Automatic sleep staging method, device, equipment and medium

Abstract

The invention relates to an automatic sleep staging method, device and equipment and a medium, and the method comprises the steps: obtaining whole-brain calcium ion imaging data to be analyzed; preprocessing the whole brain calcium ion imaging data to be analyzed to obtain optimized whole brain calcium ion imaging data; inputting the optimized whole-brain calcium ion imaging data into an automatic sleep staging model to complete automatic sleep staging; the automatic sleep staging model comprises a feature extraction part used for extracting global features and a space / time mode of data from input whole-brain calcium ion imaging data; and the classifier part is used for completing automatic sleep staging based on the extracted global features and the space / time pattern. According to the invention, high-precision sleep stage classification can be realized.

Description

Technical Field

[0001] The present invention relates to the field of brain imaging technology, and in particular to an automatic sleep staging method, device, equipment and medium. Background Art

[0002] In recent years, with the rapid development of neuroscience and imaging technology, whole-brain calcium ion imaging technology has gradually become an important means of studying brain activity. Calcium ion imaging can monitor the activity of neurons in real time and provide dynamic data with high spatial and temporal resolution, thus providing important information for studying the brain's behavioral patterns and physiological processes (such as sleep). However, whole-brain calcium ion data are usually large and complex, and traditional sleep staging methods mainly rely on manual annotation by experts or simple algorithm analysis, which is not only time-consuming and labor-intensive, but also difficult to fully utilize the complex features in the data. Therefore, the existing supervised image classification model is not suitable for the analysis of whole-brain calcium ion data. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide an automatic sleep staging method, device, equipment and medium, which can achieve high-precision sleep stage classification.

[0004] The technical solution adopted by the present invention to solve the technical problem is: to provide an automatic sleep staging method, comprising the following steps:

[0005] Acquire whole-brain calcium imaging data to be analyzed;

[0006] Preprocessing the whole-brain calcium ion imaging data to be analyzed to obtain optimized whole-brain calcium ion imaging data;

[0007] Inputting the optimized whole-brain calcium ion imaging data into an automatic sleep staging model to complete automatic sleep staging;

[0008] Wherein, the automatic sleep staging model includes:

[0009] The feature extraction part is used to extract the global features and spatial / temporal patterns of the data from the input whole-brain calcium imaging data;

[0010] The classifier part is used to complete automatic sleep staging based on the extracted global features and spatial / temporal patterns.

[0011] The preprocessing of the whole-brain calcium ion imaging data to be analyzed specifically includes: performing decimal processing on the whole-brain calcium ion imaging data to be analyzed by means of data normalization.

[0012] The feature extraction part includes:

[0013] A mask module is used to randomly mask part of the time series in the input whole-brain calcium imaging data;

[0014] The encoder module is used to extract features from the masked whole-brain calcium imaging data to obtain the global features and spatial / temporal patterns of the data;

[0015] A decoder module is used to restore the masked part of the whole-brain calcium ion imaging data based on the global features and spatial / temporal patterns.

[0016] The encoder module adopts the ViT model, and the decoder module adopts the Transformer model, which is only used during training.

[0017] The loss function used in the feature extraction part during training is a reconstruction loss function.

[0018] The technical solution adopted by the present invention to solve the technical problem is: to provide an automatic sleep staging device, comprising:

[0019] An acquisition module, used to acquire whole-brain calcium ion imaging data to be analyzed;

[0020] A preprocessing module, used for preprocessing the whole-brain calcium ion imaging data to be analyzed to obtain optimized whole-brain calcium ion imaging data;

[0021] An automatic staging module, used for inputting the optimized whole-brain calcium ion imaging data into an automatic sleep staging model to complete automatic sleep staging;

[0022] Wherein, the automatic sleep staging model includes:

[0023] The feature extraction part is used to extract the global features and spatial / temporal patterns of the data from the input whole-brain calcium imaging data;

[0024] The classifier part is used to complete automatic sleep staging based on the extracted global features and spatial / temporal patterns.

[0025] The preprocessing module includes: a decimal processing unit, which is used to perform decimal processing on the whole-brain calcium ion imaging data to be analyzed by adopting a data normalization method.

[0026] The feature extraction part includes:

[0027] A mask module is used to randomly mask part of the time series in the input whole-brain calcium imaging data;

[0028] The encoder module is used to extract features from the masked whole-brain calcium imaging data to obtain the global features and spatial / temporal patterns of the data;

[0029] A decoder module is used to restore the masked part of the whole-brain calcium ion imaging data based on the global features and spatial / temporal patterns.

[0030] The technical solution adopted by the present invention to solve its technical problem is: to provide an electronic device, including a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned automatic sleep staging method when executing the computer program.

[0031] The technical solution adopted by the present invention to solve its technical problem is: providing a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned automatic sleep staging method are implemented.

[0032] Beneficial Effects

[0033] Due to the adoption of the above technical solution, the present invention has the following advantages and positive effects compared with the prior art: the present invention performs random masking on part of the whole-brain calcium ion data, forcing the model to recover global information from incomplete input. This method can effectively learn the key patterns in the whole-brain calcium ion data, thereby utilizing the dynamic information of the global calcium ion data to automatically analyze the activity patterns of different brain regions and achieve high-precision sleep stage classification. The method of the present invention can capture more physiological information, has higher accuracy than traditional single-dimensional analysis methods, and can automatically extract features from the data without labels during the training process of the feature extraction part, thereby greatly reducing the reliance on expert annotations, which not only reduces costs, but also provides the possibility of processing large-scale data sets. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a flow chart of the automatic sleep staging method according to the first embodiment of the present invention;

[0035] Figure 2 This is a whole-brain calcium ion image reconstruction effect diagram in the first embodiment of the present invention;

[0036] Figure 3 is a characteristic distribution diagram of calcium ions in the first embodiment of the present invention;

[0037] Figure 4 is a classification accuracy index diagram of the automatic sleep staging model in the first embodiment of the present invention;

[0038] Figure 5 4 is a structural diagram of an automatic sleep staging device according to a second embodiment of the present invention. DETAILED DESCRIPTION

[0039] The present invention will be further described below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms fall within the scope limited by the appended claims of the application equally.

[0040] The first embodiment of the present invention relates to an automatic sleep staging method, such as Figure 1 As shown, the following steps are included:

[0041] Step 1: Obtain the whole-brain calcium imaging data to be analyzed. This data is usually presented in the form of a time series, including changes in calcium ion concentrations in different brain regions at different time points.

[0042] Step 2, preprocessing the whole-brain calcium ion imaging data to be analyzed to obtain optimized whole-brain calcium ion imaging data. The preprocessing in this step can be decimal processing, that is, decimal processing is performed on the whole-brain calcium ion imaging data to be analyzed by means of data normalization. The optimized whole-brain calcium ion imaging data obtained by decimal processing in this step is convenient for subsequent model input.

[0043] Step 3, input the optimized whole-brain calcium ion imaging data into the automatic sleep staging model to complete the automatic sleep staging. The automatic sleep staging model in this step includes: a feature extraction part, which is used to extract the global features and spatial / temporal patterns of the data from the input whole-brain calcium ion imaging data; and a classifier part, which is used to complete the automatic sleep staging based on the extracted global features and spatial / temporal patterns.

[0044] The feature extraction part in this embodiment includes: a mask module, which is used to randomly mask part of the time series in the input whole-brain calcium ion imaging data; an encoder module, which is used to extract features from the masked whole-brain calcium ion imaging data to obtain the global features and space / time patterns of the data; a decoder module, which is used to restore the masked part of the whole-brain calcium ion imaging data based on the global features and space / time patterns, so as to adaptively learn the dynamic features of calcium ions in different stages of sleep. Among them, the encoder module can adopt the ViT model, and the decoder module can adopt the Transformer model, and it is only used during training.

[0045] It can be seen that this embodiment forces the model to recover global information from incomplete input by randomly masking part of the whole-brain calcium ion data. This method can effectively learn the key patterns in the whole-brain calcium ion data, and then use the dynamic information of the global calcium ion data to automatically analyze the activity patterns of different brain regions and achieve high-precision sleep stage classification.

[0046] The automatic sleep staging model in this embodiment specifically includes the following steps during training:

[0047] S1, data preprocessing:

[0048] Obtain a real dataset of original whole-brain calcium ion imaging data, which contains long-term neural activity records in different sleep stages, covering rapid eye movement (REM) sleep, non-rapid eye movement sleep (NREM) and wakefulness. Perform the same decimal processing as step 2 on each whole-brain calcium ion imaging data in the real dataset.

[0049] S2, build an automatic sleep staging model and train it:

[0050] The model is built according to the structure of the automatic sleep staging model in step 3. After the model is built, the preprocessed real data set is divided into a training set and a test set. The training set is used to train the automatic sleep staging model. When training the feature extraction part, the feature extraction part is optimized by the reconstruction loss function. After the training is completed, the trained automatic sleep staging model is tested using the test set. During the test, the decoder module in the feature extraction part is stopped, and the output of the encoder module is directly input to the classifier part. If the test result meets the requirements, the training of the automatic sleep staging model is completed. If the test result does not meet the requirements, the automatic sleep staging model is retrained.

[0051] This embodiment uses a real data set from whole-brain calcium ion imaging of mice for training. The training results can better reconstruct whole-brain calcium ion imaging images. Figure 2 This is the effect of whole-brain calcium ion image reconstruction.

[0052] After being processed by the encoder part of the automatic sleep staging model, the original high-dimensional calcium ion data is compressed into a low-dimensional latent space. In this latent space, the model automatically learns and extracts features related to different sleep stages, thereby mapping the complex raw data into three main clusters or distributions, such as Figure 3 As shown, they correspond to the REM sleep, wakefulness and non-REM sleep stages respectively.

[0053] The performance evaluation of the automatic sleep staging model in this embodiment is based on the classification accuracy (measurement of the proportion of sleep stages correctly classified by the model). The experimental results show that the automatic sleep staging model in this embodiment performs well in the evaluation indicators. The specific results are as follows: Figure 4 shown.

[0054] It is not difficult to find that this implementation method realizes efficient and automated analysis of complex brain imaging data through the innovative combination of self-supervised learning and masking mechanism. It not only improves the accuracy and efficiency of sleep staging, but also provides a new technical path for the in-depth analysis of whole-brain calcium ion data.

[0055] A second embodiment of the present invention relates to an automatic sleep staging device, such as Figure 5 As shown, including:

[0056] An acquisition module, used to acquire whole-brain calcium ion imaging data to be analyzed;

[0057] A preprocessing module, used for preprocessing the whole-brain calcium ion imaging data to be analyzed to obtain optimized whole-brain calcium ion imaging data;

[0058] An automatic staging module, used for inputting the optimized whole-brain calcium ion imaging data into an automatic sleep staging model to complete automatic sleep staging;

[0059] Wherein, the automatic sleep staging model includes:

[0060] The feature extraction part is used to extract the global features and spatial / temporal patterns of the data from the input whole-brain calcium imaging data;

[0061] The classifier part is used to complete automatic sleep staging based on the extracted global features and spatial / temporal patterns.

[0062] The preprocessing module includes: a decimal processing unit, which is used to perform decimal processing on the whole-brain calcium ion imaging data to be analyzed by adopting a data normalization method.

[0063] The feature extraction part includes:

[0064] A mask module is used to randomly mask part of the time series in the input whole-brain calcium imaging data;

[0065] The encoder module is used to extract features from the masked whole-brain calcium imaging data to obtain the global features and spatial / temporal patterns of the data;

[0066] A decoder module is used to restore the masked part of the whole-brain calcium ion imaging data based on the global features and spatial / temporal patterns.

[0067] The encoder module adopts the ViT model, and the decoder module adopts the Transformer model, which is only used during training.

[0068] The loss function used in the feature extraction part during training is a reconstruction loss function.

[0069] The third embodiment of the present invention relates to an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the automatic sleep staging method of the first embodiment when executing the computer program.

[0070] A fourth embodiment of the present invention relates to a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the automatic sleep staging method of the first embodiment are implemented.

[0071] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) that contain computer-usable program code.

[0072] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0073] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction method, which is implemented in the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0074] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0075] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An automatic sleep staging method, characterized in that: The following steps are involved: Acquire whole-brain calcium imaging data to be analyzed; Preprocessing the whole-brain calcium ion imaging data to be analyzed to obtain optimized whole-brain calcium ion imaging data; Inputting the optimized whole-brain calcium ion imaging data into an automatic sleep staging model to complete automatic sleep staging; Wherein, the automatic sleep staging model includes: The feature extraction part is used to extract the global features and spatial / temporal patterns of the data from the input whole-brain calcium imaging data; The classifier part is used to complete automatic sleep staging based on the extracted global features and spatial / temporal patterns.

2. The automatic sleep staging method according to claim 1, characterized in that: The preprocessing of the whole-brain calcium ion imaging data to be analyzed specifically includes: performing decimal processing on the whole-brain calcium ion imaging data to be analyzed by means of data normalization.

3. The automatic sleep staging method according to claim 1, characterized in that: The feature extraction part includes: The mask module is used to randomly mask part of the time series in the input whole-brain calcium ion imaging data; the encoder module is used to extract features from the masked whole-brain calcium ion imaging data to obtain the global features and spatial / temporal patterns of the data; A decoder module is used to restore the masked part of the whole-brain calcium ion imaging data based on the global features and spatial / temporal patterns.

4. The automatic sleep staging method according to claim 3, characterized in that: The encoder module adopts the ViT model, and the decoder module adopts the Transformer model, which is only used during training.

5. The automatic sleep staging method according to claim 1, characterized in that: The loss function used in the feature extraction part during training is a reconstruction loss function.

6. An automatic sleep staging device, characterized in that: include: An acquisition module, used to acquire whole-brain calcium ion imaging data to be analyzed; A preprocessing module, used for preprocessing the whole-brain calcium ion imaging data to be analyzed to obtain optimized whole-brain calcium ion imaging data; An automatic staging module, used for inputting the optimized whole-brain calcium ion imaging data into an automatic sleep staging model to complete automatic sleep staging; Wherein, the automatic sleep staging model includes: The feature extraction part is used to extract the global features and spatial / temporal patterns of the data from the input whole-brain calcium imaging data; The classifier part is used to complete automatic sleep staging based on the extracted global features and spatial / temporal patterns.

7. The automatic sleep staging device according to claim 6, characterized in that: The preprocessing module includes: a decimal processing unit, which is used to perform decimal processing on the whole-brain calcium ion imaging data to be analyzed by adopting a data normalization method.

8. The automatic sleep staging device according to claim 6, characterized in that: The feature extraction part includes: The mask module is used to randomly mask part of the time series in the input whole-brain calcium ion imaging data; the encoder module is used to extract features from the masked whole-brain calcium ion imaging data to obtain the global features and spatial / temporal patterns of the data; A decoder module is used to restore the masked part of the whole-brain calcium ion imaging data based on the global features and spatial / temporal patterns.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the automatic sleep staging method according to any one of claims 1 to 5 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the automatic sleep staging method according to any one of claims 1 to 5 are implemented.

Images