CBT therapy-driven AI personalized psychological rehabilitation path generation system

Abstract

The invention relates to the technical field of psychological rehabilitation, and discloses a CBT therapy-driven AI personalized psychological rehabilitation path generation system, which comprises a data acquisition and evaluation module, a knowledge graph construction module, an AI path planning module, a dynamic adjustment module and an interactive feedback module, the data acquisition and evaluation module is used for acquiring basic information, psychological state evaluation data and cognitive behavior characteristic data of a user, and generating an initial psychological problem diagnosis result based on a CBT evaluation scale; the mapping knowledge domain construction module is used for constructing a CBT therapy mapping knowledge domain. A preliminary personalized psychological rehabilitation path is generated through a reinforcement learning algorithm, a CBT therapy is systematized and structured through a knowledge graph construction module, it is ensured that the generated rehabilitation path meets evidence-based practice standards, and the problem that a traditional AI system lacks theoretical support is solved; on the basis of multi-dimensional user data and a reinforcement learning algorithm, a rehabilitation path adaptive to individual cognitive features and problem types is generated, and the limitation of a standardized scheme is avoided.

Description

Technical Field

[0001] This invention relates to the field of psychological rehabilitation technology, specifically to an AI-driven personalized psychological rehabilitation pathway generation system driven by CBT therapy. Background Technology

[0002] Cognitive behavioral therapy (CBT) is a widely used evidence-based psychotherapy method internationally. It improves mental state by changing an individual's negative cognitions and maladaptive behaviors, and has significant efficacy in treating common psychological problems such as depression, anxiety, and phobias. However, the implementation of traditional CBT therapy has the following limitations:

[0003] With the development of artificial intelligence technology, AI-assisted psychological intervention has become an important direction for solving the above problems. However, existing AI systems mostly focus on single functions (such as emotion recognition and scale assessment), failing to deeply integrate the complete theoretical system of CBT therapy and thus unable to generate systematic and personalized rehabilitation pathways.

[0004] Existing CBT programs are mostly standardized processes, making it difficult to adapt to the individual differences of different users (such as age, cultural background, causes of psychological problems, cognitive styles, etc.), resulting in poor rehabilitation outcomes for some users. During the rehabilitation process, the user's psychological state constantly changes, but traditional programs rely on counselors to conduct regular assessments and adjustments, which suffers from long feedback cycles and untimely responses. Summary of the Invention

[0005] The purpose of this invention is to provide an AI-driven personalized psychological rehabilitation pathway generation system driven by CBT therapy to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a CBT-driven AI-based personalized psychological rehabilitation path generation system, comprising:

[0007] The data collection and evaluation module is used to collect users' basic information, psychological state evaluation data, cognitive behavioral characteristic data, and generate initial psychological problem diagnosis results based on the CBT assessment scale.

[0008] The knowledge graph construction module is used to construct a knowledge graph for CBT therapy. The knowledge graph includes the core concepts, intervention techniques, typical cases, stage goals, and corresponding implementation strategies of cognitive behavioral therapy.

[0009] The AI ​​path planning module, based on the output of the data collection and evaluation module and the CBT knowledge graph of the knowledge graph construction module, generates a preliminary personalized psychological rehabilitation path through reinforcement learning algorithms. The path includes phased intervention goals, corresponding CBT methods, practice tasks and expected effects. The CBT methods include: cognitive restructuring, behavioral activation and exposure therapy.

[0010] The dynamic adjustment module is used to collect feedback data from users in real time during the rehabilitation process. By comparing the deviation between the actual effect and the expected goal, the rehabilitation path is dynamically optimized. The feedback data includes task completion status, changes in psychological state, and cognitive adjustment effect.

[0011] The interactive feedback module visually displays the rehabilitation path to the user, receives the user's operation instructions and feedback information, and transmits them to the dynamic adjustment module. It also provides CBT-related knowledge explanations and task guidance.

[0012] Optionally, the data acquisition and evaluation module includes a standardized scale unit and a dynamic monitoring unit. The standardized scale unit integrates scales such as PHQ-9, GAD-7, and Automated Thinking Questionnaire (ATQ). The dynamic monitoring unit analyzes the user's cognitive behavior patterns through the user's daily text records, voice emotion data, and behavior logs.

[0013] Optionally, the knowledge graph in the knowledge graph construction module contains multi-level nodes. The first-level nodes are the core stages of CBT, which include psychoeducation, cognitive assessment, behavioral intervention, and relapse prevention. The second-level nodes are the intervention techniques corresponding to each stage. The third-level nodes are the implementation steps and applicable scenarios of the techniques. The synergistic relationship between the techniques is represented by the association weight between the nodes.

[0014] Optionally, the AI ​​path planning module employs an improved Deep Q-Network (DQN) algorithm, using the user's psychological state as the state space, CBT intervention technology as the action space, and the improvement value of the intervention effect as the reward function, to generate the optimal rehabilitation path sequence through iterative learning.

[0015] Optionally, the dynamic adjustment module includes a deviation analysis unit and a path correction unit. The deviation analysis unit determines the deviation value by calculating the cosine similarity between the user's actual psychological state and the expected state. When the deviation value exceeds a preset threshold, the path correction unit calls on alternative intervention techniques in the knowledge graph to correct the path.

[0016] This invention also proposes an AI-based personalized psychological rehabilitation pathway generation method based on CBT therapy, comprising the following steps:

[0017] S1: Collect user data through the data collection and evaluation module, conduct psychological state assessment, and generate an initial diagnostic report that includes the type of psychological problem, its severity, and core cognitive biases;

[0018] S2: The knowledge graph construction module calls the CBT therapy knowledge base to construct a knowledge graph that includes intervention techniques, implementation conditions, and effect relationships;

[0019] S3: The AI ​​path planning module takes the initial diagnostic report as input and generates a personalized rehabilitation path based on knowledge graphs and reinforcement learning algorithms, which includes stage goals, weekly tasks, and CBT technology application solutions.

[0020] S4: Users obtain and execute the rehabilitation path through the interactive feedback module, and the system collects user feedback data in real time;

[0021] S5: The dynamic adjustment module analyzes the deviation between the feedback data and the expected target. If the deviation is within the allowable range, the current path is maintained; if the deviation exceeds the threshold, subsequent intervention techniques and task difficulty are automatically adjusted.

[0022] S6: Repeat steps S4-S5 until the user's psychological state reaches the preset recovery standard, and generate a relapse prevention plan.

[0023] Optionally, the cognitive bias analysis in step S1 is implemented through natural language processing technology, including identifying and classifying negative automatic thoughts (such as "black and white" and "catastrophizing") in user text data, and calculating the frequency and intensity of each type of cognitive bias.

[0024] Optionally, the rehabilitation path generated in step S3 includes a phased division in terms of time dimension, with each phase setting 1-2 core goals and corresponding to 3-5 CBT practice tasks (such as cognitive diary recording, behavioral experiments, and progressive exposure exercises), and specifying the implementation frequency and evaluation criteria of the tasks.

[0025] Optionally, the dynamic adjustment rules in step S5 are built based on the experience of CBT experts, including: reducing the difficulty of the task and adding guidance instructions when the user fails to complete the task for two consecutive weeks; and replacing a certain intervention technique with the alternative technique with the highest relevance in the knowledge graph when the intervention technique is ineffective for three consecutive times.

[0026] Optionally, an effectiveness verification step is also included: by comparing the user's psychological assessment scale scores, the degree of improvement in cognitive bias, and the recovery status of social function before and after rehabilitation, the effectiveness of the generated personalized rehabilitation path is verified, and the verification results are fed back to the AI ​​path planning module to optimize the algorithm parameters.

[0027] Compared with existing technologies, this invention provides an AI-driven personalized psychological rehabilitation path generation system driven by CBT therapy, which has the following beneficial effects:

[0028] 1. This CBT therapy-driven AI-based personalized psychological rehabilitation path generation system systematizes and structures CBT therapy through a knowledge graph construction module, ensuring that the generated rehabilitation path conforms to evidence-based practice standards and solving the problem of traditional AI systems lacking theoretical support. Based on multi-dimensional user data and reinforcement learning algorithms, it generates rehabilitation paths that are adapted to individual cognitive characteristics and problem types, avoiding the limitations of standardized solutions.

[0029] 2. This CBT-driven AI-based personalized psychological rehabilitation pathway generation system optimizes intervention pathways in a timely manner through real-time feedback and deviation analysis, solving the problem of lagging adjustments in traditional therapies and improving rehabilitation efficiency. The system can autonomously generate and adjust pathways, reducing real-time dependence on counselors and improving the accessibility of CBT therapy. Through knowledge graphs and interactive feedback modules, the system clearly explains the design basis of rehabilitation pathways to users, enhancing user compliance. Attached Figure Description

[0030] Figure 1 This is a schematic diagram of the structure of the present invention. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] like Figure 1 As shown, this invention provides a technical solution: a CBT-driven AI-based personalized psychological rehabilitation path generation system, comprising: a data acquisition and evaluation module, a knowledge graph construction module, an AI path planning module, a dynamic adjustment module, and an interactive feedback module.

[0033] The data collection and evaluation module collects users' basic information, psychological state assessment data, and cognitive behavioral characteristic data, and generates initial psychological problem diagnosis results based on the CBT assessment scale. The knowledge graph construction module constructs a CBT therapy knowledge graph, which includes the core concepts, intervention techniques, typical cases, stage goals, and corresponding implementation strategies of cognitive behavioral therapy. The AI ​​path planning module, based on the output of the data collection and evaluation module and the CBT knowledge graph of the knowledge graph construction module, generates a preliminary personalized psychological rehabilitation path through reinforcement learning algorithms. The path includes stage intervention goals, corresponding CBT methods, practice tasks, and expected effects. CBT methods include: cognitive restructuring, behavioral activation, and exposure therapy.

[0034] The dynamic adjustment module is used to collect user feedback data in real time during the rehabilitation process. By comparing the deviation between the actual effect and the expected goal, the rehabilitation path is dynamically optimized. The feedback data includes task completion status, changes in psychological state, and cognitive adjustment effect. The interactive feedback module is used to visualize the rehabilitation path for users, receive user operation instructions and feedback information, and transmit them to the dynamic adjustment module. At the same time, it provides CBT-related knowledge explanations and task guidance.

[0035] The data collection and evaluation module includes a standardized scale unit and a dynamic monitoring unit. The standardized scale unit integrates scales such as PHQ-9, GAD-7, and Automated Thinking Questionnaire (ATQ). The dynamic monitoring unit analyzes users' cognitive and behavioral patterns through daily text records, voice emotion data, and behavior logs.

[0036] The knowledge graph in the knowledge graph construction module contains multi-level nodes. The first-level nodes are the core stages of CBT, which include psychoeducation, cognitive assessment, behavioral intervention, and relapse prevention. The second-level nodes are the intervention techniques corresponding to each stage. The third-level nodes are the implementation steps and applicable scenarios of the techniques. The synergistic relationship between the techniques is represented by the association weight between the nodes.

[0037] The AI ​​path planning module uses an improved Deep Q-Network (DQN) algorithm, which uses the user's psychological state as the state space, CBT intervention technology as the action space, and the improvement value of the intervention effect as the reward function. It generates the optimal rehabilitation path sequence through iterative learning.

[0038] The dynamic adjustment module includes a deviation analysis unit and a path correction unit. The deviation analysis unit determines the deviation value by calculating the cosine similarity between the user's actual psychological state and the expected state. When the deviation value exceeds a preset threshold, the path correction unit calls on alternative intervention techniques in the knowledge graph to correct the path.

[0039] This embodiment also proposes an AI-based personalized psychological rehabilitation pathway generation method based on CBT therapy, including the following steps:

[0040] S1: Collect user data through the data collection and evaluation module, conduct psychological state assessment, and generate an initial diagnostic report that includes the type, severity, and core cognitive biases of psychological problems. Cognitive bias analysis is achieved through natural language processing technology, including the identification and classification of negative automatic thoughts in user text data, and the calculation of the frequency and intensity of each type of cognitive bias.

[0041] S2: The knowledge graph construction module calls the CBT therapy knowledge base to construct a knowledge graph that includes intervention techniques, implementation conditions, and effect relationships;

[0042] S3: The AI ​​path planning module takes the initial diagnostic report as input and generates a personalized rehabilitation path based on knowledge graphs and reinforcement learning algorithms. This path includes stage goals, weekly tasks, and CBT technology application solutions. The generated rehabilitation path includes a time-divided stage, with each stage having 1-2 core goals and corresponding to 3-5 CBT practice tasks. The CBT practice tasks include cognitive diary recording, behavioral experiments, and progressive exposure exercises, and the frequency of task implementation and evaluation criteria are clearly defined.

[0043] S4: Users obtain and execute the rehabilitation path through the interactive feedback module, and the system collects user feedback data in real time;

[0044] S5: The dynamic adjustment module analyzes the deviation between the feedback data and the expected goal. If the deviation is within the allowable range, the current path is maintained; if the deviation exceeds the threshold, subsequent intervention techniques and task difficulty are automatically adjusted. The dynamic adjustment rules are built based on the experience of CBT experts and include: when a user fails to complete a task for two consecutive weeks, the task difficulty is reduced and guidance is added; when a certain intervention technique is ineffective for three consecutive times, it is replaced with the alternative technique with the highest relevance in the knowledge graph.

[0045] S6: Repeat steps S4-S5 until the user's psychological state reaches the preset recovery standard, and generate a relapse prevention plan.

[0046] S6. Effect Verification Step: By comparing the user's psychological assessment scale scores, the degree of improvement in cognitive bias, and the recovery status of social function before and after rehabilitation, the effectiveness of the generated personalized rehabilitation path is verified, and the verification results are fed back to the AI ​​path planning module to optimize the algorithm parameters.

[0047] As one application of this embodiment:

[0048] This embodiment systematizes and structures CBT therapy through a knowledge graph construction module, ensuring that the generated rehabilitation path conforms to evidence-based practice standards and solving the problem of traditional AI systems lacking theoretical support. Based on multi-dimensional user data and reinforcement learning algorithms, it generates rehabilitation paths that are adapted to individual cognitive characteristics and problem types, avoiding the limitations of standardized solutions.

[0049] The present invention has been described in detail above. However, modifications or improvements can be made to it, which will be obvious to those skilled in the art. Therefore, any modifications or improvements that do not depart from the spirit of the present invention are within the scope of protection of the present invention.

Claims

1. A CBT-driven AI-based personalized psychological rehabilitation pathway generation system, characterized in that, include: The data collection and evaluation module is used to collect users' basic information, psychological state evaluation data, cognitive behavioral characteristic data, and generate initial psychological problem diagnosis results based on the CBT assessment scale. The knowledge graph construction module is used to construct a knowledge graph for CBT therapy. The knowledge graph includes the core concepts, intervention techniques, typical cases, stage goals, and corresponding implementation strategies of cognitive behavioral therapy. The AI ​​path planning module, based on the output of the data collection and evaluation module and the CBT knowledge graph of the knowledge graph construction module, generates a preliminary personalized psychological rehabilitation path through reinforcement learning algorithms. The path includes phased intervention goals, corresponding CBT methods, practice tasks and expected effects. The CBT methods include: cognitive restructuring, behavioral activation and exposure therapy. The dynamic adjustment module is used to collect feedback data from users in real time during the rehabilitation process. By comparing the deviation between the actual effect and the expected goal, the rehabilitation path is dynamically optimized. The feedback data includes task completion status, changes in psychological state, and cognitive adjustment effect. The interactive feedback module visually displays the rehabilitation path to the user, receives the user's operation instructions and feedback information, and transmits them to the dynamic adjustment module. It also provides CBT-related knowledge explanations and task guidance.

2. The CBT-driven AI-based personalized psychological rehabilitation path generation system according to claim 1, characterized in that, The data acquisition and evaluation module includes a standardized scale unit and a dynamic monitoring unit. The standardized scale unit integrates scales such as PHQ-9, GAD-7, and Automated Thinking Questionnaire (ATQ). The dynamic monitoring unit analyzes the user's cognitive behavior patterns through the user's daily text records, voice emotion data, and behavior logs.

3. The CBT therapy-driven AI-based personalized psychological rehabilitation path generation system according to claim 1, characterized in that, The knowledge graph in the knowledge graph construction module contains multi-level nodes. The first-level nodes are the core stages of CBT, which include psychoeducation, cognitive assessment, behavioral intervention, and relapse prevention. The second-level nodes are the intervention techniques corresponding to each stage. The third-level nodes are the implementation steps and applicable scenarios of the techniques. The synergistic relationship between the techniques is represented by the association weight between the nodes.

4. The CBT therapy-driven AI-based personalized psychological rehabilitation path generation system according to claim 1, characterized in that, The AI ​​path planning module uses an improved Deep Q-Network (DQN) algorithm, taking the user's psychological state as the state space, CBT intervention technology as the action space, and the improvement value of the intervention effect as the reward function, and generates the optimal rehabilitation path sequence through iterative learning.

5. The CBT-driven AI-based personalized psychological rehabilitation path generation system according to claim 1, characterized in that, The dynamic adjustment module includes a deviation analysis unit and a path correction unit. The deviation analysis unit determines the deviation value by calculating the cosine similarity between the user's actual psychological state and the expected state. When the deviation value exceeds a preset threshold, the path correction unit calls on alternative intervention technology in the knowledge graph to correct the path.

6. A method for generating AI-based personalized psychological rehabilitation pathways based on CBT therapy, characterized in that, Includes the following steps: S1: Collect user data through the data collection and evaluation module, conduct psychological state assessment, and generate an initial diagnostic report that includes the type of psychological problem, its severity, and core cognitive biases; S2: The knowledge graph construction module calls the CBT therapy knowledge base to construct a knowledge graph that includes intervention techniques, implementation conditions, and effect relationships; S3: The AI ​​path planning module takes the initial diagnostic report as input and generates a personalized rehabilitation path based on knowledge graphs and reinforcement learning algorithms, which includes stage goals, weekly tasks, and CBT technology application solutions. S4: Users obtain and execute the rehabilitation path through the interactive feedback module, and the system collects user feedback data in real time; S5: The dynamic adjustment module analyzes the deviation between the feedback data and the expected target. If the deviation is within the allowable range, the current path is maintained; if the deviation exceeds the threshold, subsequent intervention techniques and task difficulty are automatically adjusted. S6: Repeat steps S4-S5 until the user's psychological state reaches the preset recovery standard, and generate a relapse prevention plan.

7. The method for generating AI-based personalized psychological rehabilitation pathways based on CBT therapy according to claim 6, characterized in that, The cognitive bias analysis in step S1 is achieved through natural language processing technology, including identifying and classifying negative automatic thoughts (such as "black and white" and "catastrophizing") in user text data, and calculating the frequency and intensity of each type of cognitive bias.

8. The method for generating AI-based personalized psychological rehabilitation pathways based on CBT therapy according to claim 6, characterized in that, The rehabilitation path generated in step S3 includes a phased division in terms of time. Each phase sets 1-2 core goals and corresponds to 3-5 CBT practice tasks (such as cognitive diary recording, behavioral experiments, and progressive exposure exercises), and clarifies the implementation frequency and evaluation criteria of the tasks.

9. The method for generating AI-based personalized psychological rehabilitation pathways based on CBT therapy according to claim 6, characterized in that, The dynamic adjustment rules in step S5 are built based on the experience of CBT experts and include: when a user fails to complete a task for two consecutive weeks, the task difficulty is reduced and guidance is increased; when an intervention technique is ineffective for three consecutive times, it is replaced with the alternative technique with the highest relevance in the knowledge graph.

10. The method for generating AI-based personalized psychological rehabilitation pathways based on CBT therapy according to claim 6, characterized in that, It also includes an effectiveness verification step: by comparing the user's psychological assessment scale scores, the degree of improvement in cognitive bias, and the recovery status of social function before and after rehabilitation, the effectiveness of the generated personalized rehabilitation path is verified, and the verification results are fed back to the AI ​​path planning module to optimize the algorithm parameters.

Images