Focused ultrasound open blood brain barrier effect prediction method and system, medium and terminal

Abstract

The invention provides a focused ultrasound blood brain barrier opening effect prediction method and system, a medium and a terminal, and the method comprises the steps: carrying out the preprocessing of all collected acoustic signal data, and constructing a training data set based on the blood brain barrier opening effect data; training a deep learning model by using the training data set to construct a blood brain barrier prediction model; and on the basis of the blood brain barrier prediction model, predicting the permeability of the blood brain barrier according to the obtained real-time acoustic signal data of the ultrasonic contrast agent in the process of regulating and controlling the blood brain barrier so as to output a blood brain barrier opening effect prediction result. The acoustic signals are analyzed based on the artificial intelligence algorithm, a magnetic resonance contrast agent does not need to be injected, the long-time scanning process is avoided, safe and rapid prediction of the blood brain barrier permeability is achieved, and the test cost is reduced. Compared with a dye tracing method, the blood brain barrier permeability prediction method based on artificial intelligence does not need to perform any operation on a target test, is convenient and rapid, and has a living body detection capability.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method, system, medium and terminal for predicting the effect of focused ultrasound on opening the blood-brain barrier. Background Art

[0002] The blood-brain barrier (BBB) ​​is a semipermeable physiological barrier in the brain that prevents harmful molecules or toxic substances from entering and exiting the blood and brain tissue at will, maintaining homeostasis in the animal body and protecting the brain. However, when drugs are needed to treat brain lesions, the blood-brain barrier greatly hinders the delivery of drugs. In order to overcome this problem, researchers are constantly exploring new methods to temporarily increase the permeability of the blood-brain barrier, thereby improving the efficiency of drug delivery in the brain.

[0003] The current methods for evaluating the opening effect of the blood-brain barrier include contrast-enhanced magnetic resonance imaging and dye tracer methods. Although they can effectively evaluate the opening effect of the blood-brain barrier, they have the following disadvantages:

[0004] (1) Enhanced T1-weighted magnetic resonance imaging (MRI) is currently the clinical gold standard for evaluating the effect of blood-brain barrier regulation. Its basic principle is that intravenously injected magnetic resonance contrast agents (such as gadolinium) can enter the brain tissue through the blood-brain barrier with increased permeability under ultrasound regulation, so that the high-brightness distribution of magnetic resonance contrast agents (such as gadolinium) in the tissue in this area can be displayed on the T1-weighted MRI image. This method has high spatial resolution and sensitivity, and can reflect the changes in blood-brain barrier permeability in real time and intuitively. However, the disadvantages of this method are that it requires expensive magnetic resonance equipment, high testing costs, long scanning time, and the injection of magnetic resonance contrast agents (such as gadolinium) that are toxic and difficult to excrete from the body, which may pose potential health risks to patients; at the same time, it is not suitable for patients with metal implants such as pacemakers and claustrophobia.

[0005] (2) Dye tracer method: First, a dye (such as Evans blue) is injected into the animal's vein. When the dye passes through the target area regulated by focused ultrasound, if the permeability of the blood-brain barrier in this area increases, the dye molecules will pass through the blood vessels and enter the tissue. The animal is then killed and the target tissue is removed. The distribution of the dye in the target tissue is detected by a spectrophotometer or fluorescence microscope to evaluate the regulation effect of the blood-brain barrier. This method requires the animal to be killed and the target tissue to be removed for testing. The whole process is cumbersome and time-consuming. In addition, this method cannot evaluate the changes in the permeability of the blood-brain barrier in real time in the living state, and has certain deficiencies in sensitivity and specificity, which limits its potential in clinical applications. Summary of the invention

[0006] In view of the shortcomings of the prior art mentioned above, the purpose of the present application is to provide a method, system, medium and terminal for predicting the effect of focused ultrasound on opening the blood-brain barrier, so as to solve the technical problems in the prior art such as the need to inject magnetic resonance contrast agents that are toxic and difficult to be excreted from the body, high testing costs, long scanning times, and the inability to evaluate the effect of focused ultrasound on opening the blood-brain barrier in real time in a living state.

[0007] To achieve the above-mentioned purpose and other related purposes, the first aspect of the present application provides a method for predicting the effect of focused ultrasound on opening the blood-brain barrier, comprising: collecting a number of acoustic signal data and a number of blood-brain barrier opening effect data of an ultrasound contrast agent in the process of regulating the blood-brain barrier; preprocessing each of the collected acoustic signal data respectively, and constructing a training data set based on the blood-brain barrier opening effect data; using the training data set to train a deep learning model to construct a blood-brain barrier prediction model; based on the blood-brain barrier prediction model, predicting the blood-brain barrier opening effect according to the acquired real-time acoustic signal data of the ultrasound contrast agent in the process of regulating the blood-brain barrier, so as to output the blood-brain barrier opening effect prediction result data.

[0008] In some embodiments of the first aspect of the present application, each of the collected acoustic signal data is preprocessed separately, and based on the blood-brain barrier opening effect data, a training data set is constructed in a manner including: performing Fourier transform preprocessing on each of the acoustic signal data to obtain a number of acoustic signal frequency domain feature data; normalizing the acoustic signal frequency domain feature data that meet the preset frequency range to obtain acoustic signal training data; matching and annotating the acoustic signal training data with the blood-brain barrier opening effect data to construct a training data set; the training data set includes multiple groups of data pairs, each group of data pairs includes acoustic signal training data and blood-brain barrier opening effect data corresponding to the acoustic signal training data.

[0009] In some embodiments of the first aspect of the present application, the blood-brain barrier prediction model includes an acoustic coding module, a feature aggregation module, and a classification module; the acoustic coding module is used to encode the real-time acoustic signal data of the input ultrasound contrast agent in the process of regulating the blood-brain barrier to obtain a first feature vector; the feature aggregation module is used to perform feature fusion processing on the first feature vector to obtain a second feature vector; the classification module is used to output the blood-brain barrier opening effect prediction result data based on the second feature vector.

[0010] In some embodiments of the first aspect of the present application, the method of using the training data set to train a deep learning model to construct a blood-brain barrier prediction model includes: training the deep learning model based on the training data set to obtain a primary blood-brain barrier prediction model, and generating primary blood-brain barrier opening effect prediction result data corresponding to the acoustic signal training data in each data pair according to the primary blood-brain barrier prediction model; testing and evaluating the deep learning model using cross-validation technology based on the primary blood-brain barrier opening effect prediction result data and blood-brain barrier opening effect data corresponding to the acoustic signal training data in each data pair; and adjusting the parameters of the deep learning model according to the test and evaluation results to construct a finally converged blood-brain barrier prediction model.

[0011] In some embodiments of the first aspect of the present application, the acoustic coding module includes a first fully connected layer and a first activation function layer.

[0012] In some embodiments of the first aspect of the present application, the feature aggregation module includes a first normalization layer and a multi-head self-attention layer.

[0013] In some embodiments of the first aspect of the present application, the acoustic coding module also includes a plurality of one-dimensional residual layers and flattening layers; after each one-dimensional residual layer is connected in sequence, the first fully connected layer and the first activation function layer are connected in sequence through the flattening layer.

[0014] To achieve the above-mentioned purpose and other related purposes, the second aspect of the present application provides a focused ultrasound blood-brain barrier opening effect prediction system, including: a data acquisition module, used to collect a number of acoustic signal data and a number of blood-brain barrier opening effect data of the ultrasound contrast agent in the process of regulating the blood-brain barrier; a blood-brain barrier prediction model construction module, used to pre-process each of the collected acoustic signal data respectively, and construct a training data set based on the blood-brain barrier opening effect data; use the training data set to train a deep learning model to construct a blood-brain barrier prediction model; a blood-brain barrier opening effect prediction module, used to predict the blood-brain barrier opening effect based on the blood-brain barrier prediction model and the acquired real-time acoustic signal data of the ultrasound contrast agent in the process of regulating the blood-brain barrier, so as to output the blood-brain barrier opening effect prediction result data.

[0015] To achieve the above-mentioned purpose and other related purposes, the third aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for predicting the effect of focused ultrasound on opening the blood-brain barrier as described above.

[0016] To achieve the above-mentioned purpose and other related purposes, the fourth aspect of the present application provides an electronic terminal, including a memory, a processor and a computer program stored in the memory, and the processor executes the computer program to implement the method for predicting the effect of focused ultrasound on opening the blood-brain barrier as described above.

[0017] As described above, the method, system, medium and terminal for predicting the effect of focused ultrasound opening of the blood-brain barrier of the present application have the following beneficial effects: by comprehensively processing and analyzing the real-time acoustic signal data of the ultrasound contrast agent in the process of regulating the blood-brain barrier, it replaces the traditional MRI examination, and does not need to inject magnetic resonance contrast agents, thereby avoiding the risks of displacement, damage or dysfunction caused by metal implants in strong magnetic fields, significantly improving the accuracy and reliability of the prediction of the opening effect of the blood-brain barrier, and avoiding a long scanning process. Since ultrasound examination usually does not require patients to enter a closed space, it effectively avoids the discomfort of claustrophobic patients such as shortness of breath and increased heartbeat during the examination, and realizes the safe and rapid prediction of the opening effect of the blood-brain barrier, reduces the test cost, and saves scanning time. In addition, compared with the dye tracer method, the blood-brain barrier opening effect prediction method based on artificial intelligence does not require any operation on the subject, is convenient and fast, and has the ability to detect in vivo. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a schematic diagram of the flow chart of a method for predicting the effect of focused ultrasound on opening the blood-brain barrier in one embodiment of the present application.

[0019] Figure 2 Shown is a schematic diagram of the process of constructing a training data set in one embodiment of the present application.

[0020] Figure 3 Shown is a schematic diagram of the structure of a blood-brain barrier prediction model in one embodiment of the present application.

[0021] Figure 4 Shown is a schematic diagram of the structure of an acoustic coding module in one embodiment of the present application.

[0022] Figure 5 Shown is a schematic structural diagram of a one-dimensional residual layer in one embodiment of the present application.

[0023] Figure 6 Shown is a structural diagram of a feature aggregation module in one embodiment of the present application.

[0024] Figure 7 Shown is a schematic diagram of a process for predicting the opening effect of the blood-brain barrier in one embodiment of the present application.

[0025] Figure 8 Shown is a schematic block diagram of a system for predicting the effect of focused ultrasound on opening the blood-brain barrier in one embodiment of the present application.

[0026] Fig. 9 Shown is a schematic diagram of the structure of an electronic terminal in one embodiment of the present application. DETAILED DESCRIPTION

[0027] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0028] Before further describing the present invention in detail, the nouns and terms involved in the embodiments of the present invention are explained. The nouns and terms involved in the embodiments of the present invention are applicable to the following interpretations:

[0029] <1> T1: represents the time required for the tissue spin to return to equilibrium (i.e. longitudinal relaxation) in a magnetic field, also known as T1 relaxation time. This is a key parameter in MRI, used to describe the specific physical behavior of tissue in a magnetic field.

[0030] <2> Deep learning model: It is a machine learning model based on artificial neural networks (ANN). Its core idea is to learn the feature representation and pattern of data through a multi-level neural network structure.

[0031] <3> Transformer model: is a model based on the self-attention mechanism that can process all elements in the input sequence in parallel.

[0032] <4> MRI (Magnetic Resonance Imaging): Magnetic resonance imaging is a medical imaging technology that uses magnetic fields and radio frequency waves to generate detailed images of the inside of the body.

[0033] The current methods for evaluating the effect of blood-brain barrier opening include contrast-enhanced magnetic resonance imaging and dye tracing. Among them, enhanced T1-weighted magnetic resonance imaging (MRI), as the current clinical gold standard for evaluating the effect of blood-brain barrier regulation, has the following defects:

[0034] (1) High cost: MRI equipment is expensive, which limits its popularity in primary medical institutions.

[0035] (2) Long scanning time: The MRI scanning process takes a long time and is not suitable for scenarios that require rapid assessment.

[0036] (3) Contrast agent toxicity: MRI contrast agents such as gadolinium have certain toxicity and are difficult to be completely excreted from the body, posing potential health risks to patients.

[0037] (4) Limited scope of application: MRI examination is contraindicated for patients with metal implants such as pacemakers and those with claustrophobia.

[0038] The dye tracer method injects dyes (such as Evans blue) into animal veins, uses the property of dye molecules to pass through blood vessels and enter tissues when the permeability of the blood-brain barrier increases, and detects the distribution of dyes in target tissues through equipment such as spectrophotometers or fluorescence microscopes to evaluate the regulation effect of the blood-brain barrier. However, it has the following defects:

[0039] (1) The process is cumbersome and time-consuming: The dye tracer method requires the animal to be killed and the target tissue to be removed for testing. The whole process is cumbersome and time-consuming.

[0040] (2) Lack of in vivo evaluation capability: This method cannot evaluate the permeability changes of the blood-brain barrier in real time in vivo, which limits its potential in clinical applications.

[0041] (3) Sensitivity and specificity: Compared with MRI imaging technology, dye tracer method has certain shortcomings in sensitivity and specificity.

[0042] In order to solve the problems in the above-mentioned background technology, the present application provides a method, system, medium and terminal for predicting the effect of focused ultrasound on opening the blood-brain barrier, which are used to solve the technical problems in the prior art such as the need to inject magnetic resonance contrast agents that are toxic and difficult to excrete from the body, high testing costs, long scanning times, and the inability to evaluate the effect of focused ultrasound on opening the blood-brain barrier in real time in a living state.

[0043] To facilitate understanding of the embodiments of the present application, first Figure 1 Detailed description. Figure 1 The flowchart of the method for predicting the effect of focused ultrasound opening of the blood-brain barrier in the embodiment of the present application is shown. The method for predicting the effect of focused ultrasound opening of the blood-brain barrier in the present embodiment mainly includes the following steps:

[0044] S101: Collecting a number of acoustic signal data and a number of blood-brain barrier opening effect data of the ultrasound contrast agent in the process of regulating the blood-brain barrier.

[0045] In this embodiment, a suitable passive cavitation detection probe is selected to ensure that its center frequency matches the expected acoustic signal to be detected. The passive cavitation detection probe is placed in a suitable position, and the ultrasound contrast agent will emit an acoustic signal outward, and this acoustic signal is closely related to the oscillation state of the ultrasound contrast agent, so as to capture the acoustic signal data emitted by the cavitation effect generated by the ultrasound contrast agent, thereby using the passive cavitation detection probe to collect the acoustic signal data of the ultrasound contrast agent in the process of regulating the blood-brain barrier. The acoustic signal data contains information about the effect of ultrasound regulation, such as whether the barrier is open or not, or whether there is tissue damage. Specifically, the ultrasound contrast agent will produce a cavitation effect under the action of focused ultrasound, which manifests as periodic oscillations, i.e., steady-state cavitation. This process will produce mechanical force acting on the blood-brain barrier. This mechanical force plus mechanisms such as cell ion channel switches can achieve a temporary enhancement of the blood-brain barrier opening effect, and the blood-brain barrier permeability changes, effectively enhancing the delivery of drug molecules. The oscillation state of the ultrasound contrast agent determines the blood-brain barrier opening effect. Therefore, passive cavitation detection probes were used to collect acoustic signal data for the prediction of blood-brain barrier opening effects.

[0046] In this embodiment, ultrasound contrast agent regulation of blood-brain barrier technology mainly uses low-frequency focused ultrasound to penetrate the skull, form a focus in the brain, and induce microbubbles in intracranial blood vessels to produce cavitation effects such as oscillation, expansion, contraction, and implosion. These cavitation effects act on the blood-brain barrier, stretching and expanding its structure, thereby achieving local opening of the blood-brain barrier, which has the advantages of being non-invasive, locally controllable, and combinable with other imaging modalities.

[0047] S102: Preprocessing each of the collected acoustic signal data respectively, and constructing a training data set based on the blood-brain barrier opening effect data.

[0048] In this embodiment, if Figure 2 As shown, a flow chart of constructing a training data set in an embodiment of the present application is shown. The collected acoustic signal data are preprocessed respectively, and the method of constructing a training data set based on the blood-brain barrier opening effect data includes:

[0049] S1021: Perform Fourier transform preprocessing on each of the acoustic signal data to obtain a number of acoustic signal frequency domain feature data.

[0050] In this embodiment, Fourier transform preprocessing is a mathematical transformation method that can decompose a signal into a superposition of sine waves (or cosine waves) of different frequencies. In acoustic signal processing, Fourier transform can convert an acoustic signal that varies with time into a frequency spectrum that varies with frequency, thereby revealing the frequency components and distribution characteristics of the signal.

[0051] In this embodiment, after the acoustic signal data are preprocessed by Fourier transform, the acoustic signal data are converted from the time domain to the frequency domain and represented as a series of complex frequency components, each of which corresponds to a specific frequency component in the signal. In the frequency domain, a variety of useful feature data can be extracted, such as spectrum amplitude, spectrum phase, spectrum energy, spectrum entropy, etc. These feature data can reflect the characteristics of the acoustic signal such as frequency distribution, energy distribution and complexity, which is of great significance for the subsequent construction of the blood-brain barrier prediction model.

[0052] S1022: Normalizing the acoustic signal frequency domain feature data that meets a preset frequency range to obtain acoustic signal training data.

[0053] In this embodiment, the normalized data has a uniform scale, which can make the deep learning model converge to the optimal solution faster during the training process. This is because the numerical differences between different features are eliminated, and the model does not need to spend extra time to adapt to data of different dimensions. The normalized data is more compact in value, which can reduce numerical errors and rounding errors in the calculation process, thereby reducing computational complexity and increasing training speed.

[0054] In this embodiment, the preset frequency range can be the full frequency domain or a specific frequency band. By normalizing the frequency domain feature data of the acoustic signal that meets the preset frequency range, the importance of these key frequency components can be more prominent, which helps the model to better capture these key information and improve the accuracy of the prediction. In practical applications, acoustic signals often contain certain noise components. By normalizing the frequency domain feature data, the impact of noise on model training can be reduced to a certain extent, thereby improving the robustness of the model.

[0055] S1023: Match and annotate the acoustic signal training data with the blood-brain barrier opening effect data to construct a training data set; the training data set includes multiple groups of data pairs, each group of data pairs includes acoustic signal training data and blood-brain barrier opening effect data corresponding to the acoustic signal training data.

[0056] In this embodiment, a digital converter is used to sample the acoustic signal data collected by the passive cavitation detection probe for subsequent storage and processing. The acoustic signal training data is synchronized with the blood-brain barrier opening effect data in time. According to the duration of the ultrasonic pulse, the acoustic signal training data is divided into several groups, each corresponding to the acoustic signal data collected during an ultrasonic pulse. The duration of each group of acoustic signal training data should be consistent with the duration of the applied ultrasonic pulse, usually several milliseconds. A unique identifier (such as a timestamp, a sequence number, etc.) is assigned to each group of acoustic signal training data. Similarly, a unique identifier is also assigned to each blood-brain barrier opening effect data. Through the synchronous trigger signal or other marking means during the treatment process, the identifier of the acoustic signal training data is matched with the identifier of the corresponding blood-brain barrier opening effect data, and the matched acoustic signal training data and the blood-brain barrier opening effect data are sorted into multiple groups of data pairs, each group of data pairs including a section of acoustic signal training data and the corresponding blood-brain barrier opening effect data to construct a training data set.

[0057] S103: Using the training data set to train a deep learning model to construct a blood-brain barrier prediction model.

[0058] In this embodiment, if Figure 3 As shown, a schematic diagram of the structure of the blood-brain barrier prediction model in the embodiment of the present application is shown. The blood-brain barrier prediction model includes an acoustic coding module, a feature aggregation module, and a classification module; the acoustic coding module is used to encode the real-time acoustic signal data of the input ultrasound contrast agent in the process of regulating the blood-brain barrier to obtain a first feature vector; the feature aggregation module is used to perform feature fusion processing on the first feature vector to obtain a second feature vector; the classification module is used to output the blood-brain barrier opening effect prediction result data according to the second feature vector.

[0059] In this embodiment, the acoustic coding module includes a first fully connected layer and a first activation function layer. The acoustic coding module takes the real-time acoustic signal data of the ultrasound contrast agent in the process of regulating the blood-brain barrier as input and encodes it. The encoding process includes linear transformation through the first fully connected layer and introduction of nonlinear factors through the first activation function layer to obtain a first eigenvector. The first eigenvector is a high-level abstract representation of the input real-time acoustic signal data, which contains key information for subsequent processing.

[0060] In this embodiment, the first fully connected layer is responsible for mapping the input real-time acoustic signal data to a new feature space and extracting feature representations that are more useful for subsequent tasks. As needed, the first fully connected layer can adjust the dimension of the feature to adapt it to the input requirements of the subsequent network layer. In this way, through feature mapping and dimensionality transformation, the first fully connected layer can learn more abstract and robust feature representations, which helps to improve the model's generalization ability for unseen data. Appropriate dimensionality transformation can reduce the amount of computation in subsequent network layers and improve the overall efficiency of the model.

[0061] In this embodiment, the first activation function layer introduces nonlinear factors into the model so that the model can learn and express complex data relationships. The nonlinear activation function in the first activation function layer enables the model to capture nonlinear relationships in the data, thereby improving the prediction performance of the model.

[0062] In this embodiment, if Figure 4 As shown, a schematic diagram of the structure of the acoustic coding module in the embodiment of the present application is shown. The acoustic coding module also includes a plurality of one-dimensional residual layers and a flattening layer; after each one-dimensional residual layer is connected in sequence, the first fully connected layer and the first activation function layer are connected in sequence through the flattening layer.

[0063] In this embodiment, if Figure 5 As shown, a schematic diagram of the structure of a one-dimensional residual layer in an embodiment of the present application is shown. The one-dimensional residual layer includes a first convolutional layer, a second activation function layer, a second convolutional layer, a third activation function layer, and a maximum pooling layer; the first convolutional layer and the second convolutional layer are both one-dimensional convolutional layers. The input of the first convolutional layer is jump-connected to the output of the second convolutional layer. The flattening layer converts the one-dimensional feature map into a one-dimensional vector so that it can be compatible with the subsequent first fully connected layer. The maximum pooling layer slides on the input feature map through a pooling window and selects the maximum value in each window as the output, reducing the dimension of the feature map (usually in the spatial dimension, i.e., width and height), thereby reducing the amount of calculation and memory usage, preventing overfitting, accelerating the training process and improving the generalization ability of the model. The first fully connected layer and the first activation function layer further convert and activate the features to enhance the nonlinear representation ability of the model.

[0064] In this embodiment, the one-dimensional residual layer can gradually extract the deep features of the real-time acoustic signal data of the input ultrasound contrast agent in the process of regulating the blood-brain barrier by stacking multiple convolutional layers and activation function layers. These deep features are more abstract and advanced than the original signal, and are more useful for subsequent classification tasks. The jump connection allows the network to learn the residual between the input and output, rather than directly learning the entire mapping, which helps to alleviate the gradient vanishing or gradient exploding problems that are prone to occur in the deep network during training, so that the network can be trained deeper and more stable. The flattening layer is a bridge connecting the first fully connected layer and the one-dimensional residual layer, so that the model can flexibly process data of different shapes.

[0065] In this embodiment, through the combination of a one-dimensional residual layer, a flattening layer, a first fully connected layer, and a first activation function layer, the acoustic coding module can efficiently extract and integrate key features of the real-time acoustic signal data of ultrasound contrast agents in the process of regulating the blood-brain barrier. Since the acoustic coding module can extract richer and more abstract feature representations, it can provide more powerful support for subsequent feature aggregation modules and classification modules, thereby improving the prediction accuracy of the blood-brain barrier opening effect. The introduction of jump connections and activation functions makes the acoustic coding module more robust in the face of noise or abnormal data, and can better cope with challenges in practical applications.

[0066] In this embodiment, if Figure 6As shown, a schematic diagram of the structure of the feature aggregation module in an embodiment of the present application is shown. The feature aggregation module includes a first normalization layer and a multi-head self-attention layer. The main function of the feature aggregation module is to perform feature fusion processing on the first feature vector. The process of feature fusion processing includes operations such as adding category tags, layer normalization, constructing jump connections, and adding elements by elements with data processed by the multi-head self-attention layer. Specifically, first, a category tag is added to the first feature vector and the layer is normalized, and then a jump connection is constructed, and the data is added element by element with the data processed by the multi-head self-attention layer. The process is repeated several times to obtain the second feature vector. In this way, through the first normalization layer, the multi-head self-attention layer, the jump connection and other technical means, the module can effectively fuse the original features and the self-attention features to obtain a more comprehensive and rich second feature vector. The multi-head self-attention mechanism enables the model to process different positions in the input sequence in parallel, capture long-distance dependencies, and thus enhance the model's perception of contextual information. The introduction of layer normalization and skip connections (residual connections) improves the stability and training efficiency of the model, reduces the risk of overfitting, and helps the model learn better feature representations, so that the second feature vector is more comprehensive and richer than the first feature vector, containing more contextual information and high-level features about the input acoustic signal. Therefore, by optimizing the settings of the feature aggregation module, the performance of the entire blood-brain barrier prediction model can be further improved.

[0067] In this embodiment, the classification module includes a second normalization layer and a second fully connected layer. The classification module can receive a second feature vector containing rich context information and high-level features, extract the category label containing all pulse frequency domain information after aggregation, perform layer normalization processing, and map the features to the final classification space through the second fully connected layer, so as to make a more accurate and efficient prediction, and output the blood-brain barrier opening effect prediction result data.

[0068] It is worth noting that the blood-brain barrier prediction model realizes the comprehensive processing and analysis of the real-time acoustic signal data of ultrasound contrast agents in the process of regulating the blood-brain barrier through the organic combination of acoustic coding module, feature aggregation module and classification module, which significantly improves the accuracy and reliability of the prediction of the blood-brain barrier opening effect. It can achieve safe and rapid prediction of the blood-brain barrier opening effect without the need to inject magnetic resonance contrast agents and avoid long scanning processes, thus reducing the test cost and saving scanning time. In addition, compared with the dye tracer method, the blood-brain barrier opening effect prediction method based on artificial intelligence does not require any operation on the subject, is convenient and fast, and has the ability to detect in vivo.

[0069] Furthermore, each module in the blood-brain barrier prediction model can be independently optimized and adjusted to adapt to different data sets and task requirements. At the same time, the modular design also facilitates the introduction of new technologies and methods in the future to further improve the performance of the model. Since the blood-brain barrier prediction model has high efficiency and accuracy in processing real-time acoustic signals, it has good real-time and practicality and can be applied to clinical auxiliary diagnosis, drug development and other fields.

[0070] In this embodiment, the method of using the training data set to train a deep learning model to construct a blood-brain barrier prediction model includes:

[0071] (1) Training the deep learning model based on the training data set to obtain a primary blood-brain barrier prediction model, and generating primary blood-brain barrier opening effect prediction result data corresponding to the acoustic signal training data in each data pair according to the primary blood-brain barrier prediction model.

[0072] (2) Based on the primary blood-brain barrier opening effect prediction result data and the blood-brain barrier opening effect data corresponding to the acoustic signal training data in each data set, the deep learning model is tested and evaluated using cross-validation technology; and the parameters of the deep learning model are adjusted according to the test and evaluation results to construct a final converged blood-brain barrier prediction model.

[0073] In this embodiment, the deep learning model includes: a Transformer model.

[0074] The Transformer model captures dependencies in the sequence through a self-attention mechanism and performs better when dealing with long-distance dependencies. It has achieved remarkable success in multiple natural language processing tasks, including machine translation, text summarization, speech recognition, and audio processing. The Transformer model mainly consists of two parts: an encoder and a decoder. The encoder consists of multiple Transformer encoder layers, each of which contains a self-attention mechanism and a feedforward neural network. The self-attention mechanism allows the encoder to consider the contextual information of the entire sequence when processing data. The decoder also consists of multiple Transformer decoder layers, but each decoder layer also contains an additional encoder-decoder attention mechanism so that the decoder can pay attention to the output of the encoder.

[0075] In this embodiment, the training data set is used to train the deep learning model, and the model learns the matching relationship between the acoustic signal training data and the blood-brain barrier opening effect data to obtain a primary blood-brain barrier prediction model. The primary blood-brain barrier prediction model has a certain generation capability and can generate corresponding primary blood-brain barrier opening effect prediction result data based on the acoustic signal training data.

[0076] In this embodiment, the cross-validation technique is an effective method for evaluating the performance of a machine learning model. It divides the training data set into multiple subsets, such as a training subset and a test subset, repeatedly trains and tests the model, and uses the training subset to train the deep learning model in each iteration of cross-validation. The test subset is used to evaluate the performance of the model, a confusion matrix is ​​constructed, and evaluation indicators such as accuracy, precision, recall, and specificity are calculated to more stably and reliably evaluate the generalization ability of the model. The confusion matrix is ​​a commonly used tool for evaluating the performance of a classification model. It calculates a series of performance indicators by calculating the difference between the predicted results of the statistical model and the actual results.

[0077] It should be noted that accuracy refers to the proportion of samples judged as positive. Precision refers to the proportion of samples judged as positive that are actually positive. It measures the proportion of samples predicted as positive that are actually positive. Recall refers to the proportion of correctly judged positive samples to the total positive samples. It measures the proportion of samples predicted as positive that are actually positive. Specificity refers to the proportion of all negative samples that are correctly predicted as negative.

[0078] S104: Based on the blood-brain barrier prediction model, according to the acquired real-time acoustic signal data of the ultrasound contrast agent in the process of regulating the blood-brain barrier, predict the blood-brain barrier opening effect to output the blood-brain barrier opening effect prediction result data.

[0079] In this embodiment, if Figure 7 As shown, a schematic diagram of the process of predicting the blood-brain barrier opening effect in an embodiment of the present application is shown. Among them, based on the blood-brain barrier prediction model, according to the acquired real-time acoustic signal data of the ultrasound contrast agent in the process of regulating the blood-brain barrier, the blood-brain barrier opening effect is predicted, and the method of outputting the blood-brain barrier opening effect prediction result data includes:

[0080] S1041: Encode the real-time acoustic signal data of the input ultrasound contrast agent in the process of regulating the blood-brain barrier to obtain a first eigenvector.

[0081] S1042: Perform feature fusion processing on the first feature vector to obtain a second feature vector.

[0082] S1043: Outputting blood-brain barrier opening effect prediction result data according to the second feature vector.

[0083] The method for predicting the effect of focused ultrasound opening of the blood-brain barrier of the present application replaces the traditional MRI examination by comprehensively processing and analyzing the real-time acoustic signal data of the ultrasound contrast agent in the process of regulating the blood-brain barrier. It does not require the injection of magnetic resonance contrast agents, thereby avoiding the risks of displacement, damage or dysfunction caused by metal implants in strong magnetic fields, significantly improving the accuracy and reliability of the prediction of the effect of blood-brain barrier opening, and avoiding a long scanning process. Since ultrasound examinations usually do not require patients to enter closed spaces, it effectively avoids the discomfort of claustrophobic patients such as shortness of breath and increased heartbeat during examinations, and achieves safe and rapid prediction of the effect of blood-brain barrier opening, reducing test costs and saving scanning time. In addition, compared with the dye tracer method, the method for predicting the effect of blood-brain barrier opening based on artificial intelligence does not require any operation on the subject, is convenient and fast, and has the ability to detect in vivo.

[0084] In the embodiments of the present application, words such as "first" and "second" are used to distinguish the same or similar items with basically the same functions and effects. For example, the first fully connected layer and the second fully connected layer are only used to distinguish different fully connected layers, and their order is not limited. Those skilled in the art can understand that words such as "first" and "second" do not limit the quantity and execution order, and words such as "first" and "second" do not necessarily limit them to be different.

[0085] It should be noted that in the embodiments of the present application, words such as "exemplary" or "for example" represent examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0086] In the embodiments of the present application, "at least one" refers to one or more, and "plurality" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can represent: a, b, c, ab, ac, bc or abc, where a, b, c can be single or multiple.

[0087] Figure 8 is a schematic block diagram of a system for predicting the effect of focused ultrasound opening of the blood-brain barrier provided in an embodiment of the present application. Figure 8 As shown, the focused ultrasound blood-brain barrier opening effect prediction system 800 includes:

[0088] The data acquisition module 801 is used to collect a number of acoustic signal data and a number of blood-brain barrier opening effect data of the ultrasound contrast agent in the process of regulating the blood-brain barrier.

[0089] The blood-brain barrier prediction model construction module 802 is used to pre-process each of the collected acoustic signal data respectively, and to construct a training data set based on the blood-brain barrier opening effect data; and to train a deep learning model using the training data set to construct a blood-brain barrier prediction model.

[0090] The blood-brain barrier opening effect prediction module 803 is used to predict the blood-brain barrier opening effect based on the blood-brain barrier prediction model and the acquired real-time acoustic signal data of the ultrasound contrast agent in the process of regulating the blood-brain barrier, so as to output the blood-brain barrier opening effect prediction result data.

[0091] It should be understood that the specific process of each module executing the above corresponding steps has been described in detail in the above method embodiment, and for the sake of brevity, it will not be repeated here.

[0092] It should also be understood that the division of modules in the embodiments of the present application is schematic and is only a logical function division. There may be other division methods in actual implementation. In addition, each functional module in each embodiment of the present application may be integrated into a processor, or may exist physically separately, or two or more modules may be integrated into one module. The above-mentioned integrated modules may be implemented in the form of hardware or in the form of software functional modules.

[0093] Fig. 9 is a schematic block diagram of an electronic terminal provided in an embodiment of the present application. Fig. 9 As shown, the electronic terminal 900 includes: at least one processor 901, a memory 902, at least one network interface 903 and a user interface 905. The various components in the device are coupled together through a bus system 904. It can be understood that the bus system 904 is used to realize the connection and communication between these components. In addition to the data bus, the bus system 904 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, Fig. 9 In the specification, various buses are labeled as bus systems.

[0094] The user interface 905 may include a display, a keyboard, a mouse, a trackball, a click gun, keys, buttons, a touch pad or a touch screen.

[0095] It is understood that the memory 902 can be a volatile memory or a non-volatile memory, and can also include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as static random access memory (SRAM), synchronous static random access memory (SSRAM). The memory described in the embodiments of the present invention is intended to include but is not limited to these and any other suitable categories of memory.

[0096] The memory 902 in the embodiment of the present invention is used to store various categories of data to support the operation of the electronic terminal 900. Examples of these data include: any executable program for operating on the electronic terminal 900, such as an operating system 9021 and an application 9022; the operating system 9021 includes various system programs, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and processing hardware-based tasks. The application 9022 may include various applications, such as a media player (Media Player), a browser (Browser), etc., for implementing various application services. The method for predicting the effect of focused ultrasound opening of the blood-brain barrier provided in the embodiment of the present invention may be included in the application 9022.

[0097] The method disclosed in the above embodiment of the present invention can be applied to the processor 901, or implemented by the processor 901. The processor 901 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit in the processor 901 or the instruction in the form of software. The above processor 901 may be a general processor, a digital signal processor (DSP, Digital Signal Processor), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The processor 901 can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiment of the present invention. The general processor 901 may be a microprocessor or any conventional processor, etc. In combination with the steps of the accessory optimization method provided in the embodiment of the present invention, it can be directly embodied as a hardware decoding processor to execute, or it can be executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium, which is located in a memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0098] In an exemplary embodiment, the electronic terminal 900 may be implemented by one or more application specific integrated circuits (ASIC), DSP, programmable logic device (PLD), complex programmable logic device (CPLD) to execute the aforementioned method.

[0099] According to the method provided in the embodiment of the present application, the present application also provides a computer-readable storage medium, which stores a program code, and when the program code is run on a computer, the computer executes Figures 1 to 7 A method according to any one of the embodiments shown.

[0100] According to the method provided in the embodiment of the present application, the present application also provides a computer program product, which includes: a computer program code, when the computer program code is run on a computer, the computer executes Figure 7 to Figure 7 A method according to any one of the embodiments shown.

[0101] The terms "component", "module", "system", etc. used in this specification are used to represent computer-related entities, hardware, firmware, a combination of hardware and software, software, or software in execution. For example, a component can be, but is not limited to, a process running on a processor, a processor, an object, an executable file, an execution thread, a program and / or a computer. By way of illustration, both applications running on a computing device and a computing device can be components. One or more components may reside in a process and / or an execution thread, and a component may be located on a computer and / or distributed between two or more computers. In addition, these components may be executed from various computer-readable media having various data structures stored thereon. Components may, for example, communicate through local and / or remote processes according to signals having one or more data packets (e.g., data from two components interacting with another component between a local system, a distributed system and / or a network, such as the Internet interacting with other systems through signals).

[0102] Those of ordinary skill in the art will appreciate that the various illustrative logical blocks and steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0103] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0104] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0105] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0106] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0107] In the above embodiments, the functions of each functional unit can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions (programs). When loading and executing computer program instructions (programs) on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. Computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, computer instructions can be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (digital subscriber line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media integrations. Available media may be magnetic media (e.g., floppy disks, hard disks, tapes), optical media (e.g., high-density digital video discs (DVDs), or semiconductor media (e.g., solid state disks (SSDs), etc.).

[0108] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

[0109] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0110] In summary, the present application provides a method, system, medium and terminal for predicting the effect of opening the blood-brain barrier by focusing ultrasound. By comprehensively processing and analyzing the real-time acoustic signal data of the ultrasound contrast agent in the process of regulating the blood-brain barrier, the traditional MRI examination is replaced, and there is no need to inject the magnetic resonance contrast agent, thereby avoiding the risks of displacement, damage or dysfunction caused by metal implants in a strong magnetic field, and significantly improving the accuracy and reliability of the prediction of the opening effect of the blood-brain barrier, avoiding the long scanning process, and because the ultrasound examination usually does not require the patient to enter a closed space, it effectively avoids the discomfort of claustrophobic patients such as shortness of breath and increased heartbeat during the examination, and realizes the safe and rapid prediction of the opening effect of the blood-brain barrier, reduces the test cost, and saves scanning time. In addition, compared with the dye tracer method, the blood-brain barrier opening effect prediction method based on artificial intelligence does not require any operation on the subject, is convenient and fast, and has the ability to detect live bodies. Therefore, the present application effectively overcomes the various shortcomings in the prior art and has a high industrial utilization value.

[0111] The above embodiments are merely illustrative of the principles and effects of the present application and are not intended to limit the present application. Anyone familiar with the technology may modify or change the above embodiments without violating the spirit and scope of the present application. Therefore, all equivalent modifications or changes made by a person of ordinary skill in the art without departing from the spirit and technical ideas disclosed in the present application shall still be covered by the claims of the present application.

Claims

1. A method for predicting the effect of focused ultrasound on opening the blood-brain barrier, characterized in that: include: Collect some acoustic signal data of ultrasound contrast agents in the process of regulating the blood-brain barrier and some blood-brain barrier opening effect data; Preprocessing each of the collected acoustic signal data respectively, and constructing a training data set based on the blood-brain barrier opening effect data; Using the training data set to train a deep learning model to build a blood-brain barrier prediction model; Based on the blood-brain barrier prediction model, the blood-brain barrier opening effect is predicted according to the acquired real-time acoustic signal data of the ultrasound contrast agent in the process of regulating the blood-brain barrier, so as to output the blood-brain barrier opening effect prediction result data.

2. The method for predicting the effect of focused ultrasound on opening the blood-brain barrier according to claim 1, characterized in that: The method of preprocessing the collected acoustic signal data respectively and constructing a training data set based on the blood-brain barrier opening effect data includes: Performing Fourier transform preprocessing on each of the acoustic signal data to obtain a number of acoustic signal frequency domain feature data; Normalizing the acoustic signal frequency domain feature data within a preset frequency range to obtain acoustic signal training data; The acoustic signal training data and the blood-brain barrier opening effect data are matched and annotated in correspondence to construct a training data set; the training data set includes multiple groups of data pairs, each group of data pairs includes the acoustic signal training data and the blood-brain barrier opening effect data corresponding to the acoustic signal training data.

3. The method for predicting the effect of focused ultrasound on opening the blood-brain barrier according to claim 1, characterized in that: The blood-brain barrier prediction model includes an acoustic coding module, a feature aggregation module, and a classification module; the acoustic coding module is used to encode the real-time acoustic signal data of the input ultrasound contrast agent in the process of regulating the blood-brain barrier to obtain a first feature vector; The feature aggregation module is used to perform feature fusion processing on the first feature vector to obtain a second feature vector; The classification module is used to output blood-brain barrier opening effect prediction result data according to the second feature vector.

4. The method for predicting the effect of focused ultrasound on opening the blood-brain barrier according to claim 2, characterized in that: Methods of using the training data set to train a deep learning model to construct a blood-brain barrier prediction model include: Training the deep learning model based on the training data set to obtain a primary blood-brain barrier prediction model, and generating primary blood-brain barrier opening effect prediction result data corresponding to the acoustic signal training data in each data pair according to the primary blood-brain barrier prediction model; According to the primary blood-brain barrier opening effect prediction result data and the blood-brain barrier opening effect data corresponding to the acoustic signal training data in each data pair, the deep learning model is tested and evaluated using cross-validation technology; according to the test and evaluation results, the parameters of the deep learning model are adjusted to construct a final converged blood-brain barrier prediction model.

5. The method for predicting the effect of focused ultrasound on opening the blood-brain barrier according to claim 3, characterized in that: The acoustic coding module includes a first fully connected layer and a first activation function layer.

6. The method for predicting the effect of focused ultrasound on opening the blood-brain barrier according to claim 3, characterized in that: The feature aggregation module includes a first normalization layer and a multi-head self-attention layer.

7. The method for predicting the effect of focused ultrasound on opening the blood-brain barrier according to claim 5, characterized in that: The acoustic coding module also includes a plurality of one-dimensional residual layers and flattening layers; after each one-dimensional residual layer is connected in sequence, the first fully connected layer and the first activation function layer are connected in sequence through the flattening layer.

8. A focused ultrasound blood-brain barrier opening effect prediction system, characterized in that: include: A data acquisition module, used to collect a number of acoustic signal data and a number of blood-brain barrier opening effect data of the ultrasound contrast agent in the process of regulating the blood-brain barrier; A blood-brain barrier prediction model building module, used to pre-process the collected acoustic signal data respectively, and to build a training data set based on the blood-brain barrier opening effect data; Using the training data set to train a deep learning model to build a blood-brain barrier prediction model; The blood-brain barrier opening effect prediction module is used to predict the blood-brain barrier opening effect based on the blood-brain barrier prediction model and the acquired real-time acoustic signal data of the ultrasound contrast agent in the process of regulating the blood-brain barrier, so as to output the blood-brain barrier opening effect prediction result data.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting the effect of focused ultrasound on opening the blood-brain barrier according to any one of claims 1 to 7 is implemented.

10. An electronic terminal comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the method for predicting the effect of focused ultrasound on opening the blood-brain barrier according to any one of claims 1 to 7.

Images