Male reproductive data analysis system based on artificial intelligence and electroencephalogram signals and implementation method

Abstract

The invention discloses an andrology reproductive data analysis system based on artificial intelligence and electroencephalogram signals and an implementation method. The method is mainly characterized by comprising the steps that an electroencephalogram signal acquisition subsystem is used for acquiring reproduction related data of an andrology testee user, electroencephalogram signals are transmitted to an analysis subsystem, the analysis subsystem processes and arranges the data for training a deep learning model, and after training is finished, the analysis subsystem sends the data to a server; and the processing module is used for reasoning and judging newly transmitted electroencephalogram signal data related to male reproduction, returning an obtained result to a user side in a structured data and interactive interface manner, processing an output result by using a text language large model according to an actual scene, and controlling terminal equipment and directly reading by a user. The electroencephalogram signals analyzed by artificial intelligence are applied to the field of andrology, can be applied to and not limited to the functions of judging the stage of the reproductive activity of the user, diagnosing and analyzing reproductive related diseases and the like, and are beneficial to and not limited to the fields of diagnosis of andrology diseases, improvement of andrology diseases, improvement of life quality of the user and the like.

Description

Technical Field

[0001] The present invention relates to a data analysis system based on artificial intelligence, and in particular to a male reproductive data analysis system based on artificial intelligence and electroencephalogram signals and an implementation method thereof. Background Art

[0002] In clinical and practical male health care, a large number of patients seek medical attention for sexual dysfunction. Premature ejaculation accounts for a significant proportion of these patients. In actual diagnostics, the assessment of patient symptoms primarily relies on the patient's self-description of symptoms and their own estimation of the time from sexual intercourse to ejaculation. This data is difficult to accurately quantify, significantly impacting the doctor's understanding of the patient's condition.

[0003] Meanwhile, the medical community often uses delay therapy to treat premature ejaculation, requiring patients to engage in sexual activity, then refrain from ejaculating and cease intercourse when ejaculation is imminent, and then resume intercourse after sexual excitement subsides. However, in practice, patients often lack self-control, which compromises the effectiveness of treatment.

[0004] Currently, there are some technological improvements in actual treatment and diagnosis. These include using recording devices to analyze sexual activity and mechanical stimulation devices to provide delay therapy. However, these methods have limitations, such as the lack of accuracy in recordings and the lack of feedback from mechanical stimulation devices. This can result in the inability to stop stimulation when the user is about to orgasm, easily leading to the failure of delay therapy.

[0005] Non-invasive brain-computer interfaces (BCIs) based on scalp EEG have been widely studied due to their promising properties. Analysis of EEG waveforms reveals that the user's EEG waveforms vary in shape at different stages of sexual activity, with particularly pronounced fluctuations near orgasm. This phenomenon makes it possible to diagnose and treat men's clinical conditions by analyzing EEG signals. Furthermore, EEG signals are multidimensional time series signals. Analyzing the stage of sexual activity within a time series is a typical classification task in the field of artificial intelligence, and can be processed using deep learning neural networks. This, combined with large text language models, provides users with more specialized and readable content. Summary of the Invention

[0006] To solve the above problems, the present invention proposes a male reproductive data analysis system based on artificial intelligence and EEG signals and an implementation method. The present invention consists of an EEG signal acquisition subsystem, an information transmission subsystem, an analysis subsystem, and a feedback subsystem. Using this system, the EEG signals of the subject during sexual behavior are collected and transmitted to the analysis subsystem. The EEG signals are preprocessed and used as data samples. They are sent to the training module for training to obtain a model that can classify and identify EEG signals. After obtaining the model, it can be used to reason about newly incoming EEG signals to complete classification and identification tasks. The results are input into the feedback subsystem, and according to the scenario requirements of the device, the corresponding structured data is returned, or text information for users to read is generated based on key data combined with a large text language model.

[0007] The EEG signal acquisition subsystem consists of an EEG signal acquisition module, which is used to acquire the user's EEG signals. This module consists of a head cover that secures to the user's head, EEG measurement electrodes, and a handheld controller. The specific placement of the electrodes is as follows: EEG measurement electrodes are placed in the user's occipital region at positions Fp1, Fp2, C3, C4, O1, O2, T3, and T4. This electrode distribution conforms to the international 10-20 system. The reference electrode is placed at the left earlobe A1, and the ground electrode is placed at the right earlobe A2.

[0008] The EEG signal acquisition module supports data sampling points of 100, 200, and 400 per second, and the bandpass filter range is 0.5-70Hz. When collecting EEG signals for model training, key time points such as orgasm need to be marked during collection.

[0009] The handheld controller is a part of the EEG signal acquisition module, and the user can click on the handheld controller to mark a specific time point in the collected EEG signal.

[0010] The information transmission subsystem can transmit the received sensor data to the analysis subsystem via wired or wireless means, wherein the wireless means include but are not limited to WIFI, Bluetooth and the like.

[0011] The analysis subsystem consists of a data preprocessing module, a training module, and an inference module. The data preprocessing module is used to organize and process the multimodal data obtained by transmission. It mainly includes but is not limited to denoising, signal enhancement, dividing the signal into data sequences of equal length according to a fixed time length, and combining the data sequences into matrix signals. The training module uses the matrix signal obtained by data preprocessing as a training sample, uses a training script to train the selected benchmark model, and finally uses the trained model for inference. The inference module uses the trained model to identify and classify the multimodal data obtained by transmission through an inference script. According to the actual task, it can be used to determine the stage of reproductive activity of the subject user, diagnose the reproductive disease of the subject user, provide feedback information for clinical treatment, etc.

[0012] The training module includes a baseline model and a training script. The baseline model is an AI model capable of processing time series signals, including but not limited to RNN, RCNN, Transformer, and Swin-Transformer. All of these models are pre-trained. The training script sets various hyperparameters for the training task and the location of the dataset, and records training logs. Once training begins, multiple rounds of training are performed using the pre-processed data. In other words, the training module does not receive data or train the model in real time; it only begins after all training data has been received.

[0013] The inference module includes a trained custom model and inference scripts. The inference scripts can process newly input pre-processed signals using either a script or an interface to perform recognition and classification tasks. When receiving data, the inference module can perform inference and generate results in real time or in a non-real-time manner.

[0014] The feedback subsystem is used to output the inferred data as structured data, and can output different types of signals based on different scenario requirements. This includes, but is not limited to, outputting control signals for devices based on the inferred data in device control scenarios. When users operate the interactive interface, the feedback subsystem uses a large text language model to process the inferred data, generating more professional and easier-to-understand feedback text with better readability, thereby enhancing the user experience.

[0015] A method for analyzing male reproductive data based on artificial intelligence and electroencephalogram (EEG) signals, comprising the following steps: S1. Connect the EEG signal acquisition subsystem, information transmission subsystem, analysis subsystem, and feedback subsystem together, and place EEG measurement electrodes on the subject's head. The electrode distribution conforms to the 10-20 system, with a reference electrode placed on the left earlobe and a ground electrode placed on the right earlobe. S2: The subject undergoes reproductive-related treatment and diagnosis activities based on the men's health care scenario. The entire process ends with the subject ejaculating and resting for a period of time. Throughout this process, the EEG signal acquisition subsystem continuously collects the user's EEG information. The EEG information used for the training scenario also needs to be annotated at key time points such as orgasm. S3, the information transmission subsystem continuously transmits all recorded information to the analysis subsystem; S4. The analysis subsystem first performs unified data preprocessing. Then, in the training scenario, after the analysis subsystem receives all the recorded information, it starts the training task; S5. After training, the analysis subsystem obtains a model that can be used for reasoning, and can then be used in reasoning scenarios. This means identifying and classifying the transmitted information. S6. After reasoning the information, the obtained results are transmitted to the feedback subsystem. The feedback subsystem generates different feedback signals according to the actual application scenario.

[0016] Furthermore, in step S2, the EEG information used in the training scenario is labeled by the test user himself through the handheld controller of the EEG information acquisition subsystem.

[0017] Furthermore, in step S4, the content of data preprocessing includes but is not limited to denoising, bandpass filtering, signal enhancement, dividing the signal into data sequences of equal length according to a fixed time length, and combining the data sequences into matrix data.

[0018] Furthermore, in step S5, the model can perform real-time or non-real-time reasoning according to demand.

[0019] Furthermore, in step S6, the generated feedback signal includes but is not limited to: when providing feedback to the control device, outputting a control signal for the device; in the scenario where the user operates the interactive system, combining the inference results with the text language model to generate text feedback with better readability and professionalism.

[0020] Compared with the existing technology, the present invention has the following advantages and beneficial effects: collecting feedback from subjects during sexual behavior through a brain-computer interface is more timely and accurate; using an artificial intelligence deep learning network to process EEG data is more accurate and simple, and has a better anti-interference effect on noise; using EEG data in the field of andrology can promote the widespread use of new technologies in andrology clinics, thereby obtaining more accurate and instructive reference results in diagnosis and treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] FIG1 is a structural diagram of the present invention.

[0022] Figure 2 shows a method for collecting and labeling EEG information for training. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the technical solutions of the present invention are described in detail below with reference to the accompanying drawings and by way of examples. It should be noted that if there are any processes or symbols that are not clearly described below, they can be understood and implemented by those skilled in the art with reference to existing technologies.

[0024] The structure of the present invention, shown in Figure 1, comprises an EEG signal acquisition subsystem, an information transmission subsystem, an analysis subsystem, and a feedback subsystem. This system collects EEG signals from subjects during sexual activity and transmits them to the analysis subsystem. The EEG signals are preprocessed and used as data samples, which are then fed into a training module for training. This training module generates a model capable of classifying and recognizing EEG signals. This model can then be used to infer newly received EEG signals, completing classification and recognition tasks.

[0025] The EEG signal acquisition subsystem consists of an EEG signal acquisition module, which is used to acquire the user's EEG signals. This module consists of a head cover that secures to the user's head, EEG measurement electrodes, and a handheld controller. The specific placement of the electrodes is as follows: EEG measurement electrodes are placed in the user's occipital region at positions Fp1, Fp2, C3, C4, O1, O2, T3, and T4. This electrode distribution conforms to the international 10-20 system. The reference electrode is placed at the left earlobe A1, and the ground electrode is placed at the right earlobe A2.

[0026] The EEG signal acquisition module supports data sampling rates of 100, 200, and 400 samples per second, with a bandpass filter range of 0.5-70Hz. When collecting EEG signals for model training, key time points, such as orgasm, need to be annotated. Figure 2 illustrates the collection and annotation method for EEG information used for training. Marking a time point is accomplished by the user clicking the handheld controller once.

[0027] The information transmission subsystem transmits the received sensor data to the analysis subsystem via wired or wireless means, wherein the wireless means include but are not limited to WIFI, Bluetooth and the like.

[0028] The analysis subsystem consists of a data preprocessing module, a training module, and an inference module. The data preprocessing module is used to organize and process the multimodal data obtained by transmission. It mainly includes but is not limited to denoising, signal enhancement, dividing the signal into data sequences of equal length according to a fixed time length, and combining the data sequences into matrix signals. The training module uses the matrix signal obtained by data preprocessing as a training sample, uses a training script to train the selected benchmark model, and finally uses the trained model for inference. The inference module uses the trained model to identify and classify the multimodal data obtained by transmission through an inference script. According to the actual task, it can be used to determine the stage of reproductive activity of the subject user, diagnose the reproductive disease of the subject user, provide feedback information for clinical treatment, etc.

[0029] The training module includes a baseline model and a training script. The baseline model is an AI model capable of processing time series signals, including but not limited to RNN, RCNN, Transformer, and Swin-Transformer. All of these models are pre-trained. The training script sets various hyperparameters for the training task and the location of the dataset, and records training logs. Once training begins, multiple rounds of training are performed using the pre-processed data. In other words, the training module does not receive data or train the model in real time; it only begins after all training data has been received.

[0030] The inference module includes a trained custom model and inference scripts. The inference scripts can process newly input pre-processed signals using either a script or an interface to perform recognition and classification tasks. When receiving data, the inference module can perform inference and generate results in real time or in a non-real-time manner.

[0031] The feedback subsystem is used to output the inferred data as structured data, and can output different types of signals based on different scenario requirements. This includes, but is not limited to, outputting control signals for devices based on the inferred data in device control scenarios. When users operate the interactive interface, the feedback subsystem uses a large text language model to process the inferred data, generating more professional and easier-to-understand feedback text with better readability, thereby enhancing the user experience.

[0032] A method for analyzing male reproductive data based on artificial intelligence and electroencephalogram (EEG) signals includes the following steps: S1. Connect the EEG signal acquisition subsystem, information transmission subsystem, analysis subsystem, and feedback subsystem together, and place EEG measurement electrodes on the subject's head. The electrode distribution conforms to the 10-20 system, with a reference electrode placed on the left earlobe and a ground electrode placed on the right earlobe. S2 is implemented on the EEG signal acquisition subsystem. The subject undergoes reproductive-related treatment and diagnosis activities based on the male diagnosis and treatment scenario. The entire process ends with the subject ejaculating and resting for a period of time. Throughout the entire process, the EEG signal acquisition subsystem continuously collects the user's EEG information. The EEG information used for training scenarios also needs to be annotated at key time points such as orgasm; S3, implemented on the information transmission subsystem. The information transmission subsystem continuously transmits all recorded information to the analysis subsystem; S4 is implemented in the analysis subsystem. The analysis subsystem first performs unified data preprocessing. Data preprocessing includes but is not limited to denoising, bandpass filtering, signal enhancement, dividing the signal into data sequences of equal length according to a fixed time length, and combining the data sequences into matrix data. Then, in the training scenario, after the analysis subsystem receives all the recorded information, it carries out the training task; For example: 1. After data preprocessing, let T be the total length of the EEG sample sequence, N be the length of the EEG sample sequence, M be the number of EEG sample channels (i.e., the number of EEG electrodes excluding the reference electrode and ground electrode), and P be the number of EEG samples in a long time period. Then, P = T / N. The generated matrix data dimensions should be N*M*P. 2. Also, let C be the total number of categories to be classified. The total number of categories can be generated based on actual scenarios. For example, to identify brain waves associated with orgasm, the classification can be divided into two categories: "orgasm" and "non-orgasm," with C = 2. 3. Based on this setup, the parameters for the training model should be set as follows: the input data dimension is N*M, the total number of data samples is P, and the output should be the probability value of each category in the classification C. The category with the highest probability is taken as the output category result; S5 is implemented in the analysis subsystem. After training, the analysis subsystem obtains a model that can be used for reasoning. This model can then be used in reasoning scenarios. This means identifying and classifying the transmitted information. The model can perform real-time or non-real-time reasoning as needed. S6. After reasoning about the information, the results are transmitted to the feedback subsystem. The feedback subsystem generates different feedback signals based on the actual application scenario. These include, but are not limited to: outputting control signals for control devices; and, in scenarios where users operate interactive systems, combining the reasoning results with a large text language model to generate more readable and professional text feedback.

[0033] The above description is merely an embodiment of the present invention and does not limit the scope of the present invention. Any simple modifications, equivalent changes, and modifications to the above embodiments based on the technical essence of the present invention, or equivalent structural or process transformations based on the contents of the present invention specification, or direct or indirect applications in other related technical fields, that do not depart from the technical solution of the present invention, are also included in the scope of the present invention.

Claims

1. A male reproductive data analysis system based on artificial intelligence and EEG signals, characterized by: It consists of an EEG signal acquisition subsystem, an information transmission subsystem, an analysis subsystem, and a feedback subsystem. Among them, the EEG signal acquisition subsystem is used to collect reproductive-related data of male subjects, including but not limited to brain-computer interface data. The information transmission subsystem transmits the sensor data to the analysis subsystem by wired or wireless means. Among them, the wireless methods include but are not limited to WIFI, Bluetooth and other methods. The analysis subsystem receives the EEG signal data and pre-processes the data, including denoising, enhancement and other methods, and then uses the data for deep learning model training. After the training is completed, the model can infer based on the input EEG signal data, determine the category of the EEG signal data sequence within a period of time, and return the result to the feedback subsystem. After receiving the recognition result, the feedback subsystem returns structured data according to the scenario requirements of the device, or generates text information for users to read based on key data combined with a large text language model.

2. According to claim 1, a male reproductive data analysis system based on artificial intelligence and EEG signals, characterized in that: The EEG signal acquisition system is composed of a brain-computer interface. The brain-computer interface consists of an EEG signal acquisition module to acquire the user's EEG signals; all signals are connected to the information transmission subsystem for sending to the analysis subsystem.

3. According to claim 2, a male reproductive data analysis system based on artificial intelligence and EEG signals, characterized in that: The EEG signal acquisition module is composed of a head cover fixed on the user's head, EEG measurement electrodes and a handheld controller. The specific placement of the electrodes is as follows: the EEG measurement electrodes are placed at the user's occipital region Fp1, Fp2, C3, C4, O1, O2, T3, T4. The above electrode distribution complies with the international 10-20 system. The reference electrode is placed at the left earlobe A1, and the ground electrode is placed at the right earlobe A2.

4. The male reproductive data analysis system based on artificial intelligence and EEG signals according to claim 2 is characterized in that: The data sampling points of the EEG signal acquisition module support 100, 200, and 400 per second, and the bandpass filter range is 0.5-70Hz.

5. The male reproductive data analysis system based on artificial intelligence and EEG signals according to claim 2 is characterized by: The handheld controller is a part of the EEG signal acquisition module. The user can click on the handheld controller to mark a specific time point in the collected EEG signal.

6. The male reproductive data analysis system based on artificial intelligence and EEG signals according to claim 1 is characterized by: The information transmission subsystem can transmit the received sensor data to the analysis subsystem via wired or wireless means, wherein the wireless means include but are not limited to WIFI, Bluetooth and the like.

7. The male reproductive data analysis system based on artificial intelligence and EEG signals according to claim 1 is characterized by: The analysis subsystem is composed of a data preprocessing module, a training module, and an inference module. The data preprocessing module is used to organize and process the transmitted EEG signals. It mainly includes but is not limited to denoising, signal enhancement, dividing the signal into data sequences of equal length according to a fixed time length, and combining the data sequences into matrix data. The training module uses the matrix signal obtained by data preprocessing as a training sample, uses a training script to train the selected benchmark model, and finally uses the trained model for inference. The inference module uses the trained model to identify and classify the transmitted multimodal data through an inference script. According to the actual task, it can be but is not limited to determining the stage of reproductive activity of the subject user, diagnosing the reproductive disease of the subject user, and providing feedback information for clinical treatment.

8. The male reproductive data analysis system based on artificial intelligence and EEG signals according to claim 6, characterized in that: The training module includes a benchmark model and a training script. The benchmark model selects an artificial intelligence model that can process time series signals, including but not limited to: RNN, RCNN, Transformer, Swin-Transformer and other models. The above models have all been pre-trained. The training script sets various hyperparameters and data set locations for the training task, and records the training log. After the training starts, multiple rounds of training are performed using the incoming pre-processed data. In other words, the training module does not receive data and train the model in real time, but only starts after receiving all the training data.

9. The male reproductive data analysis system based on artificial intelligence and EEG signals according to claim 6, characterized in that: The inference module includes the trained custom model and inference script. The inference script can process the newly input preprocessed signal in script mode or interface mode to perform recognition and classification tasks. When receiving data, the inference module can perform inference and give results in real time, or process the results in non-real time.

10. The male reproductive data analysis system based on artificial intelligence and EEG signals according to claim 1, characterized in that: The feedback subsystem is used to output the inferred data according to the specified structured data, and can output different types of signals according to different scenario requirements. Including but not limited to the control of equipment scenarios, based on the inferred data, output control signals for the equipment; when the user operates the interactive interface, the feedback subsystem calls the text language model to process the inferred data and generate more professional and friendly feedback texts with better readability, thus improving the user experience.

11. A method for analyzing male reproductive data based on artificial intelligence and electroencephalogram signals as claimed in any one of claims 1 to 9, Its characteristics include the following steps: S1. Connect the EEG signal data acquisition subsystem, information transmission subsystem, analysis subsystem, and feedback subsystem together, and place EEG measurement electrodes on the head of the subject. The electrode distribution conforms to the 10-20 system. Place a reference electrode on the left earlobe and a ground electrode on the right earlobe. S2. The subject undergoes reproductive-related treatment and diagnosis activities according to the male medical treatment scenario. The whole process ends with the subject ejaculating and resting for a period of time. During the whole process, the EEG signal acquisition subsystem continuously collects the user's EEG information. The EEG information used for the training scenario also needs to be marked at key time points such as orgasm; S3, the information transmission subsystem continuously transmits all recorded information to the analysis subsystem; S4, the analysis subsystem first performs unified data preprocessing, and then in the training scenario, after the analysis subsystem receives all the recorded information, it carries out the training task; S5. After the training is completed, the analysis subsystem obtains a model that can be used for reasoning, and can now be used in reasoning scenarios. That is, the transmitted information is identified and classified; S6. After reasoning about the information, the obtained result is transmitted to the feedback subsystem. The feedback subsystem generates different feedback signals according to the actual application scenario.

12. The male reproductive data analysis system based on artificial intelligence and EEG signals according to claim 1, characterized in that: In step S2, the EEG information used in the training scenario is labeled by the subject through the handheld controller of the EEG information acquisition subsystem.

13. The male reproductive data analysis system based on artificial intelligence and EEG signals according to claim 1, characterized in that: In step S4, the content of data preprocessing includes but is not limited to denoising, bandpass filtering, signal enhancement, dividing the signal into data sequences of equal length according to a fixed time length, and combining the data sequences into matrix data.

14. The male reproductive data analysis system based on artificial intelligence and EEG signals according to claim 1, characterized in that: In step S5, the model can perform real-time or non-real-time reasoning according to demand.

15. The male reproductive data analysis system based on artificial intelligence and EEG signals according to claim 1, characterized in that: In step S6, the generated feedback signal includes but is not limited to: when providing feedback to the control device, outputting a control signal for the device; in the scenario where the user operates the interactive system, combining the inference result with the text language model to generate text feedback with better readability and professionalism.

Images