Method, device and medium for improving individual depression using transcranial electrical stimulation

By acquiring and analyzing EEG data and behavioral performance, using deep learning models to generate personalized frequency points and stimulation points, and conducting transcranial electrical stimulation interventions for specific brain regions, solving the problem of poor personalized intervention in the existing technology. , achieving more efficient treatment of depression.

CN117919594BActive Publication Date: 2024-11-29SHENZHEN ZHONGKEHUAYI TECHNOLOGY CO LTD
View PDF 12 Cites 0 Cited by

Patent Information

Application Number
CN202311667723.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-12-05
Publication Date
2024-11-29
Estimated Expiration
2043-12-05

AI Technical Summary

Technical Problem

In the prior art, when using transcranial electrical stimulation to improve depression, personalized interventions are poor, spatial resolution is low, and traditional methods have problems with side effects and delayed onset.

Method used

By obtaining the EEG data and behavioral performance data of multiple channels, performing feature extraction and weighting calculations, combining behavioral performance differences, calling deep learning models for personalized learning, generating personalized frequency points and stimulation points, and controlling transcranial electrical stimulation. The device generates personalized electrical stimulation signals and intervenes in specific brain regions.

Benefits of technology

It improves the spatial resolution and personalized accuracy of transcranial electrical stimulation, ensures that each intervention can provide the most appropriate intervention, effectively improves depressive symptoms, and improves the treatment effect.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117919594B_ABST
    Figure CN117919594B_ABST
Patent Text Reader

Abstract

The present application provides a method and device for improving individual depression by using transcranial electrical stimulation, and relates to the field of neural stimulation technology. The method includes: obtaining EEG data and behavioral performance data under multiple channels during the target's depression-related training; extracting features from the EEG data under each channel to obtain the first EEG feature and the second EEG feature under each channel, and performing weighted calculation on the first EEG feature and the second EEG feature under each channel to obtain the weighted features of each channel; based on the behavioral performance data under each channel, capturing the differences between different behavioral performances of the target during the training process, and obtaining associated features by associating the captured behavioral performance differences with the weighted features of each channel; calling a personalized learning model to perform personalized learning on the associated features, thereby improving the target's depression. The present application solves the problem in the related art that the use of transcranial electrical stimulation to improve individual depression is not effective.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of neural stimulation technology, and more specifically, to a method, device and medium for improving individual depression using transcranial electrical stimulation. Background Art

[0002] Depression is one of the most common mental illnesses today. During an episode, there are significant emotional, cognitive, and physical symptoms, mainly manifested as low mood, anxiety, decreased pleasure, suicidal ideation and hopelessness, slow thinking, cognitive impairment, fatigue, appetite / weight changes, insomnia / somnia, psychomotor retardation / agitation. At present, the clinical intervention methods for depression include drug intervention, biophysical intervention, and psychobehavioral intervention. Among them, drug intervention has a small number of drug types, delayed onset, and the disadvantage of possibly aggravating depressive symptoms; among physiological and physical intervention methods, electroconvulsive therapy has large side effects and may cause serious complications; psychological and behavioral intervention has a stable effect, but a slow onset.

[0003] Transcranial Electrical Stimulation (TES) is a non-invasive neurostimulation technology that mainly uses electrodes to apply safe, low-intensity currents to specific brain areas to regulate cerebral cortical neural activity. Compared with the aforementioned centralized depression intervention methods, TES intervention methods are less likely to induce epilepsy, and the equipment is simple, low-cost, easy to operate, and has a better user experience.

[0004] Although a large number of studies have applied transcranial electrical stimulation, the effect of transcranial electrical stimulation on improving depression is controversial. The key to improving the intervention effect of transcranial electrical stimulation on depression lies in: implementing personalized intervention based on individual characteristics. Existing technologies can achieve personalized stimulation to a certain extent and improve training effects, but the personalized transcranial electrical stimulation used often has defects such as low spatial resolution and poor intervention effects. Summary of the invention

[0005] The present application provides a method, device, transcranial electrical stimulation device, electronic device and storage medium for improving individual depression by using transcranial electrical stimulation, which can solve the problem that the effect of using transcranial electrical stimulation to improve individual depression in the related art is not good. The technical solution is as follows:

[0006] According to one aspect of the present application, a method for improving individual depression using transcranial electrical stimulation includes: obtaining EEG data under multiple channels during a target's depression-related training, and behavioral performance data synchronized with the EEG data; each channel corresponds to a brain area of ​​the target acted on by a transcranial electrical stimulation device through different electrodes; the behavioral performance data is used to indicate the behavioral performance of the target during the training process; feature extraction is performed on the EEG data under each channel to obtain a first EEG feature and a second EEG feature under each channel, and weighted calculation is performed on the first EEG feature and the second EEG feature under each channel to obtain a weighted feature of each channel; the first EEG feature is a frequency domain expression of the EEG data in the frequency domain, and the second EEG feature is a frequency domain expression of the EEG data in the frequency domain. The electrical feature is the time-frequency domain expression of the EEG data in the time-frequency domain; based on the behavioral performance data under each channel, the differences between different behavioral performances of the target during the training process are captured, and the associated features are obtained by associating the captured behavioral performance differences with the weighted features of each channel; the personalized learning model is called to perform personalized learning on the associated features to obtain personalized frequency points and personalized stimulation points; the personalized learning model is a deep learning model that has been trained and has the ability to learn personalized frequency points and personalized stimulation points; the transcranial electrical stimulation device is controlled to generate an electrical stimulation signal at a personalized frequency point, and the electrical stimulation signal is applied to the corresponding brain area of ​​the target through the electrodes at the personalized stimulation points to improve the depression of the target.

[0007] According to one aspect of the present application, a device for improving individual depression using transcranial electrical stimulation includes: a data acquisition module for acquiring EEG data under multiple channels during a target's depression-related training, and behavioral performance data synchronized with the EEG data; each channel corresponds to a brain area of ​​the target acted on by a transcranial electrical stimulation device through different electrodes; the behavioral performance data is used to indicate the behavioral performance of the target during the training process; a data analysis module for performing feature extraction on the EEG data under each channel to obtain a first EEG feature and a second EEG feature under each channel, and performing weighted calculation on the first EEG feature and the second EEG feature under each channel to obtain a weighted feature of each channel; the first EEG feature is a frequency domain expression of the EEG data in the frequency domain, and the second EEG feature is The EEG data is expressed in the time-frequency domain; a behavior association module is used to capture the differences between different behavior performances of the target during the training process based on the behavior performance data under each channel, and obtain the association features by associating the captured behavior performance differences with the weighted features of each channel; a personalized learning module is used to call the personalized learning model to perform personalized learning on the association features to obtain personalized frequency points and personalized stimulation points; the personalized learning model is a deep learning model that has been trained and has the ability to learn personalized frequency points and personalized stimulation points; a control module is used to control the transcranial electrical stimulation device to generate an electrical stimulation signal at a personalized frequency point, and apply the electrical stimulation signal to the corresponding brain area of ​​the target through the electrodes at the personalized stimulation points to improve the depression of the target.

[0008] According to one aspect of the present application, a transcranial electrical stimulation device includes: a supporting portion; a plurality of integrated electrodes, which are arranged on the supporting portion and distributed in different brain regions of a target; a circuit, which is fixed on the supporting portion, and is used to collect EEG data of the target through the integrated electrodes while the target wears the transcranial electrical stimulation device to perform depression-related training, and in accordance with the method of using transcranial electrical stimulation to improve individual depression as described above, the electrical stimulation signal is applied to different brain regions of the target through each of the integrated electrodes to improve the target's depression.

[0009] According to one aspect of the present application, an electronic device includes at least one processor and at least one memory, wherein the memory stores computer-readable instructions; the computer-readable instructions are executed by one or more of the processors, so that the electronic device implements the method of improving individual depression using transcranial electrical stimulation as described above.

[0010] According to one aspect of the present application, a storage medium stores computer-readable instructions thereon, and the computer-readable instructions are executed by one or more processors to implement the method for improving depression in an individual using transcranial electrical stimulation as described above.

[0011] According to one aspect of the present application, a computer program product includes computer-readable instructions, which are stored in a storage medium. One or more processors of an electronic device read the computer-readable instructions from the storage medium, load and execute the computer-readable instructions, so that the electronic device implements the method of improving individual depression using transcranial electrical stimulation as described above.

[0012] The beneficial effects of the technical solution provided by this application are:

[0013] In the above technical scheme, while the target is wearing a transcranial electrical stimulation device for depression-related training, EEG data of the target under multiple channels and behavioral performance data synchronized with the EEG data are obtained, and first, feature extraction is performed on the EEG data under multiple channels to obtain a first EEG feature and a second EEG feature under each channel, and weighted calculation is performed on the first EEG feature and the second EEG feature under each channel to obtain a weighted feature of each channel, and then the behavioral performance data under each channel are combined to capture the differences between different behavioral performances of the target during the training process, and then the captured behavioral performance differences are associated with the weighted features of each channel to obtain associated features, and then based on the associated features, a personalized learning model is called to learn to obtain personalized frequency points and personalized stimulation points, and finally the transcranial electrical stimulation device is controlled to generate electrical stimulation signals at personalized frequencies, and the electrical stimulation signals are applied to the corresponding brain areas of the target through the electrodes at the personalized stimulation points, thereby achieving the purpose of improving the target's depression.

[0014] That is to say, in this intervention, not only the expression of EEG data in the frequency domain, time domain, and / or time-frequency domain under different channels is fully considered to improve the spatial resolution of the personalized transcranial electrical stimulation process, but also the normalized weights of the above-mentioned different expressions in the specified frequency band and the influence of the differences between different behavioral performances of the target during the training process on personalized learning are combined, thereby improving the accuracy of the learning process of personalized frequency points and personalized stimulation points. In the next intervention, the personalized frequency points and personalized stimulation points will change with the transcranial electrical stimulation of the target in the previous intervention, which will cause the transcranial electrical stimulation of the target in this intervention to change. Then, through the above-mentioned repeated cycles of multiple interventions, it can be ensured that the target can receive the most appropriate intervention in each training process, thereby effectively solving the problem that the related technology is not effective in improving individual depression using transcranial electrical stimulation.

[0015] In addition, the designated frequency band to which the personalized frequency points belong includes the alpha frequency band, which enables the transcranial electrical stimulation device to regulate the alpha oscillation in the brain through transcranial electrical stimulation in each intervention, thereby effectively improving multiple cognitive functions and emotional states such as memory, movement, creativity, etc. that are closely related to alpha oscillations, thereby achieving the purpose of effectively improving depression. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are only some embodiments of the present application, and those skilled in the art can obtain other drawings based on these drawings without creative work.

[0017] Figure 1 It is a schematic diagram of the implementation environment involved in this application;

[0018] Figure 1a is a schematic structural diagram of a transcranial electrical stimulation device according to an exemplary embodiment;

[0019] Figure 1b yes Figure 1 A structural block diagram of a circuit in a transcranial electrical stimulation device according to a corresponding embodiment;

[0020] Figure 2 is a flow chart showing a method for improving individual depression using transcranial electrical stimulation according to an exemplary embodiment;

[0021] Figure 2a yes Figure 2 A flowchart of step 330 in one embodiment of the corresponding embodiment;

[0022] Figure 2b yes Figure 2 A flowchart of step 350 in one embodiment of the corresponding embodiment;

[0023] Figure 3 is a schematic diagram of a network structure of a personalized learning model according to an exemplary embodiment;

[0024] Figure 4 is a flow chart showing another method for improving depression in an individual by using transcranial electrical stimulation according to an exemplary embodiment;

[0025] Figures 4a to 4d yes Figure 4 Schematic diagram of interfaces for different types of tasks involved in the corresponding embodiments;

[0026] Figure 5It is an interactive timing diagram of a method of using transcranial electrical stimulation to improve individual depression in an application scenario;

[0027] Figure 6 is a structural block diagram of a device for improving individual depression using transcranial electrical stimulation according to an exemplary embodiment;

[0028] Figure 7 is a hardware structure diagram of an electronic device according to an exemplary embodiment;

[0029] Figure 8 is a structural block diagram of an electronic device according to another exemplary embodiment. DETAILED DESCRIPTION

[0030] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and cannot be interpreted as limiting the present application.

[0031] It will be understood by those skilled in the art that, unless expressly stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present disclosure refers to the presence of the features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein includes all or any unit and all combinations of one or more associated listed items.

[0032] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0033] Figure 1 This is a schematic diagram of an implementation environment involved in a method for improving individual depression using transcranial electrical stimulation. It should be noted that this implementation environment is only an example adapted to the present application and cannot be considered as providing any limitation on the scope of use of the present application.

[0034] The implementation environment includes a transcranial electrical stimulation device 100 , a control device 200 , and a training device 300 .

[0035] The control device 200 may be an electronic device such as a desktop computer, a laptop computer, a server, or a computer cluster composed of multiple servers, or even a cloud computing center composed of multiple servers, which is not limited here.

[0036] The training device 300 can be run by a client, and the client can be used to provide depression-related training to the target. The client can be in the form of an application or a web page. Accordingly, the user interface provided by the training device 300 to the client can be in the form of a program window or a page, which is not limited here. The training device 300 can be a computer device such as a desktop computer, a laptop computer, a server, etc. that can be run by the client, or it can be a computer cluster composed of multiple servers, or even a cloud computing center composed of multiple servers, which is not limited here. For example, in this implementation environment, the training device 300 can be a laptop computer, which can be used to provide depression-related training to the target. The control device 200 establishes a communication connection with the training device 300 in advance through wired or wireless means, and realizes data transmission between the two through the communication connection. For example, the transmitted data includes but is not limited to: pushing corresponding training intervention parameters for each training of the target.

[0037] like Figure 1a As shown, the transcranial electrical stimulation device 100 includes but is not limited to: a support portion 101 , a plurality of integrated electrodes, and a circuit 102 .

[0038] Specifically, the support part 101 is made of a material that is light, strong, and elastic, which is not only ergonomic, comfortable to wear, but also suitable for different head circumferences of targets. Here, the target may include a person.

[0039] The integrated electrode, also called the integrated EEG stimulation electrode, can not only collect the EEG data of the target, but also apply the electrical stimulation signal to different brain regions of the target. In some embodiments, there are multiple integrated electrodes distributed in different brain regions of the target. Figure 1aIn the embodiment, there are six integrated electrodes, and the brain regions distributed therein correspond to the EEG10-20 system, specifically: electrode 103 corresponds to FP1, electrode 104 corresponds to FP2, electrode 105 corresponds to F3, electrode 106 corresponds to F4, electrode 107 corresponds to T8, and electrode 108 corresponds to T7. There are many ways for the electrical stimulation signals to act on the six target brain regions through these electrodes, including but not limited to: 1) pairing any two of the six electrodes to form a current loop, and outputting three groups of electrical stimulation signals to act on the six brain regions respectively; 2) respectively forming electrodes 103, electrode 105, and electrode 108 into one group, and electrodes 104, electrode 106, and electrode 107 into another group to form a current loop, outputting two groups of electrical stimulation signals, and using a 1-in 2-out or 2-in 1-out method to act on the six brain regions; 3) forming electrodes 103, electrode 105, electrode 108, electrode 104, electrode 106, and electrode 107 into one group to form a current loop to act on the six brain regions. In addition, each integrated electrode can be arranged at a certain rotation angle on the support portion 101, ensuring the target wearing tightness of different head shapes and the fit with the scalp, while reducing the contact impedance with the scalp and improving the wearing comfort.

[0040] The circuit 102 is fixed to the support part 101, and is used to collect EEG data of different channels of the target through different integrated electrodes when the target wears the transcranial electrical stimulation device to perform depression-related training, and to apply electrical stimulation signals to different brain regions of the target through each integrated electrode to improve the target's depression. Figure 1b As shown, the circuit 102 may include a control unit 201, a communication unit 202, a signal acquisition unit 203, a stimulation unit 204, and a power supply unit 205. The control unit 201 may be composed of an embedded microcontroller, and is used to control the normal operation of the transcranial electrical stimulation device 100, including but not limited to: 1) controlling the communication unit 202 to complete the communication between the various units and the communication with the control device 200; 2) controlling the signal acquisition unit 203 to complete the collection of EEG data; 3) controlling the stimulation unit 204 to output a transcranial electrical stimulation signal (hereinafter referred to as an electrical stimulation signal); 4) controlling the power supply unit 205 to output current / voltage to each unit to ensure the long-term and stable operation of the transcranial electrical stimulation device 100. The communication unit 202 may support communication methods such as Wifi and Bluetooth to realize data transmission between the various units and data transmission between the control device 200. For example, the transmitted data includes EEG data in the form of brain wave signals, electrical stimulation signals, etc.

[0041] It is worth mentioning that the circuit 102 can collect multi-channel EEG data with a collection bit width of 24 bits, and the sampling rate can be flexibly adjusted in the range of 250SPS-16KSPS. At the same time, the circuit 102 can also output various types of electrical stimulation signals. For example, according to the waveform classification, it can include sinusoidal wave signals, square wave signals, sawtooth wave signals, and trapezoidal wave signals; according to the current direction classification, it can include AC signals, DC signals, etc.; according to the noise characteristics classification, it can include noise signals, pure signals, etc. In some embodiments, the frequency of the electrical stimulation signal can be 0Hz-100Hz, and the adjustment step can be 1Hz; the stimulation intensity range can be ±2mA, and the adjustment step can be 0.01mA. In other embodiments, the frequency of the electrical stimulation signal belongs to the alpha band, that is, 8Hz-13Hz.

[0042] In addition, the circuit 102 also has functions such as real-time current monitoring and hardware overcurrent protection, which can not only ensure the effective and stable operation of the transcranial electrical stimulation device 100, but also ensure the safety of the transcranial electrical stimulation device 100.

[0043] Taking a target person as an example, in an application scenario, as the client providing training runs in the training device 300, the user can use the client to perform training related to depression. When the user performs training, first ensure that the transcranial electrical stimulation device 100 is worn correctly. The training device 300 can detect and notify the user of the contact status of each electrode in the transcranial electrical stimulation device 100 so that the user can perform training correctly.

[0044] When the user wears the transcranial electrical stimulation device 100 to perform depression-related training, the transcranial electrical stimulation device 100 can collect EEG data of multiple channels of the user. For example, six channels of EEG data can be collected through six integrated electrodes.

[0045] As the transcranial electrical stimulation device 100 and the control device 200 interact, the transcranial electrical stimulation device 100 uploads the EEG data of multiple channels of the user to the control device 200. Accordingly, the control device 200 can receive the EEG data of the multiple channels, and then obtain the personalized frequency and personalized stimulation points corresponding to the user's training based on the EEG data of the multiple channels, and then determine the electrical stimulation parameters based on the personalized frequency and personalized stimulation points, and feed the electrical stimulation parameters back to the transcranial electrical stimulation device 100.

[0046] At this time, the transcranial electrical stimulation device 100 can receive the electrical stimulation parameters issued by the control device 200, and finally generate an electrical stimulation signal at a personalized frequency according to the electrical stimulation parameters, and apply the electrical stimulation signal to the corresponding brain area of ​​the user through the electrodes at the personalized stimulation points, thereby achieving the purpose of effectively improving the user's depression.

[0047] Of course, according to actual operational needs, the control device 200 and the training device 300 can also be integrated into the same electronic device or the same cluster or center, so that the training of the target and the control of the transcranial electrical stimulation device can be completed by the same electronic device or the same cluster or center.

[0048] See also Figure 2 The present application embodiment provides a method for improving individual depression by using transcranial electrical stimulation. The method is applicable to an electronic device, which may be Figure 1 A control device 200 is shown in an implementation environment.

[0049] In the following method embodiments, for the convenience of description, the execution subject of each step of the method is taken as the control device as an example for explanation, but this does not constitute a specific limitation.

[0050] like Figure 2 As shown, the method may include the following steps:

[0051] Step 310, obtaining EEG data of the target under multiple channels during depression-related training, as well as behavioral performance data synchronized with the EEG data.

[0052] Among them, each channel corresponds to the brain area acted upon by the transcranial electrical stimulation device through different electrodes.

[0053] For the transcranial electrical stimulation device, when the target wears it for training, it can collect EEG data of the corresponding channels in real time (for example, the sampling rate is 500Hz) through different electrodes, and continuously upload the collected EEG data of multiple channels to the control device. Figure 1a In the figure, electrode 103 corresponds to FP1, and electrode 104 corresponds to FP2. Accordingly, the EEG data of one channel can be collected through electrode 103, and the EEG data of another channel can be collected through electrode 104.

[0054] Then, the control device can continuously receive the EEG data of each channel. It can be understood that considering the hardware resources and processing capabilities of the control device, the EEG data of each channel that is continuously received will first be cut into multiple segments of EEG data before subsequent processing, so that each segment of EEG data can be processed separately subsequently, thereby improving the processing efficiency of the control device. At the same time, it can also better measure the changes in the target before and after transcranial electrical stimulation.

[0055] Taking the EEG data of a certain channel as an example, the duration of a segment of EEG data is 2500ms. That is to say, at 0 o'clock, the EEG data collected by the transcranial electrical stimulation device through a certain electrode from 1000ms before 0 o'clock to 1500ms after 0 o'clock is regarded as a segment of EEG data of a certain channel.

[0056] In some embodiments, before extracting features from the EEG data of each channel, the EEG data of each channel of the target is first preprocessed. The preprocessing includes, but is not limited to, interference elimination processing. The interference elimination processing includes, but is not limited to, removing baseline drift, removing noise, removing artifacts, and the like. For example, noise may include electromyography, electrooculography, power frequency interference, and the like. In some embodiments, the interference elimination processing may include low-pass filtering, high-pass filtering, notch filtering, and the like.

[0057] For example, first, the interference of the power frequency and its harmonics is removed through a 50Hz and 100Hz 5th-order notch filter, then the baseline drift is removed through a 2nd-order 40Hz Butterworth low-pass filter, and then the low-frequency noise is removed through a 3rd-order 0.5Hz Butterworth high-pass filter, and finally the preprocessed EEG data of each channel is obtained.

[0058] In this way, only the pre-processed EEG data of each channel can be used for feature extraction, which not only avoids feature extraction of invalid EEG data and wastes hardware resources, but also fully guarantees the accuracy of feature extraction, which is conducive to more accurate and unique identification of EEG data.

[0059] Secondly, it is explained that the behavior performance data is used to indicate the behavior performance of the target during the training process. In some embodiments, the behavior performance includes one or more of the operation behavior and the reaction behavior. Accordingly, the behavior performance data is used to indicate the operation behavior and / or reaction behavior of the target during the training process. For example, the operation behavior can refer to the key operation of the target during the training process, and the reaction behavior can refer to the reaction speed, reaction time, etc. corresponding to the key operation of the target during the training process. Based on this, the behavior performance data is essentially to indicate the behavior performance of the target during the training process through the key operation and its corresponding reaction speed and reaction time. For example, if the key operation is wrong, it can indicate that the behavior performance of the target during the training process is poor or invalid, or if the reaction speed is too slow, it can also indicate that the behavior performance of the target during the cognitive training process is poor or invalid. In other embodiments, the behavior performance data includes a behavior performance label, so the behavior performance data is essentially to indicate the behavior performance of the target during the training process through the behavior performance label. For example, the behavior performance label is 0, indicating that the behavior performance of the target during the cognitive training process is poor or invalid, and the behavior performance label is 1, indicating that the behavior performance of the target during the cognitive training process is good or effective. Of course, in other embodiments, the behavior performance label is not limited to being represented by numbers, and can also be represented by strings, etc., which does not constitute a specific limitation here.

[0060] It should be noted that the synchronization of behavioral performance data and EEG data refers to the alignment of the timestamp of the EEG data and the timestamp of the behavioral performance data. Only when the timestamps of the two are aligned can the accuracy of the association between subsequent behavioral performance and EEG features be guaranteed.

[0061] Step 330, extracting features from the EEG data of each channel to obtain the first EEG feature and the second EEG feature of each channel, and performing weighted calculation on the first EEG feature and the second EEG feature of each channel to obtain the weighted feature of each channel.

[0062] First of all, it should be noted that EEG features are used to accurately and uniquely identify EEG data. It is understandable that different channels and different EEG data will have different corresponding EEG features. In some embodiments, the EEG features include a first EEG feature and a second EEG feature, the first EEG feature is a frequency domain expression of the EEG data in the frequency domain, and the second EEG feature is a time-frequency domain expression of the EEG data in the time-frequency domain.

[0063] In some embodiments, the method of extracting the first EEG feature includes but is not limited to Fourier transform, for example, using fast Fourier transform to convert EEG data into multiple frequency domain features in the frequency domain as the first EEG feature.

[0064] In some embodiments, the second EEG feature is extracted by, but not limited to, wavelet transform, empirical mode decomposition algorithm, etc. For example, continuous wavelet transform is used to obtain multiple time-frequency features of EEG data in the time-frequency domain as the second EEG feature.

[0065] Furthermore, by analyzing the EEG data of the target in each channel during the training process, the inventors found that there were obvious differences in the power values ​​of the target before and after transcranial electrical stimulation during the training process. Based on this, in some embodiments, time-frequency features are used to describe the time-frequency power values ​​of EEG data at different frequency points and / or time points.

[0066] Taking continuous wavelet transform as an example, the feature extraction process may include the following steps: using Morlet wavelet as the mother wavelet basis function, the EEG data of a certain channel is divided into M parts; based on the sampling rate and duration of the EEG data, the M parts of the EEG data are processed by continuous wavelet transform to obtain time-frequency features with a dimension of M×N. Among them, M represents the number of frequency components of the EEG data in the time-frequency domain, and N represents the number of time components of the EEG data in the time-frequency domain. In this way, it is conducive to more accurately measure the changes in power values ​​before and after transcranial electrical stimulation, and then better measure the changes in targets before and after transcranial electrical stimulation.

[0067] For example, Morlet wavelet is used as the mother wavelet basis function, the support area is [-5,5], the center frequency is 6, and the frequency scale range is [4,13]. At this time, M = 32. If the sampling rate of the EEG data of a certain channel is 500hz, and the duration of each EEG data segment is 2500ms, it is equivalent to that each EEG data segment includes 500×2.5=1250 frequency points in the time-frequency domain. Then, based on the 32 EEG data divided from a certain segment of EEG data, after continuous wavelet transform processing of the 32 EEG data, the time-frequency features of 32×1250 dimensions, namely P, can be obtained. At this time, each row represents the frequency component f of the EEG data in the time-frequency domain, and each column represents the time component t of the EEG data in the time-frequency domain. Correspondingly, P(t,f) represents the time-frequency power value corresponding to the frequency point at the time-frequency point (t,f). It can also be understood that the time-frequency feature P(t,f) is used to describe the time-frequency power value corresponding to the frequency point at the time-frequency point (t,f) in the time-frequency domain.

[0068] Secondly, it is explained that the weighted calculation of the first EEG feature and the second EEG feature is achieved by normalizing the weight of each frequency point in the specified frequency band.

[0069] Among them, the normalized weight of each frequency point in the specified frequency band is used to indicate the contribution of each frequency point to the learning process of personalized frequency points and personalized stimulation points belonging to the specified frequency band. It can be understood that the larger the normalized weight, the greater the contribution of the frequency point to the learning process of personalized frequency points and personalized stimulation points. After determining the normalized weight of each frequency point in the specified frequency band, the first EEG feature and the second EEG feature under each channel can be weighted respectively, and then the weighted first EEG feature and the weighted second EEG feature under each channel are used to obtain the weighted features of each channel.

[0070] In this way, by assigning different weights to the EEG features corresponding to different frequency points, the frequency components within the specified frequency band can be highlighted, which is conducive to the subsequent accurate learning of personalized frequency points and personalized stimulation points belonging to the specified frequency band.

[0071] Step 350 , based on the behavioral performance data of each channel, the differences between different behavioral performances of the target during the training process are captured, and the associated features are obtained by associating the captured behavioral performance differences with the weighted features of each channel.

[0072] After obtaining the behavioral performance data under each channel and the weighted features of each channel, it is possible to capture the differences between the different behavioral performances of the target during the training process, associate the target's behavioral performance with the EEG, and form associated features. It should be noted that since the behavioral performance data under each channel actually reflects the behavioral performance of the target at different frequencies of the EEG data synchronized with it, the weighted features of each channel reflect the EEG features of the target at different frequencies. Based on this, the associated features are used to describe the relationship between the behavioral performance and EEG features of the target at different frequencies. In some embodiments, the associated features include at least: EEG features that are associated with good behavioral performance and have high activity in a specified frequency band, and EEG features that are associated with poor behavioral performance and have low activity in a specified frequency band.

[0073] Compared with traditional personalized transcranial electrical stimulation, by correlating behavioral performance with EEG, it effectively avoids directly discarding EEG data with poor behavioral performance, and can make full use of the useful information contained in the EEG data to ensure the reliability of subsequent personalized frequency and personalized stimulation point learning.

[0074] Step 370: Call the personalized learning model to perform personalized learning on the associated features to obtain personalized frequency points and personalized stimulation points.

[0075] In order to implement personalized intervention based on individual characteristics, the inventor proposes to select a frequency point corresponding to the target's current training from each frequency point of the EEG data in the time-frequency domain for each channel collected by different electrodes, and this frequency point can best reflect the target's operation behavior and / or reaction behavior during the training process. In this embodiment, the personalized frequency point refers to the frequency point that can best reflect the target's operation behavior and / or reaction behavior during the training process, and the personalized stimulation point corresponds to the electrode where the EEG data at the frequency point is collected.

[0076] The inventors have found that alpha oscillations are closely related to multiple cognitive functions and emotional states such as memory, movement, and creativity. Therefore, if alpha oscillations in the brain can be regulated by transcranial electrical stimulation during intervention, the above-mentioned cognitive functions and emotional states closely related to alpha oscillations can be effectively improved. Therefore, in some embodiments, the personalized frequency point belongs to the alpha frequency band, wherein the alpha frequency band refers to the 8Hz-13Hz frequency band.

[0077] In this embodiment, the personalized frequency points and personalized stimulation points are learned based on a personalized learning model, and the personalized learning model is a deep learning model that has been trained and has the ability to learn personalized frequency points and personalized stimulation points.

[0078] Step 390, control the transcranial electrical stimulation device to generate an electrical stimulation signal at a personalized frequency, and apply the electrical stimulation signal to the corresponding brain area of ​​the target through the electrodes at the personalized stimulation points to improve the target's depression.

[0079] For the control device, after obtaining the personalized frequency and personalized stimulation point, the corresponding electrical stimulation parameters can be determined and sent to the transcranial electrical stimulation device. The electrical stimulation parameters include but are not limited to: intensity, frequency, duration, point, etc.

[0080] As for the transcranial electrical stimulation device, after receiving the electrical stimulation parameters sent by the control device, it can generate an electrical stimulation signal with the intensity in the electrical stimulation parameters at a personalized frequency point (i.e. the frequency in the electrical stimulation parameters) according to the intensity, frequency, etc. in the electrical stimulation parameters, and then use the electrodes at the personalized stimulation points (i.e. the points) to apply the electrical stimulation signal to the corresponding brain area of ​​the target according to the duration in the electrical stimulation parameters, thereby achieving the purpose of improving the target depression.

[0081] It should be understood that the control device can establish a communication connection with different transcranial electrical stimulation devices, that is, the control device can be used for a large number of users. Then, in the process of different users wearing transcranial electrical stimulation devices and facing different training devices for training, the control device can receive the EEG data of each channel of each user, and find personalized frequency points and personalized stimulation points for different users based on these EEG data. Since the personalized frequency points reflect the user's operating behavior and / or reaction behavior during the training process, different users have different operating behaviors and / or reaction behaviors in the same training process. It is impossible, so that the personalized frequency points and personalized stimulation points found by the control device will be different. At this time, for different users, the transcranial electrical stimulation at different personalized stimulation points based on the electrical stimulation signals at different personalized frequencies is bound to be different, thereby achieving the effect of personalized transcranial electrical stimulation to improve depression.

[0082] Through the above process, personalized alpha transcranial electrical stimulation is achieved to improve the emotional state and cognitive function of depression, that is, based on the personalized frequency and personalized stimulation points in this intervention, the transcranial electrical stimulation device is controlled to perform transcranial electrical stimulation on the target, which not only fully considers the expression of EEG data in the frequency domain, time domain, and / or time-frequency domain under different channels to improve the spatial resolution of the personalized transcranial electrical stimulation process, but also combines the normalized weights of the above different expressions in the specified frequency band and the influence of the differences between different behavioral performances of the target during the training process on personalized learning, thereby improving the accuracy of the learning process of personalized frequency and personalized stimulation points, and achieving the purpose of improving the target depression this time. In the next intervention, the personalized frequency and personalized stimulation points will change with the transcranial electrical stimulation of the target in the previous intervention, so that the transcranial electrical stimulation of the target in this intervention will also change. Then, through the above repeated interventions, it can be ensured that the target can receive the most appropriate intervention in each training process, thereby effectively solving the problem that the related technology uses transcranial electrical stimulation to improve individual depression.

[0083] In addition, the designated frequency band to which the personalized frequency points belong includes the alpha frequency band, which enables the transcranial electrical stimulation device to regulate the alpha oscillation in the brain through transcranial electrical stimulation in each intervention, thereby effectively improving multiple cognitive functions and emotional states such as memory, movement, creativity, etc. that are closely related to alpha oscillations, thereby achieving the purpose of effectively improving depression.

[0084] See also Figure 2a In an exemplary embodiment, step 330 may include the following steps:

[0085] Step 331, traverse the EEG data of each channel, perform Fourier transform on the EEG data traversed under the current channel, and obtain multiple frequency domain features of the EEG data in the frequency domain as the first EEG feature corresponding to the EEG data under the current channel.

[0086] Among them, each frequency domain feature corresponds to each frequency point in the frequency domain one by one.

[0087] That is, the EEG data is converted into the frequency domain by Fourier transform to obtain the frequency domain features of the EEG data at different frequency points in the frequency domain. That is, the first EEG feature includes a plurality of frequency domain features corresponding to each frequency point in the frequency domain.

[0088] Step 333, performing continuous wavelet transform on the EEG data to obtain multiple time-frequency features of the EEG data in the time-frequency domain as the second EEG feature corresponding to the EEG data in the current channel.

[0089] Among them, each time-frequency feature is used to describe the time-frequency power value of the EEG data at different frequency points and / or time points.

[0090] That is to say, the second EEG feature includes multiple time-frequency features corresponding to each frequency point and each time point in the time-frequency domain. For example, the second EEG feature may include multiple time-frequency features at different frequency points and the same time point, multiple time-frequency features at the same frequency point and different time points, or multiple time-frequency features at different frequency points and different time points.

[0091] Step 335, determine the normalized weight of each frequency point in the specified frequency band, and based on the determined normalized weight, perform weighted calculations on each frequency domain feature in the first EEG feature corresponding to the EEG data under the current channel and each time-frequency feature in the second EEG feature corresponding to the EEG data under the current channel.

[0092] Among them, the process of determining the normalized weight may include the following steps: determining the center frequency point of the specified frequency band; calculating the frequency difference between each frequency point and the center frequency point respectively; calculating the corresponding sensitivity weight of each frequency point in the specified frequency band based on the calculated frequency difference and the distribution coefficient of each frequency point in the specified frequency band; normalizing the corresponding sensitivity weight of each frequency point in the specified frequency band to obtain the corresponding normalized weight of each frequency point in the specified frequency band.

[0093] Specifically, based on the EEG data traversed under the current channel, the calculation formula of the sensitivity weight alpha_weights[f] of the frequency point F in the specified frequency band is as follows:

[0094]

[0095] Among them, f represents the frequency value of frequency point F; center_frequency represents the frequency value of the center frequency point of the specified frequency band; spread represents the distribution coefficient of each frequency point in the specified frequency band, which is used to control the sensitivity weight of each frequency point in the specified frequency band. It is worth mentioning that spread is related to all EEG data collected under each channel during the target training process, and will change accordingly with different training times.

[0096] For the EEG data traversed under the current channel, in order to ensure that the sum of the sensitivity weights of each frequency point in the specified frequency band is 1, these sensitivity weights need to be normalized before weighted calculation, as follows:

[0097]

[0098] Wherein, Norm_alpha_weights[f] represents the normalized weight of the frequency point F in the specified frequency band after normalization.

[0099] After obtaining the normalized weights of each frequency point in the specified frequency band, the first EEG feature and the second EEG feature corresponding to the EEG data in the current channel can be fused based on these normalized weights, and finally the weighted features of the EEG data traversed by the current channel at each frequency point can be obtained.

[0100] Step 337, the weighted frequency domain features and the weighted time-frequency features under the current channel are fused to obtain the weighted features of the EEG data traversed under the current channel at each frequency point.

[0101] In this embodiment, the fusion processing specifically refers to splicing the weighted frequency domain features and the weighted time domain features at the same frequency point to obtain the weighted features at the frequency point.

[0102] Specifically, the calculation formula for the weighted features of the EEG data traversed by the current channel at the frequency point F is as follows:

[0103] Alpha_F_spectrum[f]=F_spectrum[f]*Norm_alpha_weights[f];

[0104] Alpha_F_cwt[f][t]=F_cwt[f][t]*Norm_alpha_weights[f];

[0105] Fused_Features=[Alpha_F_spectrum[f]; Alpha_F_cwt[f][t]].

[0106] Where f represents the frequency value of frequency point F; Norm_alpha_weights[f] represents the normalized weight of frequency point F in the specified frequency band. F_spectrum[f] represents the frequency domain feature of frequency point F in the first EEG feature; F_cwt[f][t] represents the time-frequency feature of frequency point F at time point T in the second EEG feature. Alpha_F_spectrum[f] represents the weighted frequency domain feature of frequency point F; Alpha_F_cwt[f][t] represents the weighted time-frequency feature of frequency point F at time point T.

[0107] Of course, in other embodiments, fusion is not limited to the splicing method, and may also be processed by addition, multiplication and other operations, which does not constitute a specific limitation here.

[0108] After completing the feature fusion for each frequency point of all EEG data of each channel, the weighted features of all EEG data of each channel at each frequency point can be obtained.

[0109] Under the effect of the above-mentioned embodiment, by assigning different weights to the EEG features corresponding to different frequency points, it is possible to highlight the frequency components within the specified frequency band, which is conducive to the subsequent accurate learning of the personalized frequency points and personalized stimulation points belonging to the specified frequency band.

[0110] See also Figure 2b As shown, in an exemplary embodiment, step 350 may include the following steps:

[0111] Step 351 , based on the behavioral performance of the target during the training process indicated by the behavioral performance data under each channel, determine the behavioral performance of the EEG data under each channel at each frequency point.

[0112] It can be understood that, as far as each behavior performance data under each channel is concerned, the EEG data synchronized with the behavior performance data is different, so the behavior performance of the EEG data at the frequency point may also be different. For example, the behavior performance of EEG data A synchronized with behavior performance data A' at frequency point F is poor, the behavior performance of EEG data B synchronized with behavior performance data B' at frequency point F is good, and the behavior performance of EEG data C synchronized with behavior performance data C' at frequency point F is good.

[0113] In other words, based on the behavioral performance data under each channel, the behavioral performance of the EEG data under each channel at each frequency point will be different.

[0114] Step 353, according to the behavioral performance of the EEG data at each frequency point under each channel, the behavioral difference characteristics of each frequency point under each channel under different behavioral performances are calculated.

[0115] In some embodiments, the behavioral difference characteristics of each frequency point under each channel under different behavioral performances can be expressed as the difference in the normalized weights of the frequency point under different behavioral performances, specifically referring to: based on the EEG data collected during the target cognitive training process under a certain channel and its synchronized behavioral performance data, each frequency point of the collected EEG data in the frequency domain is traversed, and for the traversed frequency point, the behavioral performance of the frequency point in each behavioral performance data is determined; based on at least one behavioral performance data of the frequency point with good behavioral performance, the frequency point with good behavioral performance is determined in the frequency domain of the at least one behavioral performance data; According to the normalized weights under each synchronized EEG data, it is regarded as the normalized weight of the frequency point when the behavioral performance is good, and based on at least one behavioral performance data of the frequency point with poor behavioral performance, the normalized weight of the frequency point under each EEG data synchronized with the at least one behavioral performance data is determined, which is regarded as the normalized weight of the frequency point when the behavioral performance is poor; the normalized weights of the frequency point with good behavioral performance and poor behavioral performance are respectively averaged to obtain a first average value and a second average value, and finally the difference between the first average value and the second average value is calculated to obtain the behavioral difference characteristics at the frequency point.

[0116] Step 355, based on the weighted features of the EEG data at each frequency point under each channel and the behavioral difference features of each frequency point under each channel under different behavioral performances, the correlation features of the EEG data at each frequency point under each channel are calculated.

[0117] For example, the calculation formula for the related feature Behavior_related_features[f] of the EEG data traversed by the current channel at the frequency point F is as follows:

[0118] Difference_feature[f]=Mean(Norm_alpha_weights[f,Good_Trials])

[0119] -Mean(Norm_alpha_weights[f,Poor_Trials]);

[0120]

[0121] Among them, f represents the frequency value of frequency point F; Norm_alpha_weights[f,Good_Trials] represents the normalized weight of frequency point F in good behavior performance; Good_Trials represents that the behavior performance data indicates that the target has good behavior performance during training; Norm_alpha_weights[f,Poor_Trials] represents the normalized weight of frequency point F in poor behavior performance; Poor_Trials represents that the behavior performance data indicates that the target has poor behavior performance during training; Mean represents the mean; β represents the hyperparameter used to balance the importance ratio between good behavior performance and poor behavior performance. Difference_feature[f] represents the behavioral difference characteristics of frequency point F under different behavioral performances; Fused_Features[f] represents the weighted features on frequency point F.

[0122] Under the effect of the above-mentioned embodiments, the association between behavioral performance and EEG is achieved. Compared with traditional personalized transcranial electrical stimulation, it can effectively avoid directly discarding EEG data with poor behavioral performance, and make full use of the useful information contained in the EEG data, so as to ensure the reliability of subsequent personalized frequency point search.

[0123] Figure 3 A schematic diagram of the network structure of a personalized learning model in one embodiment is shown. Figure 3 In the embodiment of the present invention, the personalized learning model includes a convolutional neural network, a long short-term memory network (LSTM), and a classifier.

[0124] Among them, the convolutional neural network includes a convolution layer 772, which includes operations such as using a 3*3 convolution kernel for convolution operation, using the ReLu function as an activation function, and performing a 2*2 maximum pooling operation to reduce the feature dimension; the long short-term memory network includes a unidirectional LSTM layer 773 for capturing the time dependency in the sequence, with 64 nodes; the classifier includes a Flatten layer 774, a fully connected layer 775, and a Softmax output layer 776. Specifically, the Flatten layer 774 flattens the outputs of the convolution layer 772 and the LSTM layer 773 so as to be connected to the fully connected layer 775; the fully connected layer 775 has 128 nodes and the activation function is ReLu; the Softmax output layer 776 is essentially a Dense layer with a softmax activation function added, and its output size is the size of the alpha frequency band of each brain region, indicating the probability distribution of each brain region and each frequency point on the alpha frequency band, which is convenient for accurate learning of personalized frequency points and personalized stimulation points. It should be noted that the number of nodes and convolution kernels in each layer can be adjusted according to the actual data and tasks to optimize the performance of the model.

[0125] It is worth mentioning that, as mentioned above, the personalized learning model is essentially a deep learning model that has been trained and has the ability to learn personalized frequency points and personalized stimulation points. Therefore, in some embodiments, the personalized learning model uses the Adam optimizer to implement model training, that is, to minimize the model's loss function to optimize the model's parameters. Specifically, the cross entropy loss function is used to calculate the model loss, and the model learning rate is set to 0.001. When the model loss is minimized or the number of model iterations reaches the threshold, the model training is considered to be completed, and a personalized learning model with the ability to accurately learn personalized frequency points and personalized stimulation points can be obtained. In addition, in some embodiments, according to the actual needs of the application scenario, model training also takes into account the need to extract and transmit information at various levels of different network structures, and the network structure here does not constitute a specific limitation.

[0126] Now combined Figure 3 The network structure of the personalized learning model shown in the figure explains the learning process of personalized frequency points and personalized stimulation points as follows:

[0127] In the first step, the associated features 771 are input into the personalized learning model, and the convolutional neural network and long short-term memory network 773 in the personalized learning model are used to extract the spatiotemporal domain features.

[0128] In the second step, the classifier in the personalized learning model is used to predict the probability distribution of spatiotemporal domain features in each brain region and each frequency point in the specified frequency band, and personalized frequency points and personalized stimulation points 777 are obtained.

[0129] In the above process, the learning process of personalized frequency and personalized stimulation points based on the personalized learning model is realized. In each intervention, the personalized frequency and personalized stimulation points can change with the changes in the target's EEG data and / or behavioral performance data during the training process, thereby making the transcranial electrical stimulation of the target by the transcranial electrical stimulation device change in each intervention. Then, through the above-mentioned repeated cycles of multiple interventions, it can be ensured that the target can receive the most appropriate intervention in each training process, fully demonstrating the intervention effect of personalized alpha transcranial electrical stimulation in improving depression, thereby effectively solving the problem that related technologies use transcranial electrical stimulation to achieve poor improvement effects on individual depression.

[0130] See also Figure 4 In an exemplary embodiment, after step 370, the method may further include the following steps:

[0131] Step 510, determining a behavior performance evaluation index of the target in this training based on the operation behavior and / or reaction behavior of the target in this training process indicated by the behavior performance data.

[0132] In some embodiments, the behavior performance evaluation index includes at least one of key pressing accuracy, reaction time, and reaction speed.

[0133] Step 530: Pushing training intervention parameters corresponding to the determined behavior performance evaluation index for the next training of the target.

[0134] In some embodiments, the training intervention parameters include but are not limited to: the task type and task difficulty of the next training, etc. The task type includes but is not limited to emotional state, working memory, attention, executive control, short-term memory, etc.

[0135] For example, the emotional state intervention task uses the emotional Go / No Go task, which requires users to respond to specific emotional stimuli while inhibiting responses to other stimuli. It is one of the classic tasks for examining the cognitive control function of patients with depression. The task materials include faces with positive, negative, and neutral expressions. At the beginning of the task, the user is told the target face of this task (positive / negative / neutral). During the task, one emotional face is presented to the user each time, and the face is presented for 300ms. If the presented face is the target face, the user needs to press the designated button, the faster the better; if the presented face is a non-target face, the user does not need to press the button. The interval between the two stimulus faces is 800 to 1200ms.

[0136] Working memory training is divided into verbal working memory training and spatial working memory training. Each task has five levels of difficulty, namely N=1, N=2, N=3, N=4, and N=5. In the verbal working memory training task, considering the differences in cultural levels of different users, numbers are used as memory materials. A total of 10 Roman numerals from 0 to 9 will appear in the center of the screen in sequence. The order of the numbers is random. The user's task is to judge each of the following numbers based on the difficulty, starting from the N+1th number, to determine whether it is the same as the Nth number before it. If they are the same, quickly press the "same" key, and if they are different, quickly press the "different" key (see Figure 4a ). In the spatial working memory training, the screen will randomly present a blue square at 9 positions. The user's task is to judge each subsequent square, starting from the N+1th square, to determine whether it is in the same position as the Nth square before it, and if so, quickly press the "same" key; if not, quickly press the "different" key (see Figure 4b ).

[0137] The attention training task mainly provides indicators related to sustained attention. During the training process, a gray square will be randomly presented in the upper or lower half of the screen (see Figure 4c), only when the gray square appears in the upper half of the screen, the user should press the designated key as quickly as possible, and when the gray square appears in the lower half of the screen, the user does not press the key. In order to test sustained attention, the gray square appears in the upper half of the screen at a ratio of 1 / 4, and in the lower half of the screen at a ratio of 3 / 4.

[0138] During the executive control training, the screen will randomly display one of the four words "red, green, blue, and yellow", but the color of the font is different from the meaning of the word (see Figure 4d ), users need to ignore the interference of font color and press the keys corresponding to "red, green, blue, and yellow" according to the meaning of the words. The faster the key press, the better.

[0139] During the short-term memory training, the user will hear a string of numbers, and the number of numbers increases from 3 to 9 according to the difficulty from low to high. The task includes two parts: forward repetition and reverse repetition. In the forward repetition, the training device will first play a digital voice, and after the voice ends, the number buttons 0-9 will appear on the screen. The user must press the number buttons on the screen in the order in which the numbers appear; in the reverse task, the training device will play a digital voice, and after the voice ends, the number buttons 0-9 will appear on the screen. The user must remember the order in which the numbers appear and press the number buttons on the screen in reverse order.

[0140] Based on the above-mentioned tasks, during the training process, the behavioral performance data such as whether the user presses the button, the reaction speed of the button, and the reaction time of the button can be obtained, and then the behavioral performance evaluation indicators such as the button accuracy, reaction speed, and reaction time can be obtained, so as to determine the training results of this training. It can be understood that the training results of this training are an important factor affecting which training intervention parameters are recommended to the user for the next training. For example, if the button accuracy in the training results of this training is lower than the first set threshold, it indicates that the task difficulty of this training is too high. At this time, the user can be recommended to reduce the task difficulty; or, if the reaction time in the training results of this training is lower than the second set threshold, it indicates that the task difficulty of this training is low. At this time, the user can be recommended to increase the task difficulty. Of course, in other embodiments, it is not limited to the training results of this training. The task type and task difficulty in the historical training process completed by the user, and the training results of the completed historical training can all be used as the basis for pushing the training intervention parameters corresponding to the next training to the user, which does not constitute a specific limitation here.

[0141] Under the effect of the above-mentioned embodiments, the push of training intervention parameters for different trainings is realized, which can better help users obtain better intervention effects for depression.

[0142] Figure 5This is a schematic diagram of a method for improving individual depression using transcranial electrical stimulation in an application scenario. In this application scenario, the transcranial electrical stimulation device is a wearable headband, and the user wears the wearable headband on the head before training. The training device and the control device are deployed on the same software end, for example, the software end can be an electronic device such as a desktop computer, a laptop computer or a server.

[0143] Users select the training they are interested in through the software running on the software terminal, and enter the training process after a simple novice guidance. Specifically: users can choose to register as new users or log in as old users. The information users need to provide during registration includes but is not limited to age, gender, education level, underlying diseases, etc.; after logging in, users can choose "Start Training" or "View Historical Training". After clicking "Start Training", users enter the training page, which provides tasks including but not limited to emotional state, working memory, attention, executive control and short-term memory. Users can choose tasks related to cognitive functions or emotional states that they want to intervene in for training; when users perform a task for the first time, they first enter a novice tutorial to guide users on how to complete the task correctly. When users meet the requirements for task completion in the novice tutorial, formal tasks will be opened to users.

[0144] The wearable headband will send the EEG data of the user during training to the software in real time. The software will align the EEG data with the behavioral performance data of the user during training and perform data preprocessing. For example, data preprocessing includes interference elimination processing, etc. At this time, the software will combine the user's personal information (such as age, gender), training intervention parameters (task types and task difficulty such as emotional state tasks / memory tasks / attention tasks / control tasks), behavioral performance data during training (whether the button is correct, button reaction time, button reaction speed) and EEG data during training. Through feature extraction, feature fusion and personalized learning, the most suitable recommended parameters for the user's next training are calculated (including training intervention parameters such as task type and task difficulty, as well as electrical stimulation parameters such as point, intensity, frequency, and duration). Accordingly, the user can choose to intervene according to the recommendations of the software, or choose to perform customized intervention.

[0145] Then, in the next intervention, the software will perform the above operations again and update the recommended parameters based on the user's latest training situation, repeating the cycle to ensure that the user receives the most appropriate intervention during each training process.

[0146] For example, in terms of motor learning, 10Hz-tACS at an intensity of 1mA for only 10 minutes can significantly improve the performance of visual motor learning tasks. In terms of language processing, applying 10Hz-tACS at an intensity of 1mA to the prefrontal lobe for 20 minutes can increase the task-related neural activity in the stimulated area to the optimal level of speech performance, thereby significantly promoting the speed of speech response and significantly increasing the brain's theta power in speech decision-making (the increase in theta power is positively correlated with the promotion of behavior). For creativity, a high-order cognitive function, alpha oscillations are related to creative thinking, and alpha activity in the frontal lobe brain area selectively participates in the generation of creativity. By applying 10Hz-tACS to the frontal lobe, the creativity index of an individual can be effectively improved. This shows that even high-order cognitive functions can be regulated by non-invasive brain stimulation. Memory is one of the core functions of cognition, and the alpha band plays an important role in it. The alpha band is related to memory retention and anti-interference ability. The degree of alpha phase coupling within the frontal-parietal and frontal-temporal pathways can predict individual working memory performance. The stronger the phase synchronization of the fusiform gyrus (FG) → inferior frontal gyrus (IFG), the stronger the individual's behavioral WM ability. In addition, individuals will automatically adjust the alpha frequency to reduce distracting signals in the environment that interfere with memory. Applying a 12Hz electrical stimulation signal to the prefrontal lobe can effectively promote sleep spindle activity and improve motor memory performance.

[0147] In this application scenario, the wearable headband has the advantages of small size, low cost, and easy use. It can dynamically adjust the intervention method based on individual characteristics, allowing users to achieve personalized intervention at home. In addition, the software can conduct a comprehensive analysis based on a series of multi-dimensional information such as the user's personal information, the type and difficulty of the task selected during the training process, behavioral performance and EEG performance, and provide users with the most suitable electrical stimulation parameters and training intervention parameters to help users obtain better intervention effects for depression.

[0148] The following is an embodiment of the device of the present application, which can be used to implement the method of improving individual depression by using transcranial electrical stimulation involved in the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the method of improving individual depression by using transcranial electrical stimulation involved in the present application.

[0149] See also Figure 6 In an embodiment of the present application, a device 900 for improving individual depression using transcranial electrical stimulation is provided, including but not limited to: a data acquisition module 910, a data analysis module 930, a behavior association module 950, a personalized learning module 970 and a control module 990.

[0150] Among them, the data acquisition module 910 is used to obtain EEG data of multiple channels of the target when the target wears the transcranial electrical stimulation device for depression-related training; each channel corresponds to the brain area acted upon by the transcranial electrical stimulation device through different electrodes.

[0151] The data analysis module 930 is used to extract features from the EEG data of each channel to obtain a first EEG feature and a second EEG feature, and based on the normalized weights of the first EEG feature and the second EEG feature in the specified frequency band, perform feature fusion on the first EEG feature and the second EEG feature to obtain weighted features of each channel; the first EEG feature is the frequency domain expression of the EEG data in the frequency domain, and the second EEG feature is the time-frequency domain expression of the EEG data in the time-frequency domain.

[0152] The behavior association module 950 is used to use the behavior performance data corresponding to each channel and the EEG data, as well as the weighted features of each channel, to capture the differences between different behavior performances of the target during the training process and obtain association features; the behavior performance data is used to indicate the behavior performance of the target during the training process.

[0153] The personalized learning module 970 is used to call the personalized learning model, perform personalized learning on the associated features, and obtain personalized frequency points and personalized stimulation points; the personalized learning model is a deep learning model that has been trained and has the ability to learn personalized frequency points and personalized stimulation points.

[0154] The control module 990 is used to control the transcranial electrical stimulation device to generate electrical stimulation signals at personalized frequencies, and to apply the electrical stimulation signals to the corresponding brain regions of the target through electrodes at personalized stimulation points to improve the target's depression.

[0155] It should be noted that the device for improving individual depression through transcranial electrical stimulation provided in the above-mentioned embodiment only uses the division of the above-mentioned functional modules as an example to illustrate the improvement of depression. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device for improving individual depression through transcranial electrical stimulation will be divided into different functional modules to complete all or part of the functions described above.

[0156] In addition, the device for improving individual depression using transcranial electrical stimulation provided in the above-mentioned embodiments and the method for improving individual depression using transcranial electrical stimulation belong to the same concept, and the specific manner in which each module performs operations has been described in detail in the method embodiments and will not be repeated here.

[0157] Figure 7 A schematic diagram of the structure of an electronic device according to an exemplary embodiment. The electronic device is suitable for Figure 1A control device 200 is shown in an implementation environment.

[0158] It should be noted that the electronic device is only an example adapted to the present application and cannot be considered to provide any limitation on the scope of use of the present application. The electronic device cannot be interpreted as requiring dependence on or having Figure 7 One or more components of exemplary electronic device 2000 are shown.

[0159] The hardware structure of the electronic device 2000 may vary greatly due to differences in configuration or performance, such as Figure 7 As shown, the electronic device 2000 includes: a power supply 210 , an interface 230 , at least one memory 250 , and at least one central processing unit (CPU) 270 .

[0160] Specifically, the power supply 210 is used to provide operating voltage for each hardware device on the electronic device 2000 .

[0161] The interface 230 includes at least one wired or wireless network interface 231 for interacting with external devices. Figure 1 The interaction between the transcranial electrical stimulation device 100 and the control device 200 in an implementation environment is shown.

[0162] Of course, in other examples adapted by the present application, the interface 230 may further include at least one serial-to-parallel conversion interface 233, at least one input-output interface 235, and at least one USB interface 237, etc. Figure 7 As shown, this is not intended to be a specific limitation.

[0163] The memory 250 is a carrier for storing resources, which may be a read-only memory, a random access memory, a disk or an optical disk, etc. The resources stored thereon include an operating system 251, an application 253 and data 255, etc. The storage method may be temporary storage or permanent storage.

[0164] Among them, the operating system 251 is used to manage and control various hardware devices and application programs 253 on the electronic device 2000 to enable the central processor 270 to calculate and process the massive data 255 in the memory 250. It can be WindowsServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, etc.

[0165] The application program 253 is a computer-readable instruction that performs at least one specific task based on the operating system 251, and may include at least one module ( Figure 7(not shown), each module may respectively contain computer-readable instructions for the electronic device 2000. For example, a device for improving depression in an individual using transcranial electrical stimulation may be considered as an application 253 deployed in the electronic device 2000.

[0166] The data 255 may be photos, pictures, etc. stored in a disk, or may be EEG data, etc. stored in the memory 250 .

[0167] The central processor 270 may include one or more processors, and is configured to communicate with the memory 250 through at least one communication bus to read the computer-readable instructions stored in the memory 250, thereby realizing the operation and processing of the mass data 255 in the memory 250. For example, the method of improving individual depression by using transcranial electrical stimulation is completed by the central processor 270 reading a series of computer-readable instructions stored in the memory 250.

[0168] In addition, the present application can also be implemented through hardware circuits or hardware circuits combined with software. Therefore, the implementation of the present application is not limited to any specific hardware circuits, software, or a combination of the two.

[0169] See also Figure 8 In an embodiment of the present application, an electronic device 4000 is provided, and the electronic device 400 may include: a desktop computer, a laptop computer, a server, etc.

[0170] exist Figure 8 In the embodiment, the electronic device 4000 includes at least one processor 4001 and at least one memory 4003.

[0171] The data interaction between the processor 4001 and the memory 4003 can be realized through at least one communication bus 4002. The communication bus 4002 may include a path for transmitting data between the processor 4001 and the memory 4003. The communication bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The communication bus 4002 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 8 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0172] Optionally, the electronic device 4000 may further include a transceiver 4004, which may be used for data interaction between the electronic device and other electronic devices, such as data transmission and / or data reception. It should be noted that in actual applications, the transceiver 4004 is not limited to one, and the structure of the electronic device 4000 does not constitute a limitation on the embodiments of the present application.

[0173] Processor 4001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. Processor 4001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0174] The memory 4003 can be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program instructions or codes in the form of instructions or data structures and can be accessed by the electronic device 400, but is not limited to this.

[0175] Computer-readable instructions are stored in the memory 4003 , and the processor 4001 can read the computer-readable instructions stored in the memory 4003 through the communication bus 4002 .

[0176] The computer-readable instructions are executed by one or more processors 4001 to implement the method of improving individual depression using transcranial electrical stimulation in the above embodiments.

[0177] In addition, a storage medium is provided in an embodiment of the present application, on which computer-readable instructions are stored, and the computer-readable instructions are executed by one or more processors to implement the above method of improving individual depression using transcranial electrical stimulation.

[0178] A computer program product is provided in an embodiment of the present application. The computer program product includes computer-readable instructions. The computer-readable instructions are stored in a storage medium. One or more processors of an electronic device read the computer-readable instructions from the storage medium, load and execute the computer-readable instructions, so that the electronic device implements the above method of improving individual depression using transcranial electrical stimulation.

[0179] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0180] The above are only some implementation methods of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A transcranial electrical stimulation device, It is characterized in that include: Supporting part; A plurality of integrated electrodes are arranged on the support portion and distributed in different brain regions of the target; A circuit, fixed to the support part, is used to collect EEG data of the target through the integrated electrodes when the target wears the transcranial electrical stimulation device to perform depression-related training, and to apply electrical stimulation signals to different brain regions of the target through each of the integrated electrodes according to the method of using transcranial electrical stimulation to improve individual depression, so as to improve the depression of the target; The method of improving individual depression using transcranial electrical stimulation comprises: Acquire the EEG data of the target under multiple channels during depression-related training, and behavioral performance data synchronized with the EEG data; each channel corresponds to a brain region of the target acted on by the transcranial electrical stimulation device through different electrodes; the behavioral performance data is used to indicate the behavioral performance of the target during the training process; Performing feature extraction on the EEG data under each channel to obtain a first EEG feature and a second EEG feature under each channel, and performing weighted calculation on the first EEG feature and the second EEG feature under each channel to obtain a weighted feature of each channel; the first EEG feature is a frequency domain expression of the EEG data in the frequency domain, and the second EEG feature is a time-frequency domain expression of the EEG data in the time-frequency domain; Based on the behavioral performance data under each channel, the differences between different behavioral performances of the target during the training process are captured, and the associated features are obtained by associating the captured behavioral performance differences with the weighted features of each channel; the associated features at least include EEG features associated with the behavioral performance differences; Calling a personalized learning model to perform personalized learning on the associated features to obtain personalized frequency points and personalized stimulation points; the personalized learning model is a deep learning model that has been trained and has the ability to learn personalized frequency points and personalized stimulation points; The transcranial electrical stimulation device is controlled to generate the electrical stimulation signal at a personalized frequency, and the electrical stimulation signal is applied to the corresponding brain area of ​​the target through the electrodes at the personalized stimulation points to improve the depression of the target.

2. The device according to claim 1, It is characterized in that The feature extraction of the EEG data under each channel is performed to obtain the first EEG feature and the second EEG feature under each channel, and the weighted calculation of the first EEG feature and the second EEG feature under each channel is performed to obtain the weighted feature of each channel, including: Traversing the EEG data under each channel, performing Fourier transform on the EEG data traversed under the current channel, and obtaining a plurality of frequency domain features of the EEG data in the frequency domain as the first EEG feature corresponding to the EEG data under the current channel; each of the frequency domain features corresponds to each frequency point in the frequency domain; Performing continuous wavelet transform on the EEG data to obtain multiple time-frequency features of the EEG data in the time-frequency domain as the second EEG features corresponding to the EEG data under the current channel; each of the time-frequency features is used to describe the time-frequency power value of the EEG data at different frequency points and / or time points; According to the difference between each frequency point in the specified frequency band and the specified reference frequency point, the normalized weight of each frequency point in the specified frequency band is determined, and based on the determined normalized weight, each frequency domain feature in the first EEG feature corresponding to the EEG data under the current channel and each time-frequency feature in the second EEG feature corresponding to the EEG data under the current channel are weighted and calculated respectively; The weighted frequency domain features and the weighted time-frequency features under the current channel are fused to obtain the weighted features of the EEG data traversed under the current channel at each frequency point.

3. The device as claimed in claim 2, It is characterized in that The determining, according to the difference between each frequency point in the specified frequency band and the specified reference frequency point, the normalized weight of each frequency point in the specified frequency band includes: Determine the center frequency of the specified frequency band; Calculate the frequency difference between each frequency point and the central frequency point respectively; According to the calculated frequency difference and the distribution coefficient of each frequency point in the specified frequency band, the corresponding sensitivity weight of each frequency point in the specified frequency band is calculated; The sensitivity weight corresponding to each frequency point in the specified frequency band is normalized to obtain the normalized weight corresponding to each frequency point in the specified frequency band.

4. The device according to claim 1, It is characterized in that The method captures the differences between different behavioral performances of the target during the training process based on the behavioral performance data under each channel, and obtains associated features by associating the captured behavioral performance differences with the weighted features of each channel, including: Based on the behavioral performance of the target during the training process indicated by the behavioral performance data under each channel, frequency domain analysis is performed on the EEG data under each channel to determine the behavioral performance of the EEG data under each channel at each frequency point; According to the behavioral performance of the EEG data at each frequency point under each channel, the behavioral difference characteristics of each frequency point under each channel under different behavioral performances are calculated; According to the weighted features of the EEG data at each frequency point under each channel and the behavioral difference features of each frequency point under each channel under different behavioral performances, the correlation features of the EEG data at each frequency point under each channel are calculated.

5. The device according to claim 1, It is characterized in that The calling of the personalized learning model to perform personalized learning on the associated features to obtain personalized frequency points and personalized stimulation points includes: Inputting the correlation features of the EEG data at each frequency point under each channel into the personalized learning model, and extracting the spatiotemporal features by using the convolutional neural network and the long short-term memory network in the personalized learning model; The classifier in the personalized learning model is used to predict the probability distribution of each frequency point of the spatiotemporal domain feature in the specified frequency band to obtain the personalized frequency point, and the probability distribution of each brain region of the spatiotemporal domain feature in the specified frequency band is predicted to obtain the personalized stimulation point.

6. The device according to claim 1, It is characterized in that The behavioral performance includes one or more of operational behavior and reactive behavior; After obtaining the EEG data of the target under multiple channels during depression-related training and the behavioral performance data synchronized with the EEG data, the method further includes: Determine a behavior performance evaluation index of the target in this training based on the operation behavior and / or reaction behavior of the target in this training process indicated by the behavior performance data; Pushing training intervention parameters corresponding to the determined behavior performance evaluation index for the next training of the target.

7. The device according to claim 2, 3 or 5, It is characterized in that The designated frequency band includes an alpha frequency band; the frequency range of the alpha frequency band is 8 Hz-13 Hz.

8. The device according to any one of claims 1 to 6, It is characterized in that The number of the integrated electrodes is six, and the six integrated electrodes include: a first electrode corresponding to FP1, a second electrode corresponding to FP2, a third electrode corresponding to F3, a fourth electrode corresponding to F4, a fifth electrode corresponding to T8, and a sixth electrode corresponding to T7.

9. A storage medium having computer-readable instructions stored thereon, It is characterized in that The computer readable instructions are executed by one or more processors to implement a method for improving depression in an individual using transcranial electrical stimulation; The method of improving individual depression using transcranial electrical stimulation comprises: Acquire EEG data of a target under multiple channels during depression-related training, and behavioral performance data synchronized with the EEG data; each channel corresponds to a brain region of the target acted on by a transcranial electrical stimulation device through different electrodes; the behavioral performance data is used to indicate the behavioral performance of the target during the training process; Performing feature extraction on the EEG data under each channel to obtain a first EEG feature and a second EEG feature under each channel, and performing weighted calculation on the first EEG feature and the second EEG feature under each channel to obtain a weighted feature of each channel; the first EEG feature is a frequency domain expression of the EEG data in the frequency domain, and the second EEG feature is a time-frequency domain expression of the EEG data in the time-frequency domain; Based on the behavioral performance data under each channel, the differences between different behavioral performances of the target during the training process are captured, and the associated features are obtained by associating the captured behavioral performance differences with the weighted features of each channel; the associated features at least include EEG features associated with the behavioral performance differences; Calling a personalized learning model to perform personalized learning on the associated features to obtain personalized frequency points and personalized stimulation points; the personalized learning model is a deep learning model that has been trained and has the ability to learn personalized frequency points and personalized stimulation points; The transcranial electrical stimulation device is controlled to generate an electrical stimulation signal at a personalized frequency, and the electrical stimulation signal is applied to the corresponding brain area of ​​the target through the electrodes at the personalized stimulation points to improve the depression of the target.

Citation Information

Patent Citations

  • A portable multi-channel depression tendency evaluation system based on emotional stimulation task

    CN109157231A

  • Non-invasive closed-loop transcranial electrostimulation device

    CN109173047A

  • Progressive brain network determination method and device for transcranial electrical stimulation

    CN114464323A

  • Electroencephalogram attention recognition system and method based on feature selection

    CN114521903A

  • Personalized transcranial electrical stimulation emotion intervention method and device

    CN115487420A