Brain-computer interface data privacy protection method, device and computer equipment

Abstract

The application relates to a brain-computer interface data privacy protection method, device and computer equipment. The method comprises the following steps: acquiring brain-computer interface data of multiple users, and performing feature decoupling processing to obtain frequency information and phase information of each brain-computer interface data; encoding each frequency encoding information and each phase encoding information respectively, and training initial brain-computer interface models by decoding model parameter information obtained through each frequency encoding information; for each user, adjusting the parameter information of the initial brain-computer interface model based on the phase encoding information and the frequency encoding information of the user to obtain a brain-computer interface model of the user, adding privacy protection information in the brain-computer interface model of each user to obtain a privacy protection brain-computer interface model of each user; and determining a target global brain-computer interface model based on the privacy protection brain-computer interface models of the users. The method can improve the privacy protection effect of brain-computer interface data.

Description

Technical Field

[0001] This application relates to the fields of brain-computer interface and information privacy protection technology, and in particular to a method, apparatus and computer device for protecting the privacy of brain-computer interface data. Background Technology

[0002] Brain-computer interface (BCI) is a hardware-software collaborative communication system that enables information exchange and control of external devices without relying on the peripheral nervous system. A complete BCI system includes the user's brain, brain signal acquisition, signal encoding, a control interface, and external devices. However, considering both the potential information carried by the brain-computer data itself and the functionalities achievable by BCI technology, data security and user privacy protection are becoming increasingly important. Therefore, developing BCIs that balance neural information security and privacy protection by optimizing the acquisition, processing, and transmission of neural signals during BCI applications is of great significance.

[0003] Traditional methods for protecting the privacy of brain-computer interface (BCI) data involve constructing a privacy-preserving encoder and a user-shared encoder in the feature representation part, and correspondingly constructing a user-independent classifier and a user-shared classifier in the classifier part. This approach considers user privacy during the encoding of EEG information. However, this method is limited to the encoding of neural signals and does not consider the entire structure of the BCI system. This results in poor privacy protection of BCI data. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, computer device, computer-readable storage medium, and computer program product for protecting the privacy of brain-computer interface data in response to the above-mentioned technical problems.

[0005] Firstly, this application provides a method for protecting the privacy of brain-computer interface data. The method includes:

[0006] Brain-computer interface data from multiple users and individual brain-computer interface models are acquired, and the brain-computer interface data of each user are decoupled to obtain the frequency information and phase information of each user's brain-computer interface data.

[0007] Each frequency information and each phase information is encoded to obtain each frequency encoded information and each phase encoded information. The model parameter information obtained by decoding each frequency encoded information is used to train the individual brain-computer interface model to obtain the initial brain-computer interface model.

[0008] For each user, based on the user's phase encoding information and frequency encoding information, the parameter information of the initial brain-computer interface model is adjusted to obtain the user's brain-computer interface model, and privacy protection information is added to each user's brain-computer interface model to obtain each user's privacy-protected brain-computer interface model;

[0009] Construct global privacy protection information for a global brain-computer interface model, and determine the target global brain-computer interface model based on the privacy protection brain-computer interface model of each user and the global privacy protection information.

[0010] Optionally, the step of decoupling each of the brain-computer interface data to obtain the frequency information and phase information of each of the brain-computer interface data includes:

[0011] The brain-computer interface data of each of the brain-computer interfaces is subjected to low-frequency filtering to obtain low-frequency brain-computer interface data corresponding to each of the brain-computer interface data;

[0012] The low-frequency brain-computer interface data are decoupled to obtain the frequency information and phase information of each brain-computer interface data.

[0013] Optionally, the step of encoding each of the frequency information and each of the phase information to obtain each frequency encoded information and each phase encoded information includes:

[0014] For each user, the user's frequency information and phase information are encoded into the stimulation end to obtain comprehensive encoded information;

[0015] The comprehensive coding information of each user is split to obtain the frequency coding information and the phase coding information of each user.

[0016] Optionally, the step of training the individual brain-computer interface model using each of the frequency encoding information to obtain an initial brain-computer interface model includes:

[0017] The frequency encoding information is decoded to obtain the first model parameter information, and the individual brain-computer interface model is trained based on the first model parameter information to obtain the initial brain-computer interface model.

[0018] Optionally, adjusting the parameter information of the initial brain-computer interface model based on the user's phase encoding information and frequency encoding information to obtain the user's brain-computer interface model includes:

[0019] The phase encoding information of the user is decoded to obtain the second model parameters;

[0020] Based on the second model parameters and the first model parameters, the parameter information of the initial brain-computer interface model is adjusted to obtain the user's brain-computer interface model.

[0021] Optionally, the global privacy protection information for constructing the global brain-computer interface model includes:

[0022] Identify the protection level of each privacy-preserving brain-computer interface model, and determine the target protection level of the global brain-computer interface model based on the protection levels of all privacy-preserving brain-computer interface models.

[0023] Based on the target protection level, global privacy protection information for the global brain-computer interface model is generated through a privacy protection information generation strategy.

[0024] Secondly, this application also provides a privacy protection device for brain-computer interface data. The device includes:

[0025] The acquisition module is used to acquire brain-computer interface data from multiple users and individual brain-computer interface models, and to decouple each brain-computer interface data to obtain frequency information and phase information of each brain-computer interface data.

[0026] The training module is used to encode each of the frequency information and each of the phase information respectively to obtain each frequency encoded information and each phase encoded information, and to train the individual brain-computer interface model respectively by decoding each of the frequency encoded information to obtain the model parameter information, thereby obtaining the initial brain-computer interface model.

[0027] The adjustment module is used to adjust the parameter information of the initial brain-computer interface model for each user based on the user's phase encoding information and frequency encoding information to obtain the user's brain-computer interface model, and to add privacy protection information to each user's brain-computer interface model to obtain each user's privacy-protected brain-computer interface model.

[0028] The construction module is used to construct global privacy protection information for the global brain-computer interface model, and to determine the target global brain-computer interface model based on the privacy protection brain-computer interface model of each user and the global privacy protection information.

[0029] Optionally, the acquisition module is specifically used for:

[0030] The brain-computer interface data of each of the brain-computer interfaces is subjected to low-frequency filtering to obtain low-frequency brain-computer interface data corresponding to each of the brain-computer interface data;

[0031] The low-frequency brain-computer interface data are decoupled to obtain the frequency information and phase information of each brain-computer interface data.

[0032] Optionally, the training module is specifically used for:

[0033] For each user, the user's frequency information and phase information are encoded into the stimulation end to obtain comprehensive encoded information;

[0034] The comprehensive coding information of each user is split to obtain the frequency coding information and the phase coding information of each user.

[0035] Optionally, the training module is specifically used for:

[0036] The frequency encoding information is decoded to obtain the first model parameter information, and the individual brain-computer interface model is trained based on the first model parameter information to obtain the initial brain-computer interface model.

[0037] Optionally, the adjustment module is specifically used for:

[0038] The phase encoding information of the user is decoded to obtain the second model parameters;

[0039] Based on the second model parameters and the first model parameters, the parameter information of the initial brain-computer interface model is adjusted to obtain the user's brain-computer interface model.

[0040] Optionally, the building module is specifically used for:

[0041] Identify the protection level of each privacy-preserving brain-computer interface model, and determine the target protection level of the global brain-computer interface model based on the protection levels of all privacy-preserving brain-computer interface models.

[0042] Based on the target protection level, global privacy protection information for the global brain-computer interface model is generated through a privacy protection information generation strategy.

[0043] Thirdly, this application provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method described in any one of the first aspects.

[0044] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the steps of the method described in any one of the first aspects.

[0045] Fifthly, this application provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the steps of the method described in any one of the first aspects.

[0046] The aforementioned brain-computer interface (BCI) data privacy protection method, apparatus, and computer equipment acquire BCI data from multiple users and individual BCI models, and decouple each BCI data to obtain frequency information and phase information of each BCI data. They then encode each frequency and phase information to obtain frequency-encoded information and phase-encoded information, and decode the frequency-encoded information to obtain model parameter information, thereby training the individual BCI models to obtain initial BCI models. For each user, based on the user's phase-encoded information and frequency-encoded information, they adjust the parameter information of the initial BCI model to obtain the user's BCI model. Privacy protection information is added to each user's BCI model to obtain a privacy-protected BCI model for each user. Global privacy protection information for the global BCI model is constructed, and a target global BCI model is determined based on the privacy-protected BCI models of each user and the global privacy protection information. By decoupling and encoding user brain-computer interface (BCI) data, frequency encoding information and phase encoding information are obtained. Then, an individual BCI model is trained using only the frequency encoding information to obtain an initial BCI model. Next, the initial BCI model is adjusted individually using each user's phase and frequency encoding information to obtain a BCI model for each user. Finally, by adding privacy protection information to each user's BCI model and the global BCI model, a target global BCI model is obtained. This scheme splits user BCI data during the acquisition and training of the BCI model, thus avoiding the sharing of all user BCI data and protecting user privacy. Furthermore, by adding privacy protection information to each user's BCI model and the global BCI model, the possibility of data leakage between BCI models is avoided while ensuring the normal operation of the BCI models, thereby improving the privacy protection effect of BCI data. Attached Figure Description

[0047] Figure 1 This is a flowchart illustrating a method for protecting the privacy of brain-computer interface data in one embodiment;

[0048] Figure 2This is a schematic diagram of an interface database for generating brain-computer interface stimuli in one embodiment;

[0049] Figure 3 This is a schematic diagram based on the UCSD_SSVEP-BCI dataset in one embodiment;

[0050] Figure 4 This is a schematic diagram based on the Benchmark dataset in one embodiment;

[0051] Figure 5 This is a schematic diagram illustrating the accuracy (Acc) analysis in one embodiment;

[0052] Figure 6 This is a schematic diagram illustrating the Information Transmission Rate (ITR) analysis in one embodiment;

[0053] Figure 7 This is a schematic diagram of joint modeling across datasets (acquisition devices) in one embodiment;

[0054] Figure 8 This is a scenario diagram illustrating privacy-preserving brain-computer interface in one embodiment;

[0055] Figure 9 This is a schematic diagram of a distributed brain-computer interface based on data sharing / no data sharing in one embodiment;

[0056] Figure 10 This is an application diagram of a method for protecting the privacy of brain-computer interface data in one embodiment;

[0057] Figure 11 This is a structural block diagram of a brain-computer interface data privacy protection device in one embodiment;

[0058] Figure 12 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0060] The privacy protection method for brain-computer interface (BCI) data provided in this application embodiment can be applied to the application environment corresponding to the BCI system. This method can be applied to a terminal, a server, or a system including both a terminal and a server, and is implemented through interaction between the terminal and the server. The terminal can be, but is not limited to, various personal computers, laptops, smartphones, tablets, etc. The terminal decouples and encodes the user's BCI data to obtain frequency encoding information and phase encoding information. Then, it trains an individual BCI model using only the frequency encoding information to obtain an initial BCI model. Next, it adjusts the initial BCI model using each user's phase encoding information and frequency encoding information to obtain a BCI model corresponding to each user. Finally, by adding privacy protection information to the BCI models corresponding to each user and the global BCI model, a target global BCI model is obtained. This solution splits user brain-computer interface (BCI) data when acquiring training BCI models, thus avoiding the sharing of all user BCI data and protecting user privacy. Then, by adding privacy protection information to each user's BCI model and the global BCI model, the solution avoids the possibility of data leakage between BCI models while ensuring the normal operation of the BCI models, thereby improving the privacy protection effect of BCI data.

[0061] Among them, the privacy protection method for brain-computer interface (BCI) data addresses the problems of low efficiency in complex data management and insufficient confidentiality, integrity, and availability during data distribution and transmission in the context of global BCI models. It studies efficient and privacy-secure protection methods based on distributed BCI models, supporting efficient interactive transmission of massive multimodal heterogeneous time-varying data. It also studies the transmission method of user BCI data on the local end and the server end based on privacy-secure protection, realizing the reuse of user data between users under the condition that user data does not leave the domain, realizing joint modeling based on distributed systems, thereby improving the security protection effect of user BCI data.

[0062] In one embodiment, such as Figure 1 As shown, a method for protecting the privacy of brain-computer interface data is provided. Taking a terminal of this method as an example, the method includes the following steps:

[0063] Step S101: Obtain brain-computer interface data from multiple users and individual brain-computer interface models, and decouple each brain-computer interface data to obtain frequency information and phase information of each brain-computer interface data.

[0064] In this embodiment, the terminal collects brain-computer interface (BCI) data from each user via a BCI device. This BCI data consists of signal data information corresponding to neural signals in the user's brain. The terminal then decouples each BCI data set to obtain its frequency and phase information. The decoupled BCI data is the result of low-frequency filtering. The specific decoupling process will be described in detail later.

[0065] Step S102: Encode each frequency information and each phase information to obtain each frequency encoded information and each phase encoded information. Then, use the model parameter information obtained by decoding each frequency encoded information to train individual brain-computer interface models to obtain initial brain-computer interface models.

[0066] In this embodiment, the terminal encodes each frequency information and each phase information to obtain frequency-encoded information and phase-encoded information, respectively. Then, the terminal trains individual brain-computer interface (BCI) models using each frequency-encoded information to obtain initial BCI models. Before training the individual BCI models using the frequency-encoded information, the model parameters are first obtained by decoding the frequency-encoded information. The individual BCI models are trained using distributed machine learning, and these models are multi-user collaborative server-side BCI models. The specific training process will be explained in detail later.

[0067] Step S103: For each user, based on the user's phase encoding information and frequency encoding information, adjust the parameter information of the initial brain-computer interface model to obtain the user's brain-computer interface model, and add privacy protection information to each user's brain-computer interface model to obtain each user's privacy-protected brain-computer interface model.

[0068] In this embodiment, the terminal adjusts the parameters of the initial brain-computer interface (BCI) model for each user based on the user's phase encoding information and frequency encoding information to obtain the user's BCI model. Privacy protection information is then added to each user's BCI model to obtain a privacy-protected BCI model for each user. This privacy protection information may be, but is not limited to, noise perturbation information. This noise perturbation information is used to prevent the BCI model from leaking the user's BCI data, thereby improving the privacy protection effect of the BCI model.

[0069] Step S104: Construct global privacy protection information for the global brain-computer interface model, and determine the target global brain-computer interface model based on the privacy protection brain-computer interface model of each user and the global privacy protection information.

[0070] In this embodiment, the terminal constructs global privacy protection information for the global brain-computer interface model, and determines the target global brain-computer interface model based on the privacy-protected brain-computer interface models of each user and the global privacy protection information. The global privacy protection information and the privacy protection information are the same, but differ in magnitude.

[0071] Based on the above scheme, the user's brain-computer interface (BCI) data is decoupled and encoded to obtain frequency encoding information and phase encoding information. Then, an individual BCI model is trained using only the frequency encoding information to obtain an initial BCI model. Next, the initial BCI model is adjusted separately using each user's phase and frequency encoding information to obtain a BCI model corresponding to each user. Finally, by adding privacy protection information to each user's BCI model and the global BCI model, a target global BCI model is obtained. This scheme splits the user's BCI data during the acquisition and training of the BCI model, thus avoiding the sharing of all user BCI data and protecting user privacy. Furthermore, by adding privacy protection information to each user's BCI model and the global BCI model, the possibility of data leakage between BCI models is avoided while ensuring the normal operation of the BCI models, thereby improving the privacy protection effect of BCI data.

[0072] Optionally, each brain-computer interface (BCI) data is decoupled to obtain the frequency information and phase information of each BCI data. This includes: performing low-frequency filtering on each BCI data to obtain the corresponding low-frequency BCI data; and performing decoupling on each low-frequency BCI data to obtain the frequency information and phase information of each BCI data.

[0073] In this embodiment, the terminal performs low-frequency filtering on each brain-computer interface (BCI) data to obtain low-frequency BCI data corresponding to each BCI data. Then, the terminal decouples each low-frequency BCI data to obtain the frequency information and phase information of each BCI data. The separability of the low-frequency BCI data within the user's BCI data dimension is higher than that of the high-frequency BCI data after high-frequency filtering.

[0074] Based on the above scheme, the decoupling process after low-frequency filtering improves the splitting effect of brain-computer interface data.

[0075] Optionally, each frequency information and each phase information are encoded separately to obtain each frequency encoded information and each phase encoded information, including: for each user, encoding the user's frequency information and the user's phase information to the stimulation end to obtain comprehensive encoded information; and splitting the comprehensive encoded information of each user to obtain the frequency encoded information and the phase encoded information of each user.

[0076] In this embodiment, the terminal encodes the frequency information and phase information of each user into the stimulation end to obtain comprehensive encoded information. Then, the terminal decomposes the comprehensive encoded information for each user to obtain the frequency encoded information and the phase encoded information for each user. The frequency encoded information is public information, meaning all users use the same frequency encoding method; the phase encoded information is protected privacy information, meaning each user encodes their phase information differently. The stimulation end is a brain-computer interface database tailored to different users, such as... Figure 2 The image shows a database of interfaces generated for brain-computer interface stimulation.

[0077] Specifically, the formula corresponding to the phase information encoding method is as follows:

[0078] N freq =40,N phase =4,N target =40

[0079] N horizontal =8,N vertical =5

[0080]

[0081] Where, N freq N represents the frequency count in the target encoding of the brain-computer interface. phase N represents the number of phases in the encoding. target N represents the target number. horizontal N represents the number of lateral targets at the brain-computer interface. vertical N represents the number of vertical targets in the brain-computer interface. stim This represents the number of brain-computer interface interfaces that may be generated based on the above encoded information.

[0082] Based on the above scheme, by encoding frequency information and phase information in different ways, it is ensured that the complete stimulus encoding information of each user remains only locally on the user's device, and the distinguishability of the encoding information of different users is also increased.

[0083] Optionally, an individual brain-computer interface model is trained using the frequency encoding information to obtain an initial brain-computer interface model, including: decoding the frequency encoding information to obtain first model parameter information, and training the individual brain-computer interface model based on the first model parameter information to obtain the initial brain-computer interface model.

[0084] In this embodiment, the terminal decodes the frequency encoding information to obtain the first model parameter information, and trains the individual brain-computer interface model based on the first model parameter information to obtain the initial brain-computer interface model.

[0085] The individual brain-computer interface model is modeled using Domain Generalization (DG) modeling. The domain generalization modeling formula is shown below:

[0086]

[0087]

[0088]

[0089]

[0090] h∶X→Y

[0091]

[0092] Among them, S train S represents the domain used for model training or human-computer adaptation. test Represents the domain used for model testing or application; P XY Represents EEG data x i and y i The joint probability distribution is formed, where E represents the loss function.

[0093] Based on the above scheme, by decoding the encoded information to obtain model parameters for training, the system ensures that user data does not leave the domain and the model moves while the data remains stationary, thereby protecting user privacy during the data circulation phase.

[0094] Optionally, based on the user's phase encoding information and frequency encoding information, the parameter information of the initial brain-computer interface model is adjusted to obtain the user's brain-computer interface model, including: decoding the user's phase encoding information to obtain second model parameters; and adjusting the parameter information of the initial brain-computer interface model based on the second model parameters and the first model parameters to obtain the user's brain-computer interface model.

[0095] In this embodiment, the terminal decodes the user's phase-encoded information to obtain the second model parameters. Then, based on the second model parameters and the first model parameters, the terminal adjusts the parameter information of the initial brain-computer interface model to obtain the user's brain-computer interface model. The method of adjusting the parameter information of the initial brain-computer interface model is to fine-tune the norm of the initial brain-computer interface model by combining the second model parameters and the first model parameter data.

[0096] Based on the above approach, by fine-tuning the trained model, the performance of the model is not affected by a small amount of data, while ensuring the diversity of the brain-computer interface model for each user.

[0097] Optionally, construct global privacy protection information for the global brain-computer interface model, including: identifying the protection level of each privacy-protected brain-computer interface model, and determining the target protection level of the global brain-computer interface model based on the protection levels of all privacy-protected brain-computer interface models; and generating global privacy protection information for the global brain-computer interface model based on the target protection level and through a privacy protection information generation strategy.

[0098] In this embodiment, the terminal identifies the protection level of each privacy-preserving brain-computer interface (BCI) model and determines the target protection level of the global BCI model based on the protection levels of all privacy-preserving BCI models. Finally, based on the target protection level, the terminal generates global privacy protection information for the global BCI model through a privacy protection information generation strategy.

[0099] Based on the above scheme, by adding noise of different magnitudes to the parameters of the global model and the local model respectively, the privacy protection effect between the brain-computer interface models corresponding to each user, as well as between the brain-computer interface model and the global brain-computer interface model, is improved.

[0100] like Figure 3 , 4 As shown, the distributed joint brain-computer interface system (i.e., the global brain-computer interface model) framework is convergent on brain-computer data based on visual evoked potentials. Furthermore, the relationship between the algorithmic performance and communication efficiency of the distributed joint brain-computer interface system framework (i.e., the global brain-computer interface model) is quantitatively presented. Figure 5 , 6 As shown, the model's recognition performance was evaluated from the time window dimension. The results indicate that the distributed joint brain-computer interface system framework can reuse user data without requiring user data to leave the domain, thus achieving joint modeling based on a distributed system. Figure 7Compared to the modeling process of a brain-computer interface (BCI) model without this approach, the BCI model obtained using this method shows a significantly faster convergence speed and achieves higher recognition accuracy. Joint modeling experiments on data from different acquisition devices validate that the distributed joint BCI system framework (i.e., the global BCI model) can achieve cross-device joint modeling. After personalized encoder calibration, in benchmark dataset experiments, within a 1-second data time window, the proposed invention achieved an average recognition accuracy of 80.63%, significantly higher than the baseline method's 64.85%. Within a 1-second data time window, the proposed invention achieved an average recognition accuracy of 95.39%, outperforming the deep learning-based Conv-CA method and the second-order matrix factorization-based eTRCA method.

[0101] like Figure 8 The diagram illustrates a scenario for privacy-protecting brain-computer interfaces (BCIs). The left portion represents user-dependent BCIs, the middle portion represents user-independent BCIs, and the right portion represents the BCIs that this solution aims to protect. It's understandable that the left and middle portions of the diagram represent the original user's BCI. Since each user's BCI data directly participates in the training of the global BCI, the privacy protection for each individual user is relatively weak.

[0102] like Figure 9 As shown, this scheme decouples the brain-computer interface data of each user into frequency information and phase information, and then uses the model parameter information obtained by decoding the frequency information of each user for training the global brain-computer interface. This ensures both the trainability of the joint model and a certain degree of privacy protection for user data when training the brain-computer interface model. Figure 9 The left part of the diagram shows a distributed brain-computer interface without data sharing, which is the training method of the brain-computer interface model mentioned in this solution. Figure 9 The right-hand diagram illustrates a distributed brain-computer interface (BCI) with data sharing. This diagram represents the approach described in this scheme, which involves decoupling each user's BCI data into frequency and phase information, and then using the model parameters obtained from decoding the frequency information of each user for training the global BCI. Before this approach, the distributed BCI with data sharing can be pre-trained with publicly available shared data. It describes the joint modeling from wet electrode data for scientific research to dry electrode data for product applications. This joint modeling, based on the Domain Generalization (DG) modeling method mentioned in this scheme, achieves zero-calibration of the system without using current user data.

[0103] like Figure 10The diagram illustrates the brain-computer interface (BCI) generation method, data transmission protocol, and storage medium mentioned in this solution, which balances neural information security and user privacy protection. The figure depicts the model parameter exchange process during the training of the global BCI model. The server receives model parameter information (frequency information from the user's BCI data) transmitted from each local terminal and trains the global BCI model based on this parameter information. Each local terminal decouples the user's BCI data to obtain frequency and phase information, and then decodes the frequency information to obtain model parameters, which are transmitted to the server. The server then distributes the trained global BCI model to each local terminal, which then trains its corresponding BCI model using the full phase and frequency information.

[0104] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0105] Based on the same inventive concept, this application also provides a brain-computer interface data privacy protection device for implementing the aforementioned brain-computer interface data privacy protection method. The solution provided by this device is similar to the implementation described in the above method; therefore, the specific limitations in one or more brain-computer interface data privacy protection device embodiments provided below can be found in the limitations of the brain-computer interface data privacy protection method described above, and will not be repeated here.

[0106] In one embodiment, such as Figure 11 As shown, a privacy protection device for brain-computer interface data is provided, comprising: an acquisition module 1110, a training module 1120, an adjustment module 1130, and a construction module 1140, wherein:

[0107] The acquisition module 1110 is used to acquire brain-computer interface data from multiple users and individual brain-computer interface models, and to decouple each brain-computer interface data to obtain frequency information and phase information of each brain-computer interface data.

[0108] The training module 1120 is used to encode each of the frequency information and each of the phase information respectively to obtain each frequency encoded information and each phase encoded information, and to train the individual brain-computer interface model respectively by decoding the model parameter information obtained by each of the frequency encoded information to obtain an initial brain-computer interface model.

[0109] The adjustment module 1130 is used to adjust the parameter information of the initial brain-computer interface model for each user based on the user's phase encoding information and frequency encoding information, to obtain the user's brain-computer interface model, and to add privacy protection information to each user's brain-computer interface model to obtain each user's privacy-protected brain-computer interface model.

[0110] The construction module 1140 is used to construct global privacy protection information for the global brain-computer interface model, and to determine the target global brain-computer interface model based on the privacy protection brain-computer interface model of each user and the global privacy protection information.

[0111] Optionally, the acquisition module 1110 is specifically used for:

[0112] The brain-computer interface data of each of the brain-computer interfaces is subjected to low-frequency filtering to obtain low-frequency brain-computer interface data corresponding to each of the brain-computer interface data;

[0113] The low-frequency brain-computer interface data are decoupled to obtain the frequency information and phase information of each brain-computer interface data.

[0114] Optionally, the training module 1120 is specifically used for:

[0115] For each user, the user's frequency information and phase information are encoded into the stimulation end to obtain comprehensive encoded information;

[0116] The comprehensive coding information of each user is split to obtain the frequency coding information and the phase coding information of each user.

[0117] Optionally, the training module 1120 is specifically used for:

[0118] The frequency encoding information is decoded to obtain the first model parameter information, and the individual brain-computer interface model is trained based on the first model parameter information to obtain the initial brain-computer interface model.

[0119] Optionally, the adjustment module 1130 is specifically used for:

[0120] The phase encoding information of the user is decoded to obtain the second model parameters;

[0121] Based on the second model parameters and the first model parameters, the parameter information of the initial brain-computer interface model is adjusted to obtain the user's brain-computer interface model.

[0122] Optionally, the building module 1140 is specifically used for:

[0123] Identify the protection level of each privacy-preserving brain-computer interface model, and determine the target protection level of the global brain-computer interface model based on the protection levels of all privacy-preserving brain-computer interface models.

[0124] Based on the target protection level, global privacy protection information for the global brain-computer interface model is generated through a privacy protection information generation strategy.

[0125] The modules in the aforementioned brain-computer interface data privacy protection device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0126] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 12 As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a method for protecting the privacy of brain-computer interface data. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0127] Those skilled in the art will understand that Figure 12The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0128] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described in any one of the first aspects.

[0129] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in any one of the first aspects.

[0130] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the method described in any one of the first aspects.

[0131] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0132] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0133] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0134] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for protecting the privacy of brain-computer interface data, characterized in that, The method includes: Acquire brain-computer interface data from multiple users, as well as individual brain-computer interface models; The brain-computer interface data of each of the brain-computer interfaces is subjected to low-frequency filtering to obtain low-frequency brain-computer interface data corresponding to each of the brain-computer interface data; The low-frequency brain-computer interface data are decoupled to obtain the frequency information and phase information of each brain-computer interface data. Each frequency information and each phase information is encoded to obtain each frequency encoded information and each phase encoded information; wherein, the frequency encoded information is public information shared by each user, and the phase encoded information is private information not shared by each user; The frequency encoding information is decoded to obtain the first model parameter information, and the individual brain-computer interface model is trained based on the first model parameter information to obtain the initial brain-computer interface model. For each user, the phase encoding information of the user is decoded to obtain the second model parameters; Based on the second model parameters and the first model parameters, the parameter information of the initial brain-computer interface model is adjusted to obtain the user's brain-computer interface model; Add privacy protection information to each user's brain-computer interface model to obtain each user's privacy-protected brain-computer interface model; Identify the protection level of each privacy-preserving brain-computer interface model, and determine the target protection level of the global brain-computer interface model based on the protection levels of all privacy-preserving brain-computer interface models. Based on the target protection level, global privacy protection information of the global brain-computer interface model is generated through a privacy protection information generation strategy. Based on the privacy-preserving brain-computer interface models of each user and the global privacy-preserving information, the target global brain-computer interface model is determined.

2. The method according to claim 1, characterized in that, The process of encoding each of the frequency information and each of the phase information to obtain each frequency encoded information and each phase encoded information includes: For each user, the user's frequency information and phase information are encoded into the stimulation end to obtain comprehensive encoded information; The comprehensive coding information of each user is split to obtain the frequency coding information and the phase coding information of each user.

3. A privacy protection device for brain-computer interface data, characterized in that, The device includes: The acquisition module is used to acquire brain-computer interface data from multiple users and individual brain-computer interface models; perform low-frequency filtering on each brain-computer interface data to obtain low-frequency brain-computer interface data corresponding to each brain-computer interface data; and perform decoupling processing on each low-frequency brain-computer interface data to obtain frequency information and phase information of each brain-computer interface data. The training module is used to encode each of the frequency information and each of the phase information respectively to obtain each frequency encoded information and each phase encoded information; wherein, the frequency encoded information is public information shared by each user, and the phase encoded information is private information not shared by each user; the frequency encoded information is decoded to obtain first model parameter information, and the individual brain-computer interface model is trained based on the first model parameter information to obtain an initial brain-computer interface model; The adjustment module is used to decode the phase encoding information of each user to obtain the second model parameters; adjust the parameter information of the initial brain-computer interface model based on the second model parameters and the first model parameters to obtain the brain-computer interface model of the user; and add privacy protection information to the brain-computer interface model of each user to obtain the privacy protection brain-computer interface model of each user. A construction module is used to identify the protection level of each privacy-preserving brain-computer interface model, and determine the target protection level of the global brain-computer interface model based on the protection levels of all privacy-preserving brain-computer interface models; based on the target protection level, generate global privacy protection information of the global brain-computer interface model through a privacy protection information generation strategy; and determine the target global brain-computer interface model based on the privacy-preserving brain-computer interface model of each user and the global privacy protection information.

4. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 2.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 2.

6. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 2.

Images