Method for visualizing and quantifying glioma-induced brain network remodeling based on fMRI

Abstract

The present application relates to medical image analysis and brain network research technical field, specifically to glioma induced brain network remodeling visualization and quantitative analysis method based on fMRI. The method comprises obtaining patient fMRI and structural MRI data and preprocessing, excluding tumor area by lesion mask registration strategy, reducing quality effect interference; dividing tumor core area, peritumoral abnormal area and normal brain area; registering Yeo-17 network template to individual brain area to realize mapping; defining tumor core area as independent network unit, and 17 normal networks to form a new set; calculating whole brain voxel and network functional connection strength, and determining functional connection voxel according to threshold; quantifying intratumoral function proportion RIFR and peritumoral connection proportion RPTR, and generating visualization atlas. The present application accurately maps individual brain function network, overcomes tumor heterogeneity interference, provides repeatable quantitative index, and provides reliable imaging analysis tool for brain glioma function protection and clinical research.

Description

Technical Field

[0001] This invention relates to the field of medical image analysis and brain network research technology, and more specifically to a method for visualizing and quantitatively analyzing glioma-induced brain network remodeling based on fMRI. fMRI stands for functional magnetic resonance imaging. Background Technology

[0002] Gliomas are the most common primary malignant tumors of the central nervous system, exhibiting significant spatial heterogeneity and invasiveness in their biological characteristics. These features mean that the tumor's impact on brain tissue is not limited to the structural level but can also cause changes in long-range functional connectivity. Studies have shown that functional activities associated with normal brain networks exist within tumor regions, posing new challenges to brain region delineation based on functional connectivity.

[0003] With the development of functional magnetic resonance imaging (fMRI) and personalized brain network analysis methods, researchers have attempted to perform more refined functional subdivision and network feature extraction of glioma-related regions. However, most existing methods rely on universal templates for brain network mapping, making it difficult to adapt to structural and functional variations caused by tumor heterogeneity. This approach lacks mechanisms for handling the specific characteristics of lesion regions, resulting in insufficient accuracy in functional subdivision and an inability to accurately quantify the functional connectivity strength between intratumoral or peritumoral regions and the whole-brain network. Furthermore, the space-occupying effect caused by tumors often interferes with the image registration process, causing significant deviations in the spatial standardization stage of traditional registration strategies, affecting the accuracy of subsequent brain network mapping. In particular, when performing functional connectivity strength calculations and visualization analysis based on brain atlases, the lack of sufficient consideration of individual structural differences and lesion-induced brain network changes limits the applicability and reproducibility of current fMRI data analysis tools at the individual lesion level.

[0004] Therefore, there is an urgent need to develop a technical method that can achieve high-precision brain network mapping and quantitative analysis of functional connectivity based on individual brain networks and combined with lesion masking strategies. This method needs to overcome the technical obstacles posed by factors such as spatial distortion, functional network damage and remodeling caused by tumors, in order to improve the robustness and accuracy of fMRI data processing and network index extraction. Summary of the Invention

[0005] In view of this, the present invention provides a visualization and quantitative analysis method for glioma-induced brain network remodeling based on fMRI. It aims to solve the problems that most existing functional zoning methods rely on template atlases, making it difficult to accurately assess the individual functional connectivity characteristics affected by tumor heterogeneity, and failing to comprehensively quantify the interaction between the tumor area and the brain network. Furthermore, it lacks effective visualization and quantitative analysis methods for glioma-induced brain network remodeling. This invention achieves accurate mapping and quantification of the functional network of the tumor area and its surrounding areas in the brain of glioma patients, providing reliable methodological support for brain network research.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A method for visualizing and quantitatively analyzing glioma-induced brain network remodeling based on fMRI, characterized by the following steps:

[0008] Acquire structural magnetic resonance imaging (SMRI) and functional magnetic resonance imaging (fMRI) data from glioma patients;

[0009] The structural magnetic resonance imaging data is processed to segment and define the tumor core region, the abnormal peritumoral region, and the normal brain region;

[0010] The fMRI data is preprocessed, including spatial normalization using a lesion masking registration strategy, which excludes the tumor core area and peritumoral abnormal area during the registration process.

[0011] Based on a selected standard brain network atlas, the normal brain regions are registered to the standard brain space, realizing the mapping of multiple normal brain networks in the individual space;

[0012] The tumor core region is defined as an independent tumor network unit, which together with the multiple normal brain networks constitutes an extended functional network set.

[0013] Calculate the functional connectivity strength between the time series of each voxel in the whole brain and the average time series of each network unit in the extended functional network set;

[0014] Based on a preset functional connectivity strength threshold, functional connectivity voxels that reach the threshold are identified in each voxel, and functional connectivity voxels that connect with the normal brain network in the tumor core area and connect with the tumor network unit in the abnormal peritumoral area are counted respectively.

[0015] Based on the statistical results obtained from the functional connectivity voxels of the tumor core region and the normal brain network, and the abnormal peritumoral region and the tumor network unit, the intratumoral functional connectivity ratio (RIFR), which characterizes the degree of intratumoral functional connectivity, and the peritumoral tumor connectivity ratio (RPTR), which characterizes the degree of peritumoral functional connectivity, are calculated.

[0016] Generate and output a visual partition map of brain functional network remodeling that incorporates the spatial distribution information of the functional connectivity voxels.

[0017] For example, the selected standard brain network atlas is the Yeo17 resting-state functional network template; based on the Yeo17 resting-state functional network template, the normal brain regions are registered to the standard brain space.

[0018] In a specific implementation scheme, the lesion masking registration strategy is as follows: during the calculation of image nonlinear registration, the tumor core area and the abnormal area around the tumor are excluded from the registration cost function, and the registration transformation matrix is ​​calculated only based on the normal brain region.

[0019] For example, the tumor core area and the abnormal area around the tumor are obtained by manual delineation. In the process of calculating the nonlinear registration of the image, they are excluded from the registration cost function, and the registration transformation matrix is ​​calculated only based on the normal brain region.

[0020] In one specific implementation, the standard brain network atlas is the Yeo17 resting-state functional network template, and the extended functional network set contains 18 network units.

[0021] In one specific implementation scheme, the functional connectivity strength between the time series of each voxel in the whole brain and the average time series of each network unit in the extended functional network set is calculated using the Spearman correlation coefficient method, and the preset functional connectivity strength threshold is 0.7.

[0022] In one specific implementation scheme, the calculation of the intratumoral functional connectivity ratio (RIFR), which characterizes the degree of intratumoral functional connectivity, specifically involves: calculating the total volume (VFR) of functional connective voxels within the tumor core region that reach a threshold level of functional connectivity with any normal brain network, and calculating the ratio of the total volume (VFR) of functional connective voxels to the total volume of the tumor core region.

[0023] For example, the total number of functional connective voxels (VFR) in the tumor core region that have reached a threshold of functional connectivity with any normal brain network is counted, and the ratio of the total volume of functional connective voxels (VFR) to the total number of voxels in the tumor core region is calculated.

[0024] In one specific implementation scheme, the calculation of the peritumoral tumor connectivity ratio (RPTR), which characterizes the degree of peritumoral functional connectivity, specifically involves: calculating the total volume (VTR) of functional connective voxels in the peritumoral abnormal area that reach a threshold level with the functional connectivity strength of the tumor network unit, and calculating the ratio of the total volume (VTR) of the peritumoral tumor functional connective voxels to the total volume of the peritumoral abnormal area.

[0025] For example, the total number of functional link voxels (VTR) in the peritumoral abnormal area that have reached a threshold of functional connection strength with the tumor network unit is counted, and the ratio of the total volume of peritumoral tumor functional link voxels (VTR) to the total number of voxels in the peritumoral abnormal area is calculated.

[0026] In one specific implementation, the preprocessing of the fMRI data further includes sequential slice temporal correction, head motion correction, registration with structural images, component-based noise correction, frequency filtering, detrending processing, and spatial smoothing steps; wherein, component-based noise correction includes extracting noise signals from white matter and cerebrospinal fluid masks and performing regression.

[0027] For example, the preprocessing also includes sequentially performing slice temporal correction, head motion correction, registration with structural images, component-based noise correction, frequency filtering, detrending processing, and spatial smoothing steps; wherein, component-based noise correction includes extracting noise signals from white matter and cerebrospinal fluid masks and performing regression, and the preprocessing also includes performing interference variable regression on six head motion parameters and their first derivatives.

[0028] In one specific implementation, the frequency filtering uses a bandpass filter of 0.01-0.08Hz, and the spatial smoothing uses a Gaussian kernel with a full width at half maximum (FWHM) of 6mm.

[0029] For example, the frequency filtering uses a bandpass filter of 0.01-0.08Hz, and the spatial smoothing uses a Gaussian kernel of 6mm full width at half maximum (FWHM).

[0030] In a specific feasible implementation, a quality control step is also included after preprocessing to visually check and evaluate the structure-function image registration results, spatial normalization accuracy, and noise removal effect.

[0031] For example, a quality control step is included after preprocessing to visually inspect the structure-function image registration results and evaluate the spatial normalization accuracy and noise removal effect.

[0032] In one specific implementation, the step of segmenting and defining the tumor core area and the peritumoral abnormal area is completed based on T1-weighted enhanced images, T2-weighted images, and FLAIR image sequences.

[0033] For example, the step of segmenting and defining the tumor core area and the abnormal peritumoral area is manually delineated by a physician with neuroimaging qualifications based on T1-weighted enhanced images, T2-weighted images, and FLAIR image sequences.

[0034] Compared with existing technologies, the fMRI-based visualization and quantitative analysis method for glioma-induced brain network remodeling described in this invention is used to analyze the functional networks of the tumor area and its surrounding regions in the brain of glioma patients, capturing the phenomenon of glioma-induced brain network remodeling. Through targeted preprocessing of functional magnetic resonance imaging (fMRI) data (including lesion masking registration, slice temporal correction, head movement correction, noise removal, spatial normalization, etc.), tumor subregion segmentation (dividing the tumor core area into the abnormal peritumoral area), individualized normal brain region functional network mapping, and calculation of functional connectivity strength (using Spearman correlation coefficient evaluation), combined with the Yeo17 resting-state functional network template and automated analysis and visualization functions, it achieves individualized and accurate mapping of functional connectivity between the tumor area and normal brain networks, quantitative calculation of intratumoral and peritumoral functional connectivity (including intratumoral functional connectivity ratio RIFR, peritumoral and tumor network connectivity ratio RPTR), and intuitive visualization of brain network remodeling. This effectively improves the sensitivity and specificity of tumor-related functional connectivity identification, and the accuracy and reproducibility of brain network analysis, and has the following beneficial effects:

[0035] It can effectively avoid the interference of tumor quality effect. Through lesion masking registration strategy and network reconstruction strategy in individual space, the tumor region is treated as an independent network unit, realizing the accurate mapping of individualized brain network and improving the sensitivity and specificity of intratumoral and peritumoral functional connectivity identification.

[0036] For the first time, we systematically quantified the proportion of intratumoral and peritumoral functional connectivity and constructed two innovative indicators: the intratumoral functional connectivity ratio RIFR and the peritumoral and tumor network connectivity ratio RPTR. These indicators reflect the degree of tumor integration with the brain network and the extent of its spread, respectively. They have good reproducibility and discriminative power and can quantitatively characterize the neural integration degree, functional connectivity range and other network characteristics of gliomas.

[0037] The entire process is based on routine preoperative fMRI image data and standard atlas templates. It is easy to operate and has the advantages of visualization and quantitative output. It supports functions such as image fusion and index extraction, making it easy to integrate and use in scientific research and clinical practice. It has good versatility and reproducibility. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0039] Figure 1This is a flowchart illustrating the intratumoral and peritumoral brain network delineation in a method for visualizing and quantitatively analyzing glioma-induced brain network remodeling based on fMRI.

[0040] Figure 2 This is a diagram showing the Yeo 17 brain network segmentation results in the normal brain based on functional magnetic resonance imaging.

[0041] Figure 3 This is a map showing the intratumoral functional network distribution of different glioma subtypes. In the map, a represents case 1 (WHO grade 2 oligodendroglioma), b represents case 2 (WHO grade 3 astrocytoma), and c represents case 3 (WHO grade 4 glioblastoma).

[0042] Figure 4a Comparison of RIFR among different cognitive function groups.

[0043] Figure 4b Comparison of RPTR among different cognitive function groups.

[0044] Figure 4c Survival curves for glioma patients in the high and low RIFR groups.

[0045] Figure 4d Survival curves for glioma patients in the high and low RPTR groups. Detailed Implementation

[0046] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] This invention proposes a visualization and quantitative analysis method for glioma-induced brain network remodeling based on fMRI. The aim is to utilize fMRI data to perform individualized functional network analysis of the tumor area and its surrounding regions (i.e., intratumoral and peritumoral areas) in the brains of glioma patients, achieving a quantitative description and visualization of the brain network remodeling phenomenon induced by glioma. The obtained quantitative indicators and visualization maps can be used for subsequent functional assessment and research. The method of this invention consists of two stages: the first stage is the fMRI data preprocessing stage, which involves structural image processing, functional data correction, and regression of interference variables on the raw fMRI data to complete data quality control and standardization, providing a foundation for subsequent analysis; the second stage is the glioma-induced brain network remodeling analysis stage, which, based on the preprocessed data, achieves visualization and quantitative analysis of glioma-induced brain network remodeling through six steps, including tumor subregion segmentation and individualized brain network mapping.

[0048] Phase 1

[0049] The main preprocessing steps for fMRI data include:

[0050] We collected and organized the raw DICOM format image data of the patient's MRI T1, T1-enhanced, T2, FLAIR, and fMRI, and converted them into the corresponding NIfTI format using MRIcron software.

[0051] (1) Structural image processing: The NIfTI data of T1 and T2 were processed sequentially by skull stripping, spatial standardization, tissue segmentation and brain surface reconstruction. The output data were NIfTI format files, named T1.nii and T2.nii respectively.

[0052] (2) Functional data processing: The raw NIfTI data from fMRI were sequentially subjected to slice timing correction, motion correction, coregistration, spatial normalization (the transformation matrix obtained by registering the T1-weighted images of the brain surface reconstructed to the standard space using surface registration was applied to the functional images), and component-based noise correction. To reduce the interference of tumor quality effects on spatial normalization, a lesion-masked registration strategy was adopted: during image registration, the tumor regions obtained by manual delineation were excluded from the registration cost function to minimize the anatomical distortion caused by large tumors.

[0053] (3) Regression of interference variables: including 6 head motion parameters (three translations and three rotations) and their first derivatives. At the same time, the CompCor method was used to remove physiological noise, and noise signals were extracted from white matter and cerebrospinal fluid for regression.

[0054] (4) Frequency filtering and detrending: Bandpass filtering (0.01–0.08Hz) and linear detrending are used to remove low-frequency drift and high-frequency noise.

[0055] (5) Spatial smoothing: Spatial smoothing is performed using a Gaussian kernel with a half-width at half-height (FWHM) of 6 mm.

[0056] (6) Quality control: This includes visual inspection of structure-function image registration, accuracy assessment of spatial normalization, and overall evaluation of noise removal effectiveness. Accurately registered fMRI data are output as NIfTI format files, named fMRI.nii, for subsequent analysis.

[0057] Phase Two

[0058] like Figure 1 As shown, the fMRI-based visualization and quantitative analysis method for glioma-induced brain network remodeling described in this invention is applied to capture changes in network integration and separation caused by gliomas, and includes the following steps:

[0059] Step 1: Tumor Subregion Segmentation Figure 1 Module ①)

[0060] Using medical image analysis software such as MRIcron, two qualified neuroradiologists manually delineated lesion regions on T1-enhanced, T2-enhanced, and FLAIR sequences. The tumor was divided into two subregions: the core tumor and the peritumoral signal abnormality (PSA) region. The core tumor included the enhanced area, non-enhanced parenchymal area, and necrotic area; the peritumoral signal abnormality region included tumor infiltration and edema. The remaining brain tissue not marked as lesions was defined as normal-appearing brain (NAB), serving as the base region for subsequent brain network atlas registration. NIfTI format ROI mask files Tumor_mask.nii and PSA_mask.nii were generated for the core tumor and peritumoral signal abnormality regions, respectively, with voxel values ​​of 1 for lesion sites and 0 for other non-lesion areas.

[0061] Step Two: Mapping Individualized Functional Networks in NAB (Normal Brain Area Mapping) Figure 1 Module ②)

[0062] The Yeo17 resting-state functional network template (Yeo17 RSNs atlas) is adopted. Figure 2 Functional atlas registration was performed using the MNI standard brain space. To reduce the interference of tumor quality effects on the accuracy of atlas registration, a lesion masking registration strategy was adopted: during the nonlinear registration process, the tumor core region (Tumor_mask.nii) and the peritumoral abnormal region (PSA_mask.nii) were excluded from the registration cost function, and only the NAB region was registered to the MNI152 template. Subsequently, the Yeo17 atlas in the standard space was registered to the individual space (T1.nii), achieving individualized brain region mapping of 17 brain networks.

[0063] Step 3: Constructing the Tumor Network (Definition of Tumor Network) Figure 1 Module ③)

[0064] In the individual brain space, the tumor core region (Tumor_mask.nii) is considered an independent functional network unit, forming 18 functional networks (17 normal networks + 1 tumor network) together with the standard Yeo17 atlas. This setup allows for the assessment of whether the tumor region has functional connections with other networks in subsequent analyses, while also identifying whether the peritumoral region is functionalized by this "tumor network".

[0065] Step 4: Functional Connectivity Analysis Figure 1 Module 4)

[0066] Using the preprocessed fMRI data (fMRI.nii) from the first stage as input, the functional connectivity strength (FC) was calculated for each brain voxel and 18 functional networks (17 normal networks + 1 tumor network). Spearman correlation coefficient was used to assess the correlation between time series. In Spearman analysis, a correlation coefficient of 0.7 is considered a strong correlation; therefore, a FC threshold of ≥0.7 was set as the criterion for valid functional connectivity. If the FC value between a voxel and any network is greater than 0.7, a significant functional connection between the voxel and that network is identified, and the voxel is labeled as a "functionally connected voxel."

[0067] Step 5: Intratumoral Network Mapping and Quantification Figure 1 Module ⑤)

[0068] Correlation analysis was performed on the average time series of each voxel within the tumor core region with 17 standard RSNs. If the Spearman correlation coefficient between the voxel and any RSN was greater than 0.7, the voxel was identified as a functional connectivity voxel between the tumor and normal brain networks, and the region formed by these voxels was designated as the intratumoral functional region. The total volume of these regions was calculated to obtain the VFR (Volume of Functional Regions); and the proportion of the VFR to the entire tumor core region was calculated and denoted as the RIFR (Ratio of Intratumoral Functional Region to Core Tumor).

[0069] RIFR = VFR voxel count / total number of voxels in the tumor core area.

[0070] Step Six: Peritumoral Network Mapping Figure 1 Module 6)

[0071] For each voxel in the peritumoral abnormal region, its Spearman correlation coefficient with the tumor network is calculated. If the FC value is greater than 0.7, it is identified as a peritumoral tumor functionally connected voxel connected to the tumor network. The total volume of the regions constituted by these voxels is calculated as VTR (Volume of the tumoral regions in PSA area), and its proportion of the total PSA volume is calculated, denoted as RPTR (The Ratio of the Peritumoral Tumoral Regions to PSA volume).

[0072] RPTR = VTR voxel count / PSA voxel count.

[0073] Finally, the identified intratumoral / peritumor functional connective voxel distribution can be visualized using a functional network partitioning map. By outputting key indicators (such as VFR, RIFR, VTR, RPTR) to structured tables or databases, they can be used for subsequent scientific research analysis, clinical research reference, or machine learning models.

[0074] Figure 3 The study presents the results of functional network reclassification in three typical gliomas of different grades and subtypes, with different colors corresponding to 17 different normal brain networks.

[0075] Gliomas can be classified into three subtypes based on their molecular characteristics: oligodendroglioma, astrocytoma, and glioblastoma. Figure 3 The distribution of intratumoral functional networks labeled by this invention is shown for each subtype. Figure 3 The color of the marker corresponding to each functional network in the middle is... Figure 2 The consistent color of the Yeo17 brain network markers allows for a direct visualization at the voxel level of which regions within the tumor connect to normal brain network functions. This confirms the applicability of this invention to different glioma subtypes.

[0076] Compared to existing analytical techniques, this invention extracts quantitative indicators such as RIFR and RPTR, which can quantitatively describe the intratumoral and peritumoral network characteristics of gliomas. To analyze the clinical application value of RIFR and RPTR, this invention conducted clinical feature analysis on a larger sample.

[0077] A total of 79 patients with glioma underwent preoperative cognitive function assessment using the Montreal Cognitive Assessment Scale. Based on the scores, patients were divided into three groups: normal cognitive function (26-30 points), mild cognitive impairment (18-25 points), and moderate to severe cognitive impairment (0-17 points). Welch's one-way ANOVA was used to statistically analyze the differences in RIFR and RPTR among the three groups. Results are as follows: Figure 4a and Figure 4b As shown, the RIFR and RPTR of patients with moderate to severe cognitive impairment were significantly lower than those of patients with normal or mild cognitive impairment (P values ​​were all less than 0.01, indicating statistical differences).

[0078] Among 182 patients with available survival data, patients were divided into low RIFR and high RIFR groups, and low RPTR and high RIFR groups, based on the median RIFR and RPTR. Survival curve analysis was performed between the two groups. The results are shown in Figure 4. Patients in the high RIFR group had longer survival than those in the low RIFR group. Figure 4c (P=0.004); the survival time of patients in the high RPTR group was also longer than that in the low RPTR group ( Figure 4d (P=0.029). These results demonstrate that the RIFR and RPTR quantitative indices provided by this invention are effective parameters for quantifying the state of brain network remodeling. These parameters exhibit significant statistical differences among different patient groups, further validating their sensitivity and effectiveness as characteristic parameters of brain networks.

[0079] In a preferred embodiment of the present invention, the tumor subregion segmentation step is implemented in the following manner:

[0080] Two qualified neuroradiologists manually delineated the lesion region on T1-weighted, T2-weighted, and FLAIR sequences using MRIcron medical image analysis software. During the delineation, a strict adherence to the following principles was followed: the tumor was divided into a tumor core region and a peritumoral abnormality region. The tumor core region included the enhanced area, the non-enhanced parenchymal area, and the necrotic area; the peritumoral abnormality region included tumor infiltration and edema. The remaining brain tissue not marked as lesions was defined as the normal brain region (NAB), serving as the base area for subsequent brain network atlas registration.

[0081] This manual delineation method provides a reliable anatomical basis for subsequent lesion masking and registration strategies and precise functional analysis.

[0082] The various embodiments described in this specification are presented in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for visualizing and quantitatively analyzing glioma-induced brain network remodeling based on fMRI, characterized in that, Includes the following steps: Acquire structural magnetic resonance imaging (SMRI) and functional magnetic resonance imaging (fMRI) data from glioma patients; The structural magnetic resonance imaging data is processed to segment and define the tumor core region, the abnormal peritumoral region, and the normal brain region; The fMRI data is preprocessed, including spatial normalization using a lesion masking registration strategy, which excludes the tumor core area and peritumoral abnormal area during the registration process. Based on a selected standard brain network atlas, the normal brain regions are registered to the standard brain space, realizing the mapping of multiple normal brain networks in the individual space; The tumor core region is defined as an independent tumor network unit, which together with the multiple normal brain networks constitutes an extended functional network set. Calculate the functional connectivity strength between the time series of each voxel in the whole brain and the average time series of each network unit in the extended functional network set; Based on a preset functional connectivity strength threshold, functional connectivity voxels that reach the threshold are identified in each voxel, and functional connectivity voxels that connect with the normal brain network in the tumor core area and connect with the tumor network unit in the abnormal peritumoral area are counted respectively. Based on the statistical results obtained from the functional connectivity voxels of the tumor core region and the normal brain network, and the abnormal peritumoral region and the tumor network unit, the intratumoral functional connectivity ratio (RIFR), which characterizes the degree of intratumoral functional connectivity, and the peritumoral tumor connectivity ratio (RPTR), which characterizes the degree of peritumoral functional connectivity, are calculated. Generate and output a visual partition map of brain functional network remodeling that incorporates the spatial distribution information of the functional connectivity voxels.

2. The method for visualization and quantitative analysis of glioma-induced brain network remodeling based on fMRI according to claim 1, characterized in that, The lesion masking registration strategy is as follows: during the calculation of nonlinear image registration, the tumor core area and the abnormal area around the tumor are excluded from the registration cost function, and the registration transformation matrix is ​​calculated only based on the normal brain region.

3. The method for visualization and quantitative analysis of glioma-induced brain network remodeling based on fMRI according to claim 1, characterized in that, The standard brain network atlas is the Yeo17 resting-state functional network template, and the extended functional network set contains 18 network units.

4. The method for visualization and quantitative analysis of glioma-induced brain network remodeling based on fMRI according to claim 1, characterized in that, The functional connectivity strength between the time series of each voxel in the whole brain and the average time series of each network unit in the extended functional network set is calculated using the Spearman correlation coefficient method, and the preset functional connectivity strength threshold is 0.

7.

5. The method for visualization and quantitative analysis of glioma-induced brain network remodeling based on fMRI according to claim 1, characterized in that, The calculation of the intratumoral functional connectivity ratio (RIFR), which characterizes the degree of intratumoral functional connectivity, specifically involves: calculating the total volume (VFR) of functional connective voxels within the tumor core region that reach a threshold level of functional connectivity with any normal brain network, and then calculating the ratio of the total volume (VFR) of functional connective voxels to the total volume of the tumor core region.

6. The method for visualization and quantitative analysis of glioma-induced brain network remodeling based on fMRI according to claim 1, characterized in that, The calculation of the peritumoral tumor connectivity ratio (RPTR), which characterizes the degree of peritumoral functional connectivity, specifically involves: calculating the total volume (VTR) of functional connective voxels in the peritumoral abnormal area that reach a threshold level with the functional connectivity strength of the tumor network unit, and calculating the ratio of the total volume (VTR) of the peritumoral tumor functional connective voxels to the total volume of the peritumoral abnormal area.

7. The method for visualization and quantitative analysis of glioma-induced brain network remodeling based on fMRI according to claim 1, characterized in that, The preprocessing of the fMRI data also includes sequential steps such as slice temporal correction, head motion correction, registration with structural images, component-based noise correction, frequency filtering, detrending processing, and spatial smoothing; wherein, component-based noise correction includes extracting noise signals from white matter and cerebrospinal fluid masks and performing regression.

8. The method for visualization and quantitative analysis of glioma-induced brain network remodeling based on fMRI according to claim 7, characterized in that, The frequency filtering uses a bandpass filter of 0.01-0.08Hz, and the spatial smoothing uses a Gaussian kernel with a full width at half maximum (FWHM) of 6mm.

9. The method for visualization and quantitative analysis of glioma-induced brain network remodeling based on fMRI according to claim 1, characterized in that, The preprocessing process includes a quality control step, which involves visually inspecting and evaluating the structure-function image registration results, spatial normalization accuracy, and noise removal effectiveness.

10. The method for visualization and quantitative analysis of glioma-induced brain network remodeling based on fMRI according to claim 1, characterized in that, The steps of segmenting and defining the tumor core area and the abnormal peritumoral area are completed based on T1-weighted enhanced images, T2-weighted images, and FLAIR image sequences.

Images