CN115067878A - EEGNet-based resting state electroencephalogram consciousness disorder classification method and system - Google Patents
EEGNet-based resting state electroencephalogram consciousness disorder classification method and system Download PDFInfo
- Publication number
- CN115067878A CN115067878A CN202210596059.6A CN202210596059A CN115067878A CN 115067878 A CN115067878 A CN 115067878A CN 202210596059 A CN202210596059 A CN 202210596059A CN 115067878 A CN115067878 A CN 115067878A
- Authority
- CN
- China
- Prior art keywords
- eegnet
- layer
- electroencephalogram
- module
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/40—Detecting, measuring or recording for evaluating the nervous system
- A61B5/4076—Diagnosing or monitoring particular conditions of the nervous system
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/369—Electroencephalography [EEG]
- A61B5/372—Analysis of electroencephalograms
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/369—Electroencephalography [EEG]
- A61B5/372—Analysis of electroencephalograms
- A61B5/374—Detecting the frequency distribution of signals, e.g. detecting delta, theta, alpha, beta or gamma waves
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/725—Details of waveform analysis using specific filters therefor, e.g. Kalman or adaptive filters
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
Abstract
The invention relates to a resting state electroencephalogram consciousness disorder classification method and system based on EEGNet, wherein the classification method comprises the following steps: s1, acquiring an electroencephalogram signal data set, and dividing according to cases; s2, filtering the divided electroencephalogram signals; s3, resampling the EEG signal data after filtering processing to obtain a sample set; s4, setting an electroencephalogram threshold value, and eliminating samples with electroencephalogram values exceeding the electroencephalogram threshold value in a sample set to obtain a target sample set; dividing a target sample set of each case into a training set and a testing set; s5, constructing an EEGNet model, and training the EEGNet model by utilizing a training set; s6, performing performance evaluation on the trained EEGNet model by using a test set to obtain a target EEGNet model; and S7, inputting the electroencephalogram signals to be detected into the target EEGNet model to obtain a classification result. The invention has high classification precision and can accurately diagnose the patient.
Description
Technical Field
The invention belongs to the technical field of consciousness disorder classification, and particularly relates to a resting state electroencephalogram consciousness disorder classification method and system based on EEGNet.
Background
The disturbance of Consciousness can be divided into a Minimum Consciousness State (MCS) and an Unresponsive arousal Syndrome (UWS), and currently, clinically, a behavior scale score is generally used as a diagnosis standard of disturbance of Consciousness, but the mapping relation between behavior and Consciousness is not strict, and the disturbance of Consciousness is difficult to be accurately diagnosed only according to the behavior. Studies have found that current assessment methods based on clinical scales result in higher rates of misdiagnosis.
Disclosure of Invention
Based on the above disadvantages and shortcomings in the prior art, the present invention provides a method and system for classifying resting state electroencephalogram consciousness disorder based on EEGNet.
In order to achieve the purpose, the invention adopts the following technical scheme:
the resting state electroencephalogram consciousness disturbance classification method based on EEGNet comprises the following steps:
s1, acquiring an electroencephalogram signal data set, and dividing according to cases;
s2, filtering the divided electroencephalogram signals, reserving alpha, beta, delta and theta frequency bands and removing interference of other frequency bands;
s3, resampling the EEG signal data after filtering processing to obtain a sample set;
s4, setting an electroencephalogram threshold value, and eliminating samples with electroencephalogram values exceeding the electroencephalogram threshold value in a sample set to obtain a target sample set; dividing a target sample set of each case into a training set and a testing set;
s5, constructing an EEGNet model, and training the EEGNet model by using a training set;
s6, performing performance evaluation on the trained EEGNet model by using the test set to obtain a target EEGNet model;
and S7, inputting the electroencephalogram signals to be detected into the target EEGNet model to obtain a classification result.
Preferably, the step S2 specifically includes the following steps:
s21, carrying out notch filtering of 49 Hz-51 Hz, and removing power frequency interference of 50 Hz;
and S22, sequentially carrying out high-pass filtering of 0.5Hz and low-pass filtering of 40 Hz.
Preferably, in step S3, the resampling is performed by using a sliding window method.
As a preferred scheme, the resampling is performed on the filtered electroencephalogram signal data, and the number of obtained samples is as follows:
wherein, SampleNum is the number of samples, L is the length of the electroencephalogram signal, L is the length of the sliding window, and step is the sliding step length.
Preferably, the sliding window length is 10 seconds, and the sliding step length is 1 second.
Preferably, the electroencephalogram threshold is 200 μ V.
As a preferred scheme, the EEGNet model includes an input module, a first convolution module, a second convolution module, a third convolution module and a full connection module, which are connected in sequence, the first convolution module includes a Conv layer, a DepthwiseConv layer, an average pooling layer and a dropout layer, which are connected in sequence, the second convolution module includes a separateconv layer, an average pooling layer and a dropout layer, which are connected in sequence, the third convolution module includes a separateconv layer, an average pooling layer and a dropout layer, which are connected in sequence, and the full connection module includes an FC layer and a softmax layer, which are connected in sequence.
Preferably, the Conv layer of the first convolution module further passes through a BatchNorm layer, and the DepthwiseConv layer further passes through the BatchNorm layer and the ReLU layer in sequence;
the SeparableConv layer of the second convolution module sequentially passes through a BatchNorm layer and a ReLU layer;
the separatableconv layer of the third convolution module is then also passed through the BatchNorm layer and the ReLU layer in sequence.
Preferably, the kernel of the Conv layer of the first convolution module is 1 × 127;
the average pooling layer of the first convolution module is 1 × 4 average pooling, and the average pooling layers of the second convolution module and the third convolution module are both 1 × 8 average pooling.
The invention also provides a resting state electroencephalogram consciousness disorder classification system based on EEGNet, which applies the resting state electroencephalogram consciousness disorder classification method based on EEGNet according to any one of the above schemes, and the resting state electroencephalogram consciousness disorder classification system based on EEGNet comprises:
the dividing module is used for dividing the acquired electroencephalogram signal data set according to cases;
the filtering module is used for carrying out filtering processing on the divided electroencephalogram signals, reserving alpha, beta, delta and theta frequency bands and removing interference of other frequency bands;
the resampling module is used for resampling the EEG signal data after the filtering processing to obtain a sample set;
the elimination module is used for eliminating samples with the electroencephalogram signal value exceeding the electroencephalogram signal threshold value in the sample set to obtain a target sample set; the dividing module is further used for dividing the target sample set into a training set and a testing set;
the building module is used for building an EEGNet model;
the training module is used for training the EEGNet model by utilizing a training set;
the evaluation module is used for evaluating the performance of the trained EEGNet model by using a test set to obtain a target EEGNet model;
and the test module is used for inputting the electroencephalogram signals to be tested into the target EEGNet model to obtain a classification result.
Compared with the prior art, the invention has the beneficial effects that:
the EEGNet-based resting state electroencephalogram consciousness disorder classification method and system are high in classification precision and can accurately diagnose patients.
Drawings
FIG. 1 is a schematic structural diagram of the EEGNet model of example 1 of the present invention;
fig. 2 is a structural diagram of the resting state electroencephalogram consciousness disorder classification system based on the EEGNet in embodiment 1 of the present invention.
Detailed Description
In order to more clearly illustrate the embodiments of the present invention, the following description will explain the embodiments of the present invention with reference to the accompanying drawings. It is obvious that the drawings in the following description are only some examples of the invention, and that for a person skilled in the art, other drawings and embodiments can be derived from them without inventive effort.
Example 1:
the resting state electroencephalogram consciousness disorder classification method based on EEGNet in the embodiment specifically comprises the following steps:
s1, the medical staff installs the electroencephalogram signal acquisition module at the head of the patient, then controls the electroencephalogram signal acquisition module to acquire electroencephalogram signals through the signal analysis module, and divides the acquired data into cases;
s2, carrying out filtering processing on the EEG signals (namely EEG data) divided in the step S1, firstly carrying out notch filtering of 49 Hz-51 Hz to remove power frequency interference of 50Hz, and then carrying out high-pass filtering of 0.5Hz and low-pass filtering of 40Hz in sequence, reserving frequency bands of alpha, beta, delta and theta and removing interference of other frequency bands.
S3, resampling the filtered EEG data by sliding window to obtain a certain number of samples, where an example of the number of samples obtained for the EEG signal is as follows:
wherein, SampleNum is the number of samples, L is the length of the EEG signal, L is the length of the sliding window, and step is the sliding step length. In order to ensure that enough information exists in a single sample for distinguishing the MCS from the UWS, the single sample duration is set to 10 seconds; in order to increase the sample size and avoid information loss, resampling is performed every 1 second; i.e. a sliding window length of 10 seconds and a sliding step of 1 second.
And S4, setting a threshold, eliminating samples exceeding the threshold data, and setting the samples of the respective cases as a training set and a testing set, so that the training samples and the testing samples are not from EEG signals of the same case. To retain more EEG signals, outliers of EEG signals were culled at 200 μ V as a threshold.
S5, constructing an EEGNet model, and training the EEGNet model by using a training sample;
due to the inconsistent dimensionality of the input data, the structure of the existing EEGNet is adjusted, and the adjusted structure of the EEGNet is shown in fig. 1 and table 1.
TABLE 1 structural Table of EEGNet model
The EEGNet uses deep separable convolution instead of general convolution, and has the main characteristic of less parameter number, so the EEGNet is a simplified convolution neural network. The existing EEGNet model only has 2 convolution modules, and the EEG signal sample time of the embodiment is long, so 1 additional convolution module is added, and 3 additional convolution modules are provided. Specifically, the EEGNet model comprises an input module, a first convolution module, a second convolution module, a third convolution module and a full-connection module which are connected in sequence, wherein the first convolution module comprises a Conv layer, a DepthwiseConv layer, an average pooling layer and a dropout layer which are connected in sequence, the second convolution module comprises a SeparableConv layer, an average pooling layer and a dropout layer which are connected in sequence, the third convolution module comprises a SeparableConv layer, an average pooling layer and a dropout layer which are connected in sequence, and the full-connection module comprises an FC layer and a softmax layer which are connected in sequence.
The Conv layer of the first convolution module plays a role of a filter without using an activation function; the Conv layer is then also passed through the BatchNorm layer. The sampling frequency of an EEG signal processed by the prior EEGNet is 128Hz, a convolution kernel with the size of 1 x 64 is used in a first convolution layer of a first convolution module, and the size of the convolution kernel in a time dimension is half of the sampling frequency so as to extract frequency information above 2 Hz; on the other hand, since the sampling frequency of the EEG signal in this embodiment is 256Hz, the convolution kernel size of the Conv layer is increased to 1 × 127. The DepthwiseConv layer provides a different spatial filter for each different frequency, and then also passes through the BatchNorm layer and the ReLU layer in sequence. In addition, the average pooling layer of the first convolution module was an average pooling of 1 × 4.
The separatableconv layer of the second convolution module of this embodiment uses depth separable convolution instead of general convolution, which can effectively reduce the number of parameters of the model; then sequentially passing through a BatchNorm layer and a ReLU layer; the average pooling layer of the second convolution module uses an average pooling of size 1 x 8.
The third convolution module of this embodiment is designed to accommodate a longer sampling duration, and has substantially the same structure as the second convolution module, and continues to perform dimensionality reduction in the time dimension using an average pooling with a size of 1 × 8, and deepens the feature map depth by 2 times as much as before.
The feature map size finally input to the fully-connected module is 1 × 10 × 256, the fully-connected module includes 1 fully-connected layer (FC layer), and finally the sorted output is realized by the softmax layer.
In the training process, batch normalization is added after all convolution calculations, and dropout is set to be 0.5 at the end of all convolution modules. The number of trainable parameters of the EEGNet model is 101, 150.
S6, evaluating the performance of the EEGNet model obtained by training in the step S5 by using the test sample;
and S7, inputting electroencephalogram signals of the patient to be diagnosed to be acquired as models, and calculating classification results.
The method for classifying the resting state electroencephalogram consciousness disturbance based on the EEGNet is verified as follows:
the collected resting state electroencephalogram data of the consciousness disorder patients are 153, wherein the minimum consciousness state is 102, and the unresponsive arousal is 51. And finally, testing each model on a corresponding test set by adopting five-fold cross validation mode to finally obtain five models from 51450 samples obtained after resampling, and finally averaging, wherein the classification accuracy is 84.96%, and the area under the operation characteristic curve of a receiver for classifying cases is 91.46%. In contrast, long-short term memory networks and recurrent neural networks were validated with accuracy of 79.10% and 77.68% for classification of samples and 84.06% and 83.26% for areas under the receiver operating characteristic curve for case classification.
Corresponding to the resting state electroencephalogram consciousness disorder classification method based on EEGNet of the present embodiment, the present embodiment further provides a resting state electroencephalogram consciousness disorder classification system based on EEGNet, including:
the dividing module is used for dividing the acquired electroencephalogram signal data set according to cases; the detailed division can refer to the prior art and is not described herein.
The filtering module is used for carrying out filtering processing on the divided electroencephalogram signals, reserving alpha, beta, delta and theta frequency bands and removing interference of other frequency bands; the specific filtering process may be a specific step in the above method, which is not described herein again.
The resampling module is used for resampling the EEG signal data after the filtering processing to obtain a sample set; the specific steps in the above method can be used for weight sampling, which is not described herein again.
The elimination module is used for eliminating samples with the electroencephalogram signal value exceeding the electroencephalogram signal threshold value in the sample set to obtain a target sample set; the dividing module is further used for dividing the target sample set into a training set and a testing set;
the building module is used for building an EEGNet model; for a specific EEGNet model, reference may be made to the detailed description of the above method, which is not repeated herein.
The training module is used for training the EEGNet model by utilizing a training set;
the evaluation module is used for evaluating the performance of the trained EEGNet model by using a test set to obtain a target EEGNet model;
and the test module is used for inputting the electroencephalogram signals to be tested into the target EEGNet model to obtain a classification result.
The foregoing has outlined, rather broadly, the preferred embodiment and principles of the present invention in order that those skilled in the art may better understand the detailed description of the invention without departing from its broader aspects.
Claims (10)
1. The EEGNet-based resting state electroencephalogram consciousness disturbance classification method is characterized by comprising the following steps:
s1, acquiring an electroencephalogram signal data set, and dividing according to cases;
s2, filtering the divided electroencephalogram signals, reserving alpha, beta, delta and theta frequency bands and removing interference of other frequency bands;
s3, resampling the EEG signal data after filtering processing to obtain a sample set;
s4, setting an electroencephalogram threshold value, and removing samples with electroencephalogram values exceeding the electroencephalogram threshold value in the sample set to obtain a target sample set; dividing a target sample set of each case into a training set and a testing set;
s5, constructing an EEGNet model, and training the EEGNet model by utilizing a training set;
s6, performing performance evaluation on the trained EEGNet model by using a test set to obtain a target EEGNet model;
and S7, inputting the electroencephalogram signals to be detected into the target EEGNet model to obtain a classification result.
2. The EEGNet-based resting state electroencephalogram consciousness disorder classification method according to claim 1, wherein the step S2 specifically comprises the following steps:
s21, carrying out notch filtering of 49 Hz-51 Hz, and removing power frequency interference of 50 Hz;
and S22, sequentially carrying out high-pass filtering of 0.5Hz and low-pass filtering of 40 Hz.
3. The EEGNet-based resting state EEG-based consciousness disorder classification method of claim 1, wherein in said step S3, a sliding window manner is adopted for resampling.
4. The EEGNet-based resting state EEG-based consciousness disturbance classification method according to claim 3, wherein said resampling is performed on the EEG data after filtering processing, and the number of samples obtained is:
wherein, SampleNum is the number of samples, L is the length of the electroencephalogram signal, L is the length of the sliding window, and step is the sliding step length.
5. The EEGNet-based resting state EEG-based consciousness disorder classification method according to claim 4, wherein said sliding window length is 10 seconds and the sliding step size is 1 second.
6. The EEGNet-based resting state EEG-based consciousness disorder classification method according to claim 1, wherein said EEG signal threshold is 200 μ V.
7. The EEGNet-based resting state electroencephalogram consciousness disorder classification method according to claim 1, wherein the EEGNet model comprises an input module, a first convolution module, a second convolution module, a third convolution module and a fully connected module which are connected in sequence, the first convolution module comprises a Conv layer, a DepthWiseConv layer, an average pooling layer and a dropout layer which are connected in sequence, the second convolution module comprises a SeparableConv layer, an average pooling layer and a dropout layer which are connected in sequence, the third convolution module comprises a SeparableConv layer, an average pooling layer and a dropout layer which are connected in sequence, and the fully connected module comprises an FC layer and a softmax layer which are connected in sequence.
8. The EEGNet-based resting state EEG-based consciousness disorder classification method of claim 7, wherein said first convolution module Conv layer is further followed by a BatchNorm layer, and said DepthwiseConv layer is further followed by a BatchNorm layer and a ReLU layer in sequence;
the SeparableConv layer of the second convolution module sequentially passes through a BatchNorm layer and a ReLU layer;
the separatableconv layer of the third convolution module is then also passed through the BatchNorm layer and the ReLU layer in that order.
9. The EEGNet-based resting state EEG-based consciousness disorder classification method of claim 7 or 8, wherein said first convolution module Conv layer kernel is 1 x 127;
the average pooling layer of the first convolution module is 1 × 4 average pooling, and the average pooling layers of the second convolution module and the third convolution module are both 1 × 8 average pooling.
10. The EEGNet-based resting state EEG-based brain electrical consciousness disorder classification system is applied to the EEGNet-based resting state EEG-based brain electrical consciousness disorder classification method according to any one of claims 1 to 9, and is characterized in that the EEGNet-based resting state EEG-based brain electrical consciousness disorder classification system comprises:
the dividing module is used for dividing the acquired electroencephalogram signal data set according to cases;
the filtering module is used for filtering the divided electroencephalogram signals, reserving frequency bands of alpha, beta, delta and theta and removing interference of other frequency bands;
the resampling module is used for resampling the EEG signal data after the filtering processing to obtain a sample set;
the elimination module is used for eliminating samples with the electroencephalogram signal value exceeding the electroencephalogram signal threshold value in the sample set to obtain a target sample set; the dividing module is further used for dividing the target sample set into a training set and a testing set;
the building module is used for building an EEGNet model;
the training module is used for training the EEGNet model by utilizing a training set;
the evaluation module is used for evaluating the performance of the trained EEGNet model by using a test set to obtain a target EEGNet model;
and the test module is used for inputting the EEGNet model to be tested into the EEGNet model to obtain a classification result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210596059.6A CN115067878A (en) | 2022-05-30 | 2022-05-30 | EEGNet-based resting state electroencephalogram consciousness disorder classification method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210596059.6A CN115067878A (en) | 2022-05-30 | 2022-05-30 | EEGNet-based resting state electroencephalogram consciousness disorder classification method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115067878A true CN115067878A (en) | 2022-09-20 |
Family
ID=83250298
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210596059.6A Pending CN115067878A (en) | 2022-05-30 | 2022-05-30 | EEGNet-based resting state electroencephalogram consciousness disorder classification method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115067878A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117725490A (en) * | 2024-02-08 | 2024-03-19 | 山东大学 | Cross-test passive pitch-aware EEG automatic classification method and system |
CN117725490B (en) * | 2024-02-08 | 2024-04-26 | 山东大学 | Cross-test passive pitch-aware EEG automatic classification method and system |
-
2022
- 2022-05-30 CN CN202210596059.6A patent/CN115067878A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117725490A (en) * | 2024-02-08 | 2024-03-19 | 山东大学 | Cross-test passive pitch-aware EEG automatic classification method and system |
CN117725490B (en) * | 2024-02-08 | 2024-04-26 | 山东大学 | Cross-test passive pitch-aware EEG automatic classification method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20070260151A1 (en) | Method and device for filtering, segmenting, compressing and classifying oscillatory signals | |
CN111166327A (en) | Epilepsy diagnosis device based on single-channel electroencephalogram signal and convolutional neural network | |
CN111839506B (en) | Mental load detection method and device | |
US9477640B2 (en) | Neural signal processing and/or interface methods, architectures, apparatuses, and devices | |
Deng et al. | Complexity extraction of electroencephalograms in Alzheimer's disease with weighted-permutation entropy | |
CN112957055B (en) | EEMD-PCA-based method and device for removing motion artifacts in EEG signals | |
CN103970975A (en) | Electrocardio data processing method and electrocardio data processing system | |
CN113662560B (en) | Method for detecting seizure-like discharge between attacks, storage medium and device | |
CN114190944A (en) | Robust emotion recognition method based on electroencephalogram signals | |
CN114391846A (en) | Emotion recognition method and system based on filtering type feature selection | |
CN111067513B (en) | Sleep quality detection key brain area judgment method based on characteristic weight self-learning | |
CN111557662B (en) | Epileptic interval electroencephalogram signal processing method and device, storage medium and equipment | |
CN104622467A (en) | Method for detecting electroencephalogram signal complexity abnormity of Alzheimer disease | |
CN113723557A (en) | Depression electroencephalogram classification system based on multiband time-space convolution network | |
CN113343869A (en) | Electroencephalogram signal automatic classification and identification method based on NTFT and CNN | |
CN113197583A (en) | Electrocardiogram waveform segmentation method based on time-frequency analysis and recurrent neural network | |
CN112274145A (en) | Method and device for processing near-infrared brain function imaging data and storage medium | |
CN115067878A (en) | EEGNet-based resting state electroencephalogram consciousness disorder classification method and system | |
CN116211322A (en) | Depression recognition method and system based on machine learning electroencephalogram signals | |
CN110507299A (en) | Heart rate signal detection device and method | |
CN113598791B (en) | Consciousness disturbance classification method based on time-space convolution neural network used by resting state electroencephalogram | |
CN110516711B (en) | Training set quality evaluation method of MI-BCI system and optimization method of single training sample | |
CN111783857A (en) | Motor imagery brain-computer interface based on nonlinear network information graph | |
CN113229781A (en) | Cross-patient epilepsy detection system based on multiple views | |
CN104007823B (en) | Frequency phase hybrid decoding brain-computer interface method based on point-pass filtering and frequency phase hybrid decoding brain-computer interface device based on point-pass filtering |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |