Anesthetic dosage optimization method based on artificial intelligence

Abstract

The invention relates to the field of medical informatics, in particular to an anesthetic dosage optimization method based on artificial intelligence. Comprising the following steps: collecting real-time physiological data and static data of a patient, converting the real-time physiological data and static data of the patient into standardized feature vectors through standardization processing, and predicting the initial infusion rate of an intraoperative anesthetic by using a long-short-term memory network; collecting and standardizing the intraoperative data of the patient to obtain the standardized intraoperative data; a personalized anesthetic infusion rate is generated through a dynamic metabolism risk and nerve sensitivity joint optimization algorithm; the maximum allowable infusion rate and the minimum effective infusion rate are determined based on a clinical guide, the remaining time of an operation is obtained, safety verification is carried out, and the final infusion rate is generated in combination with the remaining time of the operation. The technical problem that the prediction of the initial infusion rate of the preoperative anesthetic is inaccurate due to the fact that the initial infusion rate is determined by depending on the experience of an anesthetist in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the field of medical informatics, and particularly to an artificial intelligence-based method for optimizing the dosage of anesthetic drugs. Background Art

[0002] Optimizing the dosage of anesthetic drugs is a crucial part of modern medicine, directly related to the safety of surgery, the recovery speed of patients, and the effective utilization of medical resources. Traditional methods for determining the dosage of anesthetic drugs mainly rely on the experience of anesthesiologists, the physiological parameters of patients (such as weight, age, gender), and standardized dosing guidelines. However, due to individual differences, disease states, and the complexity of pharmacokinetic and pharmacodynamic characteristics, traditional methods for determining the dosage of anesthetic drugs often struggle to achieve precise drug administration, leading to situations of insufficient or excessive anesthetic drug dosage. In recent years, the rapid development of artificial intelligence (AI) technology has provided new possibilities for optimizing the dosage of anesthetic drugs. The artificial intelligence-based method for determining the dosage of anesthetic drugs can achieve individualized and dynamic optimization of anesthetic drug dosage by integrating big data, machine learning, deep learning, and real-time monitoring technologies, thereby enhancing the precision and safety of anesthesia management.

[0003] However, the above existing technologies generally adopt standardized anesthetic drug infusion regimens, failing to fully consider individual patient differences, resulting in the possibility of anesthesia overdose or underdose in some patients, affecting the safety and efficacy of surgery; traditional anesthesia management relies on the empirical judgment of anesthesiologists, lacking a real-time, automated, data-driven mechanism for optimizing the dosage of anesthetic drugs; existing technologies usually do not fully consider the cumulative effect of previously infused drugs that have not fully taken effect, which may lead to excessive dosage or fluctuations, increasing the risk of overdose anesthesia. Summary of the Invention

[0004] The present invention provides an artificial intelligence-based method for optimizing the dosage of anesthetic drugs to solve the technical problems that existing technologies mostly rely on the experience of anesthesiologists to determine the initial infusion rate, resulting in inaccurate prediction of the initial infusion rate of preoperative anesthetic drugs; existing technologies are difficult to dynamically adjust the infusion rate according to the real-time physiological data of patients during surgery, lacking a comprehensive analysis of metabolic risk and neural sensitivity, resulting in a lag or inaccuracy in the adjustment of anesthetic drug dosage; existing technologies do not effectively model the onset delay of anesthetic drugs, resulting in the risk of overdose caused by recent high-dose infusions, especially in surgical scenarios with rapid physiological changes; and the lack of dynamic safety constraints based on clinical guidelines may lead to the infusion rate exceeding the safe range.

[0005] An artificial intelligence-based method for optimizing the dosage of anesthetic drugs according to the present invention specifically includes the following technical solutions: An artificial intelligence-based method for optimizing the dosage of anesthetic drugs includes the following steps: S1. Collect the real-time physiological data and static data of the patient, convert the real-time physiological data and static data of the patient into standardized feature vectors through standardization processing, and based on the standardized feature vectors, use a long short-term memory network to predict the initial infusion rate of intraoperative anesthetic drugs; S2. Collect and standardize the intraoperative data of the patient to obtain the standardized intraoperative data; based on the standardized intraoperative data, generate a personalized anesthetic drug infusion rate through a combined optimization algorithm of dynamic metabolic risk and neural sensitivity; S3. Determine the maximum allowable infusion rate and the minimum effective infusion rate based on clinical guidelines, obtain the remaining surgical time, perform safety verification based on the personalized anesthetic drug infusion rate, and combine the remaining surgical time to generate the final infusion rate.

[0006] Preferably, the S1 specifically includes: During the standardization process, calculate the mean and standard deviation of each feature based on the preoperative patient historical data and the healthy population data, and then divide the real-time physiological data and static data of the patient by the standard deviation after subtracting the mean to generate a standardized feature vector, where the standardized feature vector contains the preoperative physiological and static characteristics of the patient.

[0007] Preferably, the S1 specifically includes: Based on the standardized feature vector, use a long short-term memory network to predict the initial infusion rate of intraoperative anesthetic drugs; the long short-term memory network is trained using the preoperative physiological data of the patient and the corresponding initial infusion rate annotations to obtain a trained long short-term memory network.

[0008] Preferably, the S1 specifically includes: Input the standardized feature vector into the trained long short-term memory network, perform non-linear mapping processing on the temporal relationship of the standardized features, and use a fully connected layer to map the output of the trained long short-term memory network to the initial infusion rate value.

[0009] Preferably, the S2 specifically includes: The standardized intraoperative data includes: standardized heart rate variability, standardized bispectral index, and standardized mean arterial pressure; the combined optimization algorithm of dynamic metabolic risk and neural sensitivity generates a personalized anesthetic drug infusion rate by combining the standardized intraoperative data and the drug effect delay effect.

[0010] Preferably, the S2 specifically includes: In the process of implementing the joint optimization algorithm for dynamic metabolic risk and neural sensitivity, the intraoperative data after standardization is analyzed by evaluating the interaction effect between the standardized mean arterial pressure and the standardized bispectral index of electroencephalogram; and the standardized heart rate variability is introduced to adjust the individualized anesthesia drug infusion rate.

[0011] Preferably, the S2 specifically includes: In the process of implementing the joint optimization algorithm for dynamic metabolic risk and neural sensitivity, the pharmacodynamic delay effect calculates the influence of the onset delay of anesthesia drugs. By obtaining the final infusion rate within the time window from the intraoperative historical anesthesia drug infusion rate record, using the weighted summation method and combining with the exponential decay function, the individualized anesthesia drug infusion rate is obtained. The specific calculation formula is: , where, represents the individualized anesthesia drug infusion rate at time ; represents the initial infusion rate value, that is, the initial rate of continuous infusion during the operation, serving as the benchmark for dynamic adjustment; represents the weight adjustment coefficient; represents the standardized mean arterial pressure, interacting with the standardized bispectral index of electroencephalogram to represent the metabolic-neural synergistic effect; represents the standardized bispectral index of electroencephalogram; represents the standardized heart rate variability; represents the delay effect weight; represents the pharmacodynamic delay effect term; represents the current time; represents the time window and the historical time points within it; represents the pharmacodynamic delay time constant; represents the time step; represents the exponential decay function; represents the final infusion rate at time .

[0012] Preferably, the S3 specifically includes: In the process of generating the final infusion rate, the remaining operation time proportionally affects the individualized anesthesia drug infusion rate through the time adjustment coefficient to generate a candidate value. The calculation formula for the final infusion rate is: , where, represents the final infusion rate at time ; represents the remaining operation time; represents the time adjustment coefficient; represents the minimum effective infusion rate; represents the maximum allowable infusion rate; represents the maximum value function; represents the minimum value function; represents the generated candidate value.

[0013] Preferably, the S3 specifically includes: Compare the candidate value with the maximum allowable infusion rate and the minimum effective infusion rate: when the candidate value exceeds the maximum allowable infusion rate, take the maximum allowable infusion rate; when it is lower than the minimum effective infusion rate, take the minimum effective infusion rate; otherwise, retain the candidate value, so that the final infusion rate is within the safe range.

[0014] The beneficial effects of the technical solution of the present invention are: 1. Utilize professional medical equipment to collect the real-time physiological data of patients and the static data in the electronic medical record system, generate a standardized feature vector through standardized processing, and predict the initial infusion rate of intraoperative anesthetic drugs based on the long short-term memory network (LSTM). Predict the initial infusion rate based on the individual characteristics of the patient, reduce the overdose or underdose of anesthesia caused by empirical anesthetic drug dose selection, and through automated data collection and processing, reduce the workload of manual analysis by medical staff, and improve the efficiency and accuracy of preoperative anesthesia plan formulation.

[0015] 2. Real-time collect the intraoperative data of patients, including: heart rate variability, bispectral index, mean arterial pressure, generate standardized intraoperative data through standardized processing, adopt a combined optimization algorithm of dynamic metabolic risk and neural sensitivity, and combine the drug effect delay effect to generate a personalized anesthetic drug infusion rate. The personalized anesthetic drug infusion rate adapts to the intraoperative physiological changes of patients, avoids too deep or too shallow anesthesia, significantly reduces the risk of patient awakening or overdose anesthesia during surgery, and the precise correction of the drug effect delay effect optimizes drug distribution, reduces unnecessary drug use, reduces postoperative drug residues, and shortens the patient recovery time.

[0016] 3. Set the maximum allowable infusion rate and the minimum effective infusion rate based on clinical guidelines, perform safety verification by combining the remaining surgical time and the personalized anesthetic drug infusion rate, generate a clinically safe and reasonable final infusion rate, and integrate it with the intelligent infusion pump and the electronic medical record system. The strict safety verification effectively prevents drug overdose caused by too high anesthetic drug infusion rate or anesthesia underdose caused by too low rate, reduces intraoperative risks, combines the dynamic adjustment of the remaining surgical time, reduces postoperative drug residues, shortens the patient awakening time, improves the operating room turnover rate and resource utilization efficiency, and the seamless integration with the intelligent infusion pump and the electronic medical record system realizes the automation, real-time and digitalization of anesthesia management, reduces the operation burden of medical staff, and improves the efficiency of the clinical work process. Description of the Drawings

[0017] Figure 1 This is a flowchart of a method for optimizing the dosage of anesthetic drugs based on artificial intelligence according to the present invention. Detailed Embodiments

[0018] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0020] The following specifically describes the specific solution of a method for optimizing the dosage of anesthetic drugs based on artificial intelligence provided by the present invention with reference to the accompanying drawings.

[0021] Referring to the attached Figure 1 , which shows a flowchart of a method for optimizing the dosage of anesthetic drugs based on artificial intelligence provided by an embodiment of the present invention. The method includes the following steps: S1. Collect the real-time physiological data and static data of the patient, convert the real-time physiological data and static data of the patient into standardized feature vectors through standardization processing, and based on the standardized feature vectors, use a long short-term memory network to predict the initial infusion rate of anesthetic drugs during the operation; Collect the real-time physiological data of the patient within 1 hour before the operation through professional medical equipment, including heart rate (measured by an ECG monitor, unit: beats per minute), blood pressure (measured by a non-invasive blood pressure monitor, unit: mmHg), blood oxygen saturation (measured by a pulse oximeter, unit: percentage), and electroencephalogram signal (measured by a portable EEG device, unit: microvolt). At the same time, obtain the static data of the patient from the electronic medical record system, including age (unit: years), weight (unit: kilograms), gender (categorical variable), liver and kidney function indicators (such as creatinine clearance rate, normalized to a dimensionless value), and ASA classification (grades 1-5); Convert the real-time physiological data and static data of the patient into standardized feature vectors through standardization processing; specifically, the standardization processing process calculates the mean and standard deviation of each feature based on the historical data of the patient before the operation or the data of the healthy population, and then subtracts the mean from the real-time physiological data and static data of the patient and divides by the standard deviation to generate standardized feature vectors ; The standardized feature vector includes the patient's preoperative physiological features and static features, which reflect the patient's basal metabolism and health status and are used to predict the initial infusion rate of anesthetic drugs during surgery; Based on the standardized feature vector, a long short-term memory network (LSTM) is used to predict the initial infusion rate of anesthetic drugs during surgery, in milligrams per kilogram of body weight per hour, and is applied through an intelligent infusion pump at the start of the surgery (time point 0 minutes); The long short-term memory network (LSTM) is trained using a historical surgery dataset (including the patient's preoperative physiological data and the corresponding initial infusion rate annotations) to obtain the trained long short-term memory network; the standardized feature vector is input into the trained long short-term memory network (LSTM) to perform a non-linear mapping process on the temporal relationship of the standardized features, and then a fully connected layer (including weights and biases) is used to map the output of the trained long short-term memory network (LSTM) to an initial infusion rate value. The formula is as follows: , where, represents the initial infusion rate value; represents the weight matrix of the fully connected layer, which reflects the contribution weights of different features to the prediction of anesthetic drug dosage and is optimized through training; represents capturing the temporal dependence of the standardized feature vector, such as small fluctuations in heart rate over time; represents the bias of the fully connected layer, which is used to adjust the baseline value of anesthetic drug dosage prediction to ensure that the predicted dosage rate is within the clinical range; The output initial infusion rate is used to reflect the patient's preoperative physiological state, but cannot fully adapt to the changes in the patient's intraoperative physiological state, providing a baseline for intraoperative dynamic optimization for the patient; S2. Collect and standardize the patient's intraoperative data to obtain the standardized intraoperative data; based on the standardized intraoperative data, generate a personalized anesthetic drug infusion rate through a joint optimization algorithm for dynamic metabolic risk and neural sensitivity; During the surgery, collect the patient's intraoperative data; specifically, calculate the heart rate variability in real time through an ECG monitor, and use time domain analysis (such as the standard deviation of the heart rate sequence) or frequency domain analysis (such as the low-frequency / high-frequency ratio) to reflect the metabolic activity; measure the bispectral index of electroencephalogram through a BIS monitor to reflect the depth of anesthesia, with a range of 0 to 100, and an appropriate anesthesia depth range of 40 - 60; measure the mean arterial pressure (unit: mmHg) through an arterial pressure sensor to reflect the circulatory system pressure; Standardize the patient's intraoperative data, that is, calculate the mean and standard deviation of the patient's intraoperative data using a 10-minute time window, subtract the time window mean from the patient's intraoperative data and then divide by the standard deviation to obtain the standardized intraoperative data; Based on the standardized intraoperative data, the standardized intraoperative data is analyzed in real time through a joint optimization algorithm of dynamic metabolic risk and neural sensitivity. The standardized intraoperative data includes: standardized heart rate variability, standardized bispectral index, and standardized mean arterial pressure. The joint optimization algorithm of dynamic metabolic risk and neural sensitivity generates a personalized anesthesia drug infusion rate by combining the standardized intraoperative data and the drug effect delay effect, which is applicable to general anesthesia surgeries such as continuous propofol infusion, ensuring appropriate and safe intraoperative anesthesia depth, avoiding over-injection caused by delayed drug onset, and at the same time adapting to the dynamic changes in the patient's physiological state. The specific implementation process is as follows: Analyze the standardized intraoperative data by evaluating the interaction effect between the standardized mean arterial pressure and the standardized bispectral index, which reflects the synergistic state of the circulatory system pressure and the anesthesia depth. The higher the interaction effect between the two indicates more active metabolism or insufficient anesthesia. Further, introduce the standardized heart rate variability to adjust the personalized anesthesia drug infusion rate. During the operation, the higher the standardized heart rate variability, the more anesthetic drug dosage is required. The drug effect delay effect calculates the impact of the delayed onset of anesthesia drugs, considering the amount of drugs that have been injected but not fully effective within the past time window. Specifically, obtain the final infusion rate within the time window from the intraoperative historical anesthesia drug infusion rate record, and use the weighted summation method, combined with the exponential decay function, to weight the recent infusion volume according to the time distance. The recent infusion volume has a greater impact on the current anesthesia drug dosage, and the impact of the long-term infusion volume on the current anesthesia drug dosage gradually decreases. The calculation formula for the personalized anesthesia drug infusion rate is: , where represents the personalized anesthesia drug infusion rate at time ; represents the initial infusion rate value, that is, the initial rate of continuous infusion during the operation, which serves as the benchmark for dynamic adjustment to ensure that the personalized anesthesia drug infusion rate is based on preoperative prediction; represents the weight adjustment coefficient, which is used to balance the contribution of the patient's intraoperative physiological data to the adjustment of the personalized anesthesia drug infusion rate, and to adjust the comprehensive impact of the metabolism-neural synergistic effect and the standardized heart rate variability on the anesthesia drug dosage. The value range is from 0.1 to 0.5, and it can be specifically set according to the specific implementation scenario and is not limited here; represents the standardized mean arterial pressure, that is, the circulatory system pressure at time , which interacts with the standardized bispectral index to represent the metabolism-neural synergistic effect; represents the standardized bispectral index, that is, at time The depth of anesthesia reflects the sensitivity of the brain to anesthetic drugs. When the normalized bispectral index of electroencephalogram is at a low value (too deep), the infusion rate of anesthetic drugs needs to be reduced. When the normalized bispectral index of electroencephalogram is at a high value (too shallow), the infusion rate of anesthetic drugs needs to be increased; represents the normalized heart rate variability, that is, at time The metabolic activity reflects the dynamic balance of the sympathetic - parasympathetic nervous system. A high value indicates that the infusion rate of anesthetic drugs needs to be increased; represents the delay effect weight, which is used to adjust the impact of drug effect delay on anesthetic drug dose adjustment. The value range is from 0.1 to 0.3, and it can be specifically set according to the specific implementation scenario and is not limited here; represents the drug effect delay term, estimating the amount of drugs that have been injected but not fully effective within the past window, which is used to reduce the adjustment range of anesthetic drug dose when the recent infusion volume is high, and avoid over - injection caused by drug effect delay; represents the current time; represents the time window and the historical time points within it; represents the drug effect delay time constant, with the unit of minute, and the typical value is 1.5 min, based on clinical pharmacokinetic research; represents the time step, with the unit of minute, fixed at 1 (sampling once per minute); represents the exponential decay function, which is used to simulate the delay characteristics of drug onset; represents at time the final infusion rate; By comprehensively considering the intraoperative metabolic state, depth of anesthesia, and circulatory system pressure of the patient, a personalized anesthetic drug infusion rate is generated to adapt to the individual differences of the patient, improve the anesthetic effect, and significantly reduce the risk of over - anesthesia through the drug effect delay effect; S3. Determine the maximum allowable infusion rate and the minimum effective infusion rate based on clinical guidelines, obtain the remaining surgical time, perform safety verification based on the personalized anesthetic drug infusion rate, and combine with the remaining surgical time to generate the final infusion rate; Determine the maximum allowable infusion rate and the minimum effective infusion rate based on clinical guidelines; Obtain the remaining surgical time from the surgical planning system or real - time timer, which reflects the remaining duration before the expected end of the surgery. Since the anesthetic drug is continuously infused and the infusion rate is updated every minute, the remaining time before the expected end of the surgery is checked every minute; Based on the personalized anesthetic drug infusion rate, perform safety verification, and combine with the remaining surgical time to generate a clinically safe and reasonable final infusion rate; Specifically, the remaining surgical time proportionally affects the personalized anesthesia drug infusion rate through a time adjustment coefficient to generate a candidate value, which is compared with the maximum allowable infusion rate and the minimum effective infusion rate: if the candidate value exceeds the maximum allowable infusion rate, the maximum allowable infusion rate is taken; if it is lower than the minimum effective infusion rate, the minimum effective infusion rate is taken; otherwise, the candidate value is retained to ensure that the final infusion rate is always within the safe range, while considering the dynamic changes in the surgical process, which is in line with clinical practice; The formula for calculating the final infusion rate is: , wherein, represents the final infusion rate at time , which is used for clinical anesthesia drug administration, ensuring that the anesthesia drug infusion rate is within the safe range, optimizing drug distribution, transmitting to the infusion pump and recording; represents the remaining surgical time, which is the estimated duration from time to the end of the surgery and is used to fine-tune the personalized anesthesia drug infusion rate; represents the time adjustment coefficient, which is used to adjust the influence of the remaining surgical time on the anesthesia drug dose, with a value of 0.01, and can be specifically set according to the specific implementation scenario and is not limited here; represents the minimum effective infusion rate, with the unit of mg / kg / h, which sets the lower limit of the anesthesia drug infusion rate to ensure the anesthesia effect and is determined based on clinical guidelines; represents the maximum allowable infusion rate, with the unit of mg / kg / h, which sets the upper limit of the anesthesia drug infusion rate to avoid excessive anesthesia and is determined based on clinical guidelines; represents the maximum value function, ensuring that the anesthesia drug infusion rate is not lower than the minimum effective infusion rate; represents the minimum value function, ensuring that the anesthesia drug infusion rate does not exceed the maximum allowable infusion rate; represents the generated candidate value; Through strict safety constraint verification, it effectively prevents drug overdose caused by too high anesthesia drug infusion rate or anesthesia insufficiency caused by too low rate, reduces intraoperative risks, optimizes drug distribution by dynamically adjusting the remaining surgical time, reduces postoperative drug residues, shortens the patient's recovery time, improves surgical efficiency, and the seamless integration with the intelligent infusion pump and the electronic medical record system realizes real-time and automation, updates the anesthesia drug infusion rate every minute, which is consistent with the sampling frequency of clinical monitoring equipment, and the operation is efficient and in line with general anesthesia surgical practice.

[0022] In summary, an anesthesia drug dose optimization method based on artificial intelligence is completed.

[0023] The order of the invention embodiments is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0024] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments.

[0025] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments or equivalently replace some of the technical features. These modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention and should all be included in the protection scope of the present invention.

Claims

1. An artificial intelligence-based method for optimizing the dosage of anesthetic drugs, characterized in that, Including the following steps: S1. Collect the real-time physiological data and static data of the patient, convert the real-time physiological data and static data of the patient into standardized feature vectors through standardized processing, and based on the standardized feature vectors, use a long short-term memory network to predict the initial infusion rate of intraoperative anesthetic drugs; S2. Collect and standardize the intraoperative data of the patient to obtain the standardized intraoperative data; based on the standardized intraoperative data, generate a personalized anesthetic drug infusion rate through a combined optimization algorithm for dynamic metabolic risk and neural sensitivity; S3. Determine the maximum allowable infusion rate and the minimum effective infusion rate based on clinical guidelines, obtain the remaining surgical time, perform a safety check based on the personalized anesthetic drug infusion rate, and combine the remaining surgical time to generate the final infusion rate.

2. The anesthetic drug dosage optimization method based on artificial intelligence according to claim 1, wherein, The specific content of S1 includes: During the standardized processing, calculate the mean and standard deviation of each feature based on the preoperative patient historical data and the healthy population data, and then divide the real-time physiological data and static data of the patient by the standard deviation after subtracting the mean to generate a standardized feature vector, and the standardized feature vector includes the preoperative physiological and static characteristics of the patient.

3. The method for optimizing the dosage of anesthetic drugs based on artificial intelligence according to claim 2, characterized in that, The specific content of S1 includes: Based on the standardized feature vector, use a long short-term memory network to predict the initial infusion rate of intraoperative anesthetic drugs; the long short-term memory network is trained using the preoperative physiological data of the patient and the corresponding initial infusion rate annotations to obtain a trained long short-term memory network.

4. The method for optimizing the dosage of anesthetic drugs based on artificial intelligence according to claim 3, wherein The specific content of S1 includes: Input the standardized feature vector into the trained long short-term memory network, perform a non-linear mapping process on the temporal relationship of the standardized features, and use a fully connected layer to map the output of the trained long short-term memory network to the initial infusion rate value.

5. A method for optimizing the dosage of anesthetic drugs based on artificial intelligence according to claim 1, characterized in that, The specific content of S2 includes: The standardized intraoperative data includes: standardized heart rate variability, standardized bispectral index, and standardized mean arterial pressure; the combined optimization algorithm for dynamic metabolic risk and neural sensitivity generates a personalized anesthetic drug infusion rate by combining the standardized intraoperative data and the drug effect delay effect.

6. The method for optimizing the dosage of anesthetic drugs based on artificial intelligence according to claim 5, characterized in that, The specific content of S2 includes: During the implementation of the combined optimization algorithm for dynamic metabolic risk and neural sensitivity, analyze the standardized intraoperative data by evaluating the interaction effect between the standardized mean arterial pressure and the standardized bispectral index; and introduce the standardized heart rate variability to adjust the personalized anesthetic drug infusion rate.

7. The method for optimizing the dosage of anesthetic drugs based on artificial intelligence according to claim 6, characterized in that, The specific content of S2 includes: During the implementation of the combined optimization algorithm for dynamic metabolic risk and neural sensitivity, the drug effect delay effect calculates the impact of the onset delay of anesthetic drugs. By obtaining the final infusion rate within the time window from the intraoperative historical anesthetic drug infusion rate record, using the weighted summation method, and combining the exponential decay function, a personalized anesthetic drug infusion rate is obtained. The specific calculation formula is: ， Among them, represents the personalized anesthesia drug infusion rate at time ; represents the initial infusion rate value, that is, the initial rate of continuous infusion during the operation, serving as the benchmark for dynamic adjustment; represents the weight adjustment coefficient; represents the standardized mean arterial pressure, interacting with the standardized bispectral index to represent the metabolic-neural synergistic effect; represents the standardized bispectral index; represents the standardized heart rate variability; represents the delay effect weight; represents the drug effect delay effect term; represents the current time; represents the time window and the historical time points within it; represents the drug effect delay time constant; represents the time step; represents the exponential decay function; represents the final infusion rate at time .

8. The anesthetic drug dosage optimization method based on artificial intelligence according to claim 1, wherein, The specific content of S3 includes: During the generation of the final infusion rate, the remaining surgical time proportionally affects the personalized anesthetic drug infusion rate through a time adjustment coefficient to generate a candidate value. The calculation formula for the final infusion rate is: ， Among them, represents the final infusion rate at time ; represents the remaining operation time; represents the time adjustment coefficient; represents the minimum effective infusion rate; represents the maximum allowable infusion rate; represents the maximum value function; represents the minimum value function; represents the generated candidate value.

9. The method for optimizing the dosage of anesthetic drugs based on artificial intelligence according to claim 8, characterized in that The specific content of S3 includes: Compare the candidate value with the maximum allowable infusion rate and the minimum effective infusion rate: when the candidate value exceeds the maximum allowable infusion rate, take the maximum allowable infusion rate; when it is lower than the minimum effective infusion rate, take the minimum effective infusion rate; otherwise, retain the candidate value so that the final infusion rate is within the safe range.

Images