CN110859616A

CN110859616A - Cognitive assessment method, device and equipment of object and storage medium

Info

Publication number: CN110859616A
Application number: CN201911274298.4A
Authority: CN
Inventors: 储银雪; 丁悦; 李鑫; 凌震华; 李云霞
Original assignee: iFlytek Co Ltd
Current assignee: iFlytek Co Ltd
Priority date: 2019-12-12
Filing date: 2019-12-12
Publication date: 2020-03-06

Abstract

The application provides a cognitive assessment method, a cognitive assessment device, a cognitive assessment equipment and a readable storage medium of an object, wherein the method comprises the following steps: firstly, obtaining a target electroencephalogram of a target object, then segmenting the target electroencephalogram according to a specified length to obtain a target signal segment set consisting of target signal segments with specified lengths, then determining frequency domain characteristics corresponding to the target signal segments in the target signal segment set, and finally evaluating the cognitive condition of the target object according to the frequency domain characteristics corresponding to the target signal segments in the target signal segment set. The cognitive assessment method of the object can automatically assess the cognitive condition of the target object by utilizing the electroencephalogram signals, and the assessment basis is the electroencephalogram signals capable of reflecting the activity state of the brain of a human body, so that the assessment result cannot be influenced by the education degree of the target object, and the assessment method is high in efficiency, low in cost, convenient and fast to use and high in universality.

Description

Cognitive assessment method, device and equipment of object and storage medium

Technical Field

The present application relates to the field of smart medical technology, and in particular, to a method, an apparatus, a device, and a readable storage medium for cognitive assessment of a subject.

Background

In some cases, it is necessary to evaluate the cognitive ability of the target object, for example, as the population ages, the number of patients with alzheimer disease increases, alzheimer disease is accompanied by cognitive decline and life self-care, which greatly affects the life of the elderly, mild cognitive impairment is a state between normal aging and alzheimer disease, the elderly with mild cognitive impairment, 10% -30% of which can be converted into alzheimer disease, however, alzheimer disease is irreversible, and therefore, the cognitive ability of the elderly needs to be evaluated in time to determine the degree of illness of the patients, so that the patients can be treated with intervention, thereby delaying the progress of the illness and improving the life quality of the patients.

In the prior art, the methods for evaluating the cognitive ability of a target object are mostly as follows: the target object fills some cognitive disorder scales under the inquiry of evaluators, such as a simple mental state scale (MMSE), a Montreal cognitive assessment scale (MOCAB) and the like, and the cognitive condition of the target object is determined according to the filling condition of the target object for the scales. However, the existing cognitive assessment method is susceptible to the education degree of the object to be assessed, the reliability of the assessment result is not high, and the existing cognitive assessment method is low in assessment efficiency and high in labor cost.

Disclosure of Invention

In view of this, the present application provides a method, an apparatus, a device and a readable storage medium for cognitive assessment of an object, so as to solve the problems that an assessment method in the prior art is susceptible to the education degree of an object to be assessed, which results in low reliability of an assessment result, low assessment efficiency and high labor cost, and the technical scheme is as follows:

a method of cognitive assessment of a subject, comprising:

acquiring a target brain electrical signal of a target object;

segmenting the target electroencephalogram signal according to the specified length to obtain a target signal segment set consisting of target signal segments with the specified length;

determining frequency domain characteristics corresponding to target signal segments in the target signal segment set, wherein one target signal segment corresponds to one frequency domain characteristic;

and evaluating the cognitive condition of the target object according to the frequency domain characteristics corresponding to the target signal segments in the target signal segment set.

Optionally, the acquiring a target electroencephalogram signal of a target object includes:

acquiring an initial electroencephalogram signal acquired by an electroencephalogram signal acquisition device aiming at the target object;

filtering noise from the initial electroencephalogram signal, wherein the noise comprises a signal with frequency not within a preset frequency range and/or a signal with frequency as power frequency;

removing abnormal data from the EEG signals after noise filtering;

and determining a zero potential reference point, and generating the target electroencephalogram signal based on the zero potential reference point and the electroencephalogram signal without abnormal data.

Optionally, the removing abnormal data from the noise-filtered electroencephalogram signal includes:

removing one or more combinations of the following abnormal signal sections from the EEG signal after the noise is filtered out: the signal section with the first preset length at the head part, the signal section with the second preset length at the tail part, the signal section with the abnormal amplitude value and the signal section with the length smaller than the third preset length;

removing abnormal electrode signals from the electroencephalogram signals with the abnormal signal sections removed;

performing independent component analysis on the electroencephalogram signals without the abnormal electrode signals, removing non-electroencephalogram components from the independent component analysis results, and regenerating electroencephalogram signals based on the electroencephalogram components;

and regenerating an electrode signal aiming at the abnormal electrode signal to supplement the regenerated electroencephalogram signal, wherein the electroencephalogram signal after the electrode signal is supplemented is taken as the electroencephalogram signal after the abnormal data is removed.

Optionally, the removing the abnormal electrode signal from the electroencephalogram signal with the abnormal signal segment removed includes:

for each signal segment in the electroencephalogram signal with the abnormal signal segment removed:

determining a first parameter, a second parameter and a third parameter of the signal segment, wherein the first parameter is an average value of correlation coefficients of each electrode and other electrodes of the electroencephalogram signal acquisition equipment, the second parameter is a standard deviation inside each channel signal, and the third parameter is a hester index of each channel signal;

and determining abnormal electrode signals from the channel signals of the target signal segment according to the first parameter, the second parameter and the third parameter, and removing the abnormal electrode signals.

Optionally, the independent component analysis result includes components of each channel of the electroencephalogram signal from which the abnormal electrode signal is removed;

the removing of non-electroencephalogram components from the independent component analysis result comprises:

determining a first parameter, a second parameter, a third parameter, a fourth parameter and a fifth parameter according to the independent component analysis results, wherein the first parameter comprises a correlation coefficient of a component of each channel with the acquired electro-ocular signal, the second parameter comprises a kurtosis of the component of each channel, the third parameter comprises a gradient mean value of a change in a power spectral density of the component of each channel over frequency, the fourth parameter comprises a hester index of the component of each channel, and the fifth parameter comprises a median value of a change value of adjacent points of the component of each channel;

and determining non-electroencephalogram components from the components of each channel according to the first parameter, the second parameter, the third parameter, the fourth parameter and the fifth parameter, and removing the non-electroencephalogram components.

Optionally, determining a frequency domain characteristic corresponding to one target signal segment in the target signal segment set includes:

determining a power spectral density of each channel signal of the target signal segment;

and forming a power spectral density matrix by using the power spectral densities of the channel signals of the target signal segment, wherein the power spectral density matrix is used as the frequency domain characteristic corresponding to the target signal segment.

Optionally, the evaluating the cognitive condition of the target object according to the frequency domain characteristics corresponding to the target signal segments in the target signal segment set includes:

evaluating the cognitive condition of the target object by using the frequency domain characteristics corresponding to the target signal segments in the target signal segment set and at least one pre-established cognitive evaluation model;

the cognitive evaluation model is obtained by adopting frequency domain characteristics corresponding to training signal sections in a training signal section set of training electroencephalograms and real cognitive classification training corresponding to the training electroencephalograms, the training signal section set is composed of training signal sections with specified lengths obtained by cutting the training electroencephalograms according to the specified lengths, and one training signal section corresponds to one frequency domain characteristic.

Optionally, the evaluating the cognitive condition of the target object by using the frequency domain characteristics corresponding to the target signal segments in the target signal segment set and at least one pre-established cognitive evaluation model includes:

randomly extracting preset target signal segments from the target signal segment set;

inputting the frequency domain characteristics corresponding to the preset target signal segments into each cognitive assessment model to obtain at least one group of cognitive assessment results, wherein the group of cognitive assessment results are determined by one cognitive assessment model, and the group of cognitive assessment results comprise cognitive categories corresponding to the preset target signal segments;

determining the cognition category with the maximum probability from each group of cognition evaluation results as a target cognition category;

determining a target cognition category with the maximum probability from all target cognition categories; and taking the target cognition category with the maximum probability as the cognition condition of the target object.

Optionally, the evaluating the cognitive condition of the target object by using the frequency domain characteristics corresponding to the target signal segments in the target signal segment set and at least one pre-established cognitive evaluation model further includes:

after a target cognitive category with the maximum probability is determined from all target cognitive categories, taking the target cognitive category with the maximum probability as a primary evaluation result aiming at the target electroencephalogram signal;

judging whether the evaluation times of the target electroencephalogram signal reach preset times or not;

if not, returning to execute the random extraction of preset target signal segments from the target signal segment set;

if so, determining an evaluation result with the maximum probability from the evaluation results aiming at the preset times of the target electroencephalogram signal, and taking the evaluation result with the maximum probability as the cognitive condition of the target object.

An apparatus for cognitive assessment of a subject, comprising: the device comprises an electroencephalogram signal acquisition module, an electroencephalogram signal segmentation module, a frequency domain characteristic determination module and a cognitive evaluation module;

the electroencephalogram signal acquisition module is used for acquiring a target electroencephalogram signal of a target object;

the electroencephalogram signal segmentation module is used for segmenting the target electroencephalogram signal according to the specified length to obtain a target signal segment set consisting of target signal segments with the specified length;

the frequency domain characteristic determining module is used for determining frequency domain characteristics corresponding to target signal segments in the target signal segment set, wherein one target signal segment corresponds to one frequency domain characteristic;

and the cognition evaluation module is used for evaluating the cognition condition of the target object according to the frequency domain characteristics corresponding to the target signal segments in the target signal segment set.

A cognitive assessment device of a subject, comprising: a memory and a processor;

the memory is used for storing programs;

the processor is configured to execute the program to implement the steps of the cognitive assessment method for a subject according to any one of the above.

A readable storage medium having stored thereon a computer program which, when executed by a processor, carries out the steps of the method of cognitive assessment of a subject as defined in any one of the preceding claims.

According to the scheme, the cognitive assessment method, the cognitive assessment device, the cognitive assessment equipment and the readable storage medium of the object provided by the application are characterized in that firstly, the target electroencephalogram of the target object is obtained, then, the target electroencephalogram is segmented according to the specified length, a target signal segment set consisting of target signal segments with the specified length is obtained, then, the frequency domain characteristics corresponding to the target signal segments in the target signal segment set are determined, and finally, the cognitive condition of the target object is assessed according to the frequency domain characteristics corresponding to the target signal segments in the target signal segment set. Considering that the electroencephalogram signals can reflect the activity state of the brain of a human body, the cognitive assessment method of the object automatically assesses the cognitive condition of the target object by the electroencephalogram signals, and because the assessment basis is the electroencephalogram signals capable of reflecting the activity state of the brain of the human body, the assessment result is not affected by the education degree of the target object, and because manual participation is not needed, the assessment efficiency is improved, the labor cost is reduced, and in addition, the assessment method is convenient and fast and has strong universality.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.

Fig. 1 is a schematic flowchart of a cognitive assessment method for a subject according to an embodiment of the present application;

fig. 2 is a schematic flowchart of a process for acquiring a target electroencephalogram signal of a target object according to an embodiment of the present application;

FIG. 3 is a schematic diagram of an electroencephalogram signal and a frequency distribution of the electroencephalogram signal provided by an embodiment of the present application;

FIG. 4 is a schematic flow chart illustrating a process for establishing a cognitive assessment model according to an embodiment of the present disclosure;

fig. 5 is a schematic flow chart of an implementation manner of evaluating a cognitive situation of a target object by using frequency domain features corresponding to target signal segments in a target signal segment set and at least one pre-established cognitive evaluation model according to the embodiment of the present application;

fig. 6 is a schematic flow chart of another implementation manner of evaluating a cognitive situation of a target object by using frequency domain features corresponding to target signal segments in a target signal segment set and at least one pre-established cognitive evaluation model according to the embodiment of the present application;

fig. 7 is a schematic structural diagram of a cognitive assessment apparatus for a subject according to an embodiment of the present application;

fig. 8 is a schematic structural diagram of a cognitive assessment device for a subject according to an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, 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.

In view of the problems existing in the existing manual evaluation mode, the inventor of the present invention has conducted research, and finds that the cognitive ability of a subject to be evaluated can be evaluated by using a brain image of the subject to be evaluated, cerebrospinal fluid can be extracted from the subject to be evaluated, and the cognitive ability of the subject to be evaluated can be evaluated by analyzing specific components in the cerebrospinal fluid. However, both of these assessment modalities are affected by the medical level and are expensive, and in addition, assessing the cognitive abilities of a subject to be assessed by analyzing specific components in cerebrospinal fluid is a traumatic means.

In view of the problems of the above evaluation manners, the present inventors have continued research and finally provide a cognitive evaluation method based on electroencephalogram signals, which is not easily affected by the education degree of the subject to be evaluated, and which is convenient, non-invasive and low in cost, and which can be applied to a terminal having data processing capability (such as a PC, a notebook, a smartphone, a Pad, etc.), and can also be applied to a server (which may be one server, multiple servers, or a server cluster), and the terminal or the server can acquire the electroencephalogram signals of the subject to be evaluated, and automatically process and analyze the electroencephalogram signals to determine the cognitive condition of the subject to be evaluated (such as normal aging, mild cognitive impairment, or alzheimer disease). The following examples are provided to describe the cognitive assessment method for the subject provided by the present application.

Referring to fig. 1, a schematic flow chart of a cognitive assessment method for a subject according to an embodiment of the present application is shown, where the method may include:

step S101: and acquiring a target brain electrical signal of the target object.

Considering that the electroencephalogram signals can reflect the activity state of the brain of a human body, the electroencephalogram signals are adopted to evaluate the cognitive condition of the target object. The target brain electrical signal is a multi-channel signal, such as a 64-channel signal.

It should be noted that the target electroencephalogram signal in this embodiment is a pure electroencephalogram signal, that is, the target electroencephalogram signal does not contain noise and abnormal data.

Step S102: and segmenting the target electroencephalogram signal according to the specified length to obtain a target signal segment set consisting of target signal segments with the specified length.

Considering that the overall length of the target electroencephalogram signal is usually long and may be different in different time periods, in order to fully utilize the information of the target electroencephalogram signal, the present embodiment segments the target electroencephalogram signal into signal segments of a specified length (e.g., 5s, 10s, or 15 s), and performs subsequent processing and analysis on the signal segments.

Step S103: and determining the frequency domain characteristics corresponding to the target signal segments in the target signal segment set.

Because the electroencephalogram signal has uncertainty and randomness and does not have good analyzability in a time domain, the electroencephalogram signal is transformed to a frequency domain, and the frequency domain characteristics of the electroencephalogram signal are analyzed. Specifically, the frequency domain characteristics corresponding to the target signal segments in the target signal segment set are determined, wherein one target signal segment corresponds to one frequency domain characteristic.

The frequency domain characteristic of the signal mainly refers to the energy distribution of the signal on different frequencies, and the power spectral density can be used to describe the magnitude of the energy density of the signal on different frequency points, based on which, in this embodiment, the frequency domain characteristic corresponding to one target signal segment may be a power spectral density matrix of the target signal segment.

Specifically, the process of determining the power spectral density matrix of a target signal segment may include: determining a power spectral density of each channel signal of the target signal segment; and forming a power spectral density matrix by the power spectral densities of the channel signals of the target signal segment, wherein the power spectral density matrix is used as the power spectral density matrix of the target signal segment.

Step S104: and evaluating the cognitive condition of the target object according to the frequency domain characteristics corresponding to the target signal segments in the target signal segment set.

Optionally, the cognitive status of the target may be one of the following: normal aging, mild cognitive impairment, alzheimer's disease.

The application provides a cognitive assessment method of an object, firstly, a target electroencephalogram signal reflecting a brain activity state of the target object is obtained, in order to fully utilize information of the target electroencephalogram signal, after the target electroencephalogram signal is obtained, the target electroencephalogram signal is segmented according to a specified length, a target signal segment set composed of target signal segments of the specified length is obtained, the fact that time-domain electroencephalograms are not strong in analyzability is considered, in the embodiment, frequency domain characteristics corresponding to the target signal segments in the target signal segment set are determined, and then the cognitive condition of the target object is assessed according to the frequency domain characteristics corresponding to the target signal segments in the target signal segment set. The cognitive condition of the target object can be automatically evaluated by utilizing the electroencephalogram signal, and the evaluation basis is the electroencephalogram signal capable of reflecting the activity state of the brain of the human body, so that the evaluation result obtained through the embodiment cannot be influenced by the education degree of the target object, namely, the reliability of the evaluation result is high, and the evaluation efficiency is improved and the labor cost is reduced due to the fact that manual participation is not needed.

In another embodiment of the present application, as to the "step S101: and acquiring a target brain electrical signal of the target object for introduction.

Referring to fig. 2, a schematic flow chart of acquiring a target electroencephalogram signal of a target object is shown, which may include:

step S201: acquiring an initial electroencephalogram signal acquired by an electroencephalogram signal acquisition device aiming at a target object.

The electroencephalogram signal acquisition equipment usually has a plurality of electrodes, and a signal of one channel can be obtained through one electrode, namely, the initial electroencephalogram signal is a multi-channel signal.

Step S202: noise is filtered from the initial brain electrical signal.

The noise comprises a signal with a frequency not within a preset frequency range and/or a signal with a frequency of power frequency.

The initial brain electrical signal usually contains a large amount of extraneous noise, and the noise needs to be filtered to avoid the influence of the noise on the subsequent evaluation.

Usually, the electroencephalogram signal is only in a preset frequency range, generally in the range of [ 0.195 ] Hz, and the signal outside the range of [ 0.195 ] is noise, so the signal outside the range of [ 0.195 ] needs to be filtered. In one possible implementation, a band pass filter may be used to filter out signals in the [ 0.195 ] Hz range, and to filter out signals that are not in the [ 0.195 ] range; in order to reduce the ringing effect, in another possible implementation, a low-pass filter may be used to filter out signals in the range of (095), signals not in the range of (095), and then a high-pass filter may be used to filter out signals with frequencies of 0.1 and above, so that signals in the range of [ 0.195 ] Hz may be obtained.

In addition, besides filtering out signals which are not in the range of [ 0.195 ] Hz, signals with the frequency of power frequency (power frequency refers to the frequency of an industrial alternating current power supply) need to be filtered out, wherein the power frequency is 50Hz, and signals in the range of [ 4951 ] can be filtered out considering that only 50Hz signals cannot be filtered out normally.

Optionally, after noise is filtered from the initial electroencephalogram signal, the electroencephalogram signal after the noise is filtered may be subjected to de-trending processing, that is, the mean value of all the data is subtracted from all the electroencephalogram data. Preferably, in order to reduce the data computation amount, the trend-removed electroencephalogram signal may be downsampled, for example, if the sampling rate of the initial electroencephalogram signal is 1000Hz, the sampling rate may be reduced to 500 Hz.

Step S203: and removing abnormal data from the EEG signal after the noise is filtered.

The abnormal data may include an abnormal signal segment (for example, a signal segment with an abnormal amplitude, a signal segment with a length smaller than a preset length, etc.), and may also include an abnormal electrode signal. The abnormal electrode signal refers to a signal of an abnormal electrode. The process of removing abnormal data from the noise-filtered electroencephalogram signal can be referred to the description of the subsequent embodiment.

Step S204: and determining a zero potential reference point, and generating a target electroencephalogram signal based on the zero potential reference point and the electroencephalogram signal without abnormal data.

The electroencephalogram signal recorded after the acquisition is a relative potential value, so a neutral point needs to be found as a zero potential reference point, the embodiment uses an absolute zero reference method to find the zero potential reference point, the electroencephalogram signal without abnormal data is recalculated according to the found zero potential reference point, and the computed electroencephalogram signal is used as a target electroencephalogram signal.

The following is made for the above-described "step S203: introduction is made to the process of removing abnormal data from the electroencephalogram signals after noise filtering.

The process of removing abnormal data from the noise-filtered electroencephalogram signal may include:

and S2031, removing abnormal signal sections from the electroencephalogram signals after noise filtering.

Wherein the abnormal signal segment comprises one or more of the following signal segments in combination: the signal section of the first preset length of head, the signal section of the second preset length of tail, the signal section of the abnormal amplitude, the signal section of which the length is less than the third preset length. The first preset length, the second preset length and the third preset length can be set according to actual conditions.

Considering that the electroencephalogram signal is easy to be unstable when the electroencephalogram signal is started to be acquired and is about to be acquired, the signal section with the first preset length at the head and the signal section with the second preset length at the tail of the electroencephalogram signal after noise is filtered can be taken as abnormal signal sections to be removed.

In the process of acquiring the electroencephalogram of the target object, a situation that a part of electrodes of the electroencephalogram acquisition device are in poor contact possibly due to the influence of the body activity of the target object can cause the signal in the period of time to have overall large amplitude deviation, the amplitude of the overall signal section is abnormal and large, for the situation, a detection threshold (such as 0.25) is set in the embodiment, when the ratio of the number of channels with abnormal amplitude to the total number of channels is larger than the detection threshold, if the number of channels with abnormal amplitude detected at the same moment is larger than the preset number of channels (such as 5), the signal section with abnormal amplitude is determined to be abnormal, and the whole signal section is deleted (that is, all channel signals in the signal section are deleted). In addition, in order to reduce the influence of the abnormal boundary, a signal segment having a length smaller than a third preset length (for example, 10s) is deleted as an abnormal signal segment.

And S2032, removing the abnormal electrode signals from the electroencephalogram signals with the abnormal signal sections removed.

Specifically, the process of removing the abnormal electrode signal from the electroencephalogram signal from which the abnormal signal section is removed includes:

step a1, for each signal segment in the electroencephalogram signal with the abnormal signal segment removed: a first parameter, a second parameter, and a third parameter of the signal segment are determined.

The first parameter is the average value of the correlation coefficient of each electrode of the electroencephalogram signal acquisition equipment and other electrodes, the second parameter is the standard deviation inside each channel signal, and the third parameter is the Hurst index of each channel signal.

Assuming that the initial brain electrical signal acquired is a 64-channel signal, the first parameter comprises 64 mean values corresponding to 64 channels, the second parameter comprises 64 standard deviations corresponding to 64 channels, and the third parameter comprises 64 hestert indices corresponding to 64 channels.

Step a2, according to the first parameter, the second parameter and the third parameter, determining abnormal electrode signals from the signals of each channel of the target signal segment, and removing the abnormal electrode signals.

Specifically, a lower limit value corresponding to the first parameter may be determined according to the first parameter, an upper limit value corresponding to the second parameter may be determined according to the second parameter, an upper limit value and a lower limit value corresponding to the third parameter may be determined according to the third parameter, and then, whether an abnormal electrode signal exists in each channel signal of the signal segment may be determined according to the lower limit value corresponding to the first parameter and the first parameter, the upper limit value corresponding to the second parameter and the second parameter, and the upper limit value and the lower limit value corresponding to the third parameter and the third parameter.

Assume that the first parameter is param₁(including param)₁₁、param₁₂、_…、param_1NN is the total number of channels), and the second parameter is param₂(including param)₂₁、param₂₂、_…、param_2N) Representing the third parameter by param₃(including param)₃₁、param₃₂、_…、param_3N) Represents, then any parameter param_iUpper limit value up _ lim corresponding to (i ═ 1,2,3)_iCan be calculated by the following formula:

up_lim_i＝median(param_i)+

3*sqrt(mean((param_i-repmat(median(param_i),size(param_i,1),1)).^2)) (1)

similarly, any parameter param_iCorresponding lower bound value down _ lim_iCan be calculated by the following formula:

down_lim_i＝median(param_i)-

3*sqrt(mean((param_i-repmat(median(param_i),size(param_i,1),1)).^2)) (2)

in the above formulas (1) and (2), mean represents the median, sqrt represents the square root, repmat (x, n) represents the number of copies of n x and stores in a matrix, and size (x,1) represents the number of lines of x in the matrix.

It should be noted that the upper limit value corresponding to any parameter includes an upper limit value corresponding to each channel, and similarly, the lower limit value corresponding to any parameter includes a lower limit value corresponding to each channel.

Obtaining the lower limit value down _ lim corresponding to the first parameter₁The upper limit value up _ lim corresponding to the second parameter₂Upper limit value up _ lim corresponding to third parameter₃And a lower limit value down _ lim₃Then, the abnormal electrode signal can be determined, specifically, for the j (j is from 1 to N, N is the total channel number) channel signal in the signal segment, if the first parameter param₁Param in (1)_1jLess than the corresponding lower limit value down _ lim_1jOr, alternatively, the second parameter param₂Param in (1)_2jGreater than the corresponding upper limit value up _ lim_2jOr, alternatively, the third parameter param₃Param in (1)_3jLess than the corresponding lower limit value down _ lim_3jOr, alternatively, the third parameter param₃Param in (1)_3jGreater than the corresponding upper limit value up _ lim_3jAnd determining that the jth channel signal in the signal segment is an abnormal electrode signal.

In addition, if the number of the abnormal electrode signals included in one signal segment is greater than or equal to a preset number (for example, 64 channels, where the preset number here can be set to 20), all the channel signals in the signal segment are removed as the abnormal electrode signals, that is, the entire signal segment is removed.

And S2033, performing independent component analysis on the electroencephalogram signals without the abnormal electrode signals, removing non-electroencephalogram components from the independent component analysis results, and regenerating the electroencephalogram signals based on the electroencephalogram components.

Wherein, the process of removing the non-electroencephalogram component from the independent component analysis result may include:

and b1, determining the first parameter, the second parameter, the third parameter, the fourth parameter and the fifth parameter according to the independent component analysis result.

Wherein the first parameter comprises a correlation coefficient of a component of each channel with the acquired ocular electrical signal, the second parameter comprises a kurtosis of the component of each channel, the third parameter comprises a gradient mean of a change in a power spectral density of the component of each channel over frequency, the fourth parameter comprises a hester index of the component of each channel, and the fifth parameter comprises a median of adjacent point changes of the component of each channel.

Step b2, determining non-electroencephalogram components from the components of each channel according to the first parameter, the second parameter, the third parameter, the fourth parameter and the fifth parameter, and removing the non-electroencephalogram components.

Specifically, the non-electroencephalogram component may be determined from the components of each channel according to the upper limit value corresponding to the first parameter, the upper limit value corresponding to the second parameter, the lower limit value corresponding to the third parameter, the upper limit value and the lower limit value corresponding to the fourth parameter, the upper limit value and the lower limit value corresponding to the fifth parameter, and the upper limit value and the lower limit value corresponding to the first parameter, the upper limit value corresponding to the second parameter, the lower limit value corresponding to the third parameter, the upper limit value and the lower limit value corresponding to the fourth parameter, and the upper limit value and the lower limit value corresponding to the fifth parameter.

Assume that the first parameter is param₁(including param)₁₁、param₁₂、_…、param_1NN is the total number of channels), and the second parameter is param₂(including param)₂₁、param₂₂、_…、param_2N) Representing the third parameter by param₃(including param)₃₁、param₃₂、_…、param_3N) The fourth parameter is param₄(including param)₄₁、param₄₂、_…、param_4N) The fifth parameter is param₅(including param)₅₁、param₅₂、_…、param_5N) Represents, then any parameter param_kCorresponding upper limit value up _ lim_k(k ═ 1,2,3,4,5) can be calculated by the following formula:

up_lim_k＝median(param_k)+

3*sqrt(mean((param_k-repmat(median(param_k),size(param_k,1),1)).^2)) (3)

similarly, any parameter param_kCorresponding lower bound value down _ lim_kCan be calculated by the following formula:

down_lim_k＝median(param_k)-

3*sqrt(mean((param_k-repmat(median(param_k),size(param_k,1),1)).^2)) (4)

in the above formulas (3) and (4), mean represents the median, sqrt represents the square root, repmat (x, n) represents the number of copies of n x and stores in a matrix, and size (x,1) represents the number of lines of x in the matrix.

Obtaining the upper limit value up _ lim corresponding to the first parameter₁The upper limit value up _ lim corresponding to the second parameter₂And a lower limit value down _ lim corresponding to the third parameter₃The upper limit value up _ lim corresponding to the fourth parameter₄And a lower limit value down _ lim₄The upper limit value up _ lim corresponding to the fifth parameter₅And a lower limit value down _ lim₅Then, the non-electroencephalogram components can be determined from the components of each channel, specifically:

for the component of j (j is from 1 to N, N is the total number of channels) channel, if the first parameter param₁Param in (1)_1jGreater than the corresponding lower limit value up _ lim_1jOr, alternatively, the second parameter param₂Param in (1)_2jGreater than the corresponding upper limit value up _ lim_2jOr is orThird parameter param₃Param in (1)_3jLess than the corresponding lower limit value down _ lim_3jOr, the fourth parameter param₄Param in (1)_4jLess than the corresponding lower limit value down _ lim_4jOr, param in the fourth parameter_4jGreater than the corresponding upper limit value up _ lim_4jOr, the fifth parameter param₅Param in (1)_5jLess than the corresponding lower limit value down _ lim_5jOr, param in the fifth parameter_5jGreater than the corresponding upper limit value up _ lim_5jAnd determining that the component of the jth channel is a non-electroencephalogram component.

Step S2034, regenerating an electrode signal aiming at the abnormal electrode signal to supplement the regenerated electroencephalogram signal.

In consideration of the closer relationship between the electroencephalogram signal and the electrode signal, the present embodiment regenerates the electrode signal for the abnormal electrode signal to supplement the electroencephalogram signal obtained through S2033. Alternatively, an interpolation algorithm may be used to select normal electrode signals around the abnormal electrode signal according to the template of the electrode for interpolation to computationally generate a new electrode signal. The electroencephalogram signal to which the electrode signal is added is referred to as the "electroencephalogram signal after removing abnormal data".

After the initial electroencephalogram signal is processed, a target electroencephalogram signal can be obtained, and after the target electroencephalogram signal is obtained, the target electroencephalogram signal is segmented according to the specified length, so that a target signal segment set consisting of target signal segments with the specified length is obtained.

In a possible implementation manner, the target signal segment in the target signal segment set is a signal segment with a specified length obtained by segmenting the target electroencephalogram signal, and in another preferred implementation manner, the target signal segment in the target signal segment set is an effective signal segment in all signal segments with the specified length obtained by segmenting the target electroencephalogram signal. It should be noted that the valid new signal segment refers to a signal segment that conforms to the overall distribution trend of the target electroencephalogram signal.

Optionally, it may be determined whether the signal segment with the specified length obtained by splitting is a valid signal segment by the following method:

supposing that n signal segments with specified lengths are obtained after segmentation is carried out on a target electroencephalogram signal, wherein the n signal segments with the specified lengths are epochs respectively₁、epoch₂、…、epoch_n：

Step c1, determining the first parameter param of n epochs₁Second parameter param₂And a third parameter param₃。

Wherein the first parameter param₁Including the difference between the maximum and minimum values of each epoch, the second parameter including the difference between the mean value of each epoch and the mean value of the target brain electrical signal, and the third parameter including the variance of each epoch, it can be seen that each parameter includes n values.

Step c2, param according to the first parameter₁Determining a first parameter param₁Corresponding upper limit value up _ lim₁According to the second parameter param₂Determining a second parameter param₂Corresponding upper limit value up _ lim₂According to the third parameter param₃Determining a third parameter param₃Corresponding upper limit value up _ lim₃。

Specifically, any parameter param_xCorresponding upper limit value up _ lim_x(x ═ 1,2,3,4,5) can be calculated by the following formula:

up_lim_x＝median(param_x)+

3*sqrt(mean((param_x-repmat(median(param_x),size(param_x,1),1)).^2)) (5)

wherein, up _ lim_xIncluding an upper limit value corresponding to each epoch.

Step c2, param according to the first parameter₁And a first parameter param₁Corresponding upper limit value up _ lim₁Param according to the second parameter₂And a second parameter param₂Corresponding upper limit value up _ lim₂And param according to the third parameter₃And a third parameter param₃Corresponding upper limit value up _ lim₃Invalid signal segments are determined from the n epochs.

In particular, for epoch_i(i from 1 to n) if the first parameter param₁Param in (1)_1i(i.e., epoch)_iIs greater than the corresponding upper limit value up _ lim) is greater than the corresponding upper limit value up _ lim_1iOr, alternatively, the second parameter param₂Param in (1)_1i(i.e., epoch)_iThe difference between the mean value of (a) and the mean value of the target electroencephalogram signal) is greater than the corresponding upper limit value up _ lim_2iOr, alternatively, the third parameter param₃Param in (1)_3i(i.e., epoch)_iVariance of) is greater than the corresponding upper limit value up _ lim_3iThen determine epoch_iAre invalid signal segments.

After the target signal segment set is obtained, the frequency domain characteristics corresponding to the target signal segments in the target signal segment set can be determined, and then the cognitive condition of the target object is evaluated according to the frequency domain characteristics corresponding to the target signal segments in the target signal segment set.

Next, as to "step S104: and evaluating the cognitive condition of the target object for introduction according to the frequency domain characteristics corresponding to the target signal segments in the target signal segment set.

In a possible implementation manner, the process of evaluating the cognitive situation of the target object according to the frequency domain feature corresponding to the target signal segment in the target signal segment set may include: and evaluating the cognitive condition of the target object by utilizing the frequency domain characteristics corresponding to the target signal segments in the target signal segment set and at least one pre-established cognitive evaluation model.

The frequency domain characteristic corresponding to the target signal segment is preferably a power spectral density matrix. As shown in FIG. 3, the target EEG signal is a signal within [ 0.195 ] Hz, and the effective information therein is basically distributed only between [ 050 ] Hz, and in order to reduce the computation amount, only the power spectral density value within [ 050 ] Hz can be reserved.

The cognitive assessment model comprises a plurality of cognitive assessment models, wherein each cognitive assessment model is obtained by adopting frequency domain characteristics (preferably power spectrum density matrix) corresponding to a training signal section in a training signal section set of training electroencephalogram signals and real cognitive classification training corresponding to the training electroencephalogram signals, the training signal section set comprises training signal sections (preferably effective signal sections) with specified lengths, the training signal sections are obtained by dividing the training electroencephalogram signals according to the specified lengths, and one training signal section corresponds to one frequency domain characteristic.

The following describes the topology of the cognitive assessment model and the process of building the cognitive assessment model.

Optionally, the topology of the cognitive assessment model may include:

an input layer: inputting a power spectral density matrix with the size of N x M;

the convolutional layer 1: convolution kernel size 5 × 5, number 8, strings ═ 1,1], activation function "relu";

and (3) convolutional layer 2: same as the convolutional layer 1;

a pooling layer 1: "maxpololing" mode, pool _ size ═ 2,2], strands ═ 1,1 ];

setting dropout to be 0.25, randomly discarding 1/4 network nodes, and reducing overfitting;

and (3) convolutional layer: convolution kernel size 3 x 3, number 16, strings ═ 1,1], activation function "relu";

and (4) convolutional layer: the same as the convolutional layer 3;

and (3) a pooling layer 2: as with the pooling layer 1, dropout is set to 0.25;

full connection layer: the number of nodes 256, the function "relu" is activated, and dropout is set to 0.25;

an output layer: softmax classifier, output probabilities for 3 classes ("normal aging", "mild cognitive impairment" and "alzheimer").

Referring to fig. 4, a schematic flow chart of establishing a cognitive assessment model is shown, which may include:

step S401: the method comprises the steps of obtaining training electroencephalogram signals from a training data set, segmenting the training electroencephalogram signals according to specified lengths, and obtaining a training signal segment set formed by signal segments (preferably effective signal segments) with specified lengths.

The training data set can comprise training electroencephalograms with cognitive categories of normal aging, training electroencephalograms with cognitive categories of mild cognitive impairment and training electroencephalograms with cognitive categories of Alzheimer's disease, and preferably, the number of the three types of training electroencephalograms can be the same.

Step S402: and randomly extracting preset training signal segments from the training signal segment set, and determining frequency domain characteristics corresponding to the preset training signal segments respectively.

Frequency domain characteristics (preferably power spectral density matrix) corresponding to the preset training signal segments are real training samples of the cognitive assessment model. And the power spectral density matrix corresponding to one training signal segment consists of the power spectral densities of the channel signals in the training signal segment.

Step S403: and predicting the cognitive categories corresponding to the preset training signal segments by using the frequency domain characteristics and the cognitive assessment model corresponding to the preset training signal segments.

Step S404: and determining the cognition category with the maximum probability from the cognition categories respectively corresponding to the preset training signal segments, and updating the parameters of the cognition evaluation model according to the cognition category with the maximum probability and the real cognition category corresponding to the training electroencephalogram signal.

Specifically, the prediction loss of the cognitive evaluation model is determined according to the cognitive category with the maximum probability and the real cognitive category corresponding to the training electroencephalogram signal, and the parameters of the cognitive evaluation model are updated according to the prediction loss of the cognitive evaluation model. Alternatively, the predicted loss of the cognitive assessment model may be calculated using a cross-entropy loss function.

And carrying out repeated iterative training according to the process until the model converges.

And a plurality of cognitive assessment models for carrying out cognitive assessment on the target object can be established by adopting different training samples according to the training mode.

The method includes the following steps that the frequency domain characteristics corresponding to the target signal segments in the target signal segment set and at least one pre-established cognitive assessment model are utilized to assess the cognitive situation of the target object.

Referring to fig. 5, a schematic flow chart of an implementation manner of evaluating the cognitive status of a target object by using frequency domain features corresponding to target signal segments in a target signal segment set and at least one cognitive evaluation model established in advance is shown, and the evaluation method may include:

step S501: and randomly extracting preset target signal segments from the target signal segment set.

If q training signal segments are randomly extracted from the training signal segment set when the cognitive assessment model is trained, q target signal segments are also randomly extracted from the target signal segment set, wherein q is an integer greater than or equal to 1, and the specific value of q can be set according to the actual situation.

Step S502: and inputting the frequency domain characteristics corresponding to the preset target signal segments into each cognitive assessment model to obtain at least one group of cognitive assessment results.

The group of cognitive assessment results are determined by using a cognitive assessment model, the group of cognitive assessment results comprise cognitive categories corresponding to preset target signal segments respectively, and the cognitive category corresponding to one target signal segment can be one of the following cognitive categories: normal aging, mild cognitive impairment, senile dementia.

If q target signal segments are randomly extracted from the target signal segment set and p cognitive evaluation models exist, p groups of cognitive evaluation results can be obtained, one group of cognitive evaluation results is determined by one cognitive evaluation model, one group of cognitive evaluation results comprises cognitive categories corresponding to the q training signal segments respectively, namely one group of cognitive evaluation results comprises q cognitive categories.

For any cognitive assessment model, after the frequency domain features (preferably power spectral density matrix) corresponding to any target signal segment in preset target signal segments are input into the cognitive assessment model, the cognitive assessment model outputs three probabilities, namely the probability that the cognitive category corresponding to the target signal segment is normal aging, the probability that the cognitive category corresponding to the target signal segment is mild cognitive impairment and the probability that the cognitive category corresponding to the target signal segment is senile dementia, determines the maximum probability from the three probabilities, and determines the cognitive category corresponding to the maximum probability as the cognitive category corresponding to the target signal segment.

Step S503: and determining the cognitive category with the highest probability from each group of cognitive assessment results as a target cognitive category.

If p cognitive evaluation models are provided, p groups of cognitive evaluation results can be obtained, the cognitive category with the highest probability is determined from each group of cognitive evaluation results and serves as a target cognitive category, and therefore p target cognitive categories can be obtained.

Step S504: and determining the target cognition category with the maximum probability from all the target cognition categories, and taking the target cognition category with the maximum probability as the cognition condition of the target object.

Assuming that p target recognition categories are obtained in step S503, in this step, the target recognition category with the highest probability is determined from the p target recognition categories, and the target recognition category with the highest probability is determined as the recognition situation of the target object. Assuming that the target cognitive category with the highest probability is mild cognitive impairment, the cognitive status of the target object is mild cognitive impairment.

On the basis of the foregoing implementation manner, the present application provides another preferred implementation manner for evaluating the cognitive situation of the target object by using the frequency domain features corresponding to the target signal segments in the target signal segment set and at least one pre-established cognitive evaluation model, please refer to fig. 6, which shows a flow diagram of the implementation manner, and may include:

step S601: and randomly extracting preset target signal segments from the target signal segment set.

Step S602: and inputting the frequency domain characteristics corresponding to each target signal segment in the preset target signal segments into each cognitive evaluation model to obtain at least one group of cognitive evaluation results.

And the group of cognitive assessment results comprise cognitive categories corresponding to the preset target signal segments respectively.

Step S603: and determining the cognitive category with the highest probability from each group of cognitive assessment results as a target cognitive category.

Step S604: and determining the target cognitive category with the maximum probability from all the target cognitive categories, and taking the target cognitive category with the maximum probability as a primary evaluation result aiming at the target electroencephalogram.

The steps S601 to S604 are the same as the steps S501 to S504, and the detailed description is omitted here.

Step S605: and judging whether the evaluation frequency of the target electroencephalogram signal reaches a preset frequency, if not, returning to the step S601, and if so, executing the step S606.

Step S606: and obtaining an evaluation result with the maximum probability from the evaluation results aiming at the preset times of the target electroencephalogram signal, and taking the evaluation result as the cognitive condition of the target object.

The embodiment combines the signal segmentation and the multiple voting idea, and utilizes at least one (preferably a plurality of) cognitive evaluation models to evaluate the cognitive situation of the target object, so as to obtain better evaluation effect.

The cognitive assessment method of the object provided by the embodiment of the application can automatically preprocess the initial electroencephalogram signal acquired by electroencephalogram signal acquisition equipment to obtain a pure target electroencephalogram signal, after the target electroencephalogram signal is acquired, in order to fully utilize the information of the target electroencephalogram signal, the target electroencephalogram signal is segmented according to the specified length, a target signal segment set composed of target signal segments of the specified length is acquired, the time-domain electroencephalogram signal is considered to be poor in usually analyzability, the frequency domain characteristics (such as a power spectral density matrix) corresponding to the target signal segments in the target signal segment set are determined in the embodiment, and then the cognitive condition of the target object is assessed according to the frequency domain characteristics corresponding to the target signal segments in the target signal segment set. The cognitive assessment method provided by the embodiment of the application is not affected by the education degree of the target object, namely the reliability of the assessment result is high, and the cognitive assessment method provided by the embodiment of the application has high assessment efficiency and low cost, is convenient and fast, and is high in universality.

The following describes a cognitive assessment device for a subject provided in an embodiment of the present application, and the cognitive assessment device for a subject described below and the cognitive assessment method for a subject described above may be referred to in correspondence with each other.

Referring to fig. 7, a schematic structural diagram of an apparatus for cognitive assessment of a subject according to an embodiment of the present application is shown, where the apparatus for cognitive assessment may include: an electroencephalogram signal acquisition module 701, an electroencephalogram signal segmentation module 702, a frequency domain characteristic determination module 703 and a cognitive evaluation module 704.

The electroencephalogram signal acquisition module 701 is used for acquiring a target electroencephalogram signal of a target object.

And the electroencephalogram signal segmentation module 702 is configured to segment the target electroencephalogram signal according to a specified length to obtain a target signal segment set composed of target signal segments of specified lengths.

A frequency domain characteristic determining module 703, configured to determine a frequency domain characteristic corresponding to a target signal segment in the target signal segment set, where one target signal segment corresponds to one frequency domain characteristic.

And the cognition evaluation module 704 is configured to evaluate a cognition condition of the target object according to the frequency domain characteristics corresponding to the target signal segments in the target signal segment set.

The cognitive evaluation device of object that this application embodiment provided can utilize brain electrical signal to carry out automatic assessment to the cognitive condition of target object, because the evaluation foundation is the brain electrical signal that can reflect human brain activity state, consequently, the evaluation result can not receive the influence of the education degree that target object self receives, and the credibility of evaluation result is higher promptly, in addition, the evaluation device evaluation efficiency that this application embodiment provided is higher, and the evaluation cost is lower, and is more convenient, and the commonality is stronger.

Optionally, the electroencephalogram signal acquiring module 701 in the above embodiment may include: the device comprises an initial electroencephalogram signal acquisition module, a noise filtering module, an abnormal data removing module and a target electroencephalogram signal generation module.

And the initial electroencephalogram signal acquisition module is used for acquiring an initial electroencephalogram signal acquired by the electroencephalogram signal acquisition equipment aiming at the target object.

And the noise filtering module is used for filtering noise from the initial electroencephalogram signal, wherein the noise comprises a signal with frequency out of a preset frequency range and/or a signal with frequency as power frequency.

And the abnormal data removing module is used for removing abnormal data from the electroencephalogram signals after the noise is filtered.

And the target electroencephalogram signal generation module is used for determining a zero potential reference point and generating the target electroencephalogram signal based on the zero potential reference point and the electroencephalogram signal without abnormal data.

Optionally, the above-mentioned abnormal data removing module may include: an abnormal signal section removing submodule, an abnormal electrode signal removing submodule, a non-electroencephalogram component removing submodule and an electrode signal supplementing submodule.

An abnormal signal segment removing submodule, configured to remove, from the noise-filtered electroencephalogram signal, one or a combination of more than one of the following abnormal signal segments: the signal section with the first preset length at the head part, the signal section with the second preset length at the tail part, the signal section with the abnormal amplitude value and the signal section with the length smaller than the third preset length;

and the abnormal electrode signal removing submodule is used for removing the abnormal electrode signal from the electroencephalogram signal from which the abnormal signal section is removed.

And the non-electroencephalogram component removing submodule is used for carrying out independent component analysis on the electroencephalogram signals with the abnormal electrode signals removed, removing the non-electroencephalogram components from the independent component analysis results, and regenerating the electroencephalogram signals based on the electroencephalogram components.

And the electrode signal supplementing sub-module is used for regenerating an electrode signal aiming at the abnormal electrode signal so as to supplement the regenerated electroencephalogram signal, and the electroencephalogram signal after the electrode signal is supplemented is taken as the electroencephalogram signal after the abnormal data is removed.

Optionally, the abnormal electrode signal removing submodule is specifically configured to, for each signal segment in the electroencephalogram signal from which the abnormal signal segment is removed: determining a first parameter, a second parameter and a third parameter of the signal segment, wherein the first parameter is an average value of correlation coefficients of each electrode and other electrodes of electroencephalogram signal acquisition equipment, the second parameter is a standard deviation inside each channel signal, and the third parameter is a hester index of each channel signal; and determining abnormal electrode signals from the signals of all channels of the target signal segment according to the first parameter, the second parameter and the third parameter, and removing the abnormal electrode signals.

The independent component analysis result comprises the components of each channel of the electroencephalogram signal after the abnormal electrode signal is removed.

Optionally, the non-electroencephalogram component removal sub-module is specifically configured to determine, according to the independent component analysis result, a first parameter, a second parameter, a third parameter, a fourth parameter, and a fifth parameter, where the first parameter includes a correlation coefficient between a component of each channel and the acquired ocular electrical signal, the second parameter includes a kurtosis of the component of each channel, the third parameter includes a gradient mean of a change in a power spectral density of the component of each channel over frequency, the fourth parameter includes a hester index of the component of each channel, and the fifth parameter includes a median of change values of adjacent points of the component of each channel; and determining non-electroencephalogram components from the components of each channel according to the first parameter, the second parameter, the third parameter, the fourth parameter and the fifth parameter, and removing the non-electroencephalogram components.

If the target electroencephalogram signal is a multi-channel signal, the frequency domain characteristic determining module 703 is specifically configured to determine the power spectral density of each channel signal of a target signal segment when determining the frequency domain characteristic corresponding to one target signal segment in the target signal segment set; and forming a power spectral density matrix by using the power spectral densities of the channel signals of the target signal segment, wherein the power spectral density matrix is used as the frequency domain characteristic corresponding to the target signal segment.

Optionally, the cognitive evaluation module 704 in the foregoing embodiment is specifically configured to evaluate the cognitive situation of the target object by using the frequency domain features corresponding to the target signal segments in the target signal segment set and at least one cognitive evaluation model established in advance.

Each cognitive assessment model is obtained by adopting frequency domain characteristics corresponding to training signal sections in a training signal section set of training electroencephalograms and real cognitive classification training corresponding to the training electroencephalograms, the training signal section set is composed of training signal sections with specified lengths obtained by cutting the training electroencephalograms according to the specified lengths, and one training signal section corresponds to one frequency domain characteristic.

Optionally, the cognitive assessment module 704 may include: the system comprises a signal segment extraction submodule, a cognition evaluation submodule and a cognition condition determination submodule.

And the signal segment extraction submodule is used for randomly extracting preset target signal segments from the target signal segment set.

And the cognitive evaluation submodule is used for inputting the frequency domain characteristics corresponding to the preset target signal segments into each cognitive evaluation model so as to obtain at least one group of cognitive evaluation results.

And the cognition category determining submodule is used for determining a cognition category with the maximum probability from each group of cognition evaluation results as a target cognition category and determining a target cognition category with the maximum probability from all the target cognition categories.

And the first cognition situation determination submodule is used for determining the target cognition category with the highest probability as the cognition situation of the target object.

Optionally, the cognitive assessment module 704 may further include: an evaluation result determining submodule, a judging submodule and a second cognition condition determining submodule.

And the evaluation result determining submodule is used for taking the target cognition category with the maximum probability as a primary evaluation result aiming at the target electroencephalogram signal.

And the judgment submodule is used for judging whether the evaluation frequency of the target electroencephalogram signal reaches the preset frequency.

And the signal segment extraction submodule is also used for randomly extracting preset target signal segments from the target signal segment set when the evaluation times of the target electroencephalogram signals do not reach the preset times.

And the second cognitive condition determining submodule is used for determining an evaluation result with the maximum probability from the evaluation results aiming at the preset times of the target electroencephalogram signal when the evaluation times of the target electroencephalogram signal reach the preset times, and taking the evaluation result as the cognitive condition of the target object.

An embodiment of the present application further provides cognitive assessment equipment for an object, please refer to fig. 8, which shows a schematic structural diagram of the cognitive assessment equipment for the object, where the assessment equipment may include: at least one processor 801, at least one communication interface 802, at least one memory 803, and at least one communication bus 804;

in the embodiment of the present application, the number of the processor 801, the communication interface 802, the memory 803, and the communication bus 804 is at least one, and the processor 801, the communication interface 802, and the memory 803 complete communication with each other through the communication bus 804;

the processor 801 may be a central processing unit CPU, or an application specific Integrated circuit asic, or one or more Integrated circuits configured to implement embodiments of the present invention, or the like;

the memory 803 may include a high-speed RAM memory, and may further include a non-volatile memory (non-volatile memory) or the like, such as at least one disk memory;

wherein the memory stores a program and the processor can call the program stored in the memory, the program for:

acquiring a target brain electrical signal of a target object;

Alternatively, the detailed function and the extended function of the program may be as described above.

Embodiments of the present application further provide a readable storage medium, where a program suitable for being executed by a processor may be stored, where the program is configured to:

acquiring a target brain electrical signal of a target object;

Finally, it should also be noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.

The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.

The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for cognitive assessment of a subject, comprising:

acquiring a target brain electrical signal of a target object;

2. The method of cognitive assessment of a subject according to claim 1, wherein said acquiring a target brain electrical signal of a target subject comprises:

removing abnormal data from the EEG signals after noise filtering;

3. The method of cognitive assessment of a subject according to claim 2, wherein said removing abnormal data from said noise filtered brain electrical signal comprises:

4. The method for cognitive assessment of a subject according to claim 3, wherein said removing of abnormal electrode signals from said electroencephalogram signal after removal of abnormal signal segments comprises:

5. The method of cognitive assessment of a subject according to claim 3, wherein said independent component analysis results comprise components of each channel of said electroencephalogram signal after said abnormal electrode signal is removed;

6. The method of claim 1, wherein determining the frequency domain feature corresponding to the one of the set of target signal segments comprises:

7. The method according to claim 1, wherein the step of evaluating the cognition of the target object according to the frequency domain feature corresponding to the target signal segment in the target signal segment set comprises:

8. The method according to claim 7, wherein the estimating the cognitive status of the target object by using the frequency domain features corresponding to the target signal segments in the target signal segment set and at least one cognitive estimation model established in advance comprises:

9. The method according to claim 8, wherein the method for evaluating the cognition of the target object by using the frequency domain features corresponding to the target signal segments in the target signal segment set and at least one pre-established cognition evaluation model further comprises:

10. An apparatus for cognitive assessment of a subject, comprising: the device comprises an electroencephalogram signal acquisition module, an electroencephalogram signal segmentation module, a frequency domain characteristic determination module and a cognitive evaluation module;

11. A cognitive assessment device of a subject, comprising: a memory and a processor;

the memory is used for storing programs;

the processor, configured to execute the program, implementing the steps of the method for cognitive assessment of a subject according to any one of claims 1 to 9.

12. A readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method for cognitive assessment of a subject according to any one of claims 1 to 9.