Apparatus for mental disease diagnosis assistance and treatment, and operating method thereof

Abstract

An apparatus for mental disease diagnosis assistance and treatment, and an operating method thereof are disclosed. The method may comprise the steps of: receiving passive data and first active data from a user device; extracting behavioral characteristic data of a patient from the passive data; using a subjective emotional state prediction model that uses, as an input, the behavioral characteristic data, thereby predicting second active data of the patient; using a mental disease scale classification model that uses, as an input, the behavioral characteristic data and the second active data, thereby outputting prediction values representing a mental disease scale; and providing, to a user, a user interface indicating the prediction values representing a mental disease scale.

Description

Device for assisting in the diagnosis and treatment of mental disorders and method of operation thereof

[0001] The following embodiments relate to a device and its operating method for assisting in the diagnosis and treatment of mental illness. More specifically, the following embodiments relate to a device and its operating method for analyzing digital phenotypes based on Ecological Momentary Assessment (EMA), predicting a patient's depression scale using artificial intelligence models, and thereby assisting in the diagnosis and treatment of mental illness.

[0002] As the use of digital devices becomes more widespread, the potential for their application in the medical field is increasing. Digital phenotyping involves processing large amounts of lifestyle data to understand a patient's current condition. Digital phenotyping can have clinical implications for medical assessment and diagnosis by enabling measurement of previously difficult-to-measure areas. It can be used to derive clinical implications previously inaccessible in clinical settings or as a basis for new discoveries. Therefore, ongoing research is underway to consider and incorporate digital phenotyping.

[0003] Meanwhile, with the unprecedented COVID-19 pandemic leading to the introduction of telemedicine and prescriptions, questions are continually being raised about whether remote consultations actually deliver therapeutic benefits to patients and their cost-effectiveness. Consequently, discussions are ongoing within the industry and pharmaceutical communities regarding the acceptable scope and methods of remote consultations in various medical fields, including diabetes, chronic heart failure, and mental illness.

[0004] Until recently, the diagnosis and management of mental illness remained largely focused on face-to-face interactions, relying on self-reporting and professional consultations. In the field of psychiatry, where illness is diagnosed based on the presence or absence of symptoms, unlike other medical fields where new technologies like artificial intelligence and various testing techniques are being introduced, patient self-reporting and face-to-face consultations with professionals remain the most important testing and diagnostic methods.

[0005] However, when diagnosing mental illness, especially depression, in hospitals and other settings, there is a problem in that it takes a lot of time and money if it is done through expert consultation based on a patient scale test in the form of a questionnaire.

[0006] The problem that the present disclosure seeks to solve is to provide a device and an operating method for assisting in the diagnosis and treatment of mental disorders.

[0007] The problems to be solved by the present disclosure are not limited to the problems described above, and problems not mentioned can be clearly understood by a person having ordinary skill in the technical field to which the present disclosure pertains from this specification and the attached drawings.

[0008] According to one embodiment, a method of operating a device for assisting diagnosis and treatment of mental illness may include the steps of receiving passive data and first active data from a user device, extracting patient behavioral characteristic data from the passive data, predicting second active data of the patient using a subjective emotional state prediction model having the behavioral characteristic data as input, outputting a predicted value representing a mental illness scale using a mental illness scale classification model having the behavioral characteristic data and the second active data as input, and providing a user interface that displays the predicted value representing the mental illness scale to the user. Here, the second active data may be a predicted value of a total score of items answered by the patient in a mental health scale test.

[0009] The above passive data may include information about one or more of GPS coordinates, acceleration sensor data, tilt sensor data, screen ON time, screen OFF time, call start time, call end time, type of app being used, and usage time of the app received from the user device.

[0010] The above behavioral characteristic data may include information on at least one of the visited location, time spent at each visited location, number of steps, distance traveled, moving speed, distance covered, walking speed, activity complexity, and sleep time.

[0011] The above subjective emotional state prediction model includes a boosting-based regression algorithm, is learned using the behavioral characteristic data and the first active data, and can be learned using a feature subset determined based on RFE (Recursive Feature Elimination) among the behavioral characteristic data.

[0012] The above mental illness scale classification model includes a boosting-based classification algorithm and can be trained using the passive data or the second active data augmented using the synthetic minority oversampling technique (SMOTE).

[0013] The step of extracting patient behavioral characteristic data from the above passive data may include the step of extracting patient activity complexity data from the above passive data.

[0014] The step of extracting the activity complexity data of the patient from the passive data may include the steps of collecting the location data of the patient for all dates and all time zones during a certain period, the step of clustering the location data at a specific time for each date and selecting a first location that is located the most at the specific time, the step of generating a repetitive life pattern graph based on the first location for all time zones, the step of comparing the repetitive life pattern graph with the life pattern graph of each date and calculating the time during which the life pattern graph of each date does not match the pattern of the repetitive life pattern graph, and the step of calculating the average of the time during the certain period during which the life pattern graph does not match the pattern of the repetitive life pattern graph.

[0015] The method may further include providing recommended treatment content or recommended treatment solution to the patient based on at least one of the patient's passive data, the behavioral characteristic data, the second active data, or the predicted value representing a mental illness scale.

[0016] A computer program for mental health monitoring according to one embodiment may be a computer program stored in a computer-readable recording medium to perform the method of operating the above-described mental illness diagnosis assistance and treatment device, in combination with a computer as hardware.

[0017] According to one embodiment, a device for assisting and treating a mental illness includes a memory and at least one processor, and the processor is configured to receive passive data and first active data from a user device, extract patient behavioral characteristic data from the passive data, predict second active data of the patient using a subjective emotional state prediction model having the behavioral characteristic data as input, output a prediction value indicating the presence or absence of a mental illness using a mental illness scale classification model having the behavioral characteristic data and the second active data as input, and provide a user interface for displaying the prediction value indicating the presence or absence of a mental illness to the user.

[0018] According to one embodiment, a mental illness diagnosis assistance and treatment system may include a user device that collects passive data, first active data, and second active data from a patient, and a processor that receives the passive data and the first active data from the user device, extracts patient behavioral characteristic data from the passive data, predicts the second active data of the patient using a subjective emotional state prediction model having the behavioral characteristic data as an input, outputs a prediction value representing a mental illness scale using a mental illness scale classification model having the behavioral characteristic data and the second active data as an input, and provides a user interface that displays the prediction value representing a mental illness scale to the user.

[0019] According to the embodiments, by continuously collecting quantitative and objective data related to the patient's mental illness and utilizing it in real time, the patient can be accurately understood and evaluated.

[0020] According to embodiments, the patient's passive data, which can be easily collected without the patient's passive input, can be used to continuously predict the patient's active data without the problem of decreased compliance.

[0021] According to embodiments, by analyzing behavioral characteristic data highly related to mental health from passive data, more accurate patient behavioral characteristics can be obtained, particularly by extracting activity complexity from GPS coordinates.

[0022] According to the embodiments, a mental illness scale for each patient can be predicted with high accuracy using an artificial intelligence model from the patient's behavioral characteristic data.

[0023] According to embodiments, the most suitable recommended treatment content or recommended treatment solution can be provided to the patient based on the patient's passive data, behavioral characteristic data, etc.

[0024] The effects according to the present disclosure are not limited to the effects described above, and effects not mentioned can be clearly understood by a person having ordinary skill in the art to which the present disclosure pertains from this specification and the attached drawings.

[0025] FIG. 1 is a conceptual diagram illustrating a method for assisting in the diagnosis of mental illness according to one comparative example and one embodiment of the present disclosure.

[0026] FIG. 2 is a block diagram of a mental illness diagnosis assistance and treatment system according to one embodiment of the present disclosure.

[0027] FIG. 3 is a flowchart illustrating the operation of a mental illness diagnosis assistance and treatment device according to one embodiment of the present disclosure.

[0028] FIG. 4 is a conceptual diagram of the operation of a subjective emotional state prediction model according to one embodiment of the present disclosure.

[0029] Figure 5 is a conceptual diagram of the operation of a mental illness scale classification model according to one embodiment of the present disclosure.

[0030] Figures 6 to 9 are graphs showing an example of a method for extracting behavioral characteristic data from passive data.

[0031] FIG. 10 is a graph showing output data of a subjective emotional state prediction model according to one embodiment of the present disclosure.

[0032] FIGS. 11 to 14 illustrate user interface screens providing mental illness diagnosis assistance and treatment services according to one embodiment of the present disclosure.

[0033] Specific structural or functional descriptions of embodiments according to the concept of the present invention disclosed in this specification are merely illustrative for the purpose of explaining embodiments according to the concept of the present invention, and embodiments according to the concept of the present invention may be implemented in various forms and are not limited to the embodiments described in this specification.

[0034] Embodiments according to the concept of the present invention may have various modifications and take various forms, and thus, embodiments are illustrated in the drawings and described in detail in this specification. However, this is not intended to limit embodiments according to the concept of the present invention to specific disclosed forms, but rather includes modifications, equivalents, or alternatives that fall within the spirit and technical scope of the present invention.

[0035] While terms such as "first" or "second" may be used to describe various components, these components should not be limited by these terms. These terms are intended solely to distinguish one component from another. For example, a first component may be referred to as a "second component," and similarly, a second component may also be referred to as a "first component," without departing from the scope of the invention.

[0036] When a component is referred to as being "connected" or "connected" to another component, it should be understood that it may be directly connected or connected to that other component, but that there may be other components in between. Conversely, when a component is referred to as being "directly connected" or "directly connected" to another component, it should be understood that there are no other components in between. Expressions that describe relationships between components, such as "between," "immediately between," or "directly adjacent to," should be interpreted similarly.

[0037] The terminology used herein is for the purpose of describing specific embodiments only and is not intended to limit the present invention. The singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, it should be understood that the terms "comprises" or "has" are intended to specify the presence of a described feature, number, step, operation, component, part, or combination thereof, but do not preclude the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0038] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. Terms defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and will not be interpreted in an idealized or overly formal sense unless explicitly defined herein.

[0039] The processor in this specification may mean hardware capable of performing functions and operations according to each name described in this specification, may mean computer program code capable of performing specific functions and operations, or may mean an electronic recording medium loaded with computer program code capable of performing specific functions and operations.

[0040] In other words, a processor may mean a functional and / or structural combination of hardware for performing the technical idea of ​​the present invention and / or software for driving the hardware.

[0041] Hereinafter, embodiments will be described in detail with reference to the attached drawings. However, the scope of the patent application is not limited or restricted by these embodiments. The same reference numerals in each drawing represent the same components.

[0042] FIG. 1 is a conceptual diagram illustrating a method for assisting in the diagnosis of mental illness according to one comparative example and one embodiment of the present disclosure.

[0043] Referring to Fig. 1, a method for assisting diagnosis of mental illness according to a comparative example of the present disclosure can be conducted simultaneously with a questionnaire-based scale test (2) as a diagnostic aid tool for mental illness when a patient undergoes a professional consultation test (1) through outpatient treatment at a certain point in time, and then undergoes a professional consultation (3) through outpatient treatment at a certain point in time N months later. The patient's self-report used in the method for assisting diagnosis of mental illness according to a comparative example of the present disclosure is inaccurate and does not reflect the patient's unconsciousness. Due to incorrect symptom identification, a doctor may make an inaccurate diagnosis of the patient's mental health.

[0044] According to a method for assisting in the diagnosis of mental illness according to one embodiment of the present disclosure, digital data (4) regarding a patient's daily life can be collected via a smartphone, wearable device, or the like. Here, the digital data (4) may be referred to as passive data or digital life log data. In the present disclosure, the digital data (4) refers to objective and quantitative patient behavioral data that can be automatically generated and / or collected without the patient's intervention.

[0045] In addition, according to a method for assisting diagnosis of mental illness according to one embodiment of the present disclosure, the collected digital data (4) of the patient can be analyzed to automatically predict the mental health scale of the patient, and the predicted mental health scale can be used as an assisting tool when a user of the method for assisting diagnosis of mental illness (e.g., a psychiatrist) diagnoses whether the patient has a mental illness. The device for assisting diagnosis of mental illness can analyze the collected digital data to analyze the patient's lifestyle pattern, and teach the analyzed lifestyle pattern and the patient's mental illness evaluation score to an artificial intelligence model.

[0046] According to a method for assisting the diagnosis of mental illness, collected digital data (4) of a patient and actual questionnaire-style scale tests (2) already performed by the patient over a certain period of time can be analyzed, and a regular status report (5) can be generated using a learned artificial intelligence model. In one embodiment, the learned artificial intelligence model can be used to predict active data corresponding to the patient's actual questionnaire-style scale test (2). Here, the learned artificial intelligence model can be a subjective emotional state prediction model that uses behavioral characteristic data as input. The predicted active data can refer to response data of a self-reported questionnaire performed by the patient, or a total sum score of the response data. The self-reported questionnaire performed by the patient can include a 'mental health scale test'. Furthermore, in one embodiment, the learned artificial intelligence model can be used to predict mental health (e.g., anxiety, depression) scales. For example, the learned artificial intelligence model can output 0 or 1 as a predicted value representing a depression scale. A depression scale of 0 can indicate a patient within the normal range, and a depression scale of 1 can indicate a patient at risk for depression. Here, the learned artificial intelligence model may be a mental illness scale classification model that takes behavioral characteristic data and predicted active data as input.

[0047] A method for assisting in the diagnosis of mental illness according to one embodiment of the present disclosure is a method for collecting passive data and / or active data according to ecological momentary assessment (EMA), that is, a method of repeatedly sampling and recording the patient's current behavior and experience in real time in the patient's daily life environment, thereby enabling a more accurate understanding of the patient and a mental health-related evaluation, and further, a diagnosis of mental illness.

[0048] Therefore, according to a method for assisting in the diagnosis of mental illness according to one embodiment of the present disclosure, the patient's active data can be continuously predicted without the problem of decreased compliance by using the patient's digital data (4) that can be easily collected without the patient's passive input after the present time based on the active data and passive data for the past 2 to 4 weeks before the patient's compliance declines.

[0049] FIG. 2 is a block diagram of a mental illness diagnosis assistance and treatment system according to one embodiment of the present disclosure.

[0050] The mental illness diagnosis assistance and treatment system may include a mental illness diagnosis assistance and treatment device (100), a user device (102), and / or a database (104).

[0051] According to one embodiment, a mental illness diagnosis assistance and treatment device (100) in a mental illness diagnosis assistance and treatment system may receive passive data and / or active data from a patient. The mental illness diagnosis assistance and treatment device (100) may analyze the patient's behavioral characteristic data from the received passive data. At this time, the mental illness diagnosis assistance and treatment device (100) may collect passive data for a certain period of time (e.g., 2 weeks, 4 weeks) or collect active data performed by the patient every day for a certain period of time (e.g., 2 weeks, 4 weeks). The active data received from the patient for a certain period of time may be referred to as 'first active data' herein.

[0052] A mental illness diagnosis assistance and treatment device (100) can predict a patient's active data using a subjective emotional state prediction model (12) learned from behavioral characteristic data and active data. The patient's active data, which is the output value of the learned subjective emotional state prediction model (12), may be referred to as "second active data" in this specification.

[0053] The mental illness diagnosis assistance and treatment device (100) may be implemented as a printed circuit board (PCB) such as a motherboard, an integrated circuit (IC), or a system on chip (SoC). For example, the mental illness diagnosis assistance and treatment device (100) may be implemented as an application processor.

[0054] Additionally, the mental illness diagnosis assistance and treatment device (100) and / or user device (102) may be implemented in a personal computer (PC), server, or portable device.

[0055] The portable device may be implemented as a laptop computer, a mobile phone, a smart phone, a tablet PC, a mobile internet device (MID), a personal digital assistant (PDA), an enterprise digital assistant (EDA), a digital still camera, a digital video camera, a portable multimedia player (PMP), a personal navigation device (PND), a handheld game console, an e-book, or a smart device. The smart device may be implemented as a smart watch, a smart band, or a smart ring.

[0056] In one embodiment, the method for assisting and treating mental illness may be operated through a user device (102) in the form of an application or program. The application for assisting and treating mental illness may be an application running on a PC or a portable device. The application for assisting and treating mental illness may display a user interface (UI) or a graphical user interface (GUI) according to information processed by the device for assisting and treating mental illness (100) and / or the user device (102) through a display, or information processed by the device for assisting and treating mental illness (100) and / or the user device (102) through a display. The information processed by the device for assisting and treating mental illness (100) may include information on passive data, active data, and behavioral characteristic data. The application may include an interface linked with an existing hospital operating system.

[0057] The database (104) may be a database connected to the mental illness diagnosis assistance and treatment device (100) via wired or wireless means. In one embodiment, the database (104) may be a hospital database including basic information and / or clinical information of a patient, and / or a model database including learning data and preprocessing data. In one embodiment, the database (104) may be a cloud database storing passive data, active data, and / or data related to passive data or active data of a plurality of patients.

[0058] The mental illness diagnosis assistance and treatment device (100) includes a processor (10) and / or a memory (20). Although not shown in the drawing, the mental illness diagnosis assistance and treatment device (100) may further include a transmitter / receiver and / or an interface.

[0059] The processor (10) can extract patient behavioral characteristic data from passive data received from a user device. The processor (10) can use an artificial intelligence model to predict patient-specific active data or patient-specific mental illness scales from the patient's behavioral characteristic data.

[0060] The processor (10) may include a subjective emotional state prediction model (12). The subjective emotional state prediction model (12) may include a boosting-based model. For example, the subjective emotional state prediction model (12) may include an XGBoost regression algorithm, but it will be understood by those skilled in the art that the subjective emotional state prediction model (12) may include any boosting-based model, but is not limited thereto. The processor (10) may predict patient-specific active data using the subjective emotional state prediction model (12) including the XGBoost algorithm. Details regarding the subjective emotional state prediction model (12) will be described later with reference to FIG. 4.

[0061] The processor (10) may include a mental illness scale classification model (14). The mental illness scale classification model (14) may include a boosting-based model. For example, the mental illness scale classification model (14) may include an XGBoost classification algorithm, but it will be understood by those skilled in the art that the mental illness scale classification model (14) may include any boosting-based model, but is not limited thereto. The processor (10) may predict a mental illness scale for each patient using the mental illness scale classification model (14) including the XGBoost algorithm. Details regarding the mental illness scale classification model (14) are described below with reference to FIG. 5.

[0062] The processor (10) can provide recommended treatment content or recommended treatment solution to the patient based on at least one of the patient's passive data, behavioral characteristic data, second active data, or a predicted value representing a mental illness scale.

[0063] Recommended treatment content may be content for mental health management or treatment of mental illness appropriate for the patient's condition, as determined by the patient's passive data, behavioral characteristic data, secondary active data, or predicted values ​​representing mental illness scales. For example, recommended treatment content may include classical music content, content for meditation programs (e.g., music content, book content, video content, lecture material content), and / or content for electronic drug therapy.

[0064] Recommended treatment solutions can be solutions for mental health management or treatment of mental illness, which can be determined based on the patient's passive data, behavioral characteristic data, secondary active data, or predictive values ​​representing mental illness scales. For example, in the case of a treatment solution for a patient receiving home treatment, a treatment solution that utilizes digital phenotypes and self-reported items collected from the patient can be provided as a recommended treatment solution to assist in the rehabilitation and recovery of the patient receiving home treatment by providing a solution that includes a step-by-step intervention, a step for early intervention depending on the situation, and a step for conducting non-face-to-face face-to-face consultations in emergency cases.

[0065] Depending on the embodiment, recommended treatment content or recommended treatment solutions may be provided to patients in the form of applications or programs. Specific examples of recommended treatment content or recommended treatment solutions in the form of applications or programs are described below in Figure 14.

[0066] In some embodiments, the mental illness diagnosis assistance and treatment device may further include a recommendation module for providing recommended treatment content or recommended treatment solutions to the patient. The recommendation module is controlled by a processor (10) and may output recommended treatment content or recommended treatment solutions most suitable for the patient using an artificial intelligence model trained using the patient's passive data, behavioral characteristic data, first active data, second active data, etc.

[0067] The processor (10) can process data stored in the memory (20). The processor (10) can execute computer-readable code (e.g., software) stored in the memory (20) and instructions generated by the processor (10).

[0068] A processor (10) may be a data processing device implemented as hardware having a circuit with a physical structure for executing desired operations. For example, the desired operations may include code or instructions included in a program.

[0069] For example, a data processing device implemented in hardware may include a microprocessor, a central processing unit, a processor core, a multi-core processor, a multiprocessor, an application-specific integrated circuit (ASIC), or a field programmable gate array (FPGA).

[0070] The memory (20) can store passive data and / or active data received from the patient. For example, the memory (20) can store GPS coordinates, number of steps, sleep start time, sleep end time, call start time, call end time, type of application (APP) in use, and / or usage time of the APP received from the patient's user device (102). The memory (20) can store all or part of the subjective emotional state prediction model (12), all or part of the mental illness scale classification model (14), input values ​​or output values ​​of the subjective emotional state prediction model (12), input values ​​or output values ​​of the mental illness scale classification model (14), and variables required for calculating the subjective emotional state prediction model (12).

[0071] The memory (20) can store instructions (or programs) executable by the processor (10). For example, the instructions may include instructions for executing operations of the processor and / or operations of each component of the processor.

[0072] The memory (20) can be implemented as a volatile memory device or a non-volatile memory device.

[0073] Volatile memory devices can be implemented as dynamic random access memory (DRAM), static random access memory (SRAM), thyristor RAM (T-RAM), zero capacitor RAM (Z-RAM), or twin transistor RAM (TTRAM).

[0074] The nonvolatile memory device can be implemented as an Electrically Erasable Programmable Read-Only Memory (EEPROM), flash memory, Magnetic RAM (MRAM), Spin-Transfer Torque (STT)-MRAM, Conductive Bridging RAM (CBRAM), Ferroelectric RAM (FeRAM), Phase change RAM (PRAM), Resistive RAM (RRAM), Nanotube RRAM, Polymer RAM (PoRAM), Nano Floating Gate Memory (NFGM), holographic memory, Molecular Electronic Memory Device, or Insulator Resistance Change Memory.

[0075] FIG. 3 is a flowchart illustrating the operation of a mental illness diagnosis assistance and treatment device according to one embodiment of the present disclosure.

[0076] Referring to FIG. 3, the mental illness diagnosis assistance and treatment device (100) can receive passive data and first active data from the user device (102) (S302).

[0077] Passive data may include information about one or more of GPS coordinates received from the user device, acceleration sensor data, tilt sensor data, screen ON time, screen OFF time, call start time, call end time, type of app being used, and usage time of the app.

[0078] The first active data may refer to response data from a self-report questionnaire (e.g., a mental health scale test) administered by the patient, or the total sum score of the response data. The self-report questionnaire administered by the patient may include a "mental health scale test." The mental health scale test may be a four-item test with responses ranging from 1 to 7.

[0079] Here, the passive data received from the user device (102) may be passive data of the patient received through the user device over a certain period of time (e.g., 2 weeks, 4 weeks). The first active data received from the user device (102) may be active data corresponding to the total sum score of response data of a self-report questionnaire performed by the patient once a day over a certain period of time (e.g., 2 weeks, 4 weeks).

[0080] The mental illness diagnosis assistance and treatment device (100) can extract patient behavioral characteristic data from passive data (S304).

[0081] The behavioral characteristic data may include information on at least one of visited locations, time spent at each visited location, number of steps, distance traveled, movement speed, distance walked, walking speed, activity complexity, and sleep time.

[0082] In one embodiment, the mental illness diagnosis assistance and treatment device (100) can extract the patient's activity complexity data from passive data. The activity complexity data may be referred to as pattern mismatch time. The mental illness diagnosis assistance and treatment device (100) can collect the patient's location data for all dates and all time periods during a certain period. The mental illness diagnosis assistance and treatment device (100) can cluster the location data at specific times on each date and select the first location data that is most frequently located at the specific time. The mental illness diagnosis assistance and treatment device (100) can generate a repetitive life pattern graph based on the first location data at each time. The mental illness diagnosis assistance and treatment device (100) can compare the repetitive life pattern graph with the life pattern graph and calculate the time during which the life pattern graph is inconsistent with the pattern of the repetitive life pattern graph. The mental illness diagnosis assistance and treatment device (100) can calculate the average of the time during a certain period during which the pattern is inconsistent.

[0083] The mental illness diagnosis assistance and treatment device (100) can predict the patient's second active data using a subjective emotional state prediction model (12) that inputs behavioral characteristic data (S306).

[0084] The subjective emotional state prediction model (12) may include a boosting-based algorithm. The subjective emotional state prediction model (12) may be trained using a feature subset determined based on RFE (Recursive Feature Elimination) among the above-described behavioral characteristic data. RFE is a method of feature selection, which means a method of repeating the process of removing features with low feature importance until a desired number of features remains. The desired number of features may be the number of features included in the feature subset in the subjective emotional state prediction model (12) according to an embodiment of the present disclosure. Here, feature importance means the degree to which each feature contributes to the subjective emotional state prediction model. In addition, feature selection means a process of selecting appropriate features necessary for learning for the accuracy or speed of the model.

[0085] In one embodiment, the feature subset may consist of steps, distance traveled, places visited, and sleep time. In another embodiment, the feature subset may consist of distance traveled, speed, distance walked, and walking speed. In yet another embodiment, the feature subset may consist of pattern mismatch time (i.e., activity complexity), time spent at home, day of the week, and whether it is a weekday or weekend.

[0086] Secondary active data may be the predicted total score for items answered by the patient on a self-report questionnaire (e.g., a mental health scale). For example, this could be the predicted total score (e.g., 25) for a patient on a four-item mental health scale, with responses ranging from 1 to 7.

[0087] A mental illness diagnosis assistance and treatment device (100) can output a predicted value representing a mental illness scale by using a mental illness scale classification model (14) that inputs behavioral characteristic data and second active data (S308).

[0088] The mental illness scale classification model (14) may include a boosting-based classification algorithm. The mental illness scale classification model (14) may be trained using passive data or second active data augmented using the synthetic minority oversampling technique (SMOTE).

[0089] A mental illness diagnosis assistance and treatment device (100) can provide a user interface that displays a predicted value representing a mental illness scale to the user (S310).

[0090] A mental illness diagnosis assistance and treatment device (100) can predict a patient's active data based on the total score of a mental health scale test for a certain period predicted from a subjective emotional state prediction model (12), and can diagnose the patient's mental illness in real time and predict the prognosis based on the mental illness scale predicted using a mental illness scale classification model (14). Furthermore, the mental illness diagnosis assistance and treatment device (100) can provide the patient with recommended treatment content or recommended treatment solution based on at least one of the patient's passive data, behavioral characteristic data, second active data, or predicted value representing the mental illness scale.

[0091] FIG. 4 is a conceptual diagram of the operation of a subjective emotional state prediction model according to one embodiment of the present disclosure.

[0092] The subjective emotional state prediction model (43) of Fig. 4 can be applied to the subjective emotional state prediction model (12) of Fig. 2. Hereinafter, it will be described with reference to Fig. 2.

[0093] Referring to Fig. 4, the subjective emotional state prediction model (43) can use the patient's behavioral characteristic data (42) analyzed from passive data (41) as input data.

[0094] Passive data (41) is digital data that is digitized and quantified in devices such as smartphones and wearable devices, and can be included in digital data that is automatically collected without patient intervention. Passive data (41) related to assisted diagnosis and treatment of mental illness according to an embodiment of the present disclosure may include one or more of GPS coordinates, acceleration sensor data, tilt sensor data, screen ON time, screen OFF time, call start time, call end time, type of app (APP) being used, and app usage time.

[0095] A mental illness diagnosis assistance and treatment device (100) or processor (10) can analyze the patient's behavioral characteristic data (42) from passive data (41).

[0096] The behavioral characteristic data (42) may be data representing the behavioral characteristics of the patient's daily life, such as the patient's average activity pattern, activity regularity, time spent at each location, and sleep time. The behavioral characteristic data (42) may be data analyzed from passive data, a digital phenotype, or at least one of the passive data. In one embodiment, the behavioral characteristic data (42) may include the location stayed, activity complexity, movement speed, movement distance, visited locations, daily sleep time, and sleep irregularity. Specific details of analyzing the patient's behavioral characteristic data (42) from the passive data (41) are described below with reference to FIGS. 6 to 9.

[0097] A subjective emotional state prediction model (43) according to one embodiment of the present disclosure may include a boosting-based machine learning model, such as an XGBoost (Extreme Gradient Boosting) regression algorithm. The XGBoost regression algorithm may be implemented as a GBM (Gradient Boosting Machine) model capable of parallel learning.

[0098] According to one embodiment of the present disclosure, a subjective emotional state prediction model (43) can learn the relationship between passive data (41) automatically collected by a user device and active data collected from a patient.

[0099] The active data collected from patients may be collected through a four-item questionnaire with responses ranging from 1 to 7. By applying item response theory, which reflects the forgetting curve, to a simplified four-item diagnostic questionnaire (e.g., a mental health scale test), the accuracy of the active data collected from patients can be improved. In one embodiment, a subjective emotional state prediction model (43) can be trained using digital data (4) and active data collected over a two-week period.

[0100] The subjective emotional state prediction model (43) can output active data (44). In one embodiment, the active data (44) output from the subjective emotional state prediction model (43) can be a predicted total score of a survey (e.g., a mental health scale test) consisting of four items that can be answered from 1 to 7 points. The subjective emotional state prediction model (43) is trained using digital data (4) and active data collected over a certain period of time (e.g., 2 weeks), and then can predict the total score of the survey over a certain period of time (e.g., 1 week).

[0101] As a result of the performance evaluation of the subjective emotional state prediction model (43) according to one embodiment of the present disclosure, the Mean RMSE (Root Mean Square Error) was measured as 3±0.2, and the Mean R-Squared was measured as -2.2.

[0102] FIG. 5 is a conceptual diagram of the operation of a mental illness scale classification model according to one embodiment of the present disclosure.

[0103] The mental illness scale classification model (53) of FIG. 5 can be applied to the mental illness scale classification model (14) of FIG. 2, and the passive data (51), behavioral characteristic data (52), and active data (54) of FIG. 5 can correspond to the passive data (41), behavioral characteristic data (41), and active data (44) of FIG. 4, respectively. Hereinafter, the description will be made with reference to FIG. 2 and FIG. 4 together.

[0104] Referring to FIG. 5, the mental illness scale classification model (53) can use patient behavioral characteristic data (52) analyzed from passive data (51) as input data. In addition, the mental illness scale classification model (53) can use active data (54) as input data. Here, the active data (54) can be active data (44) output from the subjective emotional state prediction model (43), i.e., second active data. The mental illness scale classification model (53) can use the behavioral characteristic data (52) and the active data (44) output from the subjective emotional state prediction model (43) as input data.

[0105] As another example, a mental illness scale classification model (53) may use patient behavioral characteristic data (52) analyzed from passive data (51) and first active data as input data. The first active data may refer to response data from a self-report questionnaire (e.g., a mental health scale test) actually performed by the patient, or the total sum score of the response data.

[0106] Passive data (51) and behavioral characteristic data (52) are the same or similar concepts as the passive data (41) and behavioral characteristic data (52) of FIG. 4, so the description below is omitted within the overlapping scope.

[0107] A mental illness scale classification model (53) according to one embodiment of the present disclosure may include a boosting-based machine learning model, such as an XGBoost (Extreme Gradient Boosting) classification algorithm. The XGBoost classification algorithm may be implemented as a GBM (Gradient Boosting Machine) model capable of parallel learning.

[0108] According to one embodiment of the present disclosure, a mental illness scale classification model (53) can learn the relationship between passive data (41) automatically collected by a user device (102) and active data (54) predicted from a subjective emotional state prediction model (43).

[0109] The mental illness scale classification model (53) can output a mental illness scale prediction value (55).

[0110] In one embodiment, the predicted value (55) representing the mental illness scale output from the mental illness scale classification model (53) may be a predicted value representing a depression scale or a predicted value representing an anxiety scale. The predicted value (55) representing the mental illness scale may be 0 or 1 based on the cutoff score of the mental illness scale. For example, if the mental illness scale is PHQ-9, which is one of the depression scales, if the score of PHQ-9 is calculated to be less than or equal to the cutoff score of 30, 0 may be output as the predicted value, and if the score of PHQ-9 is calculated to be more than the cutoff score of 30, 1 may be output as the predicted value. The depression scale may be PHQ-9 or CESDR, and the anxiety scale may be GAD-7, but the present invention is not limited thereto, and the mental illness scale classification model (53) may be applied to predict scales for various mental illnesses.

[0111] As a result of evaluating the performance of a mental illness scale classification model (53) that outputs an anxiety scale (GAD-7) using behavioral characteristic data (52) and the first active data as input data, the Mean Accuracy was measured as 0.87503, the Mean Precision as 0.846822, the Mean F1 Score as 0.878034, and the Mean ROC-AUC as 0.947422.

[0112] As a result of evaluating the performance of a mental illness scale classification model (53) that outputs a depression scale (PHQ-9) using behavioral characteristic data (52) and the first active data as input data, the Mean Accuracy was measured as 0.946156, the Mean Precision as 0.920763, the Mean F1 Score as 0.945297, and the Mean ROC-AUC as 0.98447.

[0113] As a result of evaluating the performance of a mental illness scale classification model (53) that outputs an anxiety scale (GAD-7) using behavioral characteristic data (52) and second active data as input data, the Mean Accuracy was measured as 0.886533, the Mean Precision as 0.872277, the Mean F1 Score as 0.88892, and the Mean ROC-AUC as 0.950104.

[0114] As a result of evaluating the performance of a mental illness scale classification model (53) that outputs a depression scale (PHQ-9) using behavioral characteristic data (52) and second active data as input data, the Mean Accuracy was measured as 0.932701, the Mean Precision as 0.89645, the Mean F1 Score as 0.932692, and the Mean ROC-AUC as 0.97625.

[0115] Figures 6 to 9 are graphs illustrating an example of a method for extracting behavioral characteristic data from passive data. The following description is provided with reference to Figures 2, 4, and 5.

[0116] A mental illness diagnosis assistance and treatment device (100) or processor (10) can extract patient behavioral characteristic data (42) from passive data (41).

[0117] The mental illness diagnosis assistance and treatment device (100) can extract the number of locations the patient moved to each day from GPS coordinates collected at 5-minute intervals. For example, the number of moved locations can be extracted using DB-SCAN (Density-based Spatial Clustering of Applications with Noise), which estimates significant moved locations using the density of GPS coordinates. In this process, meaningless GPS coordinates can be excluded. The mental illness diagnosis assistance and treatment device (100) can calculate the average number of moved locations per patient based on the number of moved locations.

[0118] As another example, the mental illness diagnosis assistance and treatment device (100) can extract patient-specific walking-related characteristic data, such as average daily walking time, average walking speed, and total daily steps, from the number of steps collected at 5-minute intervals.

[0119] For example, data on ecologically impractical step counts collected at 5-minute intervals, and data on distance traveled at 5-minute intervals that cannot be walked, can be included. Passive data such as step counts and distance traveled collected at 5-minute intervals can be used to calculate distance walked, walking speed, or walking time, etc. by excluding the speed of transportation used and only using data that are determined to be walked, using the average walking speed of adults.

[0120] The mental illness diagnosis assistance and treatment device (100) can calculate the movement speed from the number of steps collected at 5-minute intervals and extract characteristic data on the time spent at each location. The mental illness diagnosis assistance and treatment device (100) can perform primary clustering based on the movement speed by date, excluding the time spent moving between locations. The mental illness diagnosis assistance and treatment device (100) can extract major locations stayed at, which can be referred to as secondary clustering. The mental illness diagnosis assistance and treatment device (100) can calculate the time spent at major locations by date and calculate the sum of the time spent at each major location. In addition, the same locations can be clustered for all dates during a certain period. This can be referred to as tertiary clustering. The mental illness diagnosis assistance and treatment device (100) can quantify the time spent at each location after the tertiary clustering. A device (100) for assisting and treating mental illness according to one embodiment of the present disclosure can perform repeated clustering to calculate specific locations and the time spent at those locations in detail. For example, according to a method for assisting and treating mental illness according to one embodiment of the present disclosure, a home and a coffee shop located directly across from the home can be extracted as separate locations, and the time spent at each location can be calculated.

[0121] As another example, the mental illness diagnosis assistance and treatment device (100) can extract patient-specific sleep-related characteristic data, such as daily sleep time and sleep time distribution (i.e., sleep irregularity), from sleep start time and sleep end time collected at 5-minute intervals.

[0122] Referring to FIGS. 6 to 9, a method for extracting activity complexity from behavioral characteristic data from GPS coordinates is described as follows.

[0123] The mental illness diagnosis assistance and treatment device (100) can calculate the average time deviating from a patient-specific repetitive pattern based on collected GPS coordinates. In the present disclosure, the average time deviating from a patient-specific repetitive pattern is defined as "activity complexity." "Time deviating from a pattern" refers to the time that is inconsistent with the pattern.

[0124] The mental illness diagnosis assistance and treatment device (100) can cluster patient locations by date and time zone during a given period. For example, the GPS coordinates of patient 117's location at 3:00 PM on all dates over a two-week period can be collected, and only the location data at 3:00 PM on all dates can be collected and clustered.

[0125] The mental illness diagnosis assistance and treatment device (100) can cluster location data at a specific time on each day during a given period, and select the first location that is located the most at the specific time. The location that is visited the most at the specific time can be referred to as the 'first location'. For example, the locations of patients at 15:00 on each day for two weeks can be represented as in the graph of FIG. 6. The mental illness diagnosis assistance and treatment device (100) can select the location or cluster on the graph that the patient visited the most at 15:00 during the given two weeks. The selected location (61) is the location that was visited the most, 13 times out of 14 times, and has an X coordinate of 35.573233 and a Y coordinate of 129.189264. In FIG. 6, the first location at 15:00 has an X coordinate of 35.573233 and a Y coordinate of 129.189264.

[0126] Referring to FIG. 7, the mental illness diagnosis assistance and treatment device (100) can determine a location for each time zone based on the first location for each time zone.

[0127] For example, the first position at 15 o'clock selected in Fig. 6 has an X coordinate of 35.573233 and a Y coordinate of 129.189264. The mental illness diagnosis assistance and treatment device (100) can determine the position for each time zone based on the first position at 0 o'clock to the first position at 24 o'clock.

[0128] Referring to FIG. 8, the mental illness diagnosis assistance and treatment device (100) can remove irregularities in the first location data for each time zone using a data smoothing technique and perform repetitive pattern generalization to generate a repetitive life pattern graph. The first location data can include information about the first location, for example, information about the X-coordinate and Y-coordinate of the first location.

[0129] Referring to FIG. 9, the mental illness diagnosis assistance and treatment device (100) can compare a repetitive life pattern graph (82) for a certain period of time with a life pattern graph (81) on a specific date, and calculate the time at which the life pattern graph (81) on a specific date does not match the pattern of the repetitive life pattern graph (82).

[0130] A method of comparing a repetitive life pattern graph (82) for a certain period of time and a life pattern graph (81) for a specific date may include a method of setting the error range of the repetitive life pattern to 15% when the specific date is May 25, 2023 and comparing whether it matches the life pattern graph (81) for May 25, 2023. Assuming that the specific date is May 25, 2023 and the certain period of time is two weeks, the time (6) at which the life pattern graph (81) for May 25, 2023 matches the pattern of the repetitive life pattern graph (82) for two weeks and the time (7) at which the life pattern graph (81) for May 25, 2023 does not match the pattern of the repetitive life pattern graph (82) for two weeks can be calculated.

[0131] In one embodiment, it is assumed that the two weeks for extracting the patient's activity complexity data from the passive data (41) are from May 14, 2023 to May 27, 2023. The mental illness diagnosis assistance and treatment device (100) can compare the repetitive life pattern graph (82) and the life pattern graph (81) for each date corresponding to 14 days from May 14 to May 27, and calculate the time for which the life pattern graph (81) does not match the pattern of the repetitive life pattern graph (82). For example, the time for which the pattern does not match may be 17.416667 hours on May 18, 2023, and the time for which the pattern does not match may be 9.166667 hours on May 25, 2023.

[0132] The mental illness diagnosis assistance and treatment device (100) can calculate the average of the times that are inconsistent with the above pattern over a certain period of time.

[0133] In one embodiment, the mental illness diagnosis assistance and treatment device (100) may calculate the average of the number of times inconsistent with the pattern over a two-week period. For example, the average of the number of times inconsistent with the pattern from May 14, 2023 to May 27, 2023 may be calculated.

[0134] The average amount of time that is inconsistent with the pattern of the repetitive life pattern graph over a period of time can be defined as 'activity complexity'.

[0135] A method for assisting in the diagnosis and treatment of mental illness according to one embodiment of the present disclosure extracts activity patterns for each patient and then compares them with repetitive patterns within the activities of each patient, thereby extracting more accurate behavioral characteristic data by taking into account differences in activity patterns between individuals.

[0136] FIG. 10 is a graph showing output data of a subjective emotional state prediction model according to one embodiment of the present disclosure.

[0137] Referring to the graph of FIG. 10, the X-axis represents a date, and the Y-axis represents output data of a subjective emotional state prediction model (43) according to an embodiment of the present disclosure. The output data of the subjective emotional state prediction model (43) may be a predicted value of the total score of the items answered by the patient in the mental health scale test. The graph corresponding to the first label is a graph representing the total score calculated based on the responses to the mental health scale test actually performed by the patient every day from March 25, 2023 to April 21, 2023. The graph corresponding to the second label is a graph representing the output data of the subjective emotional state prediction model (43), i.e., the total score that the patient is expected to obtain in the mental health scale test by date.

[0138] FIGS. 11 to 14 illustrate user interface screens providing mental illness diagnosis assistance and treatment services according to one embodiment of the present disclosure.

[0139] Referring to FIGS. 11 to 14, examples of user interface screens that provide mental illness diagnosis assistance and treatment services are shown on a user device (102) running a mobile application.

[0140] Referring to FIG. 11, the mental illness diagnosis assistance and treatment device (100) can provide a patient with a user interface through a user device (102) in which passive data (41) and / or behavioral characteristic data (42) including the number of steps, sleeping time, and moving distance are displayed.

[0141] Referring to FIG. 12, behavioral characteristic data (42) highly related to mental health (e.g., time spent at home) can be analyzed, and a user interface providing related data can be provided to the patient. For example, data related to behavioral characteristic data (42) highly related to mental health, such as daily time spent at home, average value over a certain period, weekday average value, and weekend average value, can be provided. Therefore, according to a mental illness diagnosis assistance and treatment service according to one embodiment of the present disclosure, behavioral characteristic data (42) can be analyzed through the collection of passive data (41).

[0142] Referring to FIG. 13, the mental illness diagnosis assistance and treatment device (100) can provide the patient with user interfaces for conducting a mental health scale test through the user device (102).

[0143] The user interfaces for conducting the mental health scale test may include one or more of a user interface (120) for starting the test, a user interface (122) for adding symptoms of interest, a user interface (124) for responding to questions included in the mental health scale test, and a user interface (126) for recording additional information about the patient's condition.

[0144] The Mental Health Scale test according to the embodiment of the present disclosure is based on a novel scale simplified to four items by applying the Ecological Momentary Assessment (EMA) concept. These four items may include questions about mood, appetite, sleep, and condition. Mental health diagnostic assistance and treatment services can provide the Mental Health Scale test to patients so that they can be administered daily. In addition to the four main items, the Mental Health Scale test allows for additional recording of symptoms of interest by selecting them and allowing patient diary recording, thereby compensating for the limited number of questions in the Mental Health Scale test while ensuring expandability.

[0145] Referring to FIG. 14, the mental illness diagnosis assistance and treatment device (100) can provide user interfaces for providing recommended treatment content or recommended treatment solutions to the patient through the user device (102).

[0146] Referring to Figure 14, user interfaces for providing a depression or anxiety treatment solution for perinatal women are illustrated.

[0147] User interfaces for providing a depression or anxiety treatment solution for perinatal women may include at least one of a user interface (140) for starting a scale test for perinatal women before treatment, a user interface (142) for providing various recommended treatment contents for objectively observing one's own state of mind, and a user interface (144) for starting a scale test for perinatal women after digital treatment with recommended treatment contents.

[0148] In a user interface (142) that provides various recommended treatment contents for objectively observing one's own state of mind, mindfulness contents corresponding to 1 to 4 sessions may be provided as recommended treatment contents. In an embodiment, when recommending recommended treatment contents or recommended treatment solutions using a learned artificial intelligence model, a pre-treatment scale test and a post-digital treatment scale test with recommended treatment contents may be used as learning data, and a content or solution predicted to have a large change value in a mental health scale or mental illness scale may be determined as a recommended treatment content or recommended treatment solution using a learned artificial intelligence model.

[0149] A mental illness diagnosis assistance and treatment device (100) according to one embodiment may be combined with a computer or computing device as hardware and may include a computer program stored in a computer-readable recording medium to perform S302 to S310 of FIG. 3 described above.

[0150] It can be implemented as a computing device including at least one processor that executes instructions of programs loaded into a memory, and a program including instructions described to execute S302 to S310 of the above-described drawing 3 can be loaded into the memory.

[0151] The devices described above may be implemented as hardware components, software components, and / or a combination of hardware components and software components. For example, the devices and components described in the embodiments may be implemented using one or more general-purpose computers or special-purpose computers, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing instructions and responding to them. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. Furthermore, the processing device may access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing device is sometimes described as being used singly; however, those skilled in the art will appreciate that the processing device may include multiple processing elements and / or multiple types of processing elements. For example, the processing device may include multiple processors, or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are also possible.

[0152] Software may include a computer program, code, instructions, or a combination of one or more of these, which may configure a processing device to perform a desired operation or may, independently or collectively, command the processing device. The software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual equipment, computer storage medium or device, or transmitted signal wave, for interpretation by the processing device or for providing instructions or data to the processing device. The software may also be distributed over networked computer systems and stored or executed in a distributed manner. The software and data may be stored on one or more computer-readable recording media.

[0153] The method according to the embodiment may be implemented in the form of program commands that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program commands, data files, data structures, etc., alone or in combination. The program commands recorded on the medium may be those specially designed and configured for the embodiment or may be those known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, and hardware devices specially configured to store and execute program commands, such as ROMs, RAMs, and flash memories. Examples of the program commands include not only machine language codes generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter, etc. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiment, and vice versa.

[0154] Although the embodiments described above have been described by way of limited examples and drawings, those skilled in the art will appreciate that various modifications and variations can be made based on the above teachings. For example, appropriate results can still be achieved even if the described techniques are performed in a different order than described, and / or components of the described systems, structures, devices, circuits, etc. are combined or combined in a different manner than described, or are replaced or substituted with other components or equivalents.

[0155] Therefore, other implementations, other embodiments, and equivalents to the claims also fall within the scope of the claims set forth below.

[0156] The device for assisting in the diagnosis and treatment of mental illness as described above and the method of operating the same can be applied to the medical field including the diagnosis and management of mental illness.

Claims

1. A step of receiving passive data and first active data from a user device; A step of extracting patient behavioral characteristic data from the above passive data; A step of predicting the second active data of the patient using a subjective emotional state prediction model that takes the above behavioral characteristic data as input; A step of outputting a predicted value representing a mental illness scale by using a mental illness scale classification model that inputs the above behavioral characteristic data and the second active data; and providing a user interface that displays the predicted value representing a mental illness scale to the user; Method of operation of a device for assisting in the diagnosis and treatment of mental disorders.

2. In paragraph 1, Passive data includes information about at least one of GPS coordinates, acceleration or tilt sensor data, screen ON or OFF time, call start and end time, type of app being used, or usage time of the app received from the user device. Method of operation of a device for assisting in the diagnosis and treatment of mental disorders.

3. In paragraph 2, The above behavioral characteristic data includes information on at least one of the visited location, time spent at each visited location, number of steps, moving distance, moving speed, walking distance, walking speed, activity complexity, and sleep time. Method of operation of a device for assisting in the diagnosis and treatment of mental disorders.

4. In paragraph 3, The above subjective emotional state prediction model is, Includes a boosting-based regression algorithm, Learned using the above behavioral characteristic data and the first active data, Method of operation of a device for assisting in the diagnosis and treatment of mental disorders.

5. In paragraph 4, The above subjective emotional state prediction model is, Learned using a feature subset determined based on RFE (Recursive Feature Elimination) among the above behavioral characteristic data. Method of operation of a device for assisting in the diagnosis and treatment of mental disorders.

6. In paragraph 3, The above mental illness scale classification model is, Includes a boosting-based classification algorithm, Learned using the passive data or the second active data augmented using SMOTE (synthetic minority oversampling technique). Method of operation of a device for assisting in the diagnosis and treatment of mental disorders.

7. In paragraph 6, The second active data is the predicted total score of the items answered by the patient in the mental health scale test. Method of operation of a device for assisting in the diagnosis and treatment of mental disorders.

8. In paragraph 7, The step of extracting patient behavioral characteristic data from the above passive data is: A step of extracting activity complexity data of the patient from the passive data; The step of extracting the patient's activity complexity data from the above passive data; A step of collecting the patient's location data for all dates and all time zones during a certain period; A step of clustering location data at a specific time for each date and selecting the first location that is located the most at the specific time; and A step of generating a repetitive life pattern graph based on the first location for each time zone; comprising; Method of operation of a device for assisting in the diagnosis and treatment of mental disorders.

9. In paragraph 8, The step of extracting the patient's activity complexity data from the above passive data is: A step of comparing the above repetitive life pattern graph with the life pattern graph of each date, and calculating the time at which the life pattern graph of each date is inconsistent with the pattern of the above repetitive life pattern graph; and Comprising a step of calculating the average of the time that is inconsistent with the pattern of the repetitive life pattern graph during the above period of time; Method of operation of a device for assisting in the diagnosis and treatment of mental disorders.

10. In paragraph 9, A step of providing recommended treatment content or recommended treatment solution to the patient based on at least one of the patient's passive data, the behavioral characteristic data, the second active data or the predicted value representing a mental illness scale; further comprising; Method of operation of a device for assisting in the diagnosis and treatment of mental disorders.

11. As a device for assisting in the diagnosis and treatment of mental illness, memory; and comprising at least one processor; The above processor, A device configured to receive passive data and first active data from a user device, extract patient behavioral characteristic data from the passive data, predict second active data of the patient using a subjective emotional state prediction model having the behavioral characteristic data as input, output a prediction value indicating the presence or absence of a mental illness using a mental illness scale classification model having the behavioral characteristic data and the second active data as input, and provide a user interface that displays the prediction value indicating the presence or absence of a mental illness to the user.

12. A computer program stored in a computer-readable recording medium, coupled with a computer as hardware, to perform the operating method of a mental illness diagnosis assistance and treatment device according to any one of claims 1 to 10.

13. As a mental illness diagnosis assistance and treatment system, A user device that collects passive data, first active data and second active data from a patient; and A mental illness diagnosis assistance and treatment device comprising a processor for receiving the passive data and the first active data from a user device, extracting the patient's behavioral characteristic data from the passive data, predicting the second active data of the patient using a subjective emotional state prediction model having the behavioral characteristic data as input, outputting a prediction value representing a mental illness scale using a mental illness scale classification model having the behavioral characteristic data and the second active data as input, and providing a user interface for displaying the prediction value representing a mental illness scale to the user; System.

Images