CN111407276A - Task state functional magnetic resonance individualized target positioning method - Google Patents

Abstract

The invention relates to a task state functional magnetic resonance individualized target spot positioning method, which comprises the steps of scanning functional magnetic resonance of a tested person to obtain functional magnetic resonance data; preprocessing functional magnetic resonance data; acquiring the active brain region activation degree of the tested person according to the processed functional magnetic resonance data; and taking the active surface brain area as an interested area, performing functional connection analysis based on seed points on all voxels in the interested area, calculating the functional connection between each voxel and the active deep brain area action point, and selecting the voxel with the strongest functional connection and closest to the surface of the scalp as a TMS target point. The invention has the advantages that: the brain area of the target point with the strongest connection with the activation point of the deep brain area is selected through comprehensive analysis of the functional connection strength index and the distance index, the target point selection has objective basis, is not influenced by subjective factors, and is accurate.

Description

An individualized target localization method for task-state fMRI

technical field

The invention relates to the technical field of medical imaging, in particular to a task-state functional magnetic resonance individualized target location method.

Background technique

Repetitive transcranial magnetic stimulation (rTMS) is an effective, drug-alternative treatment. rTMS utilizes pulsed magnetic fields to act on the central nervous system, affecting metabolism and neural electrical activity in the brain. Because of its non-invasive and safe characteristics, it has been approved by the FDA for clinical treatment of depression. Before performing rTMS, it is necessary to locate the therapeutic stimulation targets. The traditional "5-CM" manual positioning method is used for the positioning method. Using this positioning method, 1/3 of the targets located in the dorsolateral prefrontal lobe area. In addition, most of the stimulation targets of rTMS therapy are the dorsolateral prefrontal cortex, so it can be considered that the targets obtained by traditional localization methods are imprecise and non-individualized, which affects the therapeutic effect. The current method of functional connection also requires specialized professionals to operate step by step, which is labor-intensive and time-consuming, and uses fixed-coordinate brain regions as effect points rather than individualized activation points.

SUMMARY OF THE INVENTION

The invention mainly solves the above problems, and provides a method of finding the strongest activation point in the deep brain area activated by each subject, and using the method of functional connection to find the strongest functional connection with the activation point in the active surface brain area To determine the task-state fMRI individualized target localization method of TMS targets.

The technical solution adopted by the present invention to solve the technical problem is, a task-state functional magnetic resonance individualized target location method, comprising the following steps:

S1: scan the subject's fMRI to obtain fMRI data;

S2: preprocessing the fMRI data;

S3: Obtain the activation degree of the subject's active brain regions according to the processed fMRI data;

S4: Take the active surface brain region as the region of interest, perform functional connectivity analysis based on seed points on all voxels in the region of interest, calculate the functional connectivity between each voxel and the active deep brain region, and select the most functional connectivity. The voxels that are strong and closest to the scalp surface will be TMS targets.

The fMRI data can obtain the blood oxygen dependence level index of the brain in real time, which reflects the metabolic intensity of different brain regions, and can indirectly reflect the real-time neural activity of the brain. In this way, the degree of active brain area of the subject can be screened. The target point of rTMS is the deep brain area, while TMS technology can only directly act on the surface brain area of the scalp, and cannot reach the deep brain area. Therefore, functional connectivity analysis is needed to determine the optimal TMS targets in order to achieve the purpose of effectively regulating target brain regions.

As a preferred solution of the above solution, in the step S1, the subject needs to watch a picture that can activate the deep brain region when performing the functional magnetic resonance scan. Usually it is a picture with emotional expressions. It should be noted that this step is not to cause the emotional changes of the subjects, but to let them watch emotional expressions.

As a preferred solution of the above solution, the preprocessing in step S2 includes head movement correction, registration and spatial smoothing. A continuous fMRI experimental data has a continuous time dimension, and the head movement correction process is to remove the head movement artifacts in the time series; registration refers to the registration of spatial position and deflection direction, which is for different batches of fMRI. The spatial matching method between batches of resonance data is also a spatial matching method for registering brain imaging data templates of different subjects to a standard template; spatial smoothing is a denoising method that uses spatial filters to spatially filter functional magnetic resonance data. .

As a preferred solution of the above scheme, the TMS target selection in the step S4 includes the following steps:

S41: Filter the time series of all voxels in the region of interest and the time series of the voxels of the strongest activation point in the deep brain region;

S42: Calculate the correlation coefficient between each voxel in the region of interest and the voxel of the strongest activation point in the deep brain region as a functional connection strength indicator;

S43: Obtain the shortest distance between each voxel in the region of interest and the scalp surface as a distance indicator;

S44: Calculate the mixed index according to the distance index and the functional connection strength index;

S45: Select the voxel with the largest mixed index in the region of interest as the TMS target brain region.

By obtaining the functional connection strength between the voxels in the region of interest and the voxels with the strongest activation point in the deep brain region, and then synthesizing the shortest distance between the voxels in the region of interest and the scalp surface, the strongest point body in the region of interest and the deep brain region is screened out. The voxels of the region of interest with high strength of voxel functional connectivity and close to the scalp surface, namely the TMS target brain region.

As a preferred solution of the above solution, the filtering method adopted in step S41 is band-pass filtering.

As a preferred solution of the above scheme, the mixed index calculation formula in the step S44 is:

Mixing index = functional connection strength index + 1 / distance index * mixing coefficient

Among them, the functional connection strength index is the distance correlation coefficient, which is a dimensionless number; the distance index is the Euclidean distance, in millimeters; the mixing coefficient is a priori parameter, which is taken as 0.1.

As a preferred solution of the above scheme, after selecting the TMS target, the activation position of the amygdala and the position of the brain region of the TMS target are displayed by a visualization method. Visualization methods include three-view section views and three-dimensional semi-transparent views.

As a preferred solution of the above solution, before step S1 is performed, the subject needs to be screened for the safety of functional magnetic resonance scanning.

The advantages of the invention are: the target brain region with the strongest connection with the activation point of the deep brain region is selected through comprehensive analysis of the functional connection strength index and the distance index, the target point selection has an objective basis, is not affected by subjective factors, and the target point selection is accurate.

Description of drawings

FIG. 1 is a schematic flowchart of a task-state functional magnetic resonance individualized target localization method in an embodiment.

FIG. 2 is a schematic flow chart of the TMS target selection in the embodiment.

Detailed ways

The technical solutions of the present invention will be further described below through examples and in conjunction with the accompanying drawings.

Example:

In this embodiment, a task-state functional magnetic resonance individualized target location method is used for the preparation work before rTMS, as shown in FIG. 1 , and includes the following steps:

S1: scan the subject to obtain functional magnetic resonance data; before performing step S1, the subject needs to be screened for the safety of the functional magnetic resonance scan; before performing this step, the subject needs to complete the emotional expression The task, the emotional expression task requires the subject to watch pictures that can activate the deep brain region. The deep brain region studied in this example is the amygdala, and the corresponding picture for activating the amygdala is specifically the picture with the content of the horror expression. When viewing the picture, the subjects also need to judge the horror degree of the horror expression in the picture. If they think the expression in the picture is frightening, they should press button 1. If they think the expression in the picture is not very frightening, they should press button 2. , the emotional expression task is not to arouse the panic of the subjects, but to let the subjects look at the horror expressions, thereby activating the amygdala of the subjects. As for the judgment of the degree of panic of the horror expressions, it is only to allow the subjects to pay attention. Focusing on viewing pictures, the specific panic level judgment result has no influence on the subsequent operations of this method.

S2: Preprocessing the fMRI data, including head motion correction, registration and spatial smoothing. The head movement correction process is to remove the head movement artifacts in the time series; the registration refers to the registration of the spatial position and the deflection direction. The spatial matching method is to register the tested brain imaging data template to the standard template; the spatial smoothing is a denoising method that uses a spatial filter to spatially filter the fMRI data.

S3: According to the processed fMRI data, the activation degree of the active surface brain regions and the action points of the amygdala can be obtained; the fMRI data can obtain the blood oxygen dependence level index of the brain in real time, reflecting the metabolism of different brain regions. Intensity, which can indirectly reflect the real-time neural activity of the brain. Since the subjects performed the emotional expression task during the scan, we were interested in the location of the surface brain regions active during the task and the degree of activation of the amygdala. The analysis of amygdala activation can be divided into the following four steps:

S31: Establish a design matrix. The design matrix is a key indicator that reflects the corresponding relationship between the experimental conditions and the fMRI (functional magnetic resonance imaging) data in the time domain. The experimental conditions are expressed as a column of time series in the form of a mathematical matrix. There is a one-to-one correspondence between the above and the collected fMRI data, and this correspondence can help to establish a generalized linear model.

S32: Generalized linear model analysis, the experimental condition time series is regarded as the dependent variable, and the fMRI time series of different voxels (spatial positions) are regarded as independent variables, and a generalized linear model is established to solve the contribution weights of different voxels to the experimental conditions. The weight values correspond to voxels one-to-one, reflecting the importance of a specific voxel under a specific task. Statistical analysis of these weights can estimate the activation level.

S33: Statistical analysis. Since fMRI data often has the characteristics of large number of voxels (space) and small number of samples (time), the weight value estimated by the generalized linear model needs to be tested by statistical analysis to effectively reduce misjudgment. In this embodiment, FDR (false discovery rate) and FWE (family-wise error rate) checks are used.

S34: Result presentation, the verified weight value can be considered to reflect the real activation level of each voxel, and the value is mapped to the brain structural image through a visualization method, which can intuitively display the activation position and intensity of the amygdala.

S4: Take the active superficial brain region (dorsolateral prefrontal cortex) as the region of interest, perform functional connectivity analysis based on seed points on all voxels in it, and calculate the functional connectivity between each voxel and the amygdala action point. The strongest voxel closest to the scalp surface will be used as the TMS target. This example is the preparatory work before rTMS. In this example, the target action point of rTMS is the deep amygdala brain area, and TMS technology can only act directly. It is located in the surface brain area of the scalp and cannot reach the amygdala brain area. Therefore, it is necessary to carry out functional connectivity analysis to determine the optimal scalp stimulation brain area in order to achieve the purpose of effectively regulating the target brain area. TMS target selection is shown in Figure 2, including the following steps:

S41: Band-pass filtering is performed on the time series of all voxels in the region of interest and the time series of the most active point voxels in the amygdala, and the pass-band cutoff frequency is 0 Hz to 0.1 Hz;

S42: Calculate the correlation coefficient between each voxel in the region of interest and the voxel of the strongest activation point in the amygdala as an indicator of functional connectivity strength;

S43: Obtain the shortest distance between each voxel in the region of interest and the scalp surface as a distance index, and the distance index can be obtained directly from the fMRI scan result;

S44: Calculate the mixed index according to the distance index and the functional connection strength index, and the calculation formula of the mixed index is:

Mixing index = functional connection strength index + 1 / distance index * mixing coefficient

Among them, the functional connection strength index is the distance correlation coefficient, which is a dimensionless number; the distance index is the Euclidean distance, in millimeters; the mixing coefficient is a priori parameter, taking 0.1;

S45: Select the voxel with the largest mixed index in the region of interest as the TMS target brain region.

After determining the TMS target brain region, the activation position of the amygdala and the position of the TMS target brain region were displayed by three-view cross-sectional view or three-dimensional translucent map and other visualization methods.

The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which the present invention pertains can make various modifications or additions to the described specific embodiments or substitute in similar manners, but will not deviate from the spirit of the present invention or go beyond the definitions of the appended claims range.

Claims (8)

1. a task state fMRI individualized target location method is characterized in that: comprise the following steps: S1: scan the subject's fMRI to obtain fMRI data; S2: preprocessing the fMRI data; S3: Obtain the activation degree of the subject's active brain regions according to the processed fMRI data; S4: Take the active surface brain region as the region of interest, perform functional connectivity analysis based on seed points on all voxels in the region of interest, calculate the functional connectivity between each voxel and the active deep brain region, and select the most functional connectivity. The voxels that are strong and closest to the scalp surface will be TMS targets. 2 . A task-state fMRI individualized target location method according to claim 1 , wherein in the step S1 , when the subject performs fMRI scanning, it is necessary to watch the deep brain regions that can activate picture of. 3 . The task-state fMRI individualized target location method according to claim 1 , wherein the preprocessing in step S2 includes head movement correction, registration and spatial smoothing. 4 . 4. A task-state fMRI individualized target location method according to claim 1, wherein: in the step S4, the TMS target selection comprises the following steps: S41: Filter the time series of all voxels in the region of interest and the time series of the voxels of the strongest activation point in the deep brain region; S42: Calculate the correlation coefficient between each voxel in the region of interest and the voxel of the strongest activation point in the deep brain region as a functional connection strength indicator; S43: Obtain the shortest distance between each voxel in the region of interest and the scalp surface as a distance indicator; S44: Calculate the mixed index according to the distance index and the functional connection strength index; S45: Select the voxel with the largest mixed index in the region of interest as the TMS target brain region. 5 . A task-state functional magnetic resonance individual target location method according to claim 4 , wherein the filtering method adopted in step S41 is band-pass filtering. 6 . 6. A task-state fMRI individualized target location method according to claim 4, characterized in that: the mixed index calculation formula in the step S44 is: Mixing index = functional connection strength index + 1 / distance index * mixing coefficient Among them, the functional connection strength index is the distance correlation coefficient, which is a dimensionless number; the distance index is the Euclidean distance, in millimeters; the mixing coefficient is a priori parameter, which is taken as 0.1. 7. A task-state fMRI individualized target location method according to claim 1, characterized in that: after selecting the TMS target, the activation position of the amygdala and the brain region position of the TMS target are carried out by a visualization method. show. 8 . The task-state fMRI individualized target location method according to claim 1 , wherein, before step S1 is performed, the subject needs to be screened for safety of fMRI scanning. 9 .

Images