KR102120089B1 - Model for predicting recovery of stroke patient - Google Patents

Abstract

Information on lesions obtained from brain imaging data of stroke patients is applied to the dormant state of functional fMRI (fMRI) to determine the primary connection information of the entire brain area, and based on the primary connection information, the effect on lesions Disclosed is a method of predicting the degree of recovery of a stroke patient by obtaining secondary connection information corresponding to a connection lower than a predetermined criterion, and generating a recovery prediction model for the lesion based on the initial degree of disability and the secondary connection information do.

Description

Prediction model for exercise recovery in stroke patients {MODEL FOR PREDICTING RECOVERY OF STROKE PATIENT}

The disclosed embodiment relates to a recording medium recording a program for executing a method for predicting the degree of recovery of a stroke patient, a device for predicting a degree of recovery of a stroke patient, and a method for predicting a degree of recovery of a stroke patient in a computer.

Stroke is the cause of acquired disorders. Stroke-related disorders dramatically reduce a patient's daily life and quality of life. Therefore, an efficient rehabilitation plan can be important to the patient's future life. In particular, rehabilitation prediction can help clinicians develop individualized rehabilitation plans and help patients set more realistic goals. Brain tissue is characterized by a complex network. Therefore, the damage caused by stroke spreads through the brain tissue, and the brain structure is damaged from the major lesions, but it can also affect the function of distant brain regions.

Recent rehabilitation predictions are being studied through lesions and connectivity analyses, and several important factors have been identified. However, prediction is still a difficult task because of the diversity among individuals. Accordingly, there is a need for a more accurate prediction model capable of predicting the degree of recovery of stroke patients, regardless of individual diversity.

The disclosed embodiment is a method and apparatus for predicting a degree of recovery for improving the accuracy of recovery prediction by using secondary connection information obtained through a lesion analysis network in addition to the initial disorder degree of stroke patients in order to predict the degree of recovery of stroke patients It is about.

According to an embodiment, a method for predicting the degree of recovery of a stroke patient applies primary information about lesions obtained from brain structural image data of a stroke patient to a dormant state functional MRI (fMRI) of a normal person, thereby connecting the entire area of the brain Determining information; Obtaining secondary connection information corresponding to a connection having an influence on a lesion lower than a predetermined criterion based on the primary connection information; A recovery prediction model may be generated based on secondary connection information of the lesion and initial motor dysfunction.

In a method of predicting the degree of recovery of a stroke patient according to an embodiment, the step of acquiring secondary connection information, secondary connection to one lesion based on an average value of secondary connection information of a plurality of normal people to one lesion Information can be obtained.

According to an embodiment, the method for predicting the degree of recovery of a stroke patient further includes obtaining information about the patient's age and lesion size, and generating a final recovery prediction model includes initial disability degree and secondary connection information Together, the recovery prediction model can be generated including the patient's age and lesion size.

According to an embodiment, a method of predicting the degree of recovery of a stroke patient includes: obtaining brain image data of a stroke patient through MRI technique; And performing correction of the head movement of the patient on the obtained brain image data.

In a method of predicting the degree of recovery of a stroke patient according to an embodiment, the secondary connection information extracts connection information that is not affected by the lesion or that the degree of influence of the lesion is less than a predetermined criterion. Secondary connection information for a lesion was obtained by comparing the size of the value of each connection with the reference connection information without a lesion to obtain the number of lower connections, and this quantitative value can predict the degree of recovery of stroke patients. .

According to an embodiment, the apparatus for predicting the degree of recovery of a stroke patient determines an initial degree of impairment from brain image data of a stroke patient, and provides information about lesions obtained from the brain image data of a normal person in a dormant state fMRI (functional MRI) Applying to determine the primary connection information of the entire brain area, based on the primary connection information, to obtain secondary connection information corresponding to the connection with a lower impact on the lesion than the predetermined criterion, initial failure level and secondary connection information Basically, a processor that generates a recovery prediction model for a lesion; And a memory that stores a recovery prediction model.

In an apparatus for predicting the degree of recovery of a stroke patient according to an embodiment, the processor may acquire secondary connection information based on an average value of a plurality of secondary connection information obtained from a plurality of normal people's primary connection information.

In an apparatus for predicting the degree of recovery of a stroke patient according to an embodiment, the processor acquires information about the patient's age and the size of the lesion, and the patient's age and the size of the lesion along with the initial degree of disability and secondary connection information A recovery prediction model may be generated based on.

In an apparatus for predicting the degree of recovery of a stroke patient according to an embodiment, the processor may acquire brain image data of the stroke patient through an MRI technique and perform correction of the head motion of the patient on the obtained brain image data have.

In an apparatus for predicting the degree of recovery of a stroke patient according to an embodiment, the secondary connection information extracts connection information that is not affected by the lesion or that the degree of influence of the lesion is less than a predetermined criterion. Secondary connection information for a lesion was obtained by comparing the size of the value of each connection with the reference connection information without a lesion to obtain the number of lower connections, and this quantitative value can predict the degree of recovery of stroke patients. .

1 is a view for explaining a change in the population structure according to aging.
2 is a graph for explaining the increase and decrease rate of stroke medical personnel.
3 is a view for explaining brain damage in a patient with a stroke.
4 is a view for explaining the connectivity of the brain.
5 to 7 are diagrams for explaining an existing study performed for rehabilitation prediction.
8 to 13 are views for explaining a damaged area of the brain due to a lesion.
14 is a view for explaining the results of the lesion area of the ischemia stroke patient.
15 to 17 are views for explaining the process of extracting the secondary connection of the lesion.
18 to 20 are diagrams for explaining a method of expressing a secondary connection to a patient's lesion in a matrix.
21 is a diagram for explaining a method of predicting an increase in recovery after 3 months by a recovery prediction apparatus according to an embodiment.
22 to 24 are graphs showing results of modeling by combining the initial degree of disability and the predictor.
25 is a view for explaining a method of verifying the modeling performed by combining the initial failure degree and the predictor according to an embodiment.
26 is a graph for explaining the performance of modeling performed by combining an initial failure level and a predictor according to an embodiment.
27 is a block diagram of an apparatus for predicting recovery according to an embodiment.

Terms used in the specification will be briefly described, and the present invention will be described in detail.

The terminology used in the present invention has been selected, while considering the functions in the present invention, general terms that are currently widely used have been selected, but this may be changed according to the intention or precedent of a person skilled in the art or the appearance of new technologies. In addition, in certain cases, some terms are arbitrarily selected by the applicant, and in this case, their meanings will be described in detail in the description of the applicable invention. Therefore, the terms used in the present invention should be defined based on the meanings of the terms and the contents of the present invention, not simply the names of the terms.

When a part of the specification "includes" a certain component, this means that other components may be further included instead of excluding other components, unless specifically stated to the contrary. In addition, terms such as "... unit" and "module" described in the specification mean a unit that processes at least one function or operation, which may be implemented in hardware or software, or a combination of hardware and software. .

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

1 is a view for explaining a change in the population structure according to aging.

Referring to FIG. 1, the aging trend over time and the structure of the number of populations by age occupying the entire population are illustrated. In the future society, as the aging population progresses rapidly, the population over 55 will make up a large portion of the population.

Due to such a change in population, such as aging, the number of patients with physical disabilities is gradually increasing. For example, as the number of stroke patients increases, the number of patients with physical disabilities may increase. Therefore, the role of rehabilitation treatment for recovery of stroke patients is increasing.

Rehabilitation therapy is very important for stroke patients in order to return stroke patients to their daily lives, or to reduce the extent of initial disability as much as possible. In this regard, rehabilitation prediction can help therapists develop personalized rehabilitation plans and help patients set more realistic goals.

2 is a graph for explaining the increase and decrease rate of stroke medical personnel.

Referring to FIG. 2, it can be seen that while the number of deaths per 100,000 people is decreasing, the number of medical personnel due to stroke confirmed in the emergency room is steadily increasing. That is, it can be confirmed that the invention of stroke is increasing.

3 is a view for explaining brain damage in a patient with a stroke.

Referring to FIG. 3, brain damage spreads through a neural network, and brain damage in the brain structure mainly corresponds to a significant lesion, but stroke affects distant brain tissue in addition to the significant lesion. Can give.

4 is a view for explaining the connectivity of the brain.

Referring to FIG. 4, rest-state fMRI is based on a spontaneous low-frequency signal lower than 0.1 Hz in a blood-oxygen level dependent signal. This spontaneous low-frequency signal may be a meaningful brain activity despite no external stimulus.

In particular, the resting fMRI has fewer requirements than the fMRI based on a specific task, and thus may be more suitable for patients with neurological disorders.

5 to 7 are diagrams for explaining an existing study performed for rehabilitation prediction.

Recently, many studies have been conducted on rehabilitation prediction using lesion and linkage analysis, and some predictors have proved it. For example, FIG. 5 is a graph for explaining an approach for predicting recovery of function with lesion data over time. In addition, FIG. 6 describes a study of the effect of lesions based on node removal and edge removal according to an attack. In addition, FIG. 7 describes a study focusing on the connection of lesion-center damage.

8 to 13 are views for explaining a damaged area of the brain due to a lesion.

Referring to Figure 8, it can be seen that the brain is composed of a complex network. On the other hand, as stroke occurs, as shown in FIG. 9, certain areas of the brain may be damaged by major lesions. When a specific region of the brain is damaged by a major lesion, the primary neural network may be damaged, as shown in FIG. 10. The damage generated in the primary neural network can spread through the brain network, as shown in FIG. 11. In addition, as the damage spreads through the brain network, the secondary neural network may be affected by the lesion as shown in FIG. 12. That is, due to the spread of the damage, as shown in Figure 13, the secondary neural network can be strongly affected by the lesion.

Secondary neural networks have a larger number of connections, and can generally cover more areas of the brain when transferring information to a network structure. Secondary neural networks are not directly damaged by major lesions.

The recovery prediction apparatus according to an embodiment used a secondary connection of the lesion to predict rehabilitation. The characteristic of the secondary connection is that it is relatively strongly affected by lesions, and the connection is much larger in terms of information dissemination in the network structure and covers a large area. Also, the secondary connection represents a connection that is not directly damaged.

FIG. 14 is a view for explaining the result of imaging the lesion area of an ischemia stroke patient.

FIG. 14 shows the results obtained by performing MRI data collection and behavioral testing for 2 weeks after stroke expression for stroke patients, excluding patients with neuropsychiatric disorders other than stroke.

Movement disorders were measured on the day of MRI data acquisition using FMA (Fugl-Meyer assessment). It was measured again 3 months after the onset of stroke. Participants were instructed to close their eyes and no motion during resting-state fMRI scans.

At this time, fMRI data were obtained through a Philips ACHIEVA MR scanner.

15 to 17 are views for explaining the process of extracting the secondary connection of the lesion.

Referring to FIG. 15, the recovery prediction apparatus according to an embodiment may grasp the connectivity of the entire brain to time-series information about the lesion by moving the lesion of one patient onto resting-state fMRI data of a normal person through correlation analysis. And this link becomes the primary link to the lesion.

The recovery prediction apparatus according to an embodiment first divides the entire brain into 94 regions on the basis of the Automated Anatomical Labeling (AAL) template, and if there is a voxel connected by the primary connection of the lesion of the region, there is no time series information and lesions of the voxels, Secondary connections can be identified through correlation analysis between areas that had no primary connection to the lesion. In this case, a voxel is a 3D meaning of a pixel and is a minimum unit having a value on MRI data.

The recovery prediction apparatus may input a value related to the secondary connection into the matrix. The recovery prediction device can perform this process for all 64 normal people. In other words, the recovery prediction apparatus can extract secondary connection information of these lesions from 64 normal people for a patient's lesion. That is, 64 matrices can be obtained for one lesion.

Referring to FIG. 17, a process of obtaining a secondary linking process for a lesion from a plurality of normal persons and taking an average by the number of the plurality of normal persons is a secondary linking to the lesion.

18 to 20 are diagrams for explaining a method of obtaining a connection having a low influence on a lesion by comparing a secondary connection to a patient's lesion with a reference matrix, and expressing the result in a matrix.

Referring to FIG. 18, in the matrix, an area corresponding to an odd number may be rearranged as a lesion-side area, and an area corresponding to an even number as a key-side area.

At this time, asymmetric adjacent matrices can be modified into symmetric matrices by adding transpose matrices.

The recovery prediction apparatus according to an embodiment uses a reference network extracted from a normal person without a lesion to relatively evaluate the impact of each connection, and a value smaller than the value of the connection is set to 1 and a value larger than 0 is set to a weighting matrix (weighted) matrix) into a binary matrix, where you can count the number of connections marked as 1 among the half-section connections and use them as predictors.

The larger the number of these connections, the lower the impact of lesions on the entire brain, which can be a better condition for a better recovery and better recovery.

19 to 20, as described above, a secondary connection of each lesion is obtained, and a result of dividing a reference network and changing a small value to 1 and a large value to 0 is shown.

In addition, FIG. 21 shows the result of calculating the increase in recovery at the time of 3 months by placing the number of 1s in the half-section connections on the horizontal axis.

21 is a diagram illustrating a result of predicting an increase in recovery after 3 months by a recovery prediction apparatus according to an embodiment.

Referring to FIG. 21, it can be confirmed through the graph (D) that the proposed predictive factor can predict the recovery increase after 3 months. In addition, graphs (E) and (F) are graphs for dividing lesions into supratentorial and infratentorial, respectively, and it can be seen that they have strong predictive power, especially in patients with supratentorial lesions.

22 to 24 are graphs showing the results of modeling, including the initial disability level and predictor, and the initial disability level and predictor, and the patient's age and lesion size.

Referring to FIG. 22, according to an embodiment, the apparatus for predicting recovery may combine initial failure level and predictor through a multiple linear regression model. The recovery prediction device can increase the predictive power of rehabilitation by generating a model that predicts the patient's motor function score at the 3 month time point based on the combined results.

The recovery prediction device may perform lesion network analysis, measure the influence of the lesion on the entire brain area through brain neural network analysis, and use it as a predictor. The recovery prediction device sequentially acquired the primary and secondary neural networks of the lesion, and measured the influence of the lesion on the entire brain area. In particular, it was confirmed that the influence of the lesion was effective in predicting the increase in recovery by focusing on the secondary neural network. In addition, this predictive factor has characteristics that have higher performance and reproducibility than the previous one in combination with the degree of initial failure.

23 and 24, according to an embodiment, the apparatus for predicting recovery may combine the patient's age and the size of the lesion through a multiple linear regression model in addition to the initial disability level and the predictive factors. The recovery prediction apparatus can finally make a prediction model with an accuracy close to 0.8.

25 is a view for explaining a method of verifying the modeling performed by combining the initial failure degree and the predictor according to an embodiment.

Referring to FIG. 25, the apparatus for predicting recovery according to an embodiment may allocate test data to verify a model. The recovery prediction device can train the predictive model through the data. The recovery predictor uses LOOCV, which uses one test, and the rest as training data, and k-fold cross-validation, which uses 5 data as training data and performs the rest of the data as training data. Can be verified.

26 is a graph for explaining the performance of modeling performed by combining an initial failure level and a predictor according to an embodiment.

Referring to FIG. 26, the apparatus for predicting recovery may compare the results of performing prediction by modeling based on the degree of initial failure and the results of performing prediction by modeling based on the initial degree of failure and the predictor. As a result of comparison, it can be seen that the accuracy of the result of the prediction is improved by combining the initial obstacle level and the predictor.

The apparatus for predicting recovery according to an embodiment may finally process a spatially calm standardized image using a 6-mm full-width half-maximum Gaussian kernel. At this time, some of the signal elements to be removed can be removed using 22 linear regressions. The factor may include six head motion factors and six factors primarily derived from the motion factors. Here, each of the five factors is obtained and included through principal component analysis for white matter and cerebrospinal fluid signals. Through this signal, it is possible to extract a significant signal, including as an obstructive regressor.

In addition, the recovery prediction apparatus may perform band-pass filtering at a frequency in the range of 0.009 Hz to 0.08 Hz in order to remove integer offset and linear trend. In addition, the recovery prediction apparatus calculated the statistical dependence using the Pearson's correlation coefficient, and confirmed connectivity as the degree of progress over time was calculated. At this time, only the positive correlation coefficient can be used as the correlation coefficient.

The recovery prediction apparatus may use a seed-based approach for analyzing connectivity to voxels. In this case, the reference point of the connectivity map may be a t-value.

In addition, the recovery prediction apparatus may divide the brain into 116 regions using AAL (Automated Anatomical Labeling) to cover the entire brain. At this time, 116 zones may include 90 zones in the cerebrum and 26 zones in the cerebellum. The recovery predictor averaged 9 regions in the left cerebellum and right cerebellum respectively, and averaged 8 regions in the lobe. The brainstem zone can be drawn passively over the AAL atlas.

Therefore, a modified AAL atlas composed of 94 regions can be registered and analyzed together using SPM12 in a normalized fMRI data space.

Each lesion from the patient can be distinguished over a T1 weighted structure image associated with a high signal intensity diffusion-weighted image (DWI) obtained at the patient's first neuropathic practice.

The recovery prediction apparatus according to an embodiment may use a region obtained from lesion-seeded functional connectivity analysis to configure the next seeds, that is, second link-step connectivity, to use the secondary connection as a motion recovery prediction factor. Each step to obtain a secondary connection may be as follows.

(i) Transfer the lesion to a normal object.

(ii) Obtain lesion-seeded functional connectivity of normal subjects through a seed-based approach. (first link-step, voxel-wise connectivity, p<0.00005).

(iii) Lesion-seeded connected to voxel is divided into 7 regions based on modified AAL atlas.

(iv) Each region is set as the next seed region, and region-wise connectivity analysis is performed in the modified AAL space (second link-step, region-wise connectivity, p<0.01). At this time, the lesion and the lesion-seeded connected voxel can be masked.

(v) Calculate the correlation between the seeded region and the modified AAL regions. By inserting the value for the correlation into the column matching the next seed region, we can construct an adjacency matrix. At this time, in the matrix, the odd-numbered region may change from the left region to the same damage region (ipsilesional region), and the even-numbered region may change from the right region to the opposite damage region (contralesional region).

(vi) Asymmetric adjacent matrices obtained in step (v) can be modified into symmetric matrices by adding transposes (A_P1Hi= A_P1Hi+A_P1Hi'), if seeds are located in regions 1 and 2, region 1 And two different correlation coefficient values in region 2. Here, it is assumed that the coefficient reflects the effect of the lesion. If the next seed is located in regions 1 and 2, the effects of lesions between regions 1 and 2 can be summed up. The recovery prediction apparatus can express the effects of lesions in regions 1 and 2 as a summed value by creating an asymmetric matrix.

(vii) Adjacent matrices can be obtained from healthy objects, and all matrices can be averaged. The averaged matrix corresponds to the second order connection obtained from one lesion. The secondary connection matrix for stroke patients can be defined as follows.

[Equation 1]

는 i번째 정상 객체로부터 획득된 2차 연결 행렬을 나타내고, n은 정상 객체의 수를 나타낸다. In Equation 1 above

Denotes a second order connection matrix obtained from the i-th normal object, and n represents the number of normal objects.

The recovery predictor can compare unconnected or weak links to a reference reference link as a predictor of motor recovery to measure the effect on the brain from lesions.

As a result of comparison, the recovery prediction device is less connected, or the larger the number of weak connections, the less the lesions have on the brain nerves, and the brain nerves are in an advantageous state to recover from damage from stroke, and as a result, they are good for recovering motor nerves. You can decide to have an effect.

[Equation 2]

는 환자의 2차 연결 행렬을 나타내고,

는 참조 기준 연결 행렬을 나타내며, m은 영역의 개수를 나타낸다. 참조 연결 행렬은 병변 볼륨에 의해 마스킹 된 모든 정상 객체의 AAL 연결 행렬을 평균하여 획득될 수 있다. 참조 연결 세기 보다 강도가 낮은 연결은 병변의 약한 부분으로 정의될 수 있고, 세기가 없는 연결은 병변에 영향을 받지 않는 연결로 정의될 수 있다. 그러므로 회복 예측 장치는 강도가 낮은 연결 및 병변에 영향을 받지 않는 연결의 개수에 대응되는 예측 인자를 이용할 수 있다. In Equation 2 above,

Denotes the patient's quadratic connection matrix,

Denotes a reference reference connection matrix, and m denotes the number of regions. The reference connection matrix can be obtained by averaging the AAL connection matrix of all normal objects masked by the lesion volume. Connections with a strength lower than the reference connection strength may be defined as a weak portion of the lesion, and connections without an intensity may be defined as connections not affected by the lesion. Therefore, the recovery prediction apparatus may use a prediction factor corresponding to the number of low-intensity connections and the number of connections not affected by the lesion.

The apparatus for predicting recovery according to an embodiment may perform univariate analysis through a linear regression model. The recovery predictor examines the relationship between various variables (patient age, lesion size, baseline, predicted factor) and motor capacity recovery, and according to the survey results, the relationship between age, lesion size and predictor factor and motor capacity recovery Can decide.

In addition, the recovery prediction apparatus can classify patients into tentative and sub-tentative lesions, and examine the relationship between predictive factors and motor capacity recovery. As a result, it was confirmed that the predictive factor was particularly accurate for patients with lesions on the tent. In addition, the apparatus for predicting recovery may investigate tent lesions through multivariate analysis. The recovery predictor combines the initial degree of disability and the predictor, and as a result of performing the analysis, it is confirmed that the accuracy of predicting athletic ability is improved. In addition, as a result of analysis using a multiple linear regression model, including factors such as age, lesion size, as well as initial disability level and predictive factors, it was confirmed that the prediction accuracy was the highest. On the other hand, according to the results of the multivariate analysis, it was confirmed that the initial degree of impairment had a strong influence on the rehabilitation prediction, and that the proposed predictors also retained meaning in the multivariate analysis, indicating that the impact level was high.

The recovery prediction device according to an embodiment may apply LEOCV (Leave-one-out cross validation) and k-fold cross validation to confirm the effectiveness of the multivariate analysis through the initial failure level and the proposed predictor.

It shows the relationship between the predicted FMA score and the actual FMA score 3 months after brain hemorrhage. As a result of comparing the change in the sum of the two models, the combined model and the multivariate model, to compare the accuracy of the prediction, it can be seen that the overall interval is higher than that of the multivariate model. In particular, it can be seen that the determination constants of the two models are much larger in patients with large initial damage. In addition, it was confirmed that the combined model prediction accuracy was greater than the univariate analysis.

27 is a block diagram of a recovery prediction apparatus 2700 according to an embodiment.

Referring to FIG. 27, the recovery prediction apparatus 2700 may include a processor 2710 and a memory 2720. According to the recovery prediction method proposed in the above embodiments, the processor 2710 and the memory 2720 may operate. However, the components of the recovery prediction apparatus 2700 according to an embodiment are not limited to the above-described example. According to another embodiment, the recovery prediction apparatus 2700 may include more components or fewer components than the above-described components.

The processor 2710 may determine the primary connection information of the entire brain area by applying information about the lesion obtained from the brain image data to the dormant functional MRI (fMRI) of the normal person. Based on the primary connection information, the processor 2710 may obtain secondary connection information corresponding to a connection in which an effect on a lesion is lower than a predetermined criterion. The processor 2710 may generate a recovery prediction model for lesions based on the initial degree of failure and secondary connection information.

The processor 2710 may acquire secondary connection information based on an average value of the plurality of secondary connection information obtained from the primary connection information of a plurality of normal people. The processor 2710 may acquire information regarding a patient's age and lesion size. The processor 2710 may generate a recovery prediction model based on the patient's age and the size of the lesion along with the initial degree of disability and secondary connection information.

The processor 2710 may acquire brain neural network data of a stroke patient through an MRI brain imaging technique. The processor 2710 may perform correction of the head motion of the patient on the acquired brain image data.

The memory 2720 may store information necessary for the processor 2710 to predict the degree of recovery of a stroke patient. For example, the memory 2720 may store brain image data of a stroke patient. Also, the memory 2720 may store a recovery prediction model.

The device according to the present invention includes a processor, a memory for storing and executing program data, a permanent storage such as a disk drive, a communication port communicating with an external device, a user interface such as a touch panel, keys, buttons, and the like. Devices and the like. Methods implemented by a software module or algorithm may be stored on a computer-readable recording medium as computer-readable codes or program instructions executable on the processor. Here, as a computer-readable recording medium, a magnetic storage medium (eg, read-only memory (ROM), random-access memory (RAM), floppy disk, hard disk, etc.) and optical reading medium (eg, CD-ROM (CD-ROM) ), DVD (Digital Versatile Disc). The computer-readable recording medium can be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion. The medium is readable by a computer, stored in memory, and can be executed by a processor.

All documents, including published documents, patent applications, patents, etc., cited in the present invention may be incorporated into the present invention in the same way that each cited document is individually and specifically merged or represented as a whole in the present invention. .

For the understanding of the present invention, reference numerals have been described in preferred embodiments illustrated in the drawings, and specific terms have been used to describe the embodiments of the present invention, but the present invention is not limited by the specific terms. Can include all of the components that are ordinarily considered by those skilled in the art.

The present invention can be represented by functional block configurations and various processing steps. These functional blocks can be implemented with various numbers of hardware or/and software configurations that perform specific functions. For example, the present invention is directed to integrated circuit configurations such as memory, processing, logic, look-up tables, etc., capable of executing various functions by control of one or more microprocessors or other control devices. Can be employed. Similar to the components of the present invention that can be executed with software programming or software elements, the present invention includes C, C++, including various algorithms implemented with a combination of data structures, processes, routines or other programming configurations. , Can be implemented in programming or scripting languages such as Java, assembler, etc. Functional aspects can be implemented with algorithms running on one or more processors. In addition, the present invention may employ conventional techniques for electronic environment setting, signal processing, and/or data processing. Terms such as "mechanism", "element", "means", and "configuration" can be used broadly and are not limited to mechanical and physical configurations. The term may include the meaning of a series of routines of software in connection with a processor or the like.

The specific implementations described in the present invention are exemplary embodiments, and do not limit the scope of the present invention in any way. For brevity of the specification, descriptions of conventional electronic configurations, control systems, software, and other functional aspects of the systems may be omitted. In addition, the connection or connection members of the lines between the components shown in the drawings are illustrative examples of functional connections and/or physical or circuit connections. In the actual device, alternative or additional various functional connections, physical It can be represented as a connection, or circuit connections. In addition, unless specifically mentioned, such as "essential", "important", etc., it may not be a necessary component for the application of the present invention.

The use of the term "above" in the specification (particularly in the claims) of the present invention and similar indication terms may be in both singular and plural. In addition, in the case of describing a range in the present invention, as including the invention to which the individual values belonging to the range are applied (if there is no contrary description), each individual value constituting the range is described in the detailed description of the invention. Same as Finally, unless there is a clear or contradictory description of the steps constituting the method according to the invention, the steps can be done in a suitable order. The present invention is not necessarily limited to the description order of the above steps. The use of all examples or exemplary terms (eg, etc.) in the present invention is merely for describing the present invention in detail, and the scope of the present invention is limited due to the examples or exemplary terms, unless it is defined by the claims. It does not work. In addition, those skilled in the art can recognize that various modifications, combinations, and changes can be configured according to design conditions and factors within the scope of the appended claims or equivalents thereof.

Claims (11)

Determining, by the processor, an initial degree of disability based on brain image data of the stroke patient;
By the processor, information about the lesion obtained from the brain image data is applied to the dormant state functional MRI (fMRI) of a normal person, respectively, and the lesion of the stroke patient is set as a first seed in the brain network structure of the normal person Thereby determining the primary connection information of the normal person regarding the first plurality of regions that are primarily connected to the first seed;
By the processor, by setting the first plurality of regions in the brain network structure of the normal person as a second seed based on the primary connection information determined in the brain network structure of the normal person, secondary to the first seed Acquiring secondary connection information of the normal person with respect to a second plurality of regions connected and primarily connected to the second seed;
Determining, by the processor, the primary connection information of the normal person and obtaining the secondary connection information of the normal person, for the first normal person to the nth normal person, performing the first normal person to the nth normal person. Obtaining each secondary connection information;
Generating, by the processor, a recovery prediction model for the lesion of the stroke patient, using the secondary connection information of each of the first normal person to the n-th normal person, and the initial degree of disability of the stroke patient. and,
The n is an integer of 2 or more, characterized in that,
Generating the recovery prediction model,
Generating the recovery prediction model for the lesion by using an average value of secondary connection information of each of the first normal person to the nth normal person, and the initial disorder level,
Obtaining, by the processor, information regarding the age of the stroke patient and the size of the lesion of the stroke patient,
Generating the recovery prediction model,
By the processor, using the age of the stroke patient, the size of the lesion, the secondary connection information of each of the first normal person to the n normal person, and the initial degree of disability, predict the recovery of the lesion of the stroke patient Including creating a model,
Acquiring, by the processor, the brain image data of the stroke patient through an MRI technique; And
Further comprising, by the processor, performing correction of the head motion of the stroke patient on the obtained brain image data,
The secondary connection information,
It is characterized in that it is not affected by the lesion, or the degree affected by the lesion is less than a predetermined criterion,
Method for predicting the degree of recovery of stroke patients, characterized in that the following [Equation 1] is used to obtain the secondary connection information.
[Equation 1]

(

Denotes a quadratic connection matrix obtained from the i-th normal object, and n represents the number of normal objects) delete delete delete delete The initial degree of disability is determined based on brain imaging data of a stroke patient,
By applying the information about the lesion obtained from the brain image data to the dormant state functional MRI (fMRI) of the normal person, the first lesion is set by setting the lesion of the stroke patient as the first seed in the brain network structure of the normal person. Determining the primary connection information of the normal person with respect to the first plurality of regions primarily connected to the seed,
By setting the first plurality of regions as a second seed in the brain network structure of the normal person based on the primary connection information determined in the brain network structure of the normal person, the second seed is secondaryly connected to the second seed Acquiring secondary connection information of the normal person with respect to the second plurality of regions primarily connected to the seed,
Determining the primary connection information of the normal person and obtaining the secondary connection information of the normal person is performed for the first normal person to the nth normal person, and the secondary connection information of each of the first normal person to the nth normal person To obtain,
A processor configured to generate a recovery prediction model for the lesion of the stroke patient, using secondary connection information of each of the first normal person to the n-th normal person, and the initial degree of disability of the stroke patient; And
A memory for storing the recovery prediction model,
Wherein n is an integer of 2 or more,
The processor,
Further configured to generate the recovery prediction model for the lesion, using the average value of the secondary connection information of each of the first normal person to the n-th normal person, and the initial disorder level,
The processor,
Obtain information on the age of the stroke patient and the size of the lesion in the stroke patient,
Further using the age of the stroke patient, the size of the lesion, the secondary connection information of each of the first normal person to the n-th normal person, and the initial degree of disability, further to generate a recovery prediction model for the lesion of the stroke patient Composed,
The processor,
The brain image data of the stroke patient is obtained through MRI technique,
Further configured to perform correction of the head movement of the stroke patient on the obtained brain image data,
The secondary connection information,
Characterized in that the degree of influence from the lesion or not affected by the lesion is less than a predetermined criterion,
Device to predict the degree of recovery of stroke patients, characterized in that the following [Equation 1] is used to obtain the secondary connection information.
[Equation 1]

(

Denotes the second order connection matrix obtained from the i-th normal object, and n represents the number of normal objects.)
delete delete delete delete delete

Images