Small animal mental disease diagnosis system based on neuroimaging
Through multimodal neuroimaging technology and deep learning algorithms, the problem of inaccurate image data quality and feature extraction in the diagnosis of mental diseases in small animals is solved, high-precision diagnosis is achieved, and diagnostic costs and cycles are reduced.
Patent Information
- Application Number
- CN202510206711.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-25
- Publication Date
- 2025-06-10
AI Technical Summary
In the diagnosis of mental illness in small animals, existing neuroimaging technologies have problems such as low image data quality, inaccurate feature extraction, and poor generalization ability of diagnostic models.
Multimodal neuroimaging technology is used to collect and pre-process images, and automatic feature extraction and analysis is used for deep learning algorithms, and a personalized diagnostic model is constructed to consider individual differences in small animals to achieve accurate diagnosis.
It improves the resolution and signal-to-noise ratio of image data, reduces artificial intervention, improves the accuracy and efficiency of diagnosis, achieves more accurate diagnosis, shortens the diagnosis cycle, and reduces the diagnosis cost.
Smart Images
Figure CN120125543A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of small animal mental illness diagnosis, and specifically relates to a small animal mental illness diagnosis system based on neuroimaging. Background Art
[0002] In the field of small animal mental illness research, traditional diagnostic methods mainly rely on behavioral observation, physiological index detection, histological analysis, etc. These methods have problems such as strong subjectivity, insufficient accuracy, complex operation, and long time consumption. With the rapid development of neuroimaging technology, especially the wide application of high-precision magnetic resonance imaging (MRI) technology, it provides a new means for small animal mental illness diagnosis.
[0003] However, the application of existing neuroimaging technology in small animal mental illness diagnosis still faces many challenges, such as low-quality imaging data, inaccurate feature extraction, poor generalization ability of the diagnostic model, etc., resulting in limited accuracy and reliability of the diagnostic results. Therefore, those skilled in the art have proposed a small animal mental illness diagnosis system based on neuroimaging to solve the problems raised in the background art.
[0004] The above information disclosed in this background art is only used to increase the understanding of the background art of the present invention. Therefore, it may include prior art that is not known to those of ordinary skill in the art. Summary of the Invention
[0005] In order to solve the above technical problems, the present invention provides a small animal mental illness diagnosis system based on neuroimaging to solve the problems of low-quality imaging data, inaccurate feature extraction, and poor generalization ability of the diagnostic model in the prior art.
[0006] To achieve the above object, the present invention provides a small animal mental illness diagnosis system based on neuroimaging, including an image data acquisition and preprocessing module, which uses high-precision neuroimaging equipment to perform multimodal imaging on small animals to obtain neuroimaging data, and preprocesses the acquired imaging data, including steps such as denoising, registration, and segmentation, to improve data quality;
[0007] A feature extraction and analysis module, which is used to automatically extract and analyze the preprocessed imaging data by using deep learning algorithms to extract key neuropathological features;
[0008] A personalized diagnosis model construction module, which is based on the extracted features, uses a personalized diagnosis model to perform disease diagnosis, and adjusts parameters according to the individual differences of small animals to achieve accurate diagnosis;
[0009] A result display and report generation module, which is used to display the diagnostic results in an intuitive manner on the intelligent auxiliary diagnosis interface and generate a detailed diagnostic report;
[0010] Data management and cloud sharing are used to upload diagnostic reports and imaging data to the cloud data management platform to achieve data storage, sharing, and analysis.
[0011] Preferably, the imaging data acquisition and preprocessing module includes an imaging data acquisition unit, and the imaging data acquisition unit uses a magnetic resonance imaging (MRI) scanner with a field strength of 3T or higher to perform multimodal imaging of the small animal brain, including structural MRI, functional MRI, and diffusion tensor imaging.
[0012] Preferably, the imaging data acquisition and preprocessing module further includes a data preprocessing unit. The data preprocessing unit applies a Gaussian filter to smooth the original imaging data to reduce noise interference, uses rigid or non-rigid registration algorithms to align different-modal imaging data to the same anatomical space, and adopts an automatic segmentation algorithm based on atlas or deep learning to distinguish tissues such as brain gray matter, white matter, and cerebrospinal fluid. For the denoising step, the Gaussian filter formula is: where (x,y) is the pixel coordinate, and σ is the standard deviation, which controls the degree of smoothing.
[0013] Preferably, the feature extraction and analysis module includes a feature extraction unit. The feature extraction unit is used to extract morphological features such as brain volume and cortical thickness from structural MRI, analyze the activation patterns of brain regions using functional MRI, calculate functional connectivity, and extract white matter integrity indicators such as fiber bundle directionality and fractional anisotropy from diffusion tensor imaging data.
[0014] Preferably, the feature extraction and analysis module further includes a deep learning model training unit. The deep learning model training unit is used to collect imaging data and corresponding labels of small animals with diagnosed mental diseases, construct a training dataset, adopt a 3D convolutional neural network as the feature extraction and classification model, and use a cross-validation strategy to train the model and adjust hyperparameters to optimize performance. The convolution operation of the 3D convolutional neural network can be expressed as: where X i is the input feature map, W i is the convolution kernel, b is the bias term, * represents the convolution operation, and Y is the output feature map.
[0015] Preferably, the personalized diagnosis model construction module includes an individual feature analysis unit. The individual feature analysis unit is used to analyze individual features such as the age, gender, and genetic background of small animals and extract relevant biomarkers.
[0016] Preferably, the personalized diagnosis model construction module further includes a model personalization adjustment unit, which is used to adjust the weights and bias terms of the deep learning model based on the individual feature analysis results, construct a personalized diagnosis model, and use the transfer learning strategy to take the pre-trained model weights as the starting point and fine-tune according to individual features.
[0017] Preferably, the result display and report generation module includes a result visualization unit, which is used to display the diagnosis results in the form of 3D reconstruction, slice view, etc. on the user interface, providing intuitive image comparison and analysis.
[0018] Preferably, the result display and report generation module further includes a report generation unit, which is used to automatically generate a diagnosis report containing information such as diagnosis results, image data, and analysis conclusions.
[0019] Preferably, the data management and cloud sharing module includes a data upload unit and a data sharing and access control unit. The data upload unit is used to upload the diagnosis report and image data to the cloud server to achieve centralized storage and management of data. The data sharing and access control unit is used to set data access permissions to allow authorized users to access and download data.
[0020] Compared with the prior art, the present invention has the following beneficial effects:
[0021] By adopting multi-modal neuroimaging technology, the present invention improves the resolution and signal-to-noise ratio of image data, providing a reliable data basis for accurate diagnosis; automatically extracts and analyzes image features through deep learning algorithms, reducing human intervention and improving the accuracy and efficiency of diagnosis; constructs a personalized diagnosis model, considering the individual differences of small animals, and realizes more accurate diagnosis; at the same time, the introduction of the cloud data management platform realizes centralized storage, sharing, and collaborative analysis of data, promoting information sharing and cooperation among research teams. In summary, it not only improves the accuracy and reliability of small animal mental illness diagnosis, but also shortens the diagnosis cycle, reduces the diagnosis cost, and provides strong technical support for small animal mental illness research.
[0022] The above summary is only for the purpose of the specification and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the present invention will be readily apparent by reference to the drawings and the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] Figure 1 It is a schematic flow chart of a small animal mental illness diagnosis system based on neuroimaging in an embodiment of the present invention;
[0024] Figure 2 This is a schematic diagram of the specific process of a small animal mental illness diagnosis system based on neuroimaging in an embodiment of the present invention. Specific implementation manners
[0025] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. It should be noted that the drawings are schematic and not drawn to scale. For the sake of clarity and convenience in the figure, the relative sizes and proportions of the parts shown in the figure are exaggerated or reduced in size for illustration, and any size is only exemplary and not limiting.
[0026] Embodiment 1:
[0027] Please refer to Figure 1 - Figure 2 As shown, a small animal mental illness diagnosis system based on neuroimaging includes an image data acquisition and preprocessing module, which uses high-precision neuroimaging equipment to perform multi-modal imaging on small animals to obtain neuroimaging data, improve the accuracy and reliability of diagnosis, and preprocess the acquired image data, including steps such as denoising, registration, and segmentation, to improve the data quality;
[0028] A feature extraction and analysis module, which is used to automatically extract and analyze the preprocessed image data by using deep learning algorithms to extract key neuropathological features;
[0029] A personalized diagnosis model construction module, which based on the extracted features, uses a personalized diagnosis model to perform disease diagnosis and adjusts parameters according to the individual differences of small animals to achieve accurate diagnosis.
[0030] Specifically, the image data acquisition and preprocessing module includes an image data acquisition unit. The image data acquisition unit uses a magnetic resonance imaging (MRI) with a field strength of 3T or higher to perform multi-modal imaging of the small animal brain, including structural MRI (T1-weighted, T2-weighted), functional MRI (BOLD signal), and diffusion tensor imaging (DTI). Using high-field MRI can significantly improve the resolution and signal-to-noise ratio of the image compared with low-field equipment and capture more subtle changes in neural structures.
[0031] Furthermore, the image data acquisition and preprocessing module further includes a data preprocessing unit. The data preprocessing unit applies a Gaussian filter to smooth the original image data to reduce noise interference, uses a rigid or non-rigid registration algorithm to align different-modal image data to the same anatomical space, and adopts an automatic segmentation algorithm based on an atlas or deep learning to distinguish tissues such as cerebral gray matter, white matter, and cerebrospinal fluid. For the denoising step, the Gaussian filter formula is: where (x, y) are pixel coordinates, σ is the standard deviation that controls the smoothness. The automation of the denoising, registration, and segmentation steps reduces manual operations and improves the processing efficiency and accuracy.
[0032] Furthermore, the feature extraction and analysis module includes a feature extraction unit. The feature extraction unit is used to extract morphological features such as brain volume and cortical thickness from structural MRI, analyze the activation patterns of brain regions using functional MRI, calculate functional connectivity, extract white matter integrity indicators such as fiber bundle directionality and fractional anisotropy from diffusion tensor imaging data. Combining multiple features can more comprehensively reflect the structural and functional status of the brain. By combining structural, functional, and diffusion features, it comprehensively reflects the brain state, improves the sensitivity and specificity of diagnosis, and can identify more subtle disease signs.
[0033] Furthermore, the feature extraction and analysis module also includes a deep learning model training unit. The deep learning model training unit is used to collect imaging data of small animals with diagnosed mental diseases and their corresponding labels, construct a training dataset, use a 3D convolutional neural network as the feature extraction and classification model, and use a cross-validation strategy to train the model and adjust hyperparameters to optimize performance. The convolution operation of the 3D convolutional neural network can be expressed as: where X i is the input feature map, W i is the convolution kernel, b is the bias term, * represents the convolution operation, and Y is the output feature map. Compared with traditional feature extraction methods, deep learning can automatically learn high-dimensional feature representations, improve the accuracy of feature extraction, and has stronger generalization ability, which is applicable to different types of small animals and mental diseases.
[0034] Furthermore, the personalized diagnosis model construction module includes an individual feature analysis unit. Since individual differences in small animals have a significant impact on the diagnosis of mental diseases, individual features need to be considered for personalized adjustment. The individual feature analysis unit is used to analyze individual features such as the age, gender, and genetic background of small animals, extract relevant biomarkers, and adjust the model based on individual features, which improves the degree of personalization of diagnosis, realizes accurate diagnosis for different individuals, and improves the accuracy and applicability of diagnosis.
[0035] Furthermore, the personalized diagnosis model construction module also includes a model personalized adjustment unit. The model personalized adjustment unit is used to adjust the weights and bias terms of the deep learning model based on the results of individual feature analysis, construct a personalized diagnosis model, use a transfer learning strategy, take the weights of the pre-trained model as the starting point, and perform fine-tuning according to individual features, which accelerates model convergence, reduces training time, reflects individual differences in diseases, and improves the clinical value of diagnosis.
[0036] As can be seen from the above, the imaging data acquisition unit uses high-precision neuroimaging technology (MRI) to scan the brains of small animals to obtain high-quality imaging data. These imaging data are the basis for subsequent analysis and diagnosis. The raw imaging data collected need to go through preprocessing steps, including denoising, registration, and segmentation, etc. These steps are aimed at improving the quality and accuracy of the imaging data for subsequent feature extraction and analysis. The preprocessed imaging data are input into the deep learning model, and the model will automatically extract the key features in the imaging. These features include structural features, functional features, and diffusion features, etc., which can comprehensively reflect the state of the brains of small animals. The extracted features are used to train the deep learning model. By continuously learning and adjusting parameters, the model can gradually improve the accuracy of diagnosing mental diseases in small animals. During the training process, the model will be continuously iterated and optimized to achieve the best diagnostic effect. The trained deep learning model can perform personalized diagnosis according to the individual differences of small animals. The model will comprehensively consider factors such as the age, gender, and medical history of small animals and give more accurate and personalized diagnostic results. Using the deep learning model for feature extraction and diagnosis can significantly improve the accuracy of diagnosis. The deep learning model has powerful learning and generalization abilities, can automatically extract key features from imaging data, and give accurate diagnostic results. Compared with traditional manual diagnosis methods, it can greatly shorten the diagnostic time. The automated processing and the fast calculation of the deep learning model make the diagnostic process more efficient and convenient, and contribute to formulating more accurate and effective treatment plans and improving the treatment effect.
[0037] Embodiment 2:
[0038] Please refer to Figure 1 - Figure 2 As shown, this embodiment is basically the same as the previous one. The difference lies in that the result display and report generation module is used to display the diagnostic results in an intuitive way on the intelligent auxiliary diagnosis interface and generate a detailed diagnostic report. The interface supports real-time interaction, which is convenient for users to make manual adjustments and confirm the diagnostic results;
[0039] Data management and cloud sharing are used to upload the diagnostic report and imaging data to the cloud data management platform to realize data storage, sharing, and analysis. The platform supports multi-user collaboration, which is convenient for research teams to conduct remote collaboration and data exchange.
[0040] Specifically, the result display and report generation module includes a result visualization unit. The result visualization unit is used to display the diagnostic results in forms such as three-dimensional reconstruction and slice views on the user interface, providing intuitive image comparison and analysis. Visualization helps users understand the diagnostic results, improves the interpretability of the diagnosis, and is convenient for users to understand and accept.
[0041] Furthermore, the result display and report generation module further includes a report generation unit, which is used to automatically generate a diagnostic report containing information such as diagnostic results, imaging data, and analysis conclusions. The report should include multiple items such as basic information of small animals, diagnostic results, imaging comparison charts, analysis conclusions, and suggestions, reducing the manual writing time. The automated report generation improves work efficiency, reduces human errors, and the intuitive result display and detailed report enhance the professionalism and user-friendliness of the system.
[0042] Furthermore, the data management and cloud sharing module includes a data upload unit and a data sharing and access control unit. The data upload unit is used to upload the diagnostic report and imaging data to the cloud server to achieve centralized storage and management of data. Cloud storage facilitates data backup, remote access, and collaborative analysis, improving the security and availability of data and facilitating remote access and analysis. The data sharing and access control unit is used to set data access permissions, allowing authorized users to access and download data. Encrypted transmission and storage technologies are adopted to ensure data security. Compared with the prior art, the cloud platform promotes scientific research cooperation and accelerates the transformation of research results. The powerful data management ability provides data support for the continuous optimization and upgrade of the system.
[0043] As can be seen from the above, through the result display and report generation module, the diagnostic results, three-dimensional reconstructed brain images, and detailed diagnostic reports can be presented to customers in an intuitive way. These results help users better understand the condition of small animals and make corresponding treatment decisions. Through the data management and cloud sharing module, all the collected imaging data and diagnostic results can be centrally stored and managed, facilitating users to access and query at any time, making the management and sharing of data more convenient and secure. At the same time, the data can also be uploaded to the cloud platform to achieve cross-team and cross-regional sharing and collaboration, promoting the development of the field of small animal mental illness diagnosis. By continuously collecting and analyzing new imaging data, researchers can further improve and optimize the deep learning model to improve the accuracy and efficiency of diagnosis.
[0044] To verify the beneficial effects of this method, we conducted the following theoretical experiments and compared them with the existing technical solutions. The following are the theoretical experiment data tables and explanatory notes:
[0045]
[0046]
[0047] Explanatory notes:
[0048] 1. Image data resolution and signal-to-noise ratio:
[0049] The high-field MRI equipment adopted significantly improves the resolution and signal-to-noise ratio of image data compared with the low-field equipment in the existing technical solutions. Higher resolution and signal-to-noise ratio help capture more subtle changes in nerve structures, providing a reliable data basis for accurate diagnosis.
[0050] 2. Feature extraction accuracy:
[0051] Using deep learning algorithms to automatically extract and analyze image features improves the accuracy of feature extraction compared with the traditional feature extraction methods in the existing technical solutions. Deep learning algorithms can automatically learn high-dimensional feature representations and better capture key information in images.
[0052] 3. Diagnostic accuracy:
[0053] Based on high-precision image data and deep learning feature extraction, the diagnostic accuracy of this method is significantly higher than that of the existing technical solutions. This benefits from the powerful learning ability of the deep learning model and the precise matching of the personalized diagnostic model.
[0054] 4. Diagnostic time:
[0055] Through automated processing and the rapid calculation of the deep learning model, the diagnostic time is significantly shortened. Compared with the manual processing and traditional algorithm calculation in the existing technical solutions, this method improves work efficiency and reduces the diagnostic cycle.
[0056] 5. Adaptability of the personalized model:
[0057] The constructed personalized diagnostic model takes into account the individual differences of small animals and can better adapt to the diagnostic needs of different individuals. Compared with the general model in the existing technical solutions, the personalized model of this method has higher adaptability and accuracy.
[0058] 6. Data management efficiency:
[0059] Adopting a cloud data management platform realizes centralized storage, backup and recovery of data, improving data management efficiency. Compared with the decentralized storage and management methods in the existing technical solutions, the data management efficiency of this method is increased by 3 times.
[0060] 7. Convenience of cloud sharing:
[0061] The cloud data management platform supports information sharing and collaborative analysis among research teams, improving the convenience of cloud sharing. Compared with the manual sharing and transmission methods in the existing technical solutions, the cloud sharing of this method is more efficient and convenient.
[0062] Based on the above, through the comparison and analysis of theoretical test data, the superiority and effectiveness of the present method are proved in terms of image data quality, feature extraction accuracy, diagnosis accuracy, diagnosis time, adaptability of personalized models, data management efficiency, and convenience of cloud sharing.
[0063] All standard parts used in the present invention can be purchased from the market. Special-shaped parts can be customized according to the descriptions in the specification and the drawings. The specific connection methods of each part all adopt conventional means such as bolts, rivets, and welding that are mature in the prior art. The machinery, parts, and equipment all adopt conventional models in the prior art. Coupled with the circuit connection adopting the conventional connection method in the prior art, details are not elaborated herein. Contents not described in detail in this specification all belong to the prior art well-known to those skilled in the art.
[0064] In the attached drawings of the disclosed embodiments of the present invention, only the structures related to the disclosed embodiments are involved. Other structures can refer to the general design. Without conflict, the same embodiment and different embodiments of the present invention can be combined with each other.
[0065] Although the present invention has been described in detail with reference to the foregoing embodiments, for those skilled in the art, they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.
Claims
1. A neuroimaging-based small animal mental illness diagnosis system, characterized by: include, The imaging data acquisition and preprocessing module uses high-precision neuroimaging equipment to perform multimodal imaging on small animals, obtain neuroimaging data, and preprocess the acquired imaging data, including denoising, registration, segmentation and other steps to improve data quality; Feature extraction and analysis module, which is used to automatically extract and analyze the pre-processed image data using deep learning algorithms to extract key neuropathological features; The personalized diagnosis model building module uses the personalized diagnosis model to diagnose diseases based on the extracted features, adjusts parameters according to individual differences of small animals, and achieves accurate diagnosis; The result display and report generation module is used to display the diagnosis results in an intuitive way on the intelligent auxiliary diagnosis interface and generate a detailed diagnosis report; Data management and cloud sharing are used to upload diagnostic reports and imaging data to the cloud data management platform to achieve data storage, sharing and analysis.
2. According to claim 1, a small animal mental illness diagnosis system based on neuroimaging is characterized by: The image data acquisition and preprocessing module includes an image data acquisition unit, which uses a magnetic resonance imaging device with a field strength of 3T or higher to perform multimodal imaging of the small animal's brain, including structural MRI, functional MRI and diffusion tensor imaging.
3. The small animal mental illness diagnosis system based on neuroimaging according to claim 2, characterized in that: The image data acquisition and preprocessing module also includes a data preprocessing unit, which applies a Gaussian filter to smooth the original image data to reduce noise interference, uses a rigid or non-rigid registration algorithm to align image data of different modalities to the same anatomical space, and uses an automatic segmentation algorithm based on atlas or deep learning to distinguish brain gray matter, white matter, cerebrospinal fluid and other tissues. For the denoising step, the Gaussian filter formula is: Where (x, y) is the pixel coordinate and σ is the standard deviation, which controls the degree of smoothing.
4. The small animal mental illness diagnosis system based on neuroimaging according to claim 3 is characterized by: The feature extraction and analysis module includes a feature extraction unit, which is used to extract morphological features such as brain volume and cortical thickness from structural MRI, analyze the activation patterns of brain regions using functional MRI, calculate functional connectivity, and extract white matter integrity indicators such as fiber bundle directionality and anisotropy fraction from diffusion tensor imaging data.
5. The small animal mental illness diagnosis system based on neuroimaging according to claim 4, characterized in that: The feature extraction and analysis module also includes a deep learning model training unit, which is used to collect image data and corresponding labels of confirmed small animal mental illnesses, construct a training data set, use a 3D convolutional neural network as a feature extraction and classification model, and use a cross-validation strategy to train the model and adjust hyperparameters to optimize performance. The convolution operation of the 3D convolutional neural network can be expressed as: Where X i is the input feature map, W i is the convolution kernel, b is the bias term, * represents the convolution operation, and Y is the output feature map.
6. The small animal mental illness diagnosis system based on neuroimaging according to claim 5, characterized in that: The personalized diagnostic model building module includes an individual characteristic analysis unit, which is used to analyze individual characteristics of small animals such as age, gender, genetic background, etc., and extract relevant biomarkers.
7. The small animal mental illness diagnosis system based on neuroimaging according to claim 6, characterized in that: The personalized diagnosis model construction module also includes a model personalization adjustment unit, which is used to adjust the weights and bias items of the deep learning model based on the individual feature analysis results, construct a personalized diagnosis model, use a transfer learning strategy, take the pre-trained model weights as the starting point, and perform fine-tuning according to individual characteristics.
8. The small animal mental illness diagnosis system based on neuroimaging according to claim 7, characterized in that: The result display and report generation module includes a result visualization unit, which is used to display the diagnosis results on the user interface in the form of three-dimensional reconstruction, slice view, etc., to provide intuitive image comparison and analysis.
9. The small animal mental illness diagnosis system based on neuroimaging according to claim 8, characterized in that: The result display and report generation module also includes a report generation unit, which is used to automatically generate a diagnosis report containing information such as diagnosis results, image data, and analysis conclusions.
10. The small animal mental illness diagnosis system based on neuroimaging according to claim 9, characterized in that: The data management and cloud sharing module includes a data uploading unit and a data sharing and access control unit. The data uploading unit is used to upload diagnostic reports and imaging data to the cloud server to realize centralized storage and management of data. The data sharing and access control unit is used to set data access rights to allow authorized users to access and download data.