CN113598791A - Consciousness disturbance classification method using space-time convolution neural network based on resting electroencephalogram - Google Patents
Consciousness disturbance classification method using space-time convolution neural network based on resting electroencephalogram Download PDFInfo
- Publication number
- CN113598791A CN113598791A CN202110790906.8A CN202110790906A CN113598791A CN 113598791 A CN113598791 A CN 113598791A CN 202110790906 A CN202110790906 A CN 202110790906A CN 113598791 A CN113598791 A CN 113598791A
- Authority
- CN
- China
- Prior art keywords
- neural network
- electroencephalogram
- consciousness
- disturbance
- space
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 230000000284 resting effect Effects 0.000 title claims abstract description 23
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 9
- 238000012360 testing method Methods 0.000 claims abstract description 34
- 238000012549 training Methods 0.000 claims abstract description 30
- 238000001914 filtration Methods 0.000 claims abstract description 18
- 238000003062 neural network model Methods 0.000 claims abstract description 18
- 238000012952 Resampling Methods 0.000 claims abstract description 6
- 238000013527 convolutional neural network Methods 0.000 claims description 18
- 238000005070 sampling Methods 0.000 claims description 12
- 208000011580 syndromic disease Diseases 0.000 claims description 10
- 230000037007 arousal Effects 0.000 claims description 8
- 210000004556 brain Anatomy 0.000 claims description 7
- 238000003745 diagnosis Methods 0.000 abstract description 14
- 208000013159 conscious disturbance Diseases 0.000 abstract description 3
- 230000006399 behavior Effects 0.000 description 7
- 238000000605 extraction Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000002123 temporal effect Effects 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 3
- 208000022540 Consciousness disease Diseases 0.000 description 2
- 238000002790 cross-validation Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000004445 quantitative analysis Methods 0.000 description 2
- 206010010071 Coma Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013499 data model Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—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
- A61B5/24—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]
-
- 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
- A61B5/40—Detecting, measuring or recording for evaluating the nervous system
- A61B5/4076—Diagnosing or monitoring particular conditions of the nervous system
- A61B5/4088—Diagnosing of monitoring cognitive diseases, e.g. Alzheimer, prion diseases or dementia
-
- 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
- A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
-
- 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
- A61B5/7225—Details of analog processing, e.g. isolation amplifier, gain or sensitivity adjustment, filtering, baseline or drift compensation
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Public Health (AREA)
- Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Surgery (AREA)
- Psychiatry (AREA)
- Molecular Biology (AREA)
- Medical Informatics (AREA)
- Biophysics (AREA)
- Veterinary Medicine (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Signal Processing (AREA)
- Physiology (AREA)
- Artificial Intelligence (AREA)
- Neurology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Psychology (AREA)
- Evolutionary Computation (AREA)
- Power Engineering (AREA)
- Child & Adolescent Psychology (AREA)
- Developmental Disabilities (AREA)
- Hospice & Palliative Care (AREA)
- Mathematical Physics (AREA)
- Fuzzy Systems (AREA)
- Neurosurgery (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention belongs to the technical field of electroencephalogram application, and particularly relates to a method for classifying conscious disturbance by using a space-time convolution neural network based on resting electroencephalogram. The method comprises the following steps: s1, acquiring electroencephalogram signals of the patient with disturbance of consciousness, and randomly dividing the electroencephalogram signals into a training set and a testing set; s2, filtering the electroencephalogram signal, and resampling the training set and the test set to obtain a training sample and a test sample; s3, setting a threshold value, and removing training samples and test samples containing data exceeding the threshold value; s4, constructing a space-time convolution neural network model and training; s5, evaluating the performance of the trained model by using the test sample; and S6, acquiring electroencephalogram signals of the patient to be diagnosed as model input, and calculating a classification result. The invention has the characteristics of no dependence on the behavior reaction and the expert experience of the patient, high diagnosis accuracy and high diagnosis efficiency.
Description
Technical Field
The invention belongs to the technical field of electroencephalogram application, and particularly relates to a method for classifying conscious disturbance by using a space-time convolution neural network based on resting electroencephalogram.
Background
Disturbance of consciousness can be divided into minimal state of consciousness (MCS) and unresponsive arousal syndrome (UWS), however accurate diagnosis of a patient with disturbance of consciousness remains a problem. The revised coma-recovery scale (CRS-R) is the current standard clinical scale for the examination and assessment of disturbance of consciousness. However, the CRS-R scale is a diagnosis tool based on behaviors, depends on the behavior response of a patient and the expert experience of a clinician, and has great defects in accuracy and efficiency. While EEG has the advantages of high temporal resolution, low cost, high safety, etc., and can be deployed at the patient's bedside. The resting state refers to the state of a human being when awake, closed eye, or relaxed, and the resting EEG is acquired independent of the behavioral response of the patient. However, most of the related researches are carried out on the basis of prior knowledge for feature extraction at present, but the results are not ideal. The experts agree that current quantitative analysis of EEG studies are not sufficient and quantitative analysis of EEG is not recommended to distinguish MCS from UWS.
Deep learning is a method that focuses on learning deep-level data models, can understand and learn complex representations of a signal directly from the original signal, and has the advantage of automatically extracting high-level features required for classification. In recent years, with the increasing popularity of large EEG datasets, deep learning has been applied to the decoding and classification of EEG signals.
Therefore, it is necessary to design a method for classifying the disturbance of consciousness by using a space-time convolution neural network based on the resting electroencephalogram, which does not depend on the behavioral response and the expert experience of the patient, has high diagnosis accuracy and high diagnosis efficiency.
For example, the method for diagnosing disturbance of consciousness based on electroencephalogram signals, which is disclosed in application No. CN201910150296.8, specifically includes the following steps: s1, acquiring electroencephalogram signals: firstly, medical staff can install the electroencephalogram signal acquisition unit at each position of the head of a diagnostician, then the electroencephalogram signal acquisition unit is controlled by the central processing module to acquire the electroencephalogram signal of the head of the diagnostician, and S2 is used for denoising and filtering the electroencephalogram signal. Although the accuracy and the analysis processing speed of detection and evaluation are greatly improved, the filtering and denoising processing of the detected brain waves is realized, the interference of ocular artifacts and other signal sources is avoided, the purpose of simultaneously analyzing and processing the extracted four characteristic values is well achieved, and the automatic generation and automatic printing of a diagnosis and analysis table through an analysis algorithm after the completion of the electroencephalogram examination is realized, so that the diagnosis work of medical personnel is greatly facilitated, the defects are that the disease mechanism of the consciousness disorder is not clear at present, the information loss is easily caused by the characteristic extraction based on the priori knowledge, and the accuracy of classifying the consciousness disorder patients is low.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a consciousness disturbance classification method using a space-time convolution neural network based on a resting electroencephalogram, which solves the problems of low diagnosis accuracy and low efficiency caused by dependence on behavior reaction and expert experience of a patient in the conventional CRS-R scale method, and low diagnosis accuracy caused by easy information loss caused by electroencephalogram feature extraction based on priori knowledge.
In order to achieve the purpose, the invention adopts the following technical scheme:
the consciousness disturbance classification method using the space-time convolution neural network based on the resting electroencephalogram comprises the following steps:
s1, acquiring electroencephalogram signals of the patient with disturbance of consciousness, and randomly dividing the acquired electroencephalogram signals into a training set and a testing set;
s2, filtering the electroencephalogram signals collected in the step S1, and resampling the training set and the test set in the step S1 through a sliding window method to obtain training samples and test samples;
s3, setting a threshold value, and removing training samples and test samples containing data exceeding the threshold value;
s4, constructing a space-time convolution neural network model, and training the space-time convolution neural network model by using the training samples obtained in the step S3;
s5, evaluating the performance of the space-time convolution neural network model obtained by training in the step S4 by using the test sample obtained in the step S3;
and S6, collecting electroencephalogram signals of the patient with disturbance of consciousness to be diagnosed as model input, and directly calculating the probability that the patient with disturbance of consciousness to be diagnosed is in the minimum consciousness state and the unresponsive arousal syndrome by utilizing the space-time convolutional neural network model trained in the step S4.
Preferably, in the process of step S1, the sampling frequency of the electroencephalogram signal is 256Hz, and the acquisition time lasts for 5 minutes or more.
Preferably, the filtering process in step S2 includes the steps of:
s21, performing notch filtering of 49Hz-51Hz, high-pass filtering of 0.5Hz and low-pass filtering of 40Hz on the electroencephalogram signals collected in the step S1 in sequence.
Preferably, the sliding window method in step S2 includes the steps of:
and S22, setting the length of the sliding window as l and the sampling step length as step, and performing sliding sampling on the data of the whole section of electroencephalogram signals collected through the sliding window.
Preferably, the length of the sliding window is set to l-10 s, and the sampling step size is set to step-1 s.
Preferably, the threshold value is set to 200 μ V in step S3.
Preferably, in step S4, the process of constructing the spatio-temporal convolutional neural network model includes the following steps:
s41, the input data dimension is 16 × 2560 × 1;
s42, extracting the time characteristics of the electroencephalogram signal by using a one-dimensional convolution kernel in the time dimension, wherein the output data dimension is 16 multiplied by 640 multiplied by 32;
s43, extracting the spatial features of the electroencephalogram signals by using a one-dimensional convolution kernel in the spatial dimension, wherein the output data dimension is 1 multiplied by 640 multiplied by 128;
s44, extracting time characteristics of the global brain space characteristics by using a one-dimensional convolution kernel in the time dimension, wherein the output data dimension is 1 multiplied by 20 multiplied by 256;
and S45, realizing classification through the full connection layer and outputting a classification result.
Preferably, step S5 further includes the steps of:
s51, classifying each test sample by using the space-time convolution neural network model obtained by training in the step S4, and calculating the accuracy of the space-time convolution neural network model in classifying all the test samples;
s52, each case in the test set comprises a plurality of samples, wherein the proportion of the sample of the ith case classified as the minimum state of consciousness is piThe proportion of non-reactive wake syndrome is 1-piWith piAnd 1-piAs the classification probability of the ith case, a receiver operation characteristic curve is drawn with the non-reactive arousal syndrome as positive, based on the classification probability of each case of the test set, and the area under the curve is calculated.
Compared with the prior art, the invention has the beneficial effects that: (1) according to the invention, classification of patients with disturbance of consciousness can be realized only by keeping the patients in a resting state and not depending on behavioral response of the patients; (2) the invention does not depend on expert experience, and can give a diagnosis result only by an original electroencephalogram signal; (3) the invention has high diagnosis accuracy rate for the patient with disturbance of consciousness, and has a space for continuously improving along with the increase of data volume.
Drawings
FIG. 1 is a flow chart of an conscious disturbance classification method of the present invention using a spatiotemporal convolutional neural network based on resting electroencephalogram;
FIG. 2 is a schematic illustration of one position of an electroencephalogram electrode;
FIG. 3 is a schematic diagram of a process of the sliding window method of the present invention;
FIG. 4 is a schematic diagram of a process of constructing a spatio-temporal convolution neural network model according to 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 method for classifying disturbance of consciousness using spatio-temporal convolutional neural network based on resting electroencephalogram as shown in fig. 1, includes the steps of:
s1, acquiring electroencephalogram signals of the patient with disturbance of consciousness, and randomly dividing the acquired electroencephalogram signals into a training set and a testing set;
s2, filtering the electroencephalogram signals collected in the step S1, and resampling the training set and the test set in the step S1 through a sliding window method to obtain training samples and test samples;
s3, setting a threshold, and eliminating training samples and test samples containing data exceeding the threshold, wherein the threshold is set to be 200 mu V;
s4, constructing a space-time convolution neural network model, and training the space-time convolution neural network model by using the training samples obtained in the step S3;
s5, evaluating the performance of the space-time convolution neural network model obtained by training in the step S4 by using the test sample obtained in the step S3;
and S6, collecting electroencephalogram signals of the patient with disturbance of consciousness to be diagnosed as model input, and directly calculating the probability that the patient with disturbance of consciousness to be diagnosed is in the minimum consciousness state and the unresponsive arousal syndrome by utilizing the space-time convolutional neural network model trained in the step S4.
In step S1, the electroencephalogram signal acquired is a 16-lead electroencephalogram signal. According to the 10/20 system, electrodes were placed at Fp1, Fp2, F3, F4, F7, F8, C3, C4, T3, T4, T5, T6, P3, P4, O1, and O2, as shown in particular in fig. 2.
Further, in the process of step S1, the sampling frequency of the electroencephalogram signal is 256Hz, and the acquisition time lasts for 5 minutes or more.
Further, the filtering process in step S2 includes the following steps:
s21, performing notch filtering of 49Hz-51Hz, high-pass filtering of 0.5Hz and low-pass filtering of 40Hz on the electroencephalogram signals collected in the step S1 in sequence.
As shown in fig. 3, the sliding window method in step S2 includes the following steps:
and S22, setting the length of the sliding window as l and the sampling step length as step, and performing sliding sampling on the data of the whole section of electroencephalogram signals collected through the sliding window.
For example, the electroencephalogram data is 5 minutes or longer, and data of 10 seconds in length can be used as a sample of a model. We use a sliding window method to perform resampling, setting the window length l to 10s and the sampling step length step to 1s, and sliding the samples over the entire section of electroencephalogram data. I.e., 0-10 seconds for the 1 st sample, 1-11 seconds for the 2 nd sample, and so on. For a length of electroencephalogram data Ls, the sample size sampleNum that can be obtained is:
as shown in fig. 4, further, in step S4, the process of constructing the spatio-temporal convolutional neural network model includes the following steps:
s41, the input data dimension is 16 × 2560 × 1;
s42, extracting the time characteristics of the electroencephalogram signal by using a one-dimensional convolution kernel in the time dimension, wherein the output data dimension is 16 multiplied by 640 multiplied by 32;
s43, extracting the spatial features of the electroencephalogram signals by using a one-dimensional convolution kernel in the spatial dimension, wherein the output data dimension is 1 multiplied by 640 multiplied by 128;
s44, extracting time characteristics of the global brain space characteristics by using a one-dimensional convolution kernel in the time dimension, wherein the output data dimension is 1 multiplied by 20 multiplied by 256;
and S45, realizing classification through the full connection layer and outputting a classification result.
Through the process shown in fig. 4, the structure of the finally obtained space-time convolution neural network model is shown in the following table 1:
TABLE 1 spatio-temporal convolutional neural network model Structure
In Table 1, convolution modules 1-2 extract the temporal characteristics of each lead in the time dimension. The convolution module 3 is a spatial convolution module that integrates the features of the 16 leads of the EEG signal into the spatial features of the whole brain. The convolution modules 4-8 extract temporal features of the global brain spatial features in the temporal dimension. And finally, classifying by a full connection layer.
Further, step S5 includes the following steps:
s51, classifying each test sample by using the space-time convolution neural network model obtained by training in the step S4, and calculating the accuracy of the space-time convolution neural network model in classifying all the test samples;
s52, each case in the test set comprises a plurality of samples, wherein the proportion of the sample of the ith case classified as the minimum state of consciousness is piThe proportion of non-reactive wake syndrome is 1-piWith piAnd 1-piAs the classification probability of the ith case, a receiver operation characteristic curve is drawn with the non-reactive arousal syndrome as positive, based on the classification probability of each case of the test set, and the area under the curve is calculated.
To illustrate the effectiveness of the present invention's method of classifying disturbance of consciousness using spatio-temporal convolutional neural networks based on the resting electroencephalogram, an example is shown.
153 cases of resting electroencephalogram data were collected for the patients with disturbance of consciousness, among which 102 cases of minimal consciousness and 51 cases of unresponsive arousal were combined. By adopting 5-fold cross validation, 153 cases are divided into 5 groups, 1 group is selected as a test set each time, and the other 4 groups are selected as training sets. A total of 51450 samples are obtained by resampling 153 cases of electroencephalogram data. 5 models are finally obtained through 5 times of cross validation, each model is tested on a corresponding test set, 5 times of results are averaged, the accuracy rate of sample classification is 83.41%, and the area under the receiver operation characteristic curve of case classification is 0.90.
An article published in the book journal of Brain by Engemann et al in 2018 uses a feature extraction and machine learning method to achieve an area under the receiver operating characteristic curve of 0.78. We also validated long-short term memory networks with conventional convolutional neural networks with accuracy of 80.35% and 76.41% for sample classification, and 0.85 and 0.83 for areas under the receiver operating characteristic curve for case classification. Obviously, the space-time convolution neural network provided by the invention improves the classification performance, and has a continuously improved space along with the increase of the data volume.
According to the invention, classification of patients with disturbance of consciousness can be realized only by keeping the patients in a resting state and not depending on behavioral response of the patients; the invention does not depend on expert experience, and can give a diagnosis result only by an original electroencephalogram signal; the invention has high diagnosis accuracy rate for the patient with disturbance of consciousness, and has a space for continuously improving along with the increase of data volume.
The foregoing has outlined rather broadly the preferred embodiments and principles of the present invention and it will be appreciated that those skilled in the art may devise variations of the present invention that are within the spirit and scope of the appended claims.
Claims (8)
1. The consciousness disturbance classification method using the space-time convolution neural network based on the resting electroencephalogram is characterized by comprising the following steps of:
s1, acquiring electroencephalogram signals of the patient with disturbance of consciousness, and randomly dividing the acquired electroencephalogram signals into a training set and a testing set;
s2, filtering the electroencephalogram signals collected in the step S1, and resampling the training set and the test set in the step S1 through a sliding window method to obtain training samples and test samples;
s3, setting a threshold value, and removing training samples and test samples containing data exceeding the threshold value;
s4, constructing a space-time convolution neural network model, and training the space-time convolution neural network model by using the training samples obtained in the step S3;
s5, evaluating the performance of the space-time convolution neural network model obtained by training in the step S4 by using the test sample obtained in the step S3;
and S6, collecting electroencephalogram signals of the patient with disturbance of consciousness to be diagnosed as model input, and directly calculating the probability that the patient with disturbance of consciousness to be diagnosed is in the minimum consciousness state and the unresponsive arousal syndrome by utilizing the space-time convolutional neural network model trained in the step S4.
2. The method for classifying disturbance of consciousness using a spatiotemporal convolutional neural network based on the resting electroencephalogram of claim 1, wherein during the step S1, the sampling frequency of the electroencephalogram signal is 256Hz and the acquisition time lasts for 5 minutes or more.
3. The method for classifying disturbance of consciousness using spatio-temporal convolutional neural network based on the resting electroencephalogram of claim 1, wherein the filtering process in step S2 includes the steps of:
s21, performing notch filtering of 49Hz-51Hz, high-pass filtering of 0.5Hz and low-pass filtering of 40Hz on the electroencephalogram signals collected in the step S1 in sequence.
4. The method for classifying disturbance of consciousness using spatio-temporal convolutional neural network based on the resting electroencephalogram of claim 3, wherein the sliding window method in step S2 includes the steps of:
and S22, setting the length of the sliding window as l and the sampling step length as step, and performing sliding sampling on the data of the whole section of electroencephalogram signals collected through the sliding window.
5. The method for classifying disturbance of consciousness using spatio-temporal convolutional neural network based on resting electroencephalogram as claimed in claim 4, wherein the length of the sliding window is set to l-10 s and the sampling step is set to step-1 s.
6. The method for classifying disturbance of consciousness using spatio-temporal convolutional neural network based on the resting electroencephalogram of claim 1, wherein the threshold is set to 200 μ V in step S3.
7. The method for classifying disturbance of consciousness using spatio-temporal convolutional neural network based on resting electroencephalogram as claimed in claim 1, wherein in step S4, the process of constructing the spatio-temporal convolutional neural network model comprises the steps of:
s41, the input data dimension is 16 × 2560 × 1;
s42, extracting the time characteristics of the electroencephalogram signal by using a one-dimensional convolution kernel in the time dimension, wherein the output data dimension is 16 multiplied by 640 multiplied by 32;
s43, extracting the spatial features of the electroencephalogram signals by using a one-dimensional convolution kernel in the spatial dimension, wherein the output data dimension is 1 multiplied by 640 multiplied by 128;
s44, extracting time characteristics of the global brain space characteristics by using a one-dimensional convolution kernel in the time dimension, wherein the output data dimension is 1 multiplied by 20 multiplied by 256;
and S45, realizing classification through the full connection layer and outputting a classification result.
8. The method for classifying disturbance of consciousness using spatio-temporal convolutional neural network based on the resting electroencephalogram of claim 1, wherein the step S5 further comprises the steps of:
s51, classifying each test sample by using the space-time convolution neural network model obtained by training in the step S4, and calculating the accuracy of the space-time convolution neural network model in classifying all the test samples;
s52, each case in the test set comprises a plurality of samples, wherein the proportion of the sample of the ith case classified as the minimum state of consciousness is piThe proportion of non-reactive wake syndrome is 1-piWith piAnd 1-piAs the classification probability of the ith case, a receiver operation characteristic curve is drawn with the non-reactive arousal syndrome as positive, based on the classification probability of each case of the test set, and the area under the curve is calculated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110790906.8A CN113598791B (en) | 2021-07-13 | 2021-07-13 | Consciousness disturbance classification method based on time-space convolution neural network used by resting state electroencephalogram |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110790906.8A CN113598791B (en) | 2021-07-13 | 2021-07-13 | Consciousness disturbance classification method based on time-space convolution neural network used by resting state electroencephalogram |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113598791A true CN113598791A (en) | 2021-11-05 |
| CN113598791B CN113598791B (en) | 2024-04-02 |
Family
ID=78304555
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110790906.8A Active CN113598791B (en) | 2021-07-13 | 2021-07-13 | Consciousness disturbance classification method based on time-space convolution neural network used by resting state electroencephalogram |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113598791B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108209870A (en) * | 2017-12-25 | 2018-06-29 | 河海大学常州校区 | Long-term EEG monitoring automatic seizure detection method based on convolutional neural networks |
| CN108403111A (en) * | 2018-02-01 | 2018-08-17 | 华中科技大学 | A kind of epileptic electroencephalogram (eeg) recognition methods and system based on convolutional neural networks |
| CN109717866A (en) * | 2019-02-28 | 2019-05-07 | 温州大学 | A kind of disturbance of consciousness diagnostic method based on EEG signals |
| CN110269611A (en) * | 2019-07-31 | 2019-09-24 | 上海诺诚电气股份有限公司 | The monitoring of patient's disturbance of consciousness degree, early warning system and method |
| CN110584597A (en) * | 2019-07-15 | 2019-12-20 | 天津大学 | Multi-channel electroencephalogram signal monitoring method based on time-space convolutional neural network and application |
| CN111297380A (en) * | 2020-02-12 | 2020-06-19 | 电子科技大学 | Emotion recognition method based on space-time convolution core block |
| US20200260977A1 (en) * | 2018-09-14 | 2020-08-20 | Imedisync, Inc. | Method for diagnosing cognitive disorder, and computer program |
| KR20210034458A (en) * | 2019-09-20 | 2021-03-30 | 고려대학교 산학협력단 | Deep neural network based conscious state determination device and method |
-
2021
- 2021-07-13 CN CN202110790906.8A patent/CN113598791B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108209870A (en) * | 2017-12-25 | 2018-06-29 | 河海大学常州校区 | Long-term EEG monitoring automatic seizure detection method based on convolutional neural networks |
| CN108403111A (en) * | 2018-02-01 | 2018-08-17 | 华中科技大学 | A kind of epileptic electroencephalogram (eeg) recognition methods and system based on convolutional neural networks |
| US20200260977A1 (en) * | 2018-09-14 | 2020-08-20 | Imedisync, Inc. | Method for diagnosing cognitive disorder, and computer program |
| CN109717866A (en) * | 2019-02-28 | 2019-05-07 | 温州大学 | A kind of disturbance of consciousness diagnostic method based on EEG signals |
| CN110584597A (en) * | 2019-07-15 | 2019-12-20 | 天津大学 | Multi-channel electroencephalogram signal monitoring method based on time-space convolutional neural network and application |
| CN110269611A (en) * | 2019-07-31 | 2019-09-24 | 上海诺诚电气股份有限公司 | The monitoring of patient's disturbance of consciousness degree, early warning system and method |
| KR20210034458A (en) * | 2019-09-20 | 2021-03-30 | 고려대학교 산학협력단 | Deep neural network based conscious state determination device and method |
| CN111297380A (en) * | 2020-02-12 | 2020-06-19 | 电子科技大学 | Emotion recognition method based on space-time convolution core block |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113598791B (en) | 2024-04-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3803709B1 (en) | Detecting abnormalities in ecg signals | |
| CN108511055B (en) | Ventricular premature beat recognition system and method based on classifier fusion and diagnosis rules | |
| CN109009102B (en) | Electroencephalogram deep learning-based auxiliary diagnosis method and system | |
| CN111466876B (en) | Alzheimer's disease auxiliary diagnosis system based on fNIRS and graph neural network | |
| CN111166327A (en) | Epilepsy diagnosis device based on single-channel electroencephalogram signal and convolutional neural network | |
| KR102141185B1 (en) | A system of detecting epileptic seizure waveform based on coefficient in multi-frequency bands from electroencephalogram signals, using feature extraction method with probabilistic model and machine learning | |
| CN111839506B (en) | Mental load detection method and device | |
| CN113662560B (en) | Method for detecting seizure-like discharge between attacks, storage medium and device | |
| CN113017645B (en) | P300 signal detection method based on void convolutional neural network | |
| CN114190944A (en) | Robust emotion recognition method based on electroencephalogram signals | |
| CN115281685A (en) | Sleep stage identification method and device based on anomaly detection and computer readable storage medium | |
| CN113558640A (en) | Minimum consciousness state degree evaluation method based on electroencephalogram characteristics | |
| CN115067878A (en) | EEGNet-based resting state electroencephalogram consciousness disorder classification method and system | |
| CN112057068A (en) | Epilepsia pathological data classification method and device and storage medium | |
| CN115017960B (en) | Electroencephalogram signal classification method based on space-time combined MLP network and application | |
| CN113598791B (en) | Consciousness disturbance classification method based on time-space convolution neural network used by resting state electroencephalogram | |
| CN114532994A (en) | Automatic detection method for unsupervised electroencephalogram high-frequency oscillation signals based on convolution variational self-encoder | |
| CN112545535B (en) | Sleep-wake cycle analysis method based on amplitude integrated electroencephalogram | |
| CN116035594B (en) | Electroencephalogram artifact removing method based on segmentation-noise reduction network | |
| CN112650386B (en) | Brain-computer interface character output system based on PCA-PCN | |
| Afrah et al. | Automatic Detection of Mild Cognitive Impairment from EEG Recordings Using Discrete Wavelet Transform Leader and Ensemble Learning Methods | |
| Winnick et al. | Automated nystagmus detection: Accuracy of slow-phase and quick-phase algorithms to determine the presence of nystagmus | |
| Leal et al. | Epileptic Seizure Risk Assessment by Multi-Channel Imaging of the EEG | |
| Mondal et al. | A Comparative Study of Derivative and Fourier Magnitude Spectrum Based CNN Models for Automatic ECG Signal Quality Assessment | |
| CN115715680A (en) | Anxiety discrimination method and device based on connective tissue potential |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |