Brain development characteristic evolution analysis system and method based on multi-modal brain image data

Abstract

The invention discloses a brain development characteristic evolution analysis system and method based on multi-mode brain image data. The system comprises a brain development map construction module, a data acquisition module, a data processing module and an evolution matching evaluation module. The method comprises the following steps: collecting demographic information and multi-modal brain image data of multi-age healthy people, constructing a structure, and connecting and activating a three-dimensional group brain development map; the method comprises the following steps: processing individual data by adopting the same process, obtaining a multi-modal brain feature vector in a unified brain region space, mapping the multi-modal brain feature vector to a corresponding group map space, calculating a deviation amount and combining a stability weight to obtain a fusion deviation index, identifying an abnormal brain region according to the fusion deviation index, calculating an abnormal proportion and a risk change rate, and realizing brain development feature evolution analysis. According to the method, group evolution priori and individual multi-modal data are fused, brain development can be dynamically and accurately evaluated, and the problems of static detection, group individual disjunction, insufficient multi-modal fusion and the like are solved.

Description

Technical Field

[0001] This invention belongs to the field of computer-aided bioinformatics processing technology, specifically relating to a system and method for analyzing the evolution of brain development features based on multimodal brain imaging data. Background Technology

[0002] With the accelerating aging of the population, cognitive decline and related neurodegenerative diseases have become significant issues affecting the quality of life of middle-aged and elderly people and public health. Cognitive aging typically manifests as a multifaceted decline in function, including memory loss, reduced attention, impaired executive function, and slower information processing. Its development is characterized by its long-term, gradual nature and significant individual differences. Early prediction, accurate assessment, and dynamic monitoring of cognitive aging have become important research directions in neuroscience, medical engineering, and brain-computer interfaces. Current methods for assessing cognitive aging primarily rely on neuropsychological scales, clinical imaging examinations, and the analysis of certain biomarkers.

[0003] In recent years, with the development of multimodal brain imaging technology, researchers have gradually attempted to comprehensively analyze the cognitive aging process by integrating multiple brain signal data such as electroencephalography (EEG), functional near-infrared spectroscopy (FIR), functional magnetic resonance imaging (fMRI), and diffusion tensor imaging (DTI). However, the existing cognitive aging assessment and brain imaging analysis technology system still has the following significant problems:

[0004] 1. Limitations of static detection: Existing methods are mostly based on cross-sectional data from a single time point or a small number of time points, which is essentially a form of static detection. Due to the lack of a reference for the evolution of the entire population throughout its life cycle, it is difficult to depict the continuous dynamic trajectory of an individual's cognitive aging process, resulting in the system's inability to make trend predictions about an individual's future cognitive state.

[0005] 2. The disconnect between population atlases and individual predictions: While some brain development atlases exist, these atlases primarily serve population statistical analyses, and the results are usually presented in the form of average models. When dealing with individual data, current technologies lack effective mapping and alignment mechanisms, making it difficult to transform the evolutionary prior knowledge of large-scale populations into refined and interpretable localization assessments for individual individuals.

[0006] 3. Insufficient depth of multimodal fusion: Brain aging is a comprehensive result of structural atrophy, abnormal functional connectivity, and changes in task activation patterns. Existing technologies often rely only on single-modal or simple multimodal data splicing, making it difficult to identify heterogeneous evolutionary patterns such as "structural changes first" or "functional reorganization."

[0007] Therefore, how to construct a system that can systematically integrate large-scale population evolutionary priors with individual longitudinal multimodal data, and achieve a precise, dynamic and evolutionarily interpretable assessment scheme for the cognitive aging process, is a technical challenge that urgently needs to be addressed in this field. Summary of the Invention

[0008] This invention addresses the limitations of existing technologies, such as static detection, the disconnect between population mapping and individual prediction, and insufficient depth of multimodal fusion, by providing a system and method for analyzing the evolution of brain development features based on multimodal brain imaging data.

[0009] To achieve the above-mentioned technical objectives, the embodiments of the present invention adopt the following technical solutions.

[0010] In a first aspect, embodiments of the present invention provide a brain development feature evolution analysis system based on multimodal brain imaging data, comprising:

[0011] The brain development atlas construction module is used to acquire demographic information and multimodal brain imaging data of healthy people of multiple ages. Based on the multimodal brain imaging data of people of multiple ages, preprocessing, brain region segmentation and feature extraction are performed to extract structural features, connectivity features and activation features. Combined with the demographic information of people of multiple ages, a population brain development atlas reflecting the evolution of the brain structure, function and connectivity patterns of the population with age is constructed.

[0012] The data acquisition module is used to acquire demographic information and multimodal brain imaging data of the individual to be evaluated. The multimodal brain imaging data of the individual to be evaluated is consistent with the multimodal brain imaging data acquired by the brain development map construction module.

[0013] The data processing module is used to preprocess, segment, and extract features from the multimodal brain data acquired by the data acquisition module using the same processing methods as the brain development map construction module, so as to obtain the individual multimodal brain feature vector in a unified brain region spatial coordinate system.

[0014] The evolutionary matching assessment module is used to map the individual's multimodal brain feature vector to the corresponding population brain development atlas space based on the demographic information of the individual to be assessed. By calculating the deviation of the individual's multimodal brain feature vector from the evolutionary trajectory of the peer group, and combining the stability weights of each modality feature, a fusion deviation index is obtained. The fusion deviation index is used to quantitatively assess the degree of evolution of the individual's brain development state relative to the population baseline. Based on the fusion deviation index, abnormal brain regions are identified, the proportion of abnormal brain regions and the risk change rate are calculated, and brain development feature evolution analysis is realized, outputting the brain development feature evolution analysis results.

[0015] Furthermore, the population brain development atlas constructed by the brain development atlas construction module includes a structural development atlas, a connectivity development atlas, and an activation development atlas;

[0016] Among them, the structural development atlas is constructed based on brain region-level structural feature extraction and is used to establish a distribution model of brain region volume, gray matter volume, cortical thickness and cortical surface area as a function of age;

[0017] The connectivity development atlas is constructed based on spontaneous brain functional signals in a resting state. By utilizing the time-series correlation of these spontaneous brain functional signals, a distribution model of the connection strength between brain networks and the overall connectivity level of the whole brain as a function of age is established.

[0018] The activation development atlas is constructed based on brain functional response signals under task conditions. By extracting activation features, a distribution model of the degree of activation of brain regions and the similarity of multi-voxel patterns under specific tasks is established as a function of age.

[0019] Furthermore, the multimodal brain imaging data includes structural magnetic resonance imaging data, resting-state functional magnetic resonance imaging data, functional magnetic resonance imaging data for specific cognitive tasks, and functional near-infrared brain imaging data.

[0020] Furthermore, the stability weight in the evolutionary matching evaluation module is determined based on the fluctuation range of each modality feature in healthy individuals of the same age. The smaller the fluctuation range, the greater the weight; the larger the fluctuation range, the smaller the weight. Moreover, the sum of the weights of the same brain region under different modalities is 1.

[0021] Furthermore, when the evolutionary matching evaluation module identifies abnormal brain regions, it uses a preset abnormality discrimination threshold. When the absolute value of the fusion deviation index of a certain brain region of an individual reaches or exceeds the threshold, the brain region is determined to be an abnormal brain region, and the proportion of abnormal brain regions to all brain regions is calculated.

[0022] Secondly, embodiments of the present invention provide a method for analyzing the evolution of brain development features based on multimodal brain imaging data, comprising the following steps:

[0023] Step 1: Obtain demographic information and multimodal brain imaging data of healthy individuals across multiple age groups; preprocess, segment, and extract features from the multimodal brain imaging data of individuals across multiple age groups, extracting structural features, connectivity features, and activation features; combine the demographic information of individuals across multiple age groups to construct a population brain development atlas that reflects the evolution of brain structure, function, and connectivity patterns with age.

[0024] Step 2: Obtain demographic information and multimodal brain imaging data of the individual to be evaluated. The multimodal brain imaging data of the individual to be evaluated is consistent with the data type of multimodal brain imaging data of healthy individuals of multiple age groups.

[0025] Step 3: Using the same processing method as for multimodal brain imaging data of healthy individuals of multiple ages, preprocess, segment, and extract features from the multimodal brain data of the individual to be evaluated to obtain the individual's multimodal brain feature vector in a unified brain region spatial coordinate system;

[0026] Step 4: Based on the demographic information of the individual to be evaluated, the individual's multimodal brain feature vector is mapped to the corresponding population brain development atlas space. By calculating the deviation of the individual's multimodal brain feature vector from the evolutionary trajectory of the peer group, and combining the stability weights of each modality feature, a fusion deviation index is obtained. The degree of evolution of the individual's brain development state relative to the population baseline is quantitatively assessed based on the fusion deviation index. Abnormal brain regions are identified based on the fusion deviation index, and the proportion of abnormal brain regions and the risk change rate are calculated to realize the brain development feature evolution analysis and output the brain development feature evolution analysis results.

[0027] Further, the process of extracting the structural features in step 1 includes: segmenting the multimodal brain imaging data into brain regions using a cortical segmentation atlas; for each brain region, counting the number of voxels belonging to that brain region and determining the voxel volume of that brain region by combining the physical volume of each voxel; weighting the voxels in the brain region by combining the gray matter probability map to obtain the gray matter volume of that brain region; calculating the average Euclidean distance from the set of cortical vertex coordinates to the adjacent white matter coordinate sets in the brain region to obtain the cortical thickness of that brain region; and counting the area of ​​the cortical surface corresponding to the brain region to obtain the cortical surface area of ​​that brain region.

[0028] Furthermore, the population brain development atlas mentioned in step 1 includes a connectivity development atlas. The construction process of the connectivity development atlas includes: dividing the resting-state multimodal brain imaging data into multiple brain networks using a cortical segmentation atlas; calculating the Pearson correlation coefficient of the BOLD signal time series between any two brain networks; generating a functional connectivity matrix; and based on the functional connectivity matrix, calculating the sum of the connectivity strengths of a single brain network with all other brain networks and the average connectivity strength among all brain networks in the whole brain, and establishing a distribution model of the above connectivity strengths changing with age.

[0029] Furthermore, the population brain development atlas mentioned in step 1 includes an activation development atlas. The construction process of the activation development atlas includes: using a cortical segmentation atlas to segment the task-state multimodal brain imaging data into brain regions, statistically pooling the voxel signals in each brain region, calculating the degree of neural activity activation related to a specific task, calculating the similarity of multi-voxel patterns in the brain region, and establishing a distribution model of the degree of brain region activation and the similarity of multi-voxel patterns with age.

[0030] Furthermore, the specific method for identifying abnormal brain regions, calculating the proportion of abnormal brain regions and the risk change rate based on the fusion deviation index in step 4, and realizing the evolutionary analysis of brain development characteristics, is as follows:

[0031] A preset abnormality detection threshold is set, and the absolute value of the fusion deviation index of each brain region of the individual to be evaluated is compared with the threshold to screen out abnormal brain regions and form a set of abnormal brain regions.

[0032] The proportion of abnormal brain regions is calculated by comparing the number of brain regions in the abnormal brain region set with the total number of brain regions in the whole brain.

[0033] Based on the proportion of abnormal brain regions in at least two individual assessments and the corresponding assessment time intervals, the risk change rate is calculated. Combined with the age-average change trend of the proportion of abnormal brain regions in the population brain development atlas, the degree of deviation of the individual from the healthy population of the same age is obtained. The proportion of abnormal brain regions is mapped to the norm distribution of the healthy population of the same age to determine the age-related risk percentile to achieve cognitive risk stratification. Combining the risk change rate, the degree of deviation, and the population evolution trend, the future proportion of abnormal brain regions in the individual is predicted, the future cognitive risk level is determined, and corresponding early warning measures are appropriate.

[0034] It should be understood that the summary section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description.

[0035] Compared with the prior art, the beneficial technical effects achieved by the present invention are as follows:

[0036] This invention constructs a full-life-cycle population brain development atlas, establishing a continuous benchmark for the evolution of brain structure, functional connectivity, and task activation with age, providing a unified and stable population reference framework for individual assessment. This invention achieves deep fusion of multimodal images, uniformly modeling and weighting structural, connectivity, and activation features, overcoming the limitations of single modalities and accurately identifying brain region abnormalities and evolutionary heterogeneity. By precisely aligning the population atlas with individual data and quantifying the deviation of an individual's characteristics from their peers, it achieves individualized, quantitative, and interpretable assessment of brain development / cognitive aging.

[0037] This invention supports dynamic monitoring and trend prediction, calculating the proportion of abnormal brain regions and the rate of risk change, enabling cognitive risk stratification and future risk prediction, and providing early warning capabilities. It achieves full-process standardization and automation, providing integrated output from data processing to result visualization, thus improving the efficiency and clinical applicability of cognitive aging assessment. Attached Figure Description

[0038] The accompanying drawings described herein are for illustrative purposes only and are not intended to limit the scope of the invention in any way. Furthermore, the shapes and proportions of the components in the drawings are merely illustrative to aid in understanding the invention and are not intended to specifically limit the shapes and proportions of the components. Those skilled in the art, guided by the teachings of this invention, can select various possible shapes and proportions to implement the invention according to specific circumstances. In the drawings:

[0039] Figure 1 A schematic diagram of the brain development feature evolution analysis system based on multimodal brain imaging data provided in this embodiment;

[0040] Figure 2 This is a schematic diagram of the population brain development atlas constructed in the embodiment;

[0041] Figure 3 The above is a schematic diagram of the Brainnetome cortex segmentation map in the embodiment, where (a) is a schematic diagram of the Brainnetome cortex segmentation map of the left brain region and (b) is a schematic diagram of the Brainnetome cortex segmentation map of the right brain region.

[0042] Figure 4 The following is a schematic diagram of brain region-level structural feature extraction in the embodiment, where (a) is an unprocessed raw T1-weighted MRI image, (b) is a masked view after tissue segmentation and structural reconstruction, and (c) is a schematic diagram of cortical mask visualization of the target brain region.

[0043] Figure 5 This is a schematic diagram illustrating the development trend of structural or functional indicators in the embodiments;

[0044] Figure 6 The above is a schematic diagram of the DU15NET cortical segmentation map in the embodiment. In the middle (a), it is a schematic diagram of the DU15NET cortical segmentation map in the left brain region, and in the middle (b), it is a schematic diagram of the DU15NET cortical segmentation map in the right brain region.

[0045] Figure 7 This is a schematic diagram of the functional connections in the embodiment;

[0046] Figure 8 This is a schematic diagram of the searchlight method in the embodiment;

[0047] Figure 9 A schematic diagram for identifying abnormal brain regions at the brain region scale for individuals to be evaluated;

[0048] Figure 10 This is a schematic diagram illustrating the principle of the brain development feature evolution analysis method based on multimodal brain imaging data provided in the embodiment. Detailed Implementation

[0049] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this invention.

[0050] It should be fully understood that the user's EEG signals, multimodal brain imaging data, etc. involved in this invention are all information and data authorized by the user or fully authorized by all parties. The use of user information should follow the privacy policies and practices of the industry that are generally considered to meet or exceed the requirements for maintaining user privacy. The collection, use and processing of related data need to comply with relevant laws, regulations and standards, and provide corresponding operation access points for users to choose to authorize or refuse.

[0051] In recent years, with the development of multimodal brain imaging technology, researchers have gradually attempted to comprehensively analyze the cognitive aging process by fusing multiple brain signal data such as electroencephalography (EEG), functional near-infrared spectroscopy (FIR), functional magnetic resonance imaging (fMRI), and diffusion tensor imaging (DTI). However, most existing multimodal analysis methods still remain at the level of feature splicing or simple fusion, lacking a unified population reference framework. This makes it difficult to accurately characterize the relative position of an individual's brain state within the population's developmental trajectory, thus limiting the ability to provide a refined description and prediction of the cognitive aging process.

[0052] Recent studies on brain development atlases have provided crucial insights into the structural and functional evolution of the human brain at different ages. Existing brain development atlases are typically constructed based on large-scale population samples to describe typical brain structural patterns and functional connectivity patterns at different ages. However, these atlases are mostly used for population statistical analysis and are difficult to directly serve for assessing and predicting the cognitive state of individual individuals. A systematic application scheme that effectively integrates population development atlases with individual brain states has yet to be developed.

[0053] Numerous existing studies have revealed that brain structure and function exhibit distinct stages of evolution with age, with different brain regions displaying differentiated degenerative and compensatory characteristics during cognitive aging. Furthermore, there is a significant correlation between changes in cognitive function and brain network topology. However, within the current technological framework, a comprehensive scheme for predicting and assessing cognitive aging that can systematically integrate large-scale brain development atlases and individual multimodal brain imaging data has yet to be developed.

[0054] This invention provides a brain development feature evolution analysis system and method based on multimodal brain imaging data. It independently models and collaboratively analyzes deviation patterns in the three dimensions of structure, connectivity, and activation from a physiological evolutionary perspective. This can characterize the continuous evolutionary trajectory of an individual's cognitive aging process, enabling trend analysis and prediction of future cognitive states. Furthermore, it is effectively combined with an individual cognitive prediction system to provide a refined and interpretable basis for the localization and assessment of cognitive aging for a single individual.

[0055] Example 1: A brain development feature evolution analysis system based on multimodal brain imaging data. This system uses a brain development atlas as a population prior reference framework and achieves accurate prediction and dynamic assessment of the cognitive aging process through multimodal brain imaging data acquisition and individual longitudinal modeling. Specifically, it includes a brain development atlas construction module, a data acquisition module, a data processing module, and an evolutionary matching assessment module.

[0056] The brain development atlas construction module is used to acquire demographic information and multimodal brain imaging data from multiple age groups. Based on this data, it constructs a population brain development atlas reflecting the age-related evolution of brain structure, function, and connectivity patterns, providing a group-based prior reference framework for individual cognitive aging assessment. The module uses a large-scale population sample as its foundation, combined with demographic information from multiple age groups, such as grouping samples according to preset age intervals. It then performs unified preprocessing, brain region segmentation, and feature extraction on the multimodal brain data for each age group to construct structural, connectivity, and activation atlases.

[0057] Data Acquisition Module: This module acquires demographic information and multimodal brain imaging data of the individuals to be evaluated. The multimodal brain imaging data of the individuals to be evaluated is consistent with the data type acquired by the brain development atlas construction module. The module uses individuals as the acquisition target and simultaneously acquires demographic information and multimodal brain imaging data of the individuals to be evaluated through a standardized acquisition process. The multimodal brain imaging data includes, but is not limited to: structural magnetic resonance imaging (MRI) data, functional magnetic resonance imaging (fMRI) data, and functional near-infrared spectroscopy (FIR) brain imaging data. Structural MRI data is used to characterize brain region volume, cortical thickness, and brain structural integrity; fMRI and FIR data are used to characterize brain region activation patterns and functional connectivity features.

[0058] Demographic information is used to characterize the basic attributes of an individual, including at least date of birth or age, and may also include information such as gender, years of education, lifestyle, medical history and family history.

[0059] Data Processing Module: This module uses the same methods as the Brain Development Mapping Module for preprocessing, brain region segmentation, and feature extraction. It performs unified preprocessing, standardization, and feature construction on the multimodal brain data of the individuals to be evaluated obtained by the data acquisition module to obtain individual multimodal brain feature vectors.

[0060] In this embodiment, corresponding preprocessing procedures are applied to multimodal brain data from different sources, including but not limited to: denoising, bias field correction, brain tissue segmentation, and spatial registration processing of structural magnetic resonance imaging data to obtain brain region volume, cortical thickness, and gray and white matter distribution characteristics in standard brain space; temporal drift correction, motion artifact removal, signal filtering, and standardization processing of functional magnetic resonance imaging data and functional near-infrared brain imaging data to obtain stable brain region activation intensity and functional connectivity characteristics; after completing single-modal preprocessing, spatial alignment and temporal synchronization processing are further performed on different modal data, and multimodal brain features are mapped to the same brain region spatial coordinate system through a unified brain region segmentation template, thereby realizing the construction of the correspondence between multimodal features at the brain region level.

[0061] Evolutionary Matching Assessment Module: This module maps the individual's multimodal brain feature vectors to the corresponding population brain development atlas space based on the individual's demographic information. It calculates the deviation of the individual's multimodal brain feature vectors from the evolutionary trajectory of the same-age population and obtains a fusion deviation index by combining the stability weights of each modality feature. Based on the fusion deviation index, it quantitatively assesses the degree of evolution of the individual's brain development state relative to the population baseline. It identifies abnormal brain regions based on the fusion deviation index, calculates the proportion of abnormal brain regions and the risk change rate, realizes the evolutionary analysis of brain development features, and outputs the results of the brain development feature evolutionary analysis.

[0062] In this embodiment, the evolutionary matching evaluation module takes the individual's demographic information and individual multimodal brain feature vector output by the data processing module as input, and matches the individual's multimodal brain feature vector with the continuous population brain development map corresponding to the individual's demographic information generated by the brain development map construction module. In this embodiment, the relative position of the individual in the brain development trajectory can be determined by distance measurement, similarity calculation and probability distribution mapping.

[0063] In some embodiments, such as Figure 1 As shown, the brain development feature evolution analysis system based on multimodal brain imaging data also includes a display device, which is electrically connected to the evolution matching evaluation module, and is used to receive the brain development feature evolution analysis results output by the evolution matching evaluation module and to visualize them.

[0064] In this embodiment, the display device can be a liquid crystal display, an OLED display, a touch screen, or other display device with image / data display capabilities. The display device is communicatively connected to the evolutionary matching assessment module in the system to display at least one of the stratification and prediction or intermediate processing results of cognitive risk, so that users can view, monitor, or operate it.

[0065] Example 2: Based on the same inventive concept as the brain development feature evolution analysis system based on multimodal brain imaging data provided in Example 1, this example provides a method for brain development feature evolution analysis based on multimodal brain imaging data.

[0066] In this embodiment, the method includes the following steps:

[0067] Step 1: Obtain demographic information and multimodal brain imaging data of healthy individuals across multiple age groups; preprocess, segment, and extract features from the multimodal brain imaging data of individuals across multiple age groups, extracting structural features, connectivity features, and activation features; combine the demographic information of individuals across multiple age groups to construct a population brain development atlas that reflects the evolution of brain structure, function, and connectivity patterns with age.

[0068] In this embodiment, a large-scale brain development atlas covering the entire lifespan is constructed based on healthy population data from multiple publicly available brain imaging databases. This atlas characterizes the population evolution patterns of human brain structure, connectivity, and cognitive ability-related activation as they change with age. Specifically, healthy subject samples of different age groups are selected from publicly available databases, and their demographic information (especially birth dates) is statistically recorded. Based on these subjects' birth dates, their specific age information is calculated, as shown in Formula 1. The date of data collection. This refers to the specific date of birth.

[0069] (Formula 1);

[0070] in Indicates the first The age of the subjects at the time of the MRI scan.

[0071] Subsequently, multimodal brain imaging data corresponding to the given age were acquired. This multimodal brain imaging data included at least structural magnetic resonance imaging (MRI) data, resting-state functional magnetic resonance imaging (fMRI) data, and fMRI data for specific cognitive tasks. The brain imaging data from the aforementioned healthy individuals underwent standardized data preprocessing and quality control, including denoising, registration, segmentation, and standard spatial normalization.

[0072] In some embodiments, the final data and information set, as shown in Formula 2, is obtained for each subject in the selected public database. .in For the subject's structural magnetic resonance imaging data, This refers to the resting-state functional magnetic resonance imaging data of the subjects. Functional magnetic resonance imaging (fMRI) data for specific cognitive tasks of the subjects. Based on demographic information.

[0073] (Formula 2).

[0074] Next, large-scale structural developmental maps, connectivity developmental maps, and activation developmental maps will be constructed respectively, such as Figure 2 As shown.

[0075] (1) For structural development atlases, the Brainnetome cortical segmentation atlas is first used to segment the user's structural magnetic resonance imaging data into brain regions, as shown in Formula 3 and Figure 3 As shown. Among them. This is a collection of brain regions segmented from structural magnetic resonance imaging data using Brainnetome cortical segmentation atlas, containing a total of Each brain region.

[0076] (Formula 3);

[0077] This is the first brain region segmented from structural magnetic resonance imaging data using Brainnetome cortical segmentation atlas. The first brain region segmentation after using Brainnetome cortical segmentation atlas to segment structural magnetic resonance imaging data. Each brain region.

[0078] Brain region-level structural features are extracted from structural magnetic resonance imaging (MRI) data to achieve the transformation from voxel-level image data to brain region-level structural representation. Voxel signals within each brain region are statistically aggregated and combined with tissue segmentation results, cortical surface reconstruction results, and spatial deformation information to comprehensively represent the brain region from multiple structural dimensions, such as... Figure 4 As shown.

[0079] The structural feature extraction process is not limited to a single indicator. In some embodiments, the structural state of brain regions can be characterized from multiple complementary dimensions:

[0080] Voxel volume: By statistically analyzing the number of voxels contained in a brain region and their physical volume, the overall size change of the brain region on a spatial scale is characterized.

[0081] In some embodiments, the voxel volume is determined using formula 4, wherein Representing the The first subject The volume of each brain region Representing the The first subject Each brain region, For the i-th subject, the m-th brain region Any voxel, For voxel volume, statistics belong to The number of voxels in a brain region, and the volume of that brain region calculated based on the voxel volume:

[0082] (Formula 4).

[0083] Gray matter volume: Combining gray matter probability maps or gray matter density maps, it quantifies the distribution of neural tissue content in brain regions and is used to reflect macroscopic changes in neuronal and synaptic structures.

[0084] In some embodiments, the gray matter volume is determined using formula 5, wherein... Representing the The first subject The gray matter volume of each brain region is calculated based on statistical volume, according to the... The first subject Gray matter probability map of individual units To calculate the volume of gray matter in this brain region:

[0085] (Formula 5).

[0086] Cortical thickness: Based on the results of cortical surface reconstruction, the thickness of the corresponding cortical regions in the brain is statistically aggregated to characterize the changes in cortical microstructure with age or disease evolution.

[0087] In some embodiments, the cortical thickness is determined using formulas 6 and 7, wherein Representing the The first subject Average cortical thickness of each brain region Representative belongs to The set of vertex coordinates of brain regions is calculated by calculating the coordinates of each vertex in this set. The distance from the coordinates of each vertex in the adjacent white matter coordinate set European distance Finally, calculate the average value.

[0088] (Formula 6);

[0089] (Formula 7).

[0090] Cortical surface area: By statistically analyzing the cortical surface area corresponding to brain regions, the structural characteristics of brain regions in terms of cortical unfolding morphology are characterized.

[0091] In some embodiments, the cortical surface area can be determined using Formula 8, wherein Representing the The first subject Cortical surface area of ​​each brain region Representative belongs to The set of triangular faces of brain regions, the area of ​​a single triangular face is defined as... .

[0092] (Formula 8).

[0093] After the above feature extraction process, we obtain the feature extraction result for the first... A set of structural features of the subjects , Representing the Voxel volumes of all brain regions in each subject Representing the Gray matter volume of all brain regions in each subject Representing the Cortical thickness in all brain regions of each subject Representing the The cortical surface area of ​​all brain regions of each subject is shown in Formula 9. For each indicator, the first... The first subject The structural features of each brain region are denoted as Based on the first The age of the participants By fitting the trajectory of each type of structural feature in each brain region with age, mean trajectory functions were constructed respectively. and variance trajectory function As shown in Formulas 10 and 11. Based on the mean trajectory and variance trajectory, the first... Modeling any structural index of a brain region as a conditional distribution under any age condition allows for the formation of a complete developmental distribution model of that structural index within a continuous age space, as shown in Equation 12 and... Figure 5 As shown.

[0094] Figure 5 The central vertical axis represents any structural or functional index of any brain region. The three points from left to right represent three subjects who are at the normal age trend, slightly younger than the normal age trend, and significantly younger than the normal age trend, respectively.

[0095] (Formula 9);

[0096] (Formula 10);

[0097] (Formula 11);

[0098] (Formula 12).

[0099] (2) For connectivity development atlases, the DU15NET cortical segmentation atlas is first used to segment the user's resting-state functional magnetic resonance imaging data into brain networks, as shown in Equation 13 and... Figure 6 As shown. Among them. This is a collection of brain networks obtained after segmenting resting-state magnetic resonance imaging data using the DU15NET cortical segmentation atlas. A total of A brain network.

[0100] (Formula 13);

[0101] This is the first brain network segmented from resting-state MRI data using the DU15NET cortical segmentation atlas. The first brain network segmentation of resting-state magnetic resonance imaging data using the DU15NET cortical segmentation atlas. A brain network.

[0102] Brain region-level functional features were extracted from resting-state functional magnetic resonance imaging (fMRI) data to characterize the intensity of spontaneous brain activity and functional connectivity between brain regions at the brain network scale. Voxel signals within each brain network were statistically aggregated, and the time-series correlation between any two networks was calculated to obtain the functional connectivity matrix, as shown in Equation 14. Figure 7 As shown in Formula 14 Representing the The first subject The brain network and the first The functional connectivity strength of individual brain networks Representing the The first subject Resting-state BOLD signal time series of a brain network Representing the Resting-state BOLD signal time series of the nth brain network of a subject This represents the Pearson correlation function. Then... Perform Fisher-Z transform to obtain As shown in Formula 15.

[0103] (Formula 14);

[0104] (Equation 15).

[0105] Subsequently, based on the functional connectivity matrix, various features used to measure the functional connectivity state are calculated:

[0106] Brain network connectivity strength is characterized by statistically analyzing the connectivity strength between a single brain network and all other brain networks. As shown in Equation 16... Representing the The first subject The strength of brain network connectivity in a brain network.

[0107] (Formula 16);

[0108] Overall connectivity strength, by summarizing all connectivity metrics, characterizes the overall connectivity level of the entire brain. As shown in Formula 17. Representing the The overall connectivity strength of the subjects.

[0109] (Equation 17).

[0110] After the above feature extraction process, we obtain the feature extraction result for the first... resting-state feature set of each subject , Representing the The sum of the connection strengths of all brain networks of each subject. Representing the The overall connectivity strength of the subjects is shown in Equation 18. For each index, the first... The first subject The resting-state characteristics of a brain network are denoted as follows: Based on the first The age of the participants By fitting the trajectory of each type of resting-state feature of each brain region with age, mean trajectory functions were constructed respectively. and variance trajectory function As shown in Formulas 19 and 20. Based on the mean trajectory and variance trajectory, the first... Any resting-state index of a brain network can be modeled as a conditional distribution under any age condition, thereby forming a complete developmental distribution model of the resting-state index in a continuous age space, as shown in Equation 21.

[0111] (Formula 18);

[0112] (Formula 19);

[0113] (Formula 20);

[0114] (Formula 21).

[0115] (3) For the activation developmental atlas, the Brainnetome cortical segmentation atlas, consistent with the structural developmental atlas, is used to segment the user's functional magnetic resonance imaging (fMRI) data into brain regions. Brain region-level functional features are extracted from the task-state fMRI data to characterize task-dependent activity intensity at a task-related scale within the brain regions. Voxel signals within each brain region are statistically aggregated to calculate task-related neural activity, yielding the activation level of each brain region, as shown in Equation 22. Representing the Individual elements in time The BOLD signal at that time. The design matrix is ​​the result of the convolution of the hemodynamic response function and the event function. For noise terms, This is the regression coefficient of the voxel, used to characterize the degree of activation associated with the task event.

[0116] (Formula 22).

[0117] Subsequently, based on whole-brain activation, various activation features were calculated to measure task relevance:

[0118] Brain region activation level: By statistically analyzing the activation level of a specific brain region, the activation level of that brain region is characterized in relation to the task. As shown in Formula 23, Representing the The first subject Activation intensity of individual brain regions For the first The first subject A set of voxels from each brain region The voxels belong to this brain region. This represents the degree of activation of the voxel;

[0119] .

[0120] Multi-voxel pattern similarity: This characterizes differences in neural representations by calculating the multi-voxel pattern similarity across all voxels within a specific brain region. As shown in equations 24 and 25. Representing the The first subject Multi-voxel pattern similarity of voxels, based on task conditions and baseline conditions Calculated Representative of the first Individual element as the center, condition as The set of voxel activation levels, limited by the radius of the searchlight method. Representative of the first Individual element as the center, condition as The set of voxel activation levels, limited by the radius of the searchlight method. For the first The first subject A set of voxels from each brain region The voxels belong to this brain region. For the first The first subject Multi-voxel pattern similarity in individual brain regions. Voxel selection in the searchlight method, such as... Figure 8 As shown.

[0121] (Formula 24);

[0122] (Equation 25).

[0123] After the above feature extraction process, we obtain the feature extraction result for the first... Activation feature set of each subject , Representing the The sum of activation intensity in all brain regions of each subject Representing the The sum of multi-voxel pattern similarities across all brain regions of each subject, as shown in Equation 26. For each metric, the first... The first subject The task-state characteristics of each brain region are denoted as follows: Based on the first The age of the participants We fitted the trajectory of each brain region's task-state features with age development, and constructed mean trajectory functions for each region. and variance trajectory function As shown in Formulas 27 and 28. Based on the mean trajectory and variance trajectory, the first... Any task-state index in a brain region can be modeled as a conditional distribution under any age condition, thereby forming a complete developmental distribution model of the task-state index in a continuous age space, as shown in Equation 29.

[0124] (Formula 26);

[0125] (Formula 27);

[0126] (Formula 28);

[0127] (Formula 29).

[0128] Having gone through (1) to (3) above, we have obtained conditional distribution models of the characteristics (structural features, resting-state connectivity features, and task-state activation features) of the three modalities as a function of age. In each conditional distribution model that varies with age, we can, based on the first... The range of fluctuation of a feature of any modality in a brain region determines its weight in subsequent evaluation. If the range of fluctuation is small, it indicates that this indicator is relatively consistent among healthy individuals of the same age (low noise, stable, and highly predictable), and should be given a greater weight; conversely, it indicates that this indicator is relatively dispersed among healthy individuals of the same age (high noise, large individual differences), and this modality should have a smaller weight in subsequent evaluation.

[0129] As shown in formulas 30 and 31 Representing the The first mode Any characteristic indicator of a brain region changes with age The stability weights of the change Representing the The first mode Any characteristic indicator of a brain region changes with age The standard deviation of change To avoid constants with infinitely large weights, normalization is then performed to obtain the result at the th... The first mode Normalized brain regions with age Change in stability weights The weights of this brain region are summed to 1 across the three modalities. A set representing modes (including structured mode, resting mode, and task mode).

[0130] (Formula 30);

[0131] (Equation 31).

[0132] Step 2: Obtain demographic information and multimodal brain imaging data of the individual to be evaluated. The multimodal brain imaging data of the individual to be evaluated is consistent with the data type of multimodal brain imaging data of healthy individuals of various ages.

[0133] Step 3: Using the same processing method as for multimodal brain imaging data of healthy individuals of multiple ages, preprocess, segment, and extract features from the multimodal brain data of the individual to be evaluated to obtain the individual's multimodal brain feature vector in a unified brain region spatial coordinate system. The method is the same as the process of constructing a population brain development atlas.

[0134] Step 4: Based on the demographic information of the individual to be evaluated, the individual's multimodal brain feature vector is mapped to the corresponding population brain development atlas space. By calculating the deviation of the individual's multimodal brain feature vector from the evolutionary trajectory of the peer group, and combining the stability weights of each modality feature, a fusion deviation index is obtained. The fusion deviation index is used to quantitatively assess the degree of evolution of the individual's brain development state relative to the population baseline. Based on the fusion deviation index, abnormal brain regions are identified, the proportion of abnormal brain regions and the risk change rate are calculated, and the brain development feature evolution analysis is realized, outputting the brain development feature evolution analysis results.

[0135] This invention constructs a large-scale brain development atlas covering the entire life cycle, establishing a population evolution benchmark for healthy individuals as they age across multiple neural dimensions, including structure, connectivity, and functional activation. This atlas provides a stable statistical reference for subsequent individual assessments.

[0136] In the embodiment, for the first One individual to be evaluated, whose age is Under these conditions, the observed values ​​of structural features, connectivity features, and activation features are obtained respectively, and matched with the mean trajectory and variance trajectory in the corresponding developmental atlas. As shown in Equations 32 and 33, Representing the The individuals to be evaluated in the first The first mode The fusion deviation index of individual brain regions compared to large-scale brain development atlases, Representing the The individuals to be evaluated in the first The first mode Any characteristic indicator of a brain region Representative at Age when The first mode The mean of any characteristic index of a brain region, Representative at Age when The first mode The standard deviation of any characteristic index of a brain region Representing the The individual to be evaluated in the 1st multimodal fusion... The fusion deviation index of individual brain regions compared to large-scale brain development atlases, The set representing modalities (including structured state, resting state, and task state), Representative at the The first mode After normalization of individual brain regions Stability weights at different ages.

[0137] (Formula 32);

[0138] (Equation 33).

[0139] After obtaining the fusion deviation index in the embodiments, the present invention further adopts the following risk assessment method to transform it into interpretable risk evidence, thereby realizing the detection and stratification of an individual's current cognitive risk.

[0140] In some embodiments, for the first For each individual to be evaluated, the first step is to identify a set of abnormal brain regions at the brain region scale. (See Equation 34 and...) Figure 9 As shown, This is an abnormal collection of brain regions. This is the anomaly detection threshold (default setting is 1.96 or 2.58). If the first... The individual to be evaluated in the 1st multimodal fusion... Fusion deviation index of individual brain regions relative to large-scale brain development atlases If it is greater than or equal to the anomaly detection threshold, then It was identified as an abnormal brain region. The next step was to calculate... The proportion of abnormal brain regions in individuals to be evaluated As shown in Formula 35. This represents the total number of brain regions that have been segmented.

[0141] Furthermore, in order to achieve a stratified assessment of an individual's current cognitive risk, this invention maps the proportion of abnormal brain regions to the norm distribution of healthy individuals of the same age to obtain the risk percentiles of the same age, thereby achieving quantitative detection and risk stratification of an individual's current risk while ensuring interpretability.

[0142] (Formula 34);

[0143] (Formula 35).

[0144] After obtaining the fusion deviation index and risk percentile, this invention further predicts the trend of individual future cognitive risk. In some embodiments, such as Figure 10 As shown, for the first For an individual to be evaluated, if they have undergone at least two risk assessments to determine their current risk status, then the individual's rate of risk change or trend can be further defined. For some individuals, both assessments may appear normal, but if the rate of increase significantly deviates from the general trend in large-scale brain development atlases, then they also possess a higher risk. Calculate the... The rate of risk change or the rate of anomalous spread of the individual to be assessed. As shown in formulas 36 to 38, Representing the The individual to be evaluated at the first evaluation time The proportion of abnormal brain regions at that time Representing the The individuals to be evaluated will be assessed during the second evaluation period. The proportion of abnormal brain regions at different ages. In large-scale population brain development atlases, the derivative of the average trend of the proportion of abnormal brain regions with age with respect to age. This indicates that large-scale brain development atlases are available at different ages. The rate of natural change in the vicinity, therefore for the first The degree of deviation of an individual's current age x from the natural rate of change of the population is defined as follows: If this value is greater than 0, it indicates that the individual's abnormality is spreading faster than the average level of healthy individuals of the same age, and even if they are currently in the normal stratification, there is a potential risk. This is for a future assessment. After introducing corrections for herd trends, it can be predicted as follows: .

[0145] (Formula 36);

[0146] (Formula 37);

[0147] (Formula 38).

[0148] After obtaining the individual's future cognitive risk prediction results, the system first prospectively stratifies the individual based on the risk prediction levels within different future time windows. If the prediction results show that the individual has a high probability of entering a high-risk range in the short term (e.g., within one year), the system classifies them as a high-risk individual and automatically triggers an early warning mechanism, including generating a risk alert report, reminding doctors or management platforms to pay attention, and suggesting shortening the retesting cycle. If the prediction results show that the risk is within the current normal range, but the upward trend deviates significantly from the average change trajectory of healthy people of the same age, the system marks them as an abnormal trend individual. Even if they have not yet reached the high-risk threshold, the system recommends taking intervention strategies in advance, thereby achieving true early warning.

[0149] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or physical entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, or a tablet computer, or any combination of these devices.

[0150] The above provides a detailed description of the brain development feature evolution analysis system and method based on multimodal brain imaging data provided by this invention. Specific examples have been used to illustrate the principles and implementation methods of this invention. The above description of the embodiments is only for the purpose of helping to understand the concept of this invention and should not be construed as a limitation on the scope of protection of this invention.

Claims

1. A brain development feature evolution analysis system based on multimodal brain imaging data, characterized in that, include: The brain development atlas construction module is used to acquire demographic information and multimodal brain imaging data from healthy individuals of multiple ages. Based on multimodal brain imaging data of people of multiple ages, preprocessing, brain region segmentation and feature extraction are performed to extract structural features, connectivity features and activation features. Combined with demographic information of people of multiple ages, a population brain development atlas reflecting the evolution of brain structure, function and connectivity patterns with age is constructed. The data acquisition module is used to acquire demographic information and multimodal brain imaging data of the individual to be evaluated. The multimodal brain imaging data of the individual to be evaluated is consistent with the multimodal brain imaging data acquired by the brain development map construction module. The data processing module is used to preprocess, segment, and extract features from the multimodal brain data acquired by the data acquisition module using the same processing method as the brain development map construction module, so as to obtain the individual multimodal brain feature vector of the individual in a unified brain region spatial coordinate system. The evolutionary matching assessment module is used to map the individual's multimodal brain feature vector to the corresponding population brain development atlas space based on the demographic information of the individual to be assessed. By calculating the deviation of the individual's multimodal brain feature vector from the evolutionary trajectory of the peer group, and combining the stability weights of each modality feature, a fusion deviation index is obtained. The fusion deviation index is used to quantitatively assess the degree of evolution of the individual's brain development state relative to the population baseline. Based on the fusion deviation index, abnormal brain regions are identified, the proportion of abnormal brain regions and the risk change rate are calculated, and brain development feature evolution analysis is realized, outputting the brain development feature evolution analysis results.

2. The brain development feature evolution analysis system according to claim 1, characterized in that, The brain development atlas constructed by the brain development atlas construction module includes a structural development atlas, a connectivity development atlas, and an activation development atlas. Among them, the structural development atlas is constructed based on brain region-level structural feature extraction and is used to establish a distribution model of brain region volume, gray matter volume, cortical thickness and cortical surface area as a function of age; The connectivity development atlas is constructed based on spontaneous brain functional signals in a resting state. By utilizing the time-series correlation of these spontaneous brain functional signals, a distribution model of the connection strength between brain networks and the overall connectivity level of the whole brain as a function of age is established. The activation development atlas is constructed based on brain functional response signals under task conditions. By extracting activation features, a distribution model of the degree of activation of brain regions and the similarity of multi-voxel patterns under specific tasks is established as a function of age.

3. The brain development feature evolution analysis system according to claim 1, characterized in that, The multimodal brain imaging data includes structural magnetic resonance imaging (fMRI) data, resting-state functional magnetic resonance imaging (fMRI) data, fMRI data for specific cognitive tasks, and functional near-infrared brain imaging (FIN) data.

4. The brain development feature evolution analysis system according to claim 1, characterized in that, The stability weights in the evolutionary matching evaluation module are determined based on the fluctuation range of each modality feature in healthy individuals of the same age. The smaller the fluctuation range, the greater the weight; the larger the fluctuation range, the smaller the weight. Furthermore, the sum of the weights of the same brain region under different modalities is 1.

5. The brain development feature evolution analysis system according to claim 1, characterized in that, When the evolutionary matching evaluation module identifies abnormal brain regions, it uses a preset abnormality discrimination threshold. When the absolute value of the fusion deviation index of a certain brain region of an individual reaches or exceeds the threshold, the brain region is determined to be an abnormal brain region, and the proportion of abnormal brain regions to all brain regions is calculated.

6. A method for analyzing the evolution of brain development features based on multimodal brain imaging data, characterized in that, Includes the following steps: Step 1: Obtain demographic information and multimodal brain imaging data of healthy individuals across multiple age groups; Based on multimodal brain imaging data of people of multiple ages, preprocessing, brain region segmentation and feature extraction are performed to extract structural features, connectivity features and activation features. Combined with demographic information of people of multiple ages, a population brain development atlas reflecting the evolution of brain structure, function and connectivity patterns with age is constructed. Step 2: Obtain demographic information and multimodal brain imaging data of the individual to be evaluated. The multimodal brain imaging data of the individual to be evaluated is consistent with the data type of multimodal brain imaging data of healthy individuals of multiple age groups. Step 3: Using the same processing method as for multimodal brain imaging data of healthy individuals of multiple ages, preprocess, segment, and extract features from the multimodal brain data of the individual to be evaluated to obtain the individual's multimodal brain feature vector in a unified brain region spatial coordinate system; Step 4: Based on the demographic information of the individual to be evaluated, the individual's multimodal brain feature vector is mapped to the corresponding population brain development atlas space. By calculating the deviation of the individual's multimodal brain feature vector from the evolutionary trajectory of the peer group, and combining the stability weights of each modality feature, a fusion deviation index is obtained. The degree of evolution of the individual's brain development state relative to the population baseline is quantitatively assessed based on the fusion deviation index. Abnormal brain regions are identified based on the fusion deviation index, and the proportion of abnormal brain regions and the risk change rate are calculated to realize the brain development feature evolution analysis and output the brain development feature evolution analysis results.

7. The method for analyzing the evolution of brain development features according to claim 6, characterized in that, Step 1 involves extracting the structural features by: segmenting the multimodal brain imaging data into brain regions using a cortical segmentation atlas; for each brain region, counting the number of voxels belonging to that region and determining the voxel volume of that region by combining the physical volume of each voxel; weighting the voxels within the brain region using a gray matter probability map to obtain the gray matter volume of that region; calculating the average Euclidean distance from the set of cortical vertex coordinates to the adjacent white matter coordinate sets within the brain region to obtain the cortical thickness of that region; and counting the area of ​​the cortical surface corresponding to the brain region to obtain the cortical surface area of ​​that region.

8. The method for analyzing the evolution of brain development features according to claim 6, characterized in that, The population brain development atlas mentioned in step 1 includes a connectivity development atlas. The construction process of the connectivity development atlas includes: dividing the resting-state multimodal brain imaging data into multiple brain networks using a cortical segmentation atlas; calculating the Pearson correlation coefficient of the BOLD signal time series between any two brain networks; generating a functional connectivity matrix; and calculating the sum of the connectivity strengths of a single brain network with all other brain networks and the average connectivity strength among all brain networks in the whole brain based on the functional connectivity matrix, and establishing a distribution model of the above connectivity strengths changing with age.

9. The method for analyzing the evolution of brain development features according to claim 6, characterized in that, The population brain development atlas mentioned in step 1 includes an activation development atlas. The construction process of the activation development atlas includes: using a cortical segmentation atlas to segment the task-state multimodal brain imaging data into brain regions, statistically pooling the voxel signals in each brain region, calculating the degree of neural activity activation related to a specific task, calculating the similarity of multi-voxel patterns in the brain region, and establishing a distribution model of the degree of brain region activation and the similarity of multi-voxel patterns with age.

10. The method for analyzing the evolution of brain development features according to claim 6, characterized in that, The specific method for identifying abnormal brain regions, calculating the proportion of abnormal brain regions and the risk change rate based on the fusion deviation index in step 4, and realizing the evolutionary analysis of brain development characteristics is as follows: A preset abnormality detection threshold is set, and the absolute value of the fusion deviation index of each brain region of the individual to be evaluated is compared with the threshold to screen out abnormal brain regions and form a set of abnormal brain regions. The proportion of abnormal brain regions is calculated by comparing the number of brain regions in the abnormal brain region set with the total number of brain regions in the whole brain. Based on the proportion of abnormal brain regions in at least two individual assessments and the corresponding assessment time intervals, the risk change rate is calculated, and combined with the age-average change trend of the proportion of abnormal brain regions in the population brain development atlas, the degree of deviation of the individual from the changes of healthy people of the same age is obtained. Mapping the proportion of abnormal brain regions to the norm distribution of healthy people of the same age, and determining the risk percentile of the same age to achieve cognitive risk stratification; By combining the rate of change of risk, the degree of deviation of change, and the evolutionary trend of the population, the proportion of abnormal brain regions in an individual in the future can be predicted, the level of future cognitive risk can be determined, and corresponding early warning measures can be taken.

Images