Old people cognitive health monitoring system based on virtual reality

Abstract

The invention discloses an old people cognitive health monitoring system based on virtual reality, and relates to the related technical field of virtual health monitoring technologies. Comprising a multi-modal sensing acquisition module, a virtual reality environment generation module, a cognitive nerve feedback module, a real-time feedback module, a deep learning analysis module, a personalized training plan generation module, a data storage and management module, a remote medical platform and a virtual assistant and coach module. The system can help the elderly to improve the work memory ability through a special work memory training task and a cognitive nerve feedback module, and the elderly can be more efficient during daily life activities by improving the work memory and the attention basic cognitive function aiming at the common attention distraction problem of the elderly, so that the life quality is improved, and the elderly can be helped to take care of the elderly. The cognitive neural feedback module not only pays attention to a cognitive function, but also helps the elderly to better manage emotions by adjusting an emotional network, promotes social interaction, and reduces aloneness and anxiety.

Description

Technical Field

[0001] The present invention relates to the technical field related to virtual health monitoring technology, and in particular to a cognitive health monitoring system for the elderly based on virtual reality. Background Art

[0002] As the global population aging problem intensifies, the health status and quality of life of the elderly population have gradually become the focus of social attention. As an important part of the health of the elderly, the decline in cognitive health not only affects the individual's daily living ability, but may also lead to a series of psychological and social problems. Therefore, the monitoring and intervention of the cognitive health of the elderly is particularly important. Traditional methods of cognitive health monitoring for the elderly rely on face-to-face assessment and training, which is not only time-consuming and labor-intensive, but also limited by the number and regional distribution of professionals.

[0003] In recent years, the rapid development of virtual reality (VR) technology has provided new possibilities for solving the above problems. VR technology can create an immersive simulation environment and provide users with a new interactive experience through highly simulated visual, auditory and tactile stimulation. In the field of cognitive health, VR technology has begun to be tried for the assessment and rehabilitation training of cognitive impairment, and has shown good application prospects.

[0004] However, the current VR cognitive health monitoring system still has some shortcomings: 1. It cannot improve the poor cognitive ability and memory of the elderly, which makes them prone to loneliness and anxiety. 2. There is a lack of analysis of the health data of the elderly and assessment of potential risks. The elderly are too dependent on hospitals, and excessive consumption of medical resources has increased social costs.

[0005] Therefore, it is necessary to propose a cognitive health monitoring system for the elderly based on virtual reality to solve the above problems. Summary of the invention

[0006] 1. Technical issues to be resolved

[0007] In view of the shortcomings of the existing technology, the present invention provides a cognitive health monitoring system for the elderly based on virtual reality, which solves the problem that the existing health monitoring system cannot improve the elderly's poor cognitive ability and memory, which makes the elderly prone to loneliness and anxiety, lacks analysis of the elderly's health data and assessment of potential risks, and the elderly are too dependent on hospitals, and excessive consumption of medical resources leads to increased social costs.

[0008] (II) Technical solution

[0009] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0010] A virtual reality-based cognitive health monitoring system for the elderly, the system comprising:

[0011] Multimodal sensor acquisition module: including VR helmets, wearable devices, and home sensors to comprehensively collect users' physiological and behavioral data;

[0012] Virtual reality environment generation module: Create a variety of immersive scenes, including simulations of daily life scenes, memory games, and spatial navigation tasks;

[0013] Cognitive neurofeedback module: monitors the user's brain activity in real time through an electroencephalogram (EEG) headband neural monitoring device, and feeds this information back to the training environment of the virtual reality environment generation module to adjust the difficulty and content of the training;

[0014] Real-time feedback module: The real-time feedback module is an important part of the neurofeedback module. It provides real-time feedback to users in the form of visual or auditory feedback after decoding brain activity patterns. Users learn how to regulate their brain activities to achieve cognitive enhancement. This feedback mechanism not only promotes the self-regulation ability of users, but also provides researchers with instant data support to adjust training plans.

[0015] Deep learning analysis module: uses deep learning networks to analyze the user's performance data, physiological feedback, and environmental data, identifies subtle changes in cognitive ability, and predicts the risk of cognitive decline, and transmits the information to the server terminal;

[0016] Personalized training plan generation module: Combines deep learning analysis results to develop highly personalized training plans and dynamically adjusts difficulty and content;

[0017] Data storage and management module: stores user data and connects with the medical and health database to provide doctors with auxiliary diagnosis information;

[0018] Telemedicine platform: Provides a remote monitoring and guidance platform for medical professionals to access patient data in real time and provide professional advice and intervention;

[0019] Virtual Assistant and Coaching Module: Built-in intelligent virtual assistant provides real-time feedback, encouragement and support, and adjusts the interaction method based on the user's emotional and cognitive state.

[0020] Optionally, the virtual reality environment generation module includes hardware equipment, software algorithms and user interfaces;

[0021] The hardware equipment includes a virtual reality helmet and a controller. The virtual reality helmet is the core hardware of the virtual reality environment generation module and is responsible for presenting the virtual environment in front of the user. The virtual reality helmet is equipped with a display screen, a head motion tracking sensor, and a viewing angle adjustment function to ensure that the user can obtain a highly realistic visual experience. The controller is used for the user to interact with the virtual environment, such as operating virtual objects or navigating. The controller includes a handheld device and a somatosensory sensor to detect the user's movements and convert them into operation instructions in the virtual environment.

[0022] The software algorithm includes 3D modeling and physics engine. 3D modeling uses computer graphics technology to create 3D objects and scenes in virtual environments, involving geometric modeling, texture mapping and lighting model algorithms. The geometric model of a 3D object is represented by a triangular mesh, and each triangle is defined by three vertices (x1, y1, z1), (x2, y2, z2), (x3, y3, z3); ​​the physics engine is used to simulate physical phenomena in the virtual environment, including gravity, collision and friction. The physics engine ensures that the behavior of the virtual environment conforms to the physical laws of the real world and improves the user's immersion. For example, the formula of a collision detection algorithm is expressed as:

[0023] d≤r1+T2

[0024] Where d is the distance between the centers of the two objects, r1 and r2 are their radii respectively;

[0025] The user interface includes an interactive interface and a feedback system. The interactive interface allows users to interact with the virtual environment through a graphical interface or other natural interactive methods, including a menu system, gesture recognition, and voice command functions, enabling users to intuitively operate the virtual world. The feedback system feeds back information in the virtual environment to the user, including visual, auditory, and tactile feedback, using headphones or speakers to provide stereo audio, and using vibrating gloves or seats to provide tactile feedback.

[0026] Optionally, the cognitive neurofeedback module includes an electroencephalogram (EEG) acquisition device, a signal processing algorithm unit, a deep learning technology unit, and a neuroscience research theoretical model:

[0027] The EEG acquisition equipment is used to collect EEG signal data. The number and distribution of sensors in the equipment are also crucial to ensure comprehensive coverage of brain activity and effective filtering of artifacts. The EEG system and sensors are the prerequisites for achieving accurate feedback.

[0028] The signal processing algorithm unit is responsible for filtering out electrical signals of non-neural origin (such as artifacts produced by blinking) from the collected EEG signals, and decoding brain patterns related to cognitive functions. It uses blind source separation technology (ICA) to filter out artifacts that interfere with alpha waves, thereby improving signal quality and increasing effective feedback time. The application of deep learning technology in the neurofeedback module is mainly reflected in the accuracy of real-time decoding of brain patterns. It is used to calibrate the specific brain patterns of the subjects and adapt to the brain activity patterns of specific subjects during online training. This technology supports personalized feedback plans, making the training effect more accurate and efficient. The theoretical model of neuroscience research is the ternary neural system model, which describes the role of emotion, reward, and cognitive control systems in decision-making behavior, and reveals how various parts of the brain work together to achieve cognitive control.

[0029] Optionally, the deep learning technique is a feedforward neural network (FFNN), in which information flows in one direction, from the input layer to the output layer, does not contain feedback connections, and is used for function approximation and classification tasks. A feedforward neural network with a single hidden layer, the hidden layer output is expressed as:

[0030] h=f(Wx+b)

[0031] Among them, the input vector is x, the weight matrix is ​​W, the bias vector is b, and the activation function is f;

[0032] The final output y is calculated through another linear transformation and activation function:

[0033] y=9(Uh+v)

[0034] Among them, U and v are the weight and bias of the output layer respectively, and g is the activation function of the output layer.

[0035] Optionally, the ternary neural system model is a key theoretical framework for explaining and simulating complex behaviors of cognitive control, affective response and reward processing, and is composed of three interconnected subsystems, including a cognitive control network (CCN), an affective network (AN) and a reward network (RN);

[0036] The cognitive control network is responsible for high-level cognitive functions such as planning, decision-making, working memory, and attention control. The cognitive control network involves the prefrontal cortex, dorsolateral prefrontal cortex, and anterior cingulate cortex. The function of the cognitive control network is executive function, such as attention regulation, error detection, task switching, and goal maintenance. In mathematical models, the functions of the cognitive control network are described by linear and nonlinear equations. For example, the load of working memory is expressed as the weighted sum of input signals, which is processed by the activation function to form the output signal, which is expressed by the formula: y = f(w1x1+w2x2+...+wnxn-θ), where y is the output signal, xn is the input signal, wn is the weight, θ is the threshold, and f is the activation function;

[0037] The affective network is responsible for the generation and regulation of emotions, including emotional responses to environmental stimuli. It includes the amygdala, ventral striatum, and hippocampus regions, which process and regulate emotional responses. For example, the amygdala plays a key role in emotional responses to fear and anxiety, while the ventral striatum is involved in processing rewards and pleasure. The mathematical model of the affective network focuses on how to convert external stimuli into emotional responses, which is achieved through a threshold-based activation function that determines whether a specific emotional response is activated. For example, the activation function of the fear response is expressed as: F = g(s-φ), where F is the intensity of the fear response, s is the amount of external threat stimulus, φ is the minimum amount of stimulus required to activate the response, and g is a function that describes the relationship between the intensity of the response and the amount of stimulus.

[0038] The reward network is responsible for processing reward and incentive information from the external environment, including the mesolimbic system, nucleus accumbens and tegmental area. These areas are responsible for processing reward prediction, incentives and feelings of happiness. When an individual encounters something satisfying or pleasant, the reward network will be activated and release dopamine, a neurotransmitter associated with pleasure and satisfaction. The mathematical model of the reward network is based on reinforcement learning theory, in which the reward prediction error (Reward Prediction Error) formula is expressed as: RPEt = Rt-V(St), where Rt is the actual reward received at time t, and V(St) is the expected reward value in state S at time t.

[0039] Optionally, the real-time feedback mechanism includes sensor nodes, an Internet of Things gateway, a cloud platform and a user terminal;

[0040] Sensor nodes include physiological parameter sensors, motion sensors and environmental sensors. Physiological parameter sensors (such as heart rate, blood pressure, and blood sugar sensors) are used to collect key physiological indicators of the elderly. Motion sensors (such as accelerometers and gyroscopes) are used to capture the movement information of the elderly. Environmental sensors (such as temperature, humidity, and air quality sensors) are used to monitor the living environment of the elderly.

[0041] The IoT gateway includes data transmission and preprocessing, and threshold detection method. Data transmission and preprocessing are connected to sensor nodes through the network, receive and preprocess data, and then upload it to the cloud platform; the threshold detection method is used to detect the movement state of the human body and upload the relevant data to the server terminal;

[0042] The cloud platform stores, processes and analyzes the collected data, and pushes health status reports to users through application software;

[0043] The user terminal and the mobile phone application receive data through wireless communication, analyze, process and display the data, and give reasonable health advice based on the numerical range. The integrated video call function allows the elderly to have face-to-face consultations with doctors at home, and doctors to diagnose remotely.

[0044] Optionally, the personalized training plan generation module automatically generates a customized training plan based on the specific needs and cognitive abilities of each elderly person, including comprehensive assessment tools, data analysis and processing systems, and personalized algorithm technology, to achieve scientific, efficient, and personalized rehabilitation services;

[0045] Comprehensive assessment tools include internationally validated scales, such as the Montreal Cognitive Assessment (MoCA) and the Mini-Mental State Examination (MMSE), which provide measures of cognitive function;

[0046] Data analysis and processing system, which automatically generates personalized training plans based on the evaluation results, covering multiple cognitive areas such as memory, attention, language, and executive function;

[0047] Personalized algorithm technology, using classification and regression tree (CART) analysis, conducts in-depth analysis of the cognitive data of the elderly to determine a personalized training plan. The algorithm can dynamically adjust the difficulty and content of the training tasks based on the elderly's training feedback and progress, ensuring that the training is always challenging and attractive;

[0048] Optionally, the telemedicine platform includes sensor nodes, data transmission and preprocessing modules, a cloud platform, and user terminals;

[0049] The sensor nodes collect human physiological parameters and movement information through 4G terminal devices, and upload the relevant data to the data transmission and preprocessing module; the data transmission and preprocessing module connects to the sensor nodes through the network, receives and preprocesses the data, and then uploads it to the cloud platform; the cloud platform is responsible for data storage, processing and analysis to ensure data security and reliability; the application software and service units in the cloud platform push health status reports to users and medical staff through application software, and support remote diagnosis and health management services; the user terminal, the Android mobile phone application receives data through wireless communication, parses, processes and displays the data, and gives reasonable health advice. Medical staff obtain the health data of the elderly in real time through terminal devices and provide professional telemedicine services.

[0050] Optionally, the virtual assistant and coaching module includes a physiological parameter acquisition module, a data analysis module, a remote monitoring module, and a feedback suggestion unit;

[0051] The physiological parameter acquisition module uses 4G terminal equipment to collect physiological parameters and motion information; the data analysis module uses the threshold detection method to determine the motion state of the human body and uploads the relevant data to the server terminal; the remote monitoring module forwards the uploaded data to the feedback suggestion unit through the server terminal so that the guardian can receive the user's data remotely; the feedback suggestion unit receives data through wireless communication, parses, processes and displays the data, gives reasonable health advice, and sends the data to the user terminal.

[0052] (III) Beneficial effects

[0053] The present invention provides a cognitive health monitoring system for the elderly based on virtual reality, which has the following beneficial effects:

[0054] 1. The cognitive neurofeedback module of the present invention provides customized training tasks according to the specific needs and cognitive abilities of each elderly person. This personalized approach can more effectively promote the improvement of specific cognitive functions. Through special working memory training tasks, the cognitive neurofeedback module can help the elderly improve their working memory abilities. It targets the common problem of distraction in the elderly and improves working memory and basic cognitive functions of attention. The elderly will be more adept at daily activities, thereby improving their quality of life. The cognitive neurofeedback module not only focuses on cognitive functions, but also helps the elderly better manage emotions, promote social interactions, and reduce loneliness and anxiety by regulating emotional networks.

[0055] 2. The present invention uses biosensor technology to enable the system to monitor the key physiological indicators of the elderly, such as heart rate and blood pressure, 24 hours a day, and immediately issue an alarm when an abnormality is detected. Whether it is sudden high blood pressure at night or blood sugar fluctuations caused by daily diet, it can be discovered and notified to relevant personnel in a timely manner to ensure the health and safety of the elderly. Using big data and deep learning technology, the system can conduct personalized risk assessments based on the health history data of each elderly person and predict potential health problems;

[0056] 3. With the system of the present invention, the elderly can have face-to-face consultations with doctors at home and enjoy professional medical services without leaving home. Doctors can view the health data of the elderly through the platform and conduct remote diagnosis, which greatly improves the efficiency and quality of medical services. The application of virtual assistants and coaching modules saves medical resources and reduces dependence on professional medical staff. The system provides an effective solution for social health management and reduces social costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a module schematic diagram of the health monitoring system of the present invention;

[0058] Figure 2 It is a module schematic diagram of the telemedicine platform of the present invention;

[0059] Figure 3 A schematic diagram of a virtual assistant and a coaching module of the present invention;

[0060] Figure 4 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0061] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0062] Example 1

[0063] See also Figure 1-3 , a cognitive health monitoring system for the elderly based on virtual reality, the system includes:

[0064] Multimodal sensor acquisition module: including VR helmets, wearable devices (such as heart rate monitors, accelerometers), and home sensors (such as motion detectors and sleep monitors), comprehensively collecting users' physiological and behavioral data;

[0065] Virtual reality environment generation module: Create a variety of immersive scenes, including simulations of daily life scenes, memory games, and spatial navigation tasks;

[0066] Cognitive neurofeedback module: monitors the user's brain activity in real time through an electroencephalogram (EEG) headband neural monitoring device, and feeds this information back to the training environment of the virtual reality environment generation module to adjust the difficulty and content of the training;

[0067] Real-time feedback module: The real-time feedback module is an important part of the neurofeedback module. It provides real-time feedback to users in the form of visual or auditory feedback after decoding brain activity patterns. Users learn how to regulate their brain activities to achieve cognitive enhancement. This feedback mechanism not only promotes the self-regulation ability of users, but also provides researchers with instant data support to adjust training plans.

[0068] Deep learning analysis module: uses deep learning networks to analyze the user's performance data, physiological feedback, and environmental data, identifies subtle changes in cognitive ability, and predicts the risk of cognitive decline, and transmits the information to the server terminal;

[0069] Personalized training plan generation module: Combines deep learning analysis results to develop highly personalized training plans and dynamically adjusts difficulty and content;

[0070] Data storage and management module: stores user data and connects with the medical and health database to provide doctors with auxiliary diagnosis information;

[0071] Telemedicine platform: Provides a remote monitoring and guidance platform for medical professionals to access patient data in real time and provide professional advice and intervention;

[0072] Virtual Assistant and Coaching Module: Built-in intelligent virtual assistant provides real-time feedback, encouragement and support, and adjusts the interaction method based on the user's emotional and cognitive state.

[0073] In this embodiment, virtual reality technology can create an immersive environment, making the elderly feel as if they are in the real world during training, thereby greatly improving their participation and interest. Virtual reality technology can simulate various scenes in real life, such as shopping and cooking, to help the elderly practice daily living skills in a safe environment. The system can adjust the difficulty and content of training according to the cognitive level and progress of the elderly, ensuring that the training always meets their needs and abilities;

[0074] By collecting and analyzing training data, the system can objectively evaluate the improvement of the cognitive function of the elderly and provide therapists with a scientific basis for adjustment. The training content can gradually transition from simple to complex, ensuring that the elderly can gradually improve their abilities without feeling stressed. Virtual reality technology is not only used in cognitive rehabilitation, but can also be combined with physical rehabilitation and psychological intervention to achieve a comprehensive health management plan. Through systematic and interesting training, virtual reality technology can help delay the decline of cognitive function in the elderly and improve their quality of life.

[0075] Example 2

[0076] This embodiment is a further optimization based on the embodiment 1. Specifically, the virtual reality environment generation module provides an immersive cognitive training and assessment environment for the elderly by constructing a simulation environment. The virtual reality environment generation module includes hardware equipment, software algorithms and user interfaces.

[0077] The hardware equipment includes a virtual reality helmet and a controller. The virtual reality helmet is responsible for presenting the virtual environment in front of the user. The virtual reality helmet is equipped with a display screen, a head motion tracking sensor, and a viewing angle adjustment function to ensure that the user can obtain a highly realistic visual experience. The controller is used for the user to interact with the virtual environment, such as operating virtual objects or navigating. The controller includes a handheld device and a somatosensory sensor to detect the user's movements and convert them into operating instructions in the virtual environment.

[0078] The software algorithm includes 3D modeling and physics engine. 3D modeling uses computer graphics technology to create 3D objects and scenes in virtual environments, involving geometric modeling, texture mapping and lighting model algorithms. The geometric model of a 3D object is represented by a triangular mesh, and each triangle is defined by three vertices (x1, y1, z1), (x2, y2, z2), (x3, y3, z3); ​​the physics engine is used to simulate physical phenomena in the virtual environment, including gravity, collision and friction. The physics engine ensures that the behavior of the virtual environment conforms to the physical laws of the real world and improves the user's immersion. For example, the formula of a collision detection algorithm is expressed as:

[0079] d≤r1+r2

[0080] Where d is the distance between the centers of the two objects, r1 and r2 are their radii respectively;

[0081] The user interface includes an interactive interface and a feedback system. The interactive interface allows the user to interact with the virtual environment through a graphical interface or other natural interactive methods, including a menu system, gesture recognition, and voice command functions, so that the user can intuitively operate the virtual world. The feedback system feeds back information in the virtual environment to the user, including visual, auditory, and tactile feedback, using headphones or speakers to provide stereo audio, and using vibration gloves or seats to provide tactile feedback;

[0082] In a virtual reality environment, the position and orientation of an object are described by a three-dimensional transformation matrix, which is expressed as:

[0083]

[0084] Where (x, y, z) is the projection model of the object's translation in virtual space;

[0085] To convert a three-dimensional scene into an image on a two-dimensional screen, use the perspective projection model. The perspective projection matrix is ​​expressed as:

[0086]

[0087] Among them, a, b, c, and d are parameters that determine the field of view and viewing angle;

[0088] Combine data from different sensors, such as EEG signals and eye tracking, to more comprehensively assess the user's cognitive state, which requires complex data fusion algorithms such as Kalman filters or Bayesian networks; build models based on physiological parameters and behavioral data to assess the user's cognitive ability, for example, by analyzing the power spectral density of specific frequency bands in the EEG signal to assess the user's attention level.

[0089] In this embodiment, the virtual reality environment generation module can accurately simulate various real scenes, so that the elderly can perform cognitive training in a safe and controllable virtual space. By using VR helmets and advanced gesture recognition technology, the elderly can interact with objects and characters in the virtual environment. This interactive performance greatly enhances their sense of participation and immersion. The virtual reality system can monitor the user's operation in real time and provide immediate feedback, which not only helps the elderly to understand their performance in a timely manner, but also enables them to quickly adjust during the training process. According to the learning progress and ability of the elderly, the virtual reality environment generation module can provide personalized training plans and tasks to ensure that everyone can train at a pace that suits them.

[0090] Compared with traditional training methods, the virtual reality environment provides a risk-free training venue where the elderly can freely try various tasks without worrying about the dangers in real life. The virtual environment can provide emotional support for elderly people with cognitive impairment, helping them cope with anxiety and loneliness, while stimulating their memory and thinking abilities through daily activities and communication. By simulating various scenarios in real life, the elderly can practice various life skills in a virtual environment, such as dressing and eating, thereby improving their self-care ability. The virtual reality environment generation module provides great convenience and effectiveness for cognitive health monitoring and training of the elderly through its rich virtual scenes, real-time feedback mechanism, controllable and safe training environment, and enhanced motivation and participation.

[0091] Example 3

[0092] This embodiment is a further optimization based on the embodiment 1. Specifically, the cognitive neurofeedback module is a complex and advanced system that improves cognitive function by monitoring and regulating brain activity. The cognitive neurofeedback module includes an electroencephalogram (EEG) acquisition device, a signal processing algorithm unit, a deep learning technology unit, and a neuroscience research theoretical model.

[0093] EEG acquisition equipment is used to collect EEG signal data. The number and distribution of sensors in the equipment are also crucial to ensure comprehensive coverage of brain activity and effective filtering of artifacts. EEG systems and sensors are the prerequisites for achieving accurate feedback.

[0094] The signal processing algorithm unit is responsible for filtering out electrical signals of non-neural origin from the collected EEG signals and decoding brain patterns related to cognitive functions. It uses blind source separation technology (ICA) to filter out artifacts that interfere with alpha waves, thereby improving signal quality and increasing effective feedback time.

[0095] The application of deep learning technology units in the neurofeedback module is mainly reflected in the accuracy of real-time decoding of brain patterns, which is used to calibrate the subject's specific brain patterns and adapt to the brain activity patterns of specific subjects during online training. This technology supports personalized feedback solutions, making the training effect more accurate and efficient;

[0096] The theoretical model of neuroscience research is the triadic neural system model, which describes the role of emotion, reward, and cognitive control systems in decision-making behavior and reveals how various parts of the brain work together to achieve cognitive control;

[0097] The deep learning technique is a feed-forward neural network (FFNN). A feed-forward neural network has a one-way flow of information, from the input layer to the output layer, and does not contain feedback connections. It is used for function approximation and classification tasks. A feed-forward neural network with a single hidden layer, the hidden layer output is expressed as:

[0098] h=f(Wx++b)

[0099] Among them, the input vector is x, the weight matrix is ​​W, the bias vector is b, and the activation function is f;

[0100] The final output y is calculated through another linear transformation and activation function:

[0101] y=g(Uh+v)

[0102] Among them, U and v are the weight and bias of the output layer respectively, and g is the activation function of the output layer;

[0103] Since the information flow in the feedforward neural network is unidirectional, from the input layer to the output layer, this enables the network to respond quickly to the input data and is suitable for cognitive health monitoring tasks that require immediate feedback. Feedforward neural networks do not involve feedback connections within or between layers, which simplifies the network structure and reduces computational complexity. They are suitable for processing static data sets and classification problems. Feedforward neural networks extract high-level features of the data layer by layer through a multi-layer structure, and are particularly suitable for extracting meaningful information from complex health monitoring data.

[0104] In this embodiment, the cognitive neurofeedback module provides customized training tasks according to the specific needs and cognitive abilities of each elderly person. This personalized approach can more effectively promote the improvement of specific cognitive functions. Through special working memory training tasks, the cognitive neurofeedback module can help the elderly improve their working memory ability and address the common problem of distraction among the elderly. Using EEG equipment, the cognitive neurofeedback module can capture and analyze the elderly's EEG signals in real time, provide timely feedback, and enable trainees to understand their own brain activity status. Combined with the Internet of Things technology, medical professionals can remotely monitor the cognitive status of the elderly, and promptly discover and intervene in potential cognitive problems.

[0105] By improving basic cognitive functions such as working memory and attention, the elderly will be more adept at carrying out daily life activities, thereby improving their quality of life. The cognitive neurofeedback module not only focuses on cognitive functions, but also helps the elderly better manage emotions by regulating emotional networks, promoting social interactions, and reducing loneliness and anxiety.

[0106] Example 4

[0107] This embodiment is a further optimization based on the embodiment 1. Specifically, the ternary neural system model is a key theoretical framework for explaining and simulating complex behaviors of cognitive control, emotional response and reward processing, and is composed of three interconnected subsystems, including a cognitive control network (CCN), an emotional network (AN) and a reward network (RN);

[0108] The cognitive control network is responsible for high-level cognitive functions such as planning, decision-making, working memory, and attention control. The cognitive control network involves the prefrontal cortex, dorsolateral prefrontal cortex, and anterior cingulate cortex. The function of the cognitive control network is executive function, such as attention regulation, error detection, task switching, and goal maintenance. In mathematical models, the functions of the cognitive control network are described by linear and nonlinear equations. For example, the load of working memory is expressed as the weighted sum of input signals, which is processed by the activation function to form the output signal, which is expressed by the formula: y = f(w1x1+w2x2+...+wnxn-θ), where y is the output signal, xn is the input signal, wn is the weight, θ is the threshold, and f is the activation function;

[0109] The affective network is responsible for the generation and regulation of emotions, including emotional responses to environmental stimuli. It includes the amygdala, ventral striatum, and hippocampus regions, which process and regulate emotional responses. For example, the amygdala plays a key role in emotional responses to fear and anxiety, while the ventral striatum is involved in processing rewards and pleasure. The mathematical model of the affective network focuses on how to convert external stimuli into emotional responses, which is achieved through a threshold-based activation function that determines whether a specific emotional response is activated. For example, the activation function of the fear response is expressed as: F = g(s-φ), where F is the intensity of the fear response, s is the amount of external threat stimulus, φ is the minimum amount of stimulus required to activate the response, and g is a function that describes the relationship between the intensity of the response and the amount of stimulus.

[0110] The reward network is responsible for processing reward and incentive information from the external environment, including the mesolimbic system, nucleus accumbens and tegmental area. These areas are responsible for processing reward prediction, incentives and feelings of happiness. When an individual encounters something satisfying or pleasant, the reward network will be activated and release dopamine, a neurotransmitter associated with pleasure and satisfaction. The mathematical model of the reward network is based on reinforcement learning theory, in which the reward prediction error (Reward Prediction Error) formula is expressed as: RPEt = Rt-V(St), where Rt is the actual reward received at time t, and V(St) is the expected reward value in state S at time t.

[0111] In this embodiment, the ternary neural system model can comprehensively collect and analyze the neural activity data of the elderly by paying attention to the emotional network, reward network and cognitive control network at the same time. For example, in the risk decision-making task, the model can simultaneously capture the brain activity patterns related to emotional response, reward processing and cognitive control. Combined with the Internet of Things technology and artificial intelligence algorithms, the ternary neural system model can realize real-time monitoring and instant feedback of the neural activities of the elderly, and provide personalized cognitive training programs; according to the specific cognitive needs of the elderly, the ternary neural system model can generate customized training tasks and conduct targeted training for specific cognitive deficiencies. For example, for the elderly with reduced working memory ability, specific working memory training tasks can be designed;

[0112] The system can dynamically adjust the difficulty of the training tasks according to the training progress and performance of the elderly, ensuring that the training is always at an appropriate challenge level to avoid frustration or boredom. By designing the training tasks as games, the ternary nervous system model can stimulate the interest and motivation of the elderly, making them more willing to participate in the training. The ternary nervous system model can also simulate social interaction scenarios to provide emotional support for the elderly, helping them cope with anxiety and loneliness, while stimulating their memory and thinking abilities through social activities; the ternary nervous system model not only focuses on the improvement of cognitive function, but also helps the elderly alleviate emotional problems such as depression and anxiety through the regulation of emotional networks, further supporting their mental health;

[0113] The ternary nervous system model can be widely used in families and communities to reduce the elderly's dependence on professional medical institutions, thereby saving medical resources and social costs. By using Internet of Things technology and cloud computing platforms, the ternary nervous system model can realize remote health monitoring and management, making it convenient for medical staff to obtain the elderly's health data in real time and strengthen their health management.

[0114] Example 5

[0115] This embodiment is a further optimization based on the embodiment 1. Specifically, the real-time feedback mechanism includes sensor nodes, IoT gateways, cloud platforms and user terminals.

[0116] Sensor nodes include physiological parameter sensors, motion sensors and environmental sensors. Physiological parameter sensors (heart rate, blood pressure, blood sugar sensors) are used to collect key physiological indicators of the elderly, motion sensors (accelerometers and gyroscopes) are used to capture the movement information of the elderly, and environmental sensors (temperature, humidity, air quality sensors) are used to monitor the living environment of the elderly.

[0117] The IoT gateway includes data transmission and preprocessing, and threshold detection method. Data transmission and preprocessing are connected to sensor nodes through the network, receive and preprocess data, and then upload it to the cloud platform; the threshold detection method is used to detect the movement state of the human body and upload the relevant data to the server terminal;

[0118] The cloud platform stores, processes and analyzes the collected data, and pushes health status reports to users through application software;

[0119] The user terminal and the mobile phone application receive data through wireless communication, analyze, process and display the data, and give reasonable health advice based on the numerical range. The integrated video call function allows the elderly to have face-to-face consultations with doctors at home, and doctors to diagnose remotely.

[0120] The mathematical formula of the threshold detection method is as follows:

[0121] if a>Txanda>TTythen motion

[0122] Where ax and ay are the components of acceleration on the x and y axes respectively, and Tx and Ty are the preset thresholds;

[0123] The real-time feedback mechanism in the cognitive health monitoring system for the elderly realizes the real-time collection, transmission, analysis and feedback of the health data of the elderly through sensor nodes, IoT gateways, cloud platforms and user terminal components. By using threshold detection methods and deep learning model mathematical formulas, the system can provide timely warnings and give personalized health advice, greatly improving the efficiency and quality of health management for the elderly.

[0124] In this embodiment, through biosensor technology, the system can monitor the key physiological indicators of the elderly, such as heart rate and blood pressure, 24 hours a day, and immediately warn when abnormalities are detected. Whether it is sudden high blood pressure at night or blood sugar fluctuations caused by daily diet, it can be discovered and notified to relevant personnel in time to ensure the health and safety of the elderly;

[0125] Relying on big data and deep learning technology, the system can conduct personalized risk assessments based on the health history data of each elderly person and predict potential health problems. The elderly can have face-to-face consultations with doctors directly at home and enjoy professional medical services without leaving home. Doctors can view the elderly's health data through the platform and conduct remote diagnosis, greatly improving the efficiency and quality of medical services.

[0126] Example 6

[0127] This embodiment is a further optimization based on the first embodiment. Specifically, the personalized training plan generation module automatically generates a customized training plan according to the specific needs and cognitive abilities of each elderly person, including a comprehensive assessment tool, a data analysis and processing system, and a personalized algorithm technology, to achieve scientific, efficient and personalized rehabilitation services.

[0128] Comprehensive assessment tools include internationally certified scales, such as the Montreal Cognitive Assessment (MoCA) and the Mini-Mental State Examination (MMSE), which provide cognitive function assessments and ensure the scientific nature of the program: These scales have been professionally validated to ensure the scientific nature and pertinence of the rehabilitation program;

[0129] Data analysis and processing system, which automatically generates personalized training plans based on the evaluation results, covering multiple cognitive areas such as memory, attention, language, and executive function;

[0130] Personalized algorithm technology, using classification and regression tree (CART) analysis, conducts in-depth analysis of the cognitive data of the elderly to determine a personalized training plan. The algorithm can dynamically adjust the difficulty and content of the training tasks based on the elderly's training feedback and progress, ensuring that the training is always challenging and attractive;

[0131] Use virtual reality technology to create an attractive training environment, improve patient participation and training results, and through interesting interactive games and activities, make the assessment and training process no longer boring, and increase the participation and interest of elderly users.

[0132] In this embodiment, the system comprehensively evaluates the cognitive function of the elderly through integrated internationally certified cognitive function assessment scales, such as MOCA and MMSE, to ensure the scientificity and pertinence of the rehabilitation program. Based on the assessment results, the system automatically generates a personalized training plan for individual needs, covering multiple cognitive areas such as memory, attention, language, and executive function to meet the specific needs of different patients.

[0133] The system can monitor the patient's training progress and performance in real time, automatically generate detailed training reports, and help therapists accurately adjust rehabilitation strategies. The system allows the elderly to complete training tasks in a relaxed and pleasant atmosphere in an interesting and interactive way, effectively improving their cognitive abilities while enjoying the fun and sense of accomplishment brought by the game. In the process of completing the tasks, the elderly can not only exercise their memory, attention, thinking and cognitive abilities, but also enjoy the fun and sense of accomplishment brought by the game.

[0134] The system can automatically generate personalized training plans based on the results of the elderly’s cognitive assessment, ensuring that each elderly person can receive the training that suits them best. The system can also monitor the elderly’s training status in real time, and adjust the training plan in a timely manner based on the training results to ensure the scientificity and effectiveness of the training.

[0135] Example 7

[0136] See also Figure 2 and Figure 3 , this embodiment is further optimized on the basis of embodiment 1, specifically, the telemedicine platform includes sensor nodes, data transmission and preprocessing modules, cloud platform, and user terminals; the sensor nodes collect human physiological parameters and motion information through 4G terminal devices, and upload relevant data to the data transmission and preprocessing module; the data transmission and preprocessing module is connected to the sensor nodes through the network, receives and preprocesses the data, and then uploads it to the cloud platform; the cloud platform is responsible for data storage, processing and analysis to ensure data security and reliability; the application software and service units in the cloud platform push health status reports to users and medical staff through application software, and support remote diagnosis and treatment and health management services; the user terminal, the Android mobile phone application terminal receives data through wireless communication, parses, processes and displays the data, and gives reasonable health advice. Medical staff obtain the health data of the elderly in real time through terminal devices and provide professional telemedicine services;

[0137] The virtual assistant and coaching module includes a physiological parameter collection module, a data analysis module, a remote monitoring module, and a feedback suggestion unit; the physiological parameter collection module uses 4G terminal equipment to collect physiological parameters and motion information; the data analysis module uses a threshold detection method to determine the motion state of the human body and uploads the relevant data to the server terminal; the remote monitoring module forwards the uploaded data to the feedback suggestion unit through the server terminal so that the guardian can remotely receive the user's data; the feedback suggestion unit receives data through wireless communication, parses, processes, and displays the data, and gives reasonable health suggestions, and sends the data to the user terminal;

[0138] Binary logistic regression analysis was used to evaluate the statistical relationship between cognitive function and influencing factors in the elderly, helping to provide a theoretical reference for the early identification of cognitive dysfunction and the establishment of intervention measures.

[0139] In this embodiment, the telemedicine platform in the elderly cognitive health monitoring system realizes the real-time collection, transmission, analysis and feedback of the elderly health data through sensor nodes, IoT gateways, cloud platforms and user terminal components. The system can provide timely warnings and give personalized health advice, greatly improving the efficiency and quality of elderly health management.

[0140] The virtual assistant and coaching module can comprehensively assess the physical condition of the elderly, including physiological indicators and emotional state, and provide a full range of health analysis. The system can intelligently analyze the collected data, provide professional health status assessment and suggestions, and combine abnormal physiological parameters and sudden changes in behavioral data to achieve timely early warning and proactive intervention in emergencies to ensure the safety of the elderly. By providing comprehensive health management, the system helps to improve the quality of life of the elderly. The application of the virtual assistant and coaching module saves medical resources and reduces dependence on professional medical staff. The system provides an effective solution for social health management and reduces social costs.

[0141] Working principle:

[0142] Users wear VR helmets and biofeedback devices, and place sensors in their home environment;

[0143] The system guides users through an initial assessment phase and records baseline data through multimodal sensing devices;

[0144] The deep learning module analyzes baseline data and generates a personalized training plan;

[0145] Users conduct cognitive training in a virtual environment, and the system monitors multimodal data in real time and adjusts the training content;

[0146] After the training, the system provides feedback reports and adjusts subsequent training based on progress;

[0147] Medical staff monitor the user's status through the telemedicine platform and provide real-time guidance;

[0148] Repeat assessments regularly to continuously track cognitive health and synchronize information with healthcare databases.

[0149] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A virtual reality-based cognitive health monitoring system for the elderly, characterized by , the system comprises: Multimodal sensor acquisition module: including VR helmets, wearable devices, and home sensors to collect users' physiological and behavioral data; Virtual reality environment generation module: Create a variety of immersive scenes, including simulations of daily life scenes, memory games, and spatial navigation tasks; Cognitive neurofeedback module: monitors the user's brain activity in real time through an electroencephalogram (EEG) headband neural monitoring device, and feeds this information back to the training environment of the virtual reality environment generation module to adjust the difficulty and content of the training; Real-time feedback module: By providing real-time feedback of decoded brain activity patterns to users in the form of visual or auditory feedback, users can learn how to regulate their own brain activity to achieve cognitive enhancement; Deep learning analysis module: uses deep learning networks to analyze the user's performance data, physiological feedback, and environmental data, identifies subtle changes in cognitive ability, and predicts the risk of cognitive decline, and transmits the information to the server terminal; Personalized training plan generation module: Combines deep learning analysis results to develop highly personalized training plans and dynamically adjusts difficulty and content; Data storage and management module: stores user data and connects with the medical and health database; Telemedicine platform: Provides a remote monitoring and guidance platform for medical professionals to access patient data in real time and provide professional advice and intervention; Virtual Assistant and Coaching Module: Built-in intelligent virtual assistant provides real-time feedback, encouragement and support, and adjusts the interaction method based on the user's emotional and cognitive state.

2. The virtual reality-based cognitive health monitoring system for the elderly according to claim 1, characterized in that: The virtual reality environment generation module includes hardware equipment, software algorithms and user interface; The hardware equipment includes a virtual reality helmet and a controller. The virtual reality helmet is responsible for presenting the virtual environment in front of the user. The virtual reality helmet is equipped with a display screen, a head motion tracking sensor and a viewing angle adjustment function. The controller is used for the user to interact with the virtual environment. The controller includes a handheld device and a somatosensory sensor to detect the user's movements and convert them into operation instructions in the virtual environment. Software algorithms include 3D modeling and physics engines. 3D modeling uses computer graphics technology to create 3D objects and scenes in a virtual environment, involving geometric modeling, texture mapping, and lighting model algorithms. The user interface includes an interactive interface and a feedback system. The interactive interface allows users to interact with the virtual environment through a graphical interface or other natural interaction methods, including a menu system, gesture recognition, and voice command functions. The feedback system feeds back information in the virtual environment to the user, including visual, auditory, and tactile feedback.

3. The virtual reality-based cognitive health monitoring system for the elderly according to claim 1, characterized in that: The cognitive neurofeedback module includes an electroencephalogram (EEG) acquisition device, a signal processing algorithm unit, a deep learning technology unit, and a neuroscience research theoretical model: The EEG acquisition equipment is used to collect EEG signal data; the signal processing algorithm unit is responsible for filtering out electrical signals of non-neural origin from the collected EEG signals, and decoding the brain patterns related to cognitive functions, using blind source separation technology (ICA) to filter out artifacts that interfere with alpha waves; deep learning technology is used to calibrate the subject's specific brain patterns and adapt to the brain activity patterns of specific subjects during online training; the theoretical model of neuroscience research is the ternary neural system model, which describes the role of emotion, reward, and cognitive control systems in decision-making behavior and reveals how various parts of the brain work together to achieve cognitive control.

4. The virtual reality-based cognitive health monitoring system for the elderly according to claim 3, characterized in that: The deep learning technique is a feed-forward neural network (FFNN), in which information flows in one direction, from the input layer to the output layer, and does not contain feedback connections. It is used for function approximation and classification tasks. A feed-forward neural network with a single hidden layer, the hidden layer output is expressed as: h=f(Wx+b), Among them, the input vector is x, the weight matrix is ​​W, the bias vector is b, and the activation function is f; The final output y is calculated through another linear transformation and activation function: y=g(Uh+v), Among them, U and v are the weight and bias of the output layer respectively, and g is the activation function of the output layer.

5. The virtual reality-based cognitive health monitoring system for the elderly according to claim 3, characterized in that: The triadic neural system model consists of three interconnected subsystems, including the cognitive control network (CCN), the affective network (AN), and the reward network (RN); The cognitive control network is responsible for high-level cognitive functions. In the mathematical model, the functions of the cognitive control network are described by linear and nonlinear equations. The load of working memory is expressed as the weighted sum of the input signals, which is processed by the activation function to form the output signal, which is expressed by the formula: y = f (w1x1 + w2x2 + ... + w n x n -θ), where y is the output signal, x n is the input signal, w n is the weight, θ is the threshold, and f is the activation function; The affective network is responsible for the generation and regulation of emotions, which is achieved through a threshold-based activation function that determines whether a specific emotional response is activated. The activation function of the fear response is expressed as: F = g(s-φ), where F is the intensity of the fear response, s is the amount of external threat stimulus, φ is the minimum amount of stimulus required to activate the response, and g is a function that describes the relationship between the intensity of the response and the amount of stimulus; The reward network is responsible for processing rewards and incentive information from the external environment. The mathematical model of the reward network is based on reinforcement learning theory, where the reward prediction error (Reward Prediction Error) formula is expressed as: RPEt = Rt-V(St), where Rt is the actual reward received at time t, and V(St) is the expected reward value under state S at time t.

6. The virtual reality-based cognitive health monitoring system for the elderly according to claim 1, characterized in that: The real-time feedback mechanism includes sensor nodes, IoT gateways, cloud platforms and user terminals; Sensor nodes include physiological parameter sensors, motion sensors and environmental sensors. Physiological parameter sensors are used to collect key physiological indicators of the elderly, motion sensors are used to capture the movement information of the elderly, and environmental sensors are used to monitor the living environment of the elderly. The IoT gateway includes data transmission and preprocessing, and threshold detection methods. Data transmission and preprocessing are connected to sensor nodes through the network, receiving and preprocessing data, and then uploading it to the cloud platform; The threshold detection method is used to detect the motion state of the human body and upload the relevant data to the server terminal; The cloud platform stores, processes and analyzes the collected data, and pushes health status reports to users through application software; The user terminal and the mobile phone application receive data through wireless communication, analyze, process and display the data, and give reasonable health advice based on the numerical range. The integrated video call function allows the elderly to have face-to-face consultations with doctors at home, and doctors to diagnose remotely.

7. The virtual reality-based cognitive health monitoring system for the elderly according to claim 1, characterized in that: The personalized training plan generation module automatically generates a customized training plan based on the specific needs and cognitive abilities of each elderly person, including a comprehensive assessment tool, a data analysis and processing system, and personalized algorithm technology; Comprehensive assessment tools include internationally validated scales, such as the Montreal Cognitive Assessment (MoCA) and the Mini-Mental State Examination (MMSE), which provide measures of cognitive function; Data analysis and processing system, which automatically generates personalized training plans based on the evaluation results, covering multiple cognitive areas such as memory, attention, language, and executive function; Personalized algorithm technology uses classification and regression tree (CART) analysis to conduct in-depth analysis of the cognitive data of the elderly to determine a personalized training plan. The algorithm can dynamically adjust the difficulty and content of training tasks based on the elderly's training feedback and progress to ensure that the training is always challenging and attractive.

8. The virtual reality-based cognitive health monitoring system for the elderly according to claim 1, characterized in that: The telemedicine platform includes sensor nodes, data transmission and preprocessing modules, cloud platforms, and user terminals; The sensor nodes collect human physiological parameters and motion information through 4G terminal devices, and upload the relevant data to the data transmission and preprocessing module; The data transmission and preprocessing module is connected to the sensor node through the network, receives and preprocesses the data, and then uploads it to the cloud platform; the cloud platform is responsible for data storage, processing and analysis to ensure data security and reliability; the application software and service units in the cloud platform push health status reports to users and medical staff through application software, supporting remote diagnosis and health management services; the user terminal, the Android mobile phone application receives data through wireless communication, parses, processes and displays the data, and gives reasonable health advice. Medical staff obtain the health data of the elderly in real time through terminal devices and provide professional telemedicine services.

9. The virtual reality-based cognitive health monitoring system for the elderly according to claim 1, characterized in that: The virtual assistant and coaching module includes a physiological parameter acquisition module, a data analysis module, a remote monitoring module, and a feedback suggestion unit; The physiological parameter acquisition module uses 4G terminal equipment to collect physiological parameters and motion information; The data analysis module uses the threshold detection method to determine the movement state of the human body and uploads the relevant data to the server terminal; the remote monitoring module forwards the uploaded data to the feedback suggestion unit through the server terminal so that the guardian can receive the user's data remotely; the feedback suggestion unit receives the data through wireless communication, parses, processes and displays the data, gives reasonable health advice, and sends the data to the user terminal.

Images