multimodal dynamic response index (MDRI) is proposed to solve the above problems. Summary of the invention

[0004] 1. Technical issues to be resolved In view of the shortcomings of the prior art, the present invention provides an anesthesia depth monitoring method based on the multimodal dynamic response index MDRI, which has the advantages of accurately monitoring the anesthesia depth, etc., and solves the problem that traditional monitoring methods such as the bispectral index BIS and the auditory evoked potential index AEP index, although they can reflect the anesthesia state to a certain extent, have limitations and cannot accurately detect the anesthesia depth.

[0005] (II) Technical solution In order to solve the problem that traditional monitoring methods such as bispectral index (BIS) and auditory evoked potential index (AEP index) can reflect the anesthesia state to a certain extent, but have limitations and cannot accurately detect the anesthesia depth, the present invention provides the following technical solutions: a method for monitoring the anesthesia depth based on multimodal dynamic response index (MDRI), comprising the following steps: S1. Collect the patient's physiological signals and apply external stimulation to obtain the physiological signal result set ; S2. Results of physiological signals Perform data preprocessing to obtain the preprocessing result set ; S3. Extract preprocessing result set The characteristics of the multimodal dynamic response index MDRI are obtained; S4. Based on the multimodal dynamic response index MDRI, combined with dynamic threshold , get the patient's anesthesia depth ; S5. According to the patient's anesthesia depth , get anesthesia depth evaluation and anesthetic dosage recommendations.

[0006] Preferably, S1 comprises the following sub-steps: S1.1. Define stimulus type as auditory stimulus , tactile stimulation and electrical stimulation ; S1.2. Defining Strength ,frequency and duration For stimulation parameters, set the parameter range and initial value. Auditory stimulation The parameter ranges of the stimulation parameters are: , and , Auditory stimulation The initial values ​​of the stimulation parameters are: , and , Tactile stimulation The parameter ranges of the stimulation parameters are: , and , Tactile stimulation The initial values ​​of the stimulation parameters are: , and , Electrical stimulation The parameter ranges of the stimulation parameters are: , and , Electrical stimulation The initial values ​​of the stimulation parameters are: , and ; S1.3. Generate auditory stimuli using the audio generation library PyDub The sound file within the parameter range is transmitted to the earphone to apply auditory stimulation, and the vibration motor is controlled by the hardware interface Raspberry Pi, and the vibration motor is attached to the patient's upper torso to apply tactile stimulation. Tactile stimulation within the parameter range, using the NeuroConn electrical stimulation device, and placing the electrodes on the NeuroConn electrical stimulation device in contact with the patient's skin to apply electrical stimulation Electrical stimulation within parameters; S1.4. Use the same sampling frequency to collect EEG, ECG, SpO2, NIBP, RR and EMG, and add a timestamp to each collected data point to obtain the physiological signal result set. .

[0007] Preferably, S2 further includes the following sub-steps: S2.1. Use adaptive filters to remove known, predictable noise sources; S2.2. Apply independent component analysis (ICA) to separate independent physiological signal components and identify and remove those components that represent noise or artifacts; S2.3. Apply Z-score normalization to all signals, i.e. , in, and are the mean and standard deviation of the signal respectively.

[0008] Preferably, S3 includes the following sub-steps: S3.1. From the preprocessing result set Extract time domain, frequency domain and nonlinear features, as well as event-related potential ERP features; S3.2. Use the Pearson correlation coefficient method to select the key features that best reflect the changes in anesthesia depth from the features in S3.1: in, is the eigenvalue, is the target variable value, and are the means of the feature and target variables respectively; S3.3. Combining key features into a composite index using a weighted approach : in, is the extracted feature, is the corresponding weight.

[0009] Preferably, the S3.1 further includes the following sub-steps: S3.1.1. Time domain feature extraction: average value: Standard Deviation: Maximum and minimum values: ( ) ( ) RMS: Skewness: Where N is the number of samples, It is Sample points; S3.1.2. Frequency domain feature extraction: Power Spectral Density Center frequency ; S3.1.3. Nonlinear feature extraction using sample entropy: in, and B are the number of vector pairs that satisfy the similarity condition, is the embedding dimension, is the similarity threshold, is the sample size.

[0010] Preferably, S4 comprises the following sub-steps: S4.1. Setting the Age Adjustment Factor , Weight Adjustment Factor , Health Adjustment Factor : in, and It is a parameter determined by clinical research, with a value range of 0~1. in, is a standard weight based on gender and height. ; S4.2. Calculation of dynamic threshold , in, It is a standard threshold determined through a large amount of clinical data; S4.3. Based on dynamic threshold Depth of anesthesia obtained with the multimodal dynamic response index (MDRI) : Depth of anesthesia .

[0011] Preferably, S5 comprises the following sub-steps: S5.1. Judgment If the size , return to "moderate anesthesia, no need to adjust the dosage of anesthetic", otherwise jump to S5.2; S5.2. Settings , and according to Calculating adjusted anesthetic doses : in, and is the baseline value and sensitivity parameter of heart rate, and It is the baseline value and sensitivity parameter of blood pressure; in, Usually 5mg or 0.5ml, For real-time anesthesia dosage; S5.3. Setting the maximum safe value of anesthetic dose and minimum safety value , get the final anesthetic adjustment amount , ; S5.4. Judgment If the size , return to "mild anesthesia, it is recommended to increase the amount of marijuana medicine, it is recommended to increase This anesthetic dosage", otherwise it returns "deep anesthesia, it is recommended to reduce the dosage of anesthetic, it is recommended to reduce This is the amount of anesthetic used".

[0012] (III) Beneficial effects Compared with the prior art, the present invention provides a method for monitoring the depth of anesthesia based on the multimodal dynamic response index MDRI, which has the following beneficial effects: This anesthesia depth monitoring method based on the multimodal dynamic response index MDRI can more accurately assess the patient's anesthetic state by integrating different physiological signals from the patient and combining the impact of external stimuli on these signals. Furthermore, by accurately monitoring the depth of anesthesia and obtaining an anesthesia depth evaluation, it can help anesthesiologists better adjust the dosage of anesthetic drugs, ensure that patients maintain an appropriate state of anesthesia during surgery, and reduce the risk of too shallow or too deep anesthesia. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 This is a flow chart of a method for monitoring depth of anesthesia based on the multimodal dynamic response index MDRI proposed in the present invention.

[0014] Figure 2 A schematic diagram of a process for obtaining anesthesia depth evaluation and anesthetic dosage recommendation based on a multimodal dynamic response index MDRI anesthesia depth monitoring method proposed in the present invention DETAILED DESCRIPTION

[0015] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0016] See also Figure 1 , a method for monitoring the depth of anesthesia based on a multimodal dynamic response index MDRI, comprising the following steps: S1. Collect the patient's physiological signals and apply external stimulation to obtain the physiological signal result set ; S2. Results of physiological signals Perform data preprocessing to obtain the preprocessing result set ; S3. Extract preprocessing result set The characteristics of the multimodal dynamic response index MDRI are obtained; S4. Based on the multimodal dynamic response index MDRI, combined with dynamic threshold , get the patient's anesthesia depth ; S5. According to the patient's anesthesia depth , get anesthesia depth evaluation and anesthetic dosage recommendations.

[0017] In this embodiment, S1 includes the following sub-steps: S1.1. Define stimulus type as auditory stimulus , tactile stimulation and electrical stimulation ; S1.2. Defining Strength ,frequency and duration For stimulation parameters, set the parameter range and initial value. Auditory stimulation The parameter ranges of the stimulation parameters are: , and , Auditory stimulation The initial values ​​of the stimulation parameters are: , and , Tactile stimulation The parameter ranges of the stimulation parameters are: , and , Tactile stimulation The initial values ​​of the stimulation parameters are: , and , Electrical stimulation The parameter ranges of the stimulation parameters are: , and , Electrical stimulation The initial values ​​of the stimulation parameters are: , and ; S1.3. Generate auditory stimuli using the audio generation library PyDub The sound file within the parameter range is transmitted to the earphone to apply auditory stimulation, and the vibration motor is controlled by the hardware interface Raspberry Pi, and the vibration motor is attached to the patient's upper torso to apply tactile stimulation. Tactile stimulation within the parameter range, using the NeuroConn electrical stimulation device, and placing the electrodes on the NeuroConn electrical stimulation device in contact with the patient's skin to apply electrical stimulation Electrical stimulation within parameters; S1.4. Use the same sampling frequency to collect EEG, ECG, SpO2, NIBP, RR and EMG, and add a timestamp to each collected data point to obtain the physiological signal result set. .

[0018] EEG: uses a high-density electrode cap to record electrical activity in the brain; ECG: Records the electrical activity of the heart through chest leads or other appropriately placed electrodes; SpO2: measured using a pulse oximeter; NIBP: measured by cuff blood pressure monitor; Respiratory rate: obtained using respiratory induction plethysmograph (RIP); EMG: Surface electrodes placed on muscle groups record electromyographic signals.

[0019] In this embodiment, S2 also includes the following sub-steps: S2.1. Use adaptive filters to remove known, predictable noise sources; S2.1.1. Setting the initial weight vector ; S2.1.2. For each iteration Update the weight vector: ; in, is the step size parameter, which controls the learning rate; Error signal, i.e. desired output With the actual output The difference between is the input signal vector.

[0020] S2.2. Apply independent component analysis (ICA) to separate independent physiological signal components and identify and remove those components that represent noise or artifacts; S2.2.1. Subtract the mean from the original data to make the data zero-mean; S2.2.2. Preprocess the data by methods such as principal component analysis (PCA) so that the covariance between components is the unit matrix; S2.2.3. Use negentropy maximization to find independent components; S2.2.4. Apply the found unmixing moment to the original data to obtain independent components.

[0021] The mathematical basis of ICA is to find a set of transformation matrices , so that the observed signal can be decomposed into independent components : in, is the observed mixed signal matrix, is the mixing matrix, is the independent component matrix. The goal of ICA is to estimate the inverse matrix W=A−1, so from Recovery .

[0022] S2.3. Apply Z-score normalization to all signals, i.e. ; in, and are the mean and standard deviation of the signal respectively.

[0023] In this embodiment, S3 includes the following sub-steps: S3.1. From the preprocessing result set Extract time domain, frequency domain and nonlinear features, as well as event-related potential ERP features; S3.2. Use the Pearson correlation coefficient method to select the key features that best reflect the changes in anesthesia depth from the features in S3.1: in, is the eigenvalue, is the target variable value, and are the means of the feature and target variables respectively; S3.3. Combining key features into a composite index using a weighted approach : in, is the extracted feature, is the corresponding weight.

[0024] In this embodiment, S3.1 also includes the following sub-steps: S3.1.1. Time domain feature extraction: average value: Standard Deviation: Maximum and minimum values: ( ) ( ) RMS: Skewness: Where N is the number of samples, It is Sample points; S3.1.2. Frequency domain feature extraction: Power Spectral Density Center frequency ; S3.1.3. Nonlinear feature extraction using sample entropy: in, and B are the number of vector pairs that satisfy the similarity condition, is the embedding dimension, is the similarity threshold, is the sample size.

[0025] In this embodiment, S4 includes the following sub-steps: S4.1. Setting the Age Adjustment Factor , Weight Adjustment Factor , Health Adjustment Factor : in, and It is a parameter determined by clinical research. When the age is much greater than or less than When , the adjustment factor gradually approaches 1 or 0. in, is a standard weight based on gender and height. ; in, and It is a parameter determined by clinical research. When the age is much greater than or less than When , the adjustment factor gradually approaches 1 or 0; S4.2. Calculation of dynamic threshold , in, It is a standard threshold determined through a large amount of clinical data; S4.3. Based on dynamic threshold Depth of anesthesia obtained with the multimodal dynamic response index (MDRI) : Depth of anesthesia .

[0026] In this embodiment, S5 includes the following sub-steps: S5.1. Judgment If the size , return to "moderate anesthesia, no need to adjust the dosage of anesthetic", otherwise jump to S5.2; S5.2. Settings , and according to Calculating adjusted anesthetic doses : in, and is the baseline value and sensitivity parameter of heart rate, and It is the baseline value and sensitivity parameter of blood pressure; in, Usually 5mg or 0.5ml, The actual amount of anesthesia used depends on the anesthetic drug used.

[0027] The way to obtain is: Keep the patient in a quiet, relaxed state and avoid any stimulation; Continuously measure the patient's resting heart rate using a heart rate monitoring device (such as an electrocardiograph or wearable device); Record your heart rate every few minutes for at least 10-15 minutes to ensure that the data is stable; Calculate the average of all measurements as ; The way to obtain is: Through clinical trials, we analyzed the changes in heart rate under different anesthesia depths; Collect data from a large number of patients and record changes in heart rate at different depths of anesthesia; Calculate the standard deviation of heart rate variation As , The formula is: .

[0028] The way to obtain is: Before the operation, let the patient be in a quiet and relaxed state and avoid any stimulation; Continuously measure the patient's resting blood pressure using a blood pressure monitoring device (e.g., electronic sphygmomanometer); Record systolic blood pressure (SBP) and diastolic blood pressure (DBP) every few minutes for at least 10-15 minutes to ensure that the data are stable; Calculate all measurements as .

[0029] The way to obtain is: Through clinical trials, we analyzed the changes in blood pressure at different depths of anesthesia; Collect data from a large number of patients and record blood pressure changes at different depths of anesthesia; Calculate the standard deviation of blood pressure contraction and change respectively (calculation formula is the same as the standard deviation of heart rate change), and take the average value of the standard deviation of blood pressure contraction and change as the sensitivity parameter .

[0030] S5.3. Setting the maximum safe value of anesthetic dose and minimum safety value , get the final anesthetic adjustment amount , ; S5.4. Judgment If the size , return to "mild anesthesia, it is recommended to increase the amount of marijuana medicine, it is recommended to increase This anesthetic dosage", otherwise it returns "deep anesthesia, it is recommended to reduce the dosage of anesthetic, it is recommended to reduce This is the amount of anesthetic used".

[0031] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring the depth of anesthesia based on a multimodal dynamic response index MDRI, characterized in that: The following steps are involved: S1. Collect the patient's physiological signals and apply external stimulation to obtain the physiological signal result set ; S2. Results of physiological signals Perform data preprocessing to obtain the preprocessing result set ; S3. Extract preprocessing result set The characteristics of the multimodal dynamic response index MDRI are obtained; S4. Based on the multimodal dynamic response index MDRI, combined with dynamic threshold , get the patient's anesthesia depth ; S5. According to the patient's anesthesia depth , get anesthesia depth evaluation and anesthetic dosage recommendations.

2. The method for monitoring depth of anesthesia based on multimodal dynamic response index MDRI according to claim 1, characterized in that: The S1 comprises the following sub-steps: S1.

1. Define stimulus type as auditory stimulus , tactile stimulation and electrical stimulation ; S1.

2. Defining Strength ,frequency and duration For stimulation parameters, set the parameter range and initial value. Auditory stimulation The parameter ranges of the stimulation parameters are: , and , Auditory stimulation The initial values ​​of the stimulation parameters are: , and , Tactile stimulation The parameter ranges of the stimulation parameters are: , and , Tactile stimulation The initial values ​​of the stimulation parameters are: , and , Electrical stimulation The parameter ranges of the stimulation parameters are: , and , Electrical stimulation The initial values ​​of the stimulation parameters are: , and ; S1.

3. Generate auditory stimuli using the audio generation library PyDub The sound file within the parameter range is transmitted to the earphone to apply auditory stimulation, and the vibration motor is controlled by the hardware interface Raspberry Pi, and the vibration motor is attached to the patient's upper torso to apply tactile stimulation. Tactile stimulation within the parameter range, using the NeuroConn electrical stimulation device, and placing the electrodes on the NeuroConn electrical stimulation device in contact with the patient's skin to apply electrical stimulation electrical stimulation within parameters; S1.

4. Use the same sampling frequency to collect EEG, ECG, SpO2, NIBP, RR and EMG, and add a timestamp to each collected data point to obtain the physiological signal result set. .

3. The method for monitoring the depth of anesthesia based on the multimodal dynamic response index MDRI according to claim 1, characterized in that: The S2 further comprises the following sub-steps: S2.

1. Use adaptive filters to remove known, predictable noise sources; S2.

2. Apply independent component analysis (ICA) to separate independent physiological signal components and identify and remove those components that represent noise or artifacts; S2.

3. Apply Z-score normalization to all signals, i.e. , in, and are the mean and standard deviation of the signal respectively.

4. The method for monitoring depth of anesthesia based on multimodal dynamic response index MDRI according to claim 1, characterized in that: The S3 comprises the following sub-steps: S3.

1. From the preprocessing result set Extract time domain, frequency domain and nonlinear features, as well as event-related potential ERP features; S3.

2. Use the Pearson correlation coefficient method to select the key features that best reflect the changes in anesthesia depth from the features in S3.1: in, is the eigenvalue, is the target variable value, and are the means of the feature and target variables respectively; S3.

3. Combining key features into a composite index using a weighted approach : in, is the extracted feature, is the corresponding weight.

5. The method for monitoring the depth of anesthesia based on the multimodal dynamic response index MDRI according to claim 4, characterized in that: The S3.1 also includes the following sub-steps: S3.1.

1. Time domain feature extraction: average value: Standard Deviation: Maximum and minimum values: ( ) ( ) RMS: Skewness: Where N is the number of samples, It is Sample points; S3.1.

2. Frequency domain feature extraction: Power Spectral Density Center frequency ; S3.1.

3. Nonlinear feature extraction using sample entropy: in, and B are the number of vector pairs that satisfy the similarity condition, is the embedding dimension, is the similarity threshold, is the sample size.

6. The method for monitoring depth of anesthesia based on multimodal dynamic response index MDRI according to claim 1, characterized in that: The S4 comprises the following sub-steps: S4.

1. Setting the Age Adjustment Factor , Weight Adjustment Factor , Health Adjustment Factor : in, and It is a parameter determined by clinical research, with a value range of 0~1. in, is a standard weight based on gender and height. ； S4.

2. Calculation of dynamic threshold , in, It is a standard threshold determined through a large amount of clinical data; S4.

3. Based on dynamic threshold Depth of anesthesia obtained with the multimodal dynamic response index (MDRI) : Depth of anesthesia .

7. The method for monitoring the depth of anesthesia based on the multimodal dynamic response index MDRI according to claim 1, characterized in that: The S5 comprises the following sub-steps: S5.

1. Judgment If the size , return to "moderate anesthesia, no need to adjust the dosage of anesthetic", otherwise jump to S5.2; S5.

2. Settings , and according to Calculating adjusted anesthetic doses : in, and is the baseline value and sensitivity parameter of heart rate, and It is the baseline value and sensitivity parameter of blood pressure; in, Usually 5mg or 0.5ml, For real-time anesthesia dosage; S5.

3. Setting the maximum safe value of anesthetic dose and minimum safety value , get the final anesthetic adjustment amount , ； S5.

4. Judgment If the size , return to "mild anesthesia, it is recommended to increase the amount of marijuana medicine, it is recommended to increase This anesthetic dosage", otherwise it returns "deep anesthesia, it is recommended to reduce the anesthetic dosage, it is recommended to reduce This is the amount of anesthetic used".

Images