Information processing device

Abstract

An information processing device (10) according to an embodiment includes a determination unit (102) that generates a determination result in which the mental state of a subject is determined by inputting pulsation information relating to the heartbeat of the subject and body motion information relating to body motion of the subject to a trained model. The trained model has been trained using, for a plurality of providers that provide training data, pulsation information relating to heartbeat of the providers, body motion information relating to body motion of the providers, and index information relating to specific mental tendencies of the providers.

Description

Information processing device

[0001] FIELD An embodiment of the present invention relates to an information processing device.

[0002] In the fields of psychological counseling and psychiatry, various assessment methods are used to understand a subject's mental state. For example, well-known methods for assessing depression include the Clinical Global Impression of Improvement (CGI-I), Hamilton Rating Scale for Depression-17 item score (HAMD-17), a Montgomery-Asberg Depression Rating Scale score (MADRS), Beck Depression Inventory (BDI), Patient Health Questionnaire-9 (PHQ-9), and State Trait Anxiety Inventory (STAI). These assessment methods primarily consist of question-based items. For example, doctors and counselors interview subjects and score their level of depression by answering questions based on observed symptoms and conversations. Other assessment methods score the level of depression by having subjects answer questions themselves.

[0003] JP 2018-153244 A

[0004] However, interviews with doctors or counselors cannot be conducted unless the subject himself / herself is willing to do so. It is not uncommon for mental state to be unable to be objectively assessed due to reasons such as a lack of time or high psychological barriers to interviews. On the other hand, methods in which subjects themselves answer questions tend to be subjective and lack quantitativeness due to the nature of the items, including question-type items. For this reason, for example, scores can vary significantly depending on who answers the questions, and evaluations can be unstable over several months.

[0005] An object of the present invention is to provide an information processing device that can objectively and quantitatively evaluate a mental state without requiring an interview with a doctor or counselor.

[0006] In order to solve the above-mentioned problems and achieve the object, the present invention provides an information processing device that includes a judgment unit that generates a judgment result that judges the mental state of a subject by inputting pulse information related to the subject's heart rate and body movement information related to the subject's body movements into a trained model, and the trained model is a trained model that has been trained using pulse information related to the heart rate of the provider, body movement information related to the body movements of the provider, and index information related to the provider's specific mental tendencies for multiple providers who provide training data.

[0007] According to the present invention, it is possible to provide an information processing device that can objectively and quantitatively evaluate a mental state without requiring an interview with a doctor or counselor.

[0008] FIG. 1 is a diagram illustrating an example of a schematic configuration of an information processing system. FIG. 2 is a diagram illustrating an example of a hardware configuration of an information processing device. FIG. 3 is a flowchart illustrating a processing procedure of the information processing system. FIG. 4 is a diagram illustrating three axial directions of a wearable sensor. FIG. 5 is a diagram illustrating a feature calculation process. FIG. 6 is a diagram illustrating a trained model of the information processing system. FIG. 7 is a diagram illustrating an example of a schematic configuration of an information processing system. FIG. 8A is a diagram illustrating an example of a suitable attachment position and a suitable attachment direction of a wearable sensor. FIG. 8B is a diagram illustrating an example of a suitable attachment position and a suitable attachment direction of a wearable sensor. FIG. 9 is a diagram illustrating fluctuations in LF / HF in a healthy group and a symptomatic group. FIG. 10 is a diagram illustrating fluctuations in LF / HF in a healthy group and a symptomatic group. FIG. 11 is a diagram illustrating the relationship between the magnitude of circadian fluctuations in LF / HF and BDI. FIG. 12 is a diagram illustrating a learning curve in Example 1. FIG. 13 is a diagram illustrating a learning curve in Example 2.

[0009] An information processing device according to the present invention will be described below with reference to the drawings. Note that the following embodiments are not limited to the following descriptions. Furthermore, each embodiment can be combined with other embodiments or conventional techniques as long as there is no contradiction in the processing content.

[0010] The following describes a case in which the present invention is applied to psychological counseling. That is, the case is described in which an advisor (e.g., a doctor or counselor) who provides psychological counseling uses an information processing device according to the present invention to evaluate the level of depression of a subject receiving psychological counseling. The following embodiment is not limited to psychological counseling and can be applied to various situations in which the level of depression of a subject is evaluated.

[0011] (Embodiment) First, a schematic configuration of an information processing system 1 according to the present invention will be described with reference to Fig. 1. Fig. 1 is a diagram showing an example of the schematic configuration of the information processing system 1. As shown in Fig. 1, the information processing system 1 according to the embodiment includes an information processing device 10, a wearable sensor 20, and an information processing device 30.

[0012] The information processing device 10 is a server device that provides a mental state evaluation process as a service. For example, the information processing device 10 includes an acquisition unit 101, a determination unit 102, and an output control unit 103. Note that the functions of the information processing device 10 are not limited to the acquisition unit 101, the determination unit 102, and the output control unit 103. The acquisition unit 101, the determination unit 102, and the output control unit 103 will be described later.

[0013] The wearable sensor 20 is a sensor device formed in the shape of a flexible plate and attached to the trunk of a subject. For example, the wearable sensor 20 is attached by adhesion or suction to the skin of the trunk. Due to the nature of its function, the wearable sensor 20 may have a preferred attachment position and orientation (see FIGS. 8A and 8B ). In this case, it is preferable that information (marks or letters) indicating the preferred attachment position and orientation be written on the exterior surface of the wearable sensor 20 (a position visible to the user of the wearable sensor 20). Note that FIGS. 8A and 8B are diagrams showing an example of a preferred attachment position and orientation of the wearable sensor.

[0014] For example, the wearable sensor 20 is a multi-sensor device equipped with an electrocardiograph and a motion monitor. The electrocardiograph measures time-series electrocardiographic information (electrocardiographic waveform data) based on electrical signals flowing through the subject's body. The motion monitor (acceleration sensor) measures time-series acceleration information (acceleration data) based on the subject's body movements. The wearable sensor 20 records the various measured data in its internal memory in association with the measurement time. In this way, the wearable sensor 20 detects the subject's electrocardiographic information and acceleration information. Note that the wearable sensor 20 is an example of a detection unit.

[0015] The information processing device 30 is a terminal device operated by an advisor who provides psychological counseling, and is, for example, a computer such as a personal computer or a workstation. The information processing device 30 is connected to the information processing device 10 and the wearable sensor 20 via any communication means. As the communication means, any network such as a LAN (Local Area Network) or a WAN (Wide Area Network) can be used.

[0016] A representative example of processing in the information processing system 1 shown in FIG. 1 will now be described. The wearable sensor 20 is constantly worn by the subject for a certain period of time, such as several days to several weeks, and sequentially collects electrocardiogram information and acceleration information during that period. The subject removes the wearable sensor 20 when meeting with an advisor (e.g., a doctor or counselor) and hands the removed wearable sensor 20 to the advisor. The advisor operates the information processing device 30 to access the memory inside the wearable sensor 20 and reads the electrocardiogram information and acceleration information collected over the certain period. The advisor then transmits the read electrocardiogram information and acceleration information to the information processing device 10. The information processing device 10 assesses the subject's mental state based on the electrocardiogram information and acceleration information and transmits the assessment result to the information processing device 30. The advisor views the assessment result transmitted from the information processing device 10 and takes various measures based on the assessment result.

[0017] 1 is merely an example, and the present invention is not limited thereto. For example, wearable sensor 20 may additionally include any type of sensor, such as a sphygmograph that measures pulse waves or a thermistor that measures temperature (body temperature), in addition to an electrocardiograph and a motion monitor. Furthermore, in addition to sensors that record biological information, sensors that measure parameters that represent the environment in which the subject is placed, such as illuminance and humidity, can also be used as additional sensors. These additional sensors may be configured integrally with wearable sensor 20, or may be provided as separate (individual) sensors.

[0018] Furthermore, for example, the information processing device 30 does not need to be constantly connected to the information processing device 10 and the wearable sensor 20. It is sufficient for the devices to be connected when the various types of information described above are exchanged between them. Furthermore, when information is exchanged via a recording medium or the like, the devices do not need to be connected. Furthermore, the wearable sensor 20 may transmit various types of information directly to the information processing device 10 without going through the information processing device 30.

[0019] Furthermore, it is preferable that the wearable sensor 20 is always attached to the subject, but it may be removed to the extent that it does not affect the processing of the information processing device 10. Furthermore, various measurement data recorded in the wearable sensor 20 may be automatically transmitted to the information processing device 10 periodically (e.g., once every few hours) via a network.

[0020] Next, the hardware configuration of the information processing device 10 provided in the information processing system 1 will be described with reference to Fig. 2. Fig. 2 is a diagram showing an example of the hardware configuration of the information processing device 10. Note that while Fig. 2 describes the hardware configuration of the information processing device 10, the hardware configuration of the information processing device 30 is basically the same, and therefore description thereof will be omitted.

[0021] As shown in FIG. 2 , the information processing device 10 includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, an auxiliary storage device 14, an input device 15, a display device 16, and an external I / F (Interface) 17.

[0022] The CPU 11 is a processor (processing circuit) that executes a program to comprehensively control the operation of the information processing device 10 and realize various functions of the information processing device 10. For example, the functions of the acquisition unit 101, the determination unit 102, and the output control unit 103 of the information processing device 10 are realized by the CPU 11.

[0023] The ROM 12 is a non-volatile memory that stores various data (information written during the manufacturing stage of the information processing device 10) including a program for starting up the information processing device 10. The RAM 13 is a volatile memory that has a work area for the CPU 11. The auxiliary storage device 14 stores various data such as the program executed by the CPU 11. The auxiliary storage device 14 is configured, for example, by a hard disk drive (HDD), a solid state drive (SSD), or the like.

[0024] The input device 15 is a device that allows the operator of the information processing device 10 to perform various operations. The input device 15 is configured by, for example, a mouse, a keyboard, a touch panel, or hardware keys.

[0025] The display device 16 displays various types of information. For example, the display device 16 displays image data, model data, a GUI (Graphical User Interface) for receiving various operations from an operator, medical images, etc. The display device 16 is configured as, for example, a liquid crystal display, an organic EL (Electro Luminescence) display, or a cathode ray tube display. Note that the input device 15 and the display device 16 may be integrated into one unit, for example, in the form of a touch panel.

[0026] The external I / F 17 is an interface for connecting (communicating) with any external device such as a server device 20 .

[0027] 2 is merely an example, and the present invention is not limited to this. For example, the hardware configuration of the information processing device 10 may be any configuration of a known computer or workstation.

[0028] The processing of the information processing device 10 provided in the information processing system 1 will be described with reference to Fig. 3. Fig. 3 is a flowchart showing the processing procedure of the information processing system 1. The processing procedure shown in Fig. 3 is started, for example, in response to a request from an advisor. Note that Fig. 3 will be described with reference to Figs. 1, 2, and 4 to 6 as appropriate.

[0029] 3 , the acquisition unit 101 of the information processing device 10 receives the electrocardiogram information and acceleration information of the subject (step S101). For example, the advisor acquires time-series electrocardiogram information and acceleration information from the wearable sensor 20 that has been worn by the subject for seven days (approximately 168 hours), and transmits the acquired electrocardiogram information and acceleration information to the information processing device 10. The acquisition unit 101 receives the time-series electrocardiogram information and acceleration information transmitted from the information processing device 30.

[0030] Here, the electrocardiographic information is continuous data in which potentials measured by an electrocardiograph provided in the wearable sensor 20 are continuously recorded as numerical data. Furthermore, the acceleration information is continuous data in which accelerations in each of the three axial directions, i.e., the X-axis direction, the Y-axis direction, and the Z-axis direction, are continuously recorded as numerical data. These three axial directions are defined based on the structure of the wearable sensor 20, as illustrated in FIG. 4 . In FIG. 4 , the longitudinal direction of the wearable sensor 20 corresponds to the X-axis direction, the thickness direction of the wearable sensor 20 corresponds to the Z-axis direction, and the direction perpendicular to the X-axis direction and the Z-axis direction corresponds to the Y-axis direction. Note that these three axial directions are not limited to those illustrated in FIG. 4 , and may be defined in any direction based on the structure of the wearable sensor 20. Furthermore, FIG. 4 is a diagram for explaining the three axial directions in the wearable sensor 20.

[0031] If the wearable sensor 20 is worn by the subject for seven days, the subject's electrocardiogram information and acceleration information correspond to approximately 168 hours of continuous data. The subject's electrocardiogram information and acceleration information do not necessarily have to be 168 hours long. For example, the subject's electrocardiogram information and acceleration information preferably include 48 hours or more of continuous data, and more preferably include 144 hours or more of continuous data. Furthermore, the subject's electrocardiogram information and acceleration information may be linked to the subject's behavioral history (described below) and compared at any time or period, such as by day or day of the week. As an example, the subject's electrocardiogram information and acceleration information may be compared and analyzed using a 25-hour cycle, which is considered to be the cycle of the human biological clock.

[0032] Next, the acquisition unit 101 of the information processing device 10 acquires heartbeat information of the subject based on the electrocardiogram information of the subject (step S102). For example, the acquisition unit 101 calculates heartbeat information about the subject based on the time-series electrocardiogram information detected from the subject.

[0033] Specifically, the acquisition unit 101 divides the electrocardiographic information at predetermined intervals in the time series direction to generate a plurality of divided data arranged in time series. Next, the acquisition unit 101 calculates a plurality of types of statistical values ​​for each of the plurality of divided data. Then, the acquisition unit 101 calculates beat information for each of the plurality of types of statistical values. Note that the beat information is an example of a feature representing the characteristics of the electrocardiographic information in the time series direction. Furthermore, the divided data based on the electrocardiographic information will be referred to as "first divided data," and the statistical value based on the electrocardiographic information will be referred to as "first statistical value."

[0034] Here, the calculation process of the feature quantities will be described with reference to FIG. 5 . FIG. 5 is a diagram for explaining the calculation process of the feature quantities. As shown in FIG. 5 , for example, the electrocardiogram information is continuous data including n pieces of numerical data from "numerical data 1" to "numerical data n" (n is a natural number). The acquisition unit 101 generates a plurality of divided data by dividing the n pieces of numerical data included in this continuous data at 5-minute intervals, starting from the first piece. For example, if 100 pieces of numerical data are included in a 5-minute period, the acquisition unit 101 defines "numerical data 1" to "numerical data 100" as "divided data 1" and "numerical data 101" to "numerical data 200" as "divided data 2." In this way, the acquisition unit 101 generates m pieces of divided data from "divided data 1" to "divided data m" (m is a natural number smaller than n).

[0035] Next, the acquiring unit 101 calculates multiple types of statistical values ​​for each of the m pieces of divided data. The multiple types of statistical values ​​can be arbitrarily selected from any type of statistical value, such as mean value, variance, standard deviation, maximum value, minimum value, and skewness. For example, the acquiring unit 101 calculates the average value of 100 pieces of numerical data included in divided data 1 as statistical value data 1, and calculates the average value of 100 pieces of numerical data included in divided data 2 as statistical value data 2. In this way, the acquiring unit 101 calculates the average value of each of the m pieces of divided data to calculate m pieces of statistical value data (average value data) from "statistical value data 1" to "statistical value data m." Therefore, the acquiring unit 101 calculates m pieces of statistical value data for each type of statistical value.

[0036] The acquiring unit 101 then calculates a median and an interquartile range for each of the multiple types of statistical values. For example, the acquiring unit 101 calculates a median or an interquartile range using m pieces of statistical value data corresponding to the average values ​​of the m pieces of divided data. As a result, the acquiring unit 101 calculates a median or an interquartile range as beat information from the electrocardiogram information, which is continuous data in time series. Note that the calculated median and interquartile range are calculated from the divided data divided along the time series, and therefore can be said to represent the characteristics of the electrocardiogram information in the time series direction.

[0037] In this way, the acquisition unit 101 calculates the pulsation information based on the time-series electrocardiogram information detected from the subject.

[0038] 5 is merely an example, and the present invention is not limited to this. For example, the intervals at which consecutive data is divided and the number of numeric data included in the divided data are not limited to the values ​​shown in FIG. 5, and any numeric values ​​can be set.

[0039] Although FIG. 5 illustrates a case in which electrocardiographic information (electrocardiogram waveform) is used as the processing target of the calculation process, the present invention is not limited thereto. For example, the acquisition unit 101 may execute the calculation process using values ​​related to heart rate variability obtained from the electrocardiographic information as the processing target. Specifically, the acquisition unit 101 detects the position of R waves from the electrocardiographic information and generates heart rate variability (RRI time series) by plotting the intervals between the detected R waves (RR intervals: RRI) over time. The acquisition unit 101 then calculates a power spectrum from the RRI time series and integrates the power in a predetermined frequency range. As an example, the acquisition unit 101 calculates a very low frequency (VLF) component (0 Hz to 0.04 Hz), a low frequency (LF) component (0.04 Hz to 0.15 Hz), and a high frequency (HF) component (0.15 Hz to 0.40 Hz). The acquisition unit 101 then uses the values ​​of the heart rate variability, the VLF component, the LF component, and the HF component as processing targets for the calculation process. That is, the acquisition unit 101 calculates, as the statistical value, at least one of various statistical values ​​of heart rate variability obtained from electrocardiogram information, the peak frequency of a predetermined band component in a power spectrum based on frequency analysis of heart rate variability, and the integral value of the predetermined band component. The frequency analysis includes at least one of fast Fourier transform, Lomb-Scargle periodogram (see, for example, William H. Press and George B. Rybicki, "Fast Algorithm for Spectral Analysis of Unevenly Sampled Data," Astrophysical Journal, 338: 277-280, March 1989), and autoregressive model.

[0040] The acquiring unit 101 also acquires body movement information of the subject based on the acceleration information of the subject (step S103). For example, the acquiring unit 101 calculates body movement information regarding the body movement of the subject based on the time-series acceleration information detected from the subject.

[0041] Specifically, the acquisition unit 101 divides the acceleration information at predetermined intervals in the time series direction to generate multiple divided data arranged in time series. Next, the acquisition unit 101 calculates multiple types of statistical values ​​for each of the multiple divided data. Then, the acquisition unit 101 calculates body movement information for each of the multiple types of statistical values. Note that the body movement information is an example of a feature that represents the characteristics of the acceleration information in the time series direction. Furthermore, the divided data based on the acceleration information will be referred to as "second divided data," and the statistical value based on the acceleration information will be referred to as "second statistical value."

[0042] The calculation process for body movement information is basically the same as the calculation process for pulse information described in Fig. 5 except for the use of "acceleration information," and therefore will not be described further. Note that the acceleration information is information in which accelerations in the X-axis, Y-axis, and Z-axis directions are sequentially recorded as numerical data, and therefore body movement information is also calculated for each of the X-axis, Y-axis, and Z-axis directions.

[0043] The acquiring unit 101 may also calculate the correlation between the accelerations in each of the three axial directions as a statistical value. That is, the acquiring unit 101 calculates at least one of various statistical values ​​of the accelerations in each of the three axial directions included in the acceleration information and the correlation between the accelerations in each of the three axial directions as a statistical value.

[0044] The determination unit 102 of the information processing device 10 then inputs the heartbeat information and body movement information into the trained model to generate a determination result (step S104). Note that this trained model is trained using heartbeat information related to the heartbeats of multiple providers who provide training data, body movement information related to the body movements of the providers, and index information indicating the degree of a specific mental tendency of the providers. Note that the providers include those who have a specific mental tendency, such as a depressed state, and those who have recovered from the specific mental tendency.

[0045] Here, the trained model of the information processing system 1 will be described with reference to FIG. 6. FIG. 6 is a diagram for explaining the trained model of the information processing system 1. The upper part of FIG. 6 shows the processing during training of the trained model, and the lower part of FIG. 6 shows the processing during operation of the trained model. Here, the trained model is constructed in advance and stored in a predetermined storage area (e.g., ROM 12) available to the determination unit 102. The processing of the "learning unit" for constructing the trained model will be described later.

[0046] As shown in the upper part of Figure 6, during learning, machine learning is performed using, for example, pulse information, body movement information, and index information of multiple providers 1 to N. Here, the provider's pulse information and body movement information are calculated from the provider's electrocardiogram information and acceleration information, respectively. The method for calculating the provider's pulse information and body movement information is the same as the method for calculating the subject's pulse information and body movement information described in connection with the acquisition unit 101, and therefore will not be described here. The provider's electrocardiogram information and acceleration information are continuous data acquired from the provider for a certain period of time. For example, the provider's electrocardiogram information and acceleration information preferably include continuous data for 48 hours or more, and more preferably include continuous data for 144 hours or more.

[0047] As described with reference to FIG. 4 , the acceleration information is recorded in each of three axial directions: the X-axis, the Y-axis, and the Z-axis, which are defined based on the structure of the wearable sensor 20. Because the orientation of the wearable sensor 20 when worn by the subject is not specifically defined, the three axial directions of the acceleration information do not necessarily coincide with the subject's orientation (front-back, left-right, and up-down). Furthermore, the relationship between the three axial directions of the acceleration information and the subject's orientation changes each time the wearable sensor 20 is worn by the subject. Therefore, the present invention does not take into account the above-described relationship, and instead uses the acceleration information differentiated for each of the three axial directions: the X-axis, the Y-axis, and the Z-axis, as individual training data. This eliminates the need for correction processing or the like to consider the above-described relationship, thereby reducing the processing load during learning and operation and enabling the use of a wearable sensor 20 with a simple configuration.

[0048] The provider's index information is information that indicates the degree of a specific mental tendency of the provider. For example, the index information includes at least one of objective index information based on the advisor's objective opinion of the provider and subjective index information based on the provider's subjective opinion. The objective index information is information evaluated using at least one of the Clinical Global Impression of Improvement (CGI-I), Hamilton Rating Scale for Depression-17 item score (HAMD-17), and a Montgomery-Asberg Depression Rating Scale score (MADRS). The subjective index information is information evaluated using at least one of the Beck Depression Inventory (BDI), Patient Health Questionnaire-9 (PHQ-9), and State Trait Anxiety Inventory (STAI).

[0049] The CGI-I is an index used to evaluate a subject's improvement (or worsening) and is scored by assessing the subject's symptoms compared to baseline. The HAMD-17 is a depression rating scale that assesses overall depression symptoms, including physical symptoms. The MADRS is a 10-item depression severity rating scale. The BDI is a self-administered questionnaire that assesses the severity of depressive symptoms and is scored by the subject selecting the statement that most closely matches their feelings. The PHQ-9 is a nine-item self-administered test used to detect and assess the severity of depression. The STAI is a questionnaire that measures the psychological state of anxiety and can also capture personality traits such as anxiety susceptibility. The STAI assesses STAI-S (state anxiety) and STAI-T (trait anxiety). It is preferable that the various indicator information be information from the period corresponding to the period during which the provider's heartbeat information and body movement information were acquired.

[0050] For example, a trained model is constructed by performing machine learning using XGBoost. This trained model is a model that receives pulse information and body movement information of a subject and outputs a determination result of the subject's mental state.

[0051] 6, during operation, the determination unit 102 inputs the subject's heartbeat information and body movement information into the trained model constructed by machine learning, causing the trained model to output a determination result. In this way, the determination unit 102 generates a determination result.

[0052] 6 can be executed at any time before the operation process is executed. The learning process can also be executed to update (additionally learn) an already generated trained model.

[0053] Furthermore, in the present invention, the information input to the trained model is only beat information or body movement information, and is not the value of the electrocardiogram information itself, the results of nonlinear analysis of heart rate fluctuations (entropy, fluctuations), the three-axis composite acceleration, the entropy of Y-axis acceleration, or the entropy of Z-axis acceleration. In other words, the beat information is different from the electrocardiogram waveform contained in the electrocardiogram information and the value based on nonlinear analysis of heart rate fluctuations obtained from the electrocardiogram information. The body movement information is different from the composite acceleration of accelerations in each of the three axial directions contained in the acceleration information, the entropy of acceleration in the Y-axis direction of the three axial directions, and the entropy of acceleration in the Z-axis direction of the three axial directions of the three axial directions.

[0054] Then, the output control unit 103 outputs the determination result (step S105). For example, the output control unit 103 transmits the determination result to the information processing device 30. The information processing device 30 displays the determination result transmitted from the output control unit 103 on a display device or stores it in a predetermined storage device.

[0055] As described above, in the information processing system 1 according to the present invention, the determination unit 102 of the information processing device 10 inputs pulse information related to the subject's heartbeat and body movement information related to the subject's body movement into a trained model, thereby generating a determination result of the subject's mental state. This allows the information processing system 1 to quantitatively evaluate the mental state.

[0056] For example, the information processing system 1 determines the mental state of the subject based on information detected (measured) using the wearable sensor 20. This allows the information processing system 1 to quantitatively evaluate the mental state of the subject. For example, the information processing system 1 can stably perform evaluation over several months without relying on individual subjectivity.

[0057] Furthermore, the information processing system 1 judges the mental state of the subject using a trained model that has undergone machine learning using objective index information based on the advisor's objective opinion. This allows the information processing system 1 to provide a judgment result that is close to an evaluation based on the advisor's objective opinion, even if the subject does not have a meeting with the advisor.

[0058] The number of feature quantities that can be calculated in the present invention reaches several hundred types, but among them, useful feature quantities are shown in Table 1.

[0059]

[0060] In Table 1, "feature amount" is the name of the feature amount. The name of the feature amount is determined based on the calculation method of the feature amount.

[0061] In other words, the first "interquartile range of peak frequency of HF component based on fast Fourier transform" is calculated by performing frequency analysis of heart rate fluctuations using fast Fourier transform, calculating the peak frequency of the HF component of the power spectrum obtained for each first divided data as the first statistical value, and then calculating the interquartile range of the peak frequency of each calculated first divided data as beat information.

[0062] The second "interquartile range of standard deviation of X-axis acceleration" is calculated by calculating the standard deviation of acceleration in the X-axis direction for each second divided data as the second statistical value, and then calculating the interquartile range of the standard deviation of each calculated second divided data as body movement information.

[0063] The third item, "Interquartile range of skewness of Y-axis acceleration," is calculated by calculating the skewness of acceleration in the Y-axis direction for each second divided data as a second statistical value, and then calculating the interquartile range of the skewness of each calculated second divided data as body movement information.

[0064] The fourth item, "median of minimum values ​​of X-axis acceleration," is calculated by calculating the minimum value of acceleration in the X-axis direction for each second divided data as the second statistical value, and then calculating the interquartile range of the calculated minimum values ​​for each second divided data as body movement information.

[0065] The fifth "median of minimum RRI values" is calculated by calculating the minimum value of heart rate variability for each first divided data as the first statistical value, and then calculating the median of the minimum values ​​of each calculated first divided data as the beat information.

[0066] The sixth item, "Interquartile range of maximum X-axis acceleration," is calculated by calculating the maximum value of acceleration in the X-axis direction for each second divided data as the second statistical value, and then calculating the interquartile range of the maximum value of each calculated second divided data as body movement information.

[0067] The seventh item, "Interquartile range of correlation between X-axis acceleration and Z-axis acceleration," is calculated by calculating the correlation between the acceleration in the X-axis direction and the acceleration in the Z-axis direction of each second divided data as a second statistical value, and then calculating the interquartile range of the calculated correlation of each second divided data as body movement information.

[0068] The eighth item, "median of integral values ​​(relative values) of the VLF component based on the Lomb-Scargle periodogram," is calculated by performing frequency analysis of heart rate variability using the Lomb-Scargle periodogram, calculating the integral value (relative value) of the VLF component of the power spectrum obtained for each first divided data as the first statistical value, and then calculating the median of the calculated integral values ​​(relative values) of each first divided data as the beat information. Note that this integral value (relative value) is a ratio to the sum of the integral values ​​of each frequency component. For example, the integral value (relative value) of the VLF component is calculated by dividing the integral value of the VLF component by the sum of the integral values ​​of the VLF component, LF component, and HF component.

[0069] The ninth item, "median of average X-axis acceleration," is calculated by calculating the average value of acceleration in the X-axis direction for each second divided data as the second statistical value, and then calculating the median of the average values ​​of each calculated second divided data as body movement information.

[0070] The tenth item, "the interquartile range of the integral value (relative value) of the LF component based on the autoregressive model," is calculated by performing a frequency analysis of heart rate fluctuations using the autoregressive model, calculating the integral value (relative value) of the LF component of the power spectrum obtained for each of the first divided data as the first statistical value, and then calculating the interquartile range of the calculated integral value (relative value) of each of the first divided data as the beat information. Note that the integral value (relative value) of the LF component is calculated, for example, by dividing the integral value of the LF component by the sum of the integral values ​​of the VLF component, the LF component, and the HF component.

[0071] The 11th item, "Interquartile range of correlation between X-axis acceleration and Y-axis acceleration," is calculated by calculating the correlation between the acceleration in the X-axis direction and the acceleration in the Y-axis direction of each second divided data as a second statistical value, and then calculating the interquartile range of the calculated correlation of each second divided data as body movement information.

[0072] The twelfth item, "Interquartile range of the integral value (logarithm) of the HF component based on fast Fourier transform," is obtained by performing frequency analysis of heart rate fluctuations using fast Fourier transform, calculating the integral value (logarithm) of the HF component of the power spectrum obtained for each first divided data as the first statistical value, and then calculating the interquartile range of the integral value (logarithm) of the peak frequency of each calculated first divided data as beat information. Note that this integral value (logarithm) is a logarithmic representation of the obtained integral value. For example, the integral value (logarithm) of the HF component is the common logarithm of the integral value of the HF component.

[0073] The 13th item, "Median of correlation between X-axis acceleration and Y-axis acceleration," is calculated by calculating the correlation between the acceleration in the X-axis direction and the acceleration in the Y-axis direction of each second divided data as a second statistical value, and then calculating the median of the calculated correlation between each second divided data as body movement information.

[0074] The 14th item, "Median of peak frequencies of HF components based on fast Fourier transform," is calculated by performing frequency analysis of heart rate fluctuations using fast Fourier transform, calculating the peak frequency of the HF component of the power spectrum obtained for each first divided data as the first statistical value, and then calculating the median of the peak frequencies of each calculated first divided data as beat information.

[0075] The 15th item, "median of average Y-axis acceleration," is calculated by calculating the average value of acceleration in the Y-axis direction for each second divided data as the second statistical value, and then calculating the median of the average values ​​of each calculated second divided data as body movement information.

[0076] The 16th item, "median of integral values ​​(normalized values) of LF components based on an autoregressive model," is calculated as the first statistical value by performing frequency analysis of heart rate fluctuations using an autoregressive model, calculating the integral values ​​(normalized values) of the LF components of the power spectrum obtained for each first divided data, and then calculating the median of the calculated integral values ​​(normalized values) of each first divided data as the beat information. Note that this integral value (normalized value) is a ratio to the sum of the integral values ​​of the LF component and the integral values ​​of the HF component. For example, the integral value (normalized value) of the LF component is calculated by dividing the integral value of the LF component by the sum of the integral values ​​of the LLF component and the integral value of the HF component.

[0077] The 17th item, "Interquartile range of the integrated value (logarithm) of the VLF component based on the fast Fourier transform," is obtained by performing a frequency analysis of heart rate variability using the fast Fourier transform, calculating the integrated value (logarithm) of the VLF component of the power spectrum obtained for each of the first divided data as the first statistical value, and then calculating the interquartile range of the calculated integrated value (logarithm) of each of the first divided data as the beat information. Note that, for example, the integrated value (logarithm) of the VLF component is the common logarithm of the integrated value of the VLF component.

[0078] The 18th item, "Interquartile range of the integrated value of the HF component based on the Lomb-Scargle periodogram," is obtained by performing frequency analysis of heart rate fluctuations using the Lomb-Scargle periodogram, calculating the integrated value of the HF component of the power spectrum obtained for each of the first divided data as the first statistical value, and then calculating the interquartile range of the calculated integrated value of each of the first divided data as the beat information. Note that this integrated value is the value obtained by integrating the power of each frequency domain.

[0079] It is not necessary to use all of the 18 features shown in Table 1. For example, it is preferable that the trained model of the information processing system 1 be trained and operated using at least one of the 18 features.

[0080] (Variation 1) In the above embodiment, the case where continuous data for a certain period acquired from the subject is itself used as the processing target for calculating the feature amount has been described, but the present invention is not limited to this. For example, the acquisition unit 101 may select, from the acquired continuous data for a certain period, numerical data corresponding to a period during which a specific behavior was performed (specific behavior period), and use the selected numerical data for calculating the feature amount.

[0081] That is, the acquisition unit 101 calculates heartbeat information based on electrocardiogram information corresponding to a specific behavior period among the time-series electrocardiogram information. Also, the acquisition unit 101 calculates body movement information based on acceleration information corresponding to a specific behavior period among the time-series acceleration information. Here, the specific behavior period corresponds to a period during which a predetermined type of behavior continues for a predetermined time or longer.

[0082] For example, the acquisition unit 101 generates a behavioral history of the subject by analyzing time-series acceleration information. The behavioral history is information in which the behavior of the subject is associated with the date and time when the behavior was performed. As a result, the acquisition unit 101 generates a behavioral history including various behaviors such as sleep, exercise, eating, drinking, and work. Note that any known technology can be applied to the process of generating a behavioral history by analyzing acceleration information. Furthermore, the behavioral history of work may be further subdivided, for example, into desk work, meetings, presentations, etc.

[0083] The acquisition unit 101 then references the generated behavioral history and identifies, for example, a "period in which the subject slept for one hour or more" as a specific behavioral period. As a result, the acquisition unit 101 identifies one or more "periods in which the subject slept for one hour or more" from the electrocardiogram information and acceleration information for the certain period. The acquisition unit 101 then extracts electrocardiogram information corresponding to the identified one or more specific behavioral periods from the electrocardiogram information for the certain period, and performs a feature calculation process (heartbeat information) using the electrocardiogram information for the extracted period as a processing target. The acquisition unit 101 also extracts acceleration information corresponding to the identified one or more specific behavioral periods from the acceleration information for the certain period, and performs a feature calculation process (body movement information) using the acceleration information for the extracted period as a processing target. The feature calculation process is similar to the calculation process described in FIG. 5 , so a description thereof will be omitted.

[0084] In this way, the acquisition unit 101 calculates the heartbeat information and the body movement information based on, for example, the electrocardiogram information and the acceleration information corresponding to the “period of sleeping for one hour or more.” In this way, the information processing system 1 according to the first modification can quantitatively evaluate the mental state of the subject during the period in which a specific behavior was performed.

[0085] Although the first modification has been described with reference to a case in which a specific behavioral period is identified based on time-series acceleration information, the present invention is not limited to this. For example, a specific behavioral period can also be identified by analyzing time-series electrocardiogram information. For example, the acquisition unit 101 analyzes time-series changes in heart rate based on electrocardiogram information over a certain period of time. Then, the acquisition unit 101 can generate a behavioral history including various behaviors, such as sleep and exercise, based on the time-series changes in heart rate. The acquisition unit 101 may also generate a behavioral history by analyzing a combination of time-series electrocardiogram information and acceleration information. That is, the acquisition unit 101 can generate a behavioral history of the subject based on at least one of the time-series electrocardiogram information and the time-series acceleration information, and identify a specific behavioral period.

[0086] (Modification 2) The specific behavior period may be specified as a period from a preset start time to a preset end time.

[0087] For example, a period from 11:00 PM to 6:00 AM is preset as a specific sleep behavior period. In this case, the acquisition unit 101 extracts electrocardiogram information and acceleration information corresponding to the period from 11:00 PM to 6:00 AM from the time-series electrocardiogram information and acceleration information, and performs feature calculation processing on the electrocardiogram information and acceleration information for the extracted period as processing targets.

[0088] As a result, the information processing system 1 according to the second modification can quantitatively evaluate the mental state of the subject during the period in which a specific behavior was performed.

[0089] The specific behavioral period set in Modification 2 may be specified by the operator of the information processing system 1. For example, the advisor, who is the operator, may confirm the lifestyle habits of the subject during an interview and determine that the period from 10:00 PM to 5:00 AM is the sleeping time. In this case, the advisor can operate the information processing device 30 and set the period from 10:00 PM to 5:00 AM as the specific behavioral period for sleeping.

[0090] The specific behavior period may also be manually set by the subject on a case-by-case basis. For example, since the sleeping time varies from day to day, the subject may input the sleeping time (start time and end time) by himself / herself.

[0091] (Variation 3) Numerical data may be selected from continuous data using a period specified by various conditions, instead of a period during which a specific behavior was performed (specific behavior period), and used in the calculation process of the feature amount.

[0092] For example, a specific condition period specified by any condition such as the time of sunrise and sunset, the duration of daylight, weather (such as the duration of rain), temperature (the duration when the temperature is above or below a certain level), etc. can be set, and the feature amount can be calculated using this. Note that the specific condition period and the specific behavior period are collectively referred to as a specific period.

[0093] That is, the acquiring unit 101 may calculate the heartbeat information based on electrocardiogram information corresponding to a specific period among the time-series electrocardiogram information.Furthermore, the acquiring unit 101 may calculate the body movement information based on acceleration information corresponding to a specific period among the time-series acceleration information.

[0094] Other Embodiments The present invention may be implemented in various different forms other than the above-described embodiment.

[0095] (System Configuration) In the above embodiment, the schematic configuration of the information processing system 1 is described using Fig. 1, but the present invention is not limited to this. For example, the processing functions of the acquisition unit 101, the determination unit 102, and the output control unit 103 provided in the information processing device 10 may be provided in any device provided in the information processing system 1. As an example, the acquisition unit 101 may be provided in the information processing device 30.

[0096] That is, the information processing system 1 includes a determination unit that generates a determination result of the mental state of the subject by inputting pulse information related to the subject's heartbeat and body movement information related to the subject's body movement into a trained model. The trained model is trained using pulse information related to the provider's heartbeat, body movement information related to the provider's body movement, and index information indicating the degree of a specific mental tendency of the provider for multiple providers who provide training data.

[0097] (Learning Unit) The system according to the present invention may also include a processing function for constructing the trained model described in the above embodiment.

[0098] The schematic configuration of an information processing system 2 according to the present invention will be described using Fig. 7. Fig. 7 is a diagram showing an example of the schematic configuration of the information processing system 2. As shown in Fig. 7, the information processing system 2 according to the embodiment includes a wearable sensor 20, an information processing device 30, and an information processing device 40. The configurations of the wearable sensor 20 and the information processing device 30 shown in Fig. 7 are basically the same as the configurations of the wearable sensor 20 and the information processing device 30 shown in Fig. 1, and therefore description thereof will be omitted.

[0099] The information processing device 40 is a server device that constructs a trained model for evaluating a mental state. For example, the information processing device 40 includes an acquisition unit 401, a learning unit 402, and an output control unit 403. The configurations of the acquisition unit 401 and the output control unit 403 shown in Fig. 7 are basically the same as the configurations of the acquisition unit 101 and the output control unit 103 shown in Fig. 1, and therefore description thereof will be omitted.

[0100] Here, the learning unit 402 performs machine learning using the heartbeat information, body movement information, and index information of multiple providers 1 to N to construct a trained model for evaluating the mental state. For example, the learning unit 402 performs machine learning using XGBoost, as shown in the upper part of FIG. 6 . Through this machine learning, the learning unit 402 receives the heartbeat information and body movement information of the subject and generates a trained model that outputs a determination result of the subject's mental state. The learning unit 402 stores the generated trained model in any storage area available to the determination unit 102.

[0101] Although the case where machine learning is performed using XGBoost has been described here, the present invention is not limited to this. The learning unit 402 is not limited to XGBoost and can apply any known machine learning algorithm.

[0102] Although the case where the information processing device 40, which is different from the information processing device 10, includes the learning unit 402 has been described here, the present invention is not limited to this. For example, the learning unit 402 may be included in the information processing device 10.

[0103] According to the embodiment and modifications described above, the mental state can be quantitatively evaluated.

[0104] The programs executed by the information processing systems 1 and 2 according to the above-described embodiments and modifications may be provided by being recorded in an installable or executable file format on a computer-readable recording medium such as a CD-ROM, a flexible disk (FD), a CD-R, a DVD, or a USB (Universal Serial Bus), or may be provided or distributed via a network such as the Internet. Also, various programs may be provided by being pre-installed in a non-volatile storage medium such as a ROM.

[0105] For example, the programs executed in the information processing systems 1 and 2 according to the above-described embodiments and modifications cause a computer to execute a determination process that generates a determination result of the mental state of a subject by inputting pulse information related to the subject's heartbeat and body movement information related to the subject's body movement into a trained model. This trained model is trained for multiple providers who provide training data using pulse information related to the provider's heartbeat, body movement information related to the provider's body movement, and index information indicating the degree of a specific mental tendency of the provider.

[0106] EXAMPLES The present invention will be described in more detail below based on examples, but the present invention is not limited to these examples.

[0107] Example 1 Using the method described in the above embodiment, a trained model for classifying subjects into one of two groups, a healthy group and a symptomatic group, was created, and its accuracy was verified.

[0108] First, we will explain the samples. Using the method described above, we collected data from 106 samples, including a healthy group and a symptomatic group. Each sample corresponded to continuous electrocardiogram and acceleration information data for up to 7 days (approximately 168 hours). Of the 106 samples, 63 belonged to the healthy group and 43 belonged to the symptomatic group. Discrimination between the healthy and symptomatic groups was performed comprehensively using the CGI-I, HAMD-17, MADRS, BDI, PHQ-9, STAI-S, and STAI-T. The healthy group included healthy individuals aged 20 years or older with no history of psychiatric illness and individuals diagnosed with remission of depression (depressive symptoms). The collection period for each sample is described as "up to 7 days" because some continuous data (e.g., 5 days' worth) was included.

[0109] The fluctuations in LF / HF in the healthy and symptomatic groups were confirmed for the above samples. Figure 9 shows a time series plot of the average LF / HF over a 24-hour cycle and its polynomial approximation. Figure 10 shows the 95-5th percentile of the LF / HF shown in Figure 9. The results in Figures 9 and 10 confirmed that the symptomatic group had smaller fluctuations in LF / HF throughout the day. Note that Figures 9 and 10 show the fluctuations in LF / HF in the healthy and symptomatic groups.

[0110] Furthermore, the relationship between the magnitude of diurnal fluctuations in LF / HF and BDI was confirmed for the above samples. Figure 11 shows a scatter plot, similar to Figure 10, with the 95-5th percentile of LF / HF on the vertical axis and the BDI score of each sample on the horizontal axis. The results of Figure 11 confirmed that the higher the BDI score, i.e., the higher the severity of depressive symptoms, the smaller the fluctuations in LF / HF throughout the day. Note that Figure 11 shows the relationship between the magnitude of diurnal fluctuations in LF / HF and BDI.

[0111] 9 to 11 confirm the usefulness of the above sample.

[0112] The method for constructing the trained model will now be described. The samples were the electrocardiogram information and acceleration information of the donor, and information (labels) indicating whether each sample was a healthy group or a symptomatic group was used as index information. The trained model was constructed using the method described above. XGBoost (Extreme Gradient Boosting Decision Tree) was used as the machine learning model.

[0113] The method for validating the trained model will now be described. The above samples were input as the subject's electrocardiogram information and acceleration information into the trained model constructed by the above construction method, and a determination result indicating whether the subject was in the healthy group or the symptomatic group was obtained. The obtained determination result was verified by cross-validation.

[0114] The verification results are shown in Table 2 and Fig. 12. Fig. 12 is a diagram showing the learning curve in Example 1. These results show that in Example 1, sufficient accuracy was obtained in terms of accuracy, precision, sensitivity, and specificity, and that mental states can be objectively and quantitatively evaluated without the need for interviews with a doctor or counselor.

[0115]

[0116] Example 2 Using the method described in the above embodiment, a trained model for classifying subjects into three groups based on BDI was created, and its accuracy was verified.

[0117] The samples and the method for constructing the trained model used in Example 2 were basically the same as those in Example 1, but the index information for each sample was information (label) indicating one of "healthy," "borderline," and "depression" based on the BDI value. Specifically, samples with a BDI of 10 or less were classified as the "healthy group," samples between 11 and 16 were classified as the "borderline group," and samples with a BDI of 17 or more were classified as the "depression group."

[0118] The verification method for the trained model is basically the same as in Example 1, but the judgment result by the trained model is information indicating whether the subject falls into the “healthy group,” “borderline group,” or “depressed group.”

[0119] The verification results are shown in Table 3 and Fig. 13. Fig. 13 is a diagram showing the learning curve in Example 2. When classified into the three groups shown in Table 3, the average accuracy rate was 0.65, and the standard deviation was 0.189. These results demonstrate that, also in Example 2, mental state can be objectively and quantitatively evaluated without the need for an interview with a doctor or counselor.

[0120]

[0121] REFERENCE SIGNS LIST 1, 2 Information processing system 10, 30, 40 Information processing device 11 CPU 12 ROM 13 RAM 14 Auxiliary storage device 15 Input device 16 Display device 17 External I / F 20 Wearable sensor 101 Acquisition unit 102 Determination unit 103 Output control unit 104 Learning unit

Claims

1. An information processing device comprising: a judgment unit that generates a judgment result of the mental state of a subject by inputting pulse information related to the subject's heart rate and body movement information related to the subject's body movements into a trained model, wherein the trained model is a trained model trained for a plurality of providers providing training data using pulse information related to the heart rate of the provider, body movement information related to the provider's body movements, and index information related to a specific mental tendency of the provider.

2. The information processing device according to claim 1, wherein the providers include individuals who have the specific mental tendency and individuals who have been cured of the specific mental tendency.

3. The information processing device according to claim 1, further comprising an acquisition unit that acquires the heartbeat information of the subject and the body movement information of the subject.

4. The information processing device according to claim 1, wherein the index information is information representing the degree of a specific mental tendency of the provider.

5. The information processing device according to claim 1, wherein the index information includes at least one of objective index information based on an advisor's objective opinion of the provider and subjective index information based on the provider's subjective opinion.

6. The information processing device of claim 5, wherein the objective index information is information evaluated by at least one of the Clinical Global Impression of Improvement (CGI-I), Hamilton Rating Scale for Depression-17 item score (HAMD-17), and a Montgomery-Asberg Depression Rating Scale score (MADRS), and the subjective index information is information evaluated by at least one of the Beck Depression Inventory (BDI), Patient Health Questionnaire-9 (PHQ-9), and State Trait Anxiety Inventory (STAI).

7. The information processing device of claim 3, wherein the acquisition unit calculates the pulse information based on time-series electrocardiogram information detected from the subject, and calculates the body movement information based on time-series acceleration information detected from the subject.

8. The information processing device of claim 7, wherein the acquisition unit: divides the electrocardiogram information at a predetermined interval in the time series direction to generate a plurality of first split data arranged in a time series; calculates a plurality of types of first statistical values ​​for each of the plurality of first split data; calculates the beat information representing the characteristics of the electrocardiogram information in the time series direction for each of the plurality of first statistical values; divides the acceleration information at a predetermined interval in the time series direction to generate a plurality of second split data arranged in a time series; calculates a plurality of types of second statistical values ​​for each of the plurality of second split data; and calculates the body movement information representing the characteristics of the acceleration information in the time series direction for each of the plurality of second statistical values.

9. The information processing device of claim 8, wherein the first statistical value includes at least one of various statistical values ​​of heart rate fluctuations obtained from the electrocardiogram information, a peak frequency of a specified band component in a power spectrum based on frequency analysis of the heart rate fluctuations, and an integral value of the specified band component; the pulse information includes at least one of a median and an interquartile range of the first statistical value; the second statistical value includes at least one of various statistical values ​​of acceleration in each of the three axial directions included in the acceleration information and a correlation between the accelerations in each of the three axial directions; and the body movement information includes at least one of a median and an interquartile range of the second statistical value.

10. The information processing device according to claim 9, wherein the frequency analysis includes at least one of a fast Fourier transform, a Lomb-Scargle Periodogram, and an autoregressive model.

11. The information processing device of claim 7, wherein the beat information is different from the electrocardiogram waveform contained in the electrocardiogram information and a value based on nonlinear analysis of heart rate fluctuations obtained from the electrocardiogram information, and the body movement information is different from the resultant acceleration of the acceleration in each of the three axial directions contained in the acceleration information, the entropy of the acceleration in the Y-axis direction of the three axial directions, and the entropy of the acceleration in the Z-axis direction of the three axial directions.

12. The information processing device of claim 7, wherein the electrocardiogram information of the provider includes continuous data for 48 hours or more obtained from the provider, and the acceleration information of the provider includes continuous data for 48 hours or more obtained from the provider.

13. The information processing device of claim 7, wherein the electrocardiogram information of the provider includes continuous data for 144 hours or more obtained from the provider, and the acceleration information of the provider includes continuous data for 144 hours or more obtained from the provider.

14. The information processing device of claim 7, wherein the electrocardiogram information of the subject includes continuous data obtained from the subject for 48 hours or more, and the acceleration information of the subject includes continuous data obtained from the subject for 48 hours or more.

15. The information processing device of claim 7, wherein the acquisition unit calculates the heartbeat information based on electrocardiogram information corresponding to a specific period among the time-series electrocardiogram information, and calculates the body movement information based on acceleration information corresponding to the specific period among the time-series acceleration information.

16. The information processing device according to claim 15, wherein the specific period includes a specific behavior period during which a predetermined type of behavior continues for a predetermined period of time or longer, and a specific condition period specified by any condition.

17. The information processing device of claim 15, wherein the specific period is a specific behavioral period during which a predetermined type of behavior continues for a predetermined time or longer, and the specific behavioral period corresponds to any one of a period from a predetermined start time to a predetermined end time, a period specified by an operator, and a period specified based on at least one of the time-series electrocardiogram information and the time-series acceleration information.

18. The information processing device according to claim 17, further comprising a detection unit attached to the subject and configured to detect the electrocardiogram information and the acceleration information.

19. The information processing device according to claim 18, wherein the detection unit is attached to the trunk of the subject.

20. The information processing device according to claim 18, wherein the detection unit is a flexible plate and is attached to the skin of the subject's trunk by adhesion or suction.

Images