JP2018114232A - Diagnosis apparatus, program, and diagnosis system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a diagnosis apparatus capable of easily performing objective and quantitative measurement of stress.SOLUTION: A diagnosis apparatus includes: handwriting information acquisition means for acquiring handwriting information on a user; handwriting feature amount calculation means for calculating a handwriting feature amount of the user from the handwriting information, and accumulating the handwriting feature amount of the user in an accumulation part; and stress determination means for determining whether or not the user is in a stressed state from an amount of deviation between a first handwriting feature amount calculated by the handwriting feature amount calculation means and a second handwriting feature amount based on one or more past handwriting feature amounts accumulated in the accumulation part.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a diagnostic device, a program, and a diagnostic system.

  It is known that accumulating stress may cause severe mental illness such as depression. Therefore, it is required to detect and take measures against depression in advance and early by measuring stress objectively and quantitatively.

  For example, the effect of stress on writing has already been studied, and it has been reported that the size and writing pressure are affected (see, for example, Non-Patent Documents 1 and 2).

  The measurement methods for objectively and quantitatively measuring stress are mainly known as a measurement method using a biomarker such as amylase and a measurement method using a circulatory response (LF (low frequency) / HF (high frequency) ratio of heart rate variability). ing.

  However, there is a problem that the measurement method for objectively and quantitatively measuring the stress cannot be routinely performed at home without a special measurement instrument, and needs to be performed in a hospital or the like.

  An object of an embodiment of the present invention is to provide a diagnostic apparatus that can easily perform objective and quantitative stress measurement.

  In order to achieve the above object, according to an embodiment of the present invention, handwriting information acquisition means for acquiring user handwriting information, a user handwriting feature amount is calculated from the handwriting information, and the user handwriting feature amount is stored in a storage unit. Handwritten feature amount calculation means to be stored; first handwriting feature amount calculated by the handwriting feature amount calculation means; and second handwriting based on the one or more past handwriting feature amounts stored in the storage unit The diagnostic apparatus includes: a stress determination unit that determines whether the user is in a stress state based on a deviation amount from the feature amount.

  It is possible to provide a diagnostic apparatus that can easily perform objective and quantitative stress measurement.

It is a lineblock diagram of an example of a diagnostic system concerning a 1st embodiment. It is a hardware block diagram of an example of a computer. It is a functional block diagram of an example of a diagnostic system concerning this embodiment. It is the flowchart which showed an example of the process sequence of the diagnostic system which concerns on this embodiment. It is a figure explaining an example of the calculation method of a character size. It is a figure explaining an example of the selection method of the maximum pen pressure value. It is a figure explaining an example of the stress determination method. It is the flowchart which showed an example of the process sequence of the stress determination method. It is the flowchart which showed the other example of the process sequence of the stress determination method. It is an example of the figure explaining the outline of calculation of a risk value. It is an example of the whole block diagram of a diagnostic system. It is an example of the functional block diagram explaining the function of the terminal or apparatus which the diagnostic system 100 has. The structural example of the neural network which is an example of the model of machine learning is shown. It is a flowchart figure of an example of the procedure which registers diagnostic data. It is a flowchart figure of an example of the procedure which notifies a database management apparatus that diagnostic data was updated. It is a flowchart figure of an example of the procedure which registers biomarker data. It is a flowchart figure of an example of the procedure which registers handwriting feature-value data. It is a flowchart figure of an example of the procedure which registers a diagnostic subject. It is a flowchart figure of an example of the procedure which registers recommendation information. It is a flowchart figure of an example of the procedure in which a database management apparatus performs data operation. It is a flowchart figure of an example of the procedure in which a risk value calculation apparatus calculates a risk value. It is a flowchart figure of an example of the procedure in which a log | history management apparatus registers a log | history. It is a flowchart figure of an example of the procedure in which a risk value calculation terminal calculates a risk value in time series. It is an image figure of an example of the risk value screen displayed on the display of the risk value calculation terminal. It is a flowchart figure of an example explaining the calculation method of correlation data using two estimated diagnostic data.

Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings.
[First Embodiment]
<System configuration>
FIG. 1 is a configuration diagram of an example of a diagnostic system according to the first embodiment. The diagnostic system 1 in FIG. 1 includes a liquid crystal tablet 10 and a PC (personal computer) 20. The liquid crystal tablet 10 of FIG. 1 has a display unit and a sensing unit, and pen input is possible by the instruction unit 30. The PC 20 functions as a calculation unit and becomes a host of the liquid crystal tablet 10.

  When the user writes a character on the sensing unit using the pen-shaped instruction unit 30, the sensing unit acquires the position coordinates and the writing pressure of the tip of the instruction unit 30. The position coordinates and writing pressure acquired by the sensing unit are transmitted to the PC 20. The PC 20 generates display data from the received position coordinates and writing pressure, and displays characters and a stress determination result described later on the display unit of the liquid crystal tablet 10.

  The configuration of the diagnostic system 1 in FIG. 1 is an example, and a terminal device capable of handwriting input to a touch panel, such as a tablet terminal or a smartphone, in which a display unit, a sensing unit, and a control unit are housed in one housing It may be. Further, the character input to the sensing unit may be performed by the user using the fingertip in addition to the pen-type instruction unit 30 such as a pen-type input device. Further, the liquid crystal tablet 10 and the PC 20 of the diagnostic system 1 of FIG. 1 may be connected via the Internet or a LAN (local area network).

  Further, cloud computing may be applied to the PC 20. For example, in cloud computing, hardware may be stored in one housing or divided into a plurality of housings, and resources are dynamically configured. In addition, a plurality of server functions may be constructed in a virtual environment in one PC 20, and one server function may be constructed across a plurality of PCs 20.

<Hardware configuration>
The PC 20 is configured as a computer 500 as shown in FIG. FIG. 2 is a hardware configuration diagram of an example of a computer. The computer 500 includes a CPU (Central Processing Unit) 501, a ROM (Read Only Memory) 502, a RAM (Random Access Memory) 503, and an auxiliary storage device 504. The computer 500 further includes an input unit 505, a display I / F (Interface) 506, a network I / F 507, and an external device I / F 508. Note that the components of the computer 500 are connected to each other via a bus B.

  The CPU 501 executes a program 504p, an OS (Operating System), and the like stored in the auxiliary storage device 504. The ROM 502 is a nonvolatile memory. The ROM 502 stores a system loader and data.

  The RAM 503 is a main storage device such as a DRAM (Dynamic Random Access Memory) or an SRAM (Static Random Access Memory). When executed by the CPU 501, the program 504 p stored in the auxiliary storage device 504 is expanded in the RAM 503. The RAM 503 serves as a work area for the CPU 501.

  The auxiliary storage device 504 stores a program 504p executed by the CPU 501 and various databases used when the program 504p is executed by the CPU 501. The auxiliary storage device 504 is a nonvolatile memory such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive).

  The input unit 505 is an interface for an operator to input various instructions to the computer 500. For example, the input unit 505 is a keyboard, a mouse, a touch panel, a voice input device, or the like. However, the input unit 505 may be connected as necessary.

  In response to a request from the CPU 501, the display I / F 506 displays various information included in the computer 500 on the display 510, which is a display device, in the form of a cursor, menu, window, character, image, or the like. The display I / F 506 is, for example, a graphic chip or a display I / F. However, the display 510 may be connected as necessary.

  The network I / F 507 communicates with a terminal device such as the liquid crystal tablet 10 and another computer 500 via the network. The network I / F 507 is realized by a device such as a LAN adapter, for example.

  The external device I / F 508 is an interface for connecting various storage media 520 such as a USB cable or a USB memory. The external device I / F 508 can also communicate with a terminal device such as the liquid crystal tablet 10 via a USB cable. Note that the hardware configuration of FIG. 2 is an example, and other hardware configurations may be employed.

<Software configuration>
The diagnostic system 1 according to the present embodiment is realized by, for example, the functional blocks shown in FIG. FIG. 3 is a functional block diagram of an example of the diagnostic system according to the present embodiment. The diagnostic system 1 in FIG. 3 omits functional blocks that are not necessary for the description of the present embodiment.

  The liquid crystal tablet 10 in FIG. 3 includes a sensing unit 11 and a display unit 12. 3 includes a handwriting information acquisition unit 21, a display data generation unit 22, a handwriting feature amount calculation unit 23, a handwriting feature amount storage unit 24, and a stress determination unit 25. The handwriting information acquisition unit 21, the display data generation unit 22, the handwriting feature amount calculation unit 23, the handwriting feature amount accumulation unit 24, and the stress determination unit 25 in FIG. 3 are realized by executing a program in the computer 500 in FIG. .

  The sensing unit 11 of the liquid crystal tablet 10 acquires the position coordinates and the writing pressure of the tip of the instruction unit 30 when the user writes a character using the pen-type instruction unit 30, and transmits it to the PC 20 as handwriting information described later. . The handwriting information acquisition unit 21 of the PC 20 acquires handwriting information transmitted from the liquid crystal tablet 10.

  The display data generation unit 22 generates display data of a drawing result from the acquired handwriting information, and displays the drawing result on the display unit 12 of the liquid crystal tablet 10. The handwriting feature value calculation unit 23 calculates a handwriting feature value, which will be described later, and stores the handwriting feature value in the handwriting feature value storage unit 24. The stress determination unit 25 determines a user's stress based on the handwriting feature amount stored in the handwriting feature amount storage unit 24 as described later. The display data generation unit 22 generates display data of the stress determination result and displays the drawing result on the display unit 12 of the liquid crystal tablet 10.

<Details of processing>
Next, details of processing of the diagnostic system 1 according to the present embodiment will be described. FIG. 4 is a flowchart showing an example of a processing procedure of the diagnostic system according to the present embodiment. In step S <b> 11, the handwriting information acquisition unit 21 of the PC 20 acquires handwriting information as shown in Table 1 from the sensing unit 11 of the liquid crystal tablet 10.

表１は筆跡情報の一例の構成図である。表１の筆跡情報は、時刻、ストローク番号、Ｘ座標、Ｙ座標及び筆圧を項目として有している。時刻は、センシング部１１が指示部３０の先端の位置座標及び筆圧を取得した時刻を表す。表１では時刻がミリ秒精度で記録されている例を示している。また、表１ではセンシング部１１が指示部３０の先端の位置座標及び筆圧を取得するサンプリング間隔が８ミリ秒である例を示している。

Table 1 is a configuration diagram of an example of handwriting information. The handwriting information in Table 1 includes time, stroke number, X coordinate, Y coordinate, and writing pressure as items. The time represents the time when the sensing unit 11 acquires the position coordinates and the writing pressure of the tip of the instruction unit 30. Table 1 shows an example in which the time is recorded with millisecond accuracy. Table 1 shows an example in which the sampling interval at which the sensing unit 11 acquires the position coordinates of the tip of the instruction unit 30 and the writing pressure is 8 milliseconds.

  The stroke number is an example of identification information (ID) that identifies a stroke. In the present embodiment, a sequence from when the instruction unit 30 comes into contact with the sensing unit 11 until it leaves is called a stroke.

  The X coordinate and the Y coordinate are position coordinates of the tip of the instruction unit 30 acquired by the sensing unit 11. The X and Y coordinates in Table 1 are examples represented in a screen coordinate system in which the upper left of the sensing unit 11 is the origin and the unit is the display pixel. The pen pressure is a pressure applied to the tip of the instruction unit 30 acquired by the sensing unit 11.

  Returning to step S12 of FIG. 4, the display data generation unit 22 of the PC 20 generates display data of a drawing result as a normal drawing application based on the handwriting information of Table 1. In FIG. 4, handwritten “A” is displayed as an example of display data of the drawing result. In parallel with the processing of step S12, the handwriting feature value calculation unit 23 of the PC 20 calculates handwriting feature values as shown in Table 2 in step S13 and stores them in the handwriting feature value storage unit 24, for example.

表２は筆跡特徴量の一例の構成図である。表２の筆跡特徴量は、時刻、文字番号及び文字サイズを項目として有している。時刻は、筆跡特徴量算出部２３が筆跡特徴量を算出した時刻を表す。文字番号は、表１の筆跡情報から認識された文字を識別する識別情報の一例である。文字サイズは、表１の筆跡情報から認識された文字の大きさを表す情報の一例である。

Table 2 is a configuration diagram of an example of the handwriting feature amount. The handwriting feature values in Table 2 have time, character number, and character size as items. The time represents the time when the handwriting feature value calculation unit 23 calculates the handwriting feature value. The character number is an example of identification information for identifying a character recognized from the handwriting information in Table 1. The character size is an example of information indicating the size of the character recognized from the handwriting information in Table 1.

  The handwriting feature amount calculation unit 23 groups the stroke group represented by the handwriting information in Table 1 in units of characters. In addition, the process which groups the stroke group represented by handwriting information per character can utilize the well-known technique described in Unexamined-Japanese-Patent No. 2007-58803 etc., for example. For example, Japanese Patent Application Laid-Open No. 2007-58803 proposes a character cutting method for extracting individual characters from a plurality of written input strokes without restrictions on the size and position of characters to be written.

  The handwriting feature amount calculation unit 23 assigns a character number to each stroke group grouped in character units, and further calculates a character size from the position coordinates of the stroke groups grouped in character units, for example, as shown in FIG. .

FIG. 5 is a diagram for explaining an example of a character size calculation method. The handwriting feature amount calculation unit 23 calculates the area of the rectangular region including the character from the difference value between the minimum value and the maximum value of the X coordinate and the Y coordinate of the position coordinates of the stroke group grouped in character units, To do. In FIG. 5, the difference value between the minimum value and the maximum value of the X coordinate is “20 px”, the difference value between the minimum value and the maximum value of the Y coordinate is “50 px”, and the area of the rectangular area including characters is “ An example of “1000 px ² ” is shown.

  In step S <b> 13, the handwriting feature value calculation unit 23 may calculate a handwriting feature value as shown in Table 3 and store it in the handwriting feature value accumulation unit 24.

表３は筆跡特徴量の他の例の構成図である。表３の筆跡特徴量は、時刻、文字番号及び筆圧最大値を項目として有している。時刻は、筆跡特徴量算出部２３が筆跡特徴量を算出した時刻を表す。文字番号は、表１の筆跡情報から認識された文字を識別する識別情報の一例である。筆圧最大値は、文字単位にグルーピングされたストローク群の筆圧から最大値を選択したものである。

Table 3 is a configuration diagram of another example of the handwriting feature amount. The handwriting feature values in Table 3 have time, character number, and maximum writing pressure as items. The time represents the time when the handwriting feature value calculation unit 23 calculates the handwriting feature value. The character number is an example of identification information for identifying a character recognized from the handwriting information in Table 1. The maximum writing pressure value is obtained by selecting the maximum value from the writing pressures of stroke groups grouped in character units.

  FIG. 6 is a diagram for explaining an example of a method for selecting the maximum writing pressure value. The handwriting feature amount calculation unit 23 shows, for example, the maximum writing pressure values of stroke numbers “1” to “3” grouped in units of characters. In the example of FIG. 6, the maximum writing pressure value of the stroke number “3” is selected as the maximum writing pressure value of the character.

  The handwriting feature value calculated by the handwriting feature value calculation unit 23 is not limited to the character size or the maximum writing pressure value, but can be obtained from handwriting information and is influenced by stress (mental load). If it is. The handwriting feature value accumulating unit 24 stores handwriting feature values of a plurality of users in association with information identifying the user such as a user ID.

  Returning to step S14 in FIG. 4, the stress determination unit 25 of the PC 20 performs the stress determination of the user as shown in FIG. 7 based on the data of the handwriting feature amount stored in the handwriting feature amount storage unit 24. The stress determination method in FIG. 7 is based on the premise that the character size of the user in the stress state is reduced.

  FIG. 7 is a diagram for explaining an example of a stress determination method. FIG. 7 shows a stress determination method using a character size as an example. The stress determination unit 25 calculates the average value (Ave_size) of the character sizes stored in the handwriting feature amount accumulation unit 24. The stress determination unit 25 determines that the character is in a stress state if the period when the character size becomes smaller than the threshold (Thr_T) by the threshold (Thr_size) or more than the average value (Ave_size) of the character size. The stress determination unit 25 may use a character size averaged over a span of about one week for stress determination in order to randomize the influence due to the change in the usage situation.

  FIG. 8 is a flowchart showing an example of the processing procedure of the stress determination method. FIG. 8 shows a stress determination method using a character size as an example. In step S <b> 21, the stress determination unit 25 acquires the latest character size (Now_size) calculated by the handwriting feature value calculation unit 23 from the handwriting feature value calculation unit 23 or the handwriting feature value accumulation unit 24. The latest character size (Now_size) may be a character size averaged over a span of about one week.

  In step S <b> 22, the stress determination unit 25 calculates an average value (Ave_size) of past character sizes stored in the handwriting feature amount accumulation unit 24. In step S23, the stress determination unit 25 determines whether or not the latest character size (Now_size) is smaller than the character size obtained by subtracting the threshold value (Thr_size) from the average character size (Ave_size) (True). To do.

  If the determination in Step S23 is True, the stress determination unit 25 proceeds to Step S24, and determines whether the duration of the True determination in Step S23 is equal to or greater than a threshold (Thr_T) (True).

  If the determination in step S24 is True, the stress determination unit 25 proceeds to step S25 and determines that the user is in a stress state. On the other hand, if the determination in step S23 or step S24 is False, the stress determination unit 25 determines in step S26 that the user is not in a stress state.

  In addition, in step S14 of FIG. 4, the stress determination part 25 of PC20 may detect the fall of a pen pressure maximum value and may perform a user's stress determination similarly to a character size. This stress determination method is based on the premise that the maximum writing pressure value of a user who is in a stress state is small.

  FIG. 9 is a flowchart showing another example of the processing procedure of the stress determination method. FIG. 9 shows a stress determination method using the maximum writing pressure value as an example. In step S <b> 31, the stress determination unit 25 acquires the latest writing pressure maximum value (Now_P) selected by the handwriting feature value calculation unit 23 from the handwriting feature value calculation unit 23 or the handwriting feature value accumulation unit 24. The latest writing pressure maximum value (Now_P) may be a writing pressure maximum value averaged over a span of about one week.

  In step S <b> 32, the stress determination unit 25 calculates an average value (Ave_P) of past writing pressure maximum values stored in the handwriting feature amount accumulation unit 24. In step S33, the stress determination unit 25 determines whether or not the latest writing pressure maximum value (Now_P) is smaller than the writing pressure maximum value obtained by subtracting the threshold value (Thr_P) from the average writing pressure maximum value (Ave_P) (True). (False) is determined.

  If the determination in step S33 is True, the stress determination unit 25 proceeds to step S34, and determines whether or not the determination duration of True in step S33 is equal to or greater than a threshold value (Thr_T) (True).

  If the determination in step S34 is True, the stress determination unit 25 proceeds to step S35 and determines that the user is in a stress state. On the other hand, if the determination in Step S33 or Step S34 is False, the stress determination unit 25 determines in Step S36 that the user is not in a stress state.

  Returning to step S15 in FIG. 4, the display data generation unit 22 of the PC 20 generates display data of the stress determination result for feeding back the stress determination result to the user. In FIG. 4, as an example of display data of the stress determination result, a message “Stress is high. In step S16, the display data generation unit 22 causes the display unit 12 to display the drawing result display data generated in step S12 and the stress determination result display data generated in step S15.

  The diagnosis system 1 according to the present embodiment can also be moved behind an application that writes handwritten characters in a frame and inputs characters in order to suppress fluctuations in handwriting feature values depending on the situation. As an application for writing a handwritten character in a frame and inputting the character, for example, there is a tablet PC input panel that is standardly installed in Windows (registered trademark).

As described above, in the diagnosis system 1 according to the present embodiment, using the terminal device capable of handwriting input used in daily life, the handwriting feature amount is acquired from the user's handwriting information, and the user's usual handwriting feature amount database is obtained. Build up. The diagnosis system 1 according to the present embodiment can determine whether or not the user is in a stress state by using the amount of deviation between the acquired handwriting feature value and the usual handwriting feature value stored in a database.
[Second Embodiment]
The determination of the stress state can be performed as follows, for example. In the second embodiment, depression based on stress is further determined based on the determined stress state.

<Outline of risk value calculation of this embodiment>
FIG. 10 is an example of a diagram for explaining an outline of risk value calculation. FIG. 10A shows an example of a method for estimating the correlation among handwriting feature data, biomarker data, and diagnostic data.

  The risk value calculation device 160 operating as a diagnostic device acquires handwriting feature data of depressed patients and healthy persons, biomarker data of depressed patients and healthy persons, and diagnostic data of depressed patients and healthy persons. The handwriting feature data is the same as that in the first embodiment.

  A depressed patient is a patient who has been diagnosed as having depression by a doctor, and a healthy person is a patient who has been diagnosed as not having depression or who has no fear of depression. The diagnosis data is a diagnosis result of a doctor based on “ICD-10” of WHO (World Health Organization) or “DSM-5” of the American Psychiatric Association regarding whether or not there is depression.

  A biomarker refers to a substance such as a protein that is measured in a body fluid by reflecting the existence and progression of a certain disease in its concentration. In general, a biomarker is an indicator of a specific medical condition or life state. The doctor interrogates the patient to obtain information on the patient's symptoms and determines whether or not the patient is depressed. However, the report of symptoms in the interview is based on the subjectivity of the patient, and the patient may report his / her symptoms too much or too little, and the diagnosis by the doctor may not always be appropriate. Thus, accurate diagnosis of depression (diagnosing non-depressed patients as not being depressed and diagnosing depressed patients as being depressed) is also difficult for doctors.

  Therefore, a method using a biomarker is known as a method for diagnosing depression without using such subjective judgments of doctors and patients. Biomarkers are expected to be an objective and quantitative indicator of depression. However, diagnosis using a biomarker is in the process of development, and the reliability of diagnosis using only a biomarker is not sufficient. Based on this assumption, the risk value calculation device 160 estimates the biomarker data and diagnosis data of the person to be diagnosed as follows.

  A. The risk value calculation device 160 obtains the correlation between the handwriting feature data of the depressed patient and the healthy person and the biomarker data of the depressed patient and the healthy person.

  B. Next, the risk value calculation device 160 obtains a correlation between the biomarker data of the depressed patient and the healthy person and the diagnosis data of the depressed patient and the healthy person.

  When these correlations are obtained, the biomarker data of the diagnostic subject can be estimated from the handwriting feature amount data of the diagnostic subject, and the diagnostic data of the diagnostic subject can be estimated from this biomarker data. Even if a doctor's diagnosis is subjective, if there is statistically enough diagnosis data of a depression patient or a healthy person, the diagnosis data is highly reliable. Therefore, by collecting the handwriting feature amount of the diagnosis target person, the diagnosis result of the doctor is estimated without the diagnosis target person going to the medical institution, and early diagnosis becomes possible.

  Further, in this embodiment, an objective and quantitative index called biomarker data is estimated, but it is difficult to determine whether or not a diagnosis target person is depressed only with biomarker data. However, if an objective and quantitative index of biomarker data is estimated together with diagnostic data, the adverse effects of subjective diagnosis can be suppressed, and the meaning of biomarker data by diagnostic data becomes possible (for example, depression and When diagnosed, the estimate of a particular biomarker is high). For this reason, the person to be diagnosed or the related person confirms the biomarker data and modifies the living behavior to improve the biomarker data. Therefore, the biomarker data is an objective index (motivation) for improving the living behavior. Can provide.

  FIG. 10B shows another example of a method for estimating the correlation between handwriting feature data and biomarker data or diagnostic data. As shown in FIG. 10B, the risk value calculation device 160 may estimate biomarker data and diagnostic data from handwriting feature data.

  A. The risk value calculation device 160 obtains the correlation between the handwriting feature data of the depressed patient and the healthy person and the biomarker data of the depressed patient and the healthy person.

  C. Next, the risk value calculation device 160 obtains a correlation between the handwriting feature data of the depressed patient and the healthy person and the diagnostic data of the depressed patient and the healthy person.

  When these correlations are obtained, the biomarker data of the diagnosis subject and the diagnosis data of the diagnosis subject can be estimated from the handwriting feature value data of the diagnosis subject. Therefore, early diagnosis is possible as in the case of FIG. In addition, since the diagnostic data is estimated directly from the handwriting feature data, it can be expected that the reliability of the estimated value of the diagnostic data is improved.

<Terminology>
Living behavior is behavior related to daily survival and activities. Or it can be said to be human behavior that is easily correlated with mental illness. Living behavior data is data in which living behavior is digitized. Hereinafter, handwriting feature amount data, which is an example of living behavior data, will be described as an example.

  The already diagnosed person is a person who has been diagnosed with a mental illness from a doctor or a person who has the ability to conform to a doctor. Diagnosis means that a judgment is made that the risk of mental illness is high or low according to predetermined diagnostic criteria. Alternatively, the degree of mental illness may be determined in multiple stages. Diagnostic data is data in which the contents of diagnosis are digitized. The physical information is a substance that is detected from the blood, saliva, exhalation, tears, sweat, etc. of the subject to be diagnosed. Or it can be said to be a substance in the body that is easily correlated with mental illness. Alternatively, information that can be quantified, such as blood pressure, brain waves, and heart rate, may be included.

  The measured value of the physical information is a data obtained by converting the measured value of the physical information. Hereinafter, the term biomarker will be described as an example of physical information, and the term biomarker data will be described as an example of a measurement value of physical information.

  Mental illness means that the human mind is in poor condition or mental health is impaired. In this embodiment, an early diagnosis method for depression will be described as one of mental disorders. Mental diseases include schizophrenia, gonorrhea, neurosis, personality disorder, eating disorder, psychosomatic disorder, and the like. The diagnostic method of this embodiment is also suitable for these. In addition, since there are many unexplained parts of mental illness and the definition and diagnostic criteria for mental illness are not unified, the disease name may change depending on the classification of mental illness even for the same symptoms.

<System configuration example>
FIG. 11 is an example of an overall configuration diagram of the diagnostic system. The diagnostic system 100 of FIG. 11 includes a diagnostic data management terminal 110, a biomarker data registration terminal 111, a handwriting feature amount data registration terminal 112, a risk value calculation terminal 150, a diagnosis target person registration terminal 113, a recommended information registration terminal 114, and a database management. A device 130, a history management device 140, and a risk value calculation device 160. When these are not distinguished, they are called “terminals or devices”. Each of these terminals or devices is connected to be communicable by a network N. There is no point in being divided into the right side and the left side of the network N. Also, the difference in the names of terminals and devices does not mean a difference in hardware configuration. The terminal provides a function of a user interface for inputting information or outputting information, and the apparatus provides a function of mainly processing the input information. However, the terminal may be referred to as a device, and the device may be referred to as a terminal.

  The network N is constructed by a LAN constructed in a facility where each terminal or device is installed, a provider network of a provider that connects the LAN to the Internet, a line provided by a circuit provider, and the like. In the case of having a plurality of LANs, the network N is called a WAN (Wide Area Network) or the Internet.

  The network N may be constructed by either wired or wireless, and wired and wireless may be combined. In addition, when the terminal and the apparatus are directly connected to the public line network, they can be connected to the provider network without going through the LAN. Each of these terminals or devices may have a function as an information processing device. For example, each terminal or device is called a server or a PC. Each terminal or device may be a portable portable terminal.

  The mobile terminal is a smartphone, a mobile phone, a tablet terminal, a game machine, a PDA (Personal Digital Assistant), a digital camera, a wearable PC, a notebook PC, a game machine, or the like, but is not limited thereto.

  Also, office equipment may be used as each terminal or device. Office equipment is mainly equipment used in offices, but is not limited to office use. For example, there are an image forming apparatus (printer, MFP (Multi-Function Peripheral), MFP, copier, copier, etc.), a fax apparatus, a scanner apparatus, and a copier. Further, a projector, a HUD (Head Up Display) device, an electronic blackboard, a digital signage, or the like may be used.

  Also, cloud computing may be applied to one or more of each terminal or device. The cloud is a term used when a specific hardware resource is not intended. In cloud computing, hardware does not need to be housed in a single case or provided as a single device, and hardware resources are dynamically connected and disconnected according to the load. Consists of. In cloud computing, multiple server functions are built in a virtualized environment in one information processing device, or one server function is built across multiple information processing devices. Also good.

  Next, each terminal or device will be described. The diagnostic data management terminal 110 is a terminal for a diagnostic service provider to register diagnostic data of a depressed patient and a healthy person. The diagnostic service provider is a person related to the diagnostic system 100 such as an operator, administrator, provider, or person in charge of the diagnostic system 100 described in the present embodiment. Medical personnel may be included. Further, the term diagnostic service provider used below does not mean the same person.

  The biomarker data registration terminal 111 is a terminal for a diagnostic service provider to register biomarker data of a depressed patient and a healthy person in the database management apparatus 130. The handwriting feature data registration terminal 112 is a terminal for the diagnostic service provider to register handwriting feature data of depressed patients and healthy persons in the database management device 130.

  Although the handwriting feature value data has been described above, the handwriting feature value data is collected by a handwritten input terminal carried by a depressed patient and a healthy person, and the handwriting feature value data is stored in the handwriting feature value data registration terminal 112 or the database management device 130 wirelessly or by wire. You may send it.

  The diagnosis target person registration terminal 113 is a terminal for the diagnosis service provider to register information related to the diagnosis target person. The person to be diagnosed is a person who receives a diagnostic service provided by the diagnostic system 100, that is, a person who is determined from handwriting feature amount data that there is depression or its tendency. For example, the person to be diagnosed is an employee of a company that has introduced the diagnostic system 100, a customer who visits a shopping mall, or a resident who belongs to a public organization. Information relating to the diagnosis subject is referred to as “diagnosis subject data”. An example of the diagnosis target person data will be described later. For example, it is information for specifying the diagnosis target person, contact information, and the like.

  The recommended information registration terminal 114 is a terminal for the diagnostic service provider to register recommended information for the diagnosis target person. The recommended information is life behavior or various information recommended for the diagnosis subject in order to reduce the risk of depression of the diagnosis subject. For example, it refers to advice that a person to be diagnosed should keep in mind in daily life, food to be taken, effective exercise, and the like. In this case, the diagnostic service provider is more preferably a medical staff.

  The database management device 130 is a device that stores and manages registered diagnostic data, biomarker data, handwriting feature value data, diagnosis target person data, and recommended information in a database. In addition, the database management device 130 searches the database for diagnostic data, biomarker data, handwriting feature data, diagnosis target data, and recommended information in response to requests from the risk value calculation terminal 150 and the risk value calculation device 160. It is a device provided to the request source.

  The risk value calculation terminal 150 is a terminal on which a diagnosis target person or a related person inputs handwritten feature amount data and displays a risk value of depression of the diagnosis target person. The risk value is information indicating how easily the estimated diagnosis data is likely to become depressed, provided in a form that is easy to understand to the person to be diagnosed or a person concerned.

  For example, the risk value of depression is depression / not depression, prone / not likely to be depressed, risk level of multi-stage depression such as A to E and 1 to 10, between 0 and 100 points The degree of risk of depression indicated by the numerical value of. The risk value is calculated from the diagnosis data, but the calculation method may be determined appropriately in consideration of the person to be diagnosed or its related parties. Note that the handwriting feature amount data of the diagnosis subject may be input from the risk value calculation terminal 150.

  In response to a request from the risk value calculation terminal 150, the risk value calculation device 160 calculates an estimated value of biomarker data, an estimated value of diagnostic data, and a risk value, and provides the risk value calculation terminal 150 with the estimated value It is.

  In response to a request from the risk value calculation terminal 150, the history management device 140 registers the handwriting feature amount data of the diagnosis target person in the database management apparatus 130, or the diagnosis target person used by the risk value calculation apparatus 160 for calculation. This is a device for registering handwriting feature data. The terminal or device of the diagnostic system 100 is mainly used by a diagnostic service provider. For example, the diagnostic system 100 includes a Web server prepared by a diagnostic service provider.

  The diagnostic data management terminal 110, the biomarker data registration terminal 111, the handwriting feature value data registration terminal 112, the diagnosis target person registration terminal 113, and the recommended information registration terminal 114 are accessed by the diagnostic service provider by accessing the Web server. Register. The Web server may be an FTP server or the like, and the communication protocol is not limited. The database server 130 may also serve as the Web server. The database management apparatus 130 stores various data transmitted from the Web server in the database.

  The risk value calculation terminal 150 is arranged in a place where the person to be diagnosed or the person concerned can operate. This is because a person to be diagnosed or a related person often operates the risk value calculation terminal 150. When the risk value calculation terminal 150 accesses the Web server and requests calculation of the risk value, the Web server transmits the risk value calculated by the risk value calculation device 160 to the risk value calculation terminal 150.

  The history management device 140 is classified into two types: a case where it is placed under the control of a diagnostic service provider, and a case where it is placed in a place where a diagnosis target person or a person concerned can use it. When the diagnosis target person or the related person entrusts the management of the diagnosis target person data to the diagnostic service provider, the diagnosis target provider is arranged at a location accessible by the diagnostic service provider. Since the diagnosis target person data includes privacy, when the diagnosis target person or a related person manages the data, the diagnosis target person data or the related person is arranged in a range that can be accessed.

<About hardware configuration>
The hardware configuration is the same as in FIG. As shown in FIG. 2, the terminal or device has a function as an information processing device. The hardware configuration in FIG. 2 is an example, and a hardware configuration such as a smartphone may be employed.

  For example, in the case of the risk value calculation terminal 150, there may be a use case in which a diagnosis target person or a related person accesses the risk value calculation device 160 and displays the calculated risk value using application software that operates on a smartphone.

<About functions of diagnostic system 100>
FIG. 12 is an example of a functional block diagram illustrating functions of a terminal or a device included in the diagnostic system 100.

<< Diagnostic data management terminal 110, biomarker data registration terminal 111, handwriting feature value data registration terminal 112, diagnosis target person registration terminal 113, recommended information registration terminal 114 >>
The diagnostic data management terminal 110, biomarker data registration terminal 111, handwriting feature value data registration terminal 112, diagnosis target person registration terminal 113, and recommended information registration terminal 114 are a transmission / reception unit 121, an operation reception unit 122, a display control unit 123, And a registration request unit 124. These functional units are functions or means realized by any of the constituent elements shown in FIG. 2 operating according to instructions from the CPU 501 according to the program 504p expanded from the auxiliary storage device 504 to the RAM 503. It is.

  The transmission / reception unit 121 is realized by the CPU 501 illustrated in FIG. 2 executing the program 504p, controlling the network I / F 507, and the like, and mainly transmits / receives various data to / from the database management apparatus 130 via the network N. .

  The operation receiving unit 122 is realized by the CPU 501 illustrated in FIG. 2 executing the program 504p, controlling the input unit 505, and the like, and receives operations and information input to the terminal.

  The display control unit 123 is realized by the CPU 501 illustrated in FIG. 2 executing the program 504p and controlling the display I / F 506, and displays various screens on the display 510. For example, the display control unit 123 interprets HTML data for displaying a data registration screen and a script language, and displays a Web page. Alternatively, when dedicated application software is operating on these terminals, the screen components are arranged at predetermined positions and a data registration screen is displayed.

  The registration request unit 124 is realized by the CPU 501 illustrated in FIG. 2 executing the program 504p or the like, and requests the database management apparatus 130 to register various data registered by the diagnostic service provider.

<< Database management device 130 >>
The database management apparatus 130 includes a transmission / reception unit 131, a diagnostic data management unit 132, a biomarker data management unit 133, a handwriting feature value data management unit 134, a diagnosis target person data management unit 135, and a recommended information data management unit 136. These functional units are functions or means realized by any of the constituent elements shown in FIG. 2 operating according to instructions from the CPU 501 according to the program 504p expanded from the auxiliary storage device 504 to the RAM 503. It is.

  The database management device 130 is realized by the RAM 503, the auxiliary storage device 504, and the like, and includes a storage unit 139 that stores various types of information. The storage unit 139 in FIG. 12 stores a diagnosis database 3001, a biomarker database 3002, a handwriting feature amount database 3003, a diagnosis target person database 3004, and a recommended information database 3005. Below, these databases are demonstrated.

表４は、診断データベース３００１に登録されているデータをテーブル状に示す。診断データベース３００１には、うつ病患者及び健常者の診断結果の情報が格納される。診断データベース３００１は、表形式のデータベースであり、診断結果ごとのレコードの集合を有する。

Table 4 shows the data registered in the diagnostic database 3001 in a table form. The diagnosis database 3001 stores information on diagnosis results of depressed patients and healthy individuals. The diagnosis database 3001 is a tabular database and has a set of records for each diagnosis result.

  One record has six attributes (field, item) including diagnosis ID (Identification), subject ID, diagnosis type, diagnosis time, diagnosis result, and total diagnosis result. The diagnosis ID is information for specifying one diagnosis for a certain depression patient or healthy person. Alternatively, the diagnosis may be referred to as identification information for uniquely identifying the diagnosis. ID refers to a combination of a name, a code, a character string, a numerical value, and the like used to uniquely distinguish a specific target from a plurality of targets. The same applies to the following IDs.

  The subject ID is information for identifying the diagnosed depression patient or healthy person or information for uniquely identifying the subject. The diagnosis type represents a diagnosis item that is a basis for diagnosis among a plurality of diagnosis items. The type of diagnosis is determined by “International Classification of Diseases” of WHO (World Health Organization) “ICD-10” or “DSM-5” of the American Psychiatric Association. Table 5 shows the diagnosis types of DSM-5.

  The diagnosis time represents the diagnosis time for each diagnosis type (including date). The diagnosis result represents the result of diagnosis for each diagnosis type as a value. The value is “1 or 0”, for example, corresponding to whether or not the value is applicable. A total diagnosis result shows whether it was diagnosed as a depression patient or a healthy person by one diagnosis. As described above, the diagnosis type represents the diagnosis item, and the diagnosis result represents whether or not the diagnosis item corresponds to this diagnosis item.

  For example, in the diagnosis type of No. 1 in Table 5, the diagnosis result indicates whether the depressed patient or the healthy person responded with “No or No” to “depressed mood lasting almost all day”. In addition, the doctor determines the total diagnosis result by combining the answers for each diagnosis type.

表５は、うつ病の診断基準の１つである「ＤＳＭ−５」の診断基準と診断結果種別を示す。ＤＳＭ−５では、各診断結果種別は、包含事項、除外事項、修飾事項及び選択事項に区分されている。診断データベース３００１の診断基準は、包含事項、除外事項、修飾事項又は選択事項の各要素でどのような診断が行われるかを表し、診断データベース３００１の診断結果種別は、各要素に対する診断結果を値で表す。ＤＳＭ−５には３４個の診断種別があるので、１回の診断で、一人のうつ病患者又は健常者に対し３４個のレコードが診断データベース３００１に生成される。例えば、１つ目のレコードの診断種別は「DSM-5-a-i」に相当し、該レコードの診断結果の値は、「ほとんど１日中続く抑うつ気分」に該当するか否かを数値（１，０）で表す。なお、ＤＳＭ−５の詳細については「高橋 三郎著、「ＤＳＭ−５診断面接ポケットマニュアル」、医学書院、2015年01月」を参照されたい。ただし、うつ病の診断基準は研究途上であり、診断の種別が増えたり減ったりすることがあり得る。

Table 5 shows the diagnostic criteria and diagnostic result type of “DSM-5”, which is one of the diagnostic criteria for depression. In DSM-5, each diagnosis result type is divided into inclusion items, exclusion items, modification items, and selection items. The diagnosis criteria of the diagnosis database 3001 indicate what kind of diagnosis is performed for each element of inclusion items, exclusion items, modification items, or selection items, and the diagnosis result type of the diagnosis database 3001 indicates the diagnosis result for each element. Represented by Since there are 34 diagnosis types in DSM-5, 34 records are generated in the diagnosis database 3001 for one depressed patient or healthy person in one diagnosis. For example, the diagnosis type of the first record corresponds to “DSM-5-ai”, and the value of the diagnosis result of the record corresponds to “depressed mood almost all day” as a numerical value (1 , 0). Please refer to “Saburo Takahashi,“ DSM-5 Diagnostic Interview Pocket Manual ”, Medical School, January 2015” for details of DSM-5. However, the diagnostic criteria for depression are still under study, and the type of diagnosis may increase or decrease.

  Table 5 shows DSM-5, but it can be suitably applied to this embodiment even if it is revised and diagnostic criteria such as DSM-6 are created in the future.

表６は、バイオマーカデータベース３００２に登録されているデータをテーブル状に示す。バイオマーカデータには、うつ病患者及び健常者のバイオマーカの測定値が格納される。バイオマーカデータベース３００２は、表形式のデータベースであり、１つのバイオマーカごとに１つのレコードを有する。

Table 6 shows data registered in the biomarker database 3002 in a table form. In the biomarker data, measured values of biomarkers of depressed patients and healthy individuals are stored. The biomarker database 3002 is a tabular database, and has one record for each biomarker.

  Each record has five attributes: diagnosis ID, subject ID, biomarker type, measurement time, and measurement value. The diagnosis ID indicates a diagnosis corresponding to the biomarker. For example, a depression patient or a healthy person undergoes a biomarker test together with a doctor's diagnosis. The diagnosis at this time and the measured value of the biomarker are associated by the diagnosis ID.

  The subject ID is the same as the subject ID in Table 4. The biomarker type represents a plurality of types of biomarkers. An example of a biomarker is shown in Table 7. The measurement time represents the time when the biomarker was measured (may include the date). The measured value represents the measured value of the biomarker. Thus, the measurement value of one biomarker type is associated with the diagnosis ID.

表７は、バイオマーカ種別の一例を示す。うつ病の原因の１つとしてストレスがあることが知られている。うつ病患者又は健常者が受けているストレスの大きさに相関するとされるバイオマーカ（ストレスマーカ）として表７のような生化学物質がある。表７では１５個のバイオマーカが挙げられている。ただし、うつ病に有効なバイオマーカは研究途上であり、これらのバイオマーカは今後増えることも減ることもあり得る。

Table 7 shows an example of biomarker types. It is known that stress is one of the causes of depression. There are biochemical substances as shown in Table 7 as biomarkers (stress markers) that are correlated with the magnitude of stress experienced by depressed patients or healthy individuals. Table 7 lists 15 biomarkers. However, effective biomarkers for depression are still under study, and these biomarkers may increase or decrease in the future.

  When the stress marker in Table 7 is adopted as a biomarker for depression, 15 records are registered in the biomarker database 3002 by one test. The biomarker type in the biomarker database 3002 represents the biochemical substances of these stress markers, and the measurement value of the biomarker data represents the measurement value of the biochemical substance. For biomarkers, see "Masaki Yamaguchi," Measure stress with saliva markers ", [online], [Search July 15, 2016], Internet <URL: http: //www.banyu-zaidan.or. Refer to "jp / banyu_oldsite / symp / about / info / pdf / 3-2_066_070.pdf>".

表８は筆跡特徴量データベース３００３に登録されているデータをテーブル状に示している。筆跡特徴量データベース３００３には、うつ病患者及び健常者の筆跡特徴量データが格納される。筆跡特徴量データベース３００３は表形式のデータベースであり、筆跡特徴量データごとに１つのレコードが生成される。

Table 8 shows the data registered in the handwriting feature database 3003 in a table form. The handwriting feature amount database 3003 stores handwriting feature amount data of depressed patients and healthy persons. The handwriting feature amount database 3003 is a tabular database, and one record is generated for each handwriting feature amount data.

  One record has five attributes: diagnosis ID, subject ID, handwriting feature type, measurement time, and measurement value. The diagnosis ID associates the handwriting feature quantity with the diagnosis result. The handwriting feature value accumulated until the doctor's diagnosis is received is associated with one diagnosis ID.

  The subject ID is the same as the diagnosis database 3001. The handwriting feature amount type represents a type of measurement value such as the character size or the maximum writing pressure value described above. The measured value represents a specific value (character size or maximum writing pressure value) of the handwriting feature amount type.

表９は、診断対象者データベース３００４に登録されているデータをテーブル状に示している。診断対象者データベース３００４は診断対象者に関する情報が格納される。診断対象者データベース３００４は表形式のデータベースであり、登録された診断対象者ごとに１つのレコードを有する。各レコードは、診断対象者ＩＤ、及び連絡先の２つの属性を有する。

Table 9 shows the data registered in the diagnosis subject database 3004 in a table form. The diagnosis target person database 3004 stores information related to the diagnosis target person. The diagnosis subject database 3004 is a tabular database, and has one record for each registered diagnosis subject. Each record has two attributes of a diagnosis target person ID and a contact address.

  The diagnosis subject ID is information for identifying the registered diagnosis subject or information for uniquely identifying the subject. The contact address represents a communication method and a destination for transmitting information to the person to be diagnosed, and stores, for example, an e-mail address, a telephone number, an address, and the like.

表１０は推奨情報データベース３００５に登録されているデータをテーブル状に示している。推奨情報データベース３００５には診断対象者に対する推奨情報が格納される。推奨情報データベース３００５は表形式のデータベースであり、診断対象者に適切な推奨情報がレコードごとに格納されている。

Table 10 shows data registered in the recommended information database 3005 in a table form. The recommended information database 3005 stores recommended information for the diagnosis subject. The recommended information database 3005 is a database in a tabular format, and recommended information appropriate for the diagnosis target person is stored for each record.

  Each record has three attributes: type classification, type, and recommended information. The classification is either BIOMARKER (biomarker) or DIAGNOSTIC (diagnosis). The type is a biomarker type or a diagnostic type. The recommended information represents recommended information for the diagnosis target person according to this type. For example, if the classification is BIOMARKER, it is a related blood test (attention regarding blood components), and if it is DIAGNOSTIC, it is a related food (food to be consumed).

  In addition, the recommended information can include a wide variety of information. For example, it may be information useful for improving depression, such as specific sports such as yoga, sleep time to be satisfied, specific books, movies, music, and proposals for seminars on human relations.

(Function of database management device 130)
The transmission / reception unit 131 is realized by the CPU 501 illustrated in FIG. 2 executing the program 504 p and controlling the network I / F 507, and transmits and receives various data via the network N.

  The diagnostic data management unit 132 is realized by the CPU 501 illustrated in FIG. 2 executing the program 504p. The diagnostic data management unit 132 accepts registration of diagnostic data from the diagnostic data management terminal 110 and registers it in the diagnostic database 3001. In response to a request from the risk value calculation device 160, the diagnostic data is read from the diagnosis database 3001 and transmitted to the risk value calculation device 160.

  The biomarker data management unit 133 is realized by the CPU 501 illustrated in FIG. 2 executing the program 504p or the like, receives registration of biomarker data from the biomarker data registration terminal 111, and registers it in the biomarker database 3002. In response to a request from the risk value calculation device 160, the biomarker data management unit 133 reads biomarker data from the biomarker database 3002 and transmits it to the risk value calculation device 160.

  The handwriting feature value data management unit 134 is realized by the CPU 501 illustrated in FIG. 2 executing the program 504p, and the like, accepts registration of handwriting feature value data from the handwriting feature value data registration terminal 112, and stores it in the handwriting feature value database 3003. sign up. In response to a request from the risk value calculation device 160, handwriting feature value data is read from the handwriting feature value database 3003 and transmitted to the risk value calculation device 160.

  The diagnosis subject person data management unit 135 is realized by the CPU 501 shown in FIG. 2 executing the program 504p, and the like, receives registration of diagnosis subject person data from the diagnosis subject person registration terminal 113, and registers it in the diagnosis subject person database 3004. To do. Further, in response to a request from the history management apparatus 140, the diagnosis target person data is read from the diagnosis target person database 3004 and transmitted to the history management apparatus 140.

  The recommended information data management unit 136 is realized by the CPU 501 shown in FIG. 2 executing the program 504p or the like, receives registration of recommended information from the recommended information registration terminal 114, and registers it in the recommended information database 3005. Further, in response to a request from the risk value calculation terminal 150, the recommended information data management unit 136 reads the recommended information from the recommendation information database 3005 and transmits it to the risk value calculation terminal 150.

<< History management device 140 >>
The history management device 140 includes a transmission / reception unit 141, an operation reception unit 142, a display control unit 143, and a history data management unit 144. These functional units are functions or means realized by any of the constituent elements shown in FIG. 2 operating according to instructions from the CPU 501 according to the program 504p expanded from the auxiliary storage device 504 to the RAM 503. It is. The history management device 140 is realized by the RAM 503, the auxiliary storage device 504, and the like, and includes a storage unit 149 that stores various types of information. The storage unit 149 stores a history database 4001. First, the history database 4001 will be described.

表１１は履歴データベース４００１に登録されているデータをテーブル状に示す。履歴データベース４００１には診断対象者の筆跡特徴量データが格納される。

Table 11 shows data registered in the history database 4001 in a table form. The history database 4001 stores handwriting feature amount data of the diagnosis subject.

  The history database 4001 is a tabular database, and has a record for each handwriting feature amount type of the person to be diagnosed. Each record has five attributes: risk value calculation time, diagnosis target person ID, handwriting special feature type, measurement time, and measurement value. For example, the same risk value calculation time is given to the handwriting feature amount type (2 pieces × one month) input in one registration.

  The risk value calculation time is a time (including date) when the diagnosis target person or the related person calculates the risk value of depression of the diagnosis target person. The diagnosis subject person ID is the same as the diagnosis subject person database 3004. The handwriting feature amount type, measurement time, and measurement value are the same as those in the handwriting feature amount database 3003. In Table 11, the estimated risk value is not included, but may be included. However, if there is a handwriting feature amount type and a measurement value, the risk value calculation device 160 can calculate the risk value as appropriate.

  When diagnostic data and biomarker data change, the correlation described later also changes, so that different risk values may be calculated for the same handwriting feature amount. For this reason, it is preferable that the risk value calculation apparatus 160 calculates a risk value with the current correlation.

(Function of the history management device 140)
The transmission / reception unit 141, the operation reception unit 142, and the display control unit 143 may be the same as those of the diagnostic data management terminal 110 and the like. The history data management unit 144 is realized by the CPU 501 illustrated in FIG. 2 executing the program 504p. When the risk value is calculated by the risk value calculation device 160, the history data management unit 144 acquires each attribute of the history database 4001 from the risk value calculation terminal 150 and registers it in the history database 4001. Each attribute of the person to be diagnosed used for calculating the risk value is acquired from the database management apparatus 130.

<< Risk value calculation terminal 150 >>
The risk value calculation terminal 150 includes a transmission / reception unit 151, an operation reception unit 152, a display control unit 153, a risk value calculation request unit 154, and a recommended information request unit 155. Each of these functional units is a function or means realized by any one of the constituent elements shown in FIG. 2 operating according to a command from the CPU 501 according to the program 504p expanded from the auxiliary storage device 504 to the RAM 503. is there.

  The transmission / reception unit 151, the operation reception unit 152, and the display control unit 153 may be the same as the diagnostic data management terminal 110 and the like. The risk value calculation requesting unit 154 is realized by the CPU 501 shown in FIG. 2 executing the program 504p or the like, and when the person to be diagnosed or the related person requests calculation of the risk value, the risk value calculation requesting unit 154 Transmit to the calculation device 160. The display control unit 153 of the risk value calculation terminal 150 acquires the risk value calculated by the risk value calculation device 160 and displays it on the display 510.

  The recommended information request unit 155 is realized by the CPU 501 illustrated in FIG. 2 executing the program 504p, and the like, and is associated with the handwriting feature amount data of the diagnosis target person, the estimated value of the biomarker data, and the estimated value of the diagnostic data. The database management device 130 is requested for the recommended information.

<< Risk value calculation device 160 >>
The risk value calculation device 160 includes a transmission / reception unit 161, a correlation estimation unit 162, and a risk value calculation unit 163. Each of these functional units is a function or means realized by any one of the constituent elements shown in FIG. 2 operating according to a command from the CPU 501 according to the program 504p expanded from the auxiliary storage device 504 to the RAM 503. is there.

  The risk value calculation device 160 is realized by the RAM 503, the auxiliary storage device 504, and the like, and includes a storage unit 169 that stores various types of information. The storage unit 169 stores correlation data 6001. The correlation data 6001 is data representing the correlation between handwriting feature data and biomarker data, and the correlation between biomarker data and diagnostic data.

  The correlation estimation unit 162 is realized by the CPU 501 illustrated in FIG. 2 executing the program 504p and the like, and the above correlation is performed using the diagnosis data, biomarker data, and handwriting feature data of the database management apparatus 130. Is calculated.

  The risk value calculation unit 163 is realized by the CPU 501 illustrated in FIG. 2 executing the program 504p or the like, and calculates the risk value of the diagnosis target person using the correlation data 6001.

<About correlation>
<< Calculation of correlation by regression >>
Predicting the value of the target variable T from the value of the vector X of the input variable is called a regression problem. The goal of the regression problem is to predict the value of T relative to the measured value X when given training data consisting of N observed values {Xn} (n = 1,..., N) and corresponding target values {Tn}. It is to be. A simple solution is to obtain a function f that satisfies T = f (X) for the input X. The simplest linear regression model of the function f is a linear combination of input variables. For the regression problem, see “CM Bishop,“ Pattern recognition and machine learning ”, Maruzen Publishing, December 2007”.
T = a1 · x1 + a2 · x2 + a3 · x3 + a4 · x4 + ... + an · x20
The concept of this regression problem is applied to this embodiment. For convenience of explanation, the correlation of the estimation method of FIG. 10A will be described, but the same applies to the estimation method of FIG.

  Diagnostic data having each diagnosis result as an element is vector D, biomarker data having each measurement value as an element is vector B, and handwriting feature quantity data having each measurement value as an element is vector L. As shown in Table 5, the number of elements of diagnostic data is 34, as shown in Table 7, the number of elements of biomarker data is 15, and the number of elements of handwriting feature data is 2.

When considering representing the biomarker data with the handwriting feature value data, one element of the biomarker data is combined with two elements of the handwriting feature value data. A function for obtaining biomarker data of a depressed patient and a healthy person from handwriting feature data is LtoB. When each element of the handwriting feature data is represented by L1 to L2 and each element of the biomarker data is represented by B1 to B15, a linear combination of LtoB is represented as follows.
LtoB1 ＝ a1 ・ L1 ＋ a2 ・ L2
a1 to a2 are coefficients of each element of the handwriting feature data. Therefore, the handwriting feature value data and the biomarker data can be expressed by linear combination by solving the regression problem using the training data of the depression patient and the healthy person to obtain a1 to a2. Such regression problems can be solved by the least square method.

Similar calculations are performed for each element of the biomarker data.
LtoB2 = b1 / L1 + b2 / L2
LtoB3 = c1 ・ L1 ＋ c2 ・ L2
(Omitted)
LtoB15 = o1 / L1 + o2 / L2 (1)
Next, considering that diagnostic data is represented by biomarker data, it is considered that one element of diagnostic data is linked to 15 elements of biomarker data. A function for obtaining diagnosis data of a depression patient and a healthy person from biomarker data is defined as BtoD.

When each element of biomarker data is represented by B1 to B15 and each element of diagnostic data is represented by D1 to D34, a linear combination of BtoD is represented as follows. The coefficients such as a1 to a15 are different from those in the equation (1).
BtoD1 = a1 · B1 + a2 · B2 + a3 · B3 + a4 · B4 + ... + a15 · B15
BtoD2 = b1, B1 + b2, B2 + b3, B3 + b4, B4 + ... + b15, B15
(Omitted)
BtoD34 = hh1 · B1 + hh2 · B2 + hh3 · B3 + hh4 · B4 + ... + hh15 · B15 (2)
Therefore, the correlation estimation unit 162 can determine the coefficient with reference to the diagnosis database 3001, the handwriting feature value database 3003, and the biomarker database 3002. The determined coefficient is stored in the storage unit 169 as correlation data 6001.

Similarly, the handwriting feature value data of the person to be diagnosed is a vector L ′, the biomarker data estimated value vector is B ′, and the diagnostic data estimated value vector is D ′. Then, the following relationship is obtained.
L'toB '
B'toD '
The risk value calculation unit 163 uses Equation (1) as L′ toB ′, substitutes the handwriting feature amount data of the diagnosis subject, and estimates biomarker data. Further, the risk value calculation unit 163 can estimate the diagnostic data by using the formula (2) as B′toD ′ and substituting it into the biomarker data of the person to be diagnosed.

  In the present embodiment, handwritten feature amount data, biomarker data, and diagnostic data of a depression patient and a healthy person are not distinguished. However, if the correlation between the depressed patient and the healthy person is obtained by the regression, the biomarker data and diagnostic data of the diagnosis subject who cannot know whether the patient is a depressed patient or a healthy person can be estimated.

  Moreover, although diagnostic data takes a value in the range of 0-1 for every attribute, the one where diagnostic data is close to 1 does not necessarily have a high risk of depression. For example, it is estimated that a diagnosis subject whose “modification / course and seriousness / severity”, which is one of the diagnostic criteria, is close to 1, has a high risk of depression, but “modification / course and seriousness / mild” Diagnosis subjects who are close to 1 have a low risk of depression. Regarding other attributes of the diagnostic criteria, whether the risk of depression is high depends on the attribute.

  For this reason, the risk value calculation unit 163 calculates the risk value by weighting the attribute of the diagnosis data. That is, the risk value is calculated by weighting and summing the estimated results of 34 attributes. The risk value is an index that is more likely to become depressed as the value increases. Moreover, diagnostic data itself may be treated as a risk value.

  In this manner, biomarker data and diagnostic data can be estimated even if the person to be diagnosed does not visit a medical institution, and early diagnosis of depression becomes possible.

  Further, the risk value calculation unit 163 can estimate the biomarker data and the diagnosis data from the handwriting feature amount data at another time of the diagnosis target person. Therefore, the person to be diagnosed or the related person can confirm the tendency of increase / decrease in the risk value, biomarker data, or diagnostic data to confirm whether he / she is likely to be depressed or is improving.

  In the present embodiment, the linear combination has been described. However, the coefficient may be obtained by regression using a polynomial of second or higher order.

<< Estimation by machine learning >>
Further, the correlation may be obtained by machine learning. FIG. 13 shows a configuration example of a neural network which is an example of a machine learning model. FIG. 13A shows a neural network that outputs biomarker data from handwriting feature data. The input layer 601 receives handwritten feature data (L1 to L2) of a depressed patient or a healthy person, and the output layer 603 outputs biomarker data (B1 to B15) of a depressed patient or a healthy person.

  In FIG. 13A, the number of nodes in the input layer is the number of types of handwriting feature data, and the number of nodes in the output layer is the number of types of biomarker data. The number of intermediate layers is 1, but the number of intermediate layers is appropriately designed.

  FIG. 13B is a neural network that outputs diagnostic data from biomarker data. Depressed patient or healthy person biomarker data (B1 to B15) is input to the input layer 601, and the depressed patient or healthy person diagnosis data (D1 to 34) is output from the output layer 603.

  In FIG. 13B, the number of nodes in the input layer is the number of types of biomarker data, and the number of nodes in the output layer is the number of diagnosis types of diagnostic data. The number of intermediate layers is 1, but the number of intermediate layers is appropriately designed.

  In FIGS. 13A and 13B, each layer is identified by a number in order from the left (input layer). The input layer is the first layer, the intermediate layer is the second layer, and the output layer is the third layer. In addition, it is not necessary to make the number of intermediate | middle layers the same in FIG. 13 (a) and (b). Further, it is not necessary to make the number of nodes in the intermediate layer (second layer) equal.

In the following, transmission of information from the input layer to the output layer of the neural network will be described. First, handwriting feature data (L1 to L2) is input to each node of the input layer. The input to the i-th node (i is an integer not less than 1 and not more than n) from the top of the first layer in the neural network is represented by a linear combination of the handwriting feature value data Li and the weight W1i. Here, X1i is an input value to the i-th node of the first layer, and W1i is a weight of the i-th node of the first layer.
X1i ＝ ΣW1i ・ Li
Further, the output Z1i of the i-th node (i is an integer not less than 1 and not more than n) in the first layer can be expressed as follows using the activation function f (X).
Z1i = f (X1i)
The activation function is a non-linear function and outputs a non-linear operation result with respect to the input. A sigmoid function is known as an activation function. The sigmoid function is characterized in that the output falls between 0 and 1 even when the absolute value of the input value is large. The slope at which the sigmoid function changes from 0 to 1 near x = 0 can be adjusted by the exponent of the natural logarithm e. The tanh function (output is −1 to +1) may be used as the activation function.

The input and output to the second layer are also calculated using linear combination as in the first layer.
X2i ＝ ΣW2i ・ Z1i
Z2i = f (X2i)
X2i is an input value to the i-th node in the second layer, and W2i is a weight of the i-th node in the second layer. Z2i is an output value to the i-th node in the second layer. The output of the third layer is calculated similarly. Estimated values of biomarker data (B1 to B15) are output from each node in the third layer.

  In machine learning, the weight Wji is learned in the learning phase. j is the layer number and i is the node number. The output value of the third layer is compared with biomarker data of a depressed patient or a healthy person. The biomarker data to be compared with the output value is called a teacher signal. Since the biomarker data of a depression patient or a healthy person is not a value obtained by a neural network but a measured value, it is preferable that the value calculated by the neural network and the measured value are comparable.

  Further, when there is some relationship or correlation between the handwriting feature data and the biomarker data, the biomarker data should be obtained from the handwriting feature data. Further, since it is known that a neural network can approximate an arbitrary function that outputs a value corresponding to an input, it can represent a correlation between handwriting feature amount data and biomarker data. The neural network learns so that the output of the third layer approaches the biomarker data, so that the biomarker data can be obtained from the handwriting feature data.

  Learning the neural network means updating the weight Wji. Specifically, the difference between the output of the third layer and the teacher signal is used as an error, and the weight from the input layer to the output layer is corrected by a method called an error back propagation method. The error back propagation method has been formulated and will not be described. The neural network and weight Wji learned in FIG. 13A correspond to the correlation data 6001.

  The learning phase ends when the weight is updated or the weight does not change with some teacher signals. In this state, when handwriting feature value data of the diagnosis subject is input to the neural network, an estimated value of the biomarker data of the diagnosis subject is output.

  Learning of the neural network in FIG. 13B is performed in the same manner. When there is some relationship or correlation between the biomarker data of a depressed patient or a healthy person and the diagnostic data, the diagnostic data can be obtained from the biomarker data. Biomarker data (B1 to B15) is input to each node in the input layer, and estimated values of diagnostic data (D1 to D34) are output from each node in the third layer.

  The neural network and weight Wji learned in FIG. 13B correspond to the correlation data 6001. Therefore, the neural network learns so that the diagnosis data of the depression patient or the healthy person is output from the biomarker data of the depression patient or the healthy person, thereby estimating the biomarker data calculated in FIG. An estimate of the diagnostic data is obtained from the value. Machine learning includes support vector machines and multi-class classification methods in addition to neural networks.

<Operation procedure>
Below, each operation | movement which the diagnostic system 100 performs is demonstrated.

<< Registering diagnostic data >>
FIG. 14 is a flowchart of an example of a procedure for registering diagnostic data.

  The diagnostic service provider operates the diagnostic data management terminal 110 and inputs diagnostic data for each diagnosis. The operation accepting unit 122 of the diagnostic data management terminal 110 accepts input of diagnostic data by the diagnostic service provider (S51). That is, the diagnosis service provider inputs diagnosis ID, subject ID, diagnosis time, and diagnosis result for each diagnosis type.

  When the diagnostic data is input and the diagnostic service provider performs an operation for registration, the operation reception unit 122 receives the operation, and the registration request unit 124 transmits the diagnostic data to the database management apparatus 130 (S52). The diagnostic data management unit 132 registers diagnostic data in the diagnostic database 3001. In this way, the diagnosis data of the depression patient, the healthy person, and the person to be diagnosed are registered.

<< Update of diagnostic data >>
FIG. 15 is a flowchart of an example of a procedure for notifying the database management apparatus that diagnostic data has been updated.

  The diagnostic service provider operates the diagnostic data management terminal 110 and inputs that the diagnostic data has been updated. The operation accepting unit 122 of the diagnostic data management terminal 110 accepts an input for updating by the diagnostic service provider (S53). At the time of reception, a diagnosis ID, a subject ID, a diagnosis time, and a diagnosis result for each diagnosis type regarding the updated diagnosis data are input.

  The registration request unit 124 transmits a request to update the diagnostic data of the database management device 130 with the updated diagnostic data to the database management device 130 (S54). The transmission / reception unit 131 receives the update request and updates the diagnostic database 3001.

  When the transmission / reception unit 131 receives the update request, the diagnosis data management unit 132 searches the diagnosis target person database 3004 (S55). Then, it notifies the contact information of the searched person to be diagnosed that the diagnostic data has been updated.

  Upon receiving the notification that the diagnostic data has been updated, the person to be diagnosed or the person concerned can operate the risk value calculation terminal 150 and estimate the risk value of depression again. Thus, the diagnostic system 100 of this embodiment can cope with the revision of the diagnostic criteria.

  Further, the risk value may be automatically calculated for the update of the diagnostic data as follows.

  When receiving the update request, the transmission / reception unit 131 of the database management apparatus 130 requests the risk value calculation apparatus 160 to calculate the correlation data 6001 (S55-2). Doctors and the like diagnose depression patients and healthy individuals based on new diagnostic criteria, and diagnostic data by doctors is obtained.

  When the correlation data 6001 is created, the transmission / reception unit 131 of the database management device 130 requests the risk value calculation device 160 to calculate the risk value (S55-3). The calculation request includes the subject ID of the person to be diagnosed.

  The risk value calculation unit 163 of the risk value calculation device 160 requests the history management device 140 for the diagnosis subject person ID specified by the subject ID, and changes the handwriting feature amount data of the diagnosis subject acquired from the history management device 140. The risk value is calculated by applying to the later correlation data 6001 (S55-4). In addition, this risk value is transmitted to the person to be diagnosed by e-mail or the like.

  The diagnostic data is updated because the diagnostic criteria for depression are revised. When the diagnostic criteria for depression are revised, diagnostic data may need to be updated. For example, in DSM-IV, when it is regarded as a bereavement reaction that is a kind of disease due to cultural circumstances, it becomes an exclusion criterion. On the other hand, in DMS-5, this exclusion criterion is deleted. In other words, the diagnosis result needs to be changed by the revision from DSM-IV to DSM-5.

  Therefore, according to the present embodiment, by updating diagnostic data (diagnosis result), it is possible to recalculate the risk value based on the past handwritten feature amount data of the person to be diagnosed. It can be calculated.

<< Register biomarker data >>
FIG. 16 is a flowchart of an example of a procedure for registering biomarker data.

  The diagnostic service provider operates the biomarker data registration terminal 111 and inputs biomarker data for each diagnosis. The operation accepting unit 122 of the biomarker data registration terminal 111 accepts input of biomarker data by the diagnostic service provider (S56). The diagnosis service provider inputs a diagnosis ID, a subject ID, a measurement time, and a measurement value for each biomarker type.

  When biomarker data is input and an operation for registration by the diagnostic service provider is performed, the operation reception unit 122 receives the operation, and the registration request unit 124 transmits the biomarker data to the database management device 130 (S57). The biomarker data management unit 133 registers the biomarker data in the biomarker database 3002. In this way, biomarker data of depressed patients and healthy individuals is registered.

<< Registering handwriting features >>
FIG. 17 is a flowchart of an example of a procedure for registering handwriting feature data.

  The diagnosis service provider operates the handwriting feature value data registration terminal 112 to input handwriting feature value data for each diagnosis. The operation accepting unit 122 of the handwriting feature data registration terminal 112 accepts input of handwriting feature data by the diagnostic service provider (S58). The diagnosis service provider inputs a diagnosis ID, a subject ID, a measurement time, and a measurement value for each life log type.

  When handwriting feature value data is input and an operation for registration by the diagnostic service provider is performed, the operation accepting unit 122 accepts the operation, and the registration requesting unit 124 transmits the handwriting feature value to the database management apparatus 130 (S59). The handwriting feature value data management unit 134 of the database management apparatus 130 registers the handwriting feature value data in the handwriting feature value database 3003. In this way, handwriting feature amount data of a depression patient, a healthy person, and a person to be diagnosed is registered.

<< Registration of diagnostic subjects >>
FIG. 18 is a flowchart of an example of a procedure for registering a person to be diagnosed.

  The diagnosis service provider operates the diagnosis object person registration terminal 113 and inputs diagnosis object person data. The operation reception unit 122 of the diagnosis target person registration terminal 113 receives input of diagnosis target person data from the diagnosis service provider (S60). That is, the diagnosis service provider inputs the diagnosis target person ID and the contact information.

  When the diagnosis subject person data is input and the diagnosis service provider performs an operation for registration, the operation reception unit 122 accepts the operation, and the registration request unit 124 transmits the diagnosis subject person data to the database management apparatus 130 (S61). . The diagnosis target person data management unit 135 of the database management apparatus 130 registers the diagnosis target person data in the diagnosis target person database 3004. In this way, the diagnosis subject person data is registered.

<< Registration of recommended information >>
FIG. 19 is a flowchart of an example of a procedure for registering recommended information.

  The diagnostic service provider operates the recommended information registration terminal 114 to input recommended information. The operation reception unit 122 of the recommended information registration terminal 114 receives input of recommended information from the diagnostic service provider (S62). That is, the recommended information is input.

  When the recommended information is input and the diagnostic service provider performs an operation for registration, the operation reception unit 122 receives the operation, and the registration request unit 124 transmits the recommended information to the database management apparatus 130 (S63). The recommended information data management unit 136 of the database management apparatus 130 registers the recommended information in the recommended information database 3005. In this way, recommended information is registered.

<< Data manipulation >>
FIG. 20 is a flowchart of an example of a procedure in which the database management device performs data operations. The database management apparatus 130 performs registration, retrieval, provision, and the like of data in the database in response to a request from the terminal.

  The database management device 130 receives data operation requests from the diagnostic data management terminal 110, the biomarker data registration terminal 111, the handwriting feature value data registration terminal 112, the diagnosis target person registration terminal 113, the recommended information registration terminal 114, and the risk value calculation terminal 150. Receive (S64). When requesting, a request type and request data are specified.

  The database management device 130 determines the request type (S65). For example, the request type includes registration, search, and update. The database management apparatus 130 performs data operation according to the request type (S66).

<< Risk value calculation >>
FIG. 21A is a flowchart of an example of a procedure in which the risk value calculation device calculates a risk value. FIG. 21B is a sequence diagram illustrating an example of the overall operation of the diagnostic system. Although the substantial contents are the same, here, description will be made based on a sequence diagram.

  S67: The person to be diagnosed or the related person operates the risk value calculation terminal 150 to perform an operation for calculating the risk value. The operation reception unit 152 of the risk value calculation terminal 150 receives a risk value calculation request.

  At the time of reception, the person to be diagnosed or the related person inputs handwriting feature amount data. Of the input handwriting feature value data, an arbitrary diagnosis target person ID may be specified, and a period may be specified.

  S68: The risk value calculation terminal 150 requests the history management device 140 to register the designated handwriting feature amount data and the diagnosis subject person ID in the history database 4001 (S68). The history management device 140 acquires the designated handwriting feature data from the database management device 130 and registers it in the history database 4001.

  S69-1: The transmission / reception unit 141 of the history management device 140 transmits the handwriting feature amount data and the diagnosis target person ID to the risk value calculation device 160 together with the risk value calculation request.

  S69-2: The transmission / reception unit 161 of the risk value calculation device 160 receives the handwriting feature amount data and the diagnosis target person ID from the database management device 130, and the risk value calculation unit 163 applies the correlation data 6001 to the handwriting feature amount data. Then, the estimated value of the biomarker data is calculated from the handwriting feature data.

  S70: Next, the risk value calculation unit 163 calculates the estimated value of the diagnostic data by applying the correlation data 6001 to the estimated value of the biomarker data. In the estimation method of FIG. 10B, diagnostic data is estimated from life log data.

  S71-1: The transmission / reception unit 161 of the risk value calculation device 160 transmits the estimated value of biomarker data and the estimated value of diagnostic data to the risk value calculation terminal 150.

  S71-2: The transmission / reception unit 151 of the risk value calculation terminal 150 receives the estimated value of the biomarker data and the estimated value of the diagnostic data, and the display control unit 153 displays the estimated value on the display 510.

  As described above, since the biomarker data is estimated, the person to be diagnosed or the related person can grasp the quantitative and objective index. Moreover, since the estimated value of diagnostic data is obtained, diagnostic data that a doctor or the like makes a judgment based on diagnostic criteria is obtained. Therefore, it is possible to make an early diagnosis of depression based on the diagnostic criteria of a doctor with information obtained on a daily basis, that is, handwritten feature amounts.

  S72-1: Next, the recommended information request unit 155 requests the database management device 130 for recommended information associated with the specified handwriting feature value data, the estimated value of biomarker data, and the estimated value of diagnostic data. To do.

  S72-2: The recommended information data management unit 136 of the database management apparatus 130 searches the recommended information database 3005 for relevant recommended information, and transmits the recommended information of the search result to the risk value calculation terminal 150.

  S72-3: The display control unit 153 of the risk value calculation terminal 150 displays the recommended information on the display 510.

  Note that a URL (Uniform Resource Locator) may be registered as recommended information in the recommended information database 3005. When the diagnosis target person or the related person clicks the URL, the risk value calculation terminal 150 accesses an EC (Electronic Commerce) site specified by the URL. Therefore, it becomes easy to purchase food on an EC site that is effective in improving the risk value. Further, it may be a URL of a Web site that provides information effective for improving the risk value instead of the EC site.

  It is also effective to use an affiliate mechanism. The risk value calculation device 160 provides an affiliate space in the screen information described in HTML or the like for transmitting recommended information.

  The risk value calculation terminal 150 accesses an ASP (affiliate service provider) registered in advance, and displays advertisement information of an advertiser suitable for recommended information in the affiliate space. When the person to be diagnosed clicks on this space, the risk value calculation terminal 150 accesses the URL of the advertiser. At the time of access, information specifying the risk value calculation device 160 (for example, an ASP member ID) is notified to the advertiser.

  When the diagnosis target person purchases a product on the advertiser's website, the ASP is notified of information for identifying the risk value calculation device 160 and result data (what and how many have been purchased). As a result, the performance reward is paid from the ASP to the risk value calculation device 160 (that is, the diagnosis system 100).

  Thus, the operating cost of the diagnostic system 100 can be reduced by linking the recommended information with the EC site or the like.

<< Registration of history >>
FIG. 22 is a flowchart of an example of a procedure for registering a history by the history management apparatus.

  The history management device 140 receives a history registration request from the risk value calculation terminal 150 (S73). That is, the risk value calculation time, the diagnosis target person ID, the handwriting feature amount type, the measurement time, and the measurement value related to the handwriting feature amount data for which the risk value has been calculated are transmitted.

  The transmission / reception unit 141 of the history management device 140 receives the history registration request, and the history data management unit 144 registers these data in the history database 4001 (S74).

<< Time series display of risk values >>
FIG. 23 is a flowchart of an example of a procedure by which the risk value calculation terminal calculates risk values in time series. FIG. 23B is a sequence diagram illustrating an example of the overall operation of the diagnostic system. Although the substantial contents are the same, here, description will be made based on a sequence diagram.

  S75: The person to be diagnosed or the related person operates the risk value calculation terminal 150 to perform an operation for calculating a time series risk value. The operation reception unit 152 of the risk value calculation terminal 150 receives this operation (S75).

  S76-1: Next, the person to be diagnosed or the related person performs an operation of searching for handwritten feature data by designating the person to be diagnosed. The operation reception unit 152 of the risk value calculation terminal 150 receives a diagnosis target person ID and a search request. The transmission / reception unit 151 of the risk value calculation terminal 150 requests the history management device 140 to search for the history of the handwriting feature value data associated with the diagnosis subject person ID.

  S76-2: The transmission / reception unit 141 of the history management device 140 receives the search request, and the history data management unit 144 searches for the handwriting feature amount data associated with the diagnosis subject person ID, and the risk value together with the risk value calculation time. Transmit to the calculation device 160.

  S77-1: The transmission / reception unit 151 of the risk value calculation terminal 150 receives the handwriting feature amount data and the risk value calculation time, and transmits the handwriting feature amount data and the risk value calculation time to the risk value calculation device 160. Note that the history management device 140 may transmit handwriting feature amount data and risk value calculation time directly to the risk value calculation device 160.

  S77-2: The transmission / reception unit 161 of the risk value calculation device 160 receives the handwriting feature data and the risk value calculation time, and the risk value calculation unit 163 applies the correlation data 6001 to the handwriting feature data at each risk value calculation time. To do. Thereby, the estimated value of biomarker data and the estimated value of diagnostic data are calculated from handwriting feature-value data for every risk value calculation time.

  S78-1: The transmission / reception unit 161 of the risk value calculation device 160 transmits the calculated estimated value of biomarker data and the estimated value of diagnostic data to the risk value calculation terminal 150 at each risk value calculation time.

  S78-2: The transmission / reception unit 151 of the risk value calculation terminal 150 receives these, and the display control unit 153 displays the estimated value of the biomarker data and the estimated value of the diagnostic data on the display 510 for each risk value calculation time. .

  S79-1: Also, the transmission / reception unit 151 of the risk value calculation terminal 150 requests the database management apparatus 130 for recommended information related to the handwritten feature amount data, the estimated value of the biomarker data, and the estimated value of the diagnostic data. .

  S79-2: The recommended information data management unit 136 of the database management device 130 transmits the handwritten feature amount data, the estimated value of the biomarker data, and the recommended information corresponding to the estimated value of the diagnostic data to the risk value calculating terminal 150.

  S79-3: The display control unit 153 of the risk value calculation terminal 150 displays the recommended information transmitted from the database management apparatus 130 on the display 510.

  FIG. 24 is an image diagram of an example of a risk value screen displayed on the display of the risk value calculation terminal. In FIG. 24, the risk values for each month are shown as a scatter diagram. The date and time on the X axis corresponds to the risk value calculation time. As described above, the diagnosis target person or the related person can calculate the risk value using the past handwritten feature amount data and display it in time series, so that the tendency of the risk value can be easily grasped.

  Further, the risk value screen 701 in FIG. 24 includes a risk value button 702, a biomarker button 703, and a diagnostic data button 704. The risk value button 702 is a button for displaying a risk value as shown in FIG. 24, the biomarker button 703 is a button for displaying an estimated value of the biomarker, and the diagnostic data button 704 is a diagnostic data button. It is a button for displaying an estimated value. The person to be diagnosed or the related person can display the estimated value of the biomarker and the estimated value of the diagnostic data in time series. In addition, instead of a scatter diagram, numerical values of estimated values of biomarkers and estimated values of diagnostic data can also be displayed.

<Comparison of estimated values of two diagnostic data>
Although it has been described that there are two methods for estimating diagnostic data shown in FIGS. 10A and 10B, it is also effective to estimate diagnostic data using both methods. In this case, since diagnostic data can be obtained by either method, the correlation estimation unit 162 can compare the estimated values of the two diagnostic data. Therefore, if the estimated values of the two diagnostic data are almost the same, it can be confirmed that not only the reliability of the diagnostic data is high but also the reliability of the estimated value of the biomarker data is high.

  FIG. 25 is a flowchart illustrating an example of a correlation data calculation method using two estimated diagnosis data. The procedure in FIG. 25 is executed, for example, at the timing when correlation data is calculated.

  First, when correlation data is generated, the correlation estimation unit 162 estimates diagnosis data by two estimation methods using some life log data of arbitrary diagnosis subjects (S80).

  The correlation estimation unit 162 determines whether the two diagnostic data can be regarded as the same (S81). If the determination in step S81 is Yes, the correlation estimation unit 162 employs this correlation data because the correlation data used in step S80 is highly reliable.

  When the determination in step S81 is No, the correlation estimation unit 162 calculates correlation data by reducing the types of handwriting feature data and biomarker data one by one (S82).

  Next, the correlation estimation unit 162 estimates diagnosis data by two estimation methods using some handwriting feature amount data of an arbitrary diagnosis target (S83). Thereafter, the process returns to step S81, so that the types of the handwriting feature amount data and the biomarker data are decreased one by one, and highly reliable correlation data is obtained. Note that the types of handwriting feature data and biomarker data to be reduced are determined in advance. Alternatively, any one may be reduced, two diagnostic data may be compared, and the type with which the two diagnostic data are closest may be reduced.

  Therefore, it is easy to obtain an estimated value of biomarker data and an estimated value of diagnostic data with high reliability.

  As described above, the diagnostic system 100 according to the present embodiment can estimate the biomarker data of the diagnostic subject and the diagnostic data of the diagnostic subject from the handwriting feature amount data of the diagnostic subject. Therefore, since the diagnosis result of the doctor is estimated without the person to be diagnosed going to the medical institution, early diagnosis becomes possible. In addition, since an objective and quantitative index of biomarker data is estimated, adverse effects of subjective diagnosis can be suppressed, and the diagnosis target person or related parties can check the biomarker data to improve it. Since living behavior is changed, an index (motivation) for improving living behavior can be provided.

<Other application examples>
The best mode for carrying out the present invention has been described above with reference to the embodiments. However, the present invention is not limited to these embodiments, and various modifications can be made without departing from the scope of the present invention. And substitutions can be added.

  For example, biomarker data and diagnostic data may be printed in addition to being displayed, and the output form is not limited to display. In addition, the configuration example in FIG. 12 is divided according to main functions in order to facilitate understanding of processing of the terminal and the apparatus. The present invention is not limited by the way of dividing the processing unit or the name. Further, the processing of the terminal and the device can be divided into more processing units according to the processing content. Moreover, it can also divide | segment so that one process unit may contain many processes.

  In addition, it is appropriately configured whether the terminal and the device are distributed to one device or a plurality of devices. For example, all the terminals and devices may be realized as a single device. Further, the risk value calculation terminal 150 may have the risk value calculation unit 163 and the correlation data 6001. Arbitrary one terminal and apparatus may be integrated among the structures of FIG. 11 and FIG. 12, and the other terminal and apparatus may have the functions of one terminal and apparatus. Moreover, the database which the database management apparatus 130 has should just exist in the place which the diagnostic system 100 can access, and the database management apparatus 130 does not need to have.

  The handwriting information acquisition unit 21 is an example of handwriting information acquisition means in the claims. The handwriting feature value calculation unit 23 is an example of a handwriting feature value calculation unit. The stress determination unit 25 is an example of a stress determination unit. The handwriting feature amount accumulation unit 24 is an example of an accumulation unit. The correlation estimation unit 162 is an example of an estimation unit. The biomarker data is an example of a measured value of the body information of the already diagnosed person.

DESCRIPTION OF SYMBOLS 1,100 Diagnostic system 10 Liquid crystal tablet 11 Sensing part 12 Display part 20 PC (personal computer)
DESCRIPTION OF SYMBOLS 21 Handwriting information acquisition part 22 Display data generation part 23 Handwriting feature-value calculation part 24 Handwriting feature-value accumulation part 25 Stress judgment part 30 Instruction part 110 Diagnostic data management terminal 111 Biomarker data registration terminal 112 Handwriting feature-value data registration terminal 113 Diagnosis object User registration terminal 114 recommended information registration terminal 130 database management device 140 history management device 150 risk value calculation terminal 160 risk value calculation device N network

Sadao Kato, "Fundamental study of writing pressure characteristics", Proceedings of Autumn Meeting of Japan Management Engineering Society, Volume 1979, Pages 163-164 (1979) Giora Keinan, "Can Stress be Measured by Handwriting Analysis? The Effectiveness of the Analytic Method", Applied Psychology, Volume 42, Issue 2, Pages 153-170 (1993)

Claims (7)

Handwriting information acquisition means for acquiring user handwriting information;
A handwriting feature amount calculating means for calculating the handwriting feature amount of the user from the handwriting information and storing the user's handwriting feature amount in the storage unit;
By the amount of deviation between the first handwriting feature amount calculated by the handwriting feature amount calculating means and the second handwriting feature amount based on one or more past handwriting feature amounts stored in the storage unit, Stress determination means for determining whether or not the user is in a stress state;
A diagnostic device. Using the handwriting feature amount of the already diagnosed person who has been diagnosed with a mental illness, the measured value of the physical information of the already diagnosed person, and correlation data representing the correlation between the diagnosis data regarding the diagnosis, the handwriting feature Estimating means for estimating at least one of a measured value of the user's physical information or diagnostic data related to the diagnosis from the first handwriting feature value calculated by the quantity calculating means;
The diagnostic apparatus according to claim 1, further comprising: The stress determination means includes a second size based on a character size, which is the first handwriting feature amount calculated by the handwriting feature amount calculation means, and one or more past handwriting feature amounts accumulated in the accumulation unit. The diagnosis apparatus according to claim 1, wherein it is determined whether or not the user is in a stress state based on an amount of deviation from an average value of a character size that is a handwriting feature amount. The handwriting feature value calculating means groups the input strokes included in the handwriting information in character units, and calculates the character size from the position coordinates of the input strokes grouped in character units. 3. The diagnostic device according to 3. The stress determination means uses the maximum writing pressure value of the character, which is the first handwriting feature quantity calculated by the handwriting feature quantity calculation means, and one or more past handwriting feature quantities accumulated in the accumulation unit. 3. The diagnosis according to claim 1, wherein it is determined whether or not the user is in a stress state based on a deviation amount from an average value of the maximum writing pressure value of the character that is the second handwriting feature value based on the diagnosis value. apparatus. Computer
Handwriting information acquisition means for acquiring user handwriting information,
A handwriting feature amount calculating means for calculating the handwriting feature amount of the user from the handwriting information and storing the handwriting feature amount of the user in the storage unit;
By the amount of deviation between the first handwriting feature amount calculated by the handwriting feature amount calculating means and the second handwriting feature amount based on one or more past handwriting feature amounts stored in the storage unit, Stress determination means for determining whether the user is in a stress state,
Program to function as. A diagnostic system comprising: an application installed in a terminal device capable of handwriting input; and a diagnostic device for diagnosing a stress of a user who has input characters by handwriting on the terminal device,
Handwriting information acquisition means for acquiring user handwriting information from the application;
A handwriting feature amount calculating means for calculating the handwriting feature amount of the user from the handwriting information and storing the user's handwriting feature amount in the storage unit;
By the amount of deviation between the first handwriting feature amount calculated by the handwriting feature amount calculating means and the second handwriting feature amount based on one or more past handwriting feature amounts stored in the storage unit, Stress determination means for determining whether or not the user is in a stress state;
A diagnostic system.

Images