Interventional method, device, electronic device and storage medium for deep brain stimulation

Abstract

The application provides an intervention method, device, electronic equipment and storage medium for deep brain stimulation, wherein the intervention method comprises: acquiring an electroencephalogram signal of a target user, determining an electroencephalogram feature set of the electroencephalogram signal, determining intervention-related information for the target user based on the electroencephalogram feature set, and performing intervention processing according to the intervention-related information. In this way, when the DBS intervention is performed, the intervention-related information suitable for the patient's disease condition can be automatically selected according to the electroencephalogram feature of the patient, so that the effect of precise intervention treatment is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of deep brain stimulation, in particular to a deep brain stimulation intervention method and device, an electronic device and a storage medium. BACKGROUND

[0002] Deep brain stimulation (DBS) is mainly to implant electrodes into the brain of a patient, use a pulse generator to stimulate certain neural nuclei in the deep brain, correct abnormal brain electrical circuits, and thus reduce neurological symptoms.

[0003] The existing DBS intervention is usually discharge intervention at a fixed intensity, which needs to rely on medical personnel to observe the patient's disease condition, then adjust the intensity and frequency of discharge according to experience and patient symptoms, and use this fixed intervention intensity in the next period of treatment until the next time the medical personnel adjusts.

[0004] However, when medical personnel determines the intervention intensity according to experience and patient symptoms, the determined intervention intensity is not accurate enough, and the intervention intensity of the patient in a period of time will not change with the change of the disease condition. SUMMARY

[0005] Therefore, the purpose of the present application is to provide a deep brain stimulation intervention method, device, electronic device and storage medium, which can automatically select intervention related information suitable for the patient's disease condition according to the brain wave characteristics of the patient during DBS intervention, so as to achieve the effect of precise intervention treatment.

[0006] In a first aspect, the embodiments of the present application provide a deep brain stimulation intervention method, which comprises:

[0007] Obtaining a brain electrical signal of a target user;

[0008] Determining a brain wave characteristic set of the brain electrical signal;

[0009] Based on the brain wave characteristic set, determining intervention related information for the target user;

[0010] Performing intervention processing according to the intervention related information.

[0011] In the embodiments of the present application, as an optional embodiment, the intervention related information comprises:

[0012] A disease type;

[0013] An intervention intensity for the disease type.

[0014] In the embodiments of the present application, as an optional embodiment, the step of determining the intervention-related information for the target user based on the brain wave feature set comprises:

[0015] inputting the brain wave feature set into a plurality of preset first disease identification models respectively to obtain a plurality of disease onset probabilities output by the first disease identification models respectively;

[0016] if the disease onset probability output by any one of the first disease identification models is greater than a preset first disease onset probability threshold for the any one of the first disease identification models, determining the disease type corresponding to the any one of the first disease identification models as the disease onset type;

[0017] determining the intervention intensity for the disease onset type according to the intervention intensity corresponding to each of various disease onset probabilities for the disease onset type and the disease onset probability output by the any one of the first disease identification models.

[0018] In the embodiments of the present application, as an optional embodiment, after the step of performing the intervention processing according to the intervention-related information, the intervention method further comprises:

[0019] obtaining the to-be-tested electroencephalogram signal after the intervention processing;

[0020] performing potential interference subtraction processing on the to-be-tested electroencephalogram signal to obtain a potential interference-free electroencephalogram signal;

[0021] performing the steps of determining the brain wave feature set of the electroencephalogram signal, determining the intervention-related information for the target user based on the brain wave feature set, and performing the intervention processing according to the intervention-related information on the potential interference-free electroencephalogram signal.

[0022] In the embodiments of the present application, as an optional embodiment, before the step of determining the intervention-related information for the target user based on the brain wave feature set, the intervention method further comprises:

[0023] determining the disease onset state of the target user;

[0024] the step of determining the intervention-related information for the target user based on the brain wave feature set comprises:

[0025] if the disease onset state of the target user is disease onset, determining the intervention-related information for the target user based on the brain wave feature set.

[0026] In the embodiments of the present application, as an optional embodiment, the step of determining the disease onset state of the target user comprises:

[0027] determine an onset state of the target user based on the input onset identification for the target user.

[0028] In the embodiments of the present application, as an optional embodiment, the step of determining the onset state of the target user comprises:

[0029] input the brain wave feature set into a plurality of preset second disease identification models respectively, to obtain an onset probability output by each of the plurality of second disease identification models;

[0030] if the onset probability output by any one of the second disease identification models is greater than a preset second onset probability threshold for a disease type corresponding to the any one of the second disease identification models, determine that the onset state of the target user is onset.

[0031] In the embodiments of the present application, as an optional embodiment, the step of determining the brain wave feature set of the electroencephalogram signal comprises:

[0032] preprocess the electroencephalogram signal to obtain a preprocessed signal;

[0033] convert the preprocessed signal into a frequency spectrum graph;

[0034] determine power spectrum energy values corresponding to a plurality of preset frequency bands included in the frequency spectrum graph respectively;

[0035] determine the brain wave feature set according to the power spectrum energy values corresponding to the plurality of preset frequency bands included in the frequency spectrum graph respectively.

[0036] In a second aspect, the embodiments of the present application also provide an intervention device for deep brain stimulation, the intervention device comprising:

[0037] a signal acquisition module configured to acquire an electroencephalogram signal of a target user;

[0038] a feature set determination module configured to determine a brain wave feature set of the electroencephalogram signal;

[0039] an intervention information determination module configured to determine intervention-related information for the target user based on the brain wave feature set;

[0040] an intervention processing module configured to perform intervention processing according to the intervention-related information.

[0041] In a third aspect, the embodiments of the present application also provide an electronic device, comprising a processor, a memory and a bus, the memory stores machine readable instructions executable by the processor, when the electronic device is running, the processor and the memory communicate through the bus, and the machine readable instructions are executed by the processor to perform the steps of the intervention method for deep brain stimulation as described above.

[0042] In a fourth aspect, the embodiments of the present application further provide a computer readable storage medium, which stores a computer program. When the computer program is run by a processor, the steps of the intervention method for deep brain stimulation are executed.

[0043] The embodiments of the present application provide an intervention method, device, electronic equipment and storage medium for deep brain stimulation. The intervention method comprises: acquiring an electroencephalogram of a target user, determining an electroencephalogram feature set of the electroencephalogram, determining intervention related information for the target user based on the electroencephalogram feature set, and performing intervention processing according to the intervention related information.

[0044] Compared with the existing method of DBS intervention by medical staffs according to experience and patient symptoms, the embodiments of the present application can automatically select intervention related information suitable for the patient's disease condition according to the patient's brain wave characteristics when performing DBS intervention, so as to achieve the effect of precise intervention treatment.

[0045] In order to make the above objectives, characteristics and advantages of the present application more apparent and easy to understand, the following preferred embodiments are specifically described below, and the accompanying drawings are described in detail as follows. BRIEF DESCRIPTION OF DRAWINGS

[0046] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0047] Figure 1 A structural schematic diagram of an intervention system for deep brain stimulation provided by the embodiments of the present application;

[0048] Figure 2 A flowchart of an intervention method for deep brain stimulation provided by the embodiments of the present application;

[0049] Figure 3 A flowchart of applying a first disease identification model provided by the embodiments of the present application;

[0050] Figure 4 A flowchart of applying a standard intervention model provided by the embodiments of the present application;

[0051] Figure 5 A flowchart of an intervention method for deep brain stimulation provided by the embodiments of the present application;

[0052] Figure 6A structural schematic diagram of an intervention device for deep brain stimulation provided by an embodiment of the present application.

[0053] Figure 7 A structural schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0054] In order to make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the following will be combined with the accompanying drawings for the embodiments of the present application to make a clear and complete description of the technical solutions of the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, every other embodiment obtained by a person skilled in the art without creative work falls within the scope of protection of the present application.

[0055] In addition, the described embodiments of the present application are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, every other embodiment obtained by a person skilled in the art without creative work falls within the scope of protection of the present application.

[0056] It should be noted that the term "comprising" will be used in the embodiments of the present application to indicate the presence of the features declared thereafter, but does not exclude the addition of other features.

[0057] In order to enable a person skilled in the art to use the present application, the following implementation modes are given in combination with a specific application scenario "deep brain stimulation". Deep brain stimulation (DBS) mainly implants electrodes into the brain of a patient, uses a pulse generator to stimulate certain neural nuclei in the deep brain, corrects abnormal brain electrical circuits, and thus reduces symptoms in these neural aspects. Unlike some treatment methods (ablation or radiotherapy) that permanently damage and irreversibly damage the brain, DBS does not destroy brain structures and can allow further treatment in the future. DBS can treat and intervene diseases mainly including Parkinson's disease, essential tremor, dystonia disease, epilepsy, obsessive-compulsive disorder, anorexia, senile dementia, etc.

[0058] Before the present application is proposed, the existing DBS intervention is usually a discharge intervention with fixed intensity, which needs to rely on medical staff to observe the patient's onset condition, and then adjust the intensity and frequency of the discharge according to experience and patient symptoms, and use this fixed intervention intensity in the next period of treatment until the next time the medical staff adjusts. However, such DBS intervention is an intervention treatment after the patient's onset, and cannot achieve inhibitory intervention before the patient's onset; at the same time, the intervention intensity determined by medical staff according to experience and patient symptoms is not accurate enough, and the intervention intensity of the patient in a period of time will not change with the change of the onset condition.

[0059] Based on this, the present application provides a deep brain stimulation intervention system to solve the above technical problems, which can first acquire the electroencephalogram of the target user, then determine the electroencephalogram feature set of the electroencephalogram, and then determine the intervention related information for the target user based on the electroencephalogram feature set, and finally execute the intervention processing according to the intervention related information.

[0060] In this way, the intervention system can monitor the patient's neurological disease onset in real time, automatically select the intervention related information suitable for the patient's onset condition according to the patient's electroencephalogram feature, so as to realize the intervention before the patient's onset and improve the time accuracy of DBS intervention; at the same time, the DBS intervention can automatically feedback the appropriate intervention intensity, and can achieve the effect of precise intervention treatment.

[0061] Figure 1 A structural schematic diagram of a deep brain stimulation intervention system provided by an embodiment of the present application is shown in FIG. 1. Figure 1 As shown in FIG. 1, the intervention system includes a brain-computer interface (BCI) device (not shown in the figure), a deep brain stimulation device, a terminal device 20 and a server 30. The deep brain stimulation device includes a stimulation generator 100, a neural probe 101 and a stimulator 104, the stimulator 104 is installed on the neural probe 101, the neural probe 101 is implanted into a brain area 105, the stimulation generator 100 is connected with the stimulator 104 through a lead wire 102, in application, a power supply module 103 supplies power for the stimulation generator 100, the current generated by the stimulation generator 100 acts on the neural probe 101 through the stimulator 104, so as to complete the intervention on the user. In application, the brain-computer interface (BCI) is used to collect the electroencephalogram of the user and send it to the terminal device 20, the terminal device 20 transmits the electroencephalogram to the server 30, so that the server 30 analyzes the received electroencephalogram and completes the intervention through the control of the deep brain stimulation device.

[0062] In some embodiments, the server can be a single server or a group of servers. The group of servers can be centralized or distributed (e.g., the servers can be a distributed system). In some embodiments, the server can be local or remote with respect to the terminal device. For example, the server can access information and / or data stored in the terminal device, or a database, or any combination thereof. As another example, the server can be directly connected to a database to access stored information and / or data. In some embodiments, the server can be implemented on a cloud platform.

[0063] In some embodiments, the server can include a processor. The processor can process information and / or data related to the intervention of deep brain stimulation to perform one or more functions described in the present application. For example, the processor can determine the intensity of the intervention based on the encoded information of the brain electrical signals obtained from the terminal device. In some embodiments, the processor can include one or more processing cores (e.g., a single-core processor (S) or a multi-core processor (S)). By way of example only, the processor can include a central processing unit (CPU), an application specific integrated circuit (ASIC), an application specific instruction-set processor (ASIP), a physics processing unit (PPU), a digital signal processor (DSP), a field programmable gate array (FPGA), a programmable logic device (PLD), a controller, a microcontroller unit, a reduced instruction set computing (RISC), or a microprocessor, etc., or any combination thereof.

[0064] In some embodiments, the terminal device can include, but is not limited to, any one of the following devices: a smartphone, a tablet computer, a portable computer, a desktop computer, a digital television, etc. The terminal device has installed and running an application processing program, such as an application program that analyzes the early warning algorithm of the brain electrical signals and / or a rule feature system for representing the corresponding relationship between the attack probability for the attack type and the intervention intensity. In the embodiments of the present application, the application program includes the two application programs described above.

[0065] Specifically, the brain-computer interface BCI device can be an invasive brain-computer interface BCI device or a non-invasive brain-computer interface BCI device. If it is an invasive brain-computer interface BCI device, the neural probe of the deep brain stimulation device can be used as the acquisition end of the invasive brain-computer interface BCI device, that is, the brain-computer interface BCI device acquires the electroencephalogram signal by using the neural probe of the deep brain stimulation device, and sends the electroencephalogram signal collected by the neural probe to the terminal device. In application, the neural probe is inserted into each brain region of the target user, and the contact points on the neural probe collect the electroencephalogram signal (i.e., the brain activity electrical signal); wherein the contact points on the neural probe, also known as collectors, are located at the central brain region and the neural probe tip outside the brain ventricle, composed of a platinum dot array, and the data is transmitted by the photoetching cable in the neural probe; here, the brain region refers to a brain function partition, which is composed of multiple brain function latitudes, and each brain region position in the embodiment of the present application corresponds to each type of nervous system disease, wherein the stimulated brain region is generally in the midbrain and the brain ventricle outside the periphery, and the nervous system disease includes Parkinson's disease, essential tremor, dystonia disease, epilepsy, obsessive-compulsive disorder, anorexia nervosa, and senile dementia, etc. The brain-computer interface BCI device encodes the electroencephalogram signal on the neural probe and transmits it to the terminal device, and then transmits it to the cloud server for calculation through the terminal device; wherein when the brain-computer interface BCI device encodes and transmits the electroencephalogram signal, a high-precision analog-to-digital conversion device is used to integrate the dot array signal on the multi-channel neural probe, and then convert it into a high-frequency digital signal for carrier wave, and the electroencephalogram signal is transmitted to the terminal device through a high-speed Bluetooth system; the terminal device uses a 5G network to transmit the electroencephalogram signal to the cloud server in real time for decoding; the cloud server decodes and analyzes the collected encoded electroencephalogram signal, and performs preprocessing such as denoising; if the electroencephalogram signal is collected when the DBS intervention does not occur, it does not need to be processed specially, and after applying conventional filtering and other operations, it is directly decoded and analyzed; if the electroencephalogram signal is collected when the DBS intervention occurs, it needs to be processed by a specific algorithm to reduce the potential interference caused by the DBS intervention.

[0066] After the cloud server decodes and preprocesses the electroencephalogram signal, based on the pre-stored early warning algorithm based on the deep learning model on the cloud server, the preprocessed electroencephalogram signal is analyzed to obtain the incidence probability of whether a disease occurs and whether the disease has occurred; the diseases identified here include but are not limited to Parkinson's disease, essential tremor, dystonia disease, epilepsy, obsessive-compulsive disorder, anorexia, and senile dementia; here, more first disease identification models under different disease types can be constructed as needed to achieve coverage of a wider range of disease types; for example, if the early warning algorithm determines that the incidence probability is greater than 90% or the disease has occurred, an instruction to start stimulation is sent to the deep brain stimulation device, and otherwise an instruction to stop stimulation is sent.

[0067] When the cloud server sends the instruction of starting the stimulation to the deep brain stimulation device, the stimulation with appropriate intervention intensity is selected according to the real-time brain electrical activity of the patient and the pre-stored application characteristic rule system on the cloud server.

[0068] Here, the characteristic rule system can be modified by a correction algorithm based on the disease characteristics of individual patients, wherein the definition of the characteristic rule system is as follows: since different disease states correspond to different intervention needs, standardized DBS intervention intensity is taken as a reference based on the experience of professional doctors, and different intervention intensities are selected when the disease severity is different. The individual correction algorithm is an algorithm that adjusts the intervention intensity corresponding to the disease type according to the disease data of different individuals, so as to adapt to the disease characteristics of different patients. For example, for Parkinson's disease, the disorder index of brain regions varies from person to person, and the standard unit stimulation intensity labeled by medical staff and the corresponding standard unit disorder index collected are important parameters for adjusting the intervention intensity.

[0069] The deep brain stimulation intervention system provided by the embodiments of the present application can monitor the disease state of the patient's neurological disease in real time, so as to intervene before the patient is ill and improve the time accuracy of DBS intervention; at the same time, during the DBS intervention, the appropriate intervention intensity can be automatically fed back, and the effect of precise intervention treatment can be achieved.

[0070] The deep brain stimulation intervention method provided by the embodiments of the present application will be described in detail in combination with the contents described in the deep brain stimulation intervention system shown in the above Figure 1

[0071] Please refer to Figure 2 , Figure 2 The flowchart of the deep brain stimulation intervention method provided by the embodiments of the present application is shown in Figure 2 The specific execution process is as follows:

[0072] S210, acquiring the brain electrical signal of the target user.

[0073] S220, determining the brain wave feature set of the brain electrical signal.

[0074] S230, determining the intervention related information for the target user based on the brain wave feature set.

[0075] S240, performing intervention processing according to the intervention related information.

[0076] ​The target user in the embodiments of the present application refers to a patient suffering from a nervous system disease and needing deep brain stimulation. The target user can be a patient in a normal state at the current time or a patient in a coma or a vegetative state at the current time. The brain electrical signal is a brain nerve signal, which refers to a weak potential difference change generated in different brain regions due to the activity of brain neurons. According to the brain region position and the signal observation position, the brain electrical signal can be divided into a brain cortex signal (EEG signal), an intracortical signal, a neuron peripheral signal, and the like.

[0077] The method provided in the embodiments of the present application can be applied to a brain-computer interface (BCI) device, a terminal, a server, and the like. The following describes each step of the above example method provided in the embodiments of the present application by taking the application of the method to a server as an example.

[0078] In step S210, the server acquires the brain electrical signal of the target user. The brain electrical signal is collected by a deep brain stimulation device. The neural probes on the deep brain stimulation device are arranged in different brain regions of the target user. The brain electrical signal is collected by the neural probes in multiple different brain region positions. The collected brain electrical signal is converted into a high-frequency digital signal. The converted brain electrical signal is transmitted to a terminal device through a Bluetooth system, so that the terminal device transmits the brain electrical signal to the server in real time. The collected brain electrical signal can be encoded in an encoding manner. The encoded brain electrical signal is transmitted to the terminal device, so that the terminal device transmits the encoded brain electrical signal to the server in real time. The server can decode and preprocess the encoded brain electrical signal.

[0079] In step S220, the server determines a brain wave feature set of the brain electrical signal based on the brain electrical signal of the target user. The server analyzes and processes the acquired brain electrical signal to determine the brain wave feature set of the brain electrical signal.

[0080] Specifically, the brain wave feature set of the brain electrical signal is determined by the following steps: preprocessing the brain electrical signal to obtain a preprocessed signal, converting the preprocessed signal into a frequency spectrum graph, determining power spectrum energy values corresponding to multiple preset frequency bands included in the frequency spectrum graph, and determining the brain electrical feature set according to the power spectrum energy values corresponding to the multiple preset frequency bands included in the frequency spectrum graph.

[0081] Here, due to the instability and irregularity of the electroencephalogram signal, the electroencephalogram signal usually needs to be pre-processed to obtain a certain regular signal, i.e. a pre-processed signal; wherein the pre-processing includes rejecting interference noise, filtering out low-frequency interference signals, etc.; the pre-processed signal is converted into a frequency spectrum diagram, wherein the frequency spectrum diagram is used to represent the relationship between signal frequency and energy, so as to determine the power spectrum energy values corresponding to a plurality of preset frequency bands respectively; and then, according to the power spectrum energy values corresponding to a plurality of preset frequency bands respectively included in the frequency spectrum diagram, the electroencephalogram feature set is determined, wherein the electroencephalogram feature set includes power spectrum energy values corresponding to a plurality of preset frequency bands respectively, such as power spectrum energy values of alpha waves, beta waves, theta waves, delta waves, etc.

[0082] In step S230, based on the electroencephalogram feature set, intervention-related information for the target user is determined, where the intervention-related information serves as reference information for performing intervention processing.

[0083] Specifically, the intervention-related information includes an onset type and an intervention intensity for the onset type.

[0084] Here, the onset type is at least one of the neurological diseases including Parkinson's disease, essential tremor, dystonia disease, epilepsy, obsessive-compulsive disorder, anorexia nervosa, and senile dementia. The intervention intensity refers to the size of the current stimulation applied to the brain area of the target user. After the intervention intensity is determined, the target user can be intervened according to the determined intervention intensity.

[0085] In the embodiments of the present application, based on the electroencephalogram feature set, the onset type of the target user and the intervention intensity for the onset type are determined.

[0086] When the onset type of the target user is determined based on the electroencephalogram feature set, the specific determination method includes the following two kinds:

[0087] In an optional implementation, a relationship reference table between the waveform interval corresponding to the electroencephalogram feature set and the corresponding onset type can be pre-stored on the server, wherein the relationship reference table is used to represent the relationship between the waveform of the electroencephalogram and the onset type. In this way, according to the waveform of the electroencephalogram and the relationship reference table, the onset type of the target user can be determined.

[0088] In an optional implementation, a first disease recognition model is set on the server, and the electroencephalogram feature set is input into the pre-trained first disease recognition model to obtain the onset probability of the target user under the first disease recognition model, wherein there are a plurality of first disease recognition models, and for each disease type, there is a corresponding first disease recognition model.

[0089] Specifically, the brain wave feature set is input into a plurality of preset first disease identification models respectively, to obtain a plurality of disease incidence probabilities output by the plurality of first disease identification models respectively; if a disease incidence probability output by any first disease identification model is greater than a preset first disease incidence probability threshold corresponding to the any first disease identification model, a disease type corresponding to the any first disease identification model is determined as a disease incidence type; and an intervention intensity corresponding to the disease incidence type is determined according to a plurality of preset intervention intensities corresponding to various disease incidence probabilities of the disease incidence type and the disease incidence probability output by the any first disease identification model.

[0090] Here, the first disease identification model is used to determine the disease incidence probability of the target user, and then determine the disease incidence type of the target user according to the disease incidence probability. For each disease type, there is a first disease identification model corresponding to the disease type, that is, each first disease identification model can only output a disease incidence probability corresponding to one disease type.

[0091] The disease incidence probability in the embodiments of the present application is used to represent the possibility of the target user getting sick. Getting sick usually refers to the onset of a disease in the body. Probability usually refers to a ratio or proportion. Therefore, according to the literal meaning, the disease incidence probability usually refers to the proportion of the disease behavior, which can be calculated by calculating the ratio between the historical disease duration and the total monitoring duration. When the disease incidence probability is greater than the preset disease incidence probability threshold and less than the preset upper limit of the disease incidence probability, it indicates that the target user has not yet gotten sick, but has symptoms of getting sick. When the disease incidence probability is greater than or equal to the preset upper limit of the disease incidence probability, it can be considered that the target user has gotten sick, and the target disease incidence type of the target user can be determined at this time. The preset disease incidence probability threshold represents the critical value of the target user having symptoms of getting sick, which is determined according to the historical experience of medical personnel. The preset upper limit of the disease incidence probability represents the critical value of the target user getting sick, which is also determined according to the historical experience of medical personnel. In the embodiments of the present application, the preset disease incidence probability threshold is the first disease incidence probability threshold corresponding to the first disease identification model.

[0092] It should be noted that a target user may have a plurality of disease incidence types with disease incidence probabilities output by a plurality of first disease identification models respectively greater than a preset first disease incidence probability threshold corresponding to any first disease identification model, that is, the target user may have symptoms of more than one disease. When this situation exists, the disease incidence probabilities greater than the first disease incidence probability thresholds of the first disease identification models are arranged in descending order, and the target user is intervened in these disease incidence types in descending order.

[0093] In the embodiments of the present application, each first disease identification model is trained by the following steps:

[0094] Multiple training samples are obtained for the disease type corresponding to the first disease identification model. Each training sample includes the EEG feature set of the training user's historical onset cycle and the onset probability of the historical onset cycle. The constructed neural network is trained through multiple training samples until the preset training conditions are met. Once the constructed neural network is determined to be trained, the first disease identification model is obtained.

[0095] Here, the historical onset period includes a period of time before the target user develops the disease and a period of time during the onset of the disease.

[0096] The preset training conditions can be a certain number of training iterations or the loss function corresponding to the first disease identification model tending to equilibrium.

[0097] Specifically, the system acquires disease incidence data from historical disease cycles of multiple training users. This data includes EEG feature sets for training users prior to and within the historical disease cycles. Based on the EEG signal characteristics within the historical disease cycles and the disease incidence probability (which determines if a user has developed the disease), the system determines the disease incidence probability intervals and corresponding EEG signal feature intervals for multiple training users at the time of their onset. Training users within the determined disease incidence probability intervals are considered to have already developed the disease. Furthermore, based on the relationship between the EEG signal characteristics prior to and within the historical disease cycles (which can be linear, such as a direct proportional relationship), the system divides the EEG signal characteristics within the historical disease cycles into multiple intervals. Then, based on the disease incidence probability intervals and corresponding EEG signal feature intervals at the time of onset, the system determines the disease incidence probability intervals corresponding to each EEG signal feature interval prior to the historical disease cycles. There are multiple EEG signal feature intervals prior to the historical disease cycles, and correspondingly, multiple disease incidence probability intervals.

[0098] Specifically, for the first disease identification model before the onset of illness, the EEG feature set of training users from the training samples corresponding to each disease type before the historical onset period is input into the constructed neural network for learning. The model outputs the onset probability of the training user under that disease type. The onset probability output by the model is compared with the onset probability before the historical onset period to obtain the comparison error. When the comparison error approaches 0, the neural network can be considered to have completed training, thus obtaining the trained first disease identification model. Similarly, the first disease identification model during the onset of illness is trained using the same method.

[0099] For example, the brain wave signal frequency of the training user is in a first frequency interval, the incidence probability of the training user corresponds to a first incidence probability, the brain wave signal frequency of the training user is in a second frequency interval, the incidence probability of the training user corresponds to a second incidence probability, and so on, to obtain a plurality of frequency intervals of the brain wave signal frequency and the incidence probability corresponding thereto, wherein the determination of the frequency interval is determined according to the historical working experience of the medical staff. For example, the incidence probability is divided into intervals according to a certain accuracy, such as 90%, 80%, 70%, 60%, 50%, 40%, and so on, each incidence probability corresponds to a brain wave signal feature interval, and it can be specified that the incidence probability is greater than 90% to determine that the user has an attack; the incidence probability is greater than 80% to determine that the user is intervened; and the incidence probability is less than 50% to determine that the user is not intervened. Here, the division accuracy of the incidence probability can be selected.

[0100] In an optional embodiment, modeling is performed for each disease type respectively, and for each disease type, only one first disease recognition model is obtained, which can respectively output the incidence probability of the target user before and after the attack. Specifically, a first time weight value is set in a first disease recognition model, which is used to ensure that the corresponding first disease recognition model is used within or before the historical attack period, and then, during training, the training sample includes the brain wave feature set of the training user within the historical attack period and before the historical attack period, and the incidence probability before the historical attack period and the incidence result (specified incidence probability) within the historical attack period, and the first disease recognition model is trained through the training sample.

[0101] In an optional embodiment, modeling is performed for each disease type respectively, and for each disease type, it can also be divided into a first disease recognition model before the attack and a first disease recognition model after the attack. For the first disease recognition model before the attack, the training sample includes the brain wave feature set of the training user before the historical attack period and the incidence probability before the historical attack period; for the first disease recognition model after the attack, the training sample includes the brain wave feature set of the training user within the historical attack period and the incidence result (specified incidence probability) within the historical attack period.

[0102] In an optional embodiment, the onset data of the training users are classified according to user attribute features, and the types of the attribute onset data are determined according to the types of the user attribute features; here, taking the user attribute features including age and gender as an example, first attribute onset data and second attribute onset data can be obtained, and then the first attribute onset data and the second attribute onset data are classified according to the onset types to obtain onset data corresponding to multiple onset types under the first attribute feature and onset data corresponding to multiple onset types under the second attribute feature. The first disease recognition model is grouped according to the user attribute features, and here, taking the user attribute features including age and gender as an example, the first disease recognition model is divided into two groups to obtain a first group of first disease recognition models and a second group of first disease recognition models; then, according to the types of the disease types, the first group of first disease recognition models and the second group of first disease recognition models are classified respectively to obtain multiple first disease recognition sub-models of the first group and multiple first disease recognition sub-models of the second group; the first disease recognition sub-models of the first group are trained based on the onset data corresponding to the multiple onset types under the first attribute feature; and the first disease recognition sub-models of the second group are trained based on the onset data corresponding to the multiple onset types under the second attribute feature.

[0103] For example, the first attribute feature can be a gender feature such as male and female; the second attribute feature can be an age feature such as dividing the age into multiple age ranges, 1-10 years old as an age range, 11-20 years old as an age range, 21-30 years old as an age range, and two first disease recognition sub-models are established for males and females, such as a male first disease recognition sub-model and a female first disease recognition sub-model; and multiple age first disease recognition sub-models are established for each age range. Specifically, the male first disease recognition sub-model, the female first disease recognition sub-model, and the multiple age first disease recognition sub-models all include multiple first disease recognition sub-models trained according to the disease types.

[0104] For example, in an optional embodiment, the attribute features of the target user are determined first, and then the corresponding first disease recognition sub-models are selected according to the attribute features of the target user, such as a 24-year-old female target user, a female first disease recognition sub-model and an age first disease recognition sub-model corresponding to 21-30 years old are selected, the female first disease recognition sub-model and the age first disease recognition sub-model are attached with corresponding weight values according to the actual experience of medical personnel, and then the brain wave feature set of the target user is input into the female first disease recognition sub-model and the age first disease recognition sub-model to obtain the onset probabilities of the target user under multiple disease types, and the disease types with an onset probability greater than a preset first onset probability threshold for any first disease recognition model are determined as the onset types of the target user.

[0105] In one optional implementation, a weight value range is set for the male first disease identification sub-model, the female first disease identification sub-model, and the first disease identification sub-model for each age. The weight values ​​within the weight value range are adjusted according to the attribute characteristics of the target user. For example, the weight value corresponding to the female first disease identification sub-model is increased, and the weight value corresponding to the male first disease identification sub-model is decreased, to obtain the target weight values ​​for the male first disease identification sub-model, the female first disease identification sub-model, and the first disease identification sub-model for each age. Then, the target user's EEG feature set is input into each first disease identification sub-model (male first disease identification sub-model, female first disease identification sub-model, and first disease identification sub-model for each age), and the incidence probability of the target user under each disease type is obtained by combining the results. Then, the disease type of the target user is determined according to the relationship between the incidence probability and the corresponding first incidence probability threshold.

[0106] For example, such as Figure 3 As shown, Figure 3 This is a flowchart illustrating the application of a first disease identification model as provided in an embodiment of this application. Figure 3 As shown, for each disease type, EEG signals are first collected, encoded, and transmitted to a terminal device. The terminal device then decodes and denoises the encoded EEG signals to obtain input data (EEG feature set). The input data is then fed into a model library (first disease identification model) for various disease types obtained through a pre-built neural network model framework. Feature extraction is performed using this model library (first disease identification model), and the probability of disease onset is determined based on the extracted features.

[0107] Furthermore, in this embodiment, the intervention intensity for each disease type is determined based on a pre-set intervention intensity corresponding to various disease probabilities for each disease type, and the disease probability output by any of the first disease identification models. Specific implementation schemes include the following two:

[0108] In an alternative implementation, an intervention comparison table can be established on the server to correlate the incidence probability of a disease type with the intervention intensity. In this way, the intervention intensity for the target user can be determined based on the intervention comparison table and the incidence probability of the target user corresponding to each disease type.

[0109] In an optional implementation, a pre-trained target intervention model is set up on the server. The incidence probability of the target user output by the first disease identification model is input into the pre-trained target intervention model to obtain the intervention intensity for the disease type, wherein the target intervention model is an intervention model corresponding to the disease type of the target user.

[0110] In an optional embodiment, a pre-trained target intervention model is set on the server, and the brain electrical feature set of the target user is input into the pre-trained target intervention model to obtain the intervention intensity for the disease type, wherein the target intervention model is an intervention model corresponding to the disease type of the target user.

[0111] It should be noted that a relationship between the disease probability and the degree of neurological disorder can be established, where the neurological disorder refers to a disorder of brain nerve function caused by certain stimulating factors, and the degree refers to the fluctuation range of the brain electrical signal when the target user has a neurological disorder. The greater the fluctuation range of the brain electrical signal, the greater the degree of neurological disorder, and vice versa. Furthermore, the greater the degree of neurological disorder, the greater the disease probability, and vice versa.

[0112] Specifically, the target intervention model is constructed by the following steps: establishing a standard intervention model, which is determined based on disease data of different disease types and is used to represent the corresponding relationship between the disease probability and the intervention intensity under the disease type; obtaining the disease probability and the intervention intensity for the disease type of the target user within the historical disease period of the disease type; and adjusting the standard intervention model based on the obtained disease probability and intervention intensity of the target user to obtain the target intervention model.

[0113] Here, the standard intervention model is derived based on the experience of professional doctors. Since different disease states correspond to different intervention needs, different intervention intensities are selected when the disease degree of the target user is different. Therefore, different target intervention models should be used for different target users to ensure the accuracy of the intervention, and therefore, the standard intervention model is modified using a correction algorithm to obtain a target intervention model for each target user.

[0114] For example, the standard intervention model can be modified by an individual patient disease characteristic-based correction algorithm to obtain a target intervention model for each individual patient. The individual correction algorithm is an algorithm that adjusts the parameters of the standard intervention model to fit the disease characteristics of different patients based on the disease data of different individual patients. For example, taking Parkinson's disease as an example, the disorder index of the brain region varies from person to person, and the standard unit stimulation intensity labeled by medical personnel and the corresponding standard unit disorder index collected are used as important parameters for adjusting the intervention model.

[0115] Specifically, each standard intervention model is trained by the following steps:

[0116] obtain a plurality of training samples for the disease type corresponding to the standard intervention model, each training sample including a probability of the training user getting sick in a historical sick period and a probability of the training user getting sick before the historical sick period, and an intervention intensity before the historical sick period and an intervention intensity in the historical sick period; train the constructed neural network through the plurality of training samples until a preset training condition is reached, determine that the constructed neural network is trained, and obtain the trained standard intervention model.

[0117] In an optional embodiment, modeling is performed for each disease type respectively, and for each disease type, a standard intervention model before getting sick and a standard intervention model during getting sick are further divided. For the standard intervention model before getting sick, the training sample includes the probability of the training user getting sick before the historical sick period and the intervention intensity before the historical sick period; for the standard intervention model during getting sick, the training sample includes the probability of the training user getting sick in the historical sick period and the intervention intensity in the historical sick period.

[0118] In an optional embodiment, modeling is performed for each disease type respectively, and for each disease type, only one standard intervention model is obtained, which can determine the intervention intensity for the probability of the target user getting sick before and after getting sick respectively. Specifically, a second time weight value is set in a standard intervention model, which is used to ensure that the corresponding standard intervention model is used before or in the historical sick period, and then, during training, the training sample includes the probability of the training user getting sick in the historical sick period and the probability of the training user getting sick before the historical sick period, and the intervention intensity before the historical sick period and the intervention intensity in the historical sick period, and the standard intervention model is trained through the training sample.

[0119] The preset training condition can be a certain number of training times, or the loss function corresponding to the standard intervention model tends to be balanced.

[0120] Specifically, for the standard intervention model before getting sick, the probability of the training user getting sick before the historical sick period in the training sample corresponding to each disease type is input into the constructed neural network for learning, and the intervention intensity of the training user under the disease type is output. The intervention intensity output by the model is compared with the intervention intensity before the historical sick period to obtain a comparison error, and when the comparison error approaches 0, it can be considered that the neural network training is completed, thereby obtaining the trained standard intervention model. Similarly, the same training method is also used for the standard intervention model during getting sick.

[0121] Further, after obtaining the trained standard intervention model, the standard intervention model is corrected by using a correction algorithm for each target user to obtain a target intervention model for each target user, so that the intervention degree of the target user under the current degree of neural disorder is automatically obtained by using the target intervention model and the degree of neural disorder of the target user, and the target user is intervened according to the obtained intervention degree.

[0122] For example, as shown in the following figure, Figure 4 Figure 4 A flowchart of an application of a standard intervention model is provided in the embodiments of the present application. As shown in the following figure, Figure 4 For each disease type, the EEG signal is collected in advance, the PD state recognition is performed according to the collected EEG signal, and the disease degree is determined by a professional doctor to obtain the individual heterogeneity model parameter. The individual heterogeneity model parameter is input into the individual model library (target intervention model) obtained through the pre-constructed neural network model framework. Through the target intervention model, the current intervention intensity of the target user can be obtained.

[0123] In addition, before step S220, the intervention method further includes determining the disease state of the target user, and step S220 includes determining the intervention related information for the target user based on the EEG feature set if the disease state of the target user is disease.

[0124] Here, when it is determined that the disease state of the target user is disease, the intervention related information for the target user is determined based on the EEG feature set again.

[0125] The disease state includes two kinds of disease and non-disease.

[0126] Specifically, the disease state of the target user can be determined by any of the following methods:

[0127] One: determining the disease state of the target user based on the input disease identifier of the target user.

[0128] Here, the disease identifier refers to a feature for judging whether the target user has disease, which can include a specific disease result, such as: the target user has disease, the target user has epilepsy, etc.; it can also include a preset identifier for determining that the disease has occurred, such as using "1" to represent that the target user has disease, etc. The disease identifier can be uploaded by the target user himself or by others (doctors, nurses, family members, etc.), and can be directly filled in and input on the terminal device of the intervention system or directly input on the deep brain stimulation device of the intervention system, etc.

[0129] ​Secondly, the brain wave feature set is input into a plurality of preset second disease recognition models respectively to obtain a plurality of second disease recognition model outputs of the incidence probability; if the incidence probability output by any one of the second disease recognition models is greater than a preset second incidence probability threshold corresponding to the disease type of the any one of the second disease recognition models, it is determined that the incidence state of the target user is an onset.

[0130] Here, the incidence probability output by the second disease recognition model is used to determine whether the incidence state of the target user is an onset or non-onset. The incidence probability output by the second disease recognition model is compared with the second incidence probability threshold corresponding to the disease type of the second disease recognition model, and the incidence state of the target user is determined to be an onset or non-onset according to the comparison result.

[0131] The first incidence probability threshold and the second incidence probability threshold can be set according to actual conditions. In an embodiment of the present application, as a preferred implementation, the first incidence probability threshold is greater than the second incidence probability threshold. When the first incidence probability threshold is greater than the second incidence probability threshold, it can be known that the present application determines that the target user has an onset when the incidence probability output by the second disease recognition model is greater than the second incidence probability threshold (here, a rough determination is made on whether the target user has an onset); then it continues to determine the disease type of the target user when the incidence probability output by the first disease recognition model is greater than the first incidence probability threshold (here, a detailed determination is made on the disease type of the target user), and finally determines the intervention intensity for the intervention on the target user according to the preset corresponding relationship between the incidence probability output by the first disease recognition model and the intervention intensity.

[0132] Specifically, the training method of the second disease recognition model is the same as that of the first disease recognition model, and details are referred to the training method of the first disease recognition model, which will not be repeated here.

[0133] In step S240, intervention processing is performed according to the intervention-related information. That is, according to the determined intervention-related information, the deep brain stimulation device (DBS device) is controlled to intervene in the target user.

[0134] In this step, the control signal corresponding to the intervention intensity in the intervention-related information is sent to the deep brain stimulation device, and the generator of the deep brain stimulation device generates an intervention current matching the intervention intensity. The intervention current acts on the neural probe, and the target user is stimulated by the current through the neural probe.

[0135] Here, when intervening the target user, current intervention can be adopted, and the intensity of the stimulated current can be used as the basis for judging the intensity of the stimulation. For example, taking the conventional stimulation as 1 standard unit, the intensity of the stimulation is from 5% to 195%, and each 5% is a level, forming a ladder-type stimulation option; the judgment is performed once every 10 seconds, and the intensity of the stimulation is adjusted in real time according to the situation of the diseased brain area.

[0136] In the embodiment of the present application, as a preferred embodiment, after step S230, the intervention method further includes: acquiring the to-be-tested brain electrical signal after the intervention processing is performed; performing potential interference reduction processing on the to-be-tested brain electrical signal to obtain a brain electrical signal without potential interference; performing the step of determining the brain electrical wave feature set of the brain electrical signal on the brain electrical signal without potential interference; determining the intervention-related information for the target user based on the brain electrical wave feature set; and performing the step of performing the intervention processing according to the intervention-related information.

[0137] Here, after the intervention is performed on the target user, the to-be-tested brain electrical signal of the target user is acquired, and then the to-be-tested brain electrical signal is subjected to potential interference reduction processing, and the to-be-tested brain electrical signal after the processing is subjected to steps S210 to S240, so as to realize real-time analysis of the brain electrical signal of the target user, and timely adjust the intervention intensity until the target user is in a non-diseased state.

[0138] Here, if the intervention is applied to the target user within a predetermined time period, the acquired brain electrical signal is subjected to potential interference reduction processing, the brain electrical wave feature set is determined from the brain electrical signal after the potential interference reduction processing, and the intervention-related information for the target user is determined based on the obtained brain electrical wave feature set.

[0139] It should be noted that if the intervention is not applied to the target user within a predetermined time period, the acquired brain electrical signal is directly subjected to filtering processing to obtain the brain electrical wave feature set, and the intervention-related information for the target user is determined based on the brain electrical wave feature set after the filtering processing.

[0140] Here, the potential interference reduction processing is used to reduce the potential interference caused by the DBS intervention, and a common processing is to use a specific algorithm in the art to reduce the potential interference.

[0141] For example, please refer to Figure 5 , Figure 5 The overall step flowchart of the intervention method for deep brain stimulation provided by the embodiment of the present application. As shown in Figure 5As shown in the deep brain stimulation device, the brain electrical signal is collected first, then the brain electrical signal is encoded and transmitted to the terminal device, the encoded brain electrical signal is transmitted to the server through the terminal device, the server decodes the encoded brain electrical signal, determines the incidence probability of the target user under different disease types through the second disease identification model, and determines the incidence state of the target user as incidence when the incidence probability output by any second disease identification model is greater than the preset second incidence probability threshold corresponding to the disease type of the second disease identification model; then the incidence probability of the target user is determined through the first disease identification model, and the incidence type of the target user is determined when the incidence probability output by the first disease identification model is greater than the first incidence probability threshold; finally, the intervention intensity corresponding to the incidence type of the target user is determined according to the pre-set intervention intensity corresponding to the incidence probability of the incidence type, and the intervention is applied to the target user according to the obtained intervention intensity.

[0142] It should be understood that the intervention method of deep brain stimulation provided by the embodiments of the present application processes brain electrical signals every certain period of time to adjust the intervention intensity. In this way, the above-mentioned intervention method is executed in a cycle to adjust the intervention intensity in real time according to the incidence brain area of the target user.

[0143] In summary, the intervention method of deep brain stimulation provided by the embodiments of the present application can monitor the incidence of neurological diseases of the patient in real time, so as to intervene before the patient is ill and improve the time accuracy of DBS intervention. Meanwhile, the appropriate intervention intensity and frequency can be automatically fed back during DBS intervention, so as to achieve the effect of precise intervention treatment.

[0144] Based on the same inventive concept, the embodiments of the present application also provide an intervention device for deep brain stimulation corresponding to the intervention method of deep brain stimulation. Since the principle of solving problems in the device of the embodiments of the present application is similar to the above-mentioned intervention method of deep brain stimulation, the implementation of the device can be referred to the implementation of the method, and the repeated parts will not be described herein.

[0145] Please refer to Figure 6 , Figure 6 The structure of the intervention device for deep brain stimulation provided by the embodiments of the present application is shown in the figure. As Figure 6 As shown in the intervention device 600,

[0146] The signal acquisition module 610 is configured to acquire the brain electrical signal of the target user.

[0147] The feature set determination module 620 is configured to determine the brain wave feature set of the brain electrical signal.

[0148] The intervention information determination module 630 is configured to determine intervention-related information for the target user based on the brain wave feature set.

[0149] The intervention processing module 640 is configured to perform intervention processing according to the intervention-related information.

[0150] Optionally, the intervention-related information includes:

[0151] an onset type;

[0152] an intervention intensity for the onset type.

[0153] Optionally, the intervention information determination module 630 is specifically configured to:

[0154] input the brain wave feature set into a plurality of preset first disease recognition models respectively, to obtain an onset probability output by each of the plurality of first disease recognition models;

[0155] if the onset probability output by any one of the first disease recognition models is greater than a preset first onset probability threshold corresponding to the any one of the first disease recognition models, determine a disease type corresponding to the any one of the first disease recognition models as the onset type;

[0156] determine the intervention intensity for the onset type according to intervention intensities corresponding to various onset probabilities of the onset type and the onset probability output by the any one of the first disease recognition models.

[0157] Optionally, the intervention device 600 further includes a signal interference processing module 650, which is configured to:

[0158] obtain a to-be-tested electroencephalogram signal after the intervention processing is performed;

[0159] perform potential interference subtraction processing on the to-be-tested electroencephalogram signal, to obtain a potential interference-free electroencephalogram signal;

[0160] perform the steps of determining the brain wave feature set of the electroencephalogram signal based on the brain wave feature set, determining the intervention-related information for the target user based on the brain wave feature set, and performing the intervention processing according to the intervention-related information on the potential interference-free electroencephalogram signal.

[0161] Optionally, the intervention device 600 further includes an onset state determination module 660, which is configured to determine an onset state of the target user.

[0162] The intervention information determination module 630 is configured to, if the onset state of the target user is an onset, determine the intervention-related information for the target user based on the brain wave feature set.

[0163] Optionally, the onset state determination module 660 is specifically configured to:

[0164] determine the onset state of the target user based on the input onset identification for the target user.

[0165] Optionally, the onset state determination module 660 is specifically configured to:

[0166] input the brain wave feature set into a plurality of preset second disease identification models respectively, to obtain an onset probability output by each of the plurality of second disease identification models;

[0167] if the onset probability output by any one of the plurality of second disease identification models is greater than a preset second onset probability threshold for a disease type corresponding to the any one of the plurality of second disease identification models, determine that the onset state of the target user is onset.

[0168] Optionally, the feature set determination module 620 is specifically configured to:

[0169] preprocess the electroencephalogram signal to obtain a preprocessed signal;

[0170] convert the preprocessed signal into a frequency spectrum graph;

[0171] determine power spectrum energy values corresponding to a plurality of preset frequency bands included in the frequency spectrum graph respectively;

[0172] determine the brain wave feature set according to the power spectrum energy values corresponding to the plurality of preset frequency bands included in the frequency spectrum graph respectively.

[0173] The intervention device for deep brain stimulation provided by the embodiments of the present application includes a signal acquisition module, a feature set determination module, an intervention information determination module, and an intervention processing module, wherein the signal acquisition module acquires an electroencephalogram signal of a target user, the feature set determination module determines a brain wave feature set of the electroencephalogram signal, the intervention information determination module determines intervention-related information for the target user based on the brain wave feature set, and the intervention processing module performs intervention processing according to the intervention-related information.

[0174] In this way, the embodiments of the present application can monitor the onset of the neurological disease of the patient in real time, automatically select intervention-related information suitable for the onset state of the patient according to the brain wave features of the patient, and thus intervene before the patient is ill, thereby improving the time accuracy of DBS intervention; meanwhile, during DBS intervention, the appropriate intervention intensity can be automatically fed back, and the effect of precise intervention treatment can be achieved.

[0175] Please refer to Figure 7 , Figure 7 is a structural schematic diagram of an electronic device provided by the embodiments of the present application. As shown in Figure 7 , the electronic device 700 includes a processor 710, a memory 720, and a bus 730.

[0176] The memory 720 stores machine readable instructions executable by the processor 710, when the electronic device 700 is running, the processor 710 and the memory 720 communicate through the bus 730, the machine readable instructions are executed by the processor 710, can execute the above Figure 2 The steps of the intervention method of deep brain stimulation in the method embodiment are not repeated here.

[0177] The embodiment of the application further provides a computer readable storage medium, the computer readable storage medium stores a computer program, the computer program is run by the processor, and the computer program can execute the above Figure 2 The steps of the intervention method of deep brain stimulation in the method embodiment are not repeated here.

[0178] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working process of the above-described system, device and unit can refer to the corresponding process in the foregoing method embodiment, and will not be repeated here.

[0179] In several embodiments provided in the application, it should be understood that the disclosed system, device and method can be implemented by other ways. The device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, and for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some communication interface, device or unit, which can be electrical, mechanical or other forms.

[0180] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. According to actual needs, part or all of the units can be selected to achieve the purpose of the embodiment scheme.

[0181] In addition, the functional units in each embodiment of the application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit.

[0182] If the functions are realized in the form of software function units and sold or used as independent products, they can be stored in a nonvolatile computer readable storage medium executable by a processor. Based on this understanding, the technical solutions of the present application essentially or the parts of the prior art that make contributions or parts of the technical solutions can be embodied in the form of software products. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (Read-Only Memory, ROM), a random access memory (Random Access Memory, RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0183] Finally, it should be noted that: the above-described embodiments are only specific embodiments of the present application, which are used to illustrate the technical solutions of the present application, but not to limit them. The protection scope of the present application is not limited thereto. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can make modifications or easily think of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed by the present application, or make equivalent replacements to some of the technical features. The modifications, changes or replacements do not make the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An intervention device for deep brain stimulation, characterized in that The intervention device comprises a signal acquisition module, a feature set determination module, an intervention information determination module and an intervention processing module; The signal acquisition module is configured to acquire electroencephalogram signals of a target user. The feature set determination module is configured to determine an electroencephalogram feature set of the electroencephalogram signals, the electroencephalogram feature set comprising power spectrum energy values corresponding to the electroencephalogram signals in multiple preset frequency bands respectively. The intervention information determination module is specifically configured to: input the electroencephalogram feature set into multiple first disease recognition models respectively to obtain disease incidence probabilities output by the multiple first disease recognition models respectively; for each first disease recognition model, if the disease incidence probability output by the first disease recognition model is greater than a first disease incidence probability threshold preset for the first disease recognition model, determine a disease type corresponding to the first disease recognition model as an incidence type to determine all incidence types of the target user; for each incidence type, input the disease incidence probability output by the first disease recognition model corresponding to the incidence type into a target intervention model corresponding to the incidence type of the target user to determine an intervention intensity of the target user under the incidence type; The intervention processing module is configured to perform intervention processing according to the intervention intensity under each incidence type.

2. The intervention device of claim 1, wherein, The intervention device further comprises a signal interference processing module, which is configured to: acquire electroencephalogram signals to be measured after intervention processing; perform potential interference reduction processing on the electroencephalogram signals to be measured to obtain electroencephalogram signals without potential interference; perform the steps of determining the electroencephalogram feature set of the electroencephalogram signals based on the electroencephalogram feature set, determining intervention-related information for the target user based on the intervention-related information, and performing intervention processing according to the intervention-related information.

3. The intervention device of claim 1, wherein, The intervention device further comprises an incidence state determination module, which is configured to determine the incidence state of the target user. The intervention information determination module is configured to determine intervention-related information for the target user based on the electroencephalogram feature set if the incidence state of the target user is disease.

4. The intervention device of claim 3, wherein, The incidence state determination module is specifically configured to: determine the incidence state of the target user based on the input disease identifier for the target user.

5. The intervention device of claim 3, wherein, The incidence state determination module is specifically further configured to: input the electroencephalogram feature set into multiple second disease recognition models respectively to obtain disease incidence probabilities output by the multiple second disease recognition models respectively; if the disease incidence probability output by any second disease recognition model is greater than a second disease incidence probability threshold preset for a disease type corresponding to the second disease recognition model, determine that the incidence state of the target user is disease.

6. An intervention system for deep brain stimulation, characterized in that The intervention system comprises a deep brain stimulation device, a terminal device and a server; The terminal device is configured to transmit electroencephalogram signals collected by the deep brain stimulation device to the server; The server performs the following processing: acquire electroencephalogram signals of a target user; determine an electroencephalogram feature set of the electroencephalogram signals, the electroencephalogram feature set comprising power spectrum energy values corresponding to the electroencephalogram signals in multiple preset frequency bands respectively; The brain wave feature set is respectively input into a plurality of preset first disease identification models, and a plurality of disease identification models respectively output a plurality of disease probabilities; for each first disease identification model, if the disease probability output by the first disease identification model is greater than a preset first disease probability threshold for the first disease identification model, the disease type corresponding to the first disease identification model is determined as a disease type, so as to determine all disease types of the target user; for each disease type, the disease probability output by the first disease identification model corresponding to the disease type is input into a target intervention model corresponding to the disease type of the target user, so as to determine the intervention intensity of the target user under the disease type; According to the intervention intensity under each disease type, intervention processing is performed.

Images