Information processing apparatus, information processing method, program, and storage medium

Abstract

In this information processing device, patient's information corresponding to each of at least four variables corresponding to one of first information about orientation, second information about memory, third information about memory for schedules for holidays, family gatherings, appointments, or medications in IADLs, or memory for actual tasks of eating or moving in IADLs, and fourth information about the positivity or negativity of a biomarker, is inputted to a predictive model (132) that has been trained using training data in which the at least four variables are associated with the presence or absence of symptomatic progression of mild cognitive impairment (MCI) or mild Alzheimer-type dementia (mild AD), to determine the probability of the patient's symptomatic progression of mild cognitive impairment (MCI) or mild Alzheimer-type dementia (mild AD).

Description

Technical Field

[0001] The present invention relates to an information processing apparatus, an information processing method, a program, and a storage medium.

Background Art

[0002] Conventionally, various techniques for predicting the degree of dementia symptoms have been proposed. For example, Patent Document 1 discloses that by analyzing the free conversation of a patient, the severity of dementia can be predicted without performing the Mini Mental State Examination (MMSE) used for dementia screening tests and severity evaluations.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] By the way, in recent years, in order to enhance the treatment effect on dementia patients, in addition to predicting dementia, it has been required to predict diseases for patients with mild cognitive impairment (MCI) or mild Alzheimer's type dementia (mild AD) who are at risk of dementia. However, patients with mild cognitive impairment (MCI) or mild Alzheimer's type dementia (mild AD) sometimes have milder symptoms compared to dementia patients, and it has been difficult to accurately predict the disease. Also, from the perspective of reducing the processing load on patients and devices for disease prediction, it is important to perform prediction using simple variables.

[0005] The present invention has been made in view of such circumstances, and an object thereof is to accurately predict the probability of progression of symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) and to reduce the processing load on patients and devices, and to provide an information processing apparatus, an information processing method, a program, and a storage medium.

Means for Solving the Problems

[0006] The present disclosure uses learning data in which at least four variables corresponding to any one of first information related to orientation, second information related to memory, third information related to the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the memory of actual tasks of eating or moving in IADL, and fourth information related to the positivity or negativity of biomarkers are associated with the presence or absence of symptom progression of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD). For a prediction model learned using the data, information of a patient corresponding to each of at least four variables is input to predict the probability that the patient will develop symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD). The at least four variables include a first variable that is information related to orientation regarding time among the first information, a second variable that is information related to delayed recall among the second information, a third variable that is information related to the memory of holidays, family gatherings, reservations, or medication schedules in IADL among the third information, and a fourth variable that is information related to the positivity or negativity of amyloid-β among the fourth information. An information processing apparatus is provided.

[0007] The at least four variables may be four variables consisting of the first variable, the second variable, the third variable, and the fourth variable.

[0008] The at least four variables may be five variables consisting of the first variable, the second variable, the third variable, the fourth variable, and a fifth variable that is any one of the first information, the second information, and the third information.

[0009] The fifth variable may be information related to orientation regarding location among the first information.

[0010] The fifth variable may be information regarding the memory of the name or type of the article among the second information.

[0011] The fifth variable may be information regarding the memory of the actual operation of meals in IADL among the third information.

[0012] The fifth variable may be information regarding the memory of the actual movement operation in IADL among the third information.

[0013] The present disclosure includes a step of acquiring patient information, first information regarding estimation knowledge, second information regarding memory, third information regarding the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the memory of actual operations of meals or movement in IADL, fourth information regarding the positivity or negativity of biomarkers, and at least four variables corresponding to any of them, and a prediction model learned using learning data in which the presence or absence of symptom progression of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) is associated therewith. For the prediction model, information corresponding to each of at least four variables included in the patient information is input to predict the probability that the patient will develop symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD), and a step of outputting a prediction result regarding the probability that the patient will develop symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD). At least four variables include a first variable that is information regarding estimation knowledge regarding time among the first information, a second variable that is information regarding delayed reproduction among the second information, a third variable that is information regarding the memory of holidays, family gatherings, reservations, or medication schedules in IADL among the third information, and a fourth variable that is information regarding the positivity or negativity of amyloid-β among the fourth information. An information processing method is provided.

[0014] The present disclosure causes a computer to perform a process of acquiring patient information, and first information related to estimation, second information related to memory, third information related to memory of holidays, family gatherings, reservations, or medication schedules in IADL, or memory of actual tasks of meals or movement in IADL, and fourth information related to the positivity or negativity of biomarkers, and at least four variables corresponding to any of these, and a presence or absence of progression of symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD). For a prediction model learned using learning data in which these are associated, information corresponding to each of the at least four variables included in the patient information is input, and a process of predicting the probability that a patient will develop symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) is performed, and a process of outputting a prediction result regarding the probability that the patient will develop symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) is performed. The at least four variables include a first variable that is information related to estimation of time among the first information, a second variable that is information related to delayed reproduction among the second information, a third variable that is information related to memory of holidays, family gatherings, reservations, or medication schedules in IADL among the third information, and a fourth variable that is information related to the positivity or negativity of amyloid-β among the fourth information. A program is provided.

[0015] The present disclosure provides a computer-readable non-transitory storage medium storing the program having the above configuration.

Advantages of the Invention

[0016] According to the present invention, it is possible to accurately predict the probability of progression of symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) and reduce the processing load on the patient and the device.

Brief Description of the Drawings

[0017]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

Figure 22

Figure 23

Figure 24

Figure 25

Figure 26

Figure 27

Figure 28

Figure 29

Figure 30

Figure 31

Figure 32

Figure 33

Figure 34

Figure 35

Figure 36

Figure 37

Figure 38

Figure 39

Figure 40

Figure 41

Figure 42

Figure 43

Figure 44

Figure 45

Figure 46

Figure 47

Figure 48

Figure 49

Mode for Carrying Out the Invention

[0018] Hereinafter, an embodiment of the information processing apparatus will be described with reference to the drawings. Note that the present disclosure is not limited to the embodiments described below.

[0019] As shown in FIG. 1, the information processing apparatus 100 is connected to the terminal device 10 via a communication network NW, for example. The communication network NW is constituted by, for example, a communication device, a wireless communication network, or the like.

[0020] The terminal device 10 is a device for a doctor to input patient information, and is constituted by, for example, a computer terminal device or a portable information terminal device.

[0021] The information processing apparatus 100 includes, for example, a communication unit 110, a control unit 120, and a storage unit 130. The control unit 120 is realized, for example, by a hardware processor such as a CPU (Central Processing Unit) executing a program (software). Further, some or all of these components may be realized by hardware (including a circuit unit: circuitry) such as an LSI (Large Scale Integration), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), a GPU (Graphics Processing Unit), or may be realized by cooperation between software and hardware. The program may be stored in advance in the storage unit 130 of the information processing apparatus 100, or may be stored in a removable computer-readable recording medium such as a DVD or a CD-ROM, and installed in the storage unit 130 of the information processing apparatus 100 when the computer-readable recording medium is mounted on a drive device.

[0022] The communication unit 110 includes a communication interface such as a NIC (Network Interface Card). The communication unit 110 communicates with the terminal device 10 using, for example, a cellular network, a Wi-Fi (registered trademark) network, or the like.

[0023] The control unit 120 includes, for example, an acquisition unit 121, a prediction unit 122, and an output unit 123.

[0024] The acquisition unit 121 acquires patient information transmitted from the terminal device 10 through the communication network NW. Here, the patient information is information used when predicting the probability that a patient's symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) will progress. Details of the patient information will be described later.

[0025] The prediction unit 122 inputs the patient information acquired by the acquisition unit 121 into the prediction model 132, and predicts the probability that a patient's symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) will progress.

[0026] The output unit 123 outputs information regarding the probability that a patient's symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) will progress, predicted by the prediction unit 122, to the terminal device 10 through the communication device 110. The output unit 123 may output, for example, in response to a request from the patient, information regarding the probability that a patient's symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) will progress, predicted by the prediction unit 122, to a terminal device possessed by a medical staff member such as a doctor or the patient.

[0027] Figure 2 is a diagram showing an example of the learning phase of the prediction model 132.

[0028] In the example shown in Figure 2, for each of a plurality of subjects, among the cohort data (ADNI) described later, the attribute information of the subject is used as the first data group, and the diagnostic information regarding mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) of the subject is used as the second data group, and the prediction model 132 is learned using the data extracted in association. In the present embodiment, as an example of the attribute information of the subject, as will be described later, information regarding the sense of time evaluated by MMSE (first information), information regarding delayed reproduction evaluated by MMSE (second information), information regarding memory evaluated by FAQ (third information), and information regarding Aβ positive / negative (fourth information) are preferably included.

[0029] FIG. 3 is a diagram showing an example of the evaluation phase of the prediction model 132.

[0030] In the example shown in FIG. 3, the evaluation of the prediction model 132 learned as described above is performed using the evaluation data 131. In the present embodiment, the evaluation data 131 uses three databases of ADNI (Alzheimer’s Disease Neuroimaging Initiative), J-ADNI (Japanese Alzheimer’s Disease Neuroimaging Initiative), and MissionAD (data of the Elesclomol Phase III clinical trial), and the evaluation of the prediction model 132 is performed individually for each database. ADNI is a database based on cohort data collected from multiple clinical periods to advance the clinical trials of dementia disease modifying drugs in the United States. J-ADNI is a database based on cohort data collected from multiple clinical periods to advance the clinical trials of dementia disease modifying drugs in Japan. MissionAD is a database based on the trial data independently collected by the applicant to advance the clinical trials of dementia disease modifying drugs. Since the sets of subjects are different for each database, the evaluation of the prediction model 132 using each of these databases is meaningful in determining the range of patients to which the prediction model 132 is applicable.

[0031] FIG. 4 is a diagram for explaining an example of the evaluation index of the prediction model 132.

[0032] In the example shown in FIG. 4, examples of evaluation metrics of the prediction model 132 include accuracy, sensitivity, specificity, and AUC (Area Under the ROC Curve). Accuracy is an index indicating the correct answer rate of the entire prediction regardless of positive or negative. Sensitivity is an index indicating the ratio (true positive rate) of correctly predicting positive data as positive. Specificity is an index indicating the ratio (true negative rate) of correctly predicting negative data as negative. AUC is an index calculated taking into account both sensitivity and specificity. Generally, in evaluating the prediction model 132 as a whole, evaluation based on AUC is preferred. In one embodiment, an aspect of adding sensitivity is included, and in another embodiment, aspects of adding sensitivity and specificity are included. In yet another embodiment, aspects of adding accuracy, sensitivity, and specificity are included. For each index, 0.7, 0.67, or 0.6 or more can be mentioned as the criteria for judging the conformity / inconformity of the prediction result.

[0033] FIG. 5 is a diagram showing an example of the execution phase of the prediction model 132.

[0034] In the example shown in FIG. 5, for the prediction model 132 learned as described above, patient information acquired from the terminal device 10 is input, and the probability that the patient's symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) will progress is predicted. In the present embodiment, as will be described later, as an example of the attribute information of the patient information, it is preferable to include information regarding the estimation of time evaluated by MMSE (first information), information regarding delayed reproduction evaluated by MMSE (second information), information regarding memory evaluated by FAQ (third information), and information regarding Aβ positive / negative (fourth information).

[0035] FIG. 6 is a flowchart showing an example of a process of predicting the probability that a patient's symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) will progress.

[0036] As shown in FIG. 6, the information processing apparatus 100 first acquires patient data from the terminal device 10 (step S10).

[0037] Next, the information processing apparatus 100 extracts information for diagnosis prediction from the patient data acquired in the previous step S10 (step S20).

[0038] Next, the information processing apparatus 100 inputs the information for diagnosis prediction extracted in the previous step S20 into the prediction model 132, and predicts the probability that the patient's symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) will progress (step S30).

[0039] Next, the information processing apparatus 100 outputs the prediction result regarding the probability that the patient's symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) will progress in the previous step S30 to the terminal device 10 (step S40), and ends the flowchart shown in FIG. 6.

[0040] In this embodiment, the patient information is, for example, information obtained by a doctor's interview with the patient. Examples of interview methods include MMSE (Mini-Mental State Examination), FAQ (Functional activities questionnaire), and ADAS (Alzheimer’s Disease Assessment Scale) (see Dementia Disease Diagnosis and Treatment Guidelines 2017 (https: / / www.neurology-jp.org / guidelinem / nintisyo_2017.html), Japanese Journal of Geriatrics 2011;48:431-438 (https: / / jpn-geriat-soc.or.jp / publications / other / pdf / review_geriatrics_48_5_431.pdf)).

[0041] FIG. 7 is a graph showing an example of the time change in the degree of progression of symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) predicted by the prediction model 132. In the example shown in this figure, the prediction of the onset of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) (prediction of symptom progression) at a future time point is performed over time using the prediction model 132. In this case, based on the time-series data of the cohort data (ADNI) from the past to the current time point, for example, prediction models 132 for predicting the onset of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) at each future time point such as six months later, one year later, two years later, five years later, ten years later, etc. are generated respectively. Then, by inputting the cohort data (ADNI) at the current time point into these multiple prediction models 132, it becomes possible to continuously plot the degree of progression of the symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD). And referring to the graph shown in this example, it is also possible to compare and predict the time change in the degree of progression of symptoms and the effect of medication treatment when patients with mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) receive medication treatment and when they do not.

[0042] As shown in FIG. 8, the MMSE is a type of neuropsychological test performed when dementia is suspected. The MMSE diagnoses dementia by scoring each of a plurality of diagnostic items and summing up those scores, and the lower the sum of the scores, the more advanced the symptoms are. The diagnostic items of the MMSE include, for example, orientation to time, orientation to place, repetition of object names, attention, recall of object names (delayed recall), naming of objects, repetition of sentences, three-step oral commands, reading comprehension, writing, and figure copying.

[0043] As shown in FIG. 9, the FAQ is a method for objectively evaluating the life management ability of dementia patients. The FAQ diagnoses dementia by scoring each of a plurality of diagnostic items and summing up those scores. The higher the sum of the scores, the lower the life management ability. The plurality of diagnostic items are, for example, for evaluating the ability regarding diet, memory, household management, organization, shopping, games, preparation of drinks, grasp of recent events, grasp of content from TV, and movement. <Question item> · Remembering reservations, family gatherings, holidays, and taking medications (in this specification, memory evaluated by the FAQ, or simply called FAQ memory) · Preparing a balanced diet (in this specification, diet evaluated by the FAQ, or simply called FAQ diet) · Going out to a distant place, driving a car, taking a bus: Movement (in this specification, movement evaluated by the FAQ, or simply called FAQ movement) · Shopping alone: Shopping (in this specification, shopping evaluated by the FAQ, or simply called FAQ shopping) · Household management such as bank transfers · Organizing tax records, etc. · Games that require skills · Preparation of drinks · Grasp of recent events · Paying attention to and understanding the content of TV, etc. <Score> 0: Can be done without difficulty 1: With difficulty but can be done independently 2: Assistance required 3: Full assistance

[0044] As shown in FIG. 10, ADAS is a test that is continuously performed multiple times and evaluates changes in cognitive functions based on score changes. ADAS scores each of a plurality of diagnostic items, and the higher the scores are, the more the symptoms are progressing. The plurality of diagnostic items are for evaluating abilities related to word reproduction, oral language ability, auditory comprehension of language, anomia in spontaneous speech, comprehension of oral commands, naming of fingers and objects, constructional behavior, ideomotor movement, orientation, word recognition, and reproduction of test instructions.

[0045] As shown in FIG. 11, in the present embodiment, the patient information includes, for example, information related to orientation, information related to memory, information related to IADL (Instrumental Activities of Daily Living) (see (https: / / www.mhlw.go.jp / www1 / topics / kenko21_11 / s1.html)), and information related to the positivity or negativity of amyloid β.

[0046] FIG. 12 is a diagram for explaining an example of the attribute information of the learning data, and shows the branching from the upper concept to the lower concept from the left to the right of the page. As shown in FIG. 12, the information related to orientation is information related to the ability to comprehensively judge the date, current time, location, surrounding situation, grasping of people, etc., and to grasp and understand the situation in which oneself is currently placed. For example, it includes information related to time orientation, information related to location orientation, and information related to person orientation. The information related to time orientation is, for example, information indicating the score obtained from the patient for the items related to time orientation in MMSE or ADAS. The information related to location orientation is, for example, information indicating the score obtained from the patient for the items related to location orientation in MMSE.

[0047] Information related to memory refers to information regarding a patient's ability to retain short-term or long-term memory, and includes, for example, information related to delayed recall and information related to long-term memory regarding language. Information related to delayed recall refers to information regarding the ability to recall something learned after a certain period of time has elapsed. Information related to delayed recall is, for example, information indicating the score obtained from the patient regarding items related to delayed recall in the MMSE or ADAS. Information related to delayed recall includes question items for repeating the names of objects. Information related to long-term memory regarding language refers to information regarding the ability to remember things related to language over a long period of time, and includes, for example, information regarding the ability of the subject to repeat the same words when the examiner conveys three words at one-second intervals (repeating the names of objects), information regarding the ability to state the correct name of an object while looking at the object prepared by the examiner (naming the object), and information regarding the ability to accurately remember the content conveyed by the examiner (delayed recall). Information related to long-term memory regarding language is, for example, information indicating the score obtained from the patient regarding items such as naming the object in the MMSE.

[0048] Information regarding IADL refers to information about daily life activities that involve judgment capabilities. For example, it is information for numerically evaluating capabilities related to telephone use, meal preparation, laundry, medication management, shopping, housework, transportation, and property management. Generally, after a decline in the evaluation value regarding IADL is observed, a decline in the evaluation value regarding ADL (Activities of Daily Living), which is an indicator of dementia such as eating, dressing, excretion, and bathing, is then observed. Therefore, information regarding IADL is meaningful information for grasping the signs of dementia at an early stage. Information regarding IADL is classified based on whether it is an activity of daily life that includes multiple elements, a relatively simple activity of daily life, or an activity that varies according to the culture of the region, personal preferences, or circumstances. Activities of daily life that include multiple elements are activities far from ADL. For example, they are classified into activities related to the ability to remember schedules in IADL, such as activities related to holidays, family gatherings, reservations, medication management, shopping, etc., and activities related to execution tasks in IADL, such as meal preparation and movement. Information regarding the memory of schedules in IADL is information indicating the scores obtained from patients regarding items such as holidays, family gatherings, reservations, medication management, etc. in the FAQ. Information regarding execution tasks in IADL is information indicating the scores obtained from patients regarding items such as meal preparation and movement in the FAQ. Relatively simple activities of daily life are activities close to ADL and are activities related to the ability regarding telephone use, housework, laundry, and housework. Information regarding relatively simple activities of daily life in IADL is information indicating the scores obtained from patients regarding telephone use, housework, laundry, etc. Also, activities that vary according to the culture of the region, personal preferences, or circumstances correspond to capabilities related to, for example, property management and games that require skills. Information regarding activities that vary according to the culture of the region, personal preferences, or circumstances in IADL is information indicating the scores obtained from patients regarding property management, games that require skills, etc. through neuropsychological tests (including the FAQ, etc.).

[0049] A biomarker refers to an item or substance in the body that serves as an indicator of the presence or absence of a certain disease, changes in the disease state, or the effect of treatment. Substances used as biomarkers mainly include biological data such as blood pressure, heart rate, electrocardiogram, and proteins measured in the blood. Information regarding the positivity or negativity of a biomarker may also include evaluation and determination information obtained from amyloid-β, as well as evaluation and determination information obtained from Tau, p-Tau, ApoE4, etc. Tau is a protein that accumulates in the brain in various neurodegenerative diseases including Alzheimer's disease, leading to nerve cell death and causing dementia. p-Tau refers to phosphorylated tau protein, which is an abnormal structure that appears in the brains of Alzheimer's disease patients, and the amount of its accumulation correlates with the severity of dementia. ApoE4 is the ε4 (ApoE-ε4) subtype of the apolipoprotein E (ApoE) gene and has been attracting attention as a genetic factor highly associated with AD. Information regarding the positivity or negativity of a biomarker indicates whether a certain amount of a biomarker, which is a factor highly correlated with the onset of dementia, has accumulated in the patient's brain, and is, for example, information obtained through imaging diagnosis or blood diagnosis (including CSF tests, etc.). When the biomarker is positive, it suggests that the disease state of dementia is progressing compared to when the biomarker is negative.

[0050] Next, an example of the prediction model 132 of the present embodiment will be described with reference to the drawings. As described above, examples of the evaluation indices of the prediction model 132 include accuracy, sensitivity, specificity, and AUC (Area Under the ROC Curve). The method of using the evaluation indices of the prediction model 132 is not particularly limited. In the following, as an example, when using the evaluation data of any of ADNI, J-ADNI, and MissionAD among these evaluation indices, the case where the prediction model 132 is evaluated using two evaluation indices of sensitivity and AUC will be described as an example. In the present embodiment, in the evaluation of the prediction model 132 using ADNI, J-ADNI, and MissionAD, a numerical value of "0.7", which is a generally used threshold, is applied, and the prediction model 132 that satisfies both conditions of "sensitivity ≥ 0.7" and "AUC ≥ 0.7" is taken as an example, and the prediction model 132 that does not satisfy at least one of the conditions of "sensitivity ≥ 0.7" and "AUC ≥ 0.7" is taken as a comparative example. Here, the examples of the present disclosure described below are one aspect and are not limited to the examples described below.

[0051] (Implementation conditions) For the construction and evaluation of the prediction model, first, among the ADNI data in which various information is associated with the presence or absence of symptom progression, 80% of the data is used as learning data, and a model for predicting the presence or absence of symptom progression is constructed. Further, the remaining 20% of the ADNI data, J-ADNI, and MissionAD data are used as evaluation data, the presence or absence of symptom progression based on the prediction model is compared with the actual presence or absence of symptom progression, and various evaluation indices are calculated.

[0052] (Example 1) The data used for evaluating the prediction model 132 of Example 1 is data among the learning data that was not used for constructing the prediction model 132, and includes at least four variables corresponding to any one of the first information regarding estimation, the second information regarding memory, the third information regarding the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the memory of actual operations of meals or movements in IADL, and the fourth information regarding the positive or negative of biomarkers. In other words, each of the at least four variables corresponds to any one of the first to fourth information. And the at least four variables include a first variable that is information regarding estimation related to time among the first information, a second variable that is information regarding delayed reproduction among the second information, a third variable that is information regarding the memory of holidays, family gatherings, reservations, or medication schedules in IADL among the third information, and a fourth variable that is information regarding the positive or negative of amyloid β among the fourth information.

[0053] Figures 13 and 14 are diagrams showing an example of the data used for evaluating the prediction model 132 of Example 1. In the example shown in Figures 13 and 14, the data used for evaluating the prediction model 132 includes four variables consisting of variable (1), variable (2), variable (3), and variable (4). Variable (1) is information regarding estimation of time evaluated by MMSE, variable (2) is information regarding delayed reproduction evaluated by MMSE, variable (3) is information regarding memory evaluated by FAQ, and variable (4) is information regarding Aβ positive or negative. In this example, in any evaluation of the prediction model 132 using ADNI, J-ADNI, and MissionAD, the evaluation values of both sensitivity and AUC are "0.7" or more.

[0054] (Comparative Example) FIG. 15 and FIG. 16 are diagrams showing an example of data used for evaluation of the prediction model 132 of the comparative example. In the examples shown in FIGS. 15 and 16, the data used for evaluation of the prediction model 132 includes four variables consisting of variable (1), variable (2), variable (3), and variable (4). Variables (1) to (4) correspond to any one of first information regarding estimation ability, second information regarding memory, third information regarding memory of holidays, family gatherings, reservations, or medication schedules in IADL, or memory of actual operations of meals or movements in IADL, and fourth information regarding the positive / negative of amyloid β. Variables (1) to (4) do not include information regarding estimation ability of time evaluated by MMSE among the four essential pieces of information constituting the learning data of the prediction model 132 of Example 1 shown in FIGS. 13 and 14, and are the other three essential pieces of information, i.e., information regarding delayed recall evaluated by MMSE, information regarding memory evaluated by FAQ, and information regarding Aβ positive / negative. In this example, in any evaluation of the prediction model 132 using ADNI, J-ADNI, and MissionAD, at least one of the evaluation values of sensitivity and AUC is less than "0.7".

[0055] FIG. 17 and FIG. 18 are diagrams showing an example of data used for the evaluation of the prediction model 132 of the comparative example. In the examples shown in FIGS. 17 and 18, the data used for the evaluation of the prediction model 132 includes four variables consisting of variable (1), variable (2), variable (3), and variable (4). Variables (1) to (4) correspond to either the first information regarding orientation, the second information regarding memory, the third information regarding the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the memory of actual tasks of meals or movement in IADL, or the fourth information regarding the positive / negative of amyloid β. Variables (1) to (4) do not include the information regarding memory evaluated by FAQ among the four essential information constituting the learning data of the prediction model 132 of Example 1 shown in FIGS. 13 and 14, and include the other three essential information, namely, the information regarding orientation of time evaluated by MMSE, the information regarding delayed recall evaluated by MMSE, and the information regarding Aβ positive / negative. In this example, in any evaluation of the prediction model 132 using ADNI, J-ADNI, and MissionAD, at least one of the evaluation values of sensitivity and AUC is less than "0.7".

[0056] FIG. 19 and FIG. 20 are diagrams showing an example of data used for evaluating the prediction model 132 of the comparative example. In the example shown in FIGS. 19 and 20, the data used for evaluating the prediction model 132 includes four variables consisting of variable (1), variable (2), variable (3), and variable (4). Variables (1) to (4) correspond to either the first information regarding estimation ability, the second information regarding memory, the third information regarding the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the actual performance of meals or movements in IADL, or the fourth information regarding the positive / negative of amyloid-β. Variables (1) to (4) do not include the information regarding delayed recall evaluated by MMSE among the four essential information constituting the learning data of the prediction model 132 of Example 1 shown in FIGS. 13 and 14, and include the other three essential information, namely, the information regarding estimation ability of time evaluated by MMSE, the information regarding memory evaluated by FAQ, and the information regarding Aβ positive / negative. In this example, in any evaluation of the prediction model 132 using ADNI, J-ADNI, and MissionAD, at least one of the evaluation values of sensitivity and AUC is less than "0.7".

[0057] FIG. 21 and FIG. 22 are diagrams showing an example of data used for evaluating the prediction model 132 of the comparative example. In the example shown in FIGS. 21 and 22, the data used for evaluating the prediction model 132 includes four variables consisting of variable (1), variable (2), variable (3), and variable (4). Variables (1) to (4) correspond to either the first information regarding estimation ability, the second information regarding memory, the third information regarding the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the memory of actual tasks of meals or movements in IADL, or the fourth information regarding the positive / negative of amyloid-β. Variables (1) to (4) do not include information regarding Aβ positive / negative among the four essential information constituting the learning data of the prediction model 132 of Example 1 shown in FIGS. 13 and 14, and include the other three essential information, namely, information regarding estimation ability of time evaluated by MMSE, information regarding delayed recall evaluated by MMSE, and information regarding memory evaluated by FAQ. In this example, in any evaluation of the prediction model 132 using ADNI, J-ADNI, and MissionAD, at least one of the evaluation values of sensitivity and AUC is less than "0.7".

[0058] Figs. 23 and 24 are diagrams showing an example of data used for the evaluation of the prediction model 132 of the comparative example. In the example shown in Figs. 23 and 24, the data used for the evaluation of the prediction model 132 includes four variables consisting of variable (1), variable (2), variable (3), and variable (4). Variables (1) to (4) do not correspond to any of the first information related to estimation ability, the second information related to memory, the third information related to the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the memory of actual operations of meals or movements in IADL, and the fourth information related to the positive / negative of amyloid β, and include information that does not fall into any of these categories. Variables (1) to (4) do not include information related to the estimation ability of the time evaluated by MMSE among the four essential information constituting the learning data of the prediction model 132 of Example 1 shown in Figs. 13 and 14, and include the other three essential information, namely, information related to delayed recall evaluated by MMSE, information related to memory evaluated by FAQ, and information related to Aβ positive / negative. In this example, in any of the evaluations of the prediction model 132 using ADNI, J-ADNI, and MissionAD, at least one of the evaluation values of sensitivity and AUC is less than "0.7".

[0059] FIG. 25 and FIG. 26 are diagrams showing an example of data used for the evaluation of the prediction model 132 of the comparative example. In the example shown in FIGS. 25 and 26, the data used for the evaluation of the prediction model 132 includes four variables consisting of variable (1), variable (2), variable (3), and variable (4). Variables (1) to (4) include information that does not correspond to any of the first information related to estimation ability, the second information related to memory, the third information related to the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the memory of actual operations of meals or movements in IADL, and the fourth information related to the positive or negative of amyloid β. Variables (1) to (4) do not include information related to memory evaluated by FAQ among the four essential pieces of information constituting the learning data of the prediction model 132 of Example 1 shown in FIGS. 13 and 14, and are the other three essential pieces of information, namely, information related to estimation ability of time evaluated by MMSE, information related to delayed recall evaluated by MMSE, and information related to Aβ positive or negative. In this example, in any evaluation of the prediction model 132 using ADNI, J-ADNI, and MissionAD, at least one of the evaluation values of sensitivity and AUC is less than "0.7".

[0060] Figs. 27 and 28 are diagrams showing an example of data used for the evaluation of the prediction model 132 of the comparative example. In the examples shown in Figs. 27 and 28, the data used for the evaluation of the prediction model 132 includes four variables consisting of variable (1), variable (2), variable (3), and variable (4). Variables (1) to (4) include information that does not correspond to any of the first information regarding estimation ability, the second information regarding memory, the third information regarding the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the memory of actual operations of meals or movements in IADL, and the fourth information regarding the positive or negative of amyloid β. Variables (1) to (4) do not include information regarding delayed recall evaluated by MMSE among the four essential pieces of information constituting the learning data of the prediction model 132 of Example 1 shown in Figs. 13 and 14, and include the other three essential pieces of information: information regarding estimation ability of time evaluated by MMSE, information regarding memory evaluated by FAQ, and information regarding Aβ positive or negative. In this example, in any of the evaluations of the prediction model 132 using ADNI, J-ADNI, and MissionAD, at least one of the evaluation values of sensitivity and AUC is less than "0.7".

[0061] Figures 29 and 30 are diagrams showing an example of data used for the evaluation of the prediction model 132 of the comparative example. In the examples shown in Figures 29 and 30, the data used for the evaluation of the prediction model 132 includes four variables consisting of variable (1), variable (2), variable (3), and variable (4). Variables (1) to (4) do not correspond to any of the first information related to estimation ability, the second information related to memory, the third information related to the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the memory of actual operations of meals or movements in IADL, and the fourth information related to the positive / negative of amyloid β, and include information that does not fall into any of these categories. Variables (1) to (4) do not include information related to Aβ positive / negative among the four essential information constituting the learning data of the prediction model 132 of Example 1 shown in Figures 13 and 14, and are the other three essential information, namely, information related to the estimation ability of the time evaluated by MMSE, information related to the delayed reproduction evaluated by MMSE, and information related to the memory evaluated by FAQ. In this example, in any evaluation of the prediction model 132 using ADNI, J-ADNI, and MissionAD, at least one of the evaluation values of sensitivity and AUC is less than "0.7".

[0062] (Example 2) The data used for evaluating the prediction model 132 of Example 2 includes at least four variables corresponding to any one of the first information regarding estimation ability, the second information regarding memory, the third information regarding the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the memory of actual tasks of meals or movements in IADL, and the fourth information regarding the positivity or negativity of biomarkers. In other words, each of the at least four variables corresponds to any one of the information from the first information to the fourth information. And the at least four variables are the first variable which is the information regarding estimation ability related to time among the first information, the second variable which is the information regarding delayed reproduction among the second information, the third variable which is the information regarding the memory of holidays, family gatherings, reservations, or medication schedules in IADL among the third information, the fourth variable which is the information regarding the positivity or negativity of amyloid-β among the fourth information, and the fifth variable which is any one of the information from the first information, the second information, and the third information. The fifth variable may be, for example, the information regarding estimation ability of location among the first information, the information regarding the memory of the name or type of an item among the second information, the information regarding the memory of the actual task of meals in IADL among the third information, or the information regarding the memory of the actual task of movements in IADL among the third information. In Example 2 below, an example using five variables will be described, but additional other variables corresponding to any one of the information from the first information to the fourth information may be added.

[0063] Figures 31 and 32 are diagrams showing an example of data used for the evaluation of the prediction model 132 of Example 2. In the examples shown in FIGS. 31 and 32, the data used for the evaluation of the prediction model 132 includes five variables consisting of variable (1), variable (2), variable (3), variable (4), and variable (5). In the first data set, variable (1) is information regarding the sense of time evaluated by MMSE, variable (2) is information regarding delayed recall evaluated by MMSE, variable (3) is information regarding memory evaluated by FAQ, variable (4) is information regarding diet evaluated by FAQ, and variable (5) is information regarding Aβ positive / negative. In the second data set, variable (1) is information regarding the sense of time evaluated by MMSE, variable (2) is information regarding the sense of location evaluated by MMSE, variable (3) is information regarding delayed recall evaluated by MMSE, variable (4) is information regarding memory evaluated by FAQ, and variable (5) is information regarding Aβ positive / negative. In the third data set, variable (1) is information regarding the sense of time evaluated by MMSE, variable (2) is information regarding delayed recall evaluated by MMSE, variable (3) is information regarding memory evaluated by FAQ, variable (4) is information regarding movement evaluated by FAQ, and variable (5) is information regarding Aβ positive / negative. In the fourth data set, variable (1) is information regarding the sense of time evaluated by MMSE, variable (2) is information regarding delayed recall evaluated by MMSE, variable (3) is information regarding the recall of item names evaluated by MMSE, variable (4) is information regarding memory evaluated by FAQ, and variable (5) is information regarding Aβ positive / negative. In this example, for the first to fourth data sets, in any evaluation of the prediction model 132 using ADNI, J-ADNI, and MissionAD, the evaluation values of both sensitivity and AUC are "0.7" or more.

[0064] (Comparative Example) FIG. 33 and FIG. 34 are diagrams showing an example of data used for the evaluation of the prediction model 132 of the comparative example. In the example shown in FIGS. 33 and 34, the data used for the evaluation of the prediction model 132 includes five variables consisting of variable (1), variable (2), variable (3), variable (4), and variable (5). Variables (1) to (5) correspond to any one of the first information regarding estimation ability, the second information regarding memory, the third information regarding the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the memory of actual operations of meals or movements in IADL, and the fourth information regarding the positive / negative of amyloid-β. Variables (1) to (5) do not include the information regarding the estimation ability of the time evaluated by MMSE among the four essential information constituting the learning data of the prediction model 132 of Example 2 shown in FIGS. 31 and 32, and include the other three essential information, i.e., the information regarding the delayed recall evaluated by MMSE, the information regarding memory evaluated by FAQ, and the information regarding Aβ positive / negative. In this example, in any evaluation of the prediction model 132 using ADNI, J-ADNI, and MissionAD, at least one of the evaluation values of sensitivity and AUC is less than "0.7".

[0065] FIG. 35 and FIG. 36 are diagrams showing an example of data used for evaluating the prediction model 132 of the comparative example. In the example shown in FIGS. 35 and 36, the data used for evaluating the prediction model 132 includes five variables consisting of variable (1), variable (2), variable (3), variable (4), and variable (5). Variables (1) to (5) correspond to either the first information regarding estimation ability, the second information regarding memory, the third information regarding the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the memory of actual operations of meals or movements in IADL, or the fourth information regarding the positive or negative of amyloid-β. Variables (1) to (5) do not include the information regarding memory evaluated by FAQ among the four essential information constituting the learning data of the prediction model 132 of Example 2 shown in FIGS. 31 and 32, and include the information regarding estimation ability of time evaluated by MMSE, the information regarding delayed recall evaluated by MMSE, and the information regarding Aβ positive or negative, which are the other three essential information. In this example, in any evaluation of the prediction model 132 using ADNI, J-ADNI, and MissionAD, at least one of the evaluation values of sensitivity and AUC is less than "0.7".

[0066] Figs. 37 and 38 are diagrams showing an example of data used for the evaluation of the prediction model 132 of the comparative example. In the example shown in Figs. 37 and 38, the data used for the evaluation of the prediction model 132 includes five variables consisting of variable (1), variable (2), variable (3), variable (4), and variable (5). Variables (1) to (5) correspond to any one of the first information regarding estimation ability, the second information regarding memory, the third information regarding memory of holidays, family gatherings, reservations, or medication schedules in IADL, or memory of actual operations of meals or movements in IADL, and the fourth information regarding the positive or negative of amyloid β. Variables (1) to (4) do not include information regarding delayed recall evaluated by MMSE among the four essential information constituting the learning data of the prediction model 132 of Example 2 shown in Figs. 31 and 32, and are the other three essential information, namely, information regarding estimation ability of time evaluated by MMSE, information regarding memory evaluated by FAQ, and information regarding Aβ positive or negative. In this example, in any evaluation of the prediction model 132 using ADNI, J-ADNI, and MissionAD, at least one of the evaluation values of sensitivity and AUC is less than "0.7".

[0067] Figures 39 and 40 are diagrams showing an example of data used for evaluating the prediction model 132 of the comparative example. In the examples shown in FIGS. 39 and 40, the data used for evaluating the prediction model 132 includes five variables consisting of variable (1), variable (2), variable (3), variable (4), and variable (5). Variables (1) to (5) correspond to any one of the first information regarding estimation knowledge, the second information regarding memory, the third information regarding the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the actual operation of meals or movement in IADL, and the fourth information regarding the positive / negative of amyloid β. Variables (1) to (5) do not include the information regarding Aβ positive / negative among the four essential information constituting the learning data of the prediction model 132 of Example 2 shown in FIGS. 31 and 32, and include the other three essential information, namely, the information regarding estimation knowledge of the time evaluated by MMSE, the information regarding delayed reproduction evaluated by MMSE, and the information regarding memory evaluated by FAQ. In this example, in any evaluation of the prediction model 132 using ADNI, J-ADNI, and MissionAD, at least one of the evaluation values of sensitivity and AUC is less than "0.7".

[0068] Figures 41 and 42 are diagrams showing an example of learning data used for learning the prediction model 132 of the comparative example. In the examples shown in FIGS. 41 and 42, the learning data used for learning the prediction model 132 includes five variables consisting of variable (1), variable (2), variable (3), variable (4), and variable (5). Variables (1) to (5) include information that does not correspond to any of the first information regarding estimation knowledge, the second information regarding memory, the third information regarding the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the memory of actual tasks of meals or movements in IADL, and the fourth information regarding the positive or negative of amyloid β. Variables (1) to (5) do not include information regarding the estimation knowledge of the time evaluated by MMSE among the four essential information constituting the learning data of the prediction model 132 of Example 2 shown in FIGS. 31 and 32, and include the other three essential information, namely, information regarding delayed reproduction evaluated by MMSE, information regarding memory evaluated by FAQ, and information regarding Aβ positive or negative. In this example, in any evaluation of the prediction model 132 using ADNI, J-ADNI, or MissionAD, at least one of the evaluation values of sensitivity and AUC is less than "0.7".

[0069] Figures 43 and 44 are diagrams showing an example of learning data used for learning the prediction model 132 of the comparative example. In the example shown in FIGS. 43 and 44, the learning data used for learning the prediction model 132 includes five variables consisting of variable (1), variable (2), variable (3), variable (4), and variable (5). Variables (1) to (5) include information that does not correspond to any of the first information regarding estimation knowledge, the second information regarding memory, the third information regarding the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the memory of actual operations of meals or movements in IADL, and the fourth information regarding the positive or negative of amyloid β. Variables (1) to (5) do not include information regarding memory evaluated by FAQ among the four essential information constituting the learning data of the prediction model 132 of Example 2 shown in FIGS. 31 and 32, and are the other three essential information, namely, information regarding estimation knowledge of time evaluated by MMSE, information regarding delayed reproduction evaluated by MMSE, and information regarding Aβ positive or negative. In this example, in any evaluation of the prediction model 132 using ADNI, J-ADNI, or MissionAD, at least one evaluation value of sensitivity and AUC is less than "0.7".

[0070] Figures 45 and 46 are diagrams showing an example of the learning data used for the learning of the prediction model 132 of the comparative example. In the example shown in FIGS. 45 and 46, the learning data used for the learning of the prediction model 132 includes five variables consisting of variable (1), variable (2), variable (3), variable (4), and variable (5). Variables (1) to (4) include information that does not correspond to any of the first information related to estimation knowledge, the second information related to memory, the third information related to the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the memory of actual tasks of meals or movements in IADL, and the fourth information related to the positive or negative of amyloid β. Variables (1) to (5) do not include information related to delayed reproduction evaluated by MMSE among the four essential information constituting the learning data of the prediction model 132 of Example 2 shown in FIGS. 31 and 32, and include the other three essential information, that is, information related to estimation knowledge of time evaluated by MMSE, information related to memory evaluated by FAQ, and information related to Aβ positive or negative. In this example, in any evaluation of the prediction model 132 using ADNI, J-ADNI, and MissionAD, at least one of the evaluation values of sensitivity and AUC is less than "0.7".

[0071] FIGS. 47 and 48 are diagrams showing an example of learning data used for learning the prediction model 132 of the comparative example. In the examples shown in FIGS. 47 and 48, the learning data used for learning the prediction model 132 includes five variables consisting of variable (1), variable (2), variable (3), variable (4), and variable (5). Variables (1) to (5) include information that does not correspond to any of the first information related to estimation knowledge, the second information related to memory, the third information related to the memory of holidays, family gatherings, reservations, or medication schedules in IADL, or the memory of actual tasks of meals or movements in IADL, and the fourth information related to the positive / negative of amyloid β. Variables (1) to (5) do not include information related to Aβ positive / negative among the four essential information constituting the learning data of the prediction model 132 of Example 2 shown in FIGS. 31 and 32, and include the other three essential information, that is, information related to estimation knowledge of the time evaluated by MMSE, information related to delayed reproduction evaluated by MMSE, and information related to memory evaluated by FAQ. In this example, in any evaluation of the prediction model 132 using ADNI, J-ADNI, and MissionAD, at least one of the evaluation values of sensitivity and AUC is less than "0.7".

[0072] According to the information processing apparatus 100 described above, as learning data used for learning the prediction model 132, by selecting data that can obtain a high evaluation as an evaluation based on a plurality of evaluation data 131 such as ADNI, J-ADNI, and MissionAD, which are not easily affected by differences in culture, habits, and preferences due to covering a wide area, high prediction accuracy can be achieved regardless of race, culture, etc. Further, by selecting data that can be relatively easily obtained by questionnaires for patients as learning data used for learning the prediction model 132, the time for diagnosis (questionnaire) can be shortened, the physical burden of examinations (no biological sample collection), and the number of hospital visits can be reduced, and the processing load on patients and the apparatus can be reduced.

[0073] 〔Hardware Configuration〕 FIG. 49 is a diagram showing an example of the hardware configuration of the information processing apparatus 100 according to the present embodiment. As shown in the figure, the information processing apparatus 100 includes a communication controller 100-1, a CPU 100-2, a RAM (Random Access Memory) 100-3 used as a working memory, a ROM (Read Only Memory) 100-4 that stores a boot program and the like, a storage device 100-5 such as a flash memory or an HDD (Hard Disk Drive), a drive device 100-6, etc., which are connected to each other by an internal bus or a dedicated communication line. The communication controller 100-1 communicates with components other than the information processing apparatus 100. A program 100-5a executed by the CPU 100-2 is stored in the storage device 100-5. This program is expanded in the RAM 100-3 by a DMA (Direct Memory Access) controller (not shown) or the like and executed by the CPU 100-2. Thereby, the acquisition unit 121, the prediction unit 122, and the output unit 123 are realized.

[0074] Note that the embodiments described above are for facilitating the understanding of the present invention and are not for limiting the interpretation of the present invention. The present invention can be changed / improved without departing from its gist, and equivalents thereof are also included in the present invention. That is, what those skilled in the art appropriately modify in each embodiment is also included in the scope of the present invention as long as it has the features of the present invention. Also, it goes without saying that each embodiment is an example, and partial substitution or combination of the configurations shown in different embodiments is possible, and these are also included in the scope of the present invention as long as they include the features of the present invention.

Explanation of Reference Numerals

[0075] 10... Terminal device, 100... Information processing apparatus, 110... Communication device, 120... Control unit, 121... Acquisition unit, 122... Prediction unit, 123... Output unit, 130... Storage unit, 131... Evaluation data, 132... Prediction model, NW... Communication network

Claims

1. first information related to estimation knowledge, second information related to memory, third information related to memory of holidays, family gatherings, reservations, or medication schedules in IADL, or memory of actual operations of meals or movements in IADL, at least four variables corresponding to any of the above, and presence or absence of symptom progression of mild cognitive impairment (MCI) or mild Alzheimer's type dementia (mild AD), an information processing apparatus comprising a prediction unit that inputs patient information corresponding to each of the at least four variables into a prediction model learned using learning data in which the above are associated, and predicts the probability that the patient will develop symptoms of mild cognitive impairment (MCI) or mild Alzheimer's type dementia (mild AD), wherein the at least four variables are a first variable that is information related to estimation knowledge regarding time among the first information, a second variable that is information related to delayed reproduction among the second information, a third variable that is information related to memory of holidays, family gatherings, reservations, or medication schedules in IADL among the third information, a fourth variable that is information related to the positivity or negativity of amyloid β among the fourth information and includes an information processing apparatus.

2. The at least four variables are four variables consisting of the first variable, the second variable, the third variable, and the fourth variable, The information processing apparatus according to claim 1.

3. The at least four variables are five variables consisting of the first variable, the second variable, the third variable, the fourth variable, and a fifth variable that is any one of the first information, the second information, and the third information, The information processing apparatus according to claim 1.

4. The fifth variable is information related to estimation knowledge regarding location among the first information, The information processing apparatus according to claim 3.

5. The fifth variable is information related to memory of the name or type of an item among the second information, The information processing apparatus according to claim 3.

6. The fifth variable is information related to memory of the actual operation of a meal in IADL among the third information, The information processing apparatus according to claim 3.

7. The fifth variable is information related to memory of the actual operation of movement in IADL among the third information, The information processing apparatus according to claim 3.

8. A computer acquires patient information, At least four variables corresponding to any one of first information related to orientation, second information related to memory, third information related to memory of holidays, family gatherings, reservations, or medication schedules in IADL, or memory of actual tasks of eating or moving in IADL, and fourth information related to the positivity or negativity of biomarkers, and the presence or absence of symptom progression of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD). For a prediction model trained using learning data in which these are associated, input information corresponding to each of the at least four variables included in the patient information, and predict the probability that the patient will progress the symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD). Output a prediction result regarding the probability that the patient will progress the symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD). An information processing method for executing the above. The at least four variables include a first variable which is information related to orientation regarding time among the first information. A second variable which is information related to delayed reproduction among the second information. A third variable which is information related to memory of holidays, family gatherings, reservations, or medication schedules in IADL among the third information. A fourth variable which is information related to the positivity or negativity of amyloid-β among the fourth information and include. Information processing method.

9. On a computer, Execute a process of acquiring patient information, For a prediction model trained using learning data in which at least four variables corresponding to any one of first information related to orientation, second information related to memory, third information related to memory of holidays, family gatherings, reservations, or medication schedules in IADL, or memory of actual tasks of eating or moving in IADL, and fourth information related to the positivity or negativity of biomarkers, and the presence or absence of symptom progression of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD) are associated, input information corresponding to each of the at least four variables included in the patient information, and predict the probability that the patient will progress the symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD). Execute a process of outputting a prediction result regarding the probability that the patient will progress the symptoms of mild cognitive impairment (MCI) or mild Alzheimer's disease (mild AD). The at least four variables are A first variable that is information regarding an estimate of time among the first information, A second variable that is information regarding delayed playback among the second information, A third variable that is information regarding memory of holidays, family gatherings, reservations, or medication schedules in IADL among the third information, A fourth variable that is information regarding the positive or negative of amyloid β among the fourth information including a program.

10. A computer-readable non-transitory storage medium storing the program according to Claim 9.

Images