Method and system for analyzing cognitive reserve on basis of artificial intelligence learning model by using MRI image

Abstract

According to one embodiment of the present disclosure, provided is a method for analyzing a cognitive reserve value from a brain MRI image by utilizing a system including a cognitive reserve analysis device and a user terminal. A system for analyzing cognitive reserve can predict a subject's cognitive reserve value with high accuracy by using only a brain MRI image, and can take into consideration various factors that may affect the cognitive reserve value, thereby predicting the cognitive reserve value with higher accuracy.

Description

Cognitive Ability Analysis Method and System Based on Artificial Intelligence Learning Model Using MRI Images

[0001] The present disclosure relates to a method and system for analyzing cognitive reserve, and more particularly, to a method and system for analyzing cognitive reserve based on an artificial intelligence learning model using MRI images.

[0002] This application is the result of research conducted with the support of the Korea Health Industry Development Institute (No. HU20C0271) and the National IT Industry Promotion Agency (No. RS-2022-00155966, Artificial Intelligence Convergence Innovation Talent Development (Ewha Womans University), Project Identification Number: 1711179344) with funding from the government (Ministry of Health and Welfare and Ministry of Science and ICT) in 2024.

[0003]

[0004] The material described in this section merely provides background information for the present disclosure and does not constitute prior art.

[0005] Cognitive reserve is a concept based on brain plasticity, and it implies actively coping with the brain's changes associated with aging. Lifelong experiences, such as education, work, and leisure activities, are known to enhance cognitive reserve. Previous research has shown that high cognitive reserve can reduce the risk of dementia by approximately 46%. Therefore, assessing an individual's cognitive reserve can play a crucial role in predicting the onset and severity of dementia.

[0006] Traditionally, cognitive reserve has been measured through written questionnaires that ask questions about a patient's education level, occupation, and leisure activities. The most frequently used questionnaire in Korea is the Cognitive Reserve Index (CRI), a 20-item questionnaire administered through an interview that assesses education, occupation, and leisure activities known to be related to cognitive reserve.

[0007] However, the existing method had limitations in that it could cause recall bias depending on cognitive function in elderly patients.

[0008]

[0009] Accordingly, the present disclosure aims to provide a method and system capable of predicting the occurrence and progression of dementia by measuring the cognitive reserve of an elderly subject based on brain MRI images.

[0010] The problems to be solved by the present invention are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below.

[0011]

[0012] According to one embodiment of the present disclosure, a method for analyzing a cognitive reserve value from a brain MRI image using a system including a cognitive reserve analysis device and a user terminal, the method comprising: (a1) a step in which a brain MRI image input to the cognitive reserve analysis device is preprocessed by the cognitive reserve analysis device to generate brain functional connectivity data; (a2) a step in which the brain functional connectivity data and the cognitive reserve value input to the cognitive reserve analysis device corresponding thereto are stored as a learning information dataset; (a3) ​​a step in which a learning model using the brain functional connectivity data as input data and the cognitive reserve value as output data is learned by the cognitive reserve analysis device; (b1) a step in which a brain MRI image or brain functional connectivity data is input to the user terminal; And (b2) a step of transmitting the input brain MRI image or brain functional connectivity data to the cognitive reserve analysis device, and inputting the brain functional connectivity data generated by the cognitive reserve analysis device from the input brain MRI image or the input brain functional connectivity data into the learning model, thereby outputting a predicted cognitive reserve value; is provided.

[0013] Additionally, preferably, the brain functional connectivity data according to one embodiment of the present disclosure is a 90 x 90 matrix.

[0014] In addition, preferably, the step (a1) according to one embodiment of the present disclosure includes: (a10) a step in which the cognitive reserve analysis device spatially realigns the brain MRI image, normalizes it to the same coordinate frame, spatially smooths it using a Gaussian kernel, and removes noise in a preset frequency band, thereby preprocessing the brain MRI image; and (a11) a step in which the cognitive reserve analysis device segments the brain MRI image preprocessed by the step (a10) into 90 nodes.

[0015] In addition, preferably, the step (a1) according to one embodiment of the present disclosure further includes: (a12) a step in which the cognitive ability analysis device performs a Fisher Transformation on the 90 nodes segmented by the step (a11), thereby calculating partial correlation coefficients for two different pairs among the 90 nodes; and (a13) a step in which the cognitive ability analysis device calculates a symmetric matrix of size 90 x 90 using the partial correlation coefficients calculated in the step (a12).

[0016] In addition, according to another embodiment of the present disclosure, in the method according to the above embodiment, in the step (a2), personal information is further stored in the cognitive reserve analysis device, in the step (a3), the personal information is further used as input data, and in the step (b1), when personal information is further input to the user terminal and the cognitive reserve analysis device, a predicted cognitive reserve value is output.

[0017] According to one embodiment of the present disclosure, a system for analyzing cognitive reserve is provided, comprising: an MRI device configured to capture a brain MRI image of a subject; a cognitive reserve analysis device, comprising: an input unit that receives the brain MRI image and the subject's cognitive reserve value; a processor that preprocesses the brain MRI image to generate brain functional connectome data; and a learning processor configured to train a learning model that uses the brain functional connectome as input data and the cognitive reserve value as output data; and a user terminal that receives a brain MRI image or a brain functional connectome and transmits the same to the cognitive reserve analysis device; wherein the brain functional connectome data generated from the brain MRI image input to the user terminal or the brain functional connectome data input to the user terminal is transmitted and input to the cognitive reserve analysis device, and the cognitive reserve analysis device outputs a predicted cognitive reserve value using the learning model and transmits the predicted cognitive reserve value to the user terminal.

[0018] According to another embodiment of the present disclosure, in the system according to the above embodiment, the input unit of the cognitive reserve analysis device further receives personal information of the subject, the personal information is used as input data of the learning model, the personal information is further input into the user terminal, and the user terminal transmits the input personal information to the cognitive reserve analysis device.

[0019]

[0020] As described above, according to one embodiment of the present disclosure, the cognitive reserve analysis system has the effect of being able to predict the cognitive reserve level of a subject with high accuracy using only brain MRI images.

[0021] In addition, according to another embodiment of the present disclosure, various factors that may affect the cognitive reserve value can be considered, so that the cognitive reserve value can be predicted with higher accuracy.

[0022]

[0023] FIG. 1 is a block diagram of a cognitive ability analysis system according to embodiments of the present disclosure.

[0024] FIG. 2 is a block diagram of a cognitive ability analysis device according to one embodiment of the present disclosure.

[0025] FIG. 3 is a diagram illustrating a method for generating learning information of a cognitive ability analysis device according to one embodiment of the present disclosure.

[0026] FIG. 4 illustrates a method for predicting a cognitive reserve value using a cognitive reserve analysis device according to one embodiment of the present disclosure.

[0027] FIG. 5 is a flowchart of a method using a cognitive ability analysis system according to one embodiment of the present disclosure.

[0028] FIG. 6 is a block diagram of a cognitive ability analysis device according to another embodiment of the present disclosure.

[0029] FIG. 7 is a diagram illustrating a method for generating learning information of a cognitive ability analysis device according to another embodiment of the present disclosure.

[0030] FIG. 8 is a flowchart of a method using a cognitive reserve analysis system including a cognitive reserve analysis device according to another embodiment of the present disclosure.

[0031] FIG. 9 is a result of comparing the predicted cognitive reserve value and the actual cognitive reserve value using a cognitive reserve analysis device according to one embodiment of the present disclosure.

[0032]

[0033] Hereinafter, some embodiments of the present disclosure will be described in detail with reference to exemplary drawings. When designating components in each drawing, it should be noted that, where possible, identical components will be given the same reference numbers, even if they appear in different drawings. Furthermore, when describing the present disclosure, detailed descriptions of related known components or functions will be omitted if they are deemed to obscure the gist of the present disclosure.

[0034] In describing components of embodiments according to the present disclosure, symbols such as first, second, i), ii), a), b) may be used. These symbols are only for distinguishing the components from other components, and the nature, order, or sequence of the components are not limited by the symbols. When a part in the specification is said to "include" or "have" a component, this does not mean that other components are excluded, but rather that other components may be included, unless explicitly stated otherwise.

[0035] In this disclosure, “cognitive reserve” is a concept based on brain plasticity, and is a concept that represents the brain’s ability to actively cope with changes in the brain due to aging, i.e., brain resilience.

[0036] Additionally, in the present disclosure, "Cognitive Reserve Index" refers to the results of a cognitive reserve test. Here, the cognitive reserve test may be, but is not limited to, the Lifetime of Experiences Questionnaire, the Cognitive Reserve Questionnaire, the Cognitive Reserve Scale, or the Cognitive Reserve Index questionnaire.

[0037] Additionally, in the present disclosure, a “subject” is a subject of brain MRI imaging and a subject of a cognitive reserve test.

[0038] Additionally, in the present disclosure, a “user” is a person who wishes to analyze the cognitive reserve level of a subject, and may be a medical professional, a nurse, a caregiver, or the subject himself.

[0039] Additionally, in the present disclosure, “brain functional connectivity data” refers to a set of correlation degrees assigned to all pairs of nodes selected from a plurality of nodes (signals) identified in a brain MRI image, which are characteristics of brain signals that appear in each individual.

[0040]

[0041] 1. Description of the Cognitive Ability Analysis System

[0042] FIG. 1 is a block diagram of a cognitive reserve analysis system according to embodiments of the present disclosure. FIG. 2 is a block diagram of a cognitive reserve analysis device according to one embodiment of the present disclosure.

[0043] In FIGS. 1 and 2, a cognitive ability analysis device (10) according to one embodiment is described as a representative example, but it should be noted that a cognitive ability analysis device (10') according to another embodiment is also described.

[0044] The cognitive reserve analysis system (1) according to embodiments of the present disclosure includes all or part of a cognitive reserve analysis device (10), an MRI device (20), a personal information database (30), and a user terminal (40).

[0045] The cognitive reserve analysis device (10) is configured to receive data from an MRI device (20) and / or a personal information database (30) and train a learning model capable of outputting a cognitive reserve value.

[0046] For this purpose, the cognitive ability analysis device (10) may be an electronic device capable of data processing, such as a PC, a smart device, or a server.

[0047] The MRI device (20) is configured to capture a brain MRI image of a subject, and the captured brain MRI image can be transmitted to a cognitive reserve analysis device (10). Meanwhile, in the present disclosure, it is preferable that the brain MRI image is an image captured of the brain in a resting state.

[0048] The personal information database (30) is configured to store the personal information of the subject, and when there is a request from the cognitive ability analysis device (10), the stored personal information can be transmitted to the cognitive ability analysis device (10). For this purpose, the personal information database (30) may be an EMR (Electronic Medical Records).

[0049] Meanwhile, personal information may preferably include, but is not limited to, the subject's age, gender, clinical information, and academic background information.

[0050] The user terminal (40) is electrically connected to the cognitive reserve analysis device (10), and can transmit a request to the cognitive reserve analysis device (10) and receive a response from the cognitive reserve analysis device (10).

[0051] Specifically, when a user inputs brain functional connectivity data through a user terminal (40), the user terminal (40) transmits the input data to a cognitive reserve analysis device (10), and the cognitive reserve analysis device (10) can output a predicted cognitive reserve value based on the input data. The output predicted cognitive reserve value can be transmitted back to the user terminal (40). Therefore, the user end can know the cognitive reserve value of the subject by simply inputting the subject's brain functional connectivity data.

[0052] Alternatively, when a brain MRI image is input into the user terminal (40), the input brain MRI image may be preprocessed within the cognitive reserve analysis device (10) to generate brain functional connectivity data. The cognitive reserve analysis device (10) may predict a predicted cognitive reserve value based on the generated brain functional connectivity data. The output predicted cognitive reserve value may be transmitted back to the user terminal (40). Therefore, the user end may know the cognitive reserve value of the subject simply by inputting the subject's brain MRI image.

[0053]

[0054] Referring to FIG. 2, the detailed configuration and function of the cognitive reserve analysis device (10) will be described. Meanwhile, it should be noted that in the present disclosure, the minimum configuration for cognitive reserve analysis is described, and other configurations, such as an output unit, are omitted.

[0055] A cognitive ability analysis device (10) according to one embodiment of the present disclosure includes all or part of an input unit (11), a processor (12), a learning processor (13), a control unit (14), a memory (15), and a communication unit (16).

[0056] The input unit (11) is configured to receive commands or information from the outside.

[0057] Commands or information input through the input unit (11) can be transmitted to other components of the cognitive reserve analysis device (10). In the present disclosure, brain MRI images and cognitive reserve values ​​can be input through the input unit (11).

[0058] The processor (12) can perform a data processing or calculation function, and as part of the data processing or calculation function, is configured to store commands or data received from other components in volatile memory, process the commands or data stored in the volatile memory, and store result data in non-volatile memory.

[0059] The processor (12) can preprocess the brain MRI image input to the cognitive function analysis device (10) to generate brain functional connectivity data. This is described in detail in FIG. 3.

[0060] The learning processor (13) is configured to train a learning model that uses brain functional connection data as input data and cognitive reserve values ​​as output data.

[0061] The control unit (14) is configured to control one or more other components of the cognitive reserve analysis device (10). Here, the one or more other components are, for example, a concept that includes both hardware and software. That is, the control unit (14) can control the input unit (11), processor (12), learning processor (13), memory (15), and communication unit (16) of the cognitive reserve analysis device (10).

[0062] The memory (15) is configured to temporarily or permanently store information input into the cognitive ability analysis device (10) or information processed or generated by the cognitive ability analysis device (10).

[0063] For example, the learning information data set used by the cognitive ability analysis device (10) when learning the learning model can be stored in the memory (15).

[0064] The communication unit (16) is configured for communication with an external device, and the cognitive ability analysis device (10) can transmit and receive data with the external device through the communication unit (16). Here, the external device includes an MRI device (20), a personal information database (30), and a user terminal (40).

[0065]

[0066] 2. Description of a cognitive ability analysis device and analysis method according to one embodiment.

[0067] 2.1. Description of a method for generating learning information according to one embodiment

[0068] FIG. 3 is a diagram illustrating a method for generating learning information of a cognitive ability analysis device according to one embodiment of the present disclosure.

[0069] The input unit (11) of the cognitive reserve analysis device (10) according to one embodiment of the present disclosure is equipped with a brain MRI image input module (110) and a cognitive reserve numerical input module (111).

[0070] The brain MRI image input module (110) is configured to receive an MRI image of a subject from an external device, such as an MRI device (20).

[0071] The brain MRI image input into the brain MRI image input module (110) is preprocessed by the preprocessing module (120), thereby generating brain functional connectivity data. At this time, the preprocessing module (120) is preferably a module provided in the processor (12).

[0072] The preprocessing module (120) preprocesses brain MRI images according to the process described below.

[0073] Even in a resting brain MRI image, movement may occur, so the preprocessing module (120) spatially realigns the brain MRI image to compensate for head movement.

[0074] Spatially realigned brain MRI images are normalized to the same coordinate frame.

[0075] Normalized brain MRI images are spatially smoothed using a Gaussian kernel.

[0076] Spatially smoothed brain MRI images have low-frequency noise removed using filters or statistical models.

[0077] The above process can be performed using a publicly available MRI analysis software package (e.g., SPM12).

[0078] The denoised brain MRI image is segmented into 90 nodes. This segmentation can be performed based on the Automated Anatomical Labeling (AAL) atlas tool. The node formation method based on the AAL atlas is a previously disclosed signal processing method, and its related description is omitted herein.

[0079] The preprocessing module (120) estimates functional connectivity for pairs of two different nodes among the 90 segmented nodes. Here, functional connectivity refers to a correlation estimated through a partial correlation operation between two different signals (i.e., nodes).

[0080] In the spatial realignment process, estimated head movement variables and non-neuronal fluctuations extracted from white matter and cerebrospinal fluid can be considered as confounding temporal factors and removed from functional connectivity.

[0081] Meanwhile, the preprocessing module (120) can normalize the functional connectivity between 90 nodes using Fisher Transformation. Hereinafter, the normalized value of the functional connectivity between the 90 nodes, i.e., the connection strength between two different nodes, is referred to as the “partial correlation coefficient.” Accordingly, the brain functional connectivity data represents the connection strength values ​​for all node pairs that may occur in the 90 nodes, and is preferably expressed as a 90 x 90 symmetric matrix.

[0082] The brain functional connectivity data in the form of a 90 x 90 symmetric matrix generated through the above process is used as input data for learning. The brain functional connectivity data in the form of a 90 x 90 symmetric matrix can express 4,005 features.

[0083] During the learning model learning process, the correlation between the actual cognitive reserve value and the predicted cognitive reserve value for each feature is calculated, and if the p value ≤ 0.05 is satisfied, the feature can be selected as a significant important feature.

[0084]

[0085] The cognitive reserve value input module (111) is configured to receive the subject's cognitive reserve value from an external device, such as a cognitive reserve value database (not shown).

[0086]

[0087] The brain functional connectivity data and cognitive reserve value generated according to the above process can be stored in memory (15) as a learning information dataset.

[0088] The saved learning information dataset is transmitted to the learning processor (13), and the learning processor (13) trains the learning model using the received learning information dataset.

[0089] At this time, the learning model can utilize conventional machine learning techniques such as decision trees, random forests (RF), K-nearest neighbors (KNN), naive Bayes, support vector machines (SVM), artificial neural networks, and even recent deep learning techniques. This disclosure mainly describes the results of selecting a learning model according to the RF technique, but it should be noted that the effects of this disclosure are not limited to a specific learning technique.

[0090] Meanwhile, the parameters of the learning model learned by the learning processor (13) can be updated, i.e., their functions can be further improved through an optimization process. For example, the control unit (14) can update the parameters of the learning model by comparing the predicted cognitive reserve value output based on newly input brain functional connection data to the previously learned learning model with the actual cognitive reserve value.

[0091] At this time, a learning model can be built using the scikit-learn library, and hyperparameters can be adjusted through 5-fold cross-validation during the learning process of the learning model, but it is not necessarily limited to this.

[0092]

[0093] 2.2. Method for predicting cognitive reserve value using cognitive reserve analysis device according to one embodiment

[0094] FIG. 4 illustrates a method for predicting a cognitive reserve value using a cognitive reserve analysis device according to one embodiment of the present disclosure.

[0095] The user inputs brain MRI images of a subject whose cognitive reserve level is to be predicted, or brain functional connectivity data obtained through brain MRI images, through a user terminal (40).

[0096] The input brain MRI image or brain functional connectivity data is transmitted to the cognitive ability analysis device (10), and the brain functional connectivity data obtained through the brain MRI image or the transmitted brain functional connectivity data is input to the learning model.

[0097] The learning model outputs a predicted cognitive reserve value, and the output predicted cognitive reserve value can be transmitted to the user terminal (40).

[0098] Through this, the user can predict the subject's cognitive reserve level simply by inputting brain MRI image or brain functional connectivity data into the user terminal (40), which helps predict the occurrence and progression of dementia.

[0099]

[0100] 2.3. Description of a cognitive reserve analysis method according to one embodiment

[0101] FIG. 5 is a flowchart of a method using a cognitive ability analysis system according to one embodiment of the present disclosure.

[0102] Referring to FIG. 5, a brain MRI image input into a brain MRI image input module (110) is preprocessed by a preprocessing module (120), and brain functional connectivity data is generated (S500).

[0103] At this time, the detailed preprocessing method is replaced with the description in Fig. 3.

[0104]

[0105] Brain functional connectivity data and cognitive reserve value input into the cognitive reserve value input module (111) are created as a learning information dataset and stored in the memory (15) (S510).

[0106]

[0107] A learning model that uses brain functional connectivity data as input data and cognitive reserve scores as output data is learned by a learning processor (13) (S520).

[0108]

[0109] When brain MRI images or brain functional connectivity data are input into a user terminal, the user terminal (40) transmits the brain MRI images or functional connectivity data to a cognitive reserve analysis device (10). The cognitive reserve analysis device (10) inputs brain functional connectivity data generated from the input brain MRI images or brain functional connectivity data received from the user terminal (40) into a learning model and outputs a predicted cognitive reserve value (S530).

[0110]

[0111] 3. Description of a cognitive ability analysis device according to another embodiment

[0112] 3.1. Description of learning methods according to other embodiments

[0113] FIG. 6 is a block diagram of a cognitive ability analysis device according to another embodiment of the present disclosure.

[0114] Referring to FIG. 6, the input unit (11') of the cognitive ability analysis device (10') according to another embodiment of the present disclosure may further include a personal information input module (112') in addition to the cognitive ability analysis device (10) according to one embodiment.

[0115] The personal information input module (112') may be configured to receive personal information. Here, personal information preferably includes the subject's age, gender, clinical information, and academic background information, and may be transmitted from a personal information database (30, see FIG. 1).

[0116] The preprocessing module (120') can generate brain functional connectivity data based on information input to the brain MRI image input module (110').

[0117] Brain functional connectivity data, personal information and cognitive reserve scores can be generated as a learning information dataset and stored in memory (15').

[0118] The learning information dataset stored in the memory (15') can be input into the learning processor (13') and used as data for the learning model. At this time, the brain functional connection data and personal information are expressed as a 91 x 90 matrix, and the above matrix can be used as input data for the learning model, and the cognitive reserve value can be used as output data.

[0119]

[0120] FIG. 7 is a diagram illustrating a method for generating learning information of a cognitive ability analysis device according to another embodiment of the present disclosure.

[0121] According to the method illustrated in Fig. 6, brain functional connection data from a brain MRI image input through a user terminal (40) and generated by a cognitive ability analysis device (10') can be input into the learning model through a learning processor (13') including a learning model learned.

[0122] Alternatively, when brain functional connection data is input through a user terminal (40), the input brain functional connection data is transmitted to a cognitive ability analysis device (10'), and the transmitted brain functional connection data can be input into a learning model.

[0123] When the learning model receives brain functional connectivity data as input, it outputs a predicted cognitive reserve score.

[0124]

[0125] 3.2. Description of a learning information generation method according to another embodiment

[0126] FIG. 8 is a flowchart of a method using a cognitive reserve analysis system including a cognitive reserve analysis device according to another embodiment of the present disclosure.

[0127] Referring to FIG. 8, a cognitive ability analysis method according to another embodiment of the present disclosure is described.

[0128] Step S800 is replaced by the description of step S500 in Fig. 5.

[0129]

[0130] The brain functional connection data and personal information input into the personal information input module (112'), and the cognitive reserve value input into the cognitive reserve value input module (111') are stored in the memory as a learning information dataset (S810).

[0131]

[0132] A learning model that uses brain functional connectivity data and personal information as input data and cognitive reserve value as output data is learned by a learning processor (13') (S820).

[0133]

[0134] When personal information and brain MRI image or brain functional connection data are input into the user terminal (40), the user terminal (40) transmits the personal information and brain MRI image or functional connection data to the cognitive reserve analysis device (10'). The cognitive reserve analysis device (10') inputs the personal information and brain functional connection data generated from the input brain MRI image or brain functional connection data received from the user terminal (40) into a learning model and outputs a predicted cognitive reserve value (S830).

[0135]

[0136] 4. Verification results of the cognitive reserve analysis system

[0137] In this disclosure, the following verification method is adopted to verify the performance of the cognitive ability analysis device (10), but this is only an example, and it should be noted that the population selection and learning model selection may differ from the present disclosure.

[0138] In this study, brain MRI images and cognitive reserve data were collected from 72 participants aged 60 years or older. All participants were considered to have normal cognitive function in the Seoul Neuropsychological Screening Battery (SNSB) and underwent MRI and cognitive function tests. Normal cognitive function was defined as a score of -1 standard deviation or more from the population norm on memory, attention, language, visuospatial, and frontal lobe tests.

[0139] Information about the participants is shown in Table 1 below.

[0140]

[0141] General elderly (N = 72) Age (year) 72.0 ± 6.5 Education (year) 12.3 ± 2.6 Female, No. (%) 49 (68.1) Attention (Z-score) -0.1 ± 0.9 Visuospatial function 0.2 ± 0.9 Language -0.0 ± 0.9 Memory -0.1 ± 0.7 Frontal / executive function 0.7 ± 0.9

[0142] At this time, brain MRI images were acquired as T1-weighted images and resting-state images using a commercial scanner (3.0 Tesla Philips Achieva MR scanner). Structural images were acquired using a 3D T1-weighted MPRAGE (magnetization prepared rapid gradient echo) sequence. Resting-state images were acquired using an EPI (echo planar imaging) sequence. Brain MRI was acquired with the subject's eyes closed and not thinking of anything in particular. Thereafter, MRI acquired using the scanner was processed using a cognitive reserve analysis device (10) according to an embodiment of the present disclosure.

[0143] Of the entire data set for the population, 80% was used as training data, and 20% was used as validation data.

[0144]

[0145] FIG. 9 is a result of comparing the predicted cognitive reserve value and the actual cognitive reserve value using a cognitive reserve analysis device according to one embodiment of the present disclosure.

[0146] Referring to FIG. 9, it can be seen that the predicted cognitive reserve value predicted using the cognitive reserve analysis device (10) according to one embodiment of the present disclosure has a correlation coefficient of 0.952 with the actual cognitive reserve value, showing considerably high performance.

[0147] The above description is merely an example of the technical idea of ​​the present embodiment, and those skilled in the art will appreciate that various modifications and variations can be made without departing from the essential characteristics of the present embodiment. Therefore, the present embodiments are not intended to limit the technical idea of ​​the present embodiment, but rather to explain it, and the scope of the technical idea of ​​the present embodiment is not limited by these embodiments. The scope of protection of the present embodiment should be interpreted by the claims below, and all technical ideas within a scope equivalent thereto should be interpreted as being included in the scope of rights of the present embodiment.

[0148]

[0149] (Explanation of symbols)

[0150] 1: Cognitive Ability Analysis System

[0151] 10, 10': Cognitive Ability Analysis Device

[0152] 11, 11': Input section

[0153] 110, 110': Brain MRI image input module

[0154] 111, 111': Cognitive reserve numerical input module

[0155] 112': Personal information input module

[0156] 12: Processor

[0157] 120, 120': Preprocessing module

[0158] 13, 13': Running Processor

[0159] 14: Control unit

[0160] 15, 15': Memory

[0161] 16: Communications Department

[0162] 20: MRI device

[0163] 30: Personal information database

[0164] 40: User terminal

Claims

1. A method for analyzing cognitive reserve values ​​from brain MRI images using a system including a cognitive reserve analysis device and a user terminal, (a1) A step in which brain MRI images input into the cognitive reserve analysis device are preprocessed by the cognitive reserve analysis device to generate brain functional connectivity data; (a2) A step in which the brain functional connection data and the cognitive reserve value input into the cognitive reserve analysis device corresponding thereto are stored as a learning information dataset; (a3) A step in which a learning model using the brain functional connection data as input data and the cognitive reserve value as output data is learned by a cognitive reserve analysis device; (b1) a step of inputting brain MRI image or brain functional connectivity data into the user terminal; and (b2) a step in which the input brain MRI image or brain functional connection data is transmitted to the cognitive reserve analysis device, and the brain functional connection data generated by the cognitive reserve analysis device from the input brain MRI image or the input brain functional connection data is input into the learning model, thereby outputting a predicted cognitive reserve value; including; method.

2. In paragraph 1, The above brain functional connectivity data is a 90 x 90 matrix, method.

3. In paragraph 2, The above step (a1) is: (a10) The cognitive ability analysis device preprocesses the brain MRI image by spatially realigning the brain MRI image, normalizing it to the same coordinate frame, spatially smoothing it using a Gaussian kernel, and removing noise in a preset frequency band; and (a11) a step of segmenting the brain MRI image preprocessed by the above step (a10) into 90 nodes by the above cognitive ability analysis device; method.

4. In paragraph 3, The above step (a1) is: (a12) a step of calculating partial correlation coefficients for two different pairs among the 90 nodes by performing Fisher Transformation on the 90 nodes segmented by the step (a11); and (a13) further comprising a step of calculating a symmetric matrix of size 90 x 90 using the partial correlation coefficient calculated in step (a12) by the above cognitive ability analysis device; method.

5. In paragraph 1, In the above step (a2), personal information is further stored in the cognitive ability analysis device, In the above step (a3), the personal information is further used as input data, In the above step (b1), when additional personal information is input into the user terminal and the cognitive reserve analysis device, a predicted cognitive reserve value is output. method.

6. In a system for analyzing cognitive ability, An MRI device configured to take MRI images of a subject's brain; As a cognitive ability analysis device, An input unit for receiving the brain MRI image and the subject's cognitive reserve value; A processor that preprocesses the brain MRI image to generate brain functional connectivity data; and A cognitive reserve analysis device including a learning processor configured to train a learning model using the brain functional connections as input data and the cognitive reserve value as output data; and A user terminal that receives brain MRI images or brain functional connections and transmits them to the cognitive ability analysis device; Brain functional connectivity data generated from brain MRI images input to the user terminal or brain functional connectivity data input to the user terminal are transmitted and input to the cognitive ability analysis device, The above cognitive reserve analysis device outputs a predicted cognitive reserve value using the above learning model and transmits it to the user terminal. System.

7. In paragraph 6, The input unit of the above cognitive ability analysis device further receives the subject's personal information, The above personal information is used as input data for the above learning model, Personal information is further input into the user terminal, and the user terminal transmits the input personal information to the cognitive ability analysis device. System.

Images