Cerebral region homology cross-species comparison method based on grey matter brain region and fiber bundles

Abstract

The invention discloses a brain region homology cross-species comparison method based on a grey matter brain region and a fiber bundle. The method comprises the following steps: firstly, acquiring cross-species brain image data and preprocessing; extracting a region of interest from the standard template and performing brain region division to obtain region of interest division results of different species; constructing a connection fingerprint of each sub-region and a grey matter brain region by using probability fiber tracking based on the whole brain; using voxel-based probabilistic fiber tracking to construct a connection fingerprint of each subarea and a fiber bundle; respectively obtaining effective grey matter brain region connection fingerprints and fiber bundle connection fingerprints in the species on the group level; based on the connection fingerprints, calculating similarity or difference measurement between subregions of different species to obtain cross-species homology scores; the method is verified based on the consistency of the connection mode from the homologous subregion to the whole brain function. From the perspective of brain map drawing and evolution, a new thought and a new method for cross-species homology comparison of a specific brain region subregion are provided by taking the grey matter brain region and subcortical fiber bundles as feature input in a synergistic manner.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of neuroimaging technology, and more particularly to a method for comparing subregional homology of a region of interest using cross-species cortical brain regions, subcutaneous nuclei, and fiber bundles, and more particularly to a cross-species comparison method of brain region homology based on gray matter brain regions and fiber bundles. Background Art

[0002] Brain diseases are diseases that affect the brain, spinal cord, or nervous system, including stroke, Parkinson's disease, Alzheimer's disease, meningitis, etc. These diseases may lead to cognitive decline, movement disorders, emotional problems, etc., and in severe cases, they may even endanger the patient's life. Therefore, the prevention and treatment of brain diseases are crucial to protecting people's health and improving their quality of life. Early detection and intervention of brain diseases and the adoption of reasonable treatment measures can effectively alleviate the patient's pain, delay the progression of the disease, and improve the patient's quality of life.

[0003] Research has found that many brain diseases are closely associated with specific subregions of the brain. Different brain regions have different functional characteristics, and these regions can be further divided into multiple subregions, each of which may play a specific role in specific cognitive, emotional or motor functions.

[0004] Animal models of brain diseases play an irreplaceable role in basic research and preclinical research. They can help us understand disease mechanisms, including pathophysiological changes, neuronal damage and repair processes, etc.; simulate the characteristics of human diseases and understand the occurrence and development of diseases; evaluate treatment effects and test the safety and effectiveness of new drugs and treatments. Due to ethical considerations, in some cases, it is unethical or infeasible to conduct experiments directly on humans. The use of animal models allows necessary scientific research to be conducted under the premise of following ethical norms.

[0005] Cross-species research on the brain is crucial for a deeper understanding of the mechanisms of brain structure, function, and behavior, as well as for promoting the development of the field of neuroscience. Studies have shown that during the evolution of primates, changes in the cortical volume of brain regions and changes in functional connectivity between brain regions may provide a structural basis for human-specific functions such as language and tool use. Structural and functional analysis of primates is key to understanding the development of human cognition and behavior. As close relatives of humans on the evolutionary tree, chimpanzees and macaques are important references for cross-species research on the brain, and are of great significance for exploring the homology of brain regions and subregions corresponding to humans.

[0006] Typically, cross-species comparative brain region homology studies need to first use a priori homologous brain regions between species as a reference. Currently, a large number of studies have disclosed widely recognized homologous brain regions between different species. However, there is currently a lack of strict standards for how to subdivide specific brain regions and compare the homology of their subregions across species. In addition, some studies have shown that white matter fiber bundles can be used as a standard for homology comparison, but few studies have compared the structural connections of subregions with gray matter brain regions and with subcortical fiber bundles as overall characteristics. Summary of the invention

[0007] To solve the above problems, the technical solution of the present invention is as follows: The present invention aims to propose a cross-species comparison method of brain region homology based on gray matter brain regions and fiber bundles, including obtaining standard brain magnetic resonance data of different species and preprocessing them, extracting and dividing brain regions with regions of interest, calculating the connection fingerprints of each subregion with gray matter brain regions, calculating the connection fingerprints of each subregion with fiber bundles, cross-species homology comparison and homology verification.

[0008] The method comprises obtaining standard brain magnetic resonance data of different species and preprocessing the data, respectively preprocessing T1w images, fMRI images and DTI images of the standard brain magnetic resonance data of different species to obtain preprocessed structural image data, functional image data and diffusion weighted image data;

[0009] The region of interest extraction and brain region division are to extract corresponding regions of interest (ROI) from standard brain templates of different species and divide the regions using probabilistic fiber tracking and clustering methods to obtain brain region division results of regions of interest of different species;

[0010] The connection fingerprint between each subregion and the gray matter brain region is calculated, the homology maps corresponding to the cortical brain regions and subcortical nuclei in different species are constructed, the structural connection between the subregions of the regions of interest of different species and the whole brain is obtained by probabilistic fiber tracking, the effective connections within the species are screened at the group level, and the connection fingerprint between each subregion and the gray matter brain region is extracted;

[0011] The connection fingerprint between each subregion and the fiber bundle is calculated, the fiber bundles common to different species are extracted as templates, a fiber tracking map based on voxels of each subregion is obtained based on probabilistic fiber tracking, the connection probability value between each subregion and each fiber bundle in the region of interest is calculated, effective connections within the species are screened at the group level, and the connection fingerprint between each subregion and the fiber bundle is extracted;

[0012] The cross-species homology comparison uses the connection fingerprints of each subregion of the region of interest with the gray matter brain region and the connection fingerprints with the fiber bundles as feature inputs, calculates the similarity or difference measures between the subregions of different species and scores them, and the score is used as a homology comparison criterion.

[0013] The cross-species homology verification calculates the functional connections of each subregion of different species to the whole brain based on functional image data, compares the whole-brain functional connection patterns of homologous subregions, and whether the functional connection patterns between species are consistent is used as a verification of the reliability of this method.

[0014] Furthermore, the method acquires standard brain magnetic resonance data of different species and preprocesses the data, using standard brain magnetic resonance data of different species, including structural T1w images, functional fMRI images and diffusion-weighted images DTI images; the data preprocessing process specifically includes denoising, registration, brain shell removal, brain tissue segmentation, etc. for T1w structural images; motion correction, spatial standardization, smoothing, signal extraction, normalization, etc. for fMRI functional images; denoising, registration, correction, brain tissue extraction and diffusion tensor fitting for diffusion magnetic resonance data DTI images;

[0015] Furthermore, the extraction of the corresponding region of interest and the division of the brain regions mainly involves extracting the corresponding region of interest (ROI) from the standard brain templates of different species and dividing the region. The process mainly includes:

[0016] A) Determine the region of interest (ROI) to be studied and extract the corresponding region from the standard brain template of different species;

[0017] B) Align the extracted region of interest with the diffusion space, create a template mask of the seed point of the region of interest, and use probabilistic fiber tracking to estimate the connection probability for each voxel in the mask to obtain the connection matrix of each voxel with the whole brain voxels;

[0018] C) Downsample the whole-brain connection matrix at each voxel in the region of interest to form the voxel connection matrix A within the ROI m×n (The number of voxels in the ROI is m, and the number of voxels in the whole brain is n);

[0019] D) Calculate the cross-correlation matrix A between the whole-brain connectivity matrices of all voxels in the mask m×m , and used for automatic division of brain regions;

[0020] E) using a clustering method on the above correlation matrix, clustering voxels with higher similarity together to form 2-10 groups and calculating clustering indicators (silhouette coefficient SC, error square sum SSE) of different groups, selecting the optimal solution k of the clustering indicator as the division of the region of interest in different species, and obtaining k sub-region division results of the region of interest of different species respectively;

[0021] Furthermore, in the calculation of the connection fingerprint between each sub-region and the gray matter brain region, the main process includes:

[0022] A) Based on the existing standard brain maps, a homologous map of gray matter brain regions (including cortical brain regions and subcortical nuclei) is constructed in different species;

[0023] B) Use probabilistic fiber tracking based on the whole brain for each subregion of different species, standardize, denoise, remove false positives, etc., to obtain the structural connection between each subregion of different species and the whole brain;

[0024] C) The whole brain structural connectivity is used as a template for the homology map, and the average value of the structural connectivity from the subregions to the gray matter brain region is taken as the connection strength from each subregion of the region of interest to the brain region, so as to obtain the connection fingerprint from each subregion of the region of interest to the cortical brain region and subcortical nuclei at the individual level;

[0025] D) averaging the above individual-level connection fingerprints to obtain a group-based connection fingerprint for each subregion and gray matter brain region;

[0026] E) Perform group-level screening on the connection fingerprints between the subregions and gray matter brain regions at the individual level, obtain valid connections within the species as features, and extract the connection fingerprints between each subregion and gray matter brain region based on the screening results;

[0027] Furthermore, in the calculation of the connection fingerprint between each sub-region and the fiber bundle, the main process includes:

[0028] A) Based on the fiber bundles shared by different species as the basis for cross-species comparison, each fiber bundle is reconstructed for each species separately, and the mask of each fiber bundle is constructed in MNI space and then registered to the individual to obtain the mask of the individual fiber bundle;

[0029] B) Probabilistic fiber tracking is used to obtain the voxel-based fiber tracking map of each sub-region. For each sample, all voxels corresponding to each fiber bundle are extracted based on the mask, and the connection fingerprints between each sub-region of the region of interest and each fiber bundle at the individual level are averaged.

[0030] C) averaging the above individual-level connection fingerprints to obtain group-based connection fingerprints of each subregion and subcortical fiber bundle;

[0031] D) Perform group-level screening on the connection fingerprints between subregions and fiber bundles at the individual level, select fiber bundles that have effective connections in different species as features, and extract the connection fingerprints between each subregion and fiber bundle based on the screening results;

[0032] Furthermore, in the cross-species homology comparison, the connection fingerprints of each sub-region of the above-mentioned region of interest with the gray matter brain region and the connection fingerprints with the fiber bundles are used as feature inputs, and the similarity or difference metrics (including but not limited to Euclidean distance, Manhattan distance, cosine distance and KL divergence) between the sub-regions of different species are calculated. After normalizing the Euclidean distance, Manhattan distance, cosine distance and KL divergence respectively, the reciprocal sum of the four metrics is taken as the scoring score, and the scoring score is used as the homology comparison criterion.

[0033] Furthermore, in the cross-species homology verification, the normalized whole-brain time series is obtained based on the functional image, the time series corresponding to each sub-region of the region of interest is extracted, and the functional connection of each sub-region to the whole brain of different species is calculated, and the whole-brain functional connection patterns of homologous sub-regions are compared. Whether the functional connection patterns between species are consistent is used as a verification of the reliability of this method. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The drawings described herein are used to provide further understanding of the present application and constitute a part of the present application. In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below.

[0035] Figure 1 It is a schematic diagram of a process of a cross-species comparison method of brain region homology based on gray matter brain regions and fiber bundles proposed in one embodiment of the present invention;

[0036] Figure 2 is a schematic diagram of a process of extracting a region of interest and dividing brain regions according to an embodiment of the present invention;

[0037] Figure 3 It is a schematic diagram of a flow chart of calculating the connection fingerprints between subregions and gray matter brain regions proposed in one embodiment of the present invention.

[0038] Figure 4 It is a schematic diagram of a flow chart of calculating the connection fingerprint between subregions and subcortical fiber bundles according to an embodiment of the present invention. DETAILED DESCRIPTION

[0039] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0040] Unless otherwise defined, the technical terms or scientific terms involved in this application shall have the general meaning understood by people with general skills in the technical field to which this application belongs. The words such as "one", "a", "the", "these" and the like in this application do not represent quantitative restrictions, and they can be singular or plural. The terms "include", "comprise", "have" and any variants thereof involved in this application are intended to cover non-exclusive inclusions; for example, a process, method and system, product or device comprising a series of steps or modules (units) is not limited to the listed steps or modules (units), but may include unlisted steps or modules (units), or may include other steps or modules (units) inherent to these processes, methods, products or devices. The "multiple" involved in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there can be three relationships, for example, "A and / or B" can represent: A exists alone, A and B exist at the same time, and B exists alone. Usually, the character " / " indicates that the objects associated before and after are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific ordering of the objects.

[0041] In general, this embodiment provides a cross-species comparison method of brain region homology based on gray matter brain regions and fiber bundles. Figure 1This is a flowchart of the cross-species comparison method of brain region homology based on gray matter brain regions and fiber bundles of this embodiment, and the process includes: data acquisition and preprocessing, extraction and division of brain regions of interest, calculation of connection fingerprints between subregions and gray matter brain regions, calculation of connection fingerprints between subregions and subcortical fiber bundles, cross-species homology comparison and homology verification. Among them, data acquisition and preprocessing, the cross-species data used in this example are standard brain magnetic resonance data of humans and chimpanzees, including structural T1w images, functional fMRI images and diffusion-weighted images DTI images, and the T1w images, fMRI images and diffusion-weighted image data are preprocessed respectively. Region of interest extraction and division of brain regions, extract the corresponding region of interest (ROI) from the standard brain templates of humans and chimpanzees and divide the region to obtain the division results of the regions of interest of the two species. The connection fingerprints between subregions and gray matter brain regions were calculated by probabilistic fiber tracking based on the whole brain. After standardization, denoising, and false positive removal, the structural connections between subregions and the whole brain of the two species' regions of interest were obtained. The connection fingerprints between subregions and gray matter brain regions at the individual level were screened at the group level to extract the effective connection fingerprints between each subregion and the gray matter brain region. The connection fingerprints between subregions and subcortical fiber bundles were calculated by probabilistic fiber tracking to reconstruct each fiber bundle of different species and obtain the voxel-based fiber tracking map of each subregion. The connection probability values ​​between each subregion and each fiber bundle in the region of interest were obtained by averaging all voxels in each fiber bundle. The fiber bundles with effective connections in both species were selected as features to extract the effective connection fingerprints between each subregion and the fiber bundle. For cross-species homology comparison, the connection fingerprints between each subregion and the gray matter brain region and the connection fingerprints with the fiber bundles were used as feature inputs to calculate the similarity or difference measures between each subregion of humans and chimpanzees, and the similarity measures between each subregion in the region of interest of the two species were scored.

[0042] Through the above-mentioned cross-species comparison method of brain region homology based on gray matter brain regions and fiber bundles, the structural image data, functional image data and diffusion-weighted item data in the multimodal brain imaging data of the two species were preprocessed according to data acquisition and preprocessing, and the brain regions were extracted and divided by the region of interest to obtain the division results of the regions of interest of the two species. Then, the connection fingerprints of subregions and gray matter brain regions were calculated, and the connection fingerprints of each subregion in the region of interest to the cortical brain region and subcortical nuclei were obtained by probabilistic fiber tracking, and the connection fingerprints of subregions and gray matter brain regions were screened. Then, the connection fingerprints of subregions and subcortical fiber bundles were calculated to obtain the connection fingerprints of each subregion and fiber bundle, and the connection fingerprints of subregions and fiber bundles were screened. Finally, the cross-species homology comparison was used to calculate the cross-species homology score of each subregion, and the cross-species brain region homology comparison of gray matter brain regions and fiber bundles was realized, providing new ideas and methods for cross-species brain region homology comparison.

[0043] In some of the embodiments, the brain magnetic resonance data of different species are sourced from: brain magnetic resonance data of 40 healthy humans and brain magnetic resonance data of 46 healthy chimpanzees, including structural image T1 data, functional image fMRI data and diffusion weighted image DTI data.

[0044] In some embodiments, a method for extracting regions of interest and dividing brain regions is provided. Figure 2 : is a flow chart of the method for extracting regions of interest and dividing brain regions in this embodiment, including:

[0045] Region of interest extraction: determine the region of interest to be studied and extract the corresponding region from the standard brain templates of humans and chimpanzees respectively;

[0046] The extracted region of interest is aligned with the diffusion space to create a template mask of the seed point of the region of interest. For each voxel in the mask, probabilistic fiber tracking is used to estimate the connection probability to obtain the connection matrix of each voxel with the whole brain voxels.

[0047] The whole-brain connection matrix at each voxel in the region of interest is downsampled to form the voxel connection matrix A within the ROI m×n (The number of voxels in the ROI is m, and the number of voxels in the whole brain is n);

[0048] Calculate the cross-correlation matrix A between the whole-brain connectivity matrices of all voxels in the mask m×m , and used for automatic division of brain regions;

[0049] The above correlation matrix is ​​clustered using a clustering method to cluster voxels with higher similarity together to form 2-10 groups and calculate the clustering index silhouette coefficient SC and error sum of squares SSE of different groups. The k value with the highest clustering index score is selected as the division of the region of interest in different species, and the k sub-region division results of the region of interest of different species are obtained respectively;

[0050] In some embodiments, a method for calculating the connection fingerprint between subregions and gray matter brain regions is provided. Figure 3 : is a flow chart of the method for calculating the fingerprint of the sub-region and gray matter brain region connection of this embodiment, including:

[0051] Homologous atlas construction, extracting brain atlas based on MNI standard space as template, using Desikan-KillianyAtlas for human brain and CY29 template for chimpanzee brain. Based on the correspondence between these two templates in different species, homologous atlases based on the existing standard space brain templates in different species are constructed.

[0052] The sub-region to whole-brain structural connections were constructed by using probabilistic fiber tracking based on the whole brain for each sub-region of humans and chimpanzees, and then normalized, denoised, and removed false positives to obtain the structural connections from each sub-region of interest in each sample of different species to the whole brain.

[0053] The connection fingerprints from subregions to gray matter brain regions are calculated. The homologous map is used as a template for the structural connections from each subregion to the whole brain. The average value of the structural connection in each region is taken as the connection strength from the subregion to the region. The connection fingerprints from each subregion to the gray matter brain region in the region of interest are obtained. The above-mentioned individual-level connection fingerprints are averaged to obtain the group-based connection fingerprints between subregions and gray matter brain regions.

[0054] Group-level screening: statistics were performed on the connection fingerprints from each subregion to the cortical brain region of humans and chimpanzees at the group level, and the connection fingerprints at the individual level were binarized. The average value of the connection fingerprints at the group level was taken to obtain the group probability, and the connection fingerprints with a species group probability greater than 0.6 were selected as features;

[0055] Effective connection feature extraction, based on the above-screened cortical brain area connection fingerprints, extracts the connection fingerprints from each sub-region to the gray matter brain area based on the group for subsequent analysis:

[0056] In some embodiments, a method for calculating the probability of subcortical fiber bundle connection is provided. Figure 4 : is a flow chart of the method for calculating the probability of subcortical fiber bundle connection of this embodiment, comprising:

[0057] The common fiber bundles are constructed based on the common fiber bundles of different species as the basis for cross-species comparison. Each fiber bundle is reconstructed for different species separately. The mask of each fiber bundle is constructed in the MNI space and then registered to the individual to obtain the mask of the individual fiber bundle.

[0058] The connections from subregions to common fiber bundles were constructed by using whole-brain probabilistic fiber tracking for each subregion of humans and chimpanzees, with the number of streamlines for humans and chimpanzees being 5000. The connections from subregions to the whole brain were obtained, and a threshold of 0.05% was used to remove false positives, to obtain voxel-based fiber tracking maps for each subregion. All voxels corresponding to each fiber bundle were extracted based on the mask for each sample of each species, and the connection fingerprints between each subregion and each fiber bundle in the region of interest were averaged. The above-mentioned individual-level fiber bundle connection fingerprints were averaged to obtain group-based subregion-fiber bundle connection fingerprints.

[0059] Group-level screening: At the group level, the connection fingerprints from each subregion to the common fiber bundles of humans and chimpanzees were counted separately, the connection fingerprints at the individual level were binarized, the average value of the connection fingerprints at the group level was taken to obtain the group probability, and the connection fingerprints with species group probability greater than 0.6 were screened as features.

[0060] The effective connection feature extraction is based on the above-screened fiber bundle connection fingerprints. Based on the above-screened connections, the connection fingerprints from each sub-region to the fiber bundle based on the group are extracted for subsequent analysis.

[0061] In some of the embodiments, a cross-species homology comparison method is provided, and specific comparison criteria include: calculating the Euclidean distance, Manhattan distance, cosine distance and KL divergence between subregions of regions of interest of different species, normalizing the Euclidean distance, Manhattan distance, cosine distance and KL divergence, and taking the reciprocals of the four metrics as the scoring score. The smaller the distance, the higher the score, indicating that the homology between a subregion of the species and a subregion of another species is higher.

[0062] In some of the embodiments, a cross-species homology verification method is provided, specifically including: obtaining normalized whole-brain time series from pre-processed functional images of humans and chimpanzees, extracting time series corresponding to each sub-region of the region of interest, calculating the functional connection from each sub-region to the whole brain of different species, comparing the whole-brain functional connection patterns of homologous sub-regions, and verifying whether the functional connection patterns between species are consistent as a verification of the reliability of this method.

[0063] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A cross-species comparison method of brain region homology based on gray matter brain regions and fiber bundles, characterized in that: The steps include: Acquire standard brain magnetic resonance data of different species and preprocess them to obtain preprocessed structural image data, functional image data and diffusion-weighted image data; The corresponding regions of interest (ROI) were extracted from the standard brain templates of different species and the probabilistic fiber tracking and clustering methods were used to divide the regions to obtain the division results of the regions of interest. Construct homologous maps of cortical brain regions and subcortical nuclei in different species, apply probabilistic fiber tracking to obtain the structural connections between each subregion and the whole brain, screen effective connections within species at the group level, and extract the connection fingerprints between each subregion and gray matter brain region; The fiber bundles common to different species were extracted as templates, and the voxel-based fiber tracking map of each sub-region was obtained based on probabilistic fiber tracking. The connection probability values ​​between each sub-region and each fiber bundle in the region of interest were calculated, and the probability values ​​of fiber bundles within the species were counted at the group level to extract the connection fingerprint between each sub-region and the fiber bundle. The gray matter brain regions and fiber bundle connection fingerprints of each subregion of the region of interest are used as feature inputs, and the similarity or difference measures between the subregions of different species are calculated and scored. The score is used as a criterion for homology comparison. The functional connections between each subregion and the whole brain were calculated based on the functional image data, and the consistency of the whole-brain functional connection patterns of homologous subregions in different species was used to verify the reliability of this method.

2. A cross-species comparison method of brain region homology based on gray matter brain regions and fiber bundles according to claim 1, characterized in that: The method obtains standard brain magnetic resonance data of different species and preprocesses them, using standard brain magnetic resonance data of different species, including structural T1w images, functional fMRI images and diffusion weighted images DTI images; the data preprocessing process specifically includes denoising, registration, brain shell removal, brain tissue segmentation, etc. for T1w structural images; motion correction, spatial standardization, smoothing, signal extraction, normalization, etc. for fMRI functional images; denoising, registration, correction, brain tissue extraction and diffusion tensor fitting for diffusion magnetic resonance data DTI images.

3. A cross-species comparison method of brain region homology based on gray matter brain regions and fiber bundles according to claim 1, characterized in that: The extraction of the corresponding region of interest and the division of the brain regions mainly involves extracting the corresponding region of interest (ROI) from the standard brain templates of different species and dividing the region. The process mainly includes: S31, region of interest extraction, determining the region of interest (ROI) to be studied, and extracting the corresponding regions from the standard brain templates of different species; S32, aligning the extracted region of interest with the diffusion space, creating a template mask of the seed point of the region of interest, and using probabilistic fiber tracking to estimate the connection probability for each voxel in the mask to obtain a connection matrix between each voxel and the whole brain voxels; S33, downsampling the whole brain connection matrix at each voxel in the region of interest to form the voxel connection matrix A within the ROI m×n (The number of voxels in the ROI is m, and the number of voxels in the whole brain is n); S34. Calculate the cross-correlation matrix A between the whole-brain connection matrices of all voxels in the mask m×m , and used for automatic brain region segmentation; S35. Using a clustering method on the above correlation matrix, voxels with higher similarity are clustered together to form different groups and the clustering indices of these groups are calculated. The optimal solution k of the clustering index is selected as the division of the region of interest in different species, and the k division results of the region of interest of different species are obtained respectively.

4. A cross-species comparison method of brain region homology based on gray matter brain regions and fiber bundles according to claim 1, characterized in that: The gray matter brain area connectivity calculation includes: S41. Construct homologous maps of gray matter brain regions (including cortical brain regions and subcutaneous nuclei) in different species based on existing standard brain maps; S42, using probabilistic fiber tracking based on the whole brain for each subregion of different species, standardizing, denoising, removing false positives, etc., to obtain the structural connection between each subregion of different species and the whole brain; S43, using the homology map as a template for the whole brain structural connection, taking the average value of the structural connection from the sub-region to the gray matter brain region of the individual as the connection strength from each sub-region of the region of interest to the brain region, and obtaining the connection fingerprint from each sub-region of the region of interest to the cortical brain region and subcutaneous nucleus at the individual level; S44, averaging the individual-level connection fingerprints to obtain a group-based connection fingerprint between each subregion and the gray matter brain region; S45. Perform group-level screening on the gray matter brain region connection fingerprints at the individual level to obtain effective connections within the species as features, and extract the connection fingerprints between each subregion and the gray matter brain region based on the screening results.

5. The cross-species comparison method of brain region homology based on gray matter brain regions and fiber bundles according to claim 1, characterized in that: The subcortical fiber bundle connection probability calculation includes: S51. Based on the fiber bundles shared by different species as the basis for cross-species comparison, each fiber bundle is reconstructed for each species, and the mask of each fiber bundle is constructed in MNI space and then registered to the individual to obtain the mask of the individual fiber bundle. S52, using probabilistic fiber tracking to obtain a voxel-based fiber tracking map for each sub-region, extracting all voxels corresponding to each fiber bundle based on the mask for each sample, and averaging them to obtain the connection fingerprint between each sub-region of the region of interest and each fiber bundle at the individual level; S53, averaging the individual-level connection fingerprints to obtain group-based connection fingerprints of each subregion and the subcortical fiber bundle; S54. Perform group-level screening on the fiber bundle connection fingerprints at the individual level, select fiber bundles that are effectively connected in different species as features, and extract the connection fingerprints between each subregion and the fiber bundle based on the screening results.

6. The method of screening individual level connection fingerprints at group level according to claim 4 and claim 5, characterized in that: At the group level, the connection fingerprints of each sub-region of different species were counted separately, and the connection fingerprints at the individual level were binarized. The average value of the connection fingerprints at the group level was taken to obtain the group probability, and the connection fingerprints with species group probability greater than 0.6 were selected as features.

7. A cross-species comparison method of brain region homology based on gray matter brain regions and fiber bundles according to claim 1, characterized in that: The cross-species homology comparison criterion uses the connection fingerprints of each sub-region of the region of interest obtained in claim 6 with the gray matter brain region and the connection fingerprints with the fiber bundle as feature inputs, calculates the similarity or difference measures between the sub-regions of different species and scores them, and the score is used as the homology comparison criterion.

8. The cross-species comparison method of brain region homology based on gray matter brain regions and fiber bundles according to claim 1, characterized in that: Verification of cross-species homology: Based on functional image data, the functional connections of each subregion of different species to the whole brain are calculated, and the whole-brain functional connection patterns of homologous subregions of different species are compared. Whether the functional connection patterns between species are consistent serves as a verification of the reliability of this method.

Images