CN106901727A - A kind of depression Risk Screening device based on EEG signals - Google Patents

Abstract

The invention discloses a depression risk screening device based on EEG signals, which belongs to the field of depression risk screening; it includes a central processing device and a detection device for EEG signals; it also includes a universal EEG data acquisition system, an EEG Signal preprocessing system, EEG system for screening depression risk, EEG display module and result printing module; universal EEG data acquisition system includes Fp1 electrode sensor, FpZ electrode sensor, Fp2 electrode sensor, preamplifier, 50Hz trap oscilloscope and low-pass filter; it enables individual users to wear a detection device, which is convenient for users to self-test, automatically detects EEG noise, removes EEG artifacts, reduces interference between various noises, directly processes EEG data, jumps Through the sleep staging step, the calculation error generated in the sleep staging and other steps is reduced, the accuracy of depression risk assessment is improved, the amount of information collected is large, the data information of sleep EEG is fully utilized, and the selectivity is strong when selecting features .

Description

A risk screening device for depression based on EEG signals

technical field

The present invention relates to the field of depression risk screening devices, in particular to a depression risk screening device based on electroencephalogram signals.

Background technique

Depression is a common mental illness characterized by low mood and pessimism. When the symptoms are severe, suicidal behavior may occur. Based on the classification and analysis of the symptoms, background, diagnostic basis, tools and evaluation criteria of depression, this paper finds that not only the number of depression patients is large, but also the age of onset, scope and industry are gradually expanding, which brings great benefits to society and families. bring a heavy burden. However, there are relatively few personnel engaged in the diagnosis and evaluation of depression, and the diagnosis and evaluation rely on many subjective factors, which may easily lead to misdiagnosis and missed diagnosis. Therefore, there is an urgent need to improve its diagnostic accuracy and efficiency. Through the synthesis and analysis of the research status, analysis methods, means and existing achievements of depression, the present invention finds that an important reason for the inaccurate diagnosis and low efficiency of this disease is the lack of objective and quantitative biological diagnostic indicators and scientific and effective diagnostic models. and an accurate risk screening device. EEG activity is the overall macroscopic comprehensive effect of various electrical activities of a large number of brain cell groups. There is a close relationship between the degree and area of its activity and the functional state of the brain. Therefore, it also represents a state of the brain, which can reflect and characterize the disease and health of the brain; the current EEG-based risk screening and diagnosis of depression is mainly based on What is sleep staging technology. Regarding sleep stages, there have been several classification methods. It is currently widely used in 1968. According to the waveform characteristics of EEG, EOG, and EMG signals in different sleep stages, Rechtschaffen and Kales divided adult sleep into six stages: awake stage, non-rapid eye movement stage (further divided into 1, 2, 3 , stage 4), and rapid eye movement stage, which is the staging standard (as shown in Figure 9). Calculation of sleep latency, total sleep time, arousal index, sleep stage 1, sleep stage 2, sleep stage 3, sleep stage 4, REM percentage, REM sleep cycle number, REM sleep latency, REM sleep intensity , rapid eye movement sleep density and rapid eye movement sleep time and other characteristics, and bring them into the classifier and judge the risk of depression.

There are many defects in the existing devices and collection methods for screening depression risk, mainly in:

1. Traditional EEG technology is mostly used in relatively strict environments, such as clinical treatment in hospitals or laboratories in scientific research institutes, which is not convenient for individual users, and these application scenarios have ideal experimental conditions: On the one hand, these The environment has been specially physically isolated, with good electromagnetic shielding and sound shielding; on the other hand, there are professional and trained operators. Even under such strict restrictions, the measurement of EEG signals will be affected by some factors. Impact.

2. Sleep staging standards are different, and the accuracy of sleep staging is not high enough, which makes the depression risk screening system less accurate.

3. The amount of information is relatively small, and the characteristics are not very clear, and the corresponding characteristic information for the subjects cannot be obtained to give a rationalized depression risk result.

Contents of the invention

1. technical problem to be solved

In view of the problems existing in the prior art that it is inconvenient for individual users to use, interference factors interfere with detection, detection accuracy is low, and the amount of information collected by EEG is small, the purpose of the present invention is to provide a risk screening for depression based on EEG signals The device is convenient for individual users to use, reduces the interference between various noises, improves the detection accuracy, and collects a large amount of sleep EEG information.

2. Technical solution

A depression risk screening device based on electroencephalogram signals, including an electroencephalogram signal central processing device and a detection device; the electroencephalogram signal central processing device includes a preamplifier, a 50Hz notch filter, a low-pass filter, an AD Converter, EEG signal processor, impedance detector, DC rectifier and bluetooth 2.0 radio frequency transceiver; Described detection device comprises Fp1 electrode sensor, FpZ electrode sensor, Fp2 electrode sensor, reference electrode and switch; Described detection The device is connected to the EEG signal central processing device through wires; the wires include wire a, wire b and wire c; the wire c is connected to the switch; the EEG signal processing device also includes a universal EEG data acquisition system, EEG signal preprocessing system, depression risk screening EEG system, EEG display module and result printing module; universal EEG data acquisition system includes Fp1 electrode sensor, FpZ electrode sensor, Fp2 electrode Sensor, preamplifier, 50Hz notch filter and low-pass filter, Fp1 electrode sensor, FpZ electrode sensor and Fp2 electrode sensor are connected to the preamplifier; preamplifier, 50Hz notch filter and low-pass filter are connected in turn; all The EEG signal preprocessing system includes an AD converter, an EEG signal processor, an impedance detector, a DC rectifier and a Bluetooth 2.0 radio frequency transceiver, and the AD converter, an EEG signal processor and a Bluetooth 2.0 radio frequency transceiver are sequentially connected; the EEG signal processor is electrically connected to the impedance detector, and the impedance detector is connected to the detection device through a wire a; the EEG signal processor is electrically connected to the preamplifier through a DC rectifier The EEG system for screening depression risk includes an EEG display module one, a management module, an EEG signal secondary processing module, an EEG display module two and an EEG data screening system, and the EEG display module one , a management module, an EEG secondary processing module, an EEG display module two, and an EEG data screening system are connected in sequence, and the EEG data screening system includes a feature extraction module, a feature selection module, a classification model module and Depression risk screening module; individual users of this device can wear it directly, which is convenient for users to self-test. The EEG signal preprocessing system can automatically detect EEG noise, remove ocular artifacts, reduce the interference between various noises, and directly Process the EEG data, skip the sleep staging steps, reduce the calculation errors in the sleep staging and other steps, improve the accuracy of depression risk assessment, collect a large amount of information, and the EEG data screening system makes full use of The data information of the sleep EEG is selected for feature selection, and the selectivity is strong.

Preferably, the EEG system for screening the risk of depression also includes a data acquisition modeling system, an offline analysis module and a database module, the data acquisition modeling system produces reasonable physiological index parameters, and the offline analysis module obtains after preprocessing The EEG signal is processed twice, and the database module is used to store the basic information of the labeled population and the EEG data of the latest training.

Preferably, the EEG signal preprocessing system adopts an AR model and an adaptive predictor model method to automatically detect EEG noise through the AR model method, and remove electroocular artifacts through the adaptive predictor model.

Preferably, the data collection and modeling system uses a universal EEG collection system to collect EEG data from labeled healthy people and labeled depressed people, so as to collect reasonable physiological index parameters.

Preferably, the wire a includes three wires connected to the Fp1 electrode sensor, the FpZ electrode sensor and the Fp2 electrode sensor respectively; the wire b is composed of two wires; On the BIAS port and COM port of the central processing unit.

Preferably, the detection device is helmet-shaped, with a sound-insulating layer filled in the middle, the sound-insulated layer filled is made of soft sound-insulating material, an eye mask is provided at the front end of the detection device, and soft earplugs are provided at both sides; The switch is located at the lower part of the soft earplugs; this can ensure a quiet and comfortable environment for individual users anytime and anywhere, and the filling type can adapt to different groups of people for easy adjustment of the size, and the use of soft sound-proof materials can ensure comfort and reduce external noise. The influence of the electrode sensor; the Fp1 electrode sensor is arranged at the left forehead point, the Fp2 electrode sensor is arranged at the right forehead point, the FpZ electrode sensor is arranged at the forehead center point, and the reference electrode is arranged at the back of the brain; the Fp1 electrode sensor , FpZ electrode sensor, Fp2 electrode sensor and reference electrode electrode head all use medical paste type wet electrode head, which avoids the interference of electrode contact impedance.

Preferably, the feature extraction module uses nonlinear dynamics theory, the feature selection module and classification model module both use support vector machines, and the nonlinear dynamics theory analyzes the EEG signal to extract the nonlinearity of the EEG signal. Features, the support vector machine selects and classifies the features of the EEG data.

3. Beneficial effect

Compared with the prior art, the present invention has the advantages of:

(1) Individual users of this device can wear it directly, which is convenient for users to self-test. The EEG signal preprocessing system can automatically detect EEG noise, remove ocular artifacts, reduce the interference between various noises, and directly process EEG data , skip the sleep staging step, reduce the calculation error generated in the sleep staging and other steps, improve the accuracy of depression risk assessment, collect a large amount of information, and the EEG data screening system makes full use of the sleep EEG Data information is selected for feature selection, and the selectivity is strong. The DC power supply can achieve the predetermined test function through the impedance detector and the DC rectifier.

(2) The data acquisition and modeling system generates reasonable physiological index parameters. The offline analysis module performs secondary processing on the EEG signals obtained after preprocessing. The database module is used to store the basic information of the labeled population and the latest training time. EEG data.

(3) The EEG noise is automatically detected by the AR model method, and the oculograph artifact is removed by the adaptive predictor model.

(4) Collect EEG data from labeled healthy people and labeled depressed people through the universal EEG collection system, which facilitates the collection of reasonable physiological parameters.

(5) The filling type can be adapted to different groups of people for easy size adjustment. The use of soft sound-proof materials can ensure comfort and reduce the influence of the outside world on the electrode sensor. The medical stick-type wet electrode head avoids the interference of electrode contact impedance.

(6) The nonlinear dynamics theory analyzes the EEG signals, extracts the nonlinear features of the EEG signals, and supports vector machines to select and classify the EEG data features.

Description of drawings

Fig. 1 is a structural representation of the present invention;

Fig. 2 is a block diagram of the overall screening system of the present invention

Fig. 3 is a block diagram of the universalized EEG acquisition system of the present invention;

Fig. 4 is a block diagram of the oculograph artifact removal model in the adaptive prediction period of the present invention;

Fig. 5 is a block diagram of the EEG system for screening the risk of depression of the present invention;

Fig. 6 is a block diagram of the EEG data screening system of the present invention;

Fig. 7 is a block diagram of the data acquisition modeling system of the present invention;

Fig. 8 is a diagram of the electrode position of the human brain system;

Fig. 9 is a sleep staging structure diagram;

Fig. 10 is a top view of the detection device of the present invention.

Explanation of symbols in the figure: 1. EEG signal central processing device; 2. Detection device; 3. Preamplifier; 4. 50Hz notch filter; 5. Low-pass filter; 6. AD converter; 7. EEG signal Processor; 8. Impedance detector; 9. DC rectifier; 10. Bluetooth 2.0 radio frequency transceiver; 11. Fp1 electrode sensor; 12. FpZ electrode sensor; 13. Fp2 electrode sensor; 14. Reference electrode; 15. Wire; 16. Conductor a; 17. Conductor b; 18. Switch; 19. Filling the sound insulation layer; 20. Eye mask; 21. Soft earplugs; 22. Conductor c.

detailed description

The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention; obviously, the described embodiments are only some embodiments of the present invention; rather than all embodiments. Based on the embodiments of the present invention; all other embodiments obtained by persons of ordinary skill in the art without creative work; all belong to the protection scope of the present invention.

Example 1:

Please refer to Figure 1-10, a depression risk screening device based on EEG signals, including EEG signal central processing device 1 and detection device 2; EEG signal central processing device 1 includes preamplifier 3, 50Hz notch wave Device 4, low-pass filter 5, AD converter 6, EEG signal processor 7, impedance detector 8, DC rectifier 9 and Bluetooth 2.0 radio frequency transceiver 10; detection device 2 includes Fp1 electrode sensor 11, FpZ electrode sensor 12, Fp2 electrode sensor 13, reference electrode 14 and switch 18; Detection device 2 is connected in the EEG signal central processing device 1 by wire 15; Wire 15 includes wire a16, wire b17 and wire c22; Wire c22 is connected on the switch 18 The EEG signal processing device 1 also includes a universal EEG data acquisition system, an EEG signal preprocessing system, an EEG system for screening the risk of depression, an EEG display module and a result printing module; the universal EEG data acquisition The system includes Fp1 electrode sensor 11, FpZ electrode sensor 12, Fp2 electrode sensor 13, preamplifier 3, 50Hz notch filter 4 and low-pass filter 5, Fp1 electrode sensor 11, FpZ electrode sensor 12 and Fp2 electrode sensor 13 are connected with the front The preamplifier 3, the 50Hz notch filter and the low-pass filter are connected in turn; the EEG signal preprocessing system includes an AD converter 6, an EEG signal processor 7, an impedance detector 8, a DC correction Device 9 and Bluetooth 2.0 radio frequency transceiver 10, AD converter 6, EEG signal processor 7 and Bluetooth 2.0 radio frequency transceiver 10 are connected in turn; EEG signal processor 7 is electrically connected to impedance detector 8, impedance detector 8 Connect to the detection device 2 through the wire a16; the EEG signal processor 7 is electrically connected to the preamplifier 3 through the DC rectifier 9; the wire a16 includes three wires connected to the Fp1 electrode sensor 11, the FpZ electrode sensor 12 and the Fp2 respectively The electrode sensor 13; the wire b17 is composed of two wires; the two stages of the reference electrode 14 are connected to the BIAS port and the COM port of the EEG signal central processing device 1 through the wire b17.

The detection device 2 is helmet-shaped, with a sound-insulating layer 19 in the middle, and the sound-insulation layer 19 is made of soft sound-insulating material. The front end of the detection device 2 is provided with an eye mask 20, and soft earplugs 21 are provided on both sides; the switch 18 is in a soft position. The lower part of the earplug 21; this can ensure a quiet and comfortable environment for individual users anytime and anywhere, and the filling type can be used to adapt to different groups of people for easy adjustment of the size, and the use of soft sound-proof materials can ensure comfort and reduce the influence of the outside on the electrode sensor Fp1 electrode sensor 11 is arranged on the left forehead point, Fp2 electrode sensor 13 is arranged on the right forehead point, FpZ electrode sensor 12 is arranged on the forehead center point, and reference electrode 14 is arranged on the back of the brain; Fp1 electrode sensor 11, FpZ electrode sensor 12, Fp2 The electrode heads of the electrode sensor 13 and the reference electrode 14 are all made of medical paste type wet electrode heads, which avoids the interference of electrode contact impedance.

The EEG signal preprocessing system adopts the AR model and the adaptive predictor model method to automatically detect the EEG noise through the AR model method, and the detected EEG noise is sent to the adaptive predictor after initialization through the filter, and the adaptive predictor at the same time The predictor receives the EEG signal from the EEG area, and the adaptive predictor processes the EEG noise and the EEG signal in the EEG area at the same time, and removes the EEG artifact and EEG noise through the adaptive predictor model. The EEG signal preprocessing system can automatically detect EEG noise, remove oculoelectric artifacts, and reduce the interference between various noises. The EEG signal preprocessing system directly processes EEG data, skips the sleep staging step, and reduces sleep staging. The calculation errors generated by such steps improve the accuracy of depression risk assessment.

The EEG system for screening depression risk includes EEG display module 1, management module, EEG signal secondary processing module, EEG display module 2 and EEG data screening system, EEG display module 1, management module, brain The electrical signal secondary processing module, the EEG display module 2, and the EEG signal data screening system are connected in sequence. The EEG signal secondary processing module performs secondary processing on the EEG signals obtained after preprocessing to obtain the EEG power spectrum. The real-time dynamic display of the array diagram and the histogram of EEG θ, α, β and other indicators realizes real-time monitoring of EEG indicators. Doctors can observe the changes of various indicators of the patient's EEG information in real time through this function, and intuitively and accurately grasp the patient's brain function status , to monitor the effect of feedback treatment, the EEG data screening system includes a feature extraction module, a feature selection module, a classification model module and a depression risk screening module, the feature extraction module uses nonlinear dynamics theory, the Both the feature selection module and the classification model module use support vector machines, nonlinear dynamics theory to analyze EEG signals, extract nonlinear features of EEG signals, support vector machines to select and classify EEG data features, and collect information The EEG signal data screening system makes full use of the sleep EEG data information for feature selection, with strong selectivity.

The EEG system for screening depression risk also includes a data collection and modeling system, an offline analysis module and a database module. The management module manages the screening process and subject information, including creation, query, modification and deletion of subject information , as well as the selection of diagnostic screening programs, setting the indicators and forms required by the screening program and the storage of data, the database module stores the screening program, the basic information of the subjects and the screening process and results of the system, for The management system performs query and management, the data acquisition and modeling system generates reasonable physiological index parameters, the offline analysis module performs secondary processing on the EEG signals obtained after preprocessing, and the database module is used to store the basic information and EEG data during the latest training, the data acquisition modeling system through the universal EEG acquisition system, EEG data preprocessing system, EEG data secondary processing system, feature extraction module, feature selection module and classification model The module separately collects EEG data from labeled healthy people and labeled depressed people, which is convenient for collecting reasonable physiological index parameters. The new data model can be classified in the system according to different diseases, and its data structure is simple and clear. It has good data independence, security and confidentiality, and is easy to understand and use for users. It can be cross-compared with existing data to improve the accuracy and pertinence of screening and diagnosis. The dynamic EEG of the tester is to dynamically display the waveform of the EEG signal of the subject on the monitor in real time, and realize the dynamic EEG without a pen.

Working principle: The subject first registers his own information, including the subject's number, name, gender, age, etc., puts the detection device on the subject's head, turns on the switch, and the detection device according to the subject's The size of the head, the filling layer is automatically opened, the eye mask 20 and the soft earplugs 21 are automatically attached to the eyes and ears of the subject, so that the subject gradually falls asleep in a relatively quiet environment, and the real-time EEG recording is turned on. The patient's EEG signal is extracted through the Fp1 electrode sensor 11, the FpZ electrode sensor 12 and the Fp2 electrode sensor 13, the weak EEG signal is amplified by the preamplifier 3, and the original EEG signal is processed by power frequency filtering and amplified. The signal is converted into a digital signal through the 16-bit A/D converter 6, and the EEG signal is preprocessed by calling the EEG signal preprocessing system to reduce the interference of artifacts and ensure the correctness of the feedback treatment, and then through Bluetooth 2.0 transmitted to the computer and displayed on the screen in real time. The doctor invokes the corresponding signal processing program in the EEG analysis to perform secondary processing on the collected signal, uses nonlinear dynamics theory to analyze the EEG signal, extracts the nonlinear characteristics of the EEG signal, and obtains the eigenvalue of the feedback information. Bring the eigenvalues of the information into the support vector machine, carry out risk screening, and finally output the results. During the system training process, the system will automatically store the patient's EEG signals. , replay the EEG signal, analyze and print it, and monitor the treatment process of the patient.

The above; is only a preferred embodiment of the present invention; but the protection scope of the present invention is not limited thereto; any person familiar with the technical field is within the technical scope disclosed in the present invention; according to the technical solution of the present invention Equivalent replacements or changes thereof and their improved concepts should be covered within the protection scope of the present invention.

Claims (7)

1. a kind of depression Risk Screening device based on EEG signals, it is characterised in that：Including EEG signals center processing dress Put（1）And detection means（2）；Described EEG signals central processing unit（1）Including preamplifier（3）, 50Hz trappers （4）, low pass filter（5）, converter（6）, EEG Processing device（7）, impedance detector（8）, direct current rectifier（9）With The RF transceiver of bluetooth 2.0（10）；Described detection means（2）Including Fp1 electrode sensors（11）, FpZ electrode sensors （12）, Fp2 electrode sensors（13）, reference electrode（14）And switch（18）；Described detection means（2）By wire（15）Even It is connected to EEG signals central processing unit（1）In；Described wire（15）Including wire a（16）, wire b（17）With wire c （22）；Described wire c（22）It is connected to switch（18）In；Described EEG Processing device（1）Including generalization brain electricity Data collecting system, EEG signals pretreatment system, examination depression risk brain electric system, brain electricity display module and result printing Module；Described generalization brain electric data collecting system includes Fp1 electrode sensors（11）, FpZ electrode sensors（12）、Fp2 Electrode sensor（13）, preamplifier（3）, 50Hz trappers（4）And low pass filter（5）, Fp1 electrode sensors（11）、 FpZ electrode sensors（12）With Fp2 electrode sensors（13）With preamplifier（3）It is connected；Preamplifier（3）, 50Hz fall into Ripple device（4）And low pass filter（5）It is sequentially connected；The EEG signals pretreatment system includes converter（6）, EEG signals Processor（7）, impedance detector（8）, direct current rectifier（9）With the RF transceiver of bluetooth 2.0（10）, the converter（6）、 EEG Processing device（7）With the RF transceiver of bluetooth 2.0（10）It is sequentially connected；Described EEG Processing device（7）Electrically It is connected to impedance detector（8）, impedance detector（8）By wire a（16）It is connected to detection means（2）In；Described brain electricity Signal processor（7）By direct current rectifier（9）It is electrically connected to preamplifier（3）In；The examination depression risk brain Electric system includes brain electricity display module one, management module, EEG signals after-treatment module, brain electricity display module two and brain telecommunications Number screening system, the brain electricity display module one, management module, EEG signals after-treatment module, brain electricity display module 2nd, EEG signals data screening system is sequentially connected, and the EEG signals data screening system includes characteristic extracting module, feature Selecting module, disaggregated model module and depression Risk Screening module.

2. a kind of depression Risk Screening device based on EEG signals according to claim 1, it is characterised in that：It is described Examination depression risk brain electric system also includes data acquisition modeling, off-line analysis module and DBM.

3. a kind of depression Risk Screening device based on EEG signals according to claim 1, it is characterised in that：It is described EEG signals pretreatment system uses AR models and adaptive predictor model method.

4. a kind of depression Risk Screening device based on EEG signals according to claim 2, it is characterised in that：It is described Data acquisition modeling is by generalization eeg collection system healthy population respectively to label and the depression of label Crowd carries out brain electric data collecting.

5. a kind of depression Risk Screening device based on EEG signals according to claim 1, it is characterised in that：It is described Wire a（16）Fp1 electrode sensors are connected respectively to including three electric wires（11）, FpZ electrode sensors（12）With Fp2 electrodes Sensor（13）；Described wire b（17）It is made up of two root electric wires；Reference electrode（14）Two-stage pass through wire b（17）Even It is connected on EEG signals central processing unit（1）BIAS mouthfuls and com port on.

6. a kind of depression Risk Screening device based on EEG signals according to claim 1, it is characterised in that：It is described Detection means（2）It is helmet-like, centre is provided with sound insulation by filling layer（19）, described sound insulation by filling layer（19）It is soft sound insulation material Material is made, detection means（2）Front end is provided with eyeshade（20）, both sides are provided with soft earplug（31）；Described switch（18）Place In soft earplug（31）Bottom；The Fp1 electrode sensors（11）It is arranged on left front volume point, Fp2 electrode sensors（13）Set In right forehead point, FpZ electrode sensors（12）It is arranged on forehead center point, described reference electrode（14）It is arranged on back brain Position；Described Fp1 electrode sensors（11）, FpZ electrode sensors（12）, Fp2 electrode sensors（13）And reference electrode（14） Electrode tip uses medical mounted wet electrode head.

7. a kind of depression Risk Screening device based on EEG signals according to claim 1, it is characterised in that：It is described Characteristic extracting module uses nonlinear dynamics theory, the feature selection module and disaggregated model module to use supporting vector Machine.