Electronic device for determining ptsd risk signs from psychological changes detected in ai-based mental state monitoring, and method for determining psychological risk signs using the same

Abstract

An electronic device for determining the risk of a psychological state based on an AI conversation-based analysis of individual psychological changes according to the present disclosure may include a memory for storing at least one instruction and at least one processor for executing said at least one instruction. The at least one processor may perform an AI conversation including an adaptive questionnaire for psychological analysis of a user, determine the user's symptoms by synthesizing self-reference data and monitoring data derived from said AI conversation, determine the user's risk indicators based on said symptoms, perform trend analysis, direction analysis, and severity analysis on said symptoms and said risk indicators, respectively, and calculate the risk of the user's psychological state by synthesizing the analysis results.

Description

Technology Field

[0001] The present disclosure relates to a technology for determining the risk level of a psychological state, and more specifically, to an electronic device for determining signs of PTSD risk from psychological changes detected in AI conversation-based psychological state monitoring, and a method for determining signs of psychological risk using the same. Background Technology

[0002] Recently, there has been an increasing trend of combining survey-based psychological assessment techniques, voice analysis techniques, and biosignal-based stress measurement techniques to understand users' psychological states. Conventional technologies primarily rely on fixed survey results collected at specific points in time or fragmentary biosignal values; while some technologies include natural language processing-based conversational interfaces, they have limitations in that they fail to adequately fulfill the objective of understanding the continuous flow of an individual's psychological changes. Furthermore, because conventional technologies interpret users' emotional expressions, behavioral responses, and biosignals separately, it may be difficult to precisely identify the correlation between an individual's actual psychological state and internal signals.

[0003] Conventional technology may struggle to detect latent emotional changes early on because it fails to quantitatively analyze inconsistencies between a user's verbal expressions and nonverbal signals. In particular, when users consciously conceal their emotions, simple survey techniques or language-based analysis methods may struggle to accurately reflect actual psychological deterioration. Furthermore, conventional technology suffers from structural limitations, such as its inability to comprehensively analyze changes in symptoms and risk indicators over time, and its failure to dynamically reflect directional bias toward specific disease groups or differences in symptom severity. Consequently, conventional technology faces the problem of being unable to clearly and early determine risks related to psychological issues, brain disorders, mental illnesses, and cognitive impairments.

[0004] In the field of psychological state analysis technology, there is a growing need for integrated analytical criteria capable of analyzing the flow of an individual's psychological changes over time, assessing the likelihood of progression to specific disease groups, and immediately detecting high-severity risk indicators. Furthermore, there is an emphasis on the necessity of technologies that quantitatively evaluate inconsistencies between verbal, vocal, and biometric signals, and interpret psychological fluctuations associated with specific events over the long term. Additionally, there is an expanding need for technologies that can detect psychological inflection points before a user's condition deteriorates and automatically provide appropriate behavioral interventions or psychological improvement content based on the findings. The problem to be solved

[0005] One objective of the present disclosure is to provide an electronic device that quantitatively identifies the flow of a user's actual psychological changes and consistently calculates the possibility of psychological deterioration or recovery by comprehensively analyzing the user's symptoms, signs of danger, and data changes over time, and a method for determining the risk of a psychological state using the same.

[0006] Another objective of the present disclosure is to provide an electronic device that quantitatively analyzes discrepancy signals between language signals, voice signals, and biosignals, performs time-series analysis based on event identifiers to interpret psychological changes after a specific event over the long term, and predicts an individual's future psychological changes, and a method for determining the risk of a psychological state using the same.

[0007] Another objective of the present disclosure is to provide an electronic device that realizes prevention-oriented psychological management by automatically detecting psychological inflection points and, based on the detection result, providing at least one of AI conversational intervention, cognitive behavioral training-based content, psychological stabilization content, VR-based desensitization content, and expert counseling linkage in accordance with dynamic conditions, and a method for determining the risk of a psychological state using the same.

[0008] However, the problems that this disclosure aims to solve are not limited to those mentioned above, and other unmentioned problems will be clearly understood by a person skilled in the art from the description below. means of solving the problem

[0009] An electronic device for determining the risk of a psychological state based on an AI conversation-based personal psychological change analysis according to the present disclosure for achieving the aforementioned technical problem may include a memory storing at least one instruction and at least one processor executing said at least one instruction. The at least one processor may perform an AI conversation including an adaptive questionnaire for psychological analysis of a user, determine the user's symptoms by synthesizing self-reference data and monitoring data derived from said AI conversation, determine the user's risk indicators based on said symptoms, perform trend analysis, direction analysis, and severity analysis on said symptoms and said risk indicators, respectively, and calculate the risk of the user's psychological state by synthesizing the analysis results.

[0010] In one embodiment, the at least one processor can determine the mutual correlation of the two elements by comparing the mutual change amounts of the symptom and the risk sign, and analyzing whether the symptom and the risk sign show the same increasing or decreasing trend over time or show different change patterns.

[0011] In one embodiment, the at least one processor filters noise data corresponding to transient fluctuations in a multi-data signal including at least one of survey response data, biosignal data, and cognitive function measurement data accumulated in chronological order during a trend analysis process, and analyzes whether symptoms worsen or recover by comparing the trend of the past multi-data signal with a new trend.

[0012] In one embodiment, the at least one processor may determine that a specific disease group is a risk candidate disease when a number of symptoms corresponding to a specific disease group appear continuously or repeatedly during the directional analysis process, calculate a directional score for the risk candidate disease group, and perform directional analysis by comparing it with a reference directional score.

[0013] In one embodiment, the at least one processor can accumulate and sum the symptom-specific severity scores assigned to each symptom during the severity analysis process and determine whether the severity scores accumulated over a set unit period exceed a reference severity score.

[0014] In one embodiment, the at least one processor assigns a first weight (W1) to the trend analysis result (T), a second weight (W2) to the direction analysis result (D), and a third weight (W3) to the severity analysis result (S), respectively, calculates an integrated risk score (R) based on the [formula] below, and can calculate the psychological state risk for a specific disease group based on the integrated risk score (R).

[0015] [formula]

[0016] R = W1XT + W2XD + W3XS

[0017] In one embodiment, the third weight (W3) may be greater than the second weight (W2), and the second weight (W2) may be set to be greater than the first weight (W1).

[0018] In one embodiment, the at least one processor monitors at least one of a biosignal, a language signal, a discrepancy signal, and a psychological state risk level based on a symptom change pattern or a risk sign change pattern, synthesizes the signals to detect a psychological inflection point, and can automatically provide a dynamic solution based on the detection result. The dynamic solution may include at least one of AI conversational intervention, cognitive behavioral training-based psychological improvement content, psychological stabilization content, VR-based desensitization content, and linkage with expert counseling.

[0019] In addition to this, a computer program stored on a computer-readable recording medium for implementing the present disclosure may be further provided.

[0020] In addition to this, a computer-readable recording medium for recording a computer program for implementing the present disclosure may be further provided. Effects of the invention

[0021] According to the aforementioned means for solving the problem of the present disclosure, the electronic device and the method for determining the risk of a psychological state using the same according to the present disclosure can quantitatively identify the actual psychological changes of a user by combining trend analysis, direction analysis, and severity analysis performed based on the user's symptoms, signs of danger, and time-series data. Unlike conventional technology that uses fragmentary point-in-time data, the electronic device and the method for determining the risk of a psychological state using the same according to the present disclosure determine the psychological state based on change patterns, thereby enabling the rapid and accurate determination of the possibility of psychological deterioration or recovery.

[0022] In addition, the electronic device of the present disclosure and the method for determining psychological state risk using the same can detect potential emotional states by quantitatively analyzing inconsistency signals between language signals, voice signals, and biosignals, and can predict future trends in psychological state changes based on event identifier-based time series analysis and the creation of a virtual user model.

[0023] In addition, the electronic device of the present disclosure and the method for determining psychological state risk using the same can clearly present the basis for calculating psychological state risk through an explainable artificial intelligence analysis module and can ensure transparency in the analysis process.

[0024] In addition, the electronic device and the method for determining psychological state risk using it can automatically detect psychological inflection points and, based on the detection result, immediately provide at least one of AI conversational intervention, cognitive behavioral training-based content, psychological stabilization content, VR-based desensitization content, and referral to expert counseling.

[0025] Accordingly, the electronic device of the present disclosure and the method for determining psychological state risk using the same can automatically adjust the intervention intensity and timing suitable for the user's condition, thereby improving the preventive effect and reliability of psychological management.

[0026] The effects of the present disclosure are not limited to those mentioned above, and other unmentioned effects will be clearly understood by a person skilled in the art from the description below. Brief explanation of the drawing

[0027] Figure 1 is a conceptual diagram illustrating the calculation of psychological state risk by multi-dimensionally analyzing the psychological state of a user using the AI ​​conversation of the present invention. FIG. 2 is a diagram showing the block configuration of the electronic device of the present invention. FIG. 3 is a conceptual diagram showing the overall operation of the electronic device of the present invention. FIG. 4 is a flowchart showing the sequential operation of the electronic device of the present invention. FIG. 5 is a diagram illustrating the operation of the electronic device of the present invention to calculate the risk level of a user's psychological state through analysis of symptoms and signs of danger. FIG. 6 is a diagram illustrating the operation of the electronic device of the present invention detecting a user's psychological inflection point and providing a solution for improving the psychological state. Specific details for implementing the invention

[0028] Hereinafter, embodiments are described in detail with reference to the attached drawings. However, various modifications may be made to the embodiments, and thus the scope of the patent application is not limited or restricted by these embodiments. It should be understood that all modifications, equivalents, and substitutions to the embodiments are included within the scope of the rights.

[0029] Specific structural or functional descriptions of the embodiments are disclosed for illustrative purposes only and may be modified and implemented in various forms. Accordingly, the embodiments are not limited to the specific disclosed forms, and the scope of this specification includes modifications, equivalents, or substitutions that fall within the technical concept.

[0030] Terms such as "first" or "second" may be used to describe various components, but these terms should be interpreted solely for the purpose of distinguishing one component from another. For example, the first component may be named the second component, and similarly, the second component may be named the first component.

[0031] Furthermore, terms defined in commonly used dictionaries are not interpreted ideally or excessively unless explicitly and specifically defined otherwise. In certain cases, terms have been selected at the applicant's discretion, and in such cases, their meanings will be described in detail in the relevant explanatory sections. Accordingly, terms used in this disclosure should be defined not merely by their names, but based on their meanings and the content throughout this disclosure.

[0032] When it is stated that a component is "connected" to another component, it should be understood that it may be directly connected to or joined to that other component, or that there may be other components in between.

[0033] The terms used in the embodiments are for illustrative purposes only and should not be interpreted as intended to be limiting. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, terms such as "comprising" or "having" are intended to indicate the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0034] Throughout this specification, when a part is described as "comprising" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components. Furthermore, the singular form used in this specification includes the plural form unless specifically stated otherwise in the text. Additionally, the expression "at least one of a, b, and c" described throughout this specification may encompass 'a alone,' 'b alone,' 'c alone,' 'a and b,' 'a and c,' 'b and c,' or 'a, b, and c all.'

[0035] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the embodiments pertain. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this application.

[0036] Additionally, terms such as “part,” “module,” etc., as described in this specification refer to a unit that processes at least one function or operation, which may be implemented in hardware or software, or a combination of hardware and software. Furthermore, embodiments of the present disclosure may be represented by functional block configurations and various processing steps. These functional blocks may be implemented by various numbers of hardware and / or software configurations that execute specific functions. For example, embodiments of the present disclosure may employ integrated circuit configurations such as memory, processing, logic, look-up tables, etc., which can execute various functions under the control of one or more microprocessors or other control devices.

[0037] Each block of the process flow diagrams attached to this specification and combinations of the flow diagrams may be executed by computer program instructions. Since these computer program instructions may be loaded into the processor of a general-purpose computer, a computer for special purposes, or other programmable data processing equipment, the instructions executed through the processor of the computer or other programmable data processing equipment create means for performing the functions described in the flow diagram block(s).

[0038] These computer program instructions may be stored in computer-available or computer-readable memory that can be directed toward a computer or other programmable data processing equipment to implement a function in a specific way, and the instructions stored in said computer-available or computer-readable memory may also produce a manufactured item containing instruction means that performs the function described in the flowchart block(s).

[0039] Since computer program instructions can be loaded onto a computer or other programmable data processing equipment, instructions that perform a series of operation steps on the computer or other programmable data processing equipment to create a process executed by the computer can also provide steps for executing the functions described in the flowchart block(s).

[0040] Additionally, each block may represent a module, segment, or part of code containing one or more executable instructions for executing a specified logical function(s). Furthermore, in some alternative execution examples, the functions mentioned in the blocks may occur out of order. For instance, two blocks described in succession may actually be executed substantially simultaneously, or the blocks may be executed in reverse order according to their corresponding functions.

[0041] In addition, when describing with reference to the attached drawings, identical components are assigned the same reference numeral regardless of drawing symbols, and redundant descriptions thereof are omitted. In describing the embodiments, if it is determined that a detailed description of related prior art could unnecessarily obscure the essence of the embodiments, such detailed description is omitted.

[0042] FIG. 1 is a conceptual diagram (100) for explaining how to calculate the psychological state risk level by multi-dimensionally analyzing the psychological state of a user using the AI ​​conversation of the present invention.

[0043] Referring to FIG. 1, the method for determining the psychological state risk of the present disclosure can interpret the user's psychological state in a multi-dimensional way based on a conversation the user has with an AI, and evaluate the degree of psychological risk of the user according to the result of the interpretation.

[0044] The method for determining psychological state risk of the present disclosure enables a user to transmit a response, including everyday expressions, emotional expressions, and self-report narratives, to an AI, and for the AI ​​to analyze the semantic context, narrative style, and emotional modulation of the response together, thereby identifying an individual's psychological flow in real time. The method for determining psychological state risk of the present disclosure enables the AI ​​to determine psychological characteristics that the user may not have clearly perceived themselves by interpreting not only verbal signals but also the user's non-verbal signals, reaction speed, and changes in conversation patterns together.

[0045] The method for determining psychological state risk of the present disclosure collects information regarding changes in self-referential speech, fluctuations in emotional amplitude, and the repetition of descriptions of specific symptoms based on the content of conversations between a user and an AI, and can convert and analyze said information into a pattern of change over time. The method for determining psychological state risk of the present disclosure does not rely on the results of a conversation at a single point in time, but rather interprets the amount of change in linguistic expression accumulated during the conversation process, the shift in emotional tone, and the user's change in perspective regarding problematic situations as continuous indicators, thereby specifically evaluating whether an individual's psychological state is stabilizing, deteriorating, or biased in a specific direction. Based on such flow of change, the method for determining psychological state risk of the present disclosure identifies the relationship between the user's symptoms and signs of risk, and can determine a psychological state risk level customized to the user.

[0046] The method for determining psychological state risk disclosed herein does not merely perform simple information exchange through a conversation between a user and an AI, but can interpret the content of the AI ​​conversation as an interaction that simultaneously incorporates emotional signals, behavioral signals, and linguistic signals reflecting the user's psychological characteristics. The method for determining psychological state risk disclosed herein can detect signals of inconsistency that may exist between the response expressed by the user and their actual inner state, and based on the results, can more precisely determine whether psychological risk is potentially increasing. The method for determining psychological state risk disclosed herein can capture precursory changes or psychological inflection points in advance before the user's psychological state reaches a risk level, thereby providing timely guidance, advice, and emotional stabilization content suitable for the user.

[0047] The method for assessing psychological state risk of the present disclosure can broadly identify the causes of a user's psychological changes by understanding the individual user's psychological characteristics over the long term and interpreting the correlation between the time of an event and the time of a conversation. The method for assessing psychological state risk of the present disclosure can predict future risk based on patterns such as stress responses increasing after a specific event, repetitive expressions of negative emotions, or the appearance of specific symptom clusters. Based on the prediction results, the method for assessing psychological state risk of the present disclosure can enhance user protection functions by closely monitoring an individual's psychological changes and guiding them to provide professional intervention when necessary.

[0048] The psychological state risk assessment method of the present disclosure centers on the interaction between the user and the AI, while possessing the advantage of being able to interpret an individual's psychological structure in a multidimensional manner, going beyond simple conversational analysis. The psychological state risk assessment method of the present disclosure enables early detection of psychological risk, determination of the direction of risk increase, and an integrated understanding of the causes of psychological changes, thereby providing the user with a safer and more reliable psychological management environment.

[0049] FIG. 2 is a diagram showing the block configuration of the electronic device (200) of the present invention.

[0050] Referring to FIG. 2, the AI ​​conversation-based psychological analysis function described in FIG. 1 can be provided using an electronic device (200). For example, the electronic device (200) may include a computing environment for reliably processing a conversation between a user and an AI, and for analyzing self-referential data and monitoring data generated from the conversation in real time.

[0051] The electronic device (200) may include at least one of a server-type device equipped with high-performance computing resources, a portable device in the form of a smartphone or tablet capable of constant response in a mobile environment, a wearable-linked device that collects the user's biosignals in real time and supports data transmission, an scalable analysis server based on a public cloud, or a secure cloud server capable of safely storing sensitive psychological data. Through such various forms of hardware configurations, the electronic device (200) can continuously perform functions such as AI conversation processing, adaptive survey presentation, collection of language, voice, and biosignals, storage of conversation records, and output of analysis results.

[0052] For example, the electronic device (200) may be configured as a single device to immediately process user-customized psychological analysis. As another example, the electronic device (200) may be configured with a structure in which multiple devices operate cooperatively, so that monitoring data collected from a portable device is deeply analyzed on a cloud analysis server, and the analysis results are provided back to the user terminal. Through such a configuration, the electronic device (200) can reliably and accurately perform key functions such as extracting changes in the user's psychological state, determining symptoms, evaluating signs of risk, and detecting psychological inflection points.

[0053] In one embodiment, the electronic device (200) may include an input / output interface (210), a memory (220), a processor (230), and a communication interface (240). However, the present disclosure is not limited thereto. The electronic device (200) may be configured with some of the components shown in FIG. 2 omitted, or may be configured to include other components in addition to the components shown in FIG. 2.

[0054] In one embodiment, the input / output interface (210), memory (220), processor (230), and communication interface (240) may each be physically / electrically connected to each other.

[0055] In one embodiment, the electronic device (200) may be connected to various types of external devices through an input / output interface (210). In one embodiment, the input / output interface (210) may include at least one of a wired / wireless headset port, an external charger port, a wired / wireless data port, a memory card port, a port for connecting a device equipped with an identification module (SIM), an audio I / O (Input / Output) port, or a video I / O (Input / Output) port. In one embodiment, the input / output interface (210) may include a USB (Universal Serial Bus), HDMI (High Definition Multimedia Interface), or DVI (Digital Visual Interface), etc.

[0056] In one embodiment, the memory (220) may store data used in the electronic device (200). In one embodiment, the memory (220) may store instructions, programs, or modules for the operation of the processor (230).

[0057] In one embodiment, the memory (220) may include at least one type of storage medium among a flash memory type, a hard disk type, an SSD type (Solid State Disk type), an SSD type (Silicon Disk Drive type), a multimedia card micro type, a card type memory (e.g., SD or XD memory, etc.), RAM (random access memory; RAM), SRAM (static random access memory), ROM (read-only memory; ROM), EEPROM (electrically erasable programmable read-only memory), PROM (programmable read-only memory), magnetic memory, and an optical disk.

[0058] In one embodiment, the memory (220) may include user information and components displayed on the user interface.

[0059] In one embodiment, the processor (230) may include a general-purpose processor such as a CPU (Central Processing Unit), AP (Application Processor), DSP (Digital Signal Processor), or a neural network processing processor such as an NPU (Neural Processing Unit). In one embodiment, the processor (230) may be divided by one or more processors to perform operations. In one embodiment, the processor (230) may control the operation of the electronic device (200). The processor (230) may control the operation of the electronic device (200) according to instructions stored in memory (220).

[0060] In one embodiment, the processor (230) may be implemented as a memory that stores data for an algorithm or a program that reproduces the algorithm for controlling the operation of components within the electronic device (200) of the present disclosure, and as a processor that performs the aforementioned operation using the data stored in the memory. In this case, the memory and the processor may each be implemented as separate chips. Alternatively, the memory and the processor may be implemented as a single chip.

[0061] In addition, the processor may control one or a combination of the components described above in order to implement various embodiments according to the present disclosure through the electronic device (200).

[0062] In one embodiment, the communication interface (240) may include one or more components that enable communication between an external server or an external electronic device and the electronic device (200). In one embodiment, the communication interface (240) may include at least one of a wired communication module or a wireless communication module.

[0063] The wired communication module may include various wired communication modules such as a Local Area Network (LAN) module, a Wide Area Network (WAN) module, or a Value Added Network (VAN) module, as well as various cable communication modules such as USB (Universal Serial Bus), HDMI (High Definition Multimedia Interface), DVI (Digital Visual Interface), RS-232 (recommended standard 232), power line communication, or POTS (plain old telephone service).

[0064] The wireless communication module may include a wireless communication module that supports at least one of a wireless communication method including WiBro (Wireless broadband), GSM (global System for Mobile Communication), CDMA (Code Division Multiple Access), WCDMA (Wideband Code Division Multiple Access), UMTS (universal mobile telecommunications system), TDMA (Time Division Multiple Access), LTE (Long Term Evolution), 4G, 5G, 6G, Bluetooth™, or Wi-Fi (Wireless-Fidelity).

[0065] In one embodiment, the processor (230) can obtain information from an external electronic device or server or provide information through a communication interface (240).

[0066] A processor (230) according to one embodiment can perform the operation of an electronic device (200) described below through the drawings. Specifically, the processor (230) can perform an AI conversation including an adaptive questionnaire for psychological analysis of the user, and can determine the user's symptoms by integrating self-reference data and monitoring data obtained through the conversation. Based on the determined symptom information, the processor (230) can determine risk signs corresponding to specific disease groups such as PTSD, depression, anxiety, aggression, schizophrenia, and suicide risk. For each of the symptoms and risk signs, the processor (230) can perform a trend analysis over time, a direction analysis to determine whether progression to a specific disease group occurs, and a severity analysis to cumulatively evaluate the severity of the symptoms. The processor (230) can comprehensively determine the risk level of the user's psychological state by calculating an integrated risk score by assigning different weights to the results of the trend analysis, direction analysis, and severity analysis. For example, the processor (230) can perform trend analysis, direction analysis, and severity analysis using at least one artificial intelligence model, and determine the risk level of the psychological state by synthesizing the analysis results.

[0067] Here, the artificial intelligence model used by the electronic device (200) to perform trend analysis, direction analysis, and severity analysis, and to determine the psychological state risk level through the synthesis of analysis results, may be a single model or a combination of multiple models. The artificial intelligence model may be configured based on a neural network and may include statistical learning algorithms developed in the fields of machine learning and cognitive science. The neural network may include a structure in which multiple artificial neurons are combined, and the artificial neurons may solve a target problem by adjusting weights and biases during the learning process. The neural network may include one or more layer structures including an input layer, a hidden layer, and an output layer, and the processor of the electronic device (200) may infer results from input data through the learning of the neural network.

[0068] A processor included in an electronic device (200) can generate a neural network, train or retrain the neural network, perform operations based on input data, and generate an information signal based on the results of the operations. The processor may be a single processor or a multi-processor structure, and may include at least one operation module corresponding to the type of neural network model.

[0069] For example, artificial intelligence models may include various neural network structures such as CNN, R-CNN, RPN, RNN, LSTM, GRU, DBN, RBM, Deconvolution Network, Fully Convolutional Network, Auto Encoder, Variational Auto Encoder, Sparse Auto Encoder, Markov Chain, Hopfield Network, Boltzmann Machine, Deep Residual Network, Capsule Network, Attention Network, etc. In addition, artificial intelligence models may include the latest AI algorithms such as BERT, SP-BERT, GPT family models for natural language processing, question answering models, dialogue models, ResNet for vision processing, Visual Understanding, image generation models, time series forecasting models, anomaly detection models, recommendation models, optimization models, etc.

[0070] In one embodiment, a processor included in an electronic device (200) may include an artificial intelligence model that operates based on an AI Agent structure. The processor may utilize an AI Agent model that performs goal-based behavior generation, stepwise judgment based on environmental changes, automatic configuration of complex processing procedures, and interaction with external tools. Additionally, the processor may utilize the AI ​​Agent model to stepwise determine complex inference procedures according to learned policies, perform cooperative computations between models, and dynamically perform goal-based tasks based on user state or system environment changes.

[0071] An artificial intelligence model can be composed of a single model or a multi-model structure combining multiple models with different roles, such as a prediction-only model, a feature extraction model, a generative model, a decision model, and an optimization model. In the case of a multi-model structure, the processor can select a model as needed, coordinate the flow of operations between models, or integrate the results generated by multiple models to produce a final output.

[0072] FIG. 3 is a conceptual diagram (300) showing the overall operation of the electronic device (200) of the present invention, and FIG. 4 is a flowchart showing the sequential operation of the electronic device (200) of the present invention.

[0073] Referring to FIGS. 3 and 4, the electronic device (200) of the present invention can perform an AI conversation including an adaptive survey for psychological analysis of a user (operation 410), determine the user's symptoms by synthesizing self-reference data and monitoring data derived from the AI ​​conversation (operation 420), determine the user's risk indicators based on the symptoms (operation 430), perform trend analysis, direction analysis, and severity analysis on the symptoms and the risk indicators respectively (operation 440), and calculate the psychological state risk of the user by synthesizing the analysis results (operation 450).

[0074] According to one example, in operation 410, the electronic device (200) can perform an AI conversation including an adaptive survey for psychological analysis of the user.

[0075] Here, an adaptive survey for user psychological analysis may refer to a survey procedure in which the composition of questions dynamically changes according to the user's situation, context, and emotional state. An adaptive survey can be a method of selecting questions to quickly identify users' response tendencies and individual psychological vulnerabilities. AI conversation is a process in which a user and an electronic device engage in language-based interaction, and the adaptive survey can function as a key tool for collecting psychological information within that interaction.

[0076] The electronic device (200) can determine whether the current question flow is appropriate by analyzing emotional cues, contextual cues, and subjective descriptive elements included in the initial input sentence presented by the user. The electronic device (200) can adjust the difficulty, form, and topic of the next question by interpreting elements such as the user's response length, speech rate, delay time between responses, and the ratio of positive or negative expressions. The electronic device (200) can present evaluative questions that are less burdensome than open-ended burden questions to users who are in an emotionally sensitive state, and can present questions that induce cognitive stimulation to users who show a strong intention to solve problems. The electronic device (200) can determine changes in the questioning method by checking whether the user's self-referential expressions increase or decrease during the adaptive survey process, and can present questions related to the relevant emotions intensively by analyzing emotional elements that the user repeatedly mentions in the conversation flow.

[0077] The electronic device (200) can detect situations where a discrepancy occurs between language signals and voice signals and adjust the questioning method. For example, even if the user says "I am fine," if there is a decrease in voice amplitude, a decrease in speech rate, or a change in breathing rhythm, the electronic device (200) can determine the possibility of emotional vulnerability and select questions that minimize psychological burden. The electronic device (200) does not maintain the order of questions in a fixed manner, but calculates the amount of information included in the user's response and can prioritize selecting items that are judged to provide a high amount of information. The electronic device (200) can analyze the reliability and consistency of the response to present additional confirmation questions on the same topic, and if the user exhibits evasive speech on a specific topic, it can present alternative items to alleviate psychological resistance.

[0078] The electronic device (200) can continuously detect the possibility of a sudden change in the user's state during the process of conducting an adaptive survey. The electronic device (200) can automatically recognize situations where biosignal patterns rise or fall rapidly to determine the possibility of a surge in negative emotions or psychological overload, and can lower the intensity of questions or smoothly switch the topic of conversation depending on the situation. The electronic device (200) can shift the purpose of the conversation from information gathering to emotional relief if specific linguistic expressions, such as words of emotional risk like despair, helplessness, or pressure, are repeated during the conversation. The electronic device (200) can minimize psychological threat by adjusting the approach to the topic when a topic that the user has previously shown a strong emotional reaction to reappears.

[0079] The electronic device (200) can integrate adaptive surveys and AI conversations to form a data collection base for user-customized psychological state analysis. The electronic device (200) can simultaneously collect language signals, voice signals, and bio-signals within the flow of the conversation, and can obtain various clues related to changes in the user's psychological state through the combination of the collected signals. The electronic device (200) can improve the quality of psychological information that can be used in future analysis stages by evaluating the consistency, stability, and potential for emotional distortion of the data derived during the conversation. Beyond simply recording responses obtained from the adaptive survey, the electronic device (200) can determine the relationship between the response and the user's emotional direction to construct structured information for psychological change analysis.

[0080] The electronic device (200) can select the next question by reflecting both quantitative and qualitative characteristics of the user's utterance. The electronic device (200) can determine the timing of intervention or conversation adjustment by interpreting even minute changes in the user's conversation pattern. Through this conversation-based analysis, the electronic device (200) can obtain sufficient and sophisticated psychological data necessary for subsequent steps, such as symptom judgment, risk sign analysis, and psychological state risk calculation.

[0081] For example, the electronic device (200) can determine the possibility of emotional withdrawal if the user uses shorter sentences than usual or responds with short answers to specific questions, and can select a low-burden confirmatory question. If the user's speech rate continuously slows down, the electronic device (200) can select a question with reduced difficulty and emotional stimulation intensity, considering the possibility of increased fatigue or negative emotion. If the user repeats the same words or the sentence structure becomes simple, the electronic device (200) can determine the possibility of rigidity in thought flow or topic avoidance, and can select a question that allows the user to bypass the topic and explore related elements. If the user uses an exceptionally large number of emotional expressions or the intensity of emotional words suddenly increases, the electronic device (200) can determine the possibility of emotional over-immersion and can prioritize selecting a question intended to stabilize the user's emotions. If the user stays on a specific topic for a long time or provides excessive unnecessary details, the electronic device (200) can determine the possibility of thought expansion associated with cognitive load or anxiety and select a structured question that helps organize the focus of thought.

[0082] The electronic device (200) can determine the possibility of defensive thinking when the user repeatedly avoids the topic of the question or develops the conversation in a direction unrelated to the core of the question, and can select a detour question or an emotion clarification question with low approach intensity. The electronic device (200) can analyze micro-signals such as hesitation gaps, breathing patterns, a soft voice, and changes in endings that appear during the user's speech to determine whether the user feels burdened by a specific topic, and can select a question that attempts a sequential approach to a low-burden topic. If the user voluntarily mentions an unexpected topic, the electronic device (200) can select a follow-up question that can expand on the topic by considering the possibility that the topic is highly relevant to the current psychological state.

[0083] The electronic device (200) can determine whether the user is in an emotionally unstable state by measuring the speed of contextual switching of the user's language expressions and can select questions suitable for emotional stabilization. The electronic device (200) can determine the possibility of confusion in the flow of thought by evaluating changes in the clarity of logical connection structures, such as causal relationships or chronological order, included in the user's utterances and can select questions that help organize thoughts. If signals of positive changes appearing in the user's conversation patterns, such as increased expression of self-efficacy or strengthened willingness to solve problems, are detected, the electronic device (200) can select questions that can reinforce those positive elements. The electronic device (200) can determine the most suitable type of question for the user in real time by comprehensively interpreting psychological changes combining emotional, cognitive, and behavioral elements throughout the entire conversation pattern.

[0084] According to one example, in operation 420, the electronic device (200) can determine the symptoms of the user by combining self-reference data and monitoring data derived from the AI ​​conversation.

[0085] Here, self-reference data may refer to verbal expressions in which the user directly mentions their emotions, thoughts, physical reactions, and life contexts. Monitoring data may be multimodal information that includes the user's verbal signals, voice signals, and biosignals, and provides non-verbal cues related to changes in psychological state. Symptoms may refer to emotional characteristics, behavioral characteristics, or expressions associated with specific disease groups extracted from self-reference and monitoring data. Symptom assessment may be a process of classifying the user's psychological state into a structured form by interpreting whether self-report verbal expressions correspond to physiological response signals, or whether there is a discrepancy between verbal and non-verbal expressions.

[0086] The electronic device (200) can extract initial symptom candidates by analyzing emotional words, negative emotional expressions, ruminative sentence structures, and avoidant speech patterns appearing in self-reference data. The electronic device (200) can determine the emotional authenticity of a linguistic expression by analyzing voice signals such as the user's voice pitch, speech intensity, speech speed, and trailing at the end of words. The electronic device (200) can interpret the degree of emotional consistency by checking whether biosignals such as the user's heart rate changes, breathing patterns, and muscle tension signals match the linguistic expression. When linguistic signals and non-linguistic signals conflict with each other, the electronic device (200) can detect a mismatch signal and use the said mismatch signal as a basis for determining potential symptoms.

[0087] The electronic device (200) can comprehensively determine symptoms by evaluating the repetition of patterns appearing in self-reference data and monitoring data, the increase or decrease in the amount of change, and whether there is a surge in specific emotional expressions. The electronic device (200) can identify the user's emotional direction, level of psychological burden, and whether there is an increase in negative thought flow by analyzing semantic changes in self-reference data arranged in chronological order. The electronic device (200) can evaluate whether changes in linguistic flow, such as changes in response length, simplification of sentence structure, and reduction or expansion of emotional expressions, are associated with specific symptoms. The electronic device (200) can determine emotional tension and stability by analyzing the fluctuation range of biosignals, recovery speed, and the length of the stability interval.

[0088] The electronic device (200) can determine the combination pattern as the appearance of a clear symptom when the emotional element detected in the self-reference data and the physiological response detected in the monitoring data are combined in a way that reinforces each other. The electronic device (200) can determine that a psychological stress symptom exists when an expression of psychological burden in the self-reference data and a signal of increased tension in the biosignal occur simultaneously. The electronic device (200) can construct reference information for determining the intensity of symptoms when specific emotional elements such as depression, anxiety, aggression, over-tension, and confusion appear repeatedly in the user's language expression.

[0089] The electronic device (200) can evaluate how much the current data has deviated from a past stable state by referring to user-specific history information and past conversation records. The electronic device (200) can determine changes in emotional stability, increased stress vulnerability, and persistence of negative thoughts by comparing long-term patterns of self-referenced data. The electronic device (200) can distinguish between temporary emotional responses and structural symptoms by considering both short-term changes and long-term fluctuations in monitoring data. The electronic device (200) can determine the possibility of psychological instability when the user history and current state do not match, and can construct basic information to analyze whether the user state meets the criteria necessary for determining signs of danger.

[0090] For example, the electronic device (200) may determine that symptoms related to lethargy exist if the user repeatedly mentions expressions such as "I don't want to do anything," "my head feels heavy," or "I have no energy" during the past few days, and if a pattern of lower-than-usual activity and irregular sleep rhythms are confirmed in the biosignals. The electronic device (200) may determine that symptoms related to PTSD exist if the user displays self-referential expressions such as "I keep thinking about the past," "my heart suddenly beats fast," or "I keep having the same dream" during conversation, and if a rapid increase in heart rate, increased breathing rate, and voice tremors are observed simultaneously. The electronic device (200) may determine that symptoms of anxiety exist if the user frequently uses emotional expressions such as "I feel anxious," "I feel like something is going to happen soon," or "my mind is cluttered," and if a pattern appears in which the speed of speech becomes faster than usual and the intensity of speech increases momentarily.

[0091] For example, the electronic device (200) can determine that aggression-related symptoms exist if the user repeatedly mentions emotional changes such as "I am annoyed," "I have become sensitive," or "I get angry at trivial matters" during a conversation, and if aggressive intonation or rapid fluctuations in voice pitch appear in the voice characteristics. The electronic device (200) can obtain grounds for determining schizophrenia-related symptoms if the user displays expressions such as "I lack a sense of reality," "My thoughts are not organized," or "I feel like someone is watching me," and if voice signals and biosignals repeatedly show tension and confusion. The electronic device (200) can construct judgment information necessary to determine suicide-related symptoms if the user repeatedly makes expressions such as "I don't want to live," "I keep thinking I want to disappear," or "I want to end it because it is hard," without emotional consistency, and if extreme tension patterns or rapid decreases in emotional amplitude appear in the biosignals.

[0092] The electronic device (200) can determine the complex symptom structure by combining the results of multi-mode analysis of language signals, voice signals, and biosignals. The electronic device (200) can determine whether it is a single symptom or a complex psychological problem by analyzing the combined pattern of multiple signals collected at the same time. The electronic device (200) can determine the complexity of the symptom structure by evaluating the association when a specific symptom appears in association with other symptoms. After determining the symptoms, the electronic device (200) can generate input information to perform risk indicator determination, psychological direction analysis, and psychological state risk calculation in subsequent steps.

[0093] According to one example, in operation 430, the electronic device (200) can determine the signs of danger to the user based on the symptoms.

[0094] For example, a risk sign may refer to an early signal indicating the potential onset of a specific mental disorder group. The risk sign judgment module may be an analytical configuration designed to identify risk candidates for a specific disorder group based on multiple symptom information derived during the symptom assessment process. Risk candidate disorder groups may refer to representative categories of mental health disorders such as PTSD, depression, anxiety, aggression, schizophrenia, and suicide. The risk sign change pattern may be an analytical variable reflecting the intensity, frequency, and persistence of symptoms that fluctuate over time. Directional information may be an analytical element indicating whether multiple symptoms constituting a specific disorder group increase continuously or repeatedly in the direction of that same disorder group.

[0095] The electronic device (200) can initially assign a disease group risk score corresponding to each symptom when at least one of the symptoms constituting a specific disease group is identified. The electronic device (200) uses a combination of user speech data, biosignal data, voice feature data, and self-referenced expressions collected during AI conversation processes to simultaneously consider multiple symptom information during the risk indication judgment process, and can evaluate in a structured manner the intensity of the risk signal of an individual disease group based on the information. The electronic device (200) stores a list of representative symptoms constituting each disease group and severity criteria for each symptom, based on the premise that the risk candidate disease group includes various mental health categories such as PTSD, depression, anxiety, aggression, schizophrenia, and suicide, and can calculate an initial risk score based on the extent to which the symptoms observed in the user correspond to the criteria. The electronic device (200) can determine whether the symptoms persist when specific emotional expressions increase or changes in biological responses appear repeatedly in user conversations by recording the time of symptom occurrence and the interval between occurrences, and can determine the possibility of an increase in the risk of the disease group by evaluating whether expressions related to the same disease group are continuously or gradually intensified. The electronic device (200) can improve the accuracy of the risk indication judgment by adjusting the initial risk score when the intensity, frequency, and persistence of the appearance of symptoms exceed certain criteria, and can assign a high risk indication score even for a single symptom if it is classified as having very high severity or requiring immediate intervention.

[0096] The electronic device (200) can quantitatively evaluate the likelihood of a risk sign occurring by combining multiple criteria in the risk sign judgment process. The electronic device (200) may primarily define the case where at least one of the survey score, cognitive function assessment score, and self-report emotional expression score deteriorates over time as one of the key conditions for determining a risk sign. Additionally, the electronic device (200) can focus on evaluating whether symptoms associated with a specific disease group are continuously added; for example, if a user who already has one depression-related symptom subsequently accumulates expressions of a depression-related symptom group such as "lack of energy," "loss of interest," and "persistent lethargy," it may determine that the direction of the disease group has been strengthened. The electronic device (200) can determine whether there is a risk sign by analyzing whether the score calculated from a specific survey item exceeds a pre-set threshold, and in particular, if the score of a survey item requiring immediate intervention, such as suicide risk or aggression risk, exceeds the threshold value, a high risk score may be assigned. Finally, the electronic device (200) can determine an immediate increase in risk by setting it as a top priority risk indicator if a high-risk signal directly associated with serious danger, such as suicidal expressions or expressions of aggression, is added as a single item. The electronic device (200) evaluates the likelihood of the occurrence of risk indicators step by step by considering various judgment criteria such as this simultaneously, and can improve the accuracy of the analysis performed in the next step based on the risk indicator judgment results.

[0097] In one embodiment, the electronic device (200) can identify a candidate for risk of a depressive disorder group if the user repeats expressions such as “I barely slept,” “I am tired every day,” “I have no energy,” or “I feel down all day,” and if the rate of speech decreases and the rate of facial muscle movement or breathing slows down. The electronic device (200) records the timing of the appearance of the symptoms and can determine that there is a sign of risk of depression if the same group of symptoms is repeated continuously for more than a few days, or if it increases to a level of higher intensity or frequency than in the past.

[0098] In one embodiment, the electronic device (200) may determine an anxiety disorder group as a risk candidate if the user continues to express emotions such as "anxious," "feeling nervous," "feeling tightness in the chest," or "feeling that something bad is going to happen," and if the speech rate increases, speech pauses increase, and the range of heart rate fluctuations increases. The electronic device (200) may constitute an anxiety risk indicator if two or more of the anxiety-related symptoms appear repeatedly over a certain period and show a pattern in which the same group of symptoms intensifies in a specific direction.

[0099] In one embodiment, the electronic device (200) can determine the PTSD disease group as a risk candidate if the user repeatedly expresses memory re-experience such as “old scenes keep coming to mind,” “thoughts keep coming to mind to the point where it is difficult to immerse oneself,” or “fear of similar situations,” and if a rapid increase in heart rate, irregular breathing patterns, and voice tremors appear simultaneously in the biosignals. The electronic device (200) can establish grounds for determining PTSD risk signs if two or more symptoms among memory re-experience, avoidance, and hyperarousal appear for a certain period of time or longer.

[0100] In one embodiment, the electronic device (200) can determine a candidate for aggression-related risk if the user continues to use expressions such as "can't stand it," "gets angry at trivial things," or "feels like he's about to explode" in recent conversations, and if the voice pitch changes rapidly or the intensity of speech increases in intervals. The electronic device (200) can determine a sign of aggression risk if the aggression-related signal continues to increase.

[0101] In one embodiment, the electronic device (200) may determine that the schizophrenia group is a risk candidate if the user repeats expressions of cognitive confusion such as “I feel unreal,” “My surroundings feel unfamiliar,” or “I feel like someone is watching me,” and shows a pattern in which the speech structure becomes confused or the flow of thought is interrupted. The electronic device (200) may determine that there are signs of schizophrenia risk if the frequency and intensity of cognitive confusion increase.

[0102] In one embodiment, the electronic device (200) may determine a candidate for a suicide risk group if the user repeats expressions such as "it is hard to live," "I want to stop," or "I want to disappear," or if there is a sudden change in heart rate or an irregular increase in the amplitude of tension indicators in the biosignals. The electronic device (200) may immediately assign a high risk indicator score based on a single symptom related to suicide risk.

[0103] The electronic device (200) can quantify the risk of a specific disease group by interlinking the composition of the symptom group, the repetitiveness of symptom appearance, the presence of directionality, and the severity level of each symptom during the risk indication judgment process. The electronic device (200) can assign a high risk indication score if the symptom group of the specific disease group shows a pattern of intensifying over time and multiple symptoms consistently appear in the direction of the same disease group. The electronic device (200) can determine that there is a risk indication for the disease group if the risk indication score exceeds the reference risk indication score.

[0104] The electronic device (200) can store the results of the risk indication judgment and generate a risk indication information structure to be used as basic data for subsequent trend analysis, direction analysis, and severity analysis.

[0105] FIG. 5 is a diagram illustrating the operation (500) of the electronic device (200) of the present invention calculating the psychological state risk of a user through analysis of symptoms and signs of danger.

[0106] Referring to FIG. 5, the electronic device (200) can determine the progression trend of a risk sign by connecting the temporal occurrence relationship between the symptoms and the risk sign. The electronic device (200) determines the mutual correlation between the symptoms and the risk sign based on the mutual change pattern of the symptoms and the risk sign, and can perform more precise trend analysis, direction analysis, and severity analysis based on the mutual correlation.

[0107] The electronic device (200) can calculate the time-based change amount of each symptom value and each risk sign value to analyze the correlation between symptoms and risk signs. To calculate the change amount, the electronic device (200) refers to the symptom record values ​​and risk sign record values ​​aligned at the same time interval, and can analyze the change pattern by calculating the increase, decrease, and stability interval of the two values, respectively. The electronic device (200) compares whether the risk sign value also increases in tandem during the interval where the symptom value increases, and can determine that a correlation exists if the direction of change of the two values ​​matches for a certain period of time or longer.

[0108] The electronic device (200) can determine an asynchronous change pattern between two variables when the symptom value is decreasing but the risk sign value is increasing, or when the symptom value is rising but the risk sign value remains constant or the change slows down. The electronic device (200) can refer to a pre-stored pattern of change by disease group to determine whether the asynchronous change pattern corresponds to a characteristic pattern of a specific disease group. For example, since PTSD-related signs may show a delayed rise after a specific situational stimulus, the electronic device (200) can perform an analysis that associates a pattern of symptom change and risk sign change being temporally out of sync with an increase in PTSD-related risk.

[0109] The electronic device (200) can quantitatively calculate the difference in slope between the symptom change amount and the risk sign change amount during the change amount comparison process, and if the difference in slope is below a reference threshold, it can determine that the pattern has the same increasing or decreasing trend. The electronic device (200) can determine that the correlation is low if the difference in slope is greater than the reference threshold and the direction of change is opposite or intersects irregularly. After determining whether the change pattern is synchronized, the electronic device (200) can calculate the correlation level between the symptom and the risk sign by comprehensively considering the duration of the synchronized pattern, the intensity of the reinforced change, and the repeatability of the change pattern.

[0110] The electronic device (200) can improve the reliability of risk assessment by disease group by reflecting the results of the correlation analysis in the direction analysis, trend analysis, and severity analysis performed in subsequent steps. The electronic device (200) can more accurately determine whether the risk signs of a specific disease group are actually intensifying by evaluating symptom groups with high correlation with high weights.

[0111] For example, the electronic device (200) can analyze by distinguishing between patterns in which both types of information are reinforced simultaneously, patterns in which only one side changes rapidly, or patterns in which they are linked with a certain time lag, by considering that symptom data and risk indicator data do not change independently of each other but can be closely linked. During the analysis process, the electronic device (200) can calculate the weight regarding which element among the user's emotional expression, behavioral characteristics, and biological response has a greater influence on the increase in risk indicators, and can adjust the weight in the subsequent evaluation stage based on the said weight.

[0112] For example, the electronic device (200) can primarily categorize the pattern of change according to the characteristics of the disease group, such as the delayed hyperarousal pattern of PTSD, the cumulative worsening pattern of depression over a certain period, and the acute worsening pattern of anxiety, by considering that symptoms and signs of danger may change in different ways depending on the type of specific disease group. The electronic device (200) can initially determine whether the risk is increasing by comparing the typical change pattern appearing in each disease group with the user's actual change pattern, and can evaluate the possibility of increased risk by combining multiple factors such as the direction of change, the speed of change, and the intensity of change.

[0113] According to one example, in operation 440, the electronic device (200) can perform trend analysis, direction analysis, and severity analysis for the symptoms and the signs of danger, respectively.

[0114] Trend analysis can be an analytical procedure designed to quantitatively interpret the flow of a user's psychological changes by comparing symptom values ​​and risk indicator values ​​that change over time. It can be a process for identifying increasing, decreasing, or constant trends in chronologically arranged data, and by simultaneously considering the direction and speed of change, it can interpret whether the user's psychological state is deteriorating, recovering, or exhibiting temporary fluctuations. To capture long-term changes that are difficult to judge based on information from a single point in time, trend analysis can be performed by combining multiple factors, such as the slope of accumulated data, the rate of change of the slope, and the duration of the change interval; furthermore, short-term fluctuations deemed as noise can be filtered out to extract only the overall flow. Trend analysis can reliably identify the flow by using normalized data points, even when hourly data is collected irregularly or intermittently.

[0115] For example, the electronic device (200) can determine short-term rapid fluctuations in multiple data signals, such as the user's sleep-related survey response score, heart rate variability, and activity-to-rest ratio, as noise and remove them, and then compare the average trend of the same indicator over the past two weeks with the current trend. The electronic device (200) can determine that there is a worsening trend if the sleep disorder score continues to rise and there is an accumulating trend of increased night awakening in the biosignals, and conversely, if a gradual recovery trend is confirmed over the past three days, it can interpret that there is a recovery trend. The electronic device (200) can also consider it a positive change if the stability of the biosignals increases as the user's recent stress expression frequency decreases, and can utilize the result in the next stage of risk calculation.

[0116] Directional analysis can be a procedure to identify patterns in which symptoms associated with a specific disease group concentrate in a certain direction. Directionality refers to whether multiple symptoms are biased toward the same disease group category, and representative symptom groups can be pre-established based on which category the disease group falls under—such as PTSD, depression, anxiety, aggression, schizophrenia, or suicide. Directional analysis can measure the degree of accumulation of disease group candidates by evaluating whether user utterances, vocal characteristics, and biological responses are repeated as expressions associated with specific disease groups. Directional analysis can determine whether the bias toward a disease group is evident by comprehensively considering not only individual symptom scores but also correlation patterns between symptoms related to the same disease group, consistency of occurrence intervals, and cumulative increases in occurrence intensity. Directional analysis can be effectively utilized to detect risk early, even when symptom scores are low, if expressions corresponding to a specific disease group are repeated.

[0117] For example, the electronic device (200) can determine that the directionality of the depressive disorder group is being strengthened if the user continuously mentions expressions such as "I have no energy," "I feel down all day," or "I have almost no motivation" during the AI ​​conversation, and if a pattern of reduced activity is repeated in the biosignals. Conversely, the electronic device (200) can determine that the directionality score for the anxiety disorder group is increasing if the user repeatedly mentions expressions such as "I feel anxious," "I keep feeling tense," or "My chest feels tight," and if voice characteristics such as increased speech speed or increased speech pauses appear together. The electronic device (200) calculates the directionality score based on whether only symptoms associated with a specific disorder group are continuously increasing for a certain period, and if the calculated score is higher than the reference directionality score, it can evaluate that there is a pattern in which the corresponding disorder group is biased toward a risk candidate disorder.

[0118] Severity analysis may be a procedure designed to assess whether a user's risk status exceeds a critical threshold by accumulating severity scores assigned to each symptom. Severity can be a value representing the magnitude of the impact a symptom has on a user's safety or functional ability, and the severity level for each symptom may be set differently for each disease group. For example, a symptom such as "feeling lethargic" may have a low severity, whereas a utterance like "I want to die" may have a high severity based on the single item alone, as it is directly linked to suicide risk. Severity analysis can be designed to determine risk levels in both cases where multiple mild symptoms accumulate over a short period and where a single severe symptom appears. Severity analysis may consist of a process that identifies warning signs by evaluating whether the severity score aggregated over a unit period exceeds a reference threshold; factors such as the rate of increase in severity scores, the severity accumulation cycle, and whether there are rapid changes in severity may also be considered.

[0119] For example, the electronic device (200) can repeatedly accumulate low severity scores when the user continuously reports mild symptoms, such as "difficulty sleeping," over several days, and can determine whether the accumulated severity scores are gradually approaching the standard severity score. Conversely, the electronic device (200) can immediately assign a high severity score when the user mentions severe expressions, such as "unbearable" or "wanting to disappear," even once, and immediately evaluate whether the accumulated score within a unit period exceeds a threshold. Additionally, the electronic device (200) can increase the sensitivity of risk indication judgment by additionally assigning a high severity score when fluctuations in biosignals associated with severe risk, such as a sudden increase in heart rate or respiratory instability, occur.

[0120] The electronic device (200) can establish a basis for a comprehensive risk assessment by independently interpreting the results of trend analysis, direction analysis, and severity analysis, and then determining how the results are combined with each other. The electronic device (200) can determine the consistency between the analysis results and apply a high weight to analysis results with high intercorrelation and a corrected weight to analysis results with low consistency. The electronic device (200) can clearly distinguish and evaluate short-term and long-term patterns during the analysis process to reduce false positives caused by temporary emotional changes and to prioritize the detection of actual risk increase sections.

[0121] According to one example, in operation 450, the electronic device (200) can calculate the psychological state risk of the user by synthesizing the analysis results.

[0122] Here, the integrated risk score may be a comprehensive indicator designed to calculate the risk level corresponding to a user's overall psychological state by combining trend analysis, direction analysis, and severity analysis results. The integrated risk score can serve as a value to organize complex psychological changes, which are difficult to assess based solely on individual analysis results, into a single scoring system. Furthermore, it can be utilized as a criterion for determining the intensity and urgency of risk signals by quantifying risk levels for each disease group. Even when individual analysis results indicate different directions of change, the integrated risk score can be calculated using a method that considers the importance and causal contribution of each analysis result, rather than an averaging-based combination method.

[0123] The electronic device (200) can calculate an integrated risk score by assigning weights to each analysis result. The weights may represent the relative proportion that each analysis contributes to the risk judgment, and may assign a low weight to the trend analysis reflecting the progression of symptoms, a medium weight to the directional analysis judging bias toward a specific disease group, and the highest weight to the severity analysis judging the immediate risk level. The electronic device (200) can calculate a formula-based integrated risk score by combining the weighted values ​​and determine whether the calculated risk score exceeds a reference risk score to determine whether the user's condition is in a warning stage, a risk stage, or a stage requiring emergency intervention.

[0124] For example, the electronic device (200) can calculate an integrated risk score by assigning different weights to the trend analysis result, the direction analysis result, and the severity analysis result, respectively. The trend analysis result (T) may be a value representing the long-term change flow of symptoms, the direction analysis result (D) may be a value representing the degree of symptom bias belonging to a specific disease group, and the severity analysis result (S) may be a value reflecting the intensity of the appearance and the cumulative level of severe symptoms. The electronic device (200) may apply a first weight (W1), a second weight (W2), and a third weight (W3), respectively, to reflect the difference in importance of each result.

[0125] The electronic device (200) can calculate the integrated risk score (R) according to the [formula] below after applying the first weight (W1), the second weight (W2), and the third weight (W3).

[0126] [formula]

[0127] R = W1XT + W2XD + W3XS

[0128] In one embodiment, the electronic device (200) can most sensitively assess the potential for immediate risk by prioritizing the appearance of severe symptoms or rapid psychological deterioration by setting the third weight (W3) greater than the second weight (W2). By setting the second weight (W2) greater than the first weight (W1), the electronic device (200) can clearly identify the direction of disease group risk by reflecting the phenomenon of symptom bias toward a specific disease group as a more important judgment factor than the long-term change flow. By maintaining a ratio structure, the results of the severity analysis can be set to act as the highest weight in the risk judgment process. For example, the electronic device (200) may assign a high weight by considering that the severity analysis includes important signals related to immediate safety. For example, the electronic device (200) can increase the accuracy of risk interpretation by assigning a medium weight to the process of identifying the direction of the disease group and the lowest weight to the trend analysis that judges the long-term flow.

[0129] In one embodiment, the electronic device (200) can reflect the phenomenon of symptom bias associated with a specific disease group as a key indicator for risk assessment by setting the second weight (W2) to be greater than the third weight (W3). The electronic device (200) can evaluate the repetition and concentration of such expressions as major risk signals when symptom expressions associated with one of the disease groups, such as depression, anxiety, PTSD, aggression, schizophrenia, and suicide, increase intensively over a certain period. The electronic device (200) can assign a high weight considering that directional analysis detects the possibility of progression to a specific disease group early, and can prioritize the evaluation of the possibility of disease spread if expressions related to the same disease group accumulate rapidly even if the disease group bias is weak. Through such settings, the electronic device (200) treats the direction of progression of the disease group as a more important risk assessment criterion than short-term stress response, and can focus on analyzing whether a specific category of signals consistent with a specific disease group is strengthened.

[0130] In one embodiment, the electronic device (200) can reflect the long-term change flow as the most important factor in the risk assessment process by setting the first weight (W1) to be greater than the second weight (W2) or the first weight (W1) to be greater than the third weight (W3). The electronic device (200) can focus on evaluating how long-term flows, such as temporal changes in the user's emotional expression, gradual rising or falling patterns of biosignals, and cumulative changes in cognitive function measurements, relate to the deterioration process of a specific disease group. The electronic device (200) can determine whether the user's psychological state has substantially entered a deterioration phase by prioritizing the direction of change, the slope of change, and the duration of change accumulated over a certain period, rather than a one-time severe expression or a temporary state of instability. The electronic device (200) can calculate a low risk level when the long-term flow is stable even if short-term fluctuations are large, and conversely, can determine a high risk possibility when the long-term change flow continues in the direction of deterioration even if short-term changes are weak, thereby utilizing the long-term pattern as the most important judgment criterion.

[0131] The electronic device (200) can determine whether the calculated integrated risk score (R) falls within a specific threshold range and classify whether the user's psychological state is at a low-risk, medium-risk, or high-risk stage. Additionally, the electronic device (200) can calculate an independent integrated risk score for each specific disease group to identify which category among the PTSD risk, depression risk, anxiety risk, aggression risk, schizophrenia risk, and suicide risk is rising first.

[0132] The electronic device (200) recognizes the calculated integrated risk score as an early warning signal when it enters an upward phase, and can classify the risk increase pattern as a high-risk state when the risk of the same disease group increases continuously over a certain period. Conversely, the electronic device (200) can determine that the user's psychological state is changing in a direction of stabilization when the integrated risk score shows a downward trend, and can evaluate that the recovery trend is distinct when the decrease exceeds a certain standard. For example, the electronic device (200) can distinguish between short-term and long-term increases in the process of calculating the integrated risk score and can analyze the duration and repeatability of changes in the integrated risk score so as not to overestimate temporary risk increases caused by sudden short-term emotional changes.

[0133] The electronic device (200) can perform a procedure to convert the structural meaning of the integrated risk score into an actual psychological state risk level when there is a possibility that the calculated integrated risk score may exceed a certain standard. The electronic device (200) can re-evaluate the contribution of the elements constituting the integrated risk score by considering that the integrated risk score is not a simple numerical result, but is composed of a combination of the results of trend analysis, direction analysis, and severity analysis. The electronic device (200) can interpret that there is a high possibility of immediate risk if the result of severity analysis accounts for a large proportion of the integrated risk score, can determine that psychological bias toward a specific disease group is continuing if the result of direction analysis accounts for a large proportion, and can evaluate that there is a possibility of long-term deterioration if the result of trend analysis accounts for a large proportion. When the integrated risk score reaches a certain threshold, the electronic device (200) can convert the score into a psychological state risk level by mapping it to a risk system for each disease group, and can classify the risk level for each disease group into multiple stages such as a warning stage, a risk stage, and a high-risk stage.

[0134] The electronic device (200) can classify the aggregated psychological state risk by specific disease groups to determine which category among PTSD risk, depression risk, anxiety risk, aggression risk, schizophrenia risk, and suicide risk is intensifying. If the psychological state risk of a specific disease group exceeds a threshold, the electronic device (200) can increase the importance of that disease group and update the risk judgment structure so that major signals associated with that disease group are processed with higher weight in future analysis. If the psychological state risk of a specific disease group is very high, the electronic device (200) can set that disease group as a priority so that behavioral intervention, professional counseling referral, or safety measures to be provided in subsequent steps are adjusted to suit the user's risk state.

[0135] The electronic device (200) determines whether actual intervention is necessary based on the final calculated psychological state risk level and can select subsequent intervention procedures, such as suggesting adaptive behaviors, providing psychological stabilization content, or recommending expert counseling, depending on the risk level. The electronic device (200) can immediately provide a notification to the user if the integrated risk score exceeds a threshold level, and can determine that professional intervention is required if multiple signals associated with a specific high-risk disease group are detected. The electronic device (200) can record the psychological state risk level over the long term to continuously evaluate whether there has been improvement compared to the past, and can also compare changes in risk levels before and after intervention to determine the effectiveness of the intervention content.

[0136] FIG. 6 is a diagram illustrating the operation (600) of the electronic device (200) of the present invention detecting a user's psychological inflection point and providing a solution for improving the psychological state.

[0137] Referring to FIG. 6, the electronic device (200) tracks the flow of psychological changes of the user in real time and can determine the point at which a danger signal rapidly increases or emotional balance is disrupted in the flow of psychological changes as a psychological inflection point. A psychological inflection point may refer to a moment requiring immediate psychological intervention, such as when the user's emotional stability rapidly deteriorates, when a severe danger signal appears, or when the rate of change accelerates in the direction of symptom exacerbation. To detect psychological inflection points, the electronic device (200) may simultaneously monitor at least one of biosignals, linguistic signals, inconsistency signals, and a calculated psychological state risk level. Biosignals may be physiological responses related to emotional tension, such as heart rate, breathing patterns, and muscle tone, and linguistic signals may be speech characteristics such as an increase in specific emotional expressions, changes in speech speed, or increased pauses. Inconsistency signals are signals indicating a contradiction between linguistic expressions and non-linguistic responses; for example, if a person says "I'm fine" while accompanied by vocal tremors or an increase in heart rate, this can be identified as a potential danger signal.

[0138] The electronic device (200) can determine whether a psychological inflection point is imminent by comprehensively comparing the signals, and if an inflection point is confirmed, it can immediately provide a dynamic solution. The dynamic solution may be an intervention method that is automatically determined based on the user's risk state and psychological response characteristics. For example, the electronic device (200) can determine an inflection point when the user's psychological signal exceeds a specific threshold or when a risk pattern accumulates over a certain period, and then automatically select an intervention method based on the characteristics of the inflection point and the urgency of the user's condition. For example, the electronic device (200) may determine that immediate emotional stabilization intervention is necessary by determining that there is an acute deterioration of emotion if the heart rate fluctuation range in the bio-signal increases rapidly, the frequency of negative emotional expression in the verbal signal is repeated for a short period, and inconsistency signals are detected above a certain threshold. In such cases, the electronic device (200) may determine that immediate physical and emotional stability recovery is more important than simple cognitive stimulation or question presentation, and may first provide psychological stabilization content such as deep breathing training, abdominal breathing guidance, and muscle relaxation content.

[0139] For example, AI conversational intervention can be the first initial intervention method applied when a relatively minor psychological inflection point is detected. It serves as an early buffer to act before the user's emotional state rapidly deteriorates. AI conversational intervention can provide empathetic feedback regarding negative emotions or concerns expressed by the user, present leading questions for emotional regulation, and conduct ultra-short-term thought restructuring conversations to mitigate automatic negative thoughts. AI conversational intervention can activate immediately if expressions of anxiety increase or negative emotional words appear in the user's verbal signals; furthermore, even if the user says "I'm fine," it can assess potential danger and automatically initiate a conversation focused on emotional stabilization if signs of slowed speech, increased pauses, or non-verbal tension are detected. While AI conversational intervention can provide a meaningful stabilizing effect on its own, it can also serve as an intermediate buffering step that naturally connects to subsequent, more advanced intervention modules.

[0140] For example, cognitive behavioral training-based psychological improvement content may be a structured intervention method that can be provided when cognitive distortion patterns increase in the user's linguistic signals and self-reference data. Cognitive behavioral training-based psychological improvement content can induce the user to re-evaluate their thoughts when they repeat distorted thought patterns, such as interpreting situations in extremes, assuming the future is excessively negative, or overgeneralizing experiences of failure. For example, if the user repeats absolute or total negative expressions such as "I always fail" or "Nobody cares about me," the electronic device (200) may sequentially provide an automatic thought recording tool, a thought-emotion connection exploration tool, and alternative thought presentation content. Cognitive behavioral training-based psychological improvement content is configured to allow the user to self-examine and modify their thought structure, thereby performing the function of supporting not only short-term emotional relief but also long-term thought pattern improvement.

[0141] For example, psychological stabilization content may be an immediate stabilization intervention module that can be provided when strong physiological anxiety responses, such as a rapid increase in heart rate, respiratory imbalance, or increased muscle tone, are detected in biosignals. The psychological stabilization content may consist of procedures to rapidly stabilize physical responses, such as breathing exercises, muscle relaxation, and mindfulness-based body sensation control techniques. When the electronic device (200) analyzes the user's bio-data patterns and determines a state of psychological over-excitement or a state of tension at the level of panic, it may determine that physiological stabilization should take precedence over existing conversation-based intervention and call the psychological stabilization content first. The psychological stabilization content may be configured to relieve physical tension in a short period of time through visual guides, auditory-based rhythm synchronization, or simple behavioral instructions, and can then be naturally connected to CBT content or AI conversational intervention.

[0142] For example, VR-based desensitization content may be an in-depth intervention module that can be provided when the user’s symptoms include hyperarousal reactions, expressions of trauma re-experience, or warning signs related to PTSD. VR-based desensitization content may be content that supports the user in facing traumatic elements in stages, starting from low intensity, within a safe virtual space without being placed in a dangerous situation. The electronic device (200) can automatically select VR-based desensitization content at an inflection point where PTSD indicators increase, initially present a relatively low-stimulation environment, and adjust the stimulation intensity by analyzing biological responses in real time. For example, if the user’s heart rate fluctuation range is stable, the stimulation intensity can be increased by one level, and if an increase in heart rate or irregular breathing is detected, the stimulation intensity can be immediately lowered to ensure safety. VR-based desensitization content can provide effects similar to exposure therapy conducted in a real counseling room environment, while supporting the user in experiencing psychological desensitization in a safer and more controllable environment.

[0143] For example, referral to expert counseling may be a priority safety intervention provided when very high severity signals are detected at inflection points, or when suicide-related speech, expressions of extreme despair, or a surge in warning signs occurs. Referral to expert counseling may include a procedure in which the electronic device (200) automatically recommends a counseling connection and, upon user consent, immediately connects the user to a professional counselor or medical staff. The electronic device (200) may automatically configure a risk information packet to be delivered to the counselor, providing a summary of recent conversation history, major warning signs, data on changes in vital signs, and an assessed psychological state risk level. Referral to expert counseling may be designed to ensure safety through immediate human expert intervention when the user is in a high-risk state and to support follow-up treatment plans. Referral to expert counseling has a higher priority than all automated interventions and may be utilized as an essential entry step, particularly when warning signs of suicidal thoughts or schizophrenia at the level of reality distortion are detected.

[0144] That is, the electronic device (200) can automatically select and provide at least one of a plurality of intervention options based on the results of psychological inflection point detection, and can track the effect of the provided intervention and reflect it in the subsequent inflection point judgment and solution selection process. In this way, the electronic device (200) can adaptively respond to the user's unique psychological response characteristics and risk patterns to support continuous and precise improvement of the psychological state.

[0145] Accordingly, the electronic device according to the present disclosure and the method for determining psychological state risk using the same can quantitatively identify actual psychological changes of a user by combining trend analysis, direction analysis, and severity analysis performed based on the user's symptoms, signs of danger, and time-series data. Unlike conventional technology that uses fragmentary point-in-time data, the electronic device and the method for determining psychological state risk using the same can rapidly and accurately determine the possibility of psychological deterioration or recovery because they determine the psychological state based on change patterns.

[0146] In addition, the electronic device of the present disclosure and the method for determining psychological state risk using the same can detect potential emotional states by quantitatively analyzing inconsistency signals between language signals, voice signals, and biosignals, and can predict future trends in psychological state changes based on event identifier-based time series analysis and the creation of a virtual user model.

[0147] In addition, the electronic device of the present disclosure and the method for determining psychological state risk using the same can clearly present the basis for calculating psychological state risk through an explainable artificial intelligence analysis module and can ensure transparency in the analysis process.

[0148] In addition, the electronic device and the method for determining psychological state risk using it can automatically detect psychological inflection points and, based on the detection result, immediately provide at least one of AI conversational intervention, cognitive behavioral training-based content, psychological stabilization content, VR-based desensitization content, and referral to expert counseling.

[0149] Accordingly, the electronic device of the present disclosure and the method for determining psychological state risk using the same can automatically adjust the intervention intensity and timing suitable for the user's condition, thereby improving the preventive effect and reliability of psychological management.

[0150] However, as this has been explained above, a redundant explanation thereof will be omitted.

[0151] Using the embodiments of the present invention described above, those skilled in the art will be able to easily make various changes and modifications within the scope of the essential characteristics of the present invention. The content of each claim of the patent claims may be combined with other claims that are not related by reference within the scope of what can be understood from this specification.

Claims

Claim 1 An electronic device for determining the risk level of a user's psychological state based on individual psychological changes obtained from an AI conversation, comprising: a memory for storing at least one instruction; The system includes at least one processor that executes the above at least one instruction, wherein the at least one processor performs an AI conversation including an adaptive survey for psychological analysis of a user, determines the user's symptoms by synthesizing self-referential data and monitoring data derived from the AI ​​conversation, determines the user's risk indicators based on the symptoms, performs trend analysis, direction analysis, and severity analysis for the symptoms and the risk indicators, respectively, and calculates the psychological state risk of the user by synthesizing the trend analysis, direction analysis, and severity analysis, wherein the at least one processor, in the trend analysis process, filters noise data corresponding to transient fluctuations in a multi-data signal including at least one of the user's survey response data, biosignal data, and cognitive function measurement data accumulated in chronological order, and analyzes whether the user's symptoms worsen or recover by comparing the trend of the previously stored past multi-data signal with the trend of the calculated multi-data signal, wherein the at least one processor, in the direction analysis process, if multiple symptoms corresponding to a specific disease group among the user's symptoms appear continuously or repeatedly, the specific disease group is classified as a risk candidate disease. An electronic device that determines, calculates a directional score for the above-mentioned risk candidate disease group, and performs the directional analysis by comparing the above-mentioned directional score with a reference directional score. Claim 2 An electronic device according to claim 1, wherein at least one processor analyzes the correlation between the symptom and the risk sign based on whether the symptom and the risk sign show the same increasing or decreasing trend over time, or whether the symptom and the risk sign show different change patterns, by comparing the mutual change amounts of the symptom and the risk sign. Claim 3 delete Claim 4 delete Claim 5 An electronic device according to claim 1, wherein at least one processor accumulates and sums symptom-specific severity scores assigned to each of the symptoms during the severity analysis process, and determines whether the severity score accumulated over a preset unit period exceeds a reference severity score. Claim 6 In claim 5, the electronic device [Formula] R = W1XT + W2XD + W3XS, wherein at least one processor assigns a first weight (W1) to the trend analysis result (T), a second weight (W2) to the direction analysis result (D), and a third weight (W3) to the severity analysis result (S), respectively, calculates a weighted integrated risk score (R) based on the [Formula] below, and calculates the psychological state risk for a specific disease group based on the integrated risk score (R). Claim 7 An electronic device according to claim 6, characterized in that the third weight (W3) is greater than the second weight (W2), and the second weight (W2) is greater than the first weight (W1). Claim 8 An electronic device according to claim 1, wherein the at least one processor monitors at least one of a biosignal, a language signal, a discrepancy signal, and a psychological state risk level based on a pattern of change in the symptom or a pattern of change in the risk sign, detects a psychological inflection point by synthesizing the biosignal, the language signal, the discrepancy signal, and the psychological state risk level, and automatically provides a dynamic solution based on the detection of the psychological inflection point, wherein the dynamic solution includes at least one of AI conversational intervention, cognitive behavioral training-based psychological improvement content, psychological stabilization content, VR-based desensitization content, and expert counseling linkage.

Images