Language area localization method and device based on magnetic resonance imaging and lateralization coefficient

Abstract

The application relates to a language area positioning method and device based on nuclear magnetic resonance imaging and a lateralization coefficient. The application sets a structural nuclear magnetic image and a brain function nuclear magnetic image of a detected person, registers the nuclear magnetic image and the brain function nuclear magnetic image, and outputs a visual three-dimensional brain map, combines a seed point Pearson correlation coefficient, extracts voxel counts of left and right brains in a threshold range, and combines a lateralization coefficient to distinguish whether the detected person is left brain dominant, right brain dominant or bilateral brain dominant.

Description

Technical Field

[0001] This application belongs to the field of brain detection technology, and in particular relates to a method and device for locating language areas based on magnetic resonance imaging and lateralization coefficients. Background Technology

[0002] Epilepsy accounts for a significant proportion of the global disease burden, affecting approximately 50 million people worldwide. The estimated prevalence of active epilepsy (i.e., persistent seizures or requiring treatment) in the general population is 4-10 per 1,000. An estimated 5 million people are diagnosed with epilepsy globally each year. In high-income countries, it is estimated that 49 people per 100,000 are diagnosed with epilepsy annually, and 30-40% of these develop drug-resistant epilepsy. For patients with drug-resistant epilepsy, one of the most effective treatments is surgical resection of the epileptogenic focus. Preserving language function, a uniquely human ability, is crucial during resection; therefore, the locational relationship between the language area and the epileptogenic focus must be clearly defined preoperatively. Resection of the core language area can lead to permanent language impairment in patients.

[0003] Traditional research suggests that approximately 80% of the dominant language hemisphere in humans is on the left side. Left-sided language dominance is often referred to as typical language lateralization, while right-sided or bilateral language dominance is considered atypical. Increasing research confirms that language lateralization is more likely to be atypical in epilepsy patients than in healthy individuals. Currently, the gold standard for determining language lateralization is the Wada test. Studies using the Wada test to determine the dominant language hemisphere in temporal lobe epilepsy patients show a left-sided dominant hemisphere percentage of 63%-97%. Predictors of language lateralization in temporal lobe epilepsy patients include epileptogenic focus side, hand dominance, age of onset, and family history of left-sidedness. However, the invasiveness, lack of standardization, lack of spatial resolution, and difficulties associated with application and interpretation of the Wada test significantly limit its practical application. Currently, the gold standard for clinically locating the language area is cortical electrical stimulation (CES). CE methods include subdural electrode placement, stereotactic electroencephalography (SEEG) electrodes, and intraoperative awakening placement of cortical surface electrodes. These methods are invasive, costly, and require a high degree of patient cooperation, making them unsuitable for universal clinical application. Therefore, researchers are continuously exploring non-invasive alternatives for language area localization. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method and apparatus for language region localization using a rapid and accurate non-invasive magnetic resonance imaging technique to overcome the shortcomings of the prior art.

[0005] The technical solution adopted by this invention to solve its technical problem is:

[0006] A method for localizing speech regions based on magnetic resonance imaging and lateralization coefficients includes the following steps:

[0007] S1: Have the subject close their eyes, keep their mind clear, and keep their head still. Use a magnetic resonance imaging system to scan the subject and obtain structural MRI images and functional MRI images of the brain.

[0008] S2: The structural MRI images and functional MRI images of the brain are converted and preprocessed to obtain preprocessed data and a visualized three-dimensional brain map. The Pearson correlation coefficient is calculated based on the seed point functional connectivity analysis method. A composite correspondence is then performed between the visualized three-dimensional brain map and the Pearson correlation coefficient.

[0009] S3: Set a threshold for the Pearson correlation coefficient in the visualized 3D brain map region in step S2, and extract voxel counts of the left and right hemispheres within the threshold range;

[0010] S4: Calculate LI = (LSLH - RSRH) / (LSLH + RSRH), where LSLH is the voxel count from the left brain in step S3; RSRH is the voxel count from the right brain in step S3; determine whether the language dominance is left-sided, right-sided, or bilateral based on the LI result.

[0011] Preferably, in the language region localization method of the magnetic resonance imaging technology of the present invention, the magnetic resonance imaging system in step S1 is:

[0012] When using a Philips machine: Philips Achieva 3.0T, the structural NMR images are T1 high resolution, and the scanning parameters for the structural images are: sagittal scanning, TR = 7.6ms, TE = 3.7ms, flip angle = 8°, number of layers = 180 layers, layer thickness = 1.0mm, FOV = 230mm × 180mm, matrix = 232 × 230, voxel size = 1.0mm × 1.0mm × 2.0mm;

[0013] The scanning parameters for resting-state functional magnetic resonance imaging (fMRI) images were as follows: a gradient echo-based planar echo sequence was used, the scanning direction was parallel to the straight echo, TR = 2000ms, TE = 30ms, FA = 90°, FOV = 220mm × 220mm, matrix = 72 × 73, number of slices = 32, slice thickness = 4mm, slice interval = 0, and voxel size = 3.0mm × 3.0mm × 4.0mm.

[0014] When using a general-purpose machine: Data acquisition was performed using a SIGNAArchitect superconducting MRI system and a 48-channel lead coil, scanning T1 high-resolution structural images and rs-fMRI sequences, respectively.

[0015] The scanning parameters are as follows: T1 scan was performed in the sagittal plane, TR = 6.8 ms, TE = 2.7 ms, inversion time TI = 600 ms, flip angle = 8°, slice thickness = 0.5 mm, FOV = 200 mm × 200 mm, matrix = 200 × 200, voxel size = 1.0 mm × 1.0 mm × 1.0 mm; the rs-fMRI sequence used was a gradient echo-based planar echo sequence, with the scanning direction parallel to the vertical plane.

[0016] Return, TR = 3000ms, TE = 35ms, FA = 90°, FOV = 220mm × 220mm, matrix = 80 × 80, number of layers = 36 layers, layer thickness = 4mm, layer spacing = 0, voxel size = 2.8mm × 2.8mm × 4.0mm.

[0017] Preferably, in the language region localization method of the nuclear magnetic resonance imaging technology of the present invention, step S2 specifically includes:

[0018] The preprocessing procedure for structural NMR images is as follows:

[0019] S201, Data Export: Export the raw file format from the magnetic resonance imaging system into an analyzable data format.

[0020] S202. Use FreeSurfer software to reconstruct structural MRI image data in analyzable data formats to form individualized three-dimensional brain cortex maps.

[0021] The preprocessing process for functional MRI images of the brain is as follows:

[0022] S211. Data export: Export the original file format into an analyzable data format.

[0023] S212. Head movement correction: Perform rigid body transformation in 6 directions on the functional MRI images of the brain in the analyzable data format, align the intermediate layer images of the functional MRI images of the brain at all time points, and select data with head movement translation ≤2.0mm and rotational movement ≤2.0° according to the head movement correction curve.

[0024] S213. Bandpass filter the brain functional MRI image signal to 0.01-0.08 Hz, and perform head movement regression, ventricle and white matter signal regression;

[0025] S214. The signal processed by S213 is first aligned and registered with the individualized three-dimensional brain cortex map formed in step S202. Gaussian smoothing kernel is applied to the signal in the registered brain cortex space. The smoothed data is downsampled to a cortical space template containing several voxels in each hemisphere using the mri_surf2surf function in FreeSurfer, and finally a visualized three-dimensional brain map is formed.

[0026] Preferably, for the method of localizing the language area of the nuclear magnetic resonance imaging technology of the present invention,

[0027] In the seed point functional connectivity analysis method, the seed points are selected from one or any combination of the posterior part of the superior temporal gyrus on the left and right sides of the brain, the posterior part of the middle frontal gyrus on the left and right sides, the middle temporal gyrus on the left and right sides, the posterior part of the inferior frontal gyrus on the left and right sides, the lateral part of the extraorbital region on the left and right sides, the middle part of the inferior frontal gyrus on the left and right sides, the superior temporal gyrus on the left and right sides, and the supramarginal gyrus on the left and right sides.

[0028] Preferably, for the method of localizing the language area of the nuclear magnetic resonance imaging technology of the present invention, in step S2, the Pearson correlation coefficient between each seed point ROI and the time series of each voxel in the whole brain is calculated using the corr function in matlab. The value range of the obtained Pearson correlation coefficient is [-1, 1]. The negative values in the Pearson correlation coefficient are assigned as 0, that is, the values in [-1, 0] are assigned as 0, and the range of the Pearson correlation coefficient becomes [0, 1];

[0029] In step S3, the threshold analysis method is as follows:

[0030] Select a first threshold [0.50 - 0.99], and count the voxel counts of the left brain and the right brain within the range of the first threshold respectively.

[0031] Preferably, for the method of localizing the language area of the nuclear magnetic resonance imaging technology of the present invention, if LI ≥ 0.20, it is considered that the patient has left language dominance; if LI ≤ -0.20, it is considered that the patient has right language dominance; if -0.20 < LI < 0.20, it is considered that the patient has bilateral language dominance.

[0032] The present invention also discloses a device for localizing the language area of nuclear magnetic resonance imaging technology, including:

[0033] A data acquisition module: used to use a magnetic resonance imaging system when the subject closes their eyes, keeps their mind clear, and keeps their head still, and scan the subject respectively to obtain a structural nuclear magnetic resonance image and a brain functional nuclear magnetic resonance image;

[0034] A data processing module: used to perform data conversion and preprocessing on the structural nuclear magnetic resonance image and the brain functional nuclear magnetic resonance image respectively, obtain the preprocessed data, obtain a visualized three-dimensional brain map, calculate the Pearson correlation coefficient based on the seed point functional connectivity analysis method, and screen out the regions where the Pearson correlation coefficient is shown as positive correlation in the visualized three-dimensional brain map;

[0035] A core language area determination module: used to set a threshold for the Pearson correlation coefficient in the data processing module, and extract the voxel counts of the left brain and the right brain within the threshold range; [[ID=H29]]

[0036] Classification module: used to calculate LI = (LSLH - RSRH) / (LSLH + RSRH), where LSLH is the voxel count from the left brain in step S3; RSRH is the voxel count from the right brain in step S3; and the result of LI determines whether there is left-sided language dominance, right-sided language dominance, or bilateral language dominance.

[0037] The beneficial effects of this invention are:

[0038] 1. This application obtains structural MRI images and functional MRI images of the brain by setting up a scan of the subject. The MRI images and functional MRI images of the brain are registered and a visualized three-dimensional brain map is output. By combining the Pearson correlation coefficient of the seed points, the voxel counts of the left and right hemispheres within the threshold range are extracted and the lateralization coefficient is calculated to distinguish whether the subject is left-brain dominant, right-brain dominant, or bilaterally dominant. Attached Figure Description

[0039] The technical solution of this application will be further described below with reference to the accompanying drawings and embodiments.

[0040] Figure 1 This is a flowchart of the language region localization method based on nuclear magnetic resonance imaging and lateralization coefficients in this application;

[0041] Figure 2 and Figure 3 This is a three-dimensional brain map of a patient's left and right hemispheres after setting a first threshold, which is [0.50-0.99]. Detailed Implementation

[0042] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0043] In the description of this application, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as limiting the scope of protection of this application. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0044] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art will understand the specific meaning of the above terms in this application based on the specific circumstances.

[0045] The technical solution of this application will now be described in detail with reference to the accompanying drawings and embodiments.

[0046] Example 1

[0047] This embodiment provides a language region localization method based on nuclear magnetic resonance imaging and lateralization coefficients, such as... Figure 1 As shown, it includes the following steps:

[0048] S1: Have the subject close their eyes, keep their mind clear, and keep their head still. Use a magnetic resonance imaging system to scan the subject and obtain structural MRI images and functional MRI images of the brain.

[0049] The magnetic resonance imaging system in step S1 is:

[0050] The Philips Achieva 3.0T machine was used. The structural NMR images were T1 high resolution. The scanning parameters for the structural images were as follows: sagittal scanning, TR = 7.6ms, TE = 3.7ms, flip angle = 8°, number of slices = 180, slice thickness = 1.0mm, FOV = 230mm × 180mm, matrix = 232 × 230, voxel size = 1.0mm × 1.0mm × 2.0mm.

[0051] The scanning parameters for resting-state functional magnetic resonance imaging (fMRI) images were as follows: gradient echo-planar imaging (GRE-EPI) sequence was used, the scanning direction was parallel to the straight echo, TR = 2000ms, TE = 30ms, FA = 90°, FOV = 220mm × 220mm, matrix = 72 × 73, number of slices = 32, slice thickness = 4mm, slice interval = 0, and voxel size = 3.0mm × 3.0mm × 4.0mm.

[0052] If a general-purpose machine (GE) is used in step S1: Data acquisition is performed using a SIGNA Architect superconducting MRI system and a 48-channel lead coil, scanning T1 high-resolution structural images and rs-fMRI sequences, respectively.

[0053] The scanning parameters were as follows: T1 scan was performed in the sagittal plane, TR = 6.8 ms, TE = 2.7 ms, inversion time TI = 600 ms, flip angle = 8°, slice thickness = 0.5 mm, FOV = 200 mm × 200 mm, matrix = 200 × 200, voxel size = 1.0 mm × 1.0 mm × 1.0 mm. The rs-fMRI sequence used was a gradient echo-planar imaging (GRE-EPI) sequence, with the scanning direction parallel to the rectilinear eccentricity, TR = 3000 ms, TE = 35 ms, FA = 90°, FOV = 220 mm × 220 mm, matrix = 80 × 80, number of slices = 36, slice thickness = 4 mm, slice interval = 0, voxel size = 2.8 mm × 2.8 mm × 4.0 mm.

[0054] S2: The structural MRI images and functional MRI images of the brain are converted and preprocessed to obtain preprocessed data and a visualized three-dimensional brain map. The Pearson correlation coefficient is calculated based on the seed point functional connectivity analysis method. The visualized three-dimensional brain map is combined with the Pearson correlation coefficient to obtain a Pearson correlation coefficient value for each voxel in the visualized three-dimensional brain map. The voxel in the visualized three-dimensional brain map that meets the conditions can be selected by filtering the Pearson correlation coefficient values.

[0055] The seed-based functional connectivity (FC) analysis method is as follows:

[0056] Select a hemisphere (left or right hemisphere), and the seed points (ROIs) are all from the same hemisphere. The seed points are from the posterior part of the superior temporal gyrus, the posterior part of the middle frontal gyrus, the middle temporal gyrus, the posterior part of the inferior frontal gyrus, the extraorbital lateral part, the middle part of the inferior frontal gyrus, the superior temporal gyrus, and the supramarginal gyrus. Extract the average time series within the ROI.

[0057] The Pearson correlation coefficient for each seed point ROI and each voxel time series in the whole brain was calculated using the corr function in MATLAB. The range of the obtained Pearson correlation coefficient is [-1, 1]. The negative values ​​in the Pearson correlation coefficient were assigned 0, that is, the values ​​in [-1, 0] were assigned 0, and the range of the Pearson correlation coefficient became [0, 1].

[0058] The Pearson correlation coefficients for each voxel are summed to obtain the summed Pearson correlation coefficients.

[0059] Smoothing is applied to the summed Pearson correlation coefficient to improve the signal-to-noise ratio;

[0060] It should be noted that when the correlation coefficient of two time series satisfies 0 < R ≤ 1, it indicates that the two time series are positively correlated. Moreover, as the value increases, the positive correlation becomes stronger. When it reaches 1, it is a perfect positive correlation. At this time, the signal intensities of the two time series increase or decrease simultaneously over time. Similarly, when the correlation coefficient of two time series satisfies -1 ≤ R < 0, it indicates that the two time series are negatively correlated. And as the value decreases, the negative correlation becomes stronger. When it reaches -1, it is a perfect negative correlation. At this time, the trends of the signal intensities of the two time series increasing or decreasing over time are exactly opposite.

[0061] The specific steps of S2 are as follows:

[0062] The preprocessing process of the structural MRI image is as follows:

[0063] S201. Data export: Export the original file format in the magnetic resonance imaging system into an analyzable data format. For example, compress the dicom format data into a nii.gz format file.

[0064] S202. Apply the freesurfer software to reconstruct the structural MRI image data in the analyzable data format to form an individualized three-dimensional cerebral cortex map (referring to the three-dimensional cerebral cortex map unique to the tested person).

[0065] The preprocessing process of the functional MRI image of the brain is as follows:

[0066] S211. Data export: Export the original file format into an analyzable data format. For example, compress the dicom data into a nii.gz format file.

[0067] S212. Head motion correction: Perform a six-direction rigid body transformation on the functional MRI image data in the analyzable data format to align the middle layer images of the functional MRI image data at all time points, and according to the head motion correction curve, take the head motion translation ≤ 2.0 mm and the rotational movement ≤ 2.0° data;

[0068] S213. Band-pass filter the functional MRI image signal to 0.01 - 0.08 Hz, and perform head motion regression, ventricular and white matter signal regression.

[0069] S214. First, align and register the signal processed in S213 with the individualized three-dimensional cerebral cortex map formed in step S202. Apply a Gaussian smoothing kernel with a full width at half maximum of 6 mm to the signal in the registered cerebral cortex space; the smoothed data is downsampled to a cortical space model with 2562 voxels per hemisphere using the mri_surf2surf function in FreeSurfer, and finally a visualized three-dimensional brain map is formed.

[0070] The Pearson correlation coefficient is calculated based on the seed point functional connectivity analysis method. The seed regions are selected according to the paper published by Yeo et al. in 2011 [Yeo BT, Krienen FM, Sepulcre J, Sabuncu MR, Lashkari D, Hollinshead M, et al. The organization of the human cerebral cortex estimated by intrinsic functional connectivity]. Eight regions on both sides of the left and right brains are selected respectively (the posterior part of the bilateral superior temporal gyrus, the posterior part of the bilateral middle frontal gyrus, the bilateral middle temporal gyrus, the posterior part of the bilateral inferior frontal gyrus, the bilateral lateral orbital frontal cortex, the middle part of the bilateral inferior frontal gyrus, the bilateral superior temporal gyrus, the bilateral supramarginal gyrus). The regions with positive Pearson correlation coefficients are screened out in the visualized three-dimensional brain map (that is, the language areas in the visualized three-dimensional brain map are screened out). Each region selects one side of the left brain or one side of the right brain each time;

[0071] S3: Set a threshold for the Pearson correlation coefficient in the region of the visualized three-dimensional brain map in step S2, and extract the voxel counts of the left brain and the right brain within the threshold range; first form a unilateral (left or right) visualized three-dimensional brain map, and through threshold analysis, read the number of voxels of the left brain and the right brain within the threshold range respectively.

[0072] Specifically, perform threshold analysis on the visualized three-dimensional brain map combined with the Pearson correlation coefficient values in step S2, set the first threshold, and respectively count the voxel counts of the left brain and the right brain within the first threshold;

[0073] S4: Calculate LI = (LSLH - RSRH) / (LSLH + RSRH), where LSLH is the voxel count of the left brain from step S3; RSRH is the voxel count of the right brain from step S3; Determine whether the patient has left language dominance, right language dominance, or bilateral language dominance according to the result of LI.

[0074] Further, the first threshold is [0.50 - 0.99].

[0075] Further, if LI ≥ 0.20, it is considered that the patient has left language dominance; if LI ≤ -0.20, it is considered that the patient has right language dominance; if -0.20 < LI < 0.20, it is considered that the patient has bilateral language dominance.

[0076] The above methods can already distinguish whether a patient has left-sided, right-sided, or bilateral language dominance. If further clarification is needed, after identifying the dominant language side, the threshold range can be further narrowed using a 3D brain visualization map of that side until the foreground region area falls within a certain threshold range of 0.8 cm². 2 -1.5cm 2 When defining a range, the language area is classified based on its location in the brain.

[0077] The selected threshold ranges are as follows:

[0078] [0.55-0.99], [0.60-0.99], [0.65-0.99], [0.70-0.99], [0.75-0.99], [0.80-0.99], [0.85-0.99], [0.90-0.99], [0.95-0.99].

[0079] When it's necessary to specifically identify the region, eight groups of visualized 3D brain maps, each corresponding to one of eight regions, are used. Each group is divided into left and right hemispheres, and threshold analysis is performed separately. For example, the Pearson correlation coefficients of the bilateral middle temporal gyri of the left and right hemispheres are compared. Based on the voxel lateralization coefficients within the threshold range, it is determined whether the bilateral middle temporal gyri of the left or right hemisphere is the core language region. This method is often used to distinguish between Broca's area and Wernicke's area. In this case, calculations are performed on the bilateral Broca's area and Wernicke's area of ​​the left and right hemispheres. Specifically, there are eight categories: left side, right side, bilateral, and neither side has Broca's area; left side, right side, bilateral, and neither side has Wernicke's area.

[0080] It should be noted that FreeSurfer is a toolkit for analyzing and visualizing brain structural and functional imaging. It can process MRI and fMRI data and perform brain anatomical data measurements.

[0081] Furthermore, it should be noted that the location of the core language region in this application cannot be directly used for disease diagnosis.

[0082] Example 2

[0083] This embodiment provides a language region localization device for magnetic resonance imaging (MRI) technology, including:

[0084] Data acquisition module: used to scan the subject with eyes closed, mind clear, and head still using a magnetic resonance imaging system to obtain structural MRI images and functional MRI images of the brain.

[0085] Data processing module: used to perform data conversion and preprocessing on the structural MRI images and functional MRI images of the brain respectively, to obtain preprocessed data, to obtain a visualized three-dimensional brain map, to calculate the Pearson correlation coefficient based on the seed point functional connectivity analysis method, and to perform a composite correspondence between the visualized three-dimensional brain map and the Pearson correlation coefficient;

[0086] Core Language Region Determination Module: Used in the data processing module to set a threshold for the Pearson correlation coefficient and extract voxel counts of the left and right hemispheres within the threshold range;

[0087] Classification module: used to calculate LI = (LSLH - RSRH) / (LSLH + RSRH), where LSLH is the voxel count from the left brain in step S3; RSRH is the voxel count from the right brain in step S3; and the result of LI determines whether there is left-sided language dominance, right-sided language dominance, or bilateral language dominance.

[0088] The language region localization device of the magnetic resonance imaging technology in this embodiment corresponds to the language region localization method of the magnetic resonance imaging technology in Embodiment 1.

[0089] Based on the above-described preferred embodiments according to this application, and through the foregoing description, those skilled in the art can make various changes and modifications without departing from the technical concept of this application. The technical scope of this application is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. A language area localization method based on nuclear magnetic resonance imaging and lateralization coefficient, characterized in that, The method comprises the following steps: S1: close the eyes of the detected person, keep the head clear, keep the head still, use a magnetic resonance imaging system, and scan the detected person to obtain structural nuclear magnetic image and brain function nuclear magnetic image respectively; S2: data conversion and preprocessing are performed on the structural nuclear magnetic image and the brain function nuclear magnetic image respectively to obtain preprocessed data, obtain a visual three-dimensional brain map, calculate a Pearson correlation coefficient based on a seed point functional connection analysis method, and perform complex correspondence on the visual three-dimensional brain map and the Pearson correlation coefficient; S3: a threshold value of the Pearson correlation coefficient is set in the visual three-dimensional brain map region in the S2 step, and voxel counts of the left brain and the right brain in the threshold value range are extracted; S4: LI=(LSLH-RSRH) / (LSLH+RSRH) is calculated, LSLH is the voxel count of the left brain from the S3 step, RSRH is the voxel count of the right brain from the S3 step, and whether the left language dominance or the right language dominance or the bilateral language dominance is determined according to the result of LI; In the S2 step, the seed points in the seed point functional connection analysis method are selected from one or any multiple of the following: the left and right posterior superior temporal gyrus, the left and right posterior middle frontal gyrus, the left and right middle temporal gyrus, the left and right posterior inferior frontal gyrus, the left and right lateral orbital frontal lobe, the left and right middle inferior frontal gyrus, the left and right superior temporal gyrus, and the left and right superior temporal gyrus. In the S2 step, the corr function of matlab is used to calculate the Pearson correlation coefficient of each seed point ROI and the time sequence of each voxel in the whole brain, the value of the obtained Pearson correlation coefficient ranges from -1 to 1, the negative value in the Pearson correlation coefficient is assigned as 0, that is, the value of [-1, 0] is assigned as 0, and the range of the Pearson correlation coefficient is changed to [0, 1]; In the S3 step, the threshold value analysis method is: A first threshold value [0.50-0.99] is selected, and the voxel counts of the left brain and the right brain in the first threshold value range are counted respectively; If LI is greater than or equal to 0.20, the patient is considered to have left language dominance, if LI is less than or equal to -0.20, the patient is considered to have right language dominance, and if -0.20 < LI < 0.20, the patient is considered to have bilateral language dominance.

2. The method of language area localization based on MRI and lateralization coefficient according to claim 1, characterized in that, In the S1 step, the magnetic resonance imaging system is: When a philips machine is used: Philips Achieva 3.0T, the structural nuclear magnetic image is T1 high resolution, and the scanning parameters of the structural image are: sagittal scanning is adopted, TR=7.6ms, TE=3.7ms, flip angle=8°, layer number=180 layers, layer thickness=1.0mm, FOV=230mm x 180mm, matrix=232 x 230, voxel size=1.0mm x 1.0mm x 2.0mm; The scan parameters of the resting state functional magnetic resonance imaging image are as follows: a gradient echo-based planar echo sequence is used, the scanning direction is parallel to the straight return, TR=2000 ms, TE=30 ms, FA=90°, FOV=220 mm*220 mm, matrix=72*73, number of layers=32 layers, layer thickness=4 mm, layer spacing=0, voxel size=3.0 mm*3.0 mm*4.0 mm; When a general machine is used: data acquisition uses a SIGNA Architect superconducting MRI system and a 48-channel head coil, and T1 high-resolution structural images and rs-fMRI sequences are scanned respectively; The scan parameters are as follows: T1 uses sagittal scanning, TR=6.8 ms, TE=2.7 ms, inversion time TI=600 ms, flip angle=8°, layer thickness=0.5 mm, FOV=200 mm*200 mm, matrix=200*200, voxel size=1.0 mm*1.0 mm*1.0 mm; the rs-fMRI sequence uses a gradient echo-based planar echo sequence, the scanning direction is parallel to the straight return, TR=3000 ms, TE=35 ms, FA=90°, FOV=220 mm*220 mm, matrix=80*80, number of layers=36 layers, layer thickness=4 mm, layer spacing=0, voxel size=2.8 mm*2.8 mm*4.0 mm.

3. The language area localization method based on MRI and lateralization coefficient according to claim 1 or 2, characterized in that, The S2 step specifically comprises: The preprocessing process of the structural nuclear magnetic image is as follows: S201, data export, export the original file format in the magnetic resonance imaging system into an analyzable data format; S202, reconstruct the structural nuclear magnetic image data in the analyzable data format by using the freesurfer software to form an individualized three-dimensional brain cortex map; The preprocessing process of the brain function nuclear magnetic image is as follows: S211, data export, export the original file format into an analyzable data format; S212, head motion correction, rigid body conversion in 6 directions is performed on the brain function nuclear magnetic image data in the analyzable data format, the intermediate layer image in all time point image brain function nuclear magnetic image data is aligned, and the data with head motion translation ≤2.0 mm and rotation movement ≤2.0° is taken according to the head motion correction curve; S213, band-pass filter the brain function nuclear magnetic image signal to 0.01-0.08 hertz, and perform head motion regression, ventricle and white matter signal regression; S214, align and register the signal processed in S213 with the individualized three-dimensional brain cortex map formed in the S202 step, and apply a Gaussian smoothing kernel to the signal in the registered brain cortex space; the smoothed data is down-sampled to a cortex space template containing a plurality of voxels per hemisphere using the mri_surf2surf function in FreeSurfer, and finally a visualized three-dimensional brain map is formed.

4. A language area positioning apparatus based on nuclear magnetic resonance imaging and lateralization coefficient, characterized by, It comprises: The data acquisition module is configured to use a magnetic resonance imaging system to scan a subject to obtain structural magnetic resonance images and functional magnetic resonance images of the subject's brain when the subject closes his / her eyes, keeps his / her head clear, and keeps his / her head still; The data processing module is configured to perform data conversion and preprocessing on the structural magnetic resonance images and the functional magnetic resonance images to obtain preprocessed data, to obtain a visualized three-dimensional brain map, to calculate a Pearson correlation coefficient based on a seed point functional connectivity analysis method, and to display the Pearson correlation coefficient as a positively correlated region in the visualized three-dimensional brain map; The core language area determination module is configured to set a threshold value for the Pearson correlation coefficient in the data processing module and to extract voxel counts of the left brain and the right brain within the threshold value range; The classification module is configured to calculate LI=(LSLH-RSRH) / (LSLH+RSRH), where LSLH is the voxel count of the left brain from the S3 step, and RSRH is the voxel count of the right brain from the S3 step, and to determine whether the left language dominance, the right language dominance, or the bilateral language dominance is dominant based on the result of LI; In the data processing module, the seed points in the seed point functional connectivity analysis method are selected from one or any number of the following: the left and right posterior superior temporal gyrus, the left and right posterior middle frontal gyrus, the left and right middle temporal gyrus, the left and right posterior inferior frontal gyrus, the left and right orbitofrontal lateral, the left and right middle inferior frontal gyrus, the left and right superior temporal gyrus, and the left and right superior temporal gyrus. In the data processing module, the Pearson correlation coefficient between each seed point ROI and the time series of each voxel in the brain is calculated using the corr function of matlab, and the value of the obtained Pearson correlation coefficient ranges from -1 to 1. The negative values in the Pearson correlation coefficient are assigned a value of 0, i.e., the values in the range of -1 to 0 are assigned a value of 0, and the range of the Pearson correlation coefficient is changed to 0 to 1. In the core language area determination module, the threshold value analysis method is as follows: A first threshold value [0.50-0.99] is selected, and the voxel counts of the left brain and the right brain within the first threshold value range are counted. In the classification module, if LI≥0.20, the patient is considered to have left language dominance, if LI≤-0.20, the patient is considered to have right language dominance, and if -0.20<LI<0.20, the patient is considered to have bilateral language dominance.

Images