Electromagnetic wave biological effect dynamic evaluation method and system based on multilayer sensor

Abstract

The invention relates to the technical field of electromagnetic space safety, and particularly discloses an electromagnetic wave biological effect dynamic evaluation method and system based on a multi-layer perceptor.The method comprises the steps that electromagnetic parameters of electromagnetic waves, neural electrophysiological signals of organisms and data of behavioral indexes are obtained and preprocessed; electromagnetic parameters, neural electrophysiological signals and behavioral indexes are fused through a multi-layer perceptron, and an electromagnetic wave-neural dynamic-behavioral function mapping model is constructed; the neural activity anomaly degree and the behavior function deviation degree are generated in real time based on an electromagnetic wave-neural dynamic-behavior function mapping model, and an evaluation model is constructed; and inputting real-time data of the living body into the evaluation model, and dynamically adjusting electromagnetic exposure parameters or triggering a living body protection strategy.

Description

Technical Field

[0001] The present invention relates to the field of electromagnetic space safety technology, and in particular to a method and system for dynamically evaluating electromagnetic wave biological effects based on a multi-layer perceptron. Background Art

[0002] With the surge in frequency demand, electromagnetic wave transmission power is increasing, and electromagnetic wave density is increasing. The impact of electromagnetic waves on human health is causing public concern. Studies have explored the limits of thermal effects on organisms under electromagnetic conditions, and both international and domestic research has proposed measures and systems for electromagnetic protection equipment and protection. However, the mechanisms of non-thermal effects of electromagnetic waves on organisms, as well as quantitative measurement and assessment methods, are still lacking. Furthermore, there is a lack of safety management theories, methods, and corresponding protection standards.

[0003] The main proven adverse effect of radiofrequency electromagnetic wave exposure on organisms is neural stimulation. The nervous system transmits information based on bioelectrical signals and is more susceptible to external electromagnetic influences, making it considered a key target for damage from electromagnetic radiation exposure. Studies have reported that specifically modulated electromagnetic waves can affect sleep structure, spatial memory, and balance perception in mice, and related biochemical metabolic analyses have revealed abnormalities in neurotransmitters in the hippocampus. However, existing methods mostly study overall indicators of the organism's response to electromagnetic waves, resulting in delayed evaluation results. Furthermore, it is difficult to accurately quantify the dynamic impact of the non-thermal effects of electromagnetic waves on neural activity and behavioral function. These methods rely on offline biochemical indicator analysis and are unable to dynamically respond to electromagnetic exposure risks. Summary of the Invention

[0004] To achieve the purpose of the present invention, the present application provides a method for dynamic evaluation of electromagnetic wave bioeffects based on a multi-layer perceptron, comprising:

[0005] Step S1: obtaining electromagnetic parameters of electromagnetic waves, neurophysiological signals of organisms, and behavioral index data, and performing preprocessing;

[0006] Step S2: fusing the electromagnetic parameters, neurophysiological signals, and behavioral indicators through a multilayer perceptron to construct an “electromagnetic wave-neural dynamics-behavioral function” mapping model;

[0007] Step S3: generating the degree of abnormal neural activity and the degree of behavioral function deviation in real time based on the “electromagnetic wave-neural dynamics-behavioral function” mapping model, and constructing an evaluation model;

[0008] Step S4: Inputting the real-time data of the organism into the evaluation model to dynamically adjust the electromagnetic exposure parameters or trigger the organism protection strategy.

[0009] In some specific embodiments, step S1 includes:

[0010] The electromagnetic parameters include: frequency, power density, modulation frequency, and modulation mode of the electromagnetic wave;

[0011] The neuroelectrophysiological signals include: electroencephalogram, electrocardiogram and myoelectric signals of the organism;

[0012] The behavioral indicators include: the organism's motor activity and cognitive task performance.

[0013] In some specific embodiments, step S2 includes:

[0014] A database for each modality of the electromagnetic parameters, neuroelectrophysiological signals and behavioral indicators is established respectively, and the database is divided into a training set and a test set, wherein the training set includes biological individuals whose behavioral and neural activities have undergone significant changes and are accurately recorded; the feature subset in the training set corresponding to the characteristics of each modality is input into the corresponding multilayer perceptron for training to obtain the prediction results of this round; the difference between the data features of the prediction results and the experimental records is used to evaluate whether the performance of the multilayer perceptron meets the requirements; if it does not meet the requirements, the weights of each layer of the multilayer perceptron are adjusted to continue training; if it meets the requirements, the "electromagnetic wave-neural dynamics-behavioral function" mapping model is obtained from the trained multilayer perceptron.

[0015] In some specific embodiments, step S3 includes:

[0016] The error between the output of the evaluation model and the actual observation value calculation result is measured by the mean square error loss function, and the error is determined according to the following formula:

[0017]

[0018] Where a(i) represents the output of the neurons in the i-th layer, y represents the characteristic value calculated based on the experimental monitoring value, and M represents the total number of training input samples, that is, the number of biological individuals.

[0019] In some specific embodiments, step S3 includes:

[0020] The error is calculated by backpropagation to update the weights and bias column vectors of each layer of the multilayer perceptron, and the parameters are updated using the Adam optimizer to reduce the error output by the evaluation model.

[0021] In some specific embodiments, step S4 includes:

[0022] Through cross-validation, the database is randomly divided into several groups, and trained and tested repeatedly; and by adjusting the number of neural network layers and the number of neurons in each layer in the multi-layer perceptron, the parameters of the neural network model are adjusted to determine the optimal parameter combination of the evaluation model.

[0023] To achieve the same invention purpose, the present application also provides a dynamic evaluation system for electromagnetic wave bioeffects based on a multi-layer perceptron, comprising:

[0024] Multimodal data acquisition module: used to obtain electromagnetic parameters of electromagnetic waves, neurophysiological signals of organisms and behavioral indicators, and perform preprocessing;

[0025] Feature fusion and model training module: used to fuse the electromagnetic parameters, neurophysiological signals and behavioral indicators through a multi-layer perceptron to build an "electromagnetic wave-neural dynamics-behavioral function" mapping model;

[0026] Dynamic evaluation module: used to generate the degree of neural activity abnormality and behavioral function deviation in real time based on the "electromagnetic wave-neural dynamics-behavioral function" mapping model, and build an evaluation model;

[0027] Biological protection module: used to input real-time biological data into the evaluation model, dynamically adjust electromagnetic exposure parameters or trigger biological protection strategies.

[0028] In some specific embodiments, the feature fusion and model training module is used to:

[0029] A database for each modality of the electromagnetic parameters, neuroelectrophysiological signals and behavioral indicators is established respectively, and the database is divided into a training set and a test set, wherein the training set includes biological individuals whose behavioral and neural activities have undergone significant changes and are accurately recorded; the feature subset in the training set corresponding to the characteristics of each modality is input into the corresponding multilayer perceptron for training to obtain the prediction results of this round; the difference between the data features of the prediction results and the experimental records is used to evaluate whether the performance of the multilayer perceptron meets the requirements; if it does not meet the requirements, the weights of each layer of the multilayer perceptron are adjusted to continue training; if it meets the requirements, the "electromagnetic wave-neural dynamics-behavioral function" mapping model is obtained from the trained multilayer perceptron.

[0030] In some specific embodiments, the dynamic assessment module is used to:

[0031] The error between the output of the evaluation model and the actual observation value calculation result is measured by the mean square error loss function, and the error is determined according to the following formula:

[0032]

[0033] Where a(i) represents the output of the neurons in the i-th layer, y represents the characteristic value calculated based on the experimental monitoring value, and M represents the total number of training input samples, that is, the number of biological individuals.

[0034] In some specific embodiments, the dynamic assessment module is used to:

[0035] The error is calculated by backpropagation to update the weights and bias column vectors of each layer of the multilayer perceptron, and the parameters are updated using the Adam optimizer to reduce the error output by the evaluation model.

[0036] Beneficial effects of the above technical solution:

[0037] This invention allows for objective and efficient recording and fusion analysis of multimodal neuroelectrophysiological signals. It no longer relies on biochemical metabolic indicators, improving accuracy, efficiency, and temporal resolution while simultaneously considering the interplay of multiple electrophysiological indicators and information transmission pathways. Previous studies have often focused solely on behavioral and biochemical metabolism, failing to assess real-time bioelectromagnetic changes directly affected by electromagnetic waves.

[0038] The present invention adopts a multi-layer perceptron model to fuse multimodal data features and model training. Its advantages are that it has a strong ability to process high-dimensional data. The introduction of nonlinear activation function MLP can learn complex nonlinear relationships in input data. It has strong scalability and can freely adjust the number of hidden layers and the number of neurons in each layer to adapt to problems of different complexities.

[0039] The assessment model in the present invention has a strong ability to handle individual differences. Since there are individual differences in the electromagnetic sensitivity of organisms, this needs to be taken into account in the parameter optimization of the model; the model trained based on species and individual information has the ability to provide individualized solutions for safety assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0041] Figure 1 A schematic flow chart of a method for dynamic evaluation of electromagnetic wave bioeffects based on a multi-layer perceptron according to an embodiment of the present invention;

[0042] Figure 2 A schematic diagram of the neural entrainment mechanism and behavioral feedback of a multi-layer perceptron-based dynamic assessment method for electromagnetic wave bioeffects according to one embodiment of the present invention;

[0043] Figure 3 A schematic structural diagram of a multi-layer perceptron-based electromagnetic wave bioeffect dynamic assessment system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0045] Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar symbols throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0046] Example 1

[0047] One embodiment of the present invention provides a method for dynamic evaluation of electromagnetic wave bioeffects based on a multilayer perceptron, referring to Figure 1 、 Figure 2 Shown, including:

[0048] Step S1: obtaining electromagnetic parameters of electromagnetic waves, neurophysiological signals of organisms, and behavioral index data, and performing preprocessing;

[0049] Step S2: fusing the electromagnetic parameters, neurophysiological signals, and behavioral indicators through a multilayer perceptron to construct an “electromagnetic wave-neural dynamics-behavioral function” mapping model;

[0050] Step S3: generating the degree of abnormal neural activity and the degree of behavioral function deviation in real time based on the “electromagnetic wave-neural dynamics-behavioral function” mapping model, and constructing an evaluation model;

[0051] Step S4: Inputting the real-time data of the organism into the evaluation model to dynamically adjust the electromagnetic exposure parameters or trigger the organism protection strategy.

[0052] In a specific embodiment of the present invention, step S1 includes:

[0053] The electromagnetic parameters include: electromagnetic wave carrier frequency, power density, modulation frequency, and modulation mode;

[0054] The neuroelectrophysiological signals include: electroencephalogram, electrocardiogram and myoelectric signals of the organism;

[0055] The behavioral indicators include: the organism's motor activity and cognitive task performance.

[0056] Specifically, non-primates and primates were selected for neurophysiological signal and behavioral recordings under electromagnetic wave stimulation. Based on the size and classical paradigm of C57BL / 6 mice and cynomolgus macaques, electromagnetic shielding environment and behavioral experimental design were carried out respectively. The organisms were electromagnetically irradiated in an anechoic chamber using a horn antenna, and the following parameters were recorded simultaneously:

[0057] Electromagnetic parameters: The carrier frequency (such as 2.8GHz), power density (such as 4.5W / m 2 ) and modulation frequency mode (such as 10Hz), modulation mode (square wave / sine wave).

[0058] Table 1 Examples of characteristics of specific modulated electromagnetic waves

[0059] electromagnetic characteristics Parameter range Carrier frequency 1G~3GHz Power density <![CDATA[1~10W / m 2 ]]> Modulation frequency 1~50Hz Modulation mode Square wave / sine wave Duty cycle 50％

[0060] Neuroelectrophysiological signals: Implanted electrodes are used to collect local field potential (LFP) signals from the mouse hippocampus; or whole-brain electroencephalogram (EEG) signals are collected from cynomolgus monkeys using an EEG cap. Electrocardiogram (ECG) RR interval variability and electromyography (EMG) burst frequency are simultaneously recorded. High-resolution multimodal neural signals contain rich temporal and spectral features. After preprocessing, including filtering, noise removal, and missing value filling, they are then time-aligned and time-frequency decomposition is performed to identify typical frequency bands or rhythms.

[0061] Table 2 Examples of multimodal neuroelectrophysiological features

[0062]

[0063] Behavioral Indicators: Mice were subjected to an open field test, and the total length of movement trajectory (cm) and maximum movement speed (m / s) were quantified using a video tracking system and behavioral analysis software. A reversal learning task was designed for cynomolgus monkeys, and cognitive accuracy (%) and reaction time (ms) were used as high-level behavioral indicators.

[0064] Table 3 Examples of behavioral characteristics

[0065]

[0066] In a specific embodiment of the present invention, step S2 includes:

[0067] A database for each modality of the electromagnetic parameters, neuroelectrophysiological signals and behavioral indicators is established respectively, and the database is divided into a training set and a test set, wherein the training set includes biological individuals whose behavioral and neural activities have undergone significant changes and are accurately recorded; the feature subset in the training set corresponding to the characteristics of each modality is input into the corresponding multilayer perceptron for training to obtain the prediction results of this round; the difference between the data features of the prediction results and the experimental records is used to evaluate whether the performance of the multilayer perceptron meets the requirements; if it does not meet the requirements, the weights of each layer of the multilayer perceptron are adjusted to continue training; if it meets the requirements, the "electromagnetic wave-neural dynamics-behavioral function" mapping model is obtained from the trained multilayer perceptron.

[0068] Specifically, a multilayer perceptron (MLP) is a feed-forward artificial neural network that includes an input layer, an output layer, and multiple hidden layers.

[0069] Input layer: Standardized multi-dimensional data, including: Electromagnetic characteristics: carrier frequency, power density, modulation frequency, modulation mode encoding (one-hot encoding, e.g., 4 bits, indicating the presence of modulation and modulation waveform, respectively). Neural characteristics: LFP / EEG peak value and corresponding power in each frequency band, ECG variability, EMG burst frequency. Behavioral characteristics: movement trajectory length and maximum movement speed, cognitive accuracy, and reaction time. Individual characteristics: species (coded as 0 / 1), age (months), and exposure duration (hours).

[0070] Assuming there are m samples and n features, the input layer X∈R m×n , use i to represent the number of layers of the neural network, and assume that the hidden layer H has h(i) neurons, then the weight and bias of the hidden layer can be expressed as W h ∈R n×h 、b h ∈R 1×h The output layer O outputs q label values, so the weight and bias parameters of the output layer are W o ∈R h×q 、b h ∈R 1×q The output of the hidden layer and the output of the output layer can be calculated by the following formula:

[0071] H=XW h +b h

[0072] O=HW o +b0

[0073] As can be seen from the formula, the input of each layer is the output of the previous layer. Multi-layer neural networks propagate information by continuously iterating the above formula. The above formula only performs a linear transformation on the data. An activation function is also required to perform a nonlinear transformation on each layer (except the input layer).

[0074] Hidden layer: z(i) represents the net input of the neurons in the i-th layer, a(i) represents the output of the neurons in the i-th layer, and the net activity value z(i) of the neurons in the i-th layer is calculated based on the activity value a(i-1) of the neurons in the i-1-th layer. Then, an activation function is used to calculate the activity value a(i) of the neurons in the i-th layer, that is:

[0075] z(i)=W(i)a(i-1)+b(i)

[0076] a(i)=f i (z(i))

[0077] The merge can be written as

[0078] a(i)=f i (W(i)a(i-1)+b(i)

[0079] Use the LeakyReLU function as the hidden layer activation function to avoid neuron death, alleviate gradient disappearance, and be suitable for high-dimensional feature extraction and retain negative information:

[0080] LeakyReLU(x)=max(0,x)+negtive_slope×min(0,x)

[0081] Batch Normalization and Dropout are then used to accelerate training and stabilize gradients through regularization, thereby reducing overfitting and enhancing the generalization ability of the model.

[0082] Output layer: Use the Tanh function to compress the features to [-1, 1] and directly represent the abnormality ratio.

[0083]

[0084] Abnormal degree of neural activity (-1 to 1): reflects the standardized change ratio of the degree to which the power of each frequency band of EEG deviates from the baseline. Negative values ​​represent a decrease ratio, and positive values ​​represent an increase ratio.

[0085] Behavioral function deviation (-1 to 1): percentage change in motor or cognitive accuracy (e.g., -0.3 means a 30% decrease).

[0086] In a specific embodiment of the present invention, step S3 includes:

[0087] The error between the output of the evaluation model and the actual observation value calculation result is measured by the mean square error loss function, and the error is determined according to the following formula:

[0088]

[0089] Where a(i) represents the output of the neurons in the i-th layer, y represents the characteristic value calculated based on the experimental monitoring value, and M represents the total number of training input samples, that is, the number of biological individuals.

[0090] In a specific embodiment of the present invention, step S3 includes:

[0091] The error is calculated by backpropagation to update the weights and bias column vectors of each layer of the multilayer perceptron, and the parameters are updated using the Adam optimizer to reduce the error output by the evaluation model.

[0092] Specifically, the gradient of the loss with respect to the model parameters (weights W and biases b) is calculated through backpropagation, and the parameters are then updated using the Adam optimizer. Adam efficiently optimizes model parameters through adaptive learning rates and momentum mechanisms, improving robustness. Furthermore, dynamic parameter adjustment is implemented to fine-tune hidden layer weights based on species characteristics to improve cross-species generalization.

[0093] The evaluation model also uses a cross-validation approach to randomly group the database into several groups multiple times, and conduct training and testing after each grouping, and repeated testing; in addition, by adjusting the number of neural network layers and the number of neurons in each layer in the multi-layer perceptron model, trying deeper networks (more than 3 layers) requires stronger regularization, continuously adjusting the parameters of the neural network model, and finally outputting the optimal parameter combination of the evaluation model.

[0094] In a specific embodiment of the present invention, step S4 includes:

[0095] Through cross-validation, the database is randomly divided into several groups, and trained and tested repeatedly; and by adjusting the number of neural network layers and the number of neurons in each layer in the multi-layer perceptron, the parameters of the neural network model are adjusted to determine the optimal parameter combination of the evaluation model.

[0096] Based on the model trained above, real-time data can be input to obtain personalized protection strategies, such as:

[0097] Current electromagnetic parameters (such as 2.8GHz carrier, power density 4W / m 2 , modulation waveform square wave 10 Hz, individual characteristics (cynomolgus monkey, age 36 months, exposure duration 2 h).

[0098] Output:

[0099] If the degree of neural activity abnormality is >0.6 and the degree of behavioral function deviation is >0.4, the protection strategy will be triggered. Power adjustment: electromagnetic intensity is reduced to 1W / m 2 ; Targeted shielding: activating local electromagnetic absorption materials (such as graphene coatings).

[0100] Mice were exposed to 2.8 GHz pulsed electromagnetic waves using the method of this application for evaluation:

[0101] Data collection:

[0102] Electromagnetic parameters: 2.8 GHz, square wave 10 Hz pulse modulation (duty cycle 50%), intensity 3.5 W / m 2 .

[0103] Neural signals: Hippocampal theta wave power increased by 40%, and EMG burst frequency increased by 2 times.

[0104] Behavioral indicators: open field movement distance decreased by 60%, and maximum movement speed decreased by 35%.

[0105] Model Evaluation:

[0106] After inputting the feature vector, the MLP outputs a neurological abnormality of 0.72, a behavioral function deviation of 0.55, and a trigger power reduction of 0.8W / m 2 .

[0107] Example 2

[0108] One embodiment of the present invention provides a dynamic evaluation system for electromagnetic wave bioeffects based on a multi-layer perceptron, referring to Figure 3 Shown, including:

[0109] Multimodal data acquisition module 10: used to obtain electromagnetic parameters of electromagnetic waves, neurophysiological signals of organisms and behavioral indicators, and perform preprocessing;

[0110] Feature fusion and model training module 20: used to fuse the electromagnetic parameters, neural electrophysiological signals and behavioral indicators through a multi-layer perceptron to construct an "electromagnetic wave-neural dynamics-behavioral function" mapping model;

[0111] Dynamic evaluation module 30: used to generate the degree of neural activity abnormality and the degree of behavioral function deviation in real time based on the "electromagnetic wave-neural dynamics-behavioral function" mapping model, and construct an evaluation model;

[0112] Biological protection module 40: used to input real-time biological data into the evaluation model, dynamically adjust electromagnetic exposure parameters or trigger biological protection strategies.

[0113] In a specific embodiment of the present invention, the feature fusion and model training module 20 is used to:

[0114] A database for each modality of the electromagnetic parameters, neuroelectrophysiological signals and behavioral indicators is established respectively, and the database is divided into a training set and a test set, wherein the training set includes biological individuals whose behavioral and neural activities have undergone significant changes and are accurately recorded; the feature subset in the training set corresponding to the characteristics of each modality is input into the corresponding multilayer perceptron for training to obtain the prediction results of this round; the difference between the data features of the prediction results and the experimental records is used to evaluate whether the performance of the multilayer perceptron meets the requirements; if it does not meet the requirements, the weights of each layer of the multilayer perceptron are adjusted to continue training; if it meets the requirements, the "electromagnetic wave-neural dynamics-behavioral function" mapping model is obtained from the trained multilayer perceptron.

[0115] In a specific embodiment of the present invention, the dynamic evaluation module 30 is used to:

[0116] The error between the output of the evaluation model and the actual observation value calculation result is measured by the mean square error loss function, and the error is determined according to the following formula:

[0117]

[0118] Where a(i) represents the output of the neurons in the i-th layer, y represents the characteristic value calculated based on the experimental monitoring value, and M represents the total number of training input samples, that is, the number of biological individuals.

[0119] In a specific embodiment of the present invention, the dynamic evaluation module 30 is used to:

[0120] The error is calculated by backpropagation to update the weights and bias column vectors of each layer of the multilayer perceptron, and the parameters are updated using the Adam optimizer to reduce the error output by the evaluation model.

[0121] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

[0122] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same and similar parts between the various embodiments can be referred to each other. The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing terminal device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the functions in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing terminal device to work in a specific way, so that the instructions stored in the computer-readable memory produce a product including an instruction device, which implements the functions specified in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device, so that a series of operation steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for implementing the process in the process. Figure 1 a process or multiple processes and / or boxes Figure 1The steps of the functions specified in one or more blocks. Although preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they are aware of the basic inventive concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention. Finally, it should be noted that, in this document, relational terms such as first and second, etc. are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or terminal device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article, or terminal device. Without further limitation, an element defined by the phrase "comprising a..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes the element.

[0123] The method and apparatus provided by the present invention are introduced in detail above. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method and core ideas of the present invention. At the same time, for those skilled in the art, according to the ideas of the present invention, there may be changes in the specific implementation methods and application scopes. In summary, the contents of this specification should not be understood as limiting the present invention.

[0124] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "example," "specific example," "a specific embodiment," or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the present invention. In this specification, schematic representations of terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0125] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, 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 application.

Claims

1. A method for dynamic evaluation of electromagnetic wave bioeffects based on a multilayer perceptron, characterized in that: include: Step S1: obtaining electromagnetic parameters of electromagnetic waves, neurophysiological signals of organisms, and behavioral index data, and performing preprocessing; Step S2: fusing the electromagnetic parameters, neurophysiological signals, and behavioral indicators through a multilayer perceptron to construct an "electromagnetic wave-neural dynamics-behavioral function" mapping model; Step S3: generating the degree of neural activity abnormality and the degree of behavioral function deviation in real time based on the "electromagnetic wave-neural dynamics-behavioral function" mapping model, and constructing an evaluation model; Step S4: Input the real-time data of the organism into the evaluation model, and dynamically adjust the electromagnetic exposure parameters or trigger the organism protection strategy according to the evaluation results.

2. The method for dynamic evaluation of electromagnetic wave bioeffects based on multi-layer perceptron according to claim 1, characterized in that: Step S1 includes: The electromagnetic parameters include: frequency, power density, modulation frequency, and modulation mode of the electromagnetic wave; The neuroelectrophysiological signals include: electroencephalogram, electrocardiogram and myoelectric signals of the organism; The behavioral indicators include: the organism's motor activity and cognitive task performance.

3. The method for dynamic evaluation of electromagnetic wave bioeffects based on multi-layer perceptron according to claim 1, characterized in that: Step S2 includes: A database for each modality of the electromagnetic parameters, neuroelectrophysiological signals, and behavioral indicators is established separately, and the database is divided into a training set and a test set. The training set includes biological individuals whose behavioral and neural activity have undergone significant changes and are accurately recorded. The feature subset of each modality in the training set is input into the corresponding multilayer perceptron for training to obtain the prediction results of this round. The performance of the multilayer perceptron is evaluated by evaluating the difference in data features between the prediction results and the experimental records to determine whether it meets the requirements. If it does not meet the requirements, the weights of each layer of the multilayer perceptron are adjusted to continue training. If it meets the requirements, the "electromagnetic wave-neural dynamics-behavioral function" mapping model is obtained from the trained multilayer perceptron.

4. The method for dynamic evaluation of electromagnetic wave bioeffects based on multi-layer perceptron according to claim 1, characterized in that: Step S3 further includes: The error between the output of the evaluation model and the actual observation value calculation result is measured by the mean square error loss function, and the error is determined according to the following formula: Where a(i) represents the output of the neuron in the i-th layer, y represents the characteristic value calculated based on the experimental monitoring value, and M represents the total number of training input samples.

5. The method for dynamic evaluation of electromagnetic wave bioeffects based on multi-layer perceptron according to claim 4, characterized in that: Step S3 further includes: The error is calculated by backpropagation to update the weights and bias column vectors of each layer of the multilayer perceptron, and the parameters are updated using the Adam optimizer to reduce the error output by the evaluation model.

6. The method for dynamic evaluation of electromagnetic wave bioeffects based on multi-layer perceptron according to claim 4, characterized in that: Step S3 further includes: Through cross-validation, the database is randomly divided into several sub-databases multiple times, and trained and tested repeatedly; and by adjusting the number of neural network layers and the number of neurons in each layer in the multi-layer perceptron, the parameters of the neural network model are adjusted to determine the optimal parameter combination of the evaluation model.

7. A dynamic evaluation system for electromagnetic wave bioeffects based on a multi-layer perceptron, characterized in that: include: Multimodal data acquisition module: used to obtain electromagnetic parameters of electromagnetic waves, neurophysiological signals of organisms and behavioral indicators, and perform preprocessing; Feature fusion and model training module: used to fuse the electromagnetic parameters, neurophysiological signals and behavioral indicators through a multi-layer perceptron to build an "electromagnetic wave-neural dynamics-behavioral function" mapping model; Dynamic evaluation module: used to generate the degree of neural activity abnormality and behavioral function deviation in real time based on the "electromagnetic wave-neural dynamics-behavioral function" mapping model, and build an evaluation model; Biological protection module: used to input real-time biological data into the evaluation model, dynamically adjust electromagnetic exposure parameters or trigger biological protection strategies.

8. The electromagnetic wave bioeffect dynamic assessment system based on multi-layer perceptron according to claim 7 is characterized in that: The feature fusion and model training module is used to: A database for each modality of the electromagnetic parameters, neuroelectrophysiological signals, and behavioral indicators is established separately, and the database is divided into a training set and a test set. The training set includes biological individuals whose behavioral and neural activity have undergone significant changes and are accurately recorded. The feature subset of each modality in the training set is input into the corresponding multilayer perceptron for training to obtain the prediction results of this round. The performance of the multilayer perceptron is evaluated by evaluating the difference in data features between the prediction results and the experimental records to determine whether it meets the requirements. If it does not meet the requirements, the weights of each layer of the multilayer perceptron are adjusted to continue training. If it meets the requirements, the "electromagnetic wave-neural dynamics-behavioral function" mapping model is obtained from the trained multilayer perceptron.

9. The electromagnetic wave bioeffect dynamic assessment system based on multi-layer perceptron according to claim 7, characterized in that: The dynamic evaluation module is used to: The error between the output of the evaluation model and the actual observation value calculation result is measured by the mean square error loss function, and the error is determined according to the following formula: Where a(i) represents the output of the neuron in the i-th layer, y represents the characteristic value calculated based on the experimental monitoring value, and M represents the total number of training input samples.

10. The electromagnetic wave bioeffect dynamic assessment system based on multi-layer perceptron according to claim 9, characterized in that: The dynamic evaluation module is used to: The error is calculated by backpropagation to update the weights and bias column vectors of each layer of the multilayer perceptron, and the parameters are updated using the Adam optimizer to reduce the error output by the evaluation model.

Images