Method for providing information for predicting risk of onset of dementia
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2026-02-10
- Publication Date
- 2026-08-13
Smart Images
Figure KR2026002486_13082026_PF_FP_ABST
Abstract
Description
Method of providing information for predicting the risk of developing dementia
[0001] The present invention relates to a method for providing information for predicting the risk of developing dementia, and more specifically, to a method for providing information that predicts the risk of developing dementia by measuring the abundance of oral microbial strains and applying an analysis model established based thereon.
[0002] With the rapid aging of modern society, the number of dementia patients, including those with Alzheimer's, is surging, resulting in serious social costs and a decline in the quality of life for patients and their families. In particular, as dementia is a disease that is difficult to cure once it develops, it is of paramount importance to detect and manage the risk of onset early, before symptoms appear.
[0003] To date, methods used for diagnosing dementia include imaging tests such as magnetic resonance imaging (MRI) and positron emission tomography (PET), or the measurement of amyloid beta or tau protein levels in cerebrospinal fluid; however, these methods have limitations in early prediction due to high costs, difficulty of access, and resistance to invasive tests.
[0004] With the recent surge in research on the oral-brain axis, the identification of associations between specific oral microbial strains and central nervous system inflammatory responses and amyloid accumulation has opened up the possibility of screening for dementia risk through oral microbial strain screening. For example, a mechanism has already been proposed in which periodontal disease-causing bacteria, such as Porphyromonas gingivalis, and their toxins accelerate neurodegeneration through the bloodstream.
[0005] Oral samples represent an optimal methodology for predicting dementia risk because they are very easy to collect, non-invasive, and allow for repeated monitoring. Therefore, there is a need to develop a method that can rapidly predict dementia risk at a low cost by establishing indicators based on oral microbial strains that are highly correlated with the risk of developing dementia.
[0006] While researching a method for predicting the prognosis of the risk of developing dementia, the inventors confirmed that the abundance of specific oral microorganisms increases rapidly in a group of dementia patients in the preclinical stage progressing to Alzheimer's dementia. Consequently, by distinguishing between low-risk and high-risk groups based on the abundance of said oral microorganisms, it is possible to provide information on the prognosis of dementia development, and the inventors confirmed that this demonstrates a high prognosis prediction effect even for patients not identified by genetic testing, thereby completing the present invention.
[0007] Accordingly, the objective of the present invention is to provide a method for providing information that predicts the risk of developing dementia by measuring the abundance of specific microbial strains from oral samples and applying an analysis model established based thereon.
[0008] However, the technical problems that the present invention aims to solve are not limited to those mentioned above, and other unmentioned problems will be clearly understood by those skilled in the art to which the present invention belongs from the description below.
[0009] To achieve the above objective, the present invention provides a method for providing information for predicting the risk of developing dementia, comprising the steps of: measuring the abundance of one or more strains selected from a group consisting of Parvimonas micra, Prevotella baroniae, Lachnoanaerobaculum saburreum, Campylobacter rectus, Prevotella nigrescens, Streptococcus pneumoniae, Filifactor alocis, and Abiotrophia defectiva in an oral sample collected from a subject; calculating a dementia risk score by applying an analysis model established based on the measured abundance; and providing dementia risk information based on the calculated dementia risk score.
[0010] In one embodiment of the present invention, the method may further include the step of measuring the abundance of one or more strains selected from the group consisting of Treponema denticola and Porphyromonas gingivalis.
[0011] In another aspect of the present invention, the analysis model may reflect the results of a correlation analysis with clinical dementia diagnostic stages consisting of a normal control group (HC), an amyloid-positive normal group (CN), mild cognitive impairment (MCI), and Alzheimer's dementia (AD).
[0012] In another aspect of the present invention, the analysis model may reflect the results of a correlation analysis with the state of amyloid beta accumulation in the brain (Amyloid Positive or Amyloid Negative).
[0013] In one aspect of the present invention, the analysis model may reflect the results of a correlation analysis between the brain atrophy indicator confirmed through time-series tracking of brain MRI and the abundance of the strain.
[0014] In another aspect of the present invention, the brain atrophy indicator may be the rate of decrease in volume of the hippocampus or the rate of decrease in thickness of the non-endocortical region.
[0015] In another aspect of the present invention, the dementia risk score is calculated by the following mathematical formula 1, and based on the score, the subject can be classified into a high-density group and a low-density group of target marker strains.
[0016] In another aspect of the present invention, the method may further include the step of estimating the level of GFAP, a brain inflammation marker, using the abundance information of the strain.
[0017] In another aspect of the present invention, the oral sample may be a saliva or plaque sample separated from a subject.
[0018] The method for providing information on the risk of developing dementia according to the present invention can predict the risk of developing dementia solely through non-invasive oral sample collection, thereby minimizing the burden on the patient and facilitating repeated monitoring. Since the present invention enables rapid and low-cost prediction of dementia risk, it is suitable for screening large population groups and can be utilized even in areas with limited access to medical care.
[0019] By utilizing the abundance of specific oral microbial strains as a biomarker, this invention is applicable to novel patient groups not identified by genetic testing for dementia, enabling the early identification of high-risk groups for potential dementia. In particular, it allows for the screening of dementia patients in the preclinical stage who are rapidly progressing to Alzheimer's disease, thereby providing an opportunity to detect and manage the risk of onset at an early stage prior to the onset of symptoms.
[0020] The analysis model of the present invention incorporates the results of a correlation analysis between brain atrophy indicators identified through brain MRI follow-up and oral microbial abundance, thereby providing prediction results that are highly correlated with the actual progression of brain atrophy. Furthermore, it can estimate not only dementia risk scores but also GFAP levels, a brain inflammation marker, thereby providing multidimensional information on the mechanism of dementia onset and enabling a more accurate risk assessment.
[0021] Figure 1. An association analysis between the abundance of oral microbial strains and the degree of brain atrophy in the hippocampus and entorhinal cortex was confirmed based on Microbiome Multivariable Association with Linear Models 2 (MaAsLin 2) statistical analysis, and the effect size (coefficients) were determined (15% prevalence, coefficient greater than 0.15, q-value < 0.05).
[0022] Figure 2. Results of calculating the rate of decrease in left and right hippocampus brain tissue volume, i.e., the brain atrophy rate, according to the stage of dementia progression (HC, healthy control; CN, cognitive normal (amyloid-positive), MCI, mild cognitive impairment; AD, Alzheimer's disease).
[0023] Fig. 3. Results of calculating the rate of decrease in brain tissue thickness, i.e., the brain atrophy rate, according to the stage of dementia progression (HC, healthy control; CN, cognitive normal (amyloid-positive), MCI, mild cognitive impairment; AD, Alzheimer's disease), specifically the left and right entorhinal cortex.
[0024] Fig. 4. Results of a statistical prediction model predicting GFAP biomarker levels indicating brain inflammation based on oral microbial strains, apoE genotypes, and clinical metadata (R 2 = 0.38)
[0025] Fig. 5. Results confirming the association between a specific oral microbial strain, Treponema denticola strain, and the rate of decrease in left hippocampus brain tissue volume, i.e., the rate of brain atrophy.
[0026] The present invention will be described in detail below.
[0027]
[0028] The present invention relates to a method for providing information for predicting the risk of developing dementia.
[0029] The method of the present invention comprises the steps of: measuring the abundance of microbial strains in an oral sample collected from a subject; calculating a dementia risk score by applying an analysis model established based on the measured abundance; and providing dementia onset risk information based on the calculated dementia risk score.
[0030] The inventors confirmed that it is possible to predict the prognosis of a high-risk group for dementia by measuring the abundance of specific oral microbial marker strains from saliva samples isolated from patients suspected of being at risk of dementia, and by analyzing this in conjunction with clinical dementia diagnostic stage information, brain amyloid beta accumulation status information, and brain atrophy rates confirmed through brain MRI follow-up, thereby classifying the target marker strains into high-density and low-density groups based on an arbitrary cutoff for the abundance of the microbial strains. In particular, the present invention has the advantage of being able to identify a group of dementia patients in the preclinical stage who are rapidly progressing to Alzheimer's dementia at an early stage, and is also applicable to new patient groups not identified by genetic testing for dementia onset.
[0031] Oral samples may be saliva or plaque collected from the subject's oral cavity. Saliva samples are most commonly used as they represent the optimal methodology for predicting dementia risk because they are non-invasive, very easy to collect, and allow for repeated monitoring. Saliva samples may be collected, for example, on an empty stomach in the morning, provided that food intake or brushing is prohibited for at least 30 minutes prior to collection. Plaque samples may be collected from supragingival or subgingival plaque and may be collected using a sterile curette or scaler.
[0032] The abundance of microbial strains can be measured by extracting DNA from oral samples and performing shotgun metagenomic sequencing or 16S rRNA gene sequencing. For example, commercial DNA extraction kits can be used for DNA extraction, and high-purity microbial DNA can be obtained through cell lysis, protein removal, and DNA purification processes. In the case of shotgun metagenomic sequencing, precise identification down to the strain level is possible because the entire microbial genome is randomly fragmented and sequenced.
[0033] Quality control is performed on raw sequencing data through a preprocessing stage. After removing low-quality reads and human-derived DNA information, taxonomic profiling is conducted based on the DNA sequencing reads of the remaining microbial strains. Reads matching human genome sequences are excluded, and the remaining reads are mapped to a microbial reference database to identify each strain and calculate its relative abundance. Relative abundance can be expressed as the proportion of a specific strain within the total microorganisms, for example, as a percentage or a read count ratio.
[0034] The microbial strains may be one or more selected from the group consisting of Parvimonas micra, Prevotella baroniae, Lachnoanaerobaculum saburreum, Campylobacter rectus, Prevotella nigrescens, Streptococcus pneumoniae, Filifactor alocis, and Abiotrophia defectiva. These strains are key target marker strains that have been identified as having significantly increased abundance in the high-risk group for dementia.
[0035] As one embodiment of the present invention, the abundance of one or more strains selected from the group consisting of Treponema denticola and Porphyromonas gingivalis can be further measured. These strains are known to have a particularly high association with dementia.
[0036] Fabimonas microcra is an anaerobic Gram-positive coccus in the oral cavity that is associated with periodontal disease and can cause chronic inflammation. Prevotella varonia and Prevotella nigrecens are anaerobic Gram-negative bacilli of the genus Prevotella; while they are commensal bacteria in the oral cavity, their overgrowth can promote inflammatory responses. Lachnoaerobic saburéum is an anaerobic Gram-positive bacilli involved in oral biofilm formation and may be associated with gingivitis. Campylobacter rectus is a spiral-shaped Gram-negative bacilli associated with periodontitis and can induce systemic inflammatory responses. Streptococcus pneumoniae is a Gram-positive coccus generally known as a respiratory pathogen, but it can also be detected in the oral cavity. Filipactor alosysis is an anaerobic Gram-positive bacilli associated with the severity of periodontitis and is primarily found in deep periodontal pockets in the oral cavity. Aviotropia defectiva is a trophozoite streptococcus that is a commensal bacterium in the oral cavity but can exhibit pathogenicity under certain conditions. Treponema denticola is a spiral anaerobic bacterium that is one of the major pathogens of periodontitis, and its potential to migrate to the central nervous system has been reported. Porphyromonas gingivalis is an anaerobic Gram-negative bacillus that is a representative pathogen of periodontal disease, and a mechanism has already been proposed in which its toxins travel to the brain via the bloodstream and accelerate neurodegeneration.
[0037] The method of the present invention includes the step of calculating a dementia risk score by applying an analysis model established based on measured abundance. The analysis model refers to a statistical or mathematical model that receives abundance data of oral microbial strains as input and calculates the risk of developing dementia as a quantified score. Such an analysis model can be developed and verified based on data from a pre-established reference group.
[0038] The analysis model may reflect the results of a correlation analysis between the abundance of oral microbial strains and the clinical dementia diagnostic stages, consisting of healthy controls (HC), cognitively normal with amyloid-positive groups (CN), mild cognitive impairment (MCI), and Alzheimer's disease (AD). In other words, statistical patterns derived from analyzing the relationship between microbial strain abundance and clinical dementia diagnostic stages in the reference population can be incorporated into the analysis model.
[0039] The analysis model may reflect the results of a correlation analysis between the state of amyloid beta accumulation in the brain (Amyloid Positive or Amyloid Negative) and the abundance of oral microbial strains. Amyloid beta is one of the major pathological characteristics of Alzheimer's disease, and the presence of amyloid accumulation in the brain serves as an important indicator for assessing dementia risk. Statistical patterns derived from analyzing the relationship between microbial strain abundance and amyloid accumulation status in a reference population can be incorporated into the analysis model.
[0040] The analysis model may reflect the results of a correlation analysis between brain atrophy indicators identified through time-series tracking of brain MRIs and the abundance of oral microbial strains. In other words, statistical patterns derived from analyzing the relationship between microbial strain abundance and the degree of brain atrophy in a reference population can be incorporated into the analysis model. This enables the prediction of the likelihood of brain atrophy progression and the risk of dementia based solely on the abundance data of individual subjects.
[0041] Indicators of brain atrophy can be identified through time-series follow-up of brain MRIs. The degree of brain atrophy can be determined by taking two or more brain MRIs at regular intervals for a reference group, measuring the volume or thickness of specific brain regions at each time point, and calculating the rate of decline over time. The interval for brain MRI scans can be, for example, 6 to 24 months, and a 12-month interval is generally used.
[0042] Indicators of brain atrophy may be the rate of decrease in hippocampal volume or the rate of decrease in non-endocortical thickness. The hippocampus can be measured separately for the left and right hippocampus, and the non-endocortical (entorhinal cortex) can also be measured separately for the left and right entorhinal cortex. These areas are brain regions associated with dementia and are regions where significant atrophy is observed in the early stages of Alzheimer's disease.
[0043] The volume of the hippocampus is mm 3 It can be measured in units, and non-endocortical thickness can be measured in millimeters. The rate of volume reduction can be calculated, for example, as (initial volume - volume at follow-up) / time interval, and the rate of thickness reduction can be calculated as (initial thickness - thickness at follow-up) / time interval. Brain atrophy values are negative, and a larger absolute negative value indicates a more severe degree of atrophy. In other words, a higher rate of reduction indicates that brain atrophy is progressing rapidly, suggesting a high risk of developing dementia.
[0044] Brain MRI image analysis can be performed using automated image analysis software. For example, software such as FreeSurfer, FSL, and SPM can be used to subdivide brain structures and quantify the volume and thickness of each region. These software programs utilize standardized algorithms to minimize inter-individual variability and provide objective measurements.
[0045] Correlation analysis for the development of analysis models can be performed using multivariate linear model analysis. For example, the Microbiome Multivariable Association with Linear Models 2 (MaAsLin2) statistical analysis method can be used for multivariate linear model analysis. MaAsLin2 is a statistical method optimized for evaluating associations between microbial abundance data and continuous or categorical metadata, and can rigorously determine statistical significance through multiple test correction.
[0046] Association analysis between the abundance of oral microbial strains and the degree of brain atrophy in the hippocampus and non-endocortical regions can be performed using MaAsLin2 statistical analysis, and the corresponding effect sizes (coefficients) can be derived. Statistical significance can be determined based on criteria such as a 15% prevalence, a coefficient of 0.15 or greater, and a q-value < 0.05. Prevalence indicates the percentage of the total sample in which the strain is detected, and cases exceeding 15% may be included in the analysis. The coefficient represents the effect size, and an absolute value of 0.15 or greater is considered clinically significant. The q-value is the p-value after multiple testing correction, and a value less than 0.05 is considered statistically significant.
[0047] Effect size is a numerical value indicating how much the brain atrophy index changes when the abundance of each microbial strain increases by one unit. For example, a positive effect size means that brain atrophy accelerates as the abundance of the corresponding strain increases, while a negative value may indicate a protective effect. Analysis results confirmed that brain atrophy accelerated as the abundance of the target microbial strain increased.
[0048] The analysis model may include selecting strains identified as statistically significant in multivariate linear model analysis as target markers. By selecting strains that show a significant association in at least one brain region associated with dementia (q-value < 0.05) as target marker strains, strains most relevant to predicting dementia risk can be used as key elements of the analysis model.
[0049] As one embodiment of the analysis model, a dementia risk score is calculated based on the number or ratio of strains exceeding the cutoff by comparing the cutoff set for each target marker strain with the measured abundance, and accordingly, the subjects can be classified into a target marker strain high-density group and a low-density group.
[0050] A cutoff value is set for each strain based on the abundance of target marker strains, and target strains with high densities above the cutoff and those with low densities can be distinguished in individual subjects. The cutoff value can be determined as the point that optimizes sensitivity and specificity through ROC curve analysis, after identifying the distribution of strain abundance in a reference patient population.
[0051] The score for distinguishing between the high-density group and the low-density group of target marker strains can be calculated by the following mathematical formula:
[0052] Dementia Risk Score (%) = (Number of target strains with a density above the cutoff in the patient / Number of target strains detected in the patient) × 100
[0053] Here, target strains refer to marker strains selected as statistically significant. A cutoff value is set for each strain, and the proportion of strains exceeding the cutoff in the corresponding subject is calculated and used as the dementia risk score. For example, if 8 out of 10 target strains are detected in a subject and 5 of them exceed the cutoff, the dementia risk score is (5 / 8) × 100 = 62.5%.
[0054] The higher the calculated dementia risk score, the more likely the subject is to belong to the high-density group of target marker strains, indicating a high risk of developing dementia. Subjects classified into the high-density group of target marker strains are determined to be in the high-risk group, while subjects classified into the low-density group of target marker strains can be determined to be in the low-risk group.
[0055] The reference group may consist, for example, of a healthy control group (HC), a cognitively normal group with amyloid-positive individuals (CN), patients with mild cognitive impairment (MCI), and patients with Alzheimer's disease (AD). By comparing the rate of reduction in brain atrophy within each group, differences in the degree of brain atrophy according to the stage of dementia progression can be identified.
[0056] The method of the present invention may further include the step of estimating the level of GFAP (Glial Fibrillary Acidic Protein), a brain inflammation marker, through a statistical prediction model constructed using strain abundance information. GFAP is an indicator of astrocyte activation and is a biomarker that reflects the state of brain inflammation. The higher the blood GFAP level, the more active the inflammatory response in the brain, which is associated with neurodegeneration.
[0057] Statistical prediction models for GFAP estimation can be constructed using clinical metadata such as apoE genotype, age, and sex as input variables, along with information on the abundance of oral microbial strains. These prediction models can be optimized through statistical methodologies, and various statistical techniques, such as multiple linear regression and machine learning algorithms, may be used. Prediction accuracy is expressed as the coefficient of determination (R²). 2 It can be evaluated as ), for example, R 2 GFAP values can be predicted with an accuracy of = 0.38. R 2 If the value is 0.3 or higher, it can be considered to have useful predictive power.
[0058] Estimating GFAP levels allows for the indirect assessment of a subject's brain inflammation status without actually performing blood tests, which can contribute to improving the accuracy of dementia risk assessment. Since brain inflammation is closely related to the onset and progression of dementia, predicting GFAP levels can provide additional information for screening dementia risk groups.
[0059] The method of the present invention may be characterized by predicting the rate of brain atrophy in patients. That is, by applying data on the abundance of oral microbial strains to an analysis model, it is possible to predict how quickly future brain atrophy will progress, which enables the estimation of the rate of dementia progression without repeatedly performing MRI examinations on individual subjects.
[0060] The step of providing information on the risk of developing dementia involves delivering the calculated dementia risk score to the subject or medical staff. Risk information may be provided in categories such as high-risk, medium-risk, and low-risk groups, and percentile ranking information may be provided along with specific dementia risk scores. Information delivery can be carried out in various ways, such as through written reports, transmission via electronic medical record systems, or notifications via mobile applications.
[0061] The information provided may include the relative abundance of measured major microbial strains, whether each strain exceeds a cutoff, the calculated dementia risk score, risk group classification, estimated GFAP values, and recommendations. Recommendations may include, for example, the need for further testing, lifestyle modifications, the importance of periodontal health care, and a schedule for regular follow-up.
[0062] The method of the present invention provides an opportunity for timely intervention by predicting the onset of dementia early. Subjects classified as high-risk can receive more intensive monitoring and preventive interventions, which can help delay the progression of dementia or slow the onset of symptoms. The non-invasive testing method using oral samples facilitates repeated testing, making it suitable for long-term follow-up; it also has the advantages of being less expensive and more accessible compared to conventional imaging or cerebrospinal fluid examinations. In particular, the present invention can rapidly predict dementia risk at a low cost and is highly effective in identifying potential high-risk groups for dementia early and predicting the prognosis of dementia onset.
[0063]
[0064] The present invention will be explained in more detail with reference to the following examples.
[0065]
[0066] Example 1. Profiling of microbial strain abundance in subjects at low and high risk of dementia
[0067] To determine the risk of developing dementia, DNA was extracted from saliva samples of 450 subjects and shotgun metagenomic sequencing was performed. The raw sequencing data was preprocessed to remove human-derived DNA information, and taxonomic profiling was performed based on the DNA sequencing reads of the remaining microbial strains. The abundance of oral microbial strains was calculated for 145 normal individuals, 154 amyloid-positive, cognitively unimpaired subjects, 112 subjects with mild cognitive impairment (MCI), and 39 Alzheimer's patients.
[0068]
[0069] Example 2. Classification of high-risk and low-risk dementia groups based on brain atrophy following brain MRI observation
[0070] To compare high-risk and low-risk groups for dementia, changes in the brain MRI of the subjects are measured at least twice as shown in Figure 1, and the degree of brain atrophy is calculated by determining the rate of volume decrease in the left and right hippocampus regions or the rate of thickness change in the left and right entorhinal cortex. Subsequently, high-risk and low-risk groups for dementia are classified according to the degree of brain atrophy.
[0071]
[0072] Example 3. Confirmation of changes in the abundance of target microbial strains in a high-risk dementia group based on brain atrophy.
[0073] As shown in Figures 2 and 3, brain atrophy was identified based on the rate of decrease in brain volume or head size (i.e., brain atrophy rate) in the hippocampus and entorhinal cortex of the subjects, and as shown in Figure 5, the association with the abundance of Treponema denticola strains in the oral samples of the subjects was analyzed. It was confirmed that the abundance of Treponema denticola strains increased as the brain atrophy value (the more negative the value, the more severe the atrophy) increased.
[0074]
[0075] Example 4. Confirmation of statistical significance between the abundance of target oral microbial strains and the brain atrophy score based on the rate of volume or thickness reduction of dementia-associated regions in brain MRI.
[0076] In Figure 1, the association between the abundance of oral microbial strains and the degree of brain atrophy in the hippocampus and entorhinal cortex was confirmed based on Microbiome Multivariable Association with Linear Models2 (MaAsLin2) statistical analysis, and the effect sizes (coefficients) were determined (15% prevalence, coefficient greater than 0.15, q-value less than 0.05).
[0077] As a result, at least one brain region associated with dementia showed a significant association with the abundance of one of the target microbial strains (q-value < 0.05). When the high-density and low-density groups were divided based on the abundance of the target microbial strain, the high-risk group for dementia development could be predicted. Furthermore, based on the optimized scores obtained through the SUVR statistical methodology, the levels of GFAP, a brain inflammation marker, could be predicted with high accuracy as shown in Figure 4 (R 2 =0.38)
Claims
1. A method for providing information for predicting the risk of developing dementia, comprising: a step of measuring the abundance of one or more strains selected from a group consisting of Parvimonas micra, Prevotella baroniae, Lachnoanaerobaculum saburreum, Campylobacter rectus, Prevotella nigrescens, Streptococcus pneumoniae, Filifactor alocis, and Abiotrophia defectiva in an oral sample taken from a subject; a step of calculating a dementia risk score by applying an analysis model established based on the measured abundance; and a step of providing dementia risk information based on the calculated dementia risk score.
2. A method for providing information for predicting the risk of developing dementia, comprising the step of further measuring the abundance of one or more strains selected from the group consisting of Treponema denticola and Porphyromonas gingivalis, in accordance with Claim 1.
3. A method for providing information for predicting the risk of developing dementia, wherein the analysis model of claim 1 reflects the results of a correlation analysis with clinical dementia diagnostic stages consisting of a normal control group (HC), an amyloid-positive normal group (CN), mild cognitive impairment (MCI), and Alzheimer's dementia (AD).
4. A method for providing information for predicting the risk of developing dementia, wherein the analysis model of claim 1 reflects the results of a correlation analysis with the state of amyloid beta accumulation in the brain (Amyloid Positive or Amyloid Negative).
5. A method for providing information for predicting the risk of developing dementia, wherein the analysis model of claim 1 reflects the results of a correlation analysis between a brain atrophy indicator confirmed through time-series tracking of brain MRI (Magnetic Resonance Imaging) and the abundance of the strain.
6. A method for providing information for predicting the risk of developing dementia, characterized in that, in claim 5, the brain atrophy indicator is the rate of decrease in volume of the hippocampus or the rate of decrease in thickness of the entorhinal cortex.
7. A method for providing information for predicting the risk of developing dementia according to claim 1, wherein the dementia risk score is calculated by the following mathematical formula 1, and the subject is classified into a high-density group and a low-density group based on the score: [Mathematical Formula 1] Dementia risk score (%) = (Number of target strains with a high density above the cutoff in the subject / Number of target strains detected in the subject) × 100.
8. A method for providing information for predicting the risk of developing dementia, wherein, in Claim 1, the method further comprises the step of estimating the level of GFAP (Glial Fibrillary Acidic Protein), a brain inflammation marker, using the abundance information of the strain.
9. A method for providing information for predicting the risk of developing dementia, characterized in that, in Claim 1, the oral sample is a saliva or dental plaque sample separated from a subject.