Severe game intervention fused electroencephalogram collection cap for children with autism and signal processing method

Abstract

The invention discloses an autistic child electroencephalogram collection cap fusing severe game intervention and a signal processing method.The autistic child electroencephalogram collection cap comprises a structure module, an electroencephalogram collection module, a wireless communication module, a power module, a signal processing module and a game and report module, and the electroencephalogram collection module collects electroencephalogram signals through five electrodes and two ear patch type electrodes; the wireless communication module is used for communication between the electroencephalogram cap and user terminal equipment, the power supply module is used for storing electric energy and supplying power to other modules, and the signal processing module is used for processing and analyzing electroencephalogram signals in the severe game intervention process, including filtering, slicing and normalization processing, and outputting standardized electroencephalogram data. And then the data is analyzed through an MCNN network model, and the condition of the children is diagnosed, so that an effective tool is provided for brain function research and diagnosis while the adaptability of the children with autism is improved.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of medical devices and neurophysiological signal acquisition, and particularly relates to an autism child electroencephalogram acquisition cap fusing serious game intervention and a signal processing method. BACKGROUND

[0002] Autism is a common mental disorder that affects the quality of life and mental health of hundreds of millions of people around the world, and autism children as a special group of the disease, early identification and intervention are crucial for the treatment and rehabilitation of autism children. Electroencephalogram (EEG) as a non-invasive brain function detection technology can monitor the electrical activity of the brain in real time, providing important physiological information for the diagnosis and research of autism children, however, traditional intervention methods often have many limitations, such as low participation of children in training, difficulty in quantifying training effect and continuous monitoring, etc., which seriously restricts the accurate grasp of the brain function state of autism children and the effective adjustment of subsequent treatment programs. In addition, most of the existing electroencephalogram acquisition devices are large in size and complex in operation, which not only brings discomfort to children, but also is difficult to integrate into the daily rehabilitation training scene.

[0003] For example, the utility model discloses a brain electricity acquisition cap convenient to position, which is monitored by being attached to the head of a human body to obtain brain electricity signals, but it is difficult to wear for autism children due to too many brain electricity channels for measurement and complicated wearing steps.

[0004] For example, the utility model discloses a brain electricity acquisition cap convenient to clean, which measures brain electricity signals through the parts of the brain electrode head contacted with the scalp, but the brain electricity signals are not preprocessed, so that the collected brain electricity signals need to be processed again, increasing the workload of subsequent brain electricity processing, and the corresponding judgment cannot be made in time according to the collected brain electricity signals.

[0005] For example, the utility model discloses a multi-channel brain electricity acquisition cap, which uses a silica gel brain electricity cap to acquire human brain electricity signals, however, the comfort during wearing is affected due to the need to wrap the whole head during wearing, and if it is used for autism children, it may cause strong resistance of the children.

[0006] For example, the U.S. invention patent with the publication number U S2025099009A1 discloses a brain electricity diagram headset with transmission function, which acquires brain electricity through an EEG headset, but it does not realize the acquisition of task state brain electricity of the measured person through external stimulation (such as serious game).

[0007] As a new intervention method, serious game can effectively attract the attention of children with autism and improve their participation and enthusiasm due to its interesting and interactive nature. By integrating serious game into the design of the EEG acquisition cap, the brain activity of children can be monitored in real time during their participation in the game, providing rich data for the assessment of brain function status and adjusting the difficulty and content of the game according to the EEG feedback to achieve personalized rehabilitation training. Therefore, the integration of serious game and EEG monitoring can promote each other and bring new breakthroughs to the rehabilitation training of children with autism. Therefore, the invention of a portable, easy-to-use and comfortable EEG acquisition cap that can integrate serious game intervention will greatly promote the early screening and continuous monitoring of children with autism, helping doctors better understand the brain function status of patients and thus develop more effective treatment plans. SUMMARY

[0008] The present application provides a kind of autism children's EEG acquisition cap and signal processing method of fusion serious game intervention, to realize in the serious game intervention while brain electric signal acquisition, and to brain electric signal effectively process.The application includes structural module, EEG acquisition module, wireless communication module, power module, signal processing module and game and report module.

[0009] As a further scheme of the present application: the structural module is composed of an elastic cap and a central control box. The central control box has three embedded modules inside, which are the wireless communication module, the power module and the EEG signal processing module. The central control box is provided with a charging interface on the side.

[0010] As a further scheme of the present application: the EEG acquisition module is composed of five electrodes, two ear patch electrodes, a signal amplifier, a filter and a digital-to-analog conversion circuit. The electrodes are assembled on the forehead part of the elastic cap and are attached to the head of the person, with the point positions being located at F8, Fp2, Fz, Fp1 and F7 according to the 10 / 20 system standard. The ear patch electrodes are located at A1 and A2 as reference electrodes according to the 10 / 20 system standard and are attached to the earlobe part. The signal amplifier is designed with low noise and high gain to preliminarily amplify the weak EEG signals collected by the electrodes, and the amplification factor can be adjusted to adapt to the strength changes of EEG signals in different individuals and different scenarios, ensuring stable output of the signals. The filter is used to remove noise interference in the EEG signals within a set passband frequency range (such as 0.5Hz-100Hz), so that the main frequency components of the EEG signals are retained, while irrelevant noise beyond the range is effectively filtered out, improving the signal-to-noise ratio and ensuring the accuracy of subsequent signal processing. The digital-to-analog conversion circuit is responsible for converting the analog EEG signals processed by the amplifier and filter into digital signals, ensuring that the digital signals can truly reflect the characteristics and changes of the original EEG signals, so that subsequent digital signal processing and analysis such as feature extraction, classification and identification can be smoothly carried out.

[0011] As a further scheme of the present application: the wireless communication module includes a radio frequency front end, a modem, a baseband processor, an antenna interface and a microcontroller (MCU), which is used to establish a connection with the user's terminal device.

[0012] As a further scheme of the present application: the power module is composed of a battery, a charging management chip, a charging interface and a battery management chip, wherein the charging interface is located on the side of the central control box and is used to connect with an external charging power supply, and the battery management chip sends the battery power information at certain time intervals, and if the device is not used for a period of time in the on state, the device will enter a standby state to reduce power consumption and prolong the use time.

[0013] As a further scheme of the present application: the game and report module integrates emotion regulation, cognitive training and report, emotion regulation is controlled by an emotion stabilization unit, cognitive training is controlled by a serious game unit, which is deployed on the terminal device, and is connected with the EEG acquisition cap through the same wireless frequency band under the wireless communication technology, so as to closely cooperate and realize the intervention training of autistic children, and generate an inspection report after the emotion regulation and cognitive training are completed.

[0014] As a further scheme of the present application: the EEG acquisition cap needs to be used in cooperation with the serious game of the terminal device, the terminal software integrates an emotion stabilization unit and a serious game unit, and the two units work cooperatively to realize the precise intervention training of autistic children. In the game starting stage, the emotion stabilization unit automatically plays a piece of soothing music to soothe the child's emotion, while synchronously recording the baseline EEG data; the serious game unit is automatically started after the music playing is completed, and the game content thereof is closely designed around the cognitive training task, covering training projects such as attention focusing, logic matching and social interaction scene.

[0015] As a further scheme of the present application: after the training of the serious game unit is completed, the terminal software temporarily stores the collected EEG data and the related training data in the game process in a local storage area. The local storage area is equipped with professional data analysis tools and algorithms to deeply analyze the data, which covers detailed interpretation of EEG characteristics, quantitative evaluation of training effect and dynamic analysis of emotional changes, etc. According to the analysis results, the system automatically generates a detailed training report, and the report content focuses on key information such as the emotional stability of the child in the training, the completion of the cognitive task, the trend of EEG activity change, etc. Such report can be referred by doctors, therapists and other professionals, and provides a scientific basis for the subsequent treatment scheme, so as to help realize the precise intervention and long-term tracking evaluation of autistic children.

[0016] After the user presses the power button of the EEG acquisition cap, the EEG acquisition cap is powered on immediately. The power module in the central control box starts to provide stable power supply for each embedded module. The wireless communication module quickly completes automatic initialization and actively attempts to establish a connection with the terminal device. At the same time, the EEG signal processing module enters a standby state, ready to receive signals transmitted from the EEG acquisition module at any time. On the terminal device side, the matching software continues to run stably in the background, actively and in real time monitoring the connection state of the EEG acquisition cap. Once the wireless communication module successfully completes the connection with the terminal device, the software interface will immediately pop up a prominent prompt window, clearly showing that the EEG acquisition cap has been connected, and synchronously presenting the detailed state information of the current device, including power level and signal strength and other key indicators.

[0017] When the EEG cap accurately detects the EEG signals from the measured person, the serious game software pre-installed on the game terminal device will automatically start. At this time, the emotion stabilization unit is activated first, and a carefully selected soothing music is automatically played, with a duration of 1 to 3 minutes and a frequency range of 60 to 120 Hz. The volume is strictly controlled within a comfortable range of not more than 60 dB. The music type is rich and varied, and the user can pre-select from natural sound effects or light music, etc. whose spectral energy is concentrated in the low frequency band (below 500 Hz). At the same time, the EEG acquisition module immediately goes into work, the electrodes are in close contact with the skin of the head, the signal amplifier accurately amplifies the weak EEG signals, the filter efficiently filters out various external interference signals, and the digital-analog conversion circuit immediately converts the analog signals into digital signals for processing. The converted EEG signals are quickly transmitted to the terminal device by the wireless communication module, and the terminal software not only displays the received signals in real time, but also stores them as baseline EEG data, laying a solid foundation for subsequent emotion stability evaluation. In the software interface of the terminal device, the EEG waveform graph is displayed in real time and accurately, as well as the corresponding and detailed frequency spectrum graph. The user can intuitively and comprehensively view the current EEG activity state.

[0018] After the music playing period ends, the serious game unit will automatically start. The game content is carefully designed around cognitive training tasks, covering key training projects such as attention focusing, logical matching, and social interaction scenarios. Specifically, in the attention focusing training session, children need to focus their attention on the specific target that appears on the screen from time to time and respond quickly when the target appears; in the logical matching training part, children need to accurately match the items or patterns associated with each other according to the screen prompts; and in the social interaction scenario training, children need to interact with virtual characters to complete a series of set tasks. In the process of the user engaging in serious games, the EEG acquisition module continuously collects EEG signals through five precise electrodes and stably stores them in the local storage area.

[0019] When the game is over, the professional data analysis tools and algorithms equipped in the local storage area begin to conduct in-depth analysis on the collected data. The local storage area pre-stores trained models and uses federated learning to perform small-scale calculation and model optimization at the edge to better match the device and the wearer. The analysis covers multiple dimensions such as detailed interpretation of EEG features, quantitative evaluation of game task completion, and dynamic monitoring of emotional changes. Based on the results of the analysis, the system will automatically and intelligently generate a content training report and upload it to the doctor's terminal. The report summarizes the core information such as the emotional stability of the child during the training process, the progress of cognitive task completion, and the trend of EEG activity changes. The report uses a combination of charts and text to present the training results in a lively and clear manner. In the software interface of the terminal device, users can conveniently access the generated training report by clicking the "View Report" button.

[0020] The method for processing EEG signals is as follows:

[0021] Step 1: Collect EEG signals through 5 high-precision EEG sensors, convert analog signals to digital signals, and obtain the original EEG data matrix E.

[0022] Step 2: Perform band-pass filtering on the EEG data in Step 1 to filter out unwanted signals below 0.5 Hz and above 100 Hz, and retain the effective EEG signal frequency range to obtain the filtered matrix E'.

[0023] Step 3: According to the set sampling frequency, perform slicing on the filtered EEG signals in Step 2, cutting each EEG signal into identical wavelength segments with a length of T to obtain the sliced EEG segment set S.

[0024] Step 4: Perform normalization on the EEG segment set S in Step 3 to unify the numerical range to [-1, 1], and obtain the processed EEG segment set S'.

[0025] Step 5: Output the EEG segment set S' in Step 4 to provide standardized and high-quality EEG data support for subsequent EEG feature analysis, disease diagnosis, and other applications.

[0026] Step 6: The data constructed in step 5 is input into the newly defined MCNN network model for processing. The MCNN model is improved on the basis of the traditional CNN network (LeNet-5 network). Compared with the LeNet-5 model, the MCNN model is improved in two aspects: first, the number of network layers is increased, and the MCNN contains two network layers, which can extract information in one-dimensional and two-dimensional scales of the input feature matrix; second, the idea of residual learning is introduced. Residual learning can not only help to extract deeper features, but also effectively solve the network degradation problem. In addition, the optimizer used by the model is the SGD optimizer, the learning rate is 0.001, and the decay rate is 0.1.

[0027] The calculation formula of the convolution layer can be represented as:

[0028]

[0029] wherein, is the pixel value of the input feature map, W k,l is the weight of the convolution kernel, b is the bias term, and σ is the activation function.

[0030] The calculation formula of the max pooling layer can be represented as:

[0031] P(i,j)=max(X(i×s:i×s+k-1,j×s:j×s+k-1)),

[0032] wherein, X is the input feature map, k×k is the pooling window size, and s is the step length.

[0033] The calculation formula of the residual block can be represented as:

[0034] F=BN_2(Conv_2(ReLU(BN_1(Conv_1(X)))))+X,

[0035] wherein, the input feature is X, the output feature is F, Conv_1 and Conv_2 are two 5×5 convolution operations, and usually they have the same number of convolution kernels to ensure that the number of channels of the output feature map is consistent with the input. The first convolution operation is used to extract features, and the second convolution operation is used to map the features back to the same channel space as the input. BN_1 and BN_2 are batch normalization (Batch Normalization) operations after the first and second convolutions, respectively, which are used to speed up training and stabilize the network. ReLU is an activation function used to introduce nonlinearity. + represents the addition operation of the input feature map X and the feature map after two convolutions and corresponding processing, which is the core part of the residual connection, allowing the network to learn residual mapping, making it easier to train deeper networks.

[0036] The calculation formula of the fully connected layer can be represented as:

[0037]

[0038] where x i is the input neuron, w k,i is the weight, b k is the bias term, and σ is the activation function.

[0039] Step 7: The feature information extracted in step 6 is analyzed comprehensively based on the performance of the child in the serious game (such as task completion, reaction time, etc.), and the real-time collected EEG features (such as specific frequency band power changes, abnormal waveform frequency, etc.), to make a comprehensive diagnosis of the child. The final output is a specific quantitative result, including the probability of autism (such as 80% probability of disease), and the severity of the disease based on multi-dimensional evaluation, which is classified as mild, moderate or severe autism tendency. This quantitative diagnosis result will be directly used for disease diagnosis and disease assessment, and will provide accurate and efficient standardized and high-quality EEG data support for subsequent EEG feature analysis, clinical disease diagnosis, etc. After this result, it will also be uploaded to the cloud server and can be viewed by the doctor terminal at any time. When diagnosing and classifying the child, the system uses the constructed classifier to realize, for example, using softmax regression for disease screening, and the classification process can be represented as:

[0040]

[0041] where h i is the output of the fully connected layer, W k,i is the weight, b k is the bias term, and the softmax function is used to convert the output into a probability distribution. BRIEF DESCRIPTION OF DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0043] Figure 1 is the point map of the 10 / 20 system standard.

[0044] Figure 2 is a schematic diagram of an EEG acquisition method provided by the embodiment of the present application.

[0045] Figure 3 is a schematic diagram of the relationship between each module provided by the embodiment of the present application.

[0046] Figure 4 A front view schematic diagram of an EEG acquisition cap provided by an embodiment of the present application.

[0047] Figure 5 A rear view schematic diagram of an EEG acquisition cap provided by an embodiment of the present application.

[0048] The meanings of the respective reference numerals in the drawings are as follows:

[0049] F8 electrode 1; Fp2 electrode 2; Fz electrode 3; Fp1 electrode 4; F7 electrode 5; A1 left ear sticker 6; A2 right ear sticker 7; cap 8; elastic buckle 9; charging interface 10. DETAILED DESCRIPTION

[0050] The technical solutions of the present application will be described clearly and completely below in conjunction with embodiments. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort fall within the scope of protection of the present application.

[0051] An EEG acquisition cap for autistic children fused with serious game intervention and a signal processing method are provided by an embodiment of the present application. As shown in Figure 1 In the present embodiment, five silver / silver chloride (Ag / AgCl) electrodes are selected for the EEG acquisition module and are assembled at the forehead part of the elastic cap, which can be closely and comfortably attached to the head of a person. The electrode 1, electrode 2, electrode 3, electrode 4 and electrode 5 are respectively accurately positioned at the F8, Fp2, Fz, Fp1 and F7 positions according to the international 10 / 20 system standard. Among them, the F8 electrode is located at the right frontal pole region of the head, the Fp2 electrode is located at the right frontal pole close to the midline, the Fz electrode is at the midline of the frontal pole, the Fp1 electrode is located at the left frontal pole close to the midline, and the F7 electrode is located at the left frontal pole region. The left ear sticker 6 and the right ear sticker 7 are also positioned at the A1 and A2 positions, i.e. the left and right earlobes, according to the 10 / 20 system standard, and are attached to the earlobe part as reference electrodes for providing a stable reference potential to ensure the accuracy and reliability of the collected EEG signals. The selection of these electrode points is based on their importance and representativeness in the collection of neuroelectrophysiological signals, which can effectively cover key brain areas such as the frontal lobe and capture EEG activity characteristics closely related to psychological processes such as cognition and emotion, providing a solid data foundation for subsequent EEG analysis and diagnosis of autistic children.

[0052] As shown in Figures 2 to 5 The present EEG acquisition cap includes the following modules: a structure module, an EEG acquisition module, a wireless communication module, a power module, a signal processing module and a game and report module.

[0053] In the embodiment of the application, the structural module comprises an elastic cap and a central control box, the central control box has three embedded modules inside, which are a wireless communication module, a power module and an electroencephalogram signal processing module, and the central control box is provided with a charging interface on the side.

[0054] The signal processing circuit part comprises a signal amplifier, a band-pass filter (0.5-100 Hz) and a digital-to-analog conversion circuit, which are directly embedded inside the inner shell, thereby shortening the signal transmission path and reducing noise interference.

[0055] In the embodiment of the application, a flexible shell and an elastic buckle are adopted to adapt to different head circumferences and reduce the feeling of oppression when worn; ear paste electrodes avoid the discomfort of traditional ear clips; the embedded module realizes real-time filtering and standardization without relying on external equipment, thereby improving the diagnosis efficiency; the wireless communication and power management technology optimize the endurance, support flexible use in home and clinical scenarios, and improve the portability.

[0056] In the five-channel electroencephalogram acquisition cap provided in the embodiment of the application, the structure design and process manufacturing of the five-channel electroencephalogram signal acquisition technology are adopted to improve the signal precision; the flexible dry electrode structure (flexible silicone tentacle electrode and bump electrode) is designed to improve the comfort and signal adhesion, thereby improving the portability and stability of the wearable electrode system; the wireless acquisition alternative solution solves the problems of assembly and replacement, increases the encryption memory, reading circuit and authentication mechanism, and ensures the encryption memory.

[0057] In the embodiment of the application, the wireless communication module uses low-power Bluetooth technology, can stably communicate with terminal equipment within a range of 10 meters, such as a computer, a tablet or a mobile phone; and the user can receive electroencephalogram data in real time and control the acquisition parameters by means of the matching software.

[0058] In the embodiment of the application, the battery of the power module is a rechargeable lithium polymer battery, the capacity of which is 1200 mAh, the size is 30 mm x 20 mm x 5 mm, and the battery can support continuous work for more than 5 hours. The standby current is lower than 10 μA, the acquisition mode power consumption is not more than 25 mW, the charging interface 10 supports 5V / 1A fast charging, and the charging management chip is a TI BQ25601 with overvoltage and overcurrent protection.

[0059] In the embodiment of the application, the embedded microprocessor of the signal processing module is assembled with a high-performance low-power chip, can perform real-time signal filtering, slicing and normalization processing, and the memory is used to store the preprocessed standardized electroencephalogram data (the numerical range is -1 to 1) for subsequent analysis and diagnosis.

[0060] In the embodiment of the application, the terminal game linkage is also supported, the terminal software develops interactive games, can receive in real time with a delay controlled within 20 ms, and dynamically displays the electroencephalogram signal spectrum (0.5-30 Hz).

[0061] After the user starts the EEG collection cap through the terminal software, the software first calls the emotion stabilizing module to play preset soothing music. The music lasts for 1-3 minutes, the volume is controlled below 60 dB, and the spectral energy is concentrated in the low frequency band (such as natural sound effect or light music), so as to reduce the anxiety level of the child.

[0062] During the music playing, the EEG collection module continuously records the baseline EEG signal, and the signal processing module calculates the baseline EEG power spectral density (PSD) with the help of the embedded microprocessor, and stores the calculation result as reference data.

[0063] After the music ends, the terminal software automatically switches to the serious game module. The game content is designed based on the cognitive training needs of children with autism, such as guiding children to complete attention focusing tasks through virtual scenes.

[0064] During the game process, the EEG collection module collects EEG signals in real time, the signal processing module performs band-pass filtering (0.5-100 Hz), slicing (length T=2 seconds) and normalization processing on the data, and generates a standardized EEG segment set S'.

[0065] The terminal software receives the standardized data through Bluetooth and extracts EEG features in real time, dynamically adjusts the game difficulty according to the feature values (such as increasing the number of targets or shortening the reaction time). At the same time, the software combines the baseline PSD data to evaluate the emotional stability of the child, and if anxiety features (such as abnormal high-frequency gamma wave rise) are detected, the intervention mechanism (such as pausing the game and playing a soothing voice) is triggered.

[0066] In the embodiment of the application, the embedded microprocessor of the signal processing module is equipped with a high-performance low-power chip, which can perform real-time signal filtering, slicing and normalization processing, and the memory is used to store the preprocessed standardized EEG data (numerical range -1 to 1) for subsequent analysis and diagnosis.

[0067] After the serious game unit training is completed, the terminal software temporarily stores the collected EEG data and related training data during the game process in the local storage area, and is equipped with professional data analysis tools and algorithms to perform in-depth analysis on the data, and automatically generates detailed training reports according to the analysis results. The report content includes the emotional stability of the child in the training, the completion of the cognitive task, the trend of EEG activity change and other key information. At the same time, these reports will be uploaded to the cloud server in real time, and medical staff can view the test results of the measured person through the doctor terminal.

[0068] The above only describes the preferred embodiments of the application, and does not limit the protection scope of the application, and any equivalent structural transformation made under the inventive concept of the application, or direct / indirect application in other related technical fields is included in the protection scope of the application.

Claims

1. A cap for collecting EEG data from autistic children that integrates serious play intervention, characterized in that... : It includes a structural module, an EEG acquisition module, a wireless communication module, a power supply module, a signal processing module, and a game and reporting module; Structural module: It consists of an elastic cap and a central control box. The central control box contains a wireless communication module, a power module, and an EEG signal processing module, and has a charging interface on the side. EEG acquisition module: includes five electrodes, two ear patch electrodes, a signal amplifier, a filter, and a digital-to-analog converter circuit. The five electrodes are mounted on the forehead of the elastic cap and are located at points F8, Fp2, Fz, Fp1, and F7 according to the 10 / 20 system standard. The two ear patch electrodes are located at A1 and A2 according to the 10 / 20 system standard and are attached to the earlobe as reference electrodes. Wireless communication module: including RF front-end, modem, baseband processor, antenna interface and microcontroller, used to establish connection with user terminal equipment; Power module: It consists of a battery, a charging management chip, a charging interface and a battery management chip. The charging interface is located on the side of the central control box and is used to connect to an external charging power source. The battery management chip can send battery power information at intervals, and the device will enter standby mode after a period of inactivity while it is on to reduce power consumption. The game and report module integrates emotion regulation and cognitive training, and is deployed on terminal devices. It includes an emotion stabilization unit and a serious game unit, and works with an EEG collection cap to realize intervention training for children with autism.

2. The EEG collection cap for autistic children integrating serious play intervention as described in claim 1, characterized in that: The EEG collection cap needs to be used in conjunction with the serious game on the terminal device. The terminal software integrates an emotion stabilization unit and a serious game unit. The two work together. During the game start phase, the emotion stabilization unit automatically plays soothing music to calm the child's emotions and simultaneously records baseline EEG data. The serious game unit automatically starts after the music ends. Its game content covers training items such as attention focusing, logic matching, and social interaction scenarios.

3. The EEG collection cap for autistic children integrating serious play intervention as described in claim 1, characterized in that: After the training session for the serious game unit is completed, the terminal software will temporarily store the collected EEG data and relevant training data from the game process in the local storage area. The local storage area contains pre-trained models and is equipped with professional data analysis tools and algorithms to perform in-depth data analysis. At the same time, federated learning is used to perform small-scale computations and model optimization at the edge to make the device more compatible with the wearer. Based on the analysis results, a detailed training report is automatically generated, which includes key information such as the child's emotional stability during training, cognitive task completion, and EEG activity trends.

4. A method for processing EEG signals in autistic children that integrates serious play intervention, characterized in that, Includes the following steps: Step 1: Collect EEG signals through five electrodes, convert the analog signals into digital signals, and obtain the raw EEG data matrix E; Step 2: Perform bandpass filtering on the EEG data from Step 1 to remove useless signals below 0.5Hz and above 100Hz, and obtain the filtered matrix E'; Step 3: Based on the set sampling frequency, slice the filtered EEG signal from Step 2 to obtain a set of sliced ​​EEG segments S; Step 4: Normalize the set of EEG segments S from Step 3 so that its values ​​are uniformly within the range of [-1, 1], to obtain the processed set of EEG segments S'; Step 5: Output the set of EEG segments S' from Step 4 to provide standardized, high-quality EEG data support for subsequent applications such as EEG feature analysis and disease diagnosis; Step 6: Input the data constructed in Step 5 into the MCNN network model for processing. The MCNN model is an improvement on the traditional CNN network (LeNet-5 network), which increases the number of network layers and introduces the idea of ​​residual learning. It also uses the SGD optimizer with a learning rate of 0.001 and a decay rate of 0.

1. Step 7: Integrate and analyze the feature information extracted in Step 6 for applications such as disease diagnosis and condition assessment.

5. The method for processing EEG signals in autistic children incorporating serious play intervention as described in claim 4, characterized in that: The MCNN network model in step 6 includes convolutional layers, max pooling layers, residual blocks, and fully connected layers.

6. The method for processing EEG signals in autistic children incorporating serious play intervention as described in claim 4, characterized in that: In step 7, a classifier is constructed, such as using softmax regression, to screen for diseases.

Images