Method and system for monitoring neonatal neurodevelopment

Abstract

Disclosed in the present invention are a method and a system for monitoring neonatal neurodevelopment. The method comprises the following steps: on the basis of a prediction model, predicting the probability of neurodevelopmental damage to a neonate caused by placental methylmercury, wherein independent variables of the prediction model comprise: placental methylmercury level, maternal age, maternal ethnicity, maternal dietary intake frequency of freshwater products, maternal denture installation during pregnancy, paternal genetic history, family history of hereditary diseases, paternal smoking status, occupational mercury exposure, Apgar score of the neonate at 1 minute, newborn congenital malformations, birth season, and placental weight, and on the basis of same, predicting the probability of the neonate having a neurodevelopmental disorder. The prediction model provided in the present invention has good discrimination, consistency and stability, and thus is an effective prediction model for monitoring neonatal neurodevelopment.

Description

Neonatal neurodevelopmental monitoring method and system Technical Field

[0001] The present invention relates to a method for monitoring the neurodevelopment of a newborn, and also to a corresponding system for monitoring the neurodevelopment of a newborn, belonging to the technical field of medical care informatics. Background Art

[0002] Neurodevelopmental disorders in children refer to chronic developmental brain dysfunction diseases caused by multiple genetic or acquired causes that affect brain functional areas including cognition, movement, social adaptability, behavior, etc. during the developmental period.

[0003] According to the International Classification of Diseases and the Diagnostic and Statistical Manual of Mental Disorders, childhood neurodevelopmental disorders mainly include disorder of intellectual development, developmental speech or language disorder, autism spectrum disorder, developmental learning disorder, attention deficit hyperactivity disorder, tic disorder and other neurodevelopmental disorders.

[0004] The optimal time to intervene in children with neurodevelopmental abnormalities is within the first few years of life, particularly between the ages of 0 and 3. This is because the central nervous system is highly adaptable and plastic during this period, allowing for neuronal replacement and compensation through scientific intervention, enabling early detection, diagnosis, and treatment. For premature or high-risk infants prone to neurodevelopmental delays, intervention is best initiated from birth; the earlier the treatment, the better the outcome. Therefore, it is necessary to establish a child neurodevelopmental monitoring system to facilitate early intervention.

[0005] Summary of the Invention

[0006] The primary technical problem to be solved by the present invention is to provide a method for monitoring the neurodevelopment of newborns.

[0007] Another technical problem to be solved by the present invention is to provide a neonatal neurodevelopment monitoring system.

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

[0009] According to a first aspect of an embodiment of the present invention, a method for monitoring neonatal neurodevelopment is provided, which predicts the probability of placental methylmercury causing neurodevelopmental damage to the neonate based on a prediction model;

[0010] The independent variables of the prediction model include: the probability of predicting whether a newborn has a neurodevelopmental disorder based on methylmercury in the placenta, maternal age, maternal ethnicity, maternal dietary intake of fresh seafood, maternal denture use during pregnancy, paternal genetic history, family history of genetic diseases, paternal smoking status, occupational mercury exposure, newborn's 1-minute Apgar score, newborn congenital malformations, birth season, and placental weight;

[0011] The expression of the prediction model is: Logit(Neurodevelopmental_Impairment)=2.93+0.20*Placenta_MeHg+0.12*Maternal_Age+0.04*Materal_Ethnicity+0.43*MDIFF+6.68*MIATDP-0.43*Paternal_Genetic_History-1.56*FHHD+0.95*Paternal_smoking-1.64*Occupational_Mercury_Exposure+0.94*Apgar_Score_at_1_Min-3.86*Newborn_Congenital_Malformation-2.35*Birth_Season-0.01*Placental_Weight,

[0012] Among them, Logit(Neurodevelopmental_Impairment): the probability of whether the newborn has a neurodevelopmental disorder,

[0013] Placenta_MeHg: Methylmercury in the placenta

[0014] Maternal_Age: Mother's age

[0015] Material_Ethnicity: Mother's ethnicity

[0016] Maternal_Dietary_Intake_Frequency_of_Freshwater(MDIFF): Frequency of dietary intake of freshwater products by the mother

[0017] Maternal_Installing_Artificial_Teeth_During_Pregnancy(MIATDP): Maternal installation of artificial teeth during pregnancy

[0018] Paternal_Genetic_History: Father's genetic history

[0019] Family History of Hereditary Disease (FHHD): Family History of Hereditary Disease

[0020] Paternal_smoking: Father's smoking status

[0021] Occupational_Mercury_Exposure: Occupational mercury exposure

[0022] Apgar_Score_at_1_Min: Apgar score of newborn at 1 minute

[0023] Newborn_Congenital_Malformation: Newborn congenital malformation

[0024] Birth_Season: birth season

[0025] Placental_Weight: Placental weight.

[0026] Preferably, the prediction model is constructed by the following steps:

[0027] S1: Statistical tools were used to conduct statistical analysis on the total population and male and female subgroup data;

[0028] S2: Analysis of influencing factors based on logistic regression model;

[0029] S3: Establish a prediction model based on the analysis results of the multivariate logistic regression model.

[0030] Preferably, intervention is performed when the placental methylmercury level reaches 0.98 ng / g.

[0031] Preferably, for female newborns, intervention should be carried out when the mother's placental methylmercury level reaches 1.09 ng / g.

[0032] Preferably, for male newborns, intervention is performed when the mother's placental methylmercury level reaches 0.97 ng / g.

[0033] According to a second aspect of an embodiment of the present invention, a neonatal neurodevelopmental monitoring system is provided, comprising a processor and a memory; wherein the memory is coupled to the processor and is used to store a computer program, and when the computer program is executed by the processor, the processor implements the above-mentioned neonatal neurodevelopmental monitoring method.

[0034] Compared with existing technologies, the neonatal neurodevelopmental monitoring method and system provided by this invention achieves, for the first time, early prediction and early intervention for neonatal neurodevelopmental disorders. The prediction model in this invention demonstrates excellent discrimination, consistency, and stability, making it an effective prediction model for neonatal neurodevelopmental monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] FIG1 is a nomogram of a logistic regression model used in a method for monitoring neonatal neurodevelopment in a first embodiment of the present invention;

[0036] FIG2 is a dose-response relationship diagram of the total population of the neonatal neurodevelopment monitoring method according to the first embodiment of the present invention;

[0037] FIG3 is a dose-response diagram of a female neonate according to a method for monitoring neonatal neurodevelopment according to the first embodiment of the present invention;

[0038] FIG4 is a dose-response diagram of a male neonate in the neonatal neurodevelopment monitoring method according to the first embodiment of the present invention;

[0039] 5A and 5B are ROC curve diagrams of the training set and validation set of the prediction model of the neonatal neurodevelopment monitoring method in the first embodiment of the present invention, respectively;

[0040] 6A and 6B are respectively calibration curve diagrams of the training set and the validation set of the prediction model of the neonatal neurodevelopment monitoring method in the first embodiment of the present invention;

[0041] 7A and 7B are respectively standardized benefit + confidence interval graphs of the training set and validation set of the prediction model of the neonatal neurodevelopmental monitoring method in the first embodiment of the present invention;

[0042] 8A and 8B are respectively non-standardized benefit + confidence interval graphs of the training set and validation set of the prediction model of the neonatal neurodevelopmental monitoring method in the first embodiment of the present invention;

[0043] FIG9 is a schematic structural diagram of a neonatal neurodevelopment monitoring system according to a second embodiment of the present invention. DETAILED DESCRIPTION

[0044] The technical content of the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0045] The technical concept of the embodiments of the present invention is to first establish a univariate logistic regression model. Then, using the criterion of parameter α less than 0.1, significant variables are screened and incorporated into a multivariate model. Multivariate logistic models are then constructed to adjust for different types of covariates, thereby predicting the impact of different covariates on neurodevelopmental damage caused by placental methylmercury. Based on this, a neonatal neurodevelopmental monitoring system is constructed to achieve early prediction and intervention for neurodevelopmental disorders in newborns.

[0046] First embodiment

[0047] The first embodiment of the present invention provides a method for monitoring neonatal neurodevelopment, comprising the following steps:

[0048] Selection step: In the prediction model, the corresponding model parameters are selected according to the patient's test information;

[0049] Prediction step: input the patient's test information and use the prediction model to obtain the prediction results;

[0050] Intervention steps: Provide intervention plans based on the predicted results.

[0051] Before explaining the above steps in detail, let’s first introduce the modeling method of the prediction model, which includes the following steps:

[0052] S1: Statistical analysis was performed on the overall population and male and female subgroup data using statistical tools (e.g., R language 4.3.1).

[0053] In the statistical analysis, the frequency and frequency were used to describe the qualitative variables, the mean (standard deviation) was used to describe the quantitative variables that conformed to the normal distribution, and the median (interquartile range) was used to describe the quantitative variables that did not conform to the normal distribution.

[0054] In one embodiment of the present invention, statistical analysis is performed based on the survey data from Region A and Region B (a total of 420 mother-child pairs).

[0055] S2: Analysis of influencing factors based on the logistic regression model.

[0056] In this embodiment, the logistic regression model is used because logistic regression analysis belongs to nonlinear regression and is a multiple regression analysis method for studying the relationship between the dependent variable, which is a binary classification or multinomial classification result, and certain influencing factors. Because the dependent variable is a categorical variable and the independent variable and the dependent variable are not linearly related, statistical models such as the multiple linear regression model (where the dependent variable is a continuous normally distributed variable and the independent variable and the dependent variable are linearly related) cannot be applied. Therefore, the logistic regression model has become one of the most commonly used analysis methods in medical research, particularly in epidemiological etiology research.

[0057] Logistic regression analysis has flexible requirements for independent variables; they can be binary, unordered, ordered, or quantitative. However, the independent variables must be assigned reasonable values. Using different variable assignment methods for the same data set can alter the estimated values, signs, and meanings of the parameters. Therefore, the rationality of variable assignment directly affects the effectiveness of logistic regression.

[0058] In one embodiment of the present invention, step S2 includes the following sub-steps:

[0059] S21; Establish a univariate logistic regression model, and then screen key variables and include them in the multivariate model according to the standard that the parameter α is less than 0.1;

[0060] The parameter α typically represents the slope or coefficient of the model, measuring the influence of the independent variable on the dependent variable. α is an estimated value obtained through training data and optimization algorithms. By adjusting the value of α, we can control how well the model fits the data and prevent overfitting or underfitting.

[0061] S22: A multivariate logistic model was established based on key variables and adjusted for different types of covariates to predict the effects of different types of covariates on placental methylmercury on neonatal neurodevelopmental damage.

[0062] The steps of establishing a univariate logistic regression model in step S21 are described in detail below.

[0063] In one embodiment of the present invention, the logistic regression model analysis results include the univariate logistic regression model analysis results for the total population, the univariate logistic regression model analysis results for males, and the univariate logistic regression model analysis results for females, which are described in detail below.

[0064] In this embodiment, the expression of the logistic regression model is:

[0065] Logit(Neurodevelopmental_Impairment)=2.93+0.20*Placenta_MeHg+0.12*Maternal_Age+0.04*Materal_Ethnicity+0.43*MDIFF+6.68*MIATDP-0.43*Paternal_Genetic_History- 1.56*FHHD+0.95*Paternal_smoking-1.64*Occupational_Mercury_Exposure+0.94*Apgar_Score_at_1_Min-3.86*Newborn_Congenital_Malformation-2.35*Birth_Season-0.01*Placental_Weight.

[0066] 1. Assignment of independent variables, assignment of dependent variables, and parameter estimation

[0067] Assignment of binary variables: no exposure to methylmercury is set as x = 0, exposure to methylmercury is set as x = 1, and the regression model is logit(p) = β0 + β 1x , then the formula for the non-exposed group is logit(p)=β0+β1×0, and for the exposed group is logit(p)=β0+β1×1, then the odds ratio of exposed to non-exposed subjects is OR=e β1 .

[0068] The assignment of quantitative variables is divided into the following cases:

[0069] ① Linear variables: When the independent variable has a linear relationship with logit(p), the independent variable can be directly included in the model in the form of a numerical value. In this case, e βj The meaning is the OR value when the independent variable xj changes by one unit under the premise that other independent variables remain unchanged.

[0070] ② Quadratic or multi-term variables: When the independent variable has a curvilinear relationship with logit(p), the independent variable can be raised to a quadratic or multi-term power. For example, if age x has a parabolic relationship with logit(p), the model can be fitted. logit(p) = β0 + β 1x +β2x 2 .

[0071] ③ Grouped linear variables: The independent variables are graded and then included in the model. This is applicable to the case where the independent variables are graded and have a linear relationship with logit(p). When grouped linear variables are included in the model, the model is logit(p)=β0+β 1x .

[0072] ④ Dummy variable: After the quantitative variable is grouped and hierarchical, if it is not linearly related to logit(p), it is represented by the assignment method of unordered multi-classification data, and its model construction is the same as the above-mentioned unordered multi-classification variable.

[0073] The dependent variable is assigned the same value as the categorical variable, with a "positive response" value of 1 and a "negative response" value of 0. If the order of the dependent variable y is reversed, the absolute value of the regression coefficient remains unchanged, but the sign is reversed.

[0074] 2. Steps of Logistic Regression Model Analysis

[0075] First, a univariate logistic regression model was established, and then significant variables were screened according to the standard of α less than 0.1.

[0076] Then, the significant variables were included in the multivariate logistic regression model, and multivariate logistic regression models with correction for different types of covariates were established in turn to explore the effects of different types of covariates on the neurodevelopmental damage caused by placental methylmercury.

[0077] In the total population, we explored the effects of different types of covariates on placental methylmercury on neurodevelopmental damage and multivariate analysis of neurodevelopmental damage. The established multivariate logistic regression model was as follows:

[0078] Model 1: placental methylmercury;

[0079] Model 2: placental methylmercury + father's general condition;

[0080] Model 3: placental methylmercury + father's general condition + mother's pregnancy condition;

[0081] Model 4: placental methylmercury + father's general condition + mother's pregnancy condition + biological indicators of mercury exposure before pregnancy;

[0082] Model 5: placental methylmercury + father's general condition + mother's pregnancy condition + biological indicators of mercury exposure before pregnancy + characteristics of newborn and delivery organs.

[0083] In the male population, the univariate logistic regression model showed that the inclusion of variables such as placental total mercury, placental methylmercury, paternal age, maternal age, bleeding volume, placental weight, anemia, calcium supplementation, primiparity, family member exposure, malformation, and birth season into the multivariate model led to the following multivariate logistic regression model:

[0084] Model 1: placental methylmercury;

[0085] Model 2: placental methylmercury + father's general condition;

[0086] Model 3: placental methylmercury + father's general condition + mother's pregnancy condition;

[0087] Model 4: placental methylmercury + father's general condition + mother's pregnancy condition + biological indicators of mercury exposure before pregnancy;

[0088] Model 5: placental methylmercury + father's general condition + mother's pregnancy condition + biological indicators of mercury exposure before pregnancy + characteristics of newborn and birth organs;

[0089] --: Not measurable.

[0090] In the female population, the univariate logistic regression model showed that, considering variables such as placental total mercury, umbilical cord blood total mercury, placental methylmercury, height, malformation, and birth season into the multivariate model, the established multivariate logistic regression model was as follows:

[0091] Model 1: placental methylmercury;

[0092] Model 2: placental methylmercury + biological indicators of pre-pregnancy mercury exposure;

[0093] Model 3: placental methylmercury + biological indicators of pre-pregnancy mercury exposure + characteristics of newborns and birth organs;

[0094] --: Not measurable.

[0095] Next, the multivariate logistic regression model is further explained.

[0096] (1) Multivariate logistic regression model analysis of the total population

[0097] Univariate logistic regression analysis of the total population suggested that variables such as placental total mercury, umbilical cord blood total mercury, placental methylmercury, paternal age, maternal age, placental weight, calcium supplementation, primiparity, malformation, birth season, and umbilical cord entanglement should be included in the multivariate model.

[0098] An investigation into the impact of different covariates on placental methylmercury on neurodevelopmental impairment suggests that, in the unadjusted covariate model (model 1), placental methylmercury was a risk factor for neurodevelopmental impairment. For every 1 ng / g increase in placental methylmercury concentration, the risk of neonatal neurodevelopmental impairment increased 1.23-fold (OR = 1.23, 95% CI = 1.11-1.36). After further adjustment for covariates such as "paternal general health" and "maternal pregnancy status," placental methylmercury showed no significant effect on neurodevelopmental impairment. After adjustment for biological markers of pre-pregnancy mercury exposure, the effect of placental methylmercury on neurodevelopmental impairment was significantly enhanced. For every 1 ng / g increase in placental methylmercury concentration, the risk of neonatal neurodevelopmental impairment increased 1.72-fold (OR = 1.72, 95% CI = 1.26-2.34). After adjusting for all types of covariates, for every 1 ng / g increase in placental methylmercury concentration, the risk of neonatal neurodevelopmental impairment increased by 1.39 times (OR = 1.39, 95% CI = 1.01-1.94).

[0099] Analysis of risk factors for neurodevelopmental impairment found that placental methylmercury was a risk factor for neurodevelopmental impairment (OR = 1.39, 95% CI = 1.01-1.94). Compared with neonates born between 6 and 8 months, neonates born between 9 and 11 months and between 12 and 2 months were protective factors for neurodevelopmental impairment (OR = 0.09, 95% CI = 0.03-0.29; OR = 0.02, 95% CI = 0.002-0.17).

[0100] Table 1: Effects of different types of covariates on placental methylmercury on neurodevelopmental damage and multivariate analysis of neurodevelopmental damage in the total population

[0101] β: regression coefficient; SE: standard error; OR: odds ratio; CI: confidence interval; Model 1: placental methylmercury; model 2: placental methylmercury + father's general condition; model 3: placental methylmercury + father's general condition + mother's pregnancy condition; model 4: placental methylmercury + father's general condition + mother's pregnancy condition + biological indicators of mercury exposure before pregnancy; model 5: placental methylmercury + father's general condition + mother's pregnancy condition + biological indicators of mercury exposure before pregnancy + characteristics of newborns and birth organs

[0102] (2) Multivariate logistic regression model analysis of male population

[0103] The univariate logistic regression model of the male population suggested that variables such as placental total mercury, placental methylmercury, paternal age, maternal age, bleeding volume, placental weight, anemia, calcium supplementation, primiparity, exposure of family members, deformity, and birth season should be included in the multivariate model.

[0104] Table 2: Effects of different types of covariates on placental methylmercury on neurodevelopmental damage and multivariate analysis of neurodevelopmental damage in the male population

[0105] β: regression coefficient; SE: standard error; OR: odds ratio; CI: confidence interval; Model 1: placental methylmercury; model 2: placental methylmercury + father's general condition; model 3: placental methylmercury + father's general condition + mother's pregnancy condition; model 4: placental methylmercury + father's general condition + mother's pregnancy condition + biological indicators of pre-pregnancy mercury exposure; model 5: placental methylmercury + father's general condition + mother's pregnancy condition + biological indicators of pre-pregnancy mercury exposure + characteristics of neonates and birth organs; --: unmeasurable

[0106] (3) Multivariate logistic regression model analysis of female population

[0107] The univariate logistic regression model of the female population suggested that variables such as total placental mercury, total cord blood mercury, placental methylmercury, height, deformity, and birth season should be included in the multivariate model.

[0108] Table 3: Effects of different types of covariates on placental methylmercury on neurodevelopmental damage and multivariate analysis of neurodevelopmental damage in the female population

[0109] β: regression coefficient; SE: standard error; OR: odds ratio; CI: confidence interval; Model 1: placental methylmercury; model 2: placental methylmercury + biological indicators of pre-pregnancy mercury exposure; model 3: placental methylmercury + biological indicators of pre-pregnancy mercury exposure + characteristics of neonates and birth organs; --: not measurable.

[0110] S3: Establish a prediction model based on the analysis results of the multivariate logistic regression model.

[0111] In one embodiment of the present invention, all data are first divided into a training set and a validation set at a ratio of 7:3. A lasso regression model is established in the training set to screen variables. The screened variables are incorporated into a logistic regression model, a prediction model is established, and a nomogram is plotted to visualize the model. In both the training and validation sets, a receiver operating characteristic (ROC) curve is used to determine the discrimination of the fitted model, a calibration curve is plotted to determine the calibration of the fitted model, and a clinical decision curve is plotted to determine the clinical benefit of the fitted model.

[0112] Finally, the following variables were selected as independent variables of the prediction model:

[0113] Placenta_MeHg: Methylmercury in the placenta

[0114] Maternal_Age: Mother's age

[0115] Material_Ethnicity: Mother's ethnicity

[0116] Maternal_Dietary_Intake_Frequency_of_Freshwater(MDIFF): Frequency of dietary intake of freshwater products by the mother

[0117] Maternal_Installing_Artificial_Teeth_During_Pregnancy(MIATDP): Maternal installation of artificial teeth during pregnancy

[0118] Paternal_Genetic_History: Father's genetic history

[0119] Family History of Hereditary Disease (FHHD): Family History of Hereditary Disease

[0120] Paternal_smoking: Father's smoking status

[0121] Occupational_Mercury_Exposure: Occupational mercury exposure

[0122] Apgar_Score_at_1_Min: Apgar score of newborn at 1 minute

[0123] Newborn_Congenital_Malformation: Newborn congenital malformation

[0124] Birth_Season: birth season

[0125] Placental_Weight: Placental weight.

[0126] The prediction model in this embodiment is as follows: Logit(Neurodevelopmental_Impairment)=2.93+0.20*Placenta_MeHg+0.12*Maternal_Age+0.04*Materal_Ethnicity+0.43*MDIFF+6.68*MIATDP-0.43*Paternal_Genetic_History-1.56*FHHD+0.95*Paternal_smoking-1.64*Occupational_Mercury_Exposure+0.94*Apgar_Score_at_1_Min-3.86*Newborn_Congenital_Malformation-2.35*Birth_Season-0.01*Placental_Weight.

[0127] Among them, Logit(Neurodevelopmental_Impairment) represents the probability of whether a newborn has a neurodevelopmental disorder.

[0128] Visualize the logistic regression model and create a nomogram using R or other statistical software. Based on the nomogram (see Figure 1), you can obtain scales for each predictor variable, a scale for the total score, and a probability scale. Based on the total score, find the corresponding probability on the probability scale and use it as the model's predicted probability of placental methylmercury causing neurodevelopmental damage in the newborn.

[0129] S3: Based on the prediction model, the probability of placental methylmercury causing neurodevelopmental damage to the newborn is predicted.

[0130] S4: Obtain dose-response relationships for interventions.

[0131] Dose-response graphs were constructed for the overall population and male and female subgroups to demonstrate the dose-response relationship between placental methylmercury and neurodevelopmental impairment. The dose-response relationship for the overall population is shown in Figure 2. The dose-response relationship in Figure 2 demonstrates a linear relationship between placental methylmercury concentration and the onset of neurodevelopmental impairment. When placental methylmercury concentrations exceed 0.98 ng / g, newborns begin to develop neurodevelopmental impairment (OR > 1). Therefore, intervention is initiated when placental methylmercury concentrations reach 0.98 ng / g.

[0132] Figure 3 explores the dose-response relationship by gender. A linear relationship is observed between the placental methylmercury level in female newborns and the incidence of neurodevelopmental impairment. When the placental methylmercury level exceeds 1.09 ng / g, newborns begin to experience neurodevelopmental impairment (OR > 1). Therefore, intervention should be initiated for female newborns when the mother's placental methylmercury level reaches 1.09 ng / g.

[0133] As shown in Figure 4, a linear relationship exists between the placental methylmercury level of male newborns and the occurrence of neurodevelopmental impairment. When the placental methylmercury level exceeds 0.97 ng / g, the newborn begins to experience neurodevelopmental impairment (OR > 1). Therefore, for male newborns, intervention should be initiated when the mother's placental methylmercury level reaches 0.97 ng / g.

[0134] The following describes the performance evaluation of this prediction model.

[0135] Discrimination was evaluated using the Receiver Operating Characteristic Curve (ROC) curve. The closer the AUC (Area Under Curve) for the validation and training sets is to 1, the better. As shown in Figures 5A and 5B , the AUC for the training set in this example reached 0.905, and the AUC for the validation set reached 0.863, a difference of 0.042, indicating that the prediction model has good discriminatory ability.

[0136] The results of the calibration evaluation are shown in Figures 6A and 6B. When evaluating model calibration, the closer the validation and training sets are to the middle diagonal line, the better; and the blue confidence interval should span the solid line as much as possible. As shown in the figures, the prediction model provided by the embodiment of the present invention demonstrates good calibration performance in the calibration evaluation.

[0137] Figures 7A and 7B are plots of the standardized gain + confidence interval for the validation and training sets, respectively, and Figures 8A and 8B are plots of the unstandardized gain + confidence interval for the validation and training sets, respectively. As can be seen from these figures, the prediction model provided by the embodiments of the present invention exhibits good consistency and stability. Standardized gain is a metric that measures the degree of improvement in a model's prediction performance relative to a baseline model. Calculating standardized gain on the training and validation sets can help us understand the performance improvement of the model on these two datasets. If the standardized gain on the validation set is similar to or slightly lower than that on the training set, this may indicate that the model has good generalization ability; however, if the standardized gain on the validation set is significantly lower than that on the training set, this may indicate that the model is overfitting. The confidence interval plot is used to visualize the confidence intervals of the model's performance metrics at different classification thresholds. By comparing the confidence interval plots on the training and validation sets, we can observe the stability of the model's performance at different thresholds. If the confidence interval plots on the training set and validation set are similar and overlap a lot, this may indicate that the model has good consistency across different datasets; if the two are significantly different, further adjustments to the model or optimization of the training strategy may be needed.

[0138] Second embodiment

[0139] Based on the above-mentioned neonatal neurodevelopmental monitoring method, a second embodiment of the present invention further provides a neonatal neurodevelopmental monitoring system. As shown in FIG9 , the neonatal neurodevelopmental monitoring system includes one or more processors 21 and a memory 22. The memory 22 is coupled to the processor 21 and is configured to store one or more programs. When the one or more programs are executed by the one or more processors 21, the one or more processors 21 implement the neonatal neurodevelopmental monitoring method described in the above-mentioned embodiment.

[0140] The processor 21 is used to control the overall operation of the neonatal neurodevelopmental monitoring system to complete all or part of the steps of the above-mentioned neonatal neurodevelopmental monitoring method. The processor 21 can be a central processing unit (CPU), a graphics processing unit (GPU), a field programmable gate array (FPGA), an application-specific integrated circuit (ASIC), a digital signal processing (DSP) chip, etc. The memory 22 is used to store various types of data to support the operation of the neonatal neurodevelopmental monitoring system. These data may include, for example, instructions for any application or method operating on the neonatal neurodevelopmental monitoring system, as well as application-related data. The memory 22 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, etc.

[0141] In an exemplary embodiment, the neonatal neurodevelopmental monitoring system can be implemented as a computer chip or entity, or as a product with certain functions, for executing the above-described neonatal neurodevelopmental monitoring method and achieving the same technical effects as the above-described method. A typical embodiment is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, an in-vehicle human-computer interaction device, a cellular phone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0142] In another exemplary embodiment, the present invention further provides a computer-readable storage medium comprising program instructions, which, when executed by a processor, implement the steps of the neonatal neurodevelopmental monitoring method described in any of the aforementioned embodiments. For example, the computer-readable storage medium may be the aforementioned memory comprising the program instructions, which may be executed by a system processor to perform the neonatal neurodevelopmental monitoring method described above and achieve the same technical effects as the aforementioned method.

[0143] It should be noted that the above-mentioned multiple embodiments are only examples. The technical solutions of the various embodiments can be combined and the order of the steps can be changed, all within the scope of protection of this patent.

[0144] The above describes in detail the neonatal neurodevelopmental monitoring method and system provided by the present invention. For those skilled in the art, any obvious modification made thereto without departing from the essence of the present invention will constitute an infringement of the patent rights of the present invention and will result in the corresponding legal liability.

Claims

1. A method for monitoring neonatal neurodevelopment, characterized in that: Based on a prediction model, predict the probability of neonatal neurodevelopmental impairment caused by placental methylmercury; The independent variables of the prediction model include: based on methylmercury in the placenta, maternal age, maternal ethnicity, maternal dietary intake frequency of fresh aquatic products, the situation of the mother installing dentures during pregnancy, paternal genetic history, family history of hereditary diseases, paternal smoking status, occupational mercury exposure, neonatal 1-minute Apgar score, neonatal congenital malformations, birth season, and placental weight, to predict the probability of whether a neonate has neurodevelopmental disorders; The expression of the prediction model is: Logit(Neurodevelopmental_Impairment) = 2.93 + 0.20 * Placenta_MeHg + 0.12 * Maternal_Age + 0.04 * Materal_Ethnicity + 0.43 * MDIFF + 6.68 * MIATDP - 0.43 * Paternal_Genetic_History - 1.56 * FHHD + 0.95 * Paternal_smoking - 1.64 * Occupational_Mercury_Exposure + 0.94 * Apgar_Score_at_1_Min - 3.86 * Newborn_Congenital_Malformation - 2.35 * Birth_Season - 0.01 * Placental_Weight, where Logit(Neurodevelopmental_Impairment): the probability of whether a neonate has neurodevelopmental disorders, Placenta_MeHg: methylmercury in the placenta Maternal_Age: maternal age Materal_Ethnicity: maternal ethnicity Maternal_Dietary_Intake_Frequency_of_Freshwater(MDIFF): maternal dietary intake frequency of fresh aquatic products Maternal_Installing_Artificial_Teeth_During_Pregnancy(MIATDP): the situation of the mother installing dentures during pregnancy Paternal_Genetic_History: paternal genetic history Family_History_of_Hereditary_Disease(FHHD): family history of hereditary diseases Paternal_smoking: paternal smoking status Occupational_Mercury_Exposure: occupational mercury exposure Apgar_Score_at_1_Min: neonatal 1-minute Apgar score Newborn_Congenital_Malformation: neonatal congenital malformations Birth_Season: Season of birth Placental_Weight: Placental weight.

2. The neonatal neurodevelopment monitoring method according to claim 1, characterized in that The prediction model is constructed through the following steps: S1: Use statistical tools to perform statistical analysis on the total population and male and female subgroup data; S2: Conduct analysis of influencing factors based on the logistic regression model; S3: Establish a prediction model based on the analysis results of the multi-factor logistic regression model.

3. The neonatal neurodevelopment monitoring method according to claim 1, wherein: Intervention is carried out when the methylmercury content in the placenta reaches 0.98 ng / g.

4. The neonatal neurodevelopment monitoring method according to claim 1, wherein: For female neonates, intervention is carried out when the methylmercury content in the mother's placenta reaches 1.09 ng / g.

5. The neonatal neurodevelopment monitoring method according to claim 1, wherein: For male neonates, intervention is carried out when the methylmercury content in the mother's placenta reaches 0.97 ng / g.

6. A neonatal neurodevelopment monitoring system, characterized in that It includes a processor and a memory; wherein, the memory is coupled to the processor and is used to store a computer program. When the computer program is executed by the processor, the processor implements the neonatal neurodevelopment monitoring method according to any one of claims 1 to 5.

Images