Brain function state determination method, device and storage medium

Abstract

The present invention discloses a method for judging brain function status, comprising: obtaining EEG signals in an awake state and EEG signals in other states; obtaining a reference feature set of the awake state and a feature set of other states; obtaining a normalized transformed reference feature set of the awake state and a normalized transformed feature set of other states; obtaining the spatial distance between the normalized transformed feature set of other states and the normalized transformed reference feature set of the awake state; obtaining a normalized spatial distance value; and judging the degree of change in brain function status according to the normalized spatial distance value. The present invention normalizes and quantifies the spatial distance between EEG feature sets and then judges the degree of change in brain function status according to the quantized value, thereby solving the problem that traditional EEG signal analysis methods cannot quantitatively represent the specific degree of change in brain function status. The present invention also provides a brain function status detection device and a storage medium.

Description

Technical Field

[0001] The present invention relates to the field of biomedical signal processing technology, and in particular to a method, device and storage medium for determining brain function status. Background Art

[0002] When the brain is performing a task, the brain's functional state will change over time. The brain's functional state includes: driving fatigue level, brain workload, concentration level, etc. If the degree of change in the brain's functional state can be detected and judged, intervention can be made at the appropriate time to prevent the brain from entering extreme states such as fatigue, so that the brain can better complete the task.

[0003] Wearable electroencephalographic (EEG) signal detection devices can detect EEG signals in real time. Using traditional EEG signal analysis methods, these signals can qualitatively characterize changes in brain function. For example, traditional EEG signal analysis methods can determine whether a driver's brain is alert or fatigued. While traditional EEG signal analysis methods can qualitatively characterize brain function, they cannot quantitatively characterize it. Specifically, they cannot break down brain function into specific numerical values, resulting in an inability to quantify the specific degree of change in brain function. Without further analysis of the specific degree of change in brain function, it's impossible to determine when brain intervention measures are necessary. Summary of the Invention

[0004] The present invention aims to address the problem that traditional EEG signal analysis methods are unable to quantify the specific degree of change in brain function. This invention provides a method for determining brain function status by normalizing and quantifying the spatial distances between EEG feature sets, and then determining the degree of change in brain function status based on the quantified values. This invention effectively addresses the problem that traditional EEG signal analysis methods are unable to quantify the specific degree of change in brain function status.

[0005] To solve the above technical problems, an embodiment of the present invention discloses a method for judging brain function state, including: obtaining EEG signals in a waking state and EEG signals in other states; obtaining a waking state reference feature set and other state feature sets; obtaining a normalized transformed waking state reference feature set and a normalized transformed other state feature set; obtaining the spatial distance between the normalized transformed other state feature set and the normalized transformed waking state reference feature set; obtaining a normalized spatial distance value; judging the degree of change in brain function state according to the normalized spatial distance value; obtaining the waking state reference feature set requires separately calculating the features of the EEG signals, fusing the features to obtain the waking state reference feature set, and obtaining the other state feature sets also requires separately calculating the features of the EEG signals, fusing the features to obtain the other state feature sets, and the features include: functional connectivity features, nonlinear dynamic features and power spectrum features; at least one waking state reference feature set is constructed.

[0006] By adopting the above technical solution, the brain function state can be quantitatively characterized, that is, the brain function state can be broken down into specific numerical values. Based on the specific quantitative values of the brain function state, a judgment benchmark for the degree of change in the brain function state can be established, and the degree of change in the brain function state can be effectively analyzed and judged. It is also possible to set intervention measures for the brain at specific times according to actual application conditions.

[0007] According to another specific embodiment of the present invention, after obtaining the EEG signals in the awake state and the EEG signals in other states, the EEG signals are preprocessed. The preprocessing includes: threshold denoising, baseline drift removal, and band-pass filtering. After preprocessing, theta waves, alpha waves, beta waves and gamma waves of the EEG signals are obtained.

[0008] According to another specific embodiment of the present invention, band-pass filtering uses a band-pass filter with a frequency band of 4-40 Hz. After the EEG signal is processed by the band-pass filter, the EEG signal frequency band is obtained as theta waves of 4-8 Hz, alpha waves of 8-13 Hz, beta waves of 13-30 Hz and gamma waves of 30-40 Hz.

[0009] According to another specific embodiment of the present invention, the functional connectivity feature is the mutual information MI corresponding to theta waves, alpha waves, beta waves and gamma waves. The calculation formula of the mutual information MI is:

[0010]

[0011] MI(X; Y) represents the mutual information between X and Y, where X and Y are the EEG signals of two leads in the same frequency band, p x 、p y 、p xyThey represent the probability density function of X, the probability density function of Y, and the joint probability density function of X and Y respectively.

[0012] According to another specific embodiment of the present invention, the nonlinear dynamic characteristics are approximate entropies corresponding to theta waves, alpha waves, beta waves, and gamma waves. The calculation process of the approximate entropy includes:

[0013] Let X be the EEG signal of length N in a certain frequency band of a single lead, and construct an m-dimensional vector S(j) = [s(j), s(j+1)…s(j+m-1)] of X, j = 1, 2,…N-m+1;

[0014] Calculate the distance d between the two vectors of X. The distance d is calculated as d[S(i), S(j)] = max(|s(i+k)-s(j+k)|), i = 1, 2, ... N-m+1, j = 1, 2, ... N-m+1, i ≠ j, k = 0, 1, 2 ... m-1;

[0015] Given a threshold r, count the number of vectors S(i), i=1,2,…N-m+1 with distance d[S(i),S(j)]≤r from all other vectors, and divide the number by the total number of vectors S N-m+1 to get the intermediate variable C i (r);

[0016] Calculate the intermediate variable C between all vectors d (r), get the parameters of the m-dimensional vector parameter The calculation method is

[0017] Calculate the parameters of the m+1 dimensional vector Get the approximate entropy of X

[0018] According to another specific embodiment of the present invention, the power spectrum characteristic is the ratio between the corresponding power spectra of theta wave, alpha wave, beta wave and gamma wave, and the ratios between the power spectra include: Ealpha / Ebeta, (Ealpha+Etheta) / Ebeta, (Ealpha+Etheta) / (Ebeta+Egamma), Etheta / Ebeta, where Etheta, Ealpha, Ebeta, and Egamma are the power spectra of theta wave, alpha wave, beta wave, and gamma wave, respectively.

[0019] According to another specific embodiment of the present invention, before obtaining the normalized transformed awake state reference feature set and the normalized transformed other state feature sets, the method further includes: calculating a first awake parameter using the awake state reference feature set, assuming that the awake state reference feature set is M, and the calculation formula of the first awake parameter P is:

[0020] P=[p1,p2,…p k ],k=1,2,…,n

[0021] p k =[min(Mk);max(Mk)],k=1,2,…,n

[0022] k represents the kth feature of the awake state reference feature set M, p k (1) = min(Mk), min(Mk) means taking the minimum value of Mk, p k (2) = max(Mk), where max(Mk) represents the maximum value of Mk.

[0023] According to another specific embodiment of the present invention, obtaining a normalized and transformed awake state reference feature set includes: performing a first normalization process on the awake state reference feature set using a first awake parameter, and the calculation formula is:

[0024] Mk=[Mk-p k (1)] / [p k (2)-p k (1)]

[0025] Then take the median of Mk to obtain the normalized transformed awake state reference feature set B.

[0026] According to another specific embodiment of the present invention, before using the awake state reference feature set to calculate the first awake parameter, the awake state reference feature set is subjected to feature preprocessing. The feature preprocessing includes removing the 10% maximum value and the 10% minimum value in the awake state reference feature set to reduce the error caused by abnormal values in the normalization processing.

[0027] According to another specific embodiment of the present invention, when there are multiple wakefulness state reference feature sets, only the first wakefulness parameter of the first wakefulness state reference feature set is calculated, and the wakefulness state reference feature set is normalized using the first wakefulness parameter of the first wakefulness state reference feature set to obtain a normalized transformed wakefulness state reference feature set, and, the other state feature sets are normalized using the first wakefulness parameter of the first wakefulness state reference feature set to obtain normalized transformed other state feature sets.

[0028] According to another specific embodiment of the present invention, when there are multiple awake state reference feature sets, the features of the multiple awake state reference feature sets are fused into one awake state reference feature set, and the first awake parameter is calculated using the fused awake state reference feature set.

[0029] According to another specific embodiment of the present invention, obtaining the normalized transformed other state feature set includes: performing a first normalization process on the other state feature set using the first wakefulness parameter, assuming that the other state feature set is M*, and the calculation formula is:

[0030] M * k=[M * kp k (1)] / [p k (2)-p k (1)]

[0031] Then to M * Take the median k and obtain the normalized transformed other state feature set B*.

[0032] According to another specific embodiment of the present invention, the spatial distance between the normalized transformed other state feature set B* and the normalized transformed awake state reference feature set B is obtained. The calculation formula of the spatial distance D is:

[0033]

[0034] p is a user-defined real number.

[0035] According to another specific embodiment of the present invention, when there are multiple awake state reference feature sets B, the spatial distance D is the average of the spatial distances between the normalized transformed other state feature sets B* and the multiple normalized transformed awake state reference feature sets B.

[0036] According to another specific embodiment of the present invention, obtaining a normalized spatial distance value includes: performing a second normalization process on the spatial distance D to obtain a normalized spatial distance value. The calculation formula of the normalized spatial distance value S is:

[0037]

[0038] e is a natural base, a>0, a, b, and c are user-defined real numbers, and the normalized spatial distance value S ranges from 0 to 1.

[0039] According to another specific embodiment of the present invention, the degree of change in brain function state is judged based on the normalized spatial distance value S. If the normalized spatial distance value S≥0.5, it is judged that the brain function state has changed significantly. If the normalized spatial distance value S<0.5, it is judged that the brain function state has changed slightly.

[0040] An embodiment of the present invention also discloses a brain function status detection device, including: a shell, the shell is made of a flexible material, and a through hole is provided on the first surface of the shell; a memory, the memory is used to store one or more processing execution instructions of the brain function status detection device; an FPC board, the FPC board is arranged inside the shell, and a microcontroller unit is provided on the FPC board, the microcontroller unit is used to execute the instructions in the memory to execute the brain function status judgment method; an EEG acquisition electrode, the EEG acquisition electrode is arranged on the first surface through the through hole, the EEG acquisition electrode is electrically connected to the microcontroller unit, and the EEG acquisition electrode is used to collect EEG signals from the forehead; a first ear clip and a first connecting part, the first connecting part connects the first ear clip and the shell, the first ear clip includes a symmetrically arranged first magnetic attraction part, the first magnetic attraction part is provided with a bias acquisition electrode and a reference acquisition electrode, the bias acquisition electrode and the reference acquisition electrode are electrically connected to the microcontroller unit; a second ear clip and a second connecting part, the second connecting part connects the second ear clip and the shell, the first connecting part and the second connecting part are symmetrically arranged at both ends of the shell, and the second ear clip includes a symmetrically arranged second magnetic attraction part.

[0041] Using the above technical solution, the brain function status detection device processes collected EEG signals and analyzes and determines the degree of change in brain function status based on a defined judgment standard. This device integrates EEG signal collection and analysis, making it easy to use. Brain intervention components, such as electrical stimulation components and vibration stimulation components, can also be installed on the first and second magnetic attraction portions of the brain function status detection device. Based on the determination of the degree of change in brain function status by the microcontroller unit of the brain function status detection device, intervention measures can be implemented at specific moments.

[0042] An embodiment of the present invention further discloses a storage medium having instructions stored thereon, the instructions being used for a processor to load and execute the method for determining brain function status. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a flowchart of a method for determining brain function status according to an embodiment of the present invention.

[0044] Figure 2 It is a schematic structural diagram of the portion in contact with the brain in a device for detecting brain function status according to an embodiment of the present invention.

[0045] Figure 3 It is a schematic diagram of the overall structure of a brain function status detection device according to an embodiment of the present invention.

[0046] Figure 4 This is a schematic diagram of the changes in non-normalized spatial distance values of six subjects in a method for determining brain function status according to an embodiment of the present invention.

[0047] Figure 5 This is a schematic diagram of the changes in normalized spatial distance values of six subjects in a method for determining brain function status according to an embodiment of the present invention.

[0048] Description of the accompanying drawings in the specific embodiment:

[0049] 1. Housing 12.Through hole 2.FPC board 21.Microcontroller unit 3. EEG acquisition electrodes 4. The first ear clip 41.First magnetic attraction part 5. First connection part 6. Second ear clip 61. Second magnetic attraction part 7. Second connection part DETAILED DESCRIPTION

[0050] The following describes the embodiments of the present invention by means of specific embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Although the description of the present invention will be introduced in conjunction with the preferred embodiment, this does not mean that the features of this invention are limited to this embodiment. On the contrary, the purpose of introducing the invention in conjunction with the embodiment is to cover other options or modifications that may be extended based on the claims of the present invention. It should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other unless there is a conflict.

[0051] In the description of this embodiment, it should be noted that the terms "upper", "lower", "inner", "bottom", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, or are the orientations or positional relationships in which the inventive product is usually placed when in use. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, they cannot be understood as limitations on the present invention.

[0052] The terms “first”, “second”, etc. are only used for distinguishing descriptions and should not be understood as indicating or implying relative importance.

[0053] In the description of this embodiment, it should be noted that, unless otherwise specified or limited, the terms "disposed," "connected," and "connected" should be understood broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; and internal connections between two components. Those skilled in the art will understand the specific meanings of the above terms in this embodiment based on specific circumstances.

[0054] To make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0055] The brain function state judgment method of the present invention can obtain a quantitative value of the degree of change in the brain function state, and perform quantitative analysis of the brain function state based on the quantitative value of the degree of change in the brain function state.

[0056] like Figure 1As shown, the method for determining brain function status of the present invention includes the following steps:

[0057] Step S1: Obtain EEG signals in the awake state and EEG signals in other states.

[0058] Using a non-invasive technique that doesn't require surgical intervention, the device uses electrodes placed on a clean scalp to collect the subject's EEG signals. EEG signals record changes in electrical waves during brain activity and are a comprehensive reflection of the electrophysiological activity of brain nerve cells on the cerebral cortex or scalp surface.

[0059] EEG signals in the awake state need to be obtained when the subject's brain is in an initial state of relaxation. For example, if EEG signals are used to analyze driving fatigue, then the EEG signals in the awake state refer to the EEG signals when the subject has not yet driven or has just begun driving. If EEG signals are used to analyze the degree of brain workload, then the EEG signals in the awake state refer to the EEG signals when the subject has not yet started working or has just started working. If EEG signals are used to analyze the degree of concentration, then the EEG signals in the awake state refer to the EEG signals when the subject has not yet started learning or thinking or has just begun learning or thinking.

[0060] EEG signals in other states need to be obtained after the subject's brain has been performing a specific task for a period of time and is in a non-relaxed state. For example, if EEG signals are used to analyze driving fatigue, EEG signals in other states refer to those after the subject has been driving for a period of time; if EEG signals are used to analyze the degree of brain workload, EEG signals in other states refer to those after the subject has been working for a period of time; if EEG signals are used to analyze the degree of concentration, EEG signals in other states refer to those after the subject has been studying or thinking for a period of time.

[0061] There is no strict restriction on the specific time sequence for obtaining EEG signals in the awake state and EEG signals in other states. As long as the EEG signals in different states of the subject's brain are finally obtained, it can be selected independently according to the actual operation situation. For example, in a continuous period of time, the EEG signals in the awake state when the brain is in the initial relaxed state are first obtained. As the execution time of the specific task increases, the EEG signals in other states when the brain is in a non-relaxed state are then obtained. For example, in a non-continuous period of time, the EEG signals in other states when the brain is in a non-relaxed state after performing a specific task are first obtained, and at another time, the EEG signals in the awake state when the brain is in the initial relaxed state are obtained. Preferably, the EEG signals in the awake state are obtained before the brain starts to perform the specific task.

[0062] Step S2: Obtain the awake state reference feature set and other state feature sets.

[0063] Furthermore, in this application, the EEG signals in the awake state and the EEG signals in other states are processed respectively to obtain the awake state reference feature set and the other state feature set. The EEG signals in the awake state and the EEG signals in other states are processed in the same way. Here, taking the processing of the EEG signals in the awake state as an example, the processing process includes:

[0064] Select a certain length of EEG signal sample E in the awake state, calculate the n features of the sample, and construct the feature vector A of the sample E = [a1, a2, ..., a n ](n≧1, n is an integer), where A is an n-dimensional row vector. Select m samples E1, E2, ... E m (m≧1, m is an integer), construct the corresponding eigenvectors A1, A2...A m , get the awake state reference feature set A m×n =[A1,A2,……,A m ].

[0065] The specific process involves acquiring EEG signals during the awake state and then preprocessing them. This preprocessing includes threshold denoising, baseline drift removal, and bandpass filtering with a 4-40Hz bandpass filter. This preprocessing yields the following EEG signals: theta waves in the 4-8Hz range, alpha waves in the 8-13Hz range, beta waves in the 13-30Hz range, and gamma waves in the 30-40Hz range.

[0066] Functional connectivity features, nonlinear dynamic features, and power spectrum features of EEG signals in the awake state are calculated separately, and the functional connectivity features, nonlinear dynamic features, and power spectrum features are fused to obtain a reference feature set for the awake state. At least one reference feature set for the awake state is constructed, and one or more reference feature sets may be constructed.

[0067] Furthermore, in this application, the functional connectivity feature is the mutual information MI corresponding to theta wave, alpha wave, beta wave and gamma wave, and the calculation formula of mutual information MI is:

[0068]

[0069] MI(X; Y) represents the mutual information between X and Y, where X and Y are the EEG signals of two leads in the same frequency band, p x 、p y 、p xy They represent the probability density function of X, the probability density function of Y, and the joint probability density function of X and Y respectively.

[0070] Furthermore, in the present application, the nonlinear dynamic characteristics are the approximate entropies corresponding to theta waves, alpha waves, beta waves, and gamma waves. The calculation process of the approximate entropy includes:

[0071] Let X be the EEG signal of length N in a certain frequency band of a single lead, and construct an m-dimensional vector S(j) = [s(j), s(j+1)…s(j+m-1)] of X, j = 1, 2,…N-m+1;

[0072] Calculate the distance d between the two vectors of X. The distance d is calculated as d[S(i), S(j)] = max(|s(i+k)-s(j+k)|), i = 1, 2, ... N-m+1, j = 1, 2, ... N-m+1, i ≠ j, k = 0, 1, 2 ... m-1;

[0073] Given a threshold r, count the number of vectors S(i), i=1,2,…N-m+1 with distance d[S(i),S(j)]≤r from all other vectors, and divide the number by the total number of vectors S N-m+1 to get the intermediate variable C i (r);

[0074] Calculate the intermediate variable C between all vectors d (r), get the parameters of the m-dimensional vector parameter The calculation method is

[0075] Calculate the parameters of the m+1 dimensional vector Get the approximate entropy of X

[0076] Furthermore, in the present application, the power spectrum characteristic is the ratio between the corresponding power spectra of theta waves, alpha waves, beta waves and gamma waves. The ratios between the power spectra include: Ealpha / Ebeta, (Ealpha+Etheta) / Ebeta, (Ealpha+Etheta) / (Ebeta+Egamma), Etheta / Ebeta, where Etheta, Ealpha, Ebeta, and Egamma are the power spectra of theta waves, alpha waves, beta waves, and gamma waves, respectively.

[0077] Step S3: Obtain a normalized transformed awake state reference feature set and a normalized transformed other state feature set.

[0078] Furthermore, in the present application, a first wakefulness parameter is calculated using the wakefulness reference feature set, a first normalization process is performed on the wakefulness reference feature set using the first wakefulness parameter to obtain a normalized and transformed wakefulness reference feature set, and a first normalization process is performed on other state feature sets using the first wakefulness parameter to obtain normalized and transformed other state feature sets. That is, the method for performing the first normalization process on the wakefulness reference feature set is the same as the method for performing the first normalization process on the other state feature sets, the only difference being that the parameters used in the first normalization process are both the first wakefulness parameters calculated using the wakefulness reference feature set. The processing includes:

[0079] Before calculating the first wakefulness parameter using the wakefulness reference feature set, the wakefulness reference feature set is subjected to feature preprocessing. Feature preprocessing includes removing the 10% maximum and 10% minimum values in the wakefulness reference feature set to reduce the error caused by outliers in the normalization process. When there are multiple wakefulness reference feature sets, there are two processing methods. The first processing method is to fuse the features of multiple wakefulness reference feature sets into one wakefulness reference feature set, and use the fused wakefulness reference feature set to calculate the first wakefulness parameter. The second processing method is to only calculate the first wakefulness parameter of the first wakefulness reference feature set among the multiple wakefulness reference feature sets, use the first wakefulness parameter of the first wakefulness reference feature set to normalize the wakefulness reference feature set to obtain a normalized transformed wakefulness reference feature set, and use the first wakefulness parameter of the first wakefulness reference feature set to normalize other state feature sets to obtain normalized transformed other state feature sets.

[0080] The first wakefulness parameter P is calculated using the wakefulness reference feature set. Let the wakefulness reference feature set be M, and the calculation formula for the first wakefulness parameter P is:

[0081] P=[p1,p2,…p k ],k=1,2,…,n

[0082] p k =[min(Mk);max(Mk)], k=1,2,…,n

[0083] k represents the kth feature of the awake state reference feature set M, p k (1) = min(Mk), min(Mk) means taking the minimum value of Mk, p k (2) = max(Mk), where max(Mk) represents the maximum value of Mk.

[0084] The awake state reference feature set is subjected to a first normalization process using the first awake parameter to obtain a normalized awake state reference feature set. The calculation formula is:

[0085] Mk=[Mk-pk (1)] / [p k (2)-p k (1)]

[0086] Then, the median of Mk is taken to obtain the normalized transformed awake state reference feature set B, or the mean of Mk is taken to obtain the normalized transformed awake state reference feature set B.

[0087] The first awake parameter is used to perform the first normalization processing on the other state feature set to obtain the normalized transformed other state feature set. Let the other state feature set be M*, and the calculation formula is:

[0088] M * k=[M * kp k (1)] / [p k (2)-p k (1)]

[0089] Then to M * Take the median k and get the normalized transformed state feature set B*, or, for M * Take the mean of k and obtain the normalized transformed other state feature set B*.

[0090] Step S4: Obtain the spatial distance between the normalized transformed other state feature set and the normalized transformed awake state reference feature set.

[0091] Furthermore, in this application, the spatial distance between the normalized transformed other state feature set B* and the normalized transformed awake state reference feature set B is calculated, and the spatial distance D is calculated as follows:

[0092]

[0093] p is a user-defined real number.

[0094] When there are multiple normalized transformed awake state reference feature sets B, the spatial distance D is the average of the spatial distances between the normalized transformed other state feature sets B* and the multiple normalized transformed awake state reference feature sets B.

[0095] Step S5: Obtain a normalized spatial distance value.

[0096] Furthermore, in this application, the spatial distance D is subjected to a second normalization process to obtain a normalized spatial distance value. The calculation formula of the normalized spatial distance value S is:

[0097]

[0098] e is a natural base, a>0, a, b, and c are user-defined real numbers, and the normalized spatial distance value S ranges from 0 to 1. The spatial distance value S is used to measure the degree of change in brain function state.

[0099] The second normalization function S(·) is derived by improving the sigmoid function. By adjusting the custom real numbers a, b, and c, the change in brain function state is scaled to the interval [0, 1] to eliminate the differences caused by different value intervals. It can also adjust the parameters of the brain function state change curve to better reflect the changes in brain function state. The choice of the custom real numbers a, b, and c can be adjusted according to the specific application and is not limited by the present invention.

[0100] Step S6: Determine the degree of change in brain function state based on the normalized spatial distance value.

[0101] Furthermore, in the present application, if the normalized spatial distance value S≥0.5, it is judged that the brain function state has changed significantly; if the normalized spatial distance value S<0.5, it is judged that the brain function state has changed slightly.

[0102] The embodiment of the present invention also provides a brain function status detection device, referring to Figure 2 and Figure 3 As shown, the brain function status detection device includes: a shell 1, the shell 1 is a flexible material, and the shell 1 is provided with a through hole 12; a memory (not shown in the figure), the memory is used to store one or more processing execution instructions of the brain function status detection device; an FPC board 2, the FPC board 2 is arranged inside the shell 1, and a micro control unit 21 is provided on the FPC board 2, and the micro control unit 21 is used to execute the instructions in the memory to execute the brain function status judgment method; an EEG acquisition electrode 3, the EEG acquisition electrode 3 is arranged through the through hole 12, the EEG acquisition electrode 3 is electrically connected to the micro control unit 21, and the EEG acquisition electrode 3 is used to collect EEG signals from the forehead.

[0103] The housing 1 is made of a flexible material such as silicone or rubber. It is provided with a through-hole 12 that extends through the interior of the housing 1, but not entirely. Through-hole 12 allows the EEG acquisition electrode 3 to pass through. An FPC board 2, or flexible printed circuit (FPC), is provided within the housing 1. A microcontroller unit 21 is mounted on the FPC. This microcontroller unit 21 is electrically connected to the EEG acquisition electrode 3, which is in turn electrically connected to a memory device. This allows the microcontroller unit 21 to process the collected EEG signals and execute the method for determining brain function status. The EEG acquisition electrode 3 is provided through the through-hole 12 and is used to collect EEG signals from the forehead.

[0104] The brain function status detection device directly sets the EEG acquisition electrode 3 on the surface of the shell 1, and the signal is processed by the micro control unit 21 set in the shell 1. Different from the existing technology in which the data analysis equipment and the exposed EEG acquisition electrodes without a shell are connected by a long connecting line, the entire device is miniaturized, easy to carry and use, and has a wider range of usage scenarios.

[0105] During use, the housing 1 only needs to be placed in the corresponding monitoring position to perform monitoring. Furthermore, the microcontroller unit 21 is disposed within the housing 1, and no long connecting wires are required when electrically connecting the microcontroller unit 21 to the EEG acquisition electrodes 3, making EEG signal acquisition more stable and reliable. Furthermore, since the housing and circuit board in this embodiment are both made of flexible materials, they can adapt to the shape of the user's forehead during use, fitting well to the user's forehead and making it easier to fix, further improving the accuracy of the collected EEG signals.

[0106] Moreover, compared to ordinary PCB boards, the use of FPC boards can increase the maximum area of the circuit board within the limited flexible housing space. With the same number of circuit components, the FPC board can make the overall device more miniaturized. Within the same flexible housing space, more circuit components can be placed on the FPC board than on the PCB board, facilitating the functional expansion of the device. Moreover, the FPC board is thinner, so the thickness and weight of the device can be reduced, and the wearing comfort is enhanced. Preferably, the housing 1 is silicone, which facilitates the encapsulation of the FPC through low-temperature silicone vulcanization, ensuring that the circuit components are not damaged by high temperatures, and making the housing 1 have better toughness, thereby improving production efficiency, yield rate and comfort during use.

[0107] The brain function status detection device also includes a first ear clip 4 and a first connecting portion 5. The first connecting portion 5 connects the first ear clip 4 to the housing 1. The first ear clip 4 includes a symmetrically arranged first magnetic portion 41, on which a bias acquisition electrode (not shown) and a reference acquisition electrode (not shown) are disposed. The bias acquisition electrode and the reference acquisition electrode are electrically connected to the microcontroller unit 21. The magnetic first ear clip 4 uses magnets to stably and reliably secure the bias acquisition electrode and the reference acquisition electrode to the ear, thereby increasing the stability of the collected electrical signals. Placing the bias acquisition electrode and the reference acquisition electrode on the symmetrical first magnetic portion 41 makes wearing more comfortable and ensures more stable contact between the electrodes and the ear. The ear is closer to the forehead, which allows the first connecting portion 5 to be shorter, making it easier to use. Preferably, the first connecting portion 5 is a shielded wire wrapped in silicone to improve the electrical stability of the collected electrical signals. Preferably, the first connecting portion 5 is connected to the upper portion of the housing 1 to avoid interfering with the user's vision during use, making it suitable for various application scenarios.

[0108] Reference Figure 2 and Figure 3As shown, the brain function status detection device also includes a second ear clip 6 and a second connecting portion 7. The second connecting portion 7 connects the second ear clip 6 to the housing 1. The first connecting portion 5 and the second connecting portion 7 are symmetrically arranged along the two ends of the housing 1. The second ear clip 6 includes a symmetrically arranged second magnetic portion 61. The microcontroller unit 21 is electrically connected to the second magnetic portion 61.

[0109] For ease of understanding, only as an example, taking the degree of change in brain function fatigue during driving as an example, the process of applying the brain function state judgment method to judge the degree of brain function fatigue during driving is specifically explained.

[0110] In this example, EEG signals were collected from six subjects while they performed a 90-minute simulated driving fatigue task. The subjects' EEG signals were acquired by a brain function status detection device from two EEG acquisition electrodes, Fp1 and Fp2, located on the forehead. The collected EEG signals were processed by a microcontroller unit 21, which then executed the brain function status determination method. The brain function status detection device can also transmit the EEG signals to the cloud via a Bluetooth module, which then executes the brain function status determination method to assess the subjects' fatigue status.

[0111] The brain function state detection device continuously collects EEG signals from the subject from the start of the task, obtaining EEG signals in the awake state and EEG signals in other states. The EEG signals are preprocessed, including threshold denoising, baseline drift removal, and 4-40Hz bandpass filtering, to obtain EEG signals in different frequency bands: theta waves in the 4-8Hz band, alpha waves in the 8-13Hz band, beta waves in the 13-30Hz band, and gamma waves in the 30-40Hz band. Then, 4-second EEG signals are collected every 4 seconds, and the functional connectivity characteristics, nonlinear dynamic characteristics, and power spectrum characteristics are calculated according to the aforementioned step S2.

[0112] Step S2-1: Calculate functional connectivity features:

[0113] Calculate the mutual information MI corresponding to theta waves, alpha waves, beta waves, and gamma waves, and finally get four features. The calculation formula of mutual information MI is as follows:

[0114]

[0115] MI(X;Y) represents the mutual information between X and Y, p x 、p y 、p xy They represent the probability density function of X, the probability density function of Y, and the joint probability density function of X and Y, respectively, where X and Y are the EEG signals of the two leads in the same frequency band (theta wave, alpha wave, beta wave, gamma wave).

[0116] Step S2-2: Calculate nonlinear dynamic characteristics

[0117] Calculate the approximate entropy corresponding to theta waves, alpha waves, beta waves, and gamma waves, and finally get 8 features. The steps for calculating the approximate entropy are as follows:

[0118] Step S2-2-1: Let X be an EEG signal of length N in a certain frequency band of a single lead, X = [x(1), x(2)…x(N)], and construct an m-dimensional vector S(j) = [s(j), s(j+1)…s(j+m-1)] of X, j = 1, 2,…N-m+1.

[0119] Step S2-2-2: Calculate the distance d between the two vectors of X. This distance can be regarded as the maximum value of the difference between the corresponding elements of the two vectors d[S(i), S(j)] = max(|s(i+k)-s(j+k)|), i≠j, k=0,1,2…m-1.

[0120] Step S2-2-3: Given a threshold r, count the number of vectors S(i), i=1,2,…N-m+1 with distances d[S(i),S(j)]≤r from all other vectors, and divide this number by the total number of vectors S N-m+1 to obtain the intermediate variable C i (r).

[0121] Step S2-2-4: Repeat step S2-2-3 until the intermediate variable C between all vectors is found d (r), for C d (r) Take the logarithm and find the average to get the parameters of the m-dimensional vector

[0122] Step S2-2-5: Repeat steps S2-2-1 to S2-2-4 to calculate the parameters of the m+1 dimensional vector Get the approximate entropy of X

[0123] Step S2-3: Calculate power spectrum characteristics

[0124] Calculate Ealpha / Ebeta, (Ealpha+Etheta) / Ebeta, (Ealpha+Etheta) / (Ebeta+Egamma), and Etheta / Ebeta, and finally get 8 features, among which Etheta, Ealpha, Ebeta, and Egamma are the power spectra of theta wave, alpha wave, beta wave, and gamma wave respectively.

[0125] Construct feature vectors to obtain the reference feature set of the awake state and other state feature sets. Through step S2, 4 functional connectivity features, 8 nonlinear dynamic features, and 8 power spectrum features are obtained, a total of 20 features. 100 feature vectors containing these 20 features are used to construct an EEG signal feature vector A. 100×20 , the vector A 100×20 It can represent the awake state reference feature set or other state feature set, where the vector A 100×20 Each row represents a feature vector, and each column represents a class of features.

[0126] Then, according to step S2, more awake state reference feature sets and other state feature sets are constructed. Starting with the second feature set, each feature set consists of the last 95 feature vectors of the previous feature set and the five newly calculated feature vectors. The number of awake state reference feature sets and other state feature sets can be determined based on the specific application scenario and is not limited by this invention.

[0127] Calculate the first awake parameter. When using the first awake state reference feature set to calculate the first awake parameter, take the first awake state reference feature set A 100×20 , remove A 100×20 10% of the largest samples and 10% of the smallest samples of each type of feature in , that is, a total of 20 samples are eliminated to obtain the remaining feature set M 80×20 , for M 80×20 Calculate the first awake parameter for each type of feature, here the k-th type feature Mk 80×1 For example, the corresponding first awake parameter p k , the calculation formula for k=1,2,…20 is as follows:

[0128] p k =[min(Mk 80×1 );max(Mk 80×1 )]

[0129] where p k (1) = min(Mk 80×1 ), min(Mk 80×1 ) means taking Mk 80×1 The minimum value of p k (2) = max(Mk 80×1 ), max(Mk 80×1 ) means taking Mk 80×1 The maximum value of the first awake parameter P is finally obtained 1×20 =[p1,p2,…p 20 ].

[0130] The awake state reference feature set and other state feature sets are first normalized to obtain the normalized transformed awake state reference feature set and the normalized transformed other state feature sets. 100×1 Take the first normalization process as an example, Mk 100×1 The first normalization calculation formula is as follows:

[0131] Mk 100×1 =[Mk 100×1 -p k (1)] / [p k (2)-p k (1)].

[0132] Then, the first normalized Mk 100×1 Take the median to get Bk 1×1 Repeat the first normalization step to process the awake state reference feature set M 100×20 , and finally obtain the normalized transformed awake state reference feature set B 1×20 The first normalization process of the other state feature set M* is the same as the first normalization process of the awake state reference feature set M, which will not be repeated here. The first normalization process is used to process the other state feature set M*, and finally the normalized transformed other state feature set is obtained.

[0133] Calculate the spatial distance D between the normalized transformed other state feature sets and the normalized transformed awake state reference feature set. Use the EEG signals collected initially in the awake state to construct the awake state reference feature set, and use the first five awake state reference feature sets to calculate the first awake parameter. At this time, the features in the first five awake state reference feature sets are mixed. Since the overlap rate between feature sets is 95%, the feature set A is obtained after mixing. 120×20 , for A 120×20 After removing the largest and smallest 10% of samples, we get the remaining feature set M 96×20 , M 96×20 The first awake parameter p is calculated using the formula for calculating the first awake parameter. k . Use the first awake parameter p k After performing the first normalization processing on the five awake state reference feature sets and one other state feature set, the normalized awake state reference feature set is obtained: and, other state feature sets of normalized transformation The calculation formula of the spatial distance D between the normalized transformed other state feature set and the normalized transformed awake state reference feature set is as follows:

[0134]

[0135]

[0136] Wherein p=1, p is a custom parameter, and the value of p can be independently defined according to actual application conditions, and the present invention does not impose any limitation thereto.

[0137] The spatial distance D is subjected to a second normalization process to obtain a normalized spatial distance value S. The calculation formula of the normalized spatial distance value S is as follows:

[0138]

[0139] Where e is the natural base, a, b, and c are user-defined real numbers, and a=9, b=-1, and c=2 are selected respectively, that is:

[0140]

[0141] This value can effectively reflect the degree of fatigue change of the subjects in this task.

[0142] The degree of change in brain function state is judged according to the normalized spatial distance value S. In order to facilitate the observation and judgment of the changes in the brain function state of the subjects during the fatigue driving experiment, as shown in the figure below: Figure 4 As shown in FIG, the non-normalized spatial distance values S* of the brain function states of the six subjects are represented in the figure. Figure 5 As shown in the figure, the normalized spatial distance values S of the brain function states of the six subjects are respectively represented in the figure. Figure 4 In each subject's image, the horizontal axis represents the experimental time, and the vertical axis represents the non-normalized spatial distance value S*. The S* curve reflects the degree of fatigue change in the subject's brain function state over time. The difference between the non-normalized spatial distance value S* and the origin coordinates can be used to determine the degree of change. However, because the spatial distance value S* is not normalized, the range of non-normalized spatial distance value S* for each subject is not fixed, making further correlation analysis of fatigue levels between subjects impossible.

[0143] Figure 5In the figure, the horizontal axis of each subject's image represents the time of the experiment, and the vertical axis represents the normalized spatial distance value S. The change curve of S reflects the degree of fatigue change of the subject's brain function state over time. The degree of change can be judged based on the difference between the normalized spatial distance value S and the origin coordinate. Since the spatial distance value S is normalized and the change value of the brain function state is scaled to the interval [0,1], the fatigue degree between each subject can be further correlated with analysis. It can be considered that if the normalized spatial distance value S ≥ 0.5, it is judged that the brain function state has changed significantly, and if the normalized spatial distance value S < 0.5, it is judged that the brain function state has changed slightly. The normalized spatial distance value S retains the main information on the change in the subject's fatigue degree and can more intuitively reflect the subject's fatigue degree. In other embodiments, different normalized spatial distance values can be set to judge the degree of change in the brain function state according to the actual application situation, or the normalized spatial distance value can be combined with other features to comprehensively judge the degree of change in the brain function state. The present invention is not limited thereto.

[0144] An embodiment of the present invention further provides a storage medium having a computer program that can be run on a processor stored thereon, and when the computer program is executed by the processor, any of the above methods for determining a brain function state can be implemented. The storage medium may include a computer-readable recording / storage medium, such as a random access memory (RAM), a read-only memory (ROM), a flash memory, an optical disc, a magnetic disk, a solid-state disk, and the like. According to one or more embodiments, the controller is executed by a microprocessor programmed to perform one or more operations and / or functions described herein. According to one or more embodiments, the controller is executed in whole or in part by specially configured hardware, for example, by one or more application-specific integrated circuits or ASIC(s).

[0145] Although the present invention has been illustrated and described with reference to certain preferred embodiments thereof, it should be understood by those skilled in the art that the above description is provided as a further detailed description of the present invention in conjunction with specific embodiments thereof, and that the specific implementation of the present invention is not limited to these descriptions. Those skilled in the art may make various changes in form and details, including simple deductions or substitutions, without departing from the spirit and scope of the present invention.

Claims

1. A method for determining brain function status, characterized in that: include: Obtain EEG signals in the awake state and EEG signals in other states; Obtaining a reference feature set of a wakeful state and feature sets of other states; Obtaining a normalized transformed awake state reference feature set and a normalized transformed other state feature set; Obtaining the spatial distance between the normalized transformed other state feature set and the normalized transformed awake state reference feature set; Get the normalized spatial distance value; The degree of change in brain function status is determined based on the normalized spatial distance value; Obtaining the awake state reference feature set requires separately calculating the features of the EEG signals and fusing the features to obtain the awake state reference feature set. Obtaining the feature sets of other states also requires separately calculating the features of the EEG signals and fusing the features to obtain the feature sets of other states, wherein the features include: functional connectivity features, nonlinear dynamic features, and power spectrum features. At least one awake state reference feature set is constructed.

2. The method for determining brain function status according to claim 1, wherein: After obtaining the EEG signals in the awake state and the EEG signals in other states, the EEG signals are preprocessed. The preprocessing includes: threshold denoising, baseline drift removal, and bandpass filtering. After the preprocessing, theta waves, alpha waves, beta waves, and gamma waves of the EEG signals are obtained.

3. The method for determining brain function status according to claim 2, wherein: The band-pass filtering uses a band-pass filter with a frequency band of 4-40 Hz. After the EEG signal is processed by the band-pass filter, the EEG signal frequency bands of 4-8 Hz theta wave, 8-13 Hz alpha wave, 13-30 Hz beta wave and 30-40 Hz gamma wave are obtained.

4. The method for determining brain function status according to claim 2, wherein: The functional connectivity feature is the mutual information MI corresponding to the theta wave, the alpha wave, the beta wave, and the gamma wave. The calculation formula of the mutual information MI is: MI(X; Y) represents the mutual information between X and Y, where X and Y are the EEG signals of two leads in the same frequency band, p x 、p y 、p xy They represent the probability density function of X, the probability density function of Y, and the joint probability density function of X and Y respectively.

5. The method for determining brain function status according to claim 2, wherein: The nonlinear dynamic characteristics are approximate entropies corresponding to the theta wave, the alpha wave, the beta wave, and the gamma wave. The calculation process of the approximate entropy includes: Let X be the EEG signal of length N in a certain frequency band of a single lead, and construct an m-dimensional vector S(j) = [s(j), s(j+1)…s(j+m-1)] of X, j = 1, 2,…N-m+1; Calculate the distance d between the two vectors of X, where d[S(i), S(j)] = max(|s(i+k) - s(j+k)|), i = 1, 2, ... N - m + 1, j = 1, 2, ... N - m + 1, i ≠ j, k = 0, 1, 2 ... m - 1; Given a threshold r, count the number of vectors S(i), i=1,2,…N-m+1 with distances d[S(i),S(j)]≤r between them and all other vectors, and divide the number by the total number of vectors S N-m+1 to get the intermediate variable C i (r); Calculate the intermediate variable C between all vectors d (r), get the parameters of the m-dimensional vector The parameters The calculation method is Calculate the parameters of the m+1 dimensional vector Get the approximate entropy of X 6. The method for determining brain function status according to claim 2, wherein: The power spectrum characteristic is the ratio between the corresponding power spectra of the theta wave, the alpha wave, the beta wave and the gamma wave. The ratios between the power spectra include: Ealpha / Ebeta, (Ealpha+Etheta) / Ebeta, (Ealpha+Etheta) / (Ebeta+Egamma), Etheta / Ebeta, where Etheta, Ealpha, Ebeta, and Egamma are the power spectra of the theta wave, the alpha wave, the beta wave, and the gamma wave, respectively.

7. The method for determining brain function status according to claim 1, wherein: Before obtaining the normalized transformed awake state reference feature set and the normalized transformed other state feature sets, the method further includes: calculating a first awake parameter using the awake state reference feature set. Let the awake state reference feature set be M, and the calculation formula of the first awake parameter P is: P=[p1,p2,…p k ],k=1,2,…,n p k =[min(Mk);max(Mk)],k=1,2,…,n k represents the kth feature of the awake state reference feature set M, p k (1) = min(Mk), min(Mk) means taking the minimum value of Mk, p k (2) = max(Mk), where max(Mk) represents the maximum value of Mk.

8. The method for determining brain function status according to claim 7, wherein: The step of obtaining the normalized and transformed awake state reference feature set includes performing a first normalization process on the awake state reference feature set using the first awake parameter, and the calculation formula is: Mk=[Mk-p k (1)] / [p k (2)-p k (1)] Then take the median of Mk to obtain the normalized transformed awake state reference feature set B.

9. The method for determining brain function status according to claim 7, wherein: Before using the wakefulness reference feature set to calculate the first wakefulness parameter, the wakefulness reference feature set is subjected to feature preprocessing. The feature preprocessing includes removing the 10% maximum value and the 10% minimum value in the wakefulness reference feature set to reduce the error caused by abnormal values in the normalization process.

10. The method for determining brain function status according to claim 7, wherein: When there are multiple wakefulness state reference feature sets, only the first wakefulness parameter of the first wakefulness state reference feature set is calculated, and the wakefulness state reference feature set is normalized using the first wakefulness parameter of the first wakefulness state reference feature set to obtain a normalized transformed wakefulness state reference feature set, and the other state feature sets are normalized using the first wakefulness parameter of the first wakefulness state reference feature set to obtain normalized transformed other state feature sets.

11. The method for determining brain function status according to claim 7, wherein: When there are multiple awake state reference feature sets, the features of the multiple awake state reference feature sets are fused into one awake state reference feature set, and the first awake state parameter is calculated using the fused awake state reference feature set.

12. The method for determining brain function status according to claim 8, wherein: The obtaining of the normalized transformed other state feature set includes: performing a first normalization process on the other state feature set using the first wakefulness parameter, assuming the other state feature set is M*, and the calculation formula is: M * k=[M * k-p k (1)] / [p k (2)-p k (1)] Then to M * Take the median k and obtain the normalized transformed other state feature set B*.

13. The method for determining brain function status according to claim 12, wherein: The spatial distance D between the normalized transformed other state feature set B* and the normalized transformed awake state reference feature set B is calculated as follows: p is a user-defined real number.

14. The method for determining brain function status according to claim 13, wherein: When there are multiple awake state reference feature sets B, the spatial distance D is the average of the spatial distances between the normalized transformed other state feature set B* and the multiple normalized transformed awake state reference feature sets B.

15. The method for determining brain function status according to claim 13, wherein: Obtaining the normalized spatial distance value includes: performing a second normalization process on the spatial distance D to obtain a normalized spatial distance value. The calculation formula of the normalized spatial distance value S is: e is a natural base, a>0, a, b, and c are user-defined real numbers, and the numerical range of the normalized spatial distance value S is in the interval [0-1].

16. The method for determining brain function status according to claim 15, wherein: The degree of change in the brain function state is judged according to the normalized spatial distance value S. If the normalized spatial distance value S is ≥ 0.5, it is judged that the brain function state has changed significantly. If the normalized spatial distance value S is < 0.5, it is judged that the brain function state has changed slightly.

17. A device for detecting brain function status, characterized in that: include: A shell, the shell being made of a flexible material, and a first surface of the shell being provided with a through hole; a memory, the memory being configured to store one or more processing execution instructions of the brain function status detection device; An FPC board, the FPC board being arranged inside the housing, the FPC board being provided with a microcontroller unit, the microcontroller unit being configured to execute the instructions in the memory to perform the method for determining brain function status according to any one of claims 1 to 16; an EEG collection electrode, the EEG collection electrode being disposed on the first surface through the through hole, the EEG collection electrode being electrically connected to the microcontroller unit, and being used to collect EEG signals from the forehead; a first ear clip and a first connecting portion, wherein the first connecting portion connects the first ear clip and the housing, the first ear clip includes a symmetrically arranged first magnetic portion, a bias collection electrode and a reference collection electrode are provided on the first magnetic portion, and the bias collection electrode and the reference collection electrode are electrically connected to the microcontroller unit; A second ear clip and a second connecting portion, wherein the second connecting portion connects the second ear clip and the shell, the first connecting portion and the second connecting portion are symmetrically arranged at two ends of the shell, and the second ear clip includes a symmetrically arranged second magnetic portion.

18. A storage medium, characterized in that The storage medium stores instructions, which are used by a processor to load and execute the method for determining brain function status according to any one of claims 1 to 16.

Images