CN110367976B

CN110367976B - Brain wave signal detection method, related device and storage medium

Info

Publication number: CN110367976B
Application number: CN201910697209.0A
Authority: CN
Inventors: 李悦翔; 郑冶枫
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2019-07-30
Filing date: 2019-07-30
Publication date: 2021-05-07
Anticipated expiration: 2039-07-30
Also published as: CN110367976A

Abstract

The embodiment of the application discloses a brain wave signal detection method, related equipment and a storage medium; the brain wave signal can be detected through artificial intelligence, and the method specifically comprises the following steps: collecting brain wave signals of an object to be detected under intermittent stimulation of interference signals, respectively obtaining brain area connection information of the object to be detected under the interference stimulation and brain area connection information under non-interference stimulation according to the brain wave signals, then calculating similarity between the brain area connection information under the interference stimulation and the brain area connection information under the non-interference stimulation, and determining the pathological type of the brain wave signals according to the similarity so as to generate a detection result of the object to be detected; the scheme can improve the processing efficiency and the detection accuracy.

Description

Brain wave signal detection method, related device and storage medium

Technical Field

The invention relates to the technical field of communication, in particular to a brain wave signal detection method, related equipment and a storage medium.

Background

The brain wave signal is an electric wave generated by the electrophysiological activity of brain nerve cells on the surface of the cerebral cortex or scalp, and the electrophysiological activity of the brain nerve cells can be known by detecting the change of the brain wave signal, thereby providing a basis for analyzing various mental diseases, such as anxiety disorder or depression.

In the prior art, the detection of brain wave signals generally depends on interpretation of professionals, such as doctors, but in the research and practice processes of the prior art, the inventor of the invention finds that much time is consumed and the processing efficiency is low because interpretation mainly depends on manual work; moreover, due to the technical quality and the diversity of the qualification of doctors, different people may obtain different interpretation contents for the same electroencephalogram, and even interpretation errors often occur, so the detection accuracy is not high.

Disclosure of Invention

The embodiment of the application provides a brain wave signal detection method, related equipment and a storage medium; the processing efficiency and the detection accuracy can be improved.

The embodiment of the application provides a brain wave signal detection method, which comprises the following steps:

collecting brain wave signals of an object to be detected under intermittent stimulation of interference signals;

respectively acquiring brain area connection information of the detection object under interference stimulation and brain area connection information under non-interference stimulation according to the brain wave signals;

calculating the similarity between the brain region connection information under the interference stimulation and the brain region connection information under the non-interference stimulation;

determining the type of pathology to which the brain wave signal belongs according to the similarity;

generating a detection result of the detection object based on the pathology type.

The embodiment of the present application further provides a brain wave signal detection device, including:

the acquisition unit is used for acquiring brain wave signals of an object to be detected under intermittent stimulation of interference signals;

an acquisition unit, configured to acquire brain region connection information of the detection object under interference stimulation and brain region connection information under non-interference stimulation, respectively, according to the brain wave signal;

the calculating unit is used for calculating the similarity between the brain area connection information under the interference stimulation and the brain area connection information under the non-interference stimulation;

a determination unit for determining a type of pathology to which the brain wave signal belongs, according to the similarity;

a generation unit configured to generate a detection result of the detection object based on the type of pathology.

Optionally, in some embodiments of the application, the determining unit is specifically configured to predict a pathology type to which the brain wave signal belongs according to the similarity through a post-training detection model, where the post-training detection model is trained from a plurality of brain wave signal samples labeled with the pathology type.

Optionally, in some embodiments of the present application, the obtaining unit may include a dividing subunit, a first determining subunit, and a second determining subunit, as follows:

the dividing subunit is used for dividing the brain wave signals into a first type of signal segment containing interference stimulation and a second type of signal segment containing non-interference stimulation;

the first determining subunit is configured to determine correlations between signals of different brain area channels in the first type of signal segment, so as to obtain brain area connection information of the detection object under the interference stimulation;

and the second determining subunit is used for determining the correlation among the signals of different brain area channels in the second type of signal segment to obtain the brain area connection information of the detection object under the non-interference stimulation.

Optionally, in some embodiments of the application, the dividing unit is specifically configured to acquire a label of each signal segment in the brain wave signal, and if the label indicates that the signal segment is a signal acquired under stimulation of an interference signal, classify the signal segment into a first type of signal segment; and if the label indication signal segment is the signal acquired under the stimulation of the non-interference signal, classifying the signal segment into a second type signal segment.

Optionally, in some embodiments of the application, the first determining subunit may be specifically configured to calculate a pearson correlation coefficient between signals of different brain area channels in the first type of signal segment, obtain a first correlation coefficient, and generate brain area connection information of the detection object under the interference stimulation according to the first correlation coefficient;

the second determining subunit may be specifically configured to calculate a pearson correlation coefficient between signals of different brain area channels in the second type of signal segment, obtain a second phase relation number, and generate brain area connection information of the detection object under non-interference stimulation according to the second phase relation number.

Optionally, in some embodiments of the present application, the first determining subunit may be specifically configured to construct a first correlation coefficient matrix by using a first correlation coefficient, convert the first correlation coefficient matrix into a plurality of one-dimensional feature vectors, and obtain a first feature set, where the feature vectors in the first feature set are used to reflect brain region connection information of the detection object under the interference stimulation;

the second determining subunit may be specifically configured to construct a second correlation number matrix by using a second correlation number, convert the second correlation number matrix into a plurality of one-dimensional feature vectors, and obtain a second feature set, where the feature vectors in the second feature set are used to reflect brain region connection information of the detection object under non-interfering stimulation.

Optionally, in some embodiments of the application, the first determining subunit may be specifically configured to divide the first correlation coefficient matrix into two symmetric regions, select one region from the two regions according to a preset policy, convert elements in the selected region into a one-dimensional feature vector, and obtain a first feature set;

the second determining subunit may be specifically configured to divide the second correlation matrix into two symmetric regions, select one region from the two regions according to the preset policy, and convert elements in the selected region into a one-dimensional feature vector to obtain a second feature set.

Optionally, in some embodiments of the present application, the calculating unit may be specifically configured to calculate a cosine similarity between a feature vector in the first feature set and a feature vector in the second feature set;

the determining unit may be specifically configured to predict a pathology type to which the brain wave signal belongs through a trained detection model according to the cosine similarity.

Optionally, in some embodiments of the application, the determining unit may be specifically configured to construct a cosine similarity matrix according to the calculated cosine similarity, and predict a type to which the cosine similarity matrix belongs through a detection model after training to obtain a pathological type to which the brain wave signal belongs.

Optionally, in some embodiments of the present application, the brain wave signal detecting apparatus may further include a training unit, as follows:

the acquisition unit can be further used for acquiring brain wave signal samples of a plurality of detection samples under the intermittent stimulation of interference signals, and the brain wave signal samples are labeled with pathological types;

the acquiring unit may be further configured to acquire brain region connection information of the detection sample under interference stimulation and brain region connection information of the detection sample under non-interference stimulation, respectively, according to the brain wave signal sample;

the calculation unit can be further used for calculating the similarity between the brain region connection information of the detection sample under the interference stimulation and the brain region connection information under the non-interference stimulation;

and the training unit is used for predicting the pathological type of the brain wave signal sample through a detection model according to the similarity, and converging the detection model according to the marked pathological type and the predicted pathological type to obtain the trained detection model.

Optionally, in some embodiments of the present application, the obtaining unit may be specifically configured to:

dividing the brain wave signal samples into first type signal segment samples containing interference stimulation and second type signal segment samples containing non-interference stimulation;

calculating a Pearson correlation coefficient between signals of different brain area channels in a first type of signal segment sample to obtain a first correlation coefficient sample, constructing a first correlation coefficient sample matrix by using the first correlation coefficient sample, converting the first correlation coefficient sample matrix into a plurality of one-dimensional eigenvectors to obtain a first sample characteristic set, wherein the eigenvectors in the first sample characteristic set are used for reflecting brain area connection information of the detection sample under interference stimulation;

and calculating a Pearson correlation coefficient between signals of different brain area channels in the second type of signal segment samples to obtain a second phase relation number sample, constructing a second phase relation number sample matrix by using the second phase relation number sample, converting the second phase relation number sample matrix into a plurality of one-dimensional eigenvectors to obtain a second sample characteristic set, wherein the eigenvectors in the second sample characteristic set are used for reflecting brain area connection information of the detection sample under non-interference stimulation.

Optionally, in some embodiments of the present application, the calculating unit may be specifically configured to calculate a cosine similarity between a feature vector in the first sample feature set and a feature vector in the second sample feature set;

the training unit may be specifically configured to construct a cosine similarity matrix according to the calculated cosine similarity, and predict a type to which the cosine similarity matrix belongs through a detection model, so as to obtain a pathological type to which the brain wave signal sample belongs.

Correspondingly, the embodiment of the application also provides an electronic device, which comprises a memory and a processor; the memory stores an application program, and the processor is used for running the application program in the memory to execute the operation in any one of the brain wave signal detection methods provided by the embodiments of the present application.

In addition, a storage medium is provided in an embodiment of the present application, and is characterized in that the storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor to execute the steps in any one of the brain wave signal detection methods provided in the embodiment of the present application.

According to the method and the device, the brain wave signals of the object to be detected under the intermittent stimulation of the interference signals can be collected, the brain area connection information of the object to be detected under the interference stimulation and the brain area connection information under the non-interference stimulation are respectively obtained according to the brain wave signals, then, the similarity between the brain area connection information under the interference stimulation and the brain area connection information under the non-interference stimulation is calculated, the pathological type of the brain wave signals is determined according to the similarity, and the detection result of the object to be detected is generated; because the scheme can respectively acquire the brain region connection information of the detection object under the interference stimulation and the brain region connection information of the detection object under the non-interference stimulation, and the corresponding pathological type is determined through the similarity between the two, the purpose of automatic interpretation is achieved, and therefore, compared with the existing scheme which can only rely on manual interpretation, the processing efficiency can be greatly improved, and because the scheme does not need to rely on manual interpretation, the situation of misjudgment caused by human factors can be avoided, and the detection accuracy is favorably improved.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

Fig. 1 is a schematic view of a scene of a brain wave signal detection method according to an embodiment of the present application;

fig. 2 is a flowchart of a brain wave signal detection method according to an embodiment of the present application;

FIG. 3 is an exemplary diagram of a tag in an electroencephalogram provided by an embodiment of the present application;

FIG. 4 is a schematic diagram illustrating vector stretching of a matrix of correlation coefficients according to an embodiment of the present disclosure;

FIG. 5a is a diagram illustrating a structure of a detection model provided in an embodiment of the present application;

FIG. 5b is a diagram of another exemplary structure of a detection model provided in an embodiment of the present application;

FIG. 5c is a diagram of another exemplary structure of a detection model provided in the embodiments of the present application;

FIG. 6 is a diagram of an example of an interference signal in an embodiment provided in the present application;

FIG. 7 is a diagram illustrating an example of an interfering signal and a non-interfering signal in an embodiment provided by the present application;

fig. 8 is a frame diagram of model training in a brain wave signal detection method according to an embodiment of the present application;

FIG. 9 is a graph comparing brain region connectivity information for a normal population and anxiety patients in an example provided herein;

fig. 10 is another schematic flow chart of a brain wave signal detection method according to an embodiment of the present application;

fig. 11 is a frame diagram of a brain wave signal detection method according to an embodiment of the present application;

fig. 12 is a schematic structural diagram of a brain wave signal detection apparatus according to an embodiment of the present application;

fig. 13 is another structural schematic diagram of a brain wave signal detecting apparatus according to an embodiment of the present application;

fig. 14 is a schematic structural diagram of an electronic device provided in an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

The embodiment of the application provides a brain wave signal detection method, related equipment (such as a brain wave signal detection device, electronic equipment and the like) and a storage medium. The electroencephalogram signal detection apparatus may be integrated into an electronic device, which may be a server, or a terminal, such as a medical device such as an electroencephalograph, a Personal Computer (PC), a tablet Computer, or a notebook Computer.

For example, as shown in fig. 1, the electronic device may collect brain wave signals (e.g., an electroencephalogram including brain wave signals) of a subject to be detected under intermittent stimulation of an interference signal, then respectively obtain brain region connection information of the subject under the interference stimulation and brain region connection information under non-interference stimulation according to the brain wave signals, calculate similarity between the brain region connection information under the interference stimulation and the brain region connection information under the non-interference stimulation, then determine a type of pathology to which the brain wave signals belong according to the similarity, e.g., determine whether the brain wave signals belong to anxiety disorder, determine whether the brain wave signals belong to depression disorder, or the like, and then may generate a detection result of the subject to be detected based on the type of pathology.

The object to be detected can be any living creature with a brain, such as a human, a monkey, a cat or a dog; for convenience of description, in the embodiments of the present application, description will be made taking a human as an example.

In the present embodiment, the term "interfering signal" refers to an external thing that can affect the brain waves of a living body (an object to be detected, a test sample, or the like), and the stimulus refers to a stimulus that can be felt by the living body (the object to be detected, the test sample, or the like) and cause the brain nerve cells of the living body to react. The brain nerve cells generate corresponding electric waves in response, the electric waves are called brain waves, and the brain waves are collected and expressed in the form of signals, namely brain wave signals.

For example, some preset types of test questions may be provided for the subject to be detected intermittently and asked to answer, where the test questions may be regarded as the interference signals, and the collected "brain wave signals generated by the subject to be detected when answering" may be regarded as the brain wave signals generated by the interference signal stimulation (referred to as interference stimulation for short). Since the interference signal is intermittent, the brain wave signal generated by the subject to be detected can be regarded as the brain wave signal generated by the non-interference signal stimulation (referred to as non-interference stimulation for short) when the interference signal is absent.

The following are detailed below. It should be noted that the following description of the embodiments is not intended to limit the preferred order of the embodiments.

The present embodiment will be described from the perspective of a brain wave signal detecting apparatus, which may be specifically integrated in an electronic device such as a server or a terminal, which may include a medical device such as an electroencephalograph, a PC, a tablet computer, a notebook computer, or the like.

A brain wave signal detection method includes: the method comprises the steps of collecting brain wave signals of an object to be detected under intermittent stimulation of interference signals, respectively obtaining brain area connection information of the object to be detected under the interference stimulation and brain area connection information under non-interference stimulation according to the brain wave signals, calculating similarity between the brain area connection information under the interference stimulation and the brain area connection information under the non-interference stimulation, determining the pathological type of the brain wave signals according to the similarity, and generating a detection result of the object to be detected based on the pathological type.

As shown in fig. 2, a specific flow of the brain wave signal detection method may be as follows:

101. collecting brain wave signals of an object to be detected under intermittent stimulation of interference signals.

For example, electroencephalograms (EEG) of a subject to be detected under intermittent stimulation of an interference signal may be collected, wherein the EEG includes brain wave signals.

For example, if the electroencephalogram signal detection apparatus is integrated into an electroencephalograph, then the electroencephalogram may be collected by the electroencephalograph specifically; if the electroencephalogram signal detection apparatus is integrated in other equipment that cannot directly acquire an electroencephalogram, the electroencephalogram signal detection apparatus may receive an electroencephalogram transmitted by other equipment, such as an electroencephalograph, or may receive an electroencephalogram input by a user (the electroencephalogram may be acquired by other electroencephalogram acquisition equipment), and so on.

Among them, the electroencephalogram is a graph obtained by recording spontaneous biopotentials of the brain from the scalp by a precise electronic instrument in an enlarged manner, and is a graph of spontaneous and rhythmic electrical activities of a brain nerve cell group recorded by an electrode (electrophysiological index). The electroencephalogram can record the changes of brain wave signals generated during brain activity, and is an overall reflection of the electrophysiological activity of brain nerve cells on the surface of the cerebral cortex or scalp.

Alternatively, the electroencephalogram may be acquired in various ways, for example, the electroencephalogram may be acquired by dynamic electroencephalogram monitoring (electroencephalogram monitoring) or Video electroencephalogram monitoring (VEEG, Video-EEG, also called Video electroencephalogram monitoring). The dynamic electroencephalogram monitoring is mainly recorded by electroencephalogram equipment such as an electroencephalograph, and can be continuously recorded for about 24 hours generally, while the video electroencephalogram monitoring is realized by adding synchronous video equipment on the basis of the electroencephalogram equipment, and the monitoring time can be flexibly mastered according to equipment conditions and actual requirements and can be varied from hours to days.

Optionally, when specifically acquiring brain wave signals, corresponding channels may be respectively set for different brain regions, where the channels are called brain region channels, and then electric waves generated by the corresponding brain regions may be respectively acquired through each brain region channel, and the obtained signals of the plurality of brain region channels are brain wave signals, that is, the brain wave signals acquired by the method may include signals of the plurality of brain region channels.

The brain Region (ER, Encephalic Region) is a short term for "brain function partition", and may be formed by combining a plurality of brain dimensions, and the combination may break through the medical scope, that is, the division of the brain Region is not the division of a body part, but the division of the brain Region is functional, and the specific division manner may be determined according to the requirements of the actual application, and is not described herein again.

102. And respectively acquiring brain area connection information of the detection object under interference stimulation and brain area connection information under non-interference stimulation according to the brain wave signals.

When a living being moves, the brain of the living being performs corresponding information processing at the same time, and the information processing requires cooperative work and dynamic interaction between different brain regions, so that connection relationships must be generated between different brain regions, and these information can reflect the connection relationships between different brain regions, which are referred to as brain region connection information in the embodiment of the present application.

There are various ways to "separately acquire the brain region connection information of the detection object under the interfering stimulation and the brain region connection information under the non-interfering stimulation from the brain wave signal", for example, the following may be specifically used:

(1) the brain wave signals are divided into a first type of signal segments containing interfering stimuli and a second type of signal segments containing non-interfering stimuli.

For example, in addition to recording the brain wave signals, the acquired electroencephalogram may record corresponding labels to indicate whether the section of brain wave signals is the signals acquired under the stimulation of the interference signals or the signals acquired under the stimulation of the non-interference signals; namely, the step of dividing the brain wave signals into the first type signal segments containing the interfering stimuli and the second type signal segments containing the non-interfering stimuli may include:

and obtaining a label of each signal segment in the brain wave signal, if the label indicates that the signal segment is a signal acquired under the stimulation of an interference signal, classifying the signal segment into a first type of signal segment, and if the label indicates that the signal segment is a signal acquired under the stimulation of a non-interference signal, classifying the signal segment into a second type of signal segment.

Optionally, for simplicity, only the label of the signal acquired under the stimulation of the interference signal may be identified, and if the label does not exist, the signal acquired under the stimulation of the non-interference signal is determined; for example, referring to fig. 3, in the electroencephalogram, c342, Y88, d464, and N99 are labels for signals collected under the stimulation of interfering signals (see the black vertical line portion for details).

Similarly, only the tags of the signals collected under the non-interfering signal stimulation may be identified, and if no tags exist, the signals collected under the interfering signal stimulation may be determined, and so on.

The labeling manner of the label may be specifically determined according to the requirements of the actual application, and is not described herein again.

(2) Determining the correlation among signals of different brain area channels in the first type of signal segment to obtain brain area connection information of the detection object under the interference stimulation; for example, the following may be specifically mentioned:

and calculating Pearson correlation coefficients (Pearson correlation coefficients) among signals of different brain area channels in the first type of signal segments to obtain first correlation coefficients, and generating brain area connection information of the detection object under the interference stimulation according to the first correlation coefficients.

The Pearson correlation coefficient is mainly used for measuring whether two data sets are on the same line or not, namely measuring the linear relation between distance variables. Therefore, each signal (i.e., a brain electrode signal, which is referred to as a signal in this embodiment) in a brain region channel may be regarded as a data set, for example, referring to fig. 3, taking an example that one brain region channel includes 128 signals, and each signal may include more than 10 ten thousand time point data, at this time, each brain region channel may correspond to 128 data sets, each data set may include more than 10 ten thousand elements, that is, each signal may be regarded as one data set, and each time point data may be regarded as one element.

For example, taking the example that the signals of two different brain region channels respectively correspond to two data sets x and y, the calculation formula of the pearson correlation coefficient r between the signals of the two different brain region channels (i.e. the data set x and the data set y) may be as follows:

wherein x is_iFor elements in the data set x, y_iThe number of elements in the data set y (i.e. the length of the sampled time point data) and the number of elements in the data set x and y, N may be more than 10 ten thousand, and so on.

According to the above-mentioned way of calculating the pearson correlation coefficient r, for every two brain area channels, the pearson correlation coefficients, such as 128 pearson correlation coefficients, which are consistent with the signal number can be obtained, so that the correlation coefficients can be used to construct the corresponding correlation coefficient matrix for the subsequent processing; for convenience of description, in the embodiment of the present application, the correlation coefficient matrix constructed by using the first correlation coefficient is referred to as a first correlation coefficient matrix. Namely, the step of generating the brain region connection information of the detected object under the interference stimulation according to the first correlation coefficient may include:

and constructing a first correlation coefficient matrix by using the first correlation coefficient, and converting the first correlation coefficient matrix into a plurality of one-dimensional eigenvectors to obtain a first feature set.

And the feature vectors in the first feature set are used for reflecting the brain region connection information of the detected object under the interference stimulation.

For example, taking an example that each data set includes 128 signals, at this time, a 128 × 128 first correlation coefficient matrix may be constructed, and then the first correlation coefficient matrix is converted into a plurality of one-dimensional feature vectors to obtain a first feature set.

Optionally, since the first correlation coefficient matrix is in a diagonal symmetric form, in order to improve processing efficiency and reduce consumption of computational resources, the first correlation coefficient matrix may be divided into two symmetric regions, and then, only elements in one of the regions are vector-stretched to convert the elements into one-dimensional feature vectors. That is, the step of "converting the first correlation coefficient matrix into a plurality of one-dimensional feature vectors to obtain the first feature set" may include:

dividing the first correlation coefficient matrix into two symmetrical regions, selecting one region from the two regions according to a preset strategy, converting elements in the selected region into one-dimensional eigenvectors, and obtaining a first feature set.

For example, as shown in fig. 4, if the first correlation coefficient matrix is shown in fig. 4, at this time, the first correlation coefficient matrix may be divided into two symmetric regions along a diagonal line, and then elements in the regions inside the dashed triangle in fig. 4 are converted into one-dimensional feature vectors, so as to obtain a first feature set, and so on.

(3) Determining the correlation among signals of different brain area channels in the second type of signal segment to obtain brain area connection information of the detection object under non-interference stimulation; for example, the following may be specifically mentioned:

and calculating the Pearson correlation coefficient among signals of different brain area channels in the second type of signal segment to obtain a second phase relation number, and generating brain area connection information of the detection object under non-interference stimulation according to the second phase relation number.

The formula for calculating the pearson correlation coefficient may specifically refer to the calculation of the first correlation coefficient in (2), which is not described herein again.

According to the above-mentioned way of calculating the pearson correlation coefficient r, for every two brain area channels, the pearson correlation coefficients, such as 128 pearson correlation coefficients, which are consistent with the signal number can be obtained, so that the correlation coefficients can be used to construct the corresponding correlation coefficient matrix for the subsequent processing; for convenience of description, in the embodiment of the present application, the correlation coefficient matrix constructed by using the second correlation number is referred to as a second correlation number matrix. Namely, the step of generating the brain region connection information of the detected object under the non-interference stimulation according to the second relation number may include:

and constructing a second correlation number matrix by using the second correlation numbers, and converting the second correlation number matrix into a plurality of one-dimensional eigenvectors to obtain a second feature set.

And the feature vectors in the second feature set are used for reflecting the brain region connection information of the detected object under the non-interference stimulation.

For example, taking an example that each data set includes 128 signals, at this time, a 128 × 128 second correlation matrix may be constructed, and then the second correlation matrix may be converted into a plurality of one-dimensional feature vectors to obtain a second feature set.

Similar to the vector stretching of the first correlation coefficient matrix, optionally, because the second correlation coefficient matrix is in a diagonal symmetric form, in order to improve the processing efficiency and reduce the consumption of computing resources, the second correlation coefficient matrix may be divided into two symmetric regions, and then, the vector stretching may be performed only on the element in one of the regions to convert the element therein into a one-dimensional feature vector. That is, the step of "converting the second correlation matrix into a plurality of one-dimensional feature vectors to obtain a second feature set" may include:

and dividing the second correlation matrix into two symmetrical regions, selecting one region from the two regions according to the preset strategy, and converting elements in the selected region into one-dimensional eigenvectors to obtain a second feature set.

For example, as shown in fig. 4, if the second correlation number matrix is shown in fig. 4, at this time, the second correlation number matrix may be divided into two symmetric regions along a diagonal line, and then elements in the regions inside the dashed triangle in fig. 4 are converted into one-dimensional feature vectors, so as to obtain a second feature set, and so on.

103. And calculating the similarity between the brain region connection information of the object to be detected under the interference stimulation and the brain region connection information under the non-interference stimulation, and then executing the step 104.

For example, if the feature vector reflects the brain region connection information of the object to be detected, at this time, a similarity between the feature vector in the first feature set and the feature vector in the second feature set, such as a Cosine similarity (Cosine similarity), may be specifically calculated, so as to obtain a similarity between the brain region connection information of the object to be detected under the interference stimulation and the brain region connection information of the object to be detected under the non-interference stimulation.

The formula for calculating the cosine similarity may be as follows:

wherein A is a first feature set aggregate, and B is a second feature set; a. the_iThe feature vectors in the first feature set; b is_iFor the feature vectors in the second feature set, n is the number of feature vectors in the first feature set (or may be the number of feature vectors in the second feature set, that is, the number of feature vectors in the first feature set is the same as the number of feature vectors in the second feature set).

104. And determining the pathological type of the brain wave signal according to the similarity.

The pathological type may be determined according to the requirement of the actual application, for example, anxiety, depression or obsessive-compulsive disorder, and of course, the brain wave signal may also be indicated as "abnormal", for example, non-anxiety, non-depression or non-obsessive-compulsive disorder.

For example, taking anxiety as an example, because the brain region connections of normal people and anxiety patients under the stimulation of the interfering signal may present different situations, that is, the similarity between the brain region connections of most normal people under the interfering stimulation and the brain region connections of non-interfering stimulation is low, and the brain region connections of anxiety patients under the interfering stimulation and the brain region connections of non-interfering stimulation do not present obvious difference (that is, the similarity is high), whether the object to be detected belongs to the high-focus group can be characterized by calculating the cosine similarity between the eigenvectors under the interfering stimulation and the eigenvectors under the non-interfering stimulation.

Specifically, the similarity calculated in step 103, such as the cosine similarity, may be input into a post-training detection model, and the pathology type to which the brain wave signal belongs may be predicted by using the post-training detection model. For example, the following may be specifically mentioned:

and constructing a cosine similarity matrix according to the cosine similarity obtained through calculation, and predicting the type of the cosine similarity matrix through a detection model after training to obtain the pathological type of the brain wave signal.

For example, if in step 103, the feature vectors in the first feature set and the second feature set are 128, 128 cosine similarities may be calculated, and then a cosine similarity matrix may be constructed based on the 128 cosine similarities, and then the type to which the cosine similarity matrix belongs may be predicted by the trained detection model, so as to obtain the pathological type to which the brain wave signal belongs.

The post-training detection model is formed by training a plurality of brain wave signal samples labeled with pathological types, namely, the post-training detection model can be obtained by a Machine Learning (ML) method.

Machine learning is a multi-field cross discipline and relates to a plurality of disciplines such as probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and the like. The special research on how a computer simulates or realizes the learning behavior of human beings so as to acquire new knowledge or skills and reorganize the existing knowledge structure to continuously improve the performance of the computer. Machine learning is the core of Artificial Intelligence (AI), and is the fundamental approach to make computers intelligent, which is applied across various fields of Artificial Intelligence. Machine learning and deep learning generally include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, and inductive learning. The artificial intelligence is a theory, a method, a technology and an application system which can simulate, extend and expand human intelligence by using a digital computer or a machine controlled by the digital computer, sense the environment, acquire knowledge and obtain the best result by using the knowledge. The artificial intelligence technology is a comprehensive subject and relates to the field of extensive technology, namely the technology of a hardware level and the technology of a software level. The artificial intelligence basic technology generally includes technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation/interaction systems, and mechatronics, which are not described herein again.

Optionally, the post-training detection model may be provided to the electroencephalogram signal detection apparatus after being trained by another device, or may be formed by self-training of the electroencephalogram signal detection apparatus, that is, before the step "predicting the type of pathology to which the electroencephalogram signal belongs by the post-training detection model according to the similarity", the method for detecting the electroencephalogram signal may further include steps S1 to S5, as follows:

and S1, collecting brain wave signal samples of a plurality of detection samples under the intermittent stimulation of the interference signals, wherein the detection samples are marked with pathological types.

For example, an electroencephalogram of a test sample under intermittent stimulation of an interference signal may be collected, and for convenience of description, in the embodiment of the present application, the electroencephalogram of the test sample is referred to as an electroencephalogram sample, wherein the electroencephalogram sample includes an electroencephalogram signal sample, and the electroencephalogram signal sample is labeled with a pathology type.

The electroencephalogram signal detection device may be configured to detect an electroencephalogram signal, and the electroencephalogram signal detection device may be configured to detect an electroencephalogram signal according to the electroencephalogram signal, and may further be configured to detect an electroencephalogram signal according to the electroencephalogram signal detection device.

Optionally, the detection sample may be any living body having a brain, such as a human, a monkey, a cat, or a dog, and may be specifically selected according to actual requirements; for example, if the detection object of the detection model after training is a human, "human" needs to be selected as the detection sample, and if the detection object of the detection model after training is a cat, "cat" needs to be selected as the detection sample.

In addition, in order to improve the training efficiency of the detection model, the detection samples can be selected in a targeted manner, for example, if the type of pathology to be detected is "anxiety disorder", a plurality of anxiety disorder patients can be selected as the detection samples, and of course, in order to improve the detection accuracy of the detection model, a plurality of normal persons can be selected as the negative samples in addition to a plurality of anxiety disorder patients. For another example, if the type of pathology to be detected is "depression", then a positive sample of a plurality of depression patients may be selected, and a negative sample of a plurality of normal persons may be selected.

And S2, respectively acquiring the brain area connection information of the detection sample under the interference stimulation and the brain area connection information under the non-interference stimulation according to the brain wave signal sample. For example, the following may be specifically mentioned:

A. the brain wave signal samples are divided into first type signal segment samples containing interfering stimuli and second type signal segment samples containing non-interfering stimuli.

For example, a label of each signal segment sample in the electroencephalogram signal samples is obtained, if the label indicates that the signal segment is a signal acquired under the stimulation of an interference signal, the signal segment is classified as a first type of signal segment sample, and if the label indicates that the signal segment is a signal acquired under the stimulation of a non-interference signal, the signal segment is classified as a second type of signal segment sample.

B. Determining the correlation among signals of different brain area channels in the first type of signal segment sample to obtain brain area connection information of the detection sample under interference stimulation; for example, the following may be specifically mentioned:

calculating a Pearson correlation coefficient between signals of different brain area channels in the first type of signal segment sample to obtain a first correlation coefficient sample, and generating brain area connection information of the detection sample under interference stimulation according to the first correlation coefficient sample; for example, a first correlation coefficient sample matrix may be specifically constructed by using first correlation coefficient samples, and then the first correlation coefficient sample matrix is converted into a plurality of one-dimensional feature vectors to obtain a first sample feature set.

And the feature vectors in the first sample feature set are used for reflecting the brain region connection information of the detection sample under the interference stimulation.

For a specific calculation method of the pearson correlation coefficient, reference may be made to the foregoing embodiments, which are not described herein again.

Optionally, since the first correlation coefficient sample matrix is in a diagonal symmetric form, in order to improve processing efficiency and reduce consumption of computational resources, the first correlation coefficient sample matrix may be divided into two symmetric regions, and then, only elements in one of the regions are vector-stretched to convert the elements into one-dimensional feature vectors. That is, the step of "converting the first correlation coefficient sample matrix into a plurality of one-dimensional feature vectors to obtain a first sample feature set" may include:

dividing the first correlation coefficient sample matrix into two symmetrical regions, selecting one region from the two regions according to a preset strategy, converting elements in the selected region into one-dimensional eigenvectors, and obtaining a first sample feature set.

C. Determining the correlation among signals of different brain area channels in the second type of signal segment sample to obtain brain area connection information of the detection sample under non-interference stimulation; for example, the following may be specifically mentioned:

and calculating a Pearson correlation coefficient between signals of different brain area channels in the second type of signal segment samples to obtain second phase relation number samples, and generating brain area connection information of the detection sample under non-interference stimulation according to the second phase relation number samples, for example, a second phase relation number sample matrix can be specifically constructed by using the second phase relation number samples, and the second phase relation number sample matrix is converted into a plurality of one-dimensional eigenvectors to obtain a second sample characteristic set.

And the feature vectors in the second sample feature set are used for reflecting the brain region connection information of the detection sample under the non-interference stimulation.

Optionally, because the second correlation coefficient sample matrix is in a diagonal symmetric form, in order to improve processing efficiency and reduce consumption of computing resources, the second correlation coefficient sample matrix may be divided into two symmetric regions, and then, only the element in one of the regions is vector-stretched to convert the element therein into a one-dimensional feature vector. That is, the step of "converting the second correlation coefficient sample matrix into a plurality of one-dimensional feature vectors to obtain a second sample feature set" may include:

and dividing the second correlation coefficient sample matrix into two symmetrical regions, selecting one region from the two regions according to a preset strategy, converting elements in the selected region into one-dimensional eigenvectors, and obtaining a second sample characteristic set.

Wherein, the steps B and C can be executed without any sequence.

And S3, calculating the similarity between the brain area connection information of the detection sample under the interference stimulation and the brain area connection information under the non-interference stimulation.

For example, cosine similarities between the feature vectors in the first sample feature set and the feature vectors in the second sample feature set may be specifically calculated, and the plurality of cosine similarities obtained through calculation are similarities between brain region connection information of the detection sample under interference stimulation and brain region connection information under non-interference stimulation.

The calculation formula of the cosine similarity may be as follows:

wherein C is the first sample feature set amount, and D is the second sample feature set; c_iIs a feature vector in the first sample feature set; d_iIs a feature vector in the second sample feature set, n is theThe number of feature vectors in a sample feature set (or the number of feature vectors in a second sample feature set, that is, the number of feature vectors in a first sample feature set is the same as the number of feature vectors in a second sample feature set).

S4, predicting the pathological type of the brain wave signal sample through a detection model according to the similarity; for example, the following may be specifically mentioned:

and constructing a cosine similarity matrix according to the cosine similarity obtained by calculation, and predicting the type of the cosine similarity matrix through a detection model to obtain the pathological type of the brain wave signal sample.

For example, as shown in fig. 5a, 5b, and 5c, the detection model may be a convolutional Neural Network, a Residual Neural Network (ResNet), or an image segmentation Network (VGG, Visual Geometry Group, that is, a convolutional Neural Network for performing image recognition and segmentation based on a Visual Geometry) Network, and the like.

FIG. 5a is an exemplary diagram of a convolutional neural network structure, which may include network layers such as a plurality of convolutional layers, pooling layers, and fully-connected layers; the number of convolutional layers, the size of the convolution kernel, the step size, the dimension size of each convolutional layer, and other parameters can be set according to the requirements of practical applications, for example, the dimension size of the first convolutional layer can be set to 64, the dimension size of the convolution kernel can be set to "7 × 7", the dimension sizes of the subsequent convolutional layers can be sequentially set to 64, 128, 256, 512, and the like, and the dimension sizes of the convolution kernels can be set to "3 × 3". Therefore, when the pathology type to which the brain wave signal sample belongs needs to be predicted, the cosine similarity matrix can be used as input and is led into the convolutional neural network, then each network layer of the convolutional neural network sequentially carries out convolution processing, pooling processing and full connection processing on the cosine similarity matrix, and finally the result output by the convolutional neural network is the predicted pathology type.

Similarly, fig. 5b is a structural example diagram of ResNet, fig. 5c is a structural example diagram of VGG network, when a pathology type to which brain wave signal samples belong needs to be predicted, a cosine similarity matrix can be taken as input and introduced into the ResNet or VGG network, and then the cosine similarity matrix is subjected to convolution processing, pooling processing and full connection processing by each network layer of the ResNet or VGG network in sequence, and finally, the result output by the ResNet or VGG network is the predicted pathology type, and other models are similar to this and will not be described herein again.

And S5, converging the detection model according to the marked pathology type and the predicted pathology type to obtain the trained detection model.

For example, the detection model may be converged according to the labeled pathology type and the predicted pathology type by using a preset loss function, so as to obtain a trained detection model.

That is, at this time, the network parameters of the detection model need to be adjusted, so that the labeled pathology type and the predicted pathology type can approach to each other indefinitely, and each time the adjustment is performed, it can be considered that the detection model is trained once (i.e., learning is completed once). By analogy, after a plurality of times of training (all the detection samples are processed through the steps S2-S5), the trained detection model can be obtained.

The loss function may be determined according to the requirements of the actual application, and for example, may specifically be a cross entropy (cross entropy) loss function, and the like.

105. And generating a detection result of the detection object based on the pathology type.

For example, a preset display template may be acquired, and then the pathology type is displayed according to the format of the display template, so as to obtain the detection result of the detection object.

Optionally, information such as the name, sex, age, address, contact information and/or occupation of the object to be detected can be displayed in the detection result.

As can be seen from the above, the embodiment may collect brain wave signals of an object to be detected under intermittent stimulation of an interference signal, respectively obtain brain region connection information of the object to be detected under the interference stimulation and brain region connection information under non-interference stimulation according to the brain wave signals, then calculate similarity between the brain region connection information under the interference stimulation and the brain region connection information under the non-interference stimulation, and determine a pathological type to which the brain wave signals belong according to the similarity, so as to generate a detection result of the object to be detected; because the scheme can respectively acquire the brain region connection information of the detection object under the interference stimulation and the brain region connection information of the detection object under the non-interference stimulation, and the corresponding pathological type is determined through the similarity between the two, the purpose of automatic interpretation is achieved, and therefore, compared with the existing scheme which can only rely on manual interpretation, the processing efficiency can be greatly improved, and because the scheme does not need to rely on manual interpretation, the situation of misjudgment caused by human factors can be avoided, and the detection accuracy is favorably improved.

The method according to the preceding embodiment is illustrated in further detail below by way of example.

In the present embodiment, the brain wave signal detection apparatus is specifically integrated into an electronic device, and the pathological types thereof including anxiety disorder and non-anxiety disorder will be described as an example.

And (I) training a detection model.

Firstly, the electronic equipment can acquire electroencephalograms of a plurality of detection samples under intermittent stimulation of interference signals, wherein the electroencephalograms comprise information such as electroencephalogram signal samples; for example, a plurality of common people and a plurality of anxiety patients may be selected as the detection samples, and then the detection samples are connected to the EEG signal electrodes to detect changes of the brain waves under the stimulation of the interference signals or the non-interference signals, and the brain waves are collected to obtain the electroencephalogram including the brain wave signal samples. For example, the specific detection procedure may be as follows:

setting a screen, presenting an interference signal on the screen, and requiring the detection sample to complete a corresponding task when the detection sample is viewed to have the interference signal, for example, referring to fig. 6, the detection sample may be firstly watched at the central point 800m of fig. 6(a), then the screen displays the interference signal once, and the detection sample needs to judge whether the line segment in the diamond shape in fig. 6(b) is horizontal or vertical when the interference signal appears. After the interference signal is over, the screen content is switched to fig. 6(c), the detection sample continuously annotates the center point of fig. 6(c) (the gazing time is about 1600- & 2000ms), waits for the next interference signal to appear, and so on until the detection is finished.

Optionally, in addition to displaying the interference signal intermittently, the screen may also present the interference signal and the non-interference signal according to a preset policy, for example, referring to fig. 7, the interference signal as in fig. 7(a) may be displayed, and it is required that the detection sample needs to determine whether a line segment in a diamond as in fig. 6(b) is horizontal or vertical when the interference signal occurs, after the interference signal is ended, the screen content is switched to fig. 7(b), the detection sample continues to watch the screen, waits for the occurrence of the next interference signal, and so on until the detection is ended.

It should be noted that, in addition to the electroencephalograms including the brain wave signal samples, corresponding labels are recorded to indicate whether the brain wave signal samples are the signals acquired under the stimulation of the interference signals or the signals acquired under the stimulation of the non-interference signals, for example, referring to fig. 3, in the electroencephalograms, c342, Y88, d464 and N99 are the labels of the signals acquired under the stimulation of the interference signals.

After the electroencephalograms of the detection samples under the intermittent stimulation of the interference signals are collected, the pathological types of the electroencephalograms can be labeled, for example, if the detection samples are anxiety patients, the pathological types are labeled as anxiety, if the detection samples are normal people, the pathological types are labeled as non-anxiety, and the like, and then, the electroencephalograms labeled with the pathological types can be used as training samples of the detection model, namely, the electroencephalogram samples.

Secondly, after obtaining the electroencephalogram samples, as shown in fig. 8, the electronic equipment may divide the electroencephalogram signal samples in the electroencephalogram samples into signal segment samples of a first type containing interfering stimuli and signal segment samples of a second type containing non-interfering stimuli; for example, the electronic device may obtain a label of each signal segment sample in the brain wave signal samples, classify the signal segment into a first type of signal segment sample if the label indicates that the signal segment is a signal acquired under the stimulation of an interference signal, classify the signal segment into a second type of signal segment sample if the label indicates that the signal segment is a signal acquired under the stimulation of a non-interference signal, and so on.

After dividing the electroencephalogram signal samples in the electroencephalogram samples into the first type of signal segment samples and the second type of signal segment samples, the electronic device can determine the brain region connection information of the detection samples under the interference stimulation and the brain region connection information of the detection samples under the non-interference stimulation according to the first type of signal segment samples and the second type of signal segment samples. For example, the following may be specifically mentioned:

A) interfering with brain region connectivity information under stimulation;

the electronic device calculates a pearson correlation coefficient (see the previous embodiment) between signals of different brain area channels in the first type of signal segment sample to obtain a first correlation coefficient sample, then constructs a first correlation coefficient sample matrix by using the first correlation coefficient sample, and converts the first correlation coefficient sample matrix into a plurality of one-dimensional eigenvectors to obtain a first sample feature set, wherein the eigenvectors in the first sample feature set are used for reflecting brain area connection information of the detection sample under the interference stimulation.

B) Brain region connectivity information under non-interfering stimulation;

the electronic device calculates the pearson correlation coefficients between signals of different brain area channels in the second type of signal segment samples (see the previous embodiment), obtains a second phase relation number sample, then constructs a second phase relation number sample matrix by using the second phase relation number sample, and converts the second phase relation number sample matrix into a plurality of one-dimensional eigenvectors, so as to obtain a second sample feature set, wherein the eigenvectors in the second sample feature set are used for reflecting brain area connection information of the detection sample under non-interference stimulation.

Optionally, since the correlation coefficient matrices (such as the first correlation coefficient sample matrix or the second correlation coefficient sample matrix) are in a diagonal symmetric form, in order to improve processing efficiency and reduce consumption of computing resources, the correlation coefficient matrices may be divided into two symmetric regions, and then, only elements of one of the regions are vector-stretched to convert the elements thereof into one-dimensional feature vectors.

It should be noted that, for convenience of description, only one schematic diagram of the correlation coefficient matrix and the one-dimensional feature vector is given in fig. 8, and it should be understood that fig. 8 is merely an example, the correlation coefficient matrix in fig. 8 may include the first correlation coefficient sample matrix and the second correlation coefficient sample matrix, and the one-dimensional feature vector in fig. 8 may also include the feature vector in the first sample feature set and the feature vector in the second sample feature set, which is not described herein again.

Thereafter, the electronic device may calculate cosine similarities between the feature vectors in the first sample feature set and the feature vectors in the second sample feature set, and construct a two-dimensional similarity matrix (one dimension is "interfering stimulation" and the other dimension is "non-interfering stimulation") based on the cosine similarities, so as to obtain the cosine similarity matrix, for example, see fig. 8.

Because the brain area connections of normal people and anxiety neurosis patients under the stimulation of the interference signals may present different conditions, that is, the cosine similarity between the brain area connections of most normal people under the interference stimulation and the brain area connections of non-interference stimulation is low, and the cosine similarity between the brain area connections of anxiety neurosis patients under the interference stimulation and the brain area connections of non-interference stimulation is high, for example, see fig. 9, machine learning can be performed on a preset detection model by using the cosine similarity matrixes, and a post-training detection model for detecting whether brain wave signals belong to anxiety neurosis is obtained.

For example, after the electronic device obtains the cosine similarity matrix of the current brain wave signal sample (i.e., the matrix constructed by the cosine similarity of the brain area connection under the interference stimulation and the brain area connection under the non-interference stimulation), the type to which the cosine similarity matrix belongs may be predicted by the detection model to obtain the pathological type to which the brain wave signal sample belongs, then the detection model may be converged according to the labeled pathological type and the predicted pathological type, after the convergence is completed, the current brain wave signal sample may be updated to another brain wave signal sample, and similarly, after the cosine similarity of the current brain wave signal sample is obtained and the cosine similarity matrix is constructed, the type to which the cosine similarity matrix belongs may be predicted by the detection model to obtain the pathological type to which the current brain wave signal sample belongs, then, converging the detection model according to the marked pathology type and the predicted pathology type, and repeating the steps until all brain wave signal samples are trained, and obtaining the trained detection model

And secondly, the brain wave signals of the object to be detected can be detected through the trained detection model.

For example, as shown in fig. 10, the specific flow of the text recognition method may be as follows:

201. the electronic equipment acquires an electroencephalogram of an object to be detected under intermittent stimulation of interference signals, wherein the electroencephalogram comprises brain wave signals.

For example, taking the object to be detected as zhang, the following may be specifically used:

a screen may be set to present the interference signal on the screen, and ask zhang to complete the corresponding task when observing the occurrence of the interference signal, for example, zhang san may first note the central point 800m in fig. 6(a), and then the screen displays the interference signal once, and zhang san needs to determine whether the line segment in the diamond shape in fig. 6(b) is horizontal or vertical when the interference signal occurs. After the interference signal is ended, the screen content is switched to fig. 6(c), zhang san continues to note the central point in fig. 6(c), waits for the next interference signal to appear, and so on until the detection is finished.

In the detection process, the electroencephalogram of Zhang III under the intermittent stimulation of the interference signal can be obtained by collecting the electroencephalogram generated by Zhang III.

Optionally, in addition to recording the brain wave signals, a corresponding label may be recorded in the electroencephalogram to indicate whether the section of brain wave signals is the signals acquired under the stimulation of the interfering signals or the signals acquired under the stimulation of the non-interfering signals; the labeling manner of the label may be specifically determined according to the requirements of the actual application, and is not described herein again.

202. The electronics divide the brain wave signals in the electroencephalogram into a first type of signal segment containing interfering stimuli and a second type of signal segment containing non-interfering stimuli.

For example, as shown in fig. 11, if the acquired electroencephalogram records a corresponding label, at this time, the electronic device may acquire a label of each signal segment in the electroencephalogram signal, classify the signal segment into a first type of signal segment if the label indicates that the signal segment is a signal acquired under the stimulation of an interfering signal, classify the signal segment into a second type of signal segment if the label indicates that the signal segment is a signal acquired under the stimulation of a non-interfering signal, and so on.

203. The electronic device calculates the pearson correlation coefficients between the signals of the different brain area channels in the first type of signal segment to obtain first correlation coefficients, and then executes step 204.

For example, taking the example that the signals of two different brain region channels respectively correspond to two data sets x and y, the calculation formula of the pearson correlation coefficient r between the signals of the two different brain region channels (i.e. the data set x and the data set y) can be as follows:

204. The electronic equipment constructs a first correlation coefficient matrix by using the first correlation coefficient, and converts the first correlation coefficient matrix into a plurality of one-dimensional eigenvectors to obtain a first feature set.

Optionally, as shown in fig. 11, since the first correlation coefficient matrix is in a diagonal symmetric form, in order to improve processing efficiency and reduce consumption of computing resources, the first correlation coefficient matrix may be divided into two symmetric regions, and then only elements in one of the regions (e.g., the region within the dashed triangle in fig. 11) are vector-stretched to convert the elements therein into one-dimensional feature vectors, which is detailed in the foregoing embodiments and not repeated herein.

It should be noted that, for convenience of description, only one schematic diagram of the correlation coefficient matrix and the one-dimensional feature vector is given in fig. 11, and it should be understood that the correlation coefficient matrix in fig. 11 may include the first correlation coefficient sample matrix and the second correlation coefficient sample matrix, and the one-dimensional feature vector in fig. 11 may also include the feature vector in the first sample feature set and the feature vector in the second sample feature set, which is not described herein again.

205. The electronic device calculates the pearson correlation coefficients between the signals of the different brain area channels in the second type of signal segment to obtain a second correlation coefficient, and then executes step 206.

The specific calculation method of the pearson correlation coefficient may be referred to in step 203, which is not described herein.

It should be noted that the execution order of

steps

203 and 205 may not be sequential.

206. And the electronic equipment constructs a second correlation number matrix by using the second correlation number, and converts the second correlation number matrix into a plurality of one-dimensional eigenvectors to obtain a second feature set.

Similar to the vector stretching of the first correlation coefficient matrix, optionally, as shown in fig. 11, since the second correlation coefficient matrix is in a diagonal symmetric form, in order to improve processing efficiency and reduce consumption of computing resources, the second correlation coefficient matrix may be divided into two symmetric regions, and then, only the elements in one of the regions (e.g., the region in the dashed triangle in fig. 11) are vector stretched to convert the elements therein into one-dimensional feature vectors, which may be specifically referred to the foregoing embodiment and is not described herein again.

207. The electronic device calculates cosine similarity between the feature vectors in the first feature set and the feature vectors in the second feature set. The formula for calculating the cosine similarity may be as follows:

208. And the electronic equipment constructs a cosine similarity matrix according to the cosine similarity obtained by calculation, and predicts the type of the cosine similarity matrix through a detection model after training to obtain the pathological type of the brain wave signal.

For example, if the feature vectors in the first feature set and the second feature set are 128, 128 cosine similarities may be calculated at this time, and then a cosine similarity matrix may be constructed based on the 128 cosine similarities, and then the type to which the cosine similarity matrix belongs is predicted by the post-training detection model, so as to obtain the pathological type to which the brain wave signal belongs.

For example, taking the zhangsan brain wave signal as an example, after the cosine similarity matrix is constructed, the cosine similarity matrix may be imported into the post-training detection model, the post-training detection model performs feature extraction on the cosine similarity matrix, and identifies the extracted features, and if the identification result meets the "anxiety disorder" feature, the pathological type of the zhangsan brain wave signal is determined to be "anxiety disorder"; otherwise, if the recognition result does not accord with the anxiety disorder feature, the pathological type of the Zhang III brain wave signal is determined to be non-anxiety disorder.

209. The electronic device generates a detection result of the detection object based on the pathology type.

Therefore, according to the electroencephalogram detection method and the electroencephalogram detection device, a large number of electroencephalogram samples marked with pathological types are collected, when anxiety patients and normal people receive interference information stimulation, different brain area connection conditions are used for training the detection models, and the pathological types of the brain wave signals of the objects to be detected are identified by the trained detection models, so that compared with the existing scheme which can only rely on manual interpretation, the efficiency of anxiety detection can be greatly improved, and in addition, the scheme does not need to rely on manual interpretation, therefore, the condition of misjudgment caused by human factors can be avoided, and the accuracy of detection is favorably improved.

In order to better implement the above method, embodiments of the present invention also provide a brain wave signal detecting apparatus, which may be specifically integrated in an electronic device such as a server or a terminal.

For example, as shown in fig. 12, the brain wave signal detecting apparatus may include an acquisition unit 301, an acquisition unit 302, a calculation unit 303, a determination unit 304, and a generation unit 305 as follows:

(1) an acquisition unit 301;

the acquisition unit 301 is configured to acquire brain wave signals of an object to be detected under intermittent stimulation of the interference signals.

For example, the acquisition unit 301 may be specifically configured to acquire an electroencephalogram of the object to be detected under intermittent stimulation of the interference signal, where the electroencephalogram includes an electroencephalogram signal.

(2) An acquisition unit 302;

an obtaining unit 302, configured to obtain brain region connection information of the detection object under the interfering stimulation and brain region connection information under the non-interfering stimulation respectively according to the brain wave signal.

For example, the obtaining unit 302 may include a dividing subunit, a first determining subunit, and a second determining subunit, as follows:

and the dividing subunit is used for dividing the brain wave signals into a first type of signal segments containing interference stimulation and a second type of signal segments containing non-interference stimulation.

The first determining subunit is used for determining the correlation among the signals of different brain area channels in the first signal segment to obtain the brain area connection information of the detection object under the interference stimulation;

For example, when recording the brain wave signals in the electroencephalogram, a corresponding tag may be recorded to indicate whether the section of brain wave signals is the signals collected under the stimulation of the interfering signals or the signals collected under the stimulation of the non-interfering signals, and then the time-division unit may divide the brain wave signals based on the tag, that is:

the dividing subunit is specifically configured to acquire a label of each signal segment in the brain wave signal, and if the label indicates that the signal segment is a signal acquired under the stimulation of an interference signal, classify the signal segment into a first type of signal segment; and if the label indicates that the signal segment is the signal acquired under the stimulation of the non-interference signal, classifying the signal segment into a second type of signal segment.

Alternatively, there are various ways to determine the correlation between signals of different brain region channels, for example, the following may be used:

the first determining subunit is specifically configured to calculate a pearson correlation coefficient between signals of different brain area channels in the first-class signal segment to obtain a first correlation coefficient, and generate brain area connection information of the detection object under the interference stimulation according to the first correlation coefficient; for example, a first correlation coefficient matrix may be specifically constructed by using the first correlation coefficient, and the first correlation coefficient matrix is converted into a plurality of one-dimensional feature vectors to obtain a first feature set.

The second determining subunit is specifically configured to calculate a pearson correlation coefficient between signals of different brain area channels in the second type of signal segment to obtain a second phase relation number, and generate brain area connection information of the detection object under non-interference stimulation according to the second phase relation number; for example, a second relation matrix may be specifically constructed by using the second relation, the second relation matrix is converted into a plurality of one-dimensional feature vectors to obtain a second feature set, and the feature vectors in the second feature set are used for reflecting brain region connection information of the detection object under non-interfering stimulation

The feature vectors in the first feature set are used for reflecting the brain region connection information of the detection object under the interference stimulation, and the feature vectors in the second feature set are used for reflecting the brain region connection information of the detection object under the non-interference stimulation.

Optionally, since the correlation coefficient matrix (the first correlation coefficient matrix and the second correlation coefficient matrix) is in a diagonal symmetric form, in order to improve processing efficiency and reduce consumption of computing resources, the correlation coefficient matrix may be divided into two symmetric regions, and then, only the element in one of the regions is vector-stretched to convert the element therein into a one-dimensional feature vector. Namely:

the first determining subunit may be specifically configured to divide the first correlation coefficient matrix into two symmetric regions, select one region from the two regions according to a preset policy, and convert elements in the selected region into a one-dimensional feature vector to obtain a first feature set.

(3) A calculation unit 303;

a calculating unit 303, configured to calculate a similarity between the brain region connection information under the interference stimulation and the brain region connection information under the non-interference stimulation;

for example, the calculating unit 303 may be specifically configured to calculate a cosine similarity between the feature vector in the first feature set and the feature vector in the second feature set.

(4) A determination unit 304;

a determining unit 304 for determining the type of pathology to which the brain wave signal belongs according to the similarity.

For example, the determining unit 304 is specifically configured to predict the pathology type to which the brain wave signal belongs through the post-training detection model according to the similarity, for example, the pathology type to which the brain wave signal belongs may be predicted through the post-training detection model according to the cosine similarity, which is specifically as follows:

the determining unit 304 may be specifically configured to construct a cosine similarity matrix according to the calculated cosine similarity, and predict a type to which the cosine similarity matrix belongs through a detection model after training, so as to obtain a pathological type to which the brain wave signal belongs.

(5) A generation unit 305;

a generating unit 305 for generating a detection result of the detection object based on the type of pathology.

For example, the generating unit 305 may be specifically configured to obtain a preset display template, and then display the pathology type according to a format of the display template, so as to obtain a detection result of the detection object.

The trained detection model is formed by training a plurality of brain wave signal samples marked with pathological types. The post-training detection model may be provided to the electroencephalogram signal detection apparatus after being trained by other equipment, or may be trained by the electroencephalogram signal detection apparatus itself, that is, as shown in fig. 13, the electroencephalogram signal detection apparatus may further include a training unit 306, as follows:

the collecting unit 301 may further be configured to collect brain wave signal samples of a plurality of detection samples under intermittent stimulation of an interference signal, where the brain wave signal samples are labeled with pathological types.

The obtaining unit 302 may be further configured to obtain brain region connection information of the detection sample under the interfering stimulation and brain region connection information of the detection sample under the non-interfering stimulation respectively according to the brain wave signal samples.

For example, the obtaining unit 302 may be specifically configured to divide the brain wave signal samples into a first type of signal segment samples including interference stimulation and a second type of signal segment samples including non-interference stimulation, determine correlations between signals of different brain area channels in the first type of signal segment samples, obtain brain area connection information of the detection sample under the interference stimulation, and determine correlations between signals of different brain area channels in the second type of signal segment samples, and obtain brain area connection information of the detection sample under the non-interference stimulation. For example, the following may be specifically mentioned:

calculating a Pearson correlation coefficient between signals of different brain area channels in a first type of signal segment sample to obtain a first correlation coefficient sample, constructing a first correlation coefficient sample matrix by using the first correlation coefficient sample, converting the first correlation coefficient sample matrix into a plurality of one-dimensional eigenvectors to obtain a first sample characteristic set, wherein the eigenvectors in the first sample characteristic set are used for reflecting brain area connection information of the detection sample under interference stimulation; and calculating a Pearson correlation coefficient between signals of different brain area channels in the second type of signal segment samples to obtain a second phase relation number sample, constructing a second phase relation number sample matrix by using the second phase relation number sample, converting the second phase relation number sample matrix into a plurality of one-dimensional eigenvectors to obtain a second sample characteristic set, wherein the eigenvectors in the second sample characteristic set are used for reflecting brain area connection information of the detection sample under non-interference stimulation.

The calculating unit 303 may be further configured to calculate a similarity between the brain region connection information of the detection sample under the interference stimulation and the brain region connection information under the non-interference stimulation; for example, the cosine similarity between the feature vector in the first sample feature set and the feature vector in the second sample feature set may be calculated, and the cosine similarity may be used as the similarity between the brain region connection information of the sample to be detected under the interference stimulation and the brain region connection information under the non-interference stimulation.

The training unit 306 is configured to predict a pathology type to which the brain wave signal sample belongs through a detection model according to the similarity, and converge the detection model according to the labeled pathology type and the predicted pathology type to obtain a trained detection model.

For example, the training unit 306 may be specifically configured to construct a cosine similarity matrix according to the calculated cosine similarity, predict a type to which the cosine similarity matrix belongs through a detection model to obtain a pathology type to which the brain wave signal sample belongs, and then converge the detection model according to the labeled pathology type and the predicted pathology type to obtain a post-training detection model.

In a specific implementation, the above units may be implemented as independent entities, or may be combined arbitrarily to be implemented as the same or several entities, and the specific implementation of the above units may refer to the foregoing method embodiments, which are not described herein again.

As can be seen from the above, in this embodiment, the acquisition unit 301 may acquire the brain wave signal of the object to be detected under the intermittent stimulation of the interference signal, the acquisition unit 302 then respectively acquires the brain region connection information of the object to be detected under the interference stimulation and the brain region connection information under the non-interference stimulation according to the brain wave signal, the calculation unit 303 calculates the similarity between the brain region connection information under the interference stimulation and the brain region connection information under the non-interference stimulation, and then the determination unit 304 determines the pathology type to which the brain wave signal belongs according to the similarity, so that the generation unit 305 generates the detection result of the object to be detected based on the pathology type; because the scheme can respectively acquire the brain region connection information of the detection object under the interference stimulation and the brain region connection information of the detection object under the non-interference stimulation, and the corresponding pathological type is determined through the similarity between the two, the purpose of automatic interpretation is achieved, and therefore, compared with the existing scheme which can only rely on manual interpretation, the processing efficiency can be greatly improved, and because the scheme does not need to rely on manual interpretation, the situation of misjudgment caused by human factors can be avoided, and the detection accuracy is favorably improved.

The embodiment of the present invention further provides an electronic device, which may be integrated with any one of the brain wave signal detection apparatuses provided in the embodiments of the present invention, and the electronic device may be a server, or a terminal, such as a PC, a tablet computer, or a notebook computer, or may be an intelligent medical device.

For example, as shown in fig. 14, a schematic structural diagram of an electronic device according to an embodiment of the present application is shown, specifically:

the electronic device may include components such as a processor 401 of one or more processing cores, memory 402 of one or more computer-readable storage media, a power supply 403, and an input unit 404. Those skilled in the art will appreciate that the electronic device configuration shown in fig. 14 does not constitute a limitation of the electronic device and may include more or fewer components than those shown, or some components may be combined, or a different arrangement of components. Wherein:

the processor 401 is a control center of the electronic device, connects various parts of the whole electronic device by various interfaces and lines, performs various functions of the electronic device and processes data by running or executing software programs and/or modules stored in the memory 402 and calling data stored in the memory 402, thereby performing overall monitoring of the electronic device. Optionally, processor 401 may include one or more processing cores; preferably, the processor 401 may integrate an application processor, which mainly handles operating systems, user interfaces, application programs, etc., and a modem processor, which mainly handles wireless communications. It will be appreciated that the modem processor described above may not be integrated into the processor 401.

The memory 402 may be used to store software programs and modules, and the processor 401 executes various functional applications and data processing by operating the software programs and modules stored in the memory 402. The memory 402 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required by at least one function (such as a sound playing function, an image playing function, etc.), and the like; the storage data area may store data created according to use of the electronic device, and the like. Further, the memory 402 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid state storage device. Accordingly, the memory 402 may also include a memory controller to provide the processor 401 access to the memory 402.

The electronic device further comprises a power supply 403 for supplying power to the various components, and preferably, the power supply 403 is logically connected to the processor 401 through a power management system, so that functions of managing charging, discharging, and power consumption are realized through the power management system. The power supply 403 may also include any component of one or more dc or ac power sources, recharging systems, power failure detection circuitry, power converters or inverters, power status indicators, and the like.

The electronic device may further include an input unit 404, and the input unit 404 may be used to receive input numeric or character information and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function control.

Although not shown, the electronic device may further include a display unit and the like, which are not described in detail herein. Specifically, in this embodiment, the processor 401 in the electronic device loads the executable file corresponding to the process of one or more application programs into the memory 402 according to the following instructions, and the processor 401 runs the application program stored in the memory 402, thereby implementing various functions as follows:

the method comprises the steps of collecting brain wave signals of an object to be detected under intermittent stimulation of interference signals, respectively obtaining brain area connection information of the object to be detected under the interference stimulation and brain area connection information under non-interference stimulation according to the brain wave signals, calculating similarity between the brain area connection information under the interference stimulation and the brain area connection information under the non-interference stimulation, determining the pathological type of the brain wave signals according to the similarity, and generating a detection result of the object to be detected based on the pathological type.

For example, the electroencephalogram signals may be specifically divided into a first type of signal segment including interference stimulation and a second type of signal segment including non-interference stimulation, the correlation between signals of different brain area channels in the first type of signal segment is determined, brain area connection information of the detection object under the interference stimulation is obtained, and the correlation between signals of different brain area channels in the second type of signal segment is determined, brain area connection information of the detection object under the non-interference stimulation is obtained; then, the similarity between the brain region connection information under the interference stimulation and the brain region connection information under the non-interference stimulation is calculated, the pathology type to which the brain wave signal belongs of the trained detection model is used for prediction according to the similarity, and the detection result of the detection object is generated based on the predicted pathology type.

The trained detection model is formed by training a plurality of brain wave signal samples marked with pathological types. Optionally, the trained detection model may be provided to the electronic device after being trained by other devices, or may be trained by the electronic device itself, that is, the processor 401 may also run an application program stored in the memory 402, so as to implement the following functions:

the method comprises the steps of collecting brain wave signal samples of a plurality of detection samples under intermittent stimulation of interference signals, wherein the detection samples are marked with pathological types, respectively obtaining brain area connection information of the detection samples under the interference stimulation and brain area connection information under the non-interference stimulation according to the brain wave signal samples, calculating the similarity between the brain area connection information of the detection samples under the interference stimulation and the brain area connection information under the non-interference stimulation, predicting the pathological type to which the brain wave signal samples belong through a detection model according to the similarity, converging the detection model according to the marked pathological type and the predicted pathological type, and obtaining the detection model after training.

The above operations can be implemented in the foregoing embodiments, and are not described in detail herein.

As can be seen from the above, the electronic device of this embodiment may collect brain wave signals of an object to be detected under intermittent stimulation of an interference signal, respectively obtain brain region connection information of the object to be detected under the interference stimulation and brain region connection information under non-interference stimulation according to the brain wave signals, then calculate similarity between the brain region connection information under the interference stimulation and the brain region connection information under the non-interference stimulation, and determine a pathological type to which the brain wave signals belong according to the similarity, so as to generate a detection result of the object to be detected; because the scheme can respectively acquire the brain region connection information of the detection object under the interference stimulation and the brain region connection information of the detection object under the non-interference stimulation, and the corresponding pathological type is determined through the similarity between the two, the purpose of automatic interpretation is achieved, and therefore, compared with the existing scheme which can only rely on manual interpretation, the processing efficiency can be greatly improved, and because the scheme does not need to rely on manual interpretation, the situation of misjudgment caused by human factors can be avoided, and the detection accuracy is favorably improved.

It will be understood by those skilled in the art that all or part of the steps of the methods of the above embodiments may be performed by instructions or by associated hardware controlled by the instructions, which may be stored in a computer readable storage medium and loaded and executed by a processor.

To this end, the embodiments of the present application provide a storage medium, in which a plurality of instructions are stored, and the instructions can be loaded by a processor to execute the steps in any one of the brain wave signal detection methods provided by the embodiments of the present application. For example, the instructions may perform the steps of:

The detection model after training is formed by training a plurality of brain wave signal samples marked with pathological types, namely, optionally, the instruction can also execute the following steps:

Wherein the storage medium may include: read Only Memory (ROM), Random Access Memory (RAM), magnetic or optical disks, and the like.

Since the instructions stored in the storage medium can execute the steps in any one of the brain wave signal detection methods provided in the embodiments of the present application, the beneficial effects that can be achieved by any one of the brain wave signal detection methods provided in the embodiments of the present application can be achieved, and detailed descriptions thereof are omitted here.

The brain wave signal detection method, the related devices and the storage medium provided by the embodiments of the present application are described in detail above, and the principle and the implementation of the present invention are explained in this document by applying specific examples, and the description of the above embodiments is only used to help understanding the method and the core concept of the present invention; meanwhile, for those skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.

Claims

1. A brain wave signal detection method is characterized by comprising the following steps:

dividing the brain wave signals into a first type of signal segments containing interfering stimuli and a second type of signal segments containing non-interfering stimuli;

calculating a Pearson correlation coefficient between signals of different brain area channels in a first type of signal segment to obtain a first correlation coefficient, constructing a first correlation coefficient matrix by using the first correlation coefficient, dividing the first correlation coefficient matrix into two symmetrical areas, selecting one area from the two areas according to a preset strategy, converting elements in the selected area into a one-dimensional eigenvector to obtain a first feature set, wherein the eigenvector in the first feature set is used for reflecting brain area connection information of the detection object under interference stimulation; wherein the first correlation coefficient matrix is in a diagonal symmetric form;

calculating a Pearson correlation coefficient between signals of different brain area channels in a second type of signal segment to obtain a second phase relation number, constructing a second phase relation number matrix by using the second phase relation number, dividing the second phase relation number matrix into two symmetrical areas, selecting one area from the two areas according to a preset strategy, converting elements in the selected area into one-dimensional eigenvectors to obtain a second feature set, wherein the eigenvectors in the second feature set are used for reflecting brain area connection information of the detection object under non-interference stimulation; wherein the second correlation matrix is in a diagonal symmetric form;

calculating the similarity between the characteristic vector corresponding to the brain region connection information under the interference stimulation and the characteristic vector corresponding to the brain region connection information under the non-interference stimulation;

determining the pathological type to which the brain wave signals belong according to the similarity, wherein the pathological type indicates whether the brain wave signals are abnormal or not;

and displaying the pathological type by using a preset display template format to obtain a detection result of the detection object.

2. The method according to claim 1, wherein the determining the type of pathology to which the brain wave signals belong according to the similarity includes:

and predicting the pathological type of the brain wave signal through a trained detection model according to the similarity, wherein the trained detection model is formed by training a plurality of brain wave signal samples marked with the pathological type.

3. The method according to claim 1, wherein the dividing the brain wave signals into signal segments of a first type containing interfering stimuli and signal segments of a second type containing non-interfering stimuli comprises:

acquiring a label of each signal segment in the brain wave signal;

if the label indication signal segment is a signal acquired under the stimulation of an interference signal, classifying the signal segment into a first type signal segment;

and if the label indication signal segment is the signal acquired under the stimulation of the non-interference signal, classifying the signal segment into a second type signal segment.

4. The method according to claim 2, wherein the calculating the similarity between the brain region connectivity information under the interferential stimulation and the brain region connectivity information under the non-interferential stimulation comprises:

calculating cosine similarity between the feature vectors in the first feature set and the feature vectors in the second feature set;

the predicting the pathology type to which the brain wave signal belongs through a trained detection model according to the similarity includes: and predicting the pathological type of the brain wave signal through a detection model after training according to the cosine similarity.

5. The method according to any one of claims 2 or 4, wherein before predicting the type of pathology to which the brain wave signals belong by the post-training detection model according to the similarity, the method further comprises:

collecting brain wave signal samples of a plurality of detection samples under intermittent stimulation of interference signals, wherein the brain wave signal samples are marked with pathological types;

respectively acquiring brain area connection information of the detection sample under interference stimulation and brain area connection information under non-interference stimulation according to the brain wave signal sample;

calculating the similarity between the brain area connection information of the detection sample under the interference stimulation and the brain area connection information under the non-interference stimulation;

predicting the pathological type of the brain wave signal sample through a detection model according to the similarity;

and converging the detection model according to the marked pathology type and the predicted pathology type to obtain the trained detection model.

6. A brain wave signal detection device is characterized by comprising a collecting unit, an acquiring unit, a calculating unit, a determining unit and a generating unit, wherein the acquiring unit comprises a dividing subunit, a first determining subunit and a second determining subunit;

the acquisition unit is used for acquiring brain wave signals of the object to be detected under the intermittent stimulation of the interference signals;

the first determining subunit is configured to calculate a pearson correlation coefficient between signals of different brain area channels in a first-class signal segment to obtain a first correlation coefficient, construct a first correlation coefficient matrix using the first correlation coefficient, divide the first correlation coefficient matrix into two symmetric regions, select one region from the two regions according to a preset strategy, convert elements in the selected region into a one-dimensional feature vector, and obtain a first feature set, where the feature vector in the first feature set is used to reflect brain area connection information of the detection object under the interference stimulation; wherein the first correlation coefficient matrix is in a diagonal symmetric form;

the second determining subunit is configured to calculate a pearson correlation coefficient between signals of different brain area channels in a second type of signal segment to obtain a second phase relation number, construct a second phase relation number matrix by using the second phase relation number, divide the second phase relation number matrix into two symmetrical regions, select one region from the two regions according to the preset strategy, convert elements in the selected region into a one-dimensional feature vector, and obtain a second feature set, where the feature vector in the second feature set is used to reflect brain area connection information of the detection object under non-interference stimulation; wherein the second correlation matrix is in a diagonal symmetric form;

the calculation unit is used for calculating the similarity between the feature vector corresponding to the brain region connection information under the interference stimulation and the feature vector corresponding to the brain region connection information under the non-interference stimulation;

the determining unit is used for determining the pathological type to which the brain wave signals belong according to the similarity, and the pathological type indicates whether the brain wave signals are abnormal or not;

the generation unit is used for generating a detection result of the detection object based on the pathology type.

7. The brain wave signal detecting device according to claim 6, wherein the determining unit is configured to predict a pathology type to which the brain wave signal belongs by a post-training detection model that is trained from a plurality of brain wave signal samples to which pathology types are labeled, according to the similarity.

8. The brain wave signal detecting device according to claim 6, wherein the dividing unit is configured to acquire a label of each signal segment in the brain wave signal; if the label indication signal segment is a signal acquired under the stimulation of an interference signal, classifying the signal segment into a first type signal segment; and if the label indication signal segment is the signal acquired under the stimulation of the non-interference signal, classifying the signal segment into a second type signal segment.

9. The brain wave signal detection device according to claim 7, wherein the calculation unit is configured to calculate a cosine similarity between the feature vectors in the first feature set and the feature vectors in the second feature set;

and the generating unit is used for predicting the pathological type of the brain wave signal through a detection model after training according to the cosine similarity.

10. The brain wave signal detection device according to claim 7 or 9, wherein the post-training detection model is obtained by:

11. A storage medium storing a plurality of instructions adapted to be loaded by a processor to perform the steps of the brain wave signal detecting method according to any one of claims 1 to 5.