Brain disease diagnosis model based on multi-level causal intervention network

Abstract

The present invention discloses a brain disease diagnosis model based on a multi-level causal intervention network, which relates to the field of brain functional network analysis and medical assistance research. The method specifically includes: a brain network pre-construction stage, data pre-processing and extraction of fMRI time series, using tangent Pearson embedding to construct the input features required for the model; using a 3-layer Transformer network with a self-attention mechanism to realize feature extraction of the brain network; using an HCI module to fuse multiple layers of attention and accordingly divide the brain network features into causal and non-causal parts; applying DCI to impose first-order causal constraints on the model to reduce the interference of non-causal factors, while introducing ICAC into the causal part and imposing second-order causal constraints on the model to further optimize the model; using an orthogonal clustering OCR readout function to realize the classification of brain diseases. The invention can effectively filter out non-causal interference factors and extract high-order functional connectivity information to improve the accuracy and robustness of brain disease analysis.

Description

Technical Field

[0001] The present invention relates to the field of brain function network analysis and medical assistance research, and in particular to a brain disease diagnosis model based on a multi-level causal intervention network with hierarchical causal perception capabilities. Background Art

[0002] In recent years, functional connectivity networks (FCNs) based on resting-state cerebral functional magnetic resonance imaging (rs-fMRI) have played a crucial role in the detection and analysis of neurological disorders. They measure the dependence of functional activity between any two brain regions. Abnormal changes in functional connectivity are used as a basis for diagnosing neurological disorders, helping to understand the underlying mechanisms of neurological diseases and improve diagnostic and treatment strategies. Therefore, functional connectivity derived from rs-fMRI data has become a powerful method for quantifying and mapping brain function. Building on this foundation, the introduction of advanced methods in machine learning and computational neuroscience has opened up new avenues for analyzing complex neuroimaging data and developing predictive models for ASD detection.

[0003] Analyses have shown that functional connectivity networks can serve as effective biomarkers for identifying patients with brain disorders. However, traditional methods typically capture only direct pairwise relationships between brain regions, thereby overlooking a wealth of higher-order information. Furthermore, fMRI data contain signals directly related to the subject's physiological information but are also contaminated by various external factors unrelated to the subject's physiological state, leading to the incorporation of numerous non-causal factors, which introduce noise and potentially mislead research findings. Furthermore, these non-causal factors are often heterogeneous across sites, a primary cause of domain shift in multi-site datasets.

[0004] To address these issues, we propose a brain disease diagnosis method model based on a multi-level causal intervention network. This method model integrates high-order FC and brain transformers to continuously capture more comprehensive high-order features in the data. Simultaneously, by utilizing a hierarchical causal-aware integrator as an observation lens, a series of causal interventions are implemented to guide the model to learn truly invariant causal features, achieving superior performance and excellent inter-site versatility. Extensive experiments on real-world datasets for autism spectrum disorder (ASD) and major depressive disorder (MDD) demonstrate that this method outperforms current state-of-the-art methods, demonstrating its significant clinical value. Summary of the Invention

[0005] The purpose of this invention is to propose a brain disease diagnosis model based on a multi-level causal intervention network to solve the problem that the existing technology can only capture the direct pair-wise relationship between brain regions in functional connectivity network analysis, ignoring high-order information, resulting in insufficient accuracy in brain disease identification. At the same time, fMRI data contains external factors unrelated to the subject's physiological state, mixed with non-causal interference signals, generating noise and potentially misleading research results.

[0006] To achieve the above objectives, the present invention proposes a brain disease diagnosis model based on a multi-level causal intervention network, which includes the following steps:

[0007] S1, the brain network model pre-construction stage, first pre-processing the data and extracting the fMRI time series, using the Tangent Pearson Embedding (TPE) to construct the input features required for the model;

[0008] S2, feature extraction stage, uses a 3-layer Transformer network with self-attention mechanism to extract features of the brain network;

[0009] S3, the causal brain connection division stage, uses the HCI module to fuse multiple layers of attention and divide the brain network features into causal and non-causal parts accordingly;

[0010] S4, causal constraint stage, applying DCI to impose first-order causal constraints on the model, and applying ICAC to impose second-order causal constraints on the model;

[0011] S5. Classification and recognition stage: orthogonal clustering OCR readout function is used to classify brain diseases.

[0012] Preferably, in S1, the preprocessed data is divided into 200 brain regions of interest using the CC200 template to construct a brain network matrix.

[0013] Preferably, the connection patterns between different brain regions are different. Tangent Pearson Embedding (TPE) is used to extract the characteristics of the brain network matrix. Suppose the fMRI time series input of the subject is Where t represents the time point, the Tangent Pearson Embedding (TPE) is calculated as follows:

[0014]

[0015] Among them, x i , x j , x k is the fMRI time series element, i, j, k are integer indexes, r represents the number of brain ROIs, Cov(x i ,x k ) is x i and x kThe covariance between them, σ represents the standard deviation, P ij represents the covariance between the i-th row and the j-th column of the Pearson correlation matrix, M is the number of subjects, log m is the matrix logarithm, P is the Pearson correlation matrix, P i is the element of the i-th row of the Pearson correlation matrix, and TPE is the tangent Pearson embedding matrix.

[0016] Preferably, in S2, a Transformer network is constructed for feature extraction. The data fit must be re-evaluated and the attention mechanism must be adjusted to correspond to the unique characteristics of the brain network data. The connection profile tp i As the initial node features to establish the core network, where tp i is the row corresponding to node i in the tangent Pearson embedding matrix TPE.

[0017] Preferably, the self-attention mechanism is an important module of Transformer. The traditional self-attention module is defined as:

[0018]

[0019] Among them, <.,.> represents the inner product, the input Z of the first layer 0 This is the TPE mentioned above, l represents the number of layers, Z l and Z l-1 Represent the network characteristics of layer l and layer l-1 respectively, is a learnable parameter, represent The size of the first dimension;

[0020] Use a multi-head attention mechanism with H heads working simultaneously, and combine their outputs through connections to form the overall output of the multi-head attention module, where H is an integer;

[0021] Z l-1 The i-th and j-th rows of and attention For projection and The linear kernel between the two uses the RBF nonlinear kernel. The linear kernel is improved, and the formula is as follows:

[0022]

[0023] Preferably, in S3, in the causal brain connection division stage, the necessity of attention fusion lies in that the attention matrices at different levels capture multi-scale features from local to global, and the global attention representation is obtained through the HCI method, which is formulated as follows:

[0024]

[0025] A f =α1A1+α2A2+α3A3;

[0026] Among them, ⊙ represents the element product, α1, α2, and α3 are the fusion weights of A1, A2, and A3 respectively. is the attention matrix of the first layer of the i-th head, is the attention matrix of the second layer of the i-th head, is the attention matrix of the third layer of the i-th head, A f is the fusion attention matrix.

[0027] Preferably, in order to separate the causal and non-causal parts, keep A f The first γ percent of the largest value in the γ region is set to zero, which is expressed as The remaining areas are non-causal focal points, represented by γ represents the proportion of causal features, at this time, Focusing only on non-causal regions, Focus on the causal area;

[0028] Execute the forward propagation process of Transformer, respectively and Substitute the attention matrix part in the forward propagation to obtain the candidate causal part Z C and the non-causal part Z N .

[0029] Preferably, in S4, DCI is applied to impose first-order causal constraints on the model, and feature Z is input into the readout function to obtain the output feature make and Represent the candidate causal prediction for the mth sample and the non-causal prediction for the nth sample, respectively. Use different non-causal predictions as interventions and combine them with the causal prediction to form a composite prediction. The formula is as follows:

[0030]

[0031] Where sigm represents the Sigmoid function, For composite prediction;

[0032] Then, by intervening in the loss function The first-order causal constraints are imposed on the model, as follows:

[0033]

[0034] in, is the cross entropy loss function, E n [.] means taking the expectation of n, Var n [.] means taking the variance of n, y m is the category label of sample m;

[0035] DCI can effectively alleviate the spurious correlations caused by sampling bias. In the case that the non-causal subgraph does not contain spurious correlations, this combination is unbiased, thereby increasing the possibility of learning pure causal correlations.

[0036] Preferably, in S4, ICAC is applied to impose second-order causal constraints on the model, and Z C Expand it into a column vector. Let the column vector have K elements, each element represents a neuron, where the kth neuron is z k , neuron z k To Z C The causal attribution of is defined as the value z', and the average causal effect of z' is The calculation formula is as follows:

[0037] E[Z c |do(z -k = z′)] = ∫ Z C ·p(Z c |do(z -k =z'))dZ c

[0038] ξ=∫(∫Z C ·p(Z C |do(z=z′))dZ C )dz′

[0039]

[0040] Among them, z represents the entire column vector, do(.) is the intervention method, p(.) represents the process of deep network calculation probability, z -k represents z divided by z k For all values ​​other than , ξ is the baseline value and E[.|.] is the conditional expectation;

[0041] The intervention variable z′ satisfies the uniform distribution within the range of z, that is, for the i-th element z of z′ ′i , ξ is the mean distribution of z, representing its baseline value; the do(.) intervention method only retains z each time k The value of z k The distribution of all values ​​other than ξ is eliminated by subtracting ξ, and then by traversing z k , repeat the above method to quantify the true causal value of each variable to the model;

[0042] The causal influence of all features on the i-th sample label is quantified by the invariant causal attribution ICA vector, which is the feature neuron z k The causal attribution of provides a metric, and the ICA vector of the k-th feature pair label y is defined as follows:

[0043]

[0044] By analyzing all K dimensions, the ICA vector f of the i-th sample is generated i , the formula is as follows:

[0045] f i =[(f i 1 ) T ,(f i 2 ) T ,…,(f i k ) T ,…,(f i K ) T ] T ;

[0046] In order to evaluate the difference between the ICA vectors of the input features on the output y across domains, a contrast loss function is proposed In order to match the input ICA vector with the output of all instances of the same class across domains, the loss is optimized by minimizing the distance between inputs of the same class and maximizing the distance between inputs of different classes. The inputs of the same class are regarded as positive matches of the i-th sample, and the inputs of different classes are regarded as negative samples. The contrast loss formula is as follows:

[0047]

[0048] Among them, f i 、f j and f k are the ICA vectors of samples i, j, and k, respectively, y is the label, and τ is the temperature parameter for scaling cosine similarity.

[0049] This loss function encourages the embeddings of samples with the same label to be close, while pushing the embeddings of samples with different labels apart; by carefully designing this contrastive loss and combining it with the main loss function, we ensure that the model learns to effectively maintain inter-class similarities, thereby enhancing its ability to capture potential causal relationships.

[0050] Preferably, in S5, after intervention and treatment, Z CFor the causal features most relevant to the disease, the orthogonal clustering readout function (OCR) is used to extract these features for final classification;

[0051] Set O cluster centers, each with V dimensions, and use the Softmax projection operator to calculate the probability P of assigning node i to cluster O io , the formula is as follows:

[0052]

[0053] in, Represents Z C The i-th row, u o is an orthogonal basis obtained by uniformly initializing O cluster centers using Xavier and using the Gram-Schmidt method;

[0054] The final embedding Z G is obtained through aggregate learning, and the causal feature Z C The calculated distribution ρ is calculated as follows:

[0055] Z G =ρ T Z C ;

[0056] Z G Bring it into MLP to obtain the final classification result and use the cross entropy function Constraints are performed; the readout module of the non-causal branch also uses the OCR function and passes the cross entropy loss function Conduct individual training;

[0057] Model training involves four losses, and the total loss for disease classification is expressed as:

[0058]

[0059] Among them, λ1, λ2 and λ3 are hyperparameters for balancing contributions, and ε represents the current number of training times.

[0060] In the experiment, we found that since the model cannot distinguish non-causal features well in the early stage of training, if the constraints are too strong, the feature extraction of the model will become more chaotic; on the contrary, in the later stage of training, the constraints should be strengthened to obtain better results. Based on this observation, we and Configured weights separately and This weight gradually increases as training progresses, and the three loss functions are combined according to the designed weights to achieve better overall performance.

[0061] Therefore, the present invention proposes a brain disease diagnosis model based on a multi-level causal intervention network, which has the following beneficial effects:

[0062] (1) The present invention can effectively filter out non-causal interference factors and extract high-order functional connectivity information to improve the accuracy and robustness of brain disease analysis.

[0063] (2) The present invention integrates high-order FC and brain transformers to continuously capture more comprehensive high-order features in the data, and guides the model to learn truly invariant causal features by utilizing a hierarchical causal-aware integrator as an observation lens.

[0064] (3) The present invention has both excellent performance and excellent inter-site versatility, outperforming the current most advanced method models and demonstrating its important clinical value.

[0065] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 is a flow chart of the brain disease diagnosis model based on a multi-level causal intervention network of the present invention;

[0067] Figure 2 This is the result of the biomarker visualization of the brain disease diagnostic model based on the multi-level causal intervention network of the present invention; wherein, (a) represents the 5 most important brain connections and their influence weights, (b) represents the connection strength of the 5 most important brain connections of ASD patients, and (c) represents the connection strength of the 5 most important brain connections of normal people. DETAILED DESCRIPTION

[0068] To make the technical solutions, advantages, and purposes of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below. The described embodiments are part of the embodiments of the present invention, not all of them. Based on the described embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0069] Unless otherwise defined, technical or scientific terms used in the present invention shall have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs.

[0070] like Figure 1As shown in the figure, functional connectivity networks can be used as effective biomarkers for identifying brain diseases, but traditional methods ignore high-order information, and fMRI data are susceptible to noise and inter-site domain offset. To this end, the present invention proposes a brain disease diagnosis model based on a multi-level causal intervention network. The model integrates high-order FC and brain converters to capture more comprehensive features, and performs causal intervention through hierarchical causal-aware integrators to help the model learn truly invariant causal features. Experimental results on autism spectrum disorder (ASD) and major depressive disorder (MDD) datasets are superior to existing method models and have potential for clinical application. The specific processing process includes the following steps:

[0071] (1) In the brain network pre-construction stage, data preprocessing is performed to extract fMRI time series, and Tangent Pearson Embedding (TPE) is used to construct the input features required for the model.

[0072] (2) In the feature extraction stage, a three-layer Transformer network with self-attention mechanism is used to extract features of the brain network.

[0073] (3) In the causal brain connection division stage, the HCI module is used to fuse multiple layers of attention and divide the brain network features into causal and non-causal parts accordingly.

[0074] (4) In the causal constraint stage, DCI is applied to impose first-order causal constraints on the model to reduce the interference of non-causal factors; at the same time, ICAC is introduced into the causal part to impose second-order causal constraints on the model to further optimize the model.

[0075] (5) In the classification and recognition stage, the orthogonal clustering (OCR) readout function is used to classify brain diseases.

[0076] Among them, steps (2) to (5) are an end-to-end deep network. The deep network model needs to be trained in advance, and then the brain network matrix extracted in step (1) is input into it to complete the entire process of steps (2) to (5). Therefore, the actual operation process mainly requires the following three steps:

[0077] 1. Brain network pre-construction: extract fMRI time series through data preprocessing, and use Tangent Pearson Embedding (TPE) to construct the input features required by the model.

[0078] 2. Model training: Use the training dataset to train subsequent models until the expected standards are met.

[0079] 3. Classification and recognition: Input the data to be diagnosed into the trained model to achieve classification and recognition of the corresponding brain disease.

[0080] The following three practical operation steps of the present invention are respectively described in conjunction with a specific example using the ABIDE data set:

[0081] 1) Brain network preconstruction: The ABIDE-I database includes 573 NC and 539 ASD subjects. Preprocessed rs-fMRI data are publicly accessible through the Preprocessed Connectome Project (PCP) and were preprocessed using the Configurable Connectome Analysis Pipeline (CPAC). After removing invalid samples with missing data, 1035 subjects were selected, including 530 NC and 505 ASD subjects. The brain regions were divided into 200 regions of interest (ROIs) using the Consensus Clustering 200 (CC200) template. After the above processing, the data of each sample was processed into a 200×200 matrix. The TPE algorithm was then applied to all data to obtain the required brain network matrix, where the dimension of the brain network matrix for each sample was still 200×200.

[0082] 2) Model training: The data was divided into training and test sets using 10-fold cross-validation and leave-one-out-site-out (LEAVE-ONE-OUT) strategies, respectively. Training was performed in Python using the PyTorch framework. The learning rate was set to 0.00001, and the four hyperparameters were set to γ ​​= 0.8, λ1 = 0.1, λ2 = 0.1, and λ3 = 1.4. Training was performed 100 times, and the trained model weights were saved for future use.

[0083] 3) Classification and recognition: Input the data to be diagnosed into the trained model to achieve classification and recognition of the corresponding brain disease. The effectiveness of the model can be evaluated in the following two aspects:

[0084] 1. Accuracy Evaluation

[0085] The following metrics were used to evaluate the performance of binary classification of patients with or without abnormalities: accuracy (Acc), area under the receiver operating characteristic curve (AUC), recall (Rec), specificity (Spec), F1 score (F1), and precision (Pre). The model's results using ten-fold cross-validation and leave-one-out-site validation on the ABIDE dataset, as well as comparisons with existing state-of-the-art methods, are shown in Tables 1 and 2, respectively. These results demonstrate that the proposed model demonstrates excellent disease diagnosis performance.

[0086] Table 1 Comparison results of ten-fold cross validation and existing best methods

[0087] ABIDE 10% off acc auc recall spec F1 prec RF 63.665 63.461 54.848 72.074 63.212 65.184 Support Vector Machine 66.668 66.593 63.373 69.811 66.562 66.918 MVS 67.534 67.494 66.117 68.866 67.238 67.304 BolT 68.35 74.94 63.75 72.68 70.19 68.14 FBNETGEN 67.201 67.024 60.989 73.059 66.598 64.588 bnt 71.457 71.115 66.442 75.786 70.887 73.397 TP-MIDA 70.5237 77.21968 67.707 73.206 70.441 70.714 MBT 74.781 74.051 70.9 77.202 74.107 75.707 Proposed 75.449 74.49 71.599 77.382 74.644 76.572

[0088] Table 2 Comparison results between the leave-one-site-out method and the existing best method

[0089] ABIDE leave one acc auc recall spec F1 prec RF 60.659 61.128 54.723 67.604 59.158 64.9905 Support Vector Machine 64.477 65.027 62.368 67.689 63.2845 66.2725 MVS 70.0085 69.61 66.206 73.013 68.9335 70.9995 BolT 67.8655 74.7955 59.165 75.46 69.616 65.37 FBNETGEN 70.21684 69.48368 68.39895 70.56684 68.8279 69.0895 bnt 71.6758 70.8295 70.5475 71.11 70.4163 71.463 TP-MIDA 69.2754 74.445 65.282 74.4545 69.401 70.846 MBT 74.806 74.663 69.3835 79.9425 73.819 78.879 Proposed 75.955 75.399 71.6125 79.184 75.2845 77.8785

[0090] 2. Biomarker Visualization

[0091] The causal multi-level fusion attention matrix calculated by the HCI module in the model reflects the importance of the connections between each brain region and can be used to extract the most discriminative brain region connections for disease diagnosis. These brain region connections can be used as a powerful tool for detecting ASD diseases. The average of all samples was taken, and the five largest connections were extracted as the five most important brain connections for ASD classification. The BrainNetViewer toolbox was used to visualize them, as shown in the following example: Figure 2 If you want to further compare the differences in brain connectivity between normal people and patients, you can map the extracted important brain connections onto the TPE brain connectivity matrix of normal people and the TPE brain connectivity matrix of ASD patients, take the average of the two types of patients, and use the BrainNet Viewer toolbox to visualize it as shown below: Figure 2 (b) and Figure 2 As shown in (c), by comparison, it can be found that normal people and ASD patients show obvious differences in these 5 most discriminative brain connections.

[0092] Therefore, the brain disease diagnosis model based on a multi-level causal intervention network proposed in the present invention can effectively filter out non-causal interference factors and extract high-order functional connectivity information to improve the accuracy and robustness of brain disease analysis; at the same time, by using a hierarchical causal-aware integrator as an observation lens, it guides the model to learn truly invariant causal features, demonstrating its important clinical value.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the same. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solutions of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A brain disease diagnosis model based on a multi-level causal intervention network, including the following steps: S1, brain network model pre-construction stage, first pre-process the data and extract the fMRI time series, using tangent Pearson embedding TPE The input features required to build the model; S2, feature extraction stage, uses a 3-layer Transformer network with self-attention mechanism to extract features of the brain network; S3, the causal brain connection division stage, uses a hierarchical causal-aware integrator to fuse multiple layers of attention and divide brain network features into causal and non-causal parts accordingly; S4, causal constraint stage, applying DCI to impose first-order causal constraints on the model, and applying ICAC to impose second-order causal constraints on the model; S5, classification and recognition stage, using orthogonal clustering to read out the OCR readout function to achieve the classification of brain diseases; In S3, at the causal brain connection partitioning stage, the necessity of attention fusion lies in the fact that the attention matrices at different levels capture multi-scale features from local to global, respectively, and the global attention representation is obtained through the hierarchical causal perception integrator. The formula is as follows: ； ； in, represents element-wise product, They are The fusion weight of For the The attention matrix of the first layer of the head, For the The second-layer attention matrix of the head, For the The third layer of attention matrix, To fuse the attention matrix, M is the number of subjects; To separate the causal and non-causal parts, keep The top percent with the largest median value area, and set the rest to zero, expressed as , the rest of the area is non-causal focal point, expressed as , represents the proportion of causal features, at this time, Focusing only on non-causal regions, Focus on the causal area; Execute the forward propagation process of Transformer, respectively and Substitute the attention matrix part in the forward propagation to obtain the candidate causal part and the non-causal part .

2. The brain disease diagnosis model based on a multi-level causal intervention network according to claim 1, wherein in S1, the preprocessed data is divided into 200 brain regions of interest using the CC200 template to construct a brain network matrix.

3. The brain disease diagnosis model based on the multi-level causal intervention network according to claim 2, wherein the connection patterns between different brain regions are different, and the tangent Pearson embedding is used. TPE Extract the features of the brain network matrix, the tangent Pearson embedding TPE The calculation formula is as follows: ； ； in, , , is the fMRI time series element, is an integer index, Represents the number of brain region ROIs, for and The covariance between represents the standard deviation, Represents the Pearson correlation matrix Row and The covariance between columns, is the number of subjects, is the matrix logarithm, is the Pearson correlation matrix, The Pearson correlation matrix i Elements of the row, is the tangent Pearson embedding matrix.

4. According to the brain disease diagnosis model based on the multi-level causal intervention network of claim 1, in S2, the Transformer network is constructed for feature extraction, the data fit must be re-evaluated and the attention mechanism must be adjusted to correspond to the unique characteristics of the brain network data, using the connection profile As the initial node features to establish the core network, where, is the tangent Pearson embedding matrix TPE midpoint i The corresponding row.

5. In the brain disease diagnosis model based on a multi-level causal intervention network according to claim 4, the self-attention mechanism is an important module of the Transformer. The traditional self-attention module is defined as: ； in, represents the inner product, Represents the number of layers, and Representing the Layer and The network characteristics of the layer, is a learnable parameter, represent The size of the first dimension; Use include H The multi-head attention mechanism works simultaneously, and its output is combined through connections to form the overall output of the multi-head attention module, where H is an integer; Will No. Row and Rows are represented as and ,attention For projection and The linear kernel between the two uses the RBF nonlinear kernel. The linear kernel is improved, and the formula is as follows: 。 6. According to the brain disease diagnosis model based on the multi-level causal intervention network of claim 1, in S4, DCI is applied to impose first-order causal constraints on the model, and the feature Z Input into the readout function to get the output features ;make Representing the The candidate causal predictions of the samples and the We use different non-causal predictions as interventions and combine them with the causal predictions to form a composite prediction, which is formulated as follows: ； in represents the Sigmoid function, For composite prediction; Then, by intervening in the loss function The first-order causal constraints are imposed on the model, as follows: in, is the cross entropy loss function, Express n Take expectations, Express Take the variance, For samples The category label of .

7. According to the brain disease diagnosis model based on the multi-level causal intervention network of claim 1, in S4, ICAC is applied to impose second-order causal constraints on the model, Expand it into a column vector, let the column vector have elements, each element represents a neuron, where The neurons are , neurons right The causal attribution is defined as the value of , The average causal effect is , which is calculated as follows: in, represents the entire column vector, For intervention methods, (.) is the process of calculating the probability of the deep network, represent Medium All values ​​except is the baseline value, is the conditional expectation; Intervention variables In the range of z, the distribution is uniform, that is, No. Elements , , is the mean distribution of z, representing its baseline value; Intervention methods only retain The value of The distribution of all values ​​other than Eliminate it, and then traverse , repeat the above method to quantify the true causal value of each variable to the model; Quantify the invariant causal attribution ICA vector for all features The causal influence of sample labels is the feature neuron provides a measure of causal attribution, the first feature pair labels The ICA vector definition formula is as follows: ； Through all Dimensional analysis generates the ICA vector of samples , the formula is as follows: ； To evaluate the input features at the output The difference between the ICA vectors on the cross-domain is analyzed, and the contrast loss function is proposed. ; In order to match the input ICA vector with the output of all instances of the same class across domains, the loss is optimized by minimizing the distance between inputs of the same class and maximizing the distance between inputs of different classes, and the inputs of the same class are regarded as Positive matching of samples, treating inputs of different classes as negative samples, the contrast loss formula is as follows: in, 、 and The samples are and The ICA vector, For labels, is the temperature parameter for scaling the cosine similarity.

8. The brain disease diagnosis model based on a multi-level causal intervention network according to claim 1, wherein in S5, after intervention and treatment, For the causal features most relevant to the disease, the orthogonal clustering readout function (OCR) is used to extract these features for final classification; set up O cluster centers, each with V dimensions, and uses the Softmax projection operator to calculate the assignment of node i to cluster O Probability , the formula is as follows: ； in, express No. OK, By using Xavier uniform initialization Cluster centers and orthogonal basis obtained using Gram-Schmidt method; Final embedding It is obtained through aggregate learning, which is composed of causal features With the calculated distribution The calculation formula is as follows: ； Will Bring it into MLP to obtain the final classification result and use the cross entropy function Constraints are performed; the readout module of the non-causal branch also uses the OCR function and passes the cross entropy loss function Conduct individual training; Model training involves four losses, and the total loss for disease classification is expressed as: ； in, , and is a hyperparameter that balances the contribution, Represents the current number of training times.

Images