Brain age prediction method and system, and storage medium

Abstract

The present invention belongs to the technical field of brain age prediction. Specifically disclosed are a brain age prediction method and system, and a storage medium. The method comprises: collecting electroencephalogram data of a subject in a resting state; inputting the electroencephalogram data into three pre-trained models to respectively obtain three predicted ages, i.e. predicted age I, predicted age II, and predicted age III; and determining whether the three predicted ages all belong to the same age group, thereby performing processing depending on the situation. The present invention integrates prediction results of three models, and aims to combine the advantages of three models to make the prediction results more accurate.

Description

A brain age prediction method and system, storage medium Priority application

[0001] This application claims priority to Chinese Invention Patent Application No. 2024111349857, filed August 16, 2024, entitled “A brain age prediction method and system, storage medium,” which is incorporated by reference in its entirety. TECHNICAL FIELD

[0002] The present invention belongs to the technical field of brain age prediction, and particularly relates to a brain age prediction method and system, computer readable storage medium. BACKGROUND

[0003] A person's age can be distinguished by actual age (calculated by the length of life), physiological age (the age of the function of other important organs except the brain), and brain age (BA), which is used to evaluate the degree of brain aging. The brain will gradually atrophy and function decline with age, and the speed and trajectory of this change vary significantly among brain regions and individuals. Regarding the concept of differential aging of the brain, for example, a 50-year-old person can “have a 20-year-old heart,” or a 30-year-old smoker's lungs “work like an 80-year-old,” which is manifested in clinical practice and is gaining scientific traction. Similarly, while brain function decline is a “normal” part of aging, some people apparently age faster than others. The degree of deviation between the brain aging trajectory of an individual subject and the typical brain aging trajectory can help diagnose, make treatment decisions, judge prognosis, and early identify brain-based diseases, especially neurodegenerative diseases.

[0004] There is no standard method or device for brain age measurement in the clinic at present. The judgment is mainly based on the subjective judgment of the doctor on the brain function of the individual object, including cognitive function (logical thinking function, recognition function, etc.), motor function, sensory function (hearing, vision, etc.). Secondly, some intelligence scales are also used in the clinic to judge the cognitive function of individual objects. Although this method quantifies the standard, it can only reflect a certain aspect of brain function and cannot fully represent the brain age. Recently, magnetic resonance imaging (MRI) studies have identified structural biomarkers of aging, such as axonal demyelination and reduction in gray matter and white matter volume, which contribute to the estimation of brain age index of general nervous system health and function. However, there are still two limitations in the practical application of structural MRI to predict brain age. One limitation is that the biomarkers of brain structure aging are not always well correlated with cognitive function. The little-known phenomenon of cognitive reserve and neural reserve is an example, that is, the brain showing Alzheimer's disease-related amyloid pathology and structural damage can also maintain a high level of function. Another limitation is the cost. It is well known that MRI is costly, and it would be very expensive to conduct extensive early and repeated screening, and it is unlikely to be paid by insurance companies.

[0005] The electroencephalogram signal is closely related to age as an important aspect of objectively reflecting the brain function state. The brain wave development and change can be observed with increasing age. Brain waves can show obvious differences at different age stages. For example, the brain waves of infants are more significantly affected by various factors than those of adults, and obvious brain wave abnormalities are more likely to occur, and the range of abnormalities is also wider, but the corresponding disappearance is faster than that of adults. By adulthood, brain waves gradually stabilize, and after middle age, brain waves change accordingly with the gradual decline of brain function. In addition, electroencephalogram imaging is not a brain structure but a brain function, and has been proven to be effective in capturing the contribution of cognitive reserve to brain health and cognitive function. The hardware cost of electroencephalogram is relatively economical and affordable, and it is convenient to conduct extensive screening and use in multiple scenarios. SUMMARY

[0006] The purpose of the present application is to provide a brain age prediction method and system, and a computer readable storage medium, which partially solve or alleviate the above-mentioned deficiencies in the prior art, and comprehensively predict the results of three models to make the final predicted age more consistent with the actual brain age.

[0007] In order to solve the above-mentioned technical problems, the present application specifically adopts the following technical solutions:

[0008] A brain age prediction method, comprising:

[0009] Collecting electroencephalogram data of the subject in a resting state;

[0010] The EEG data is input into the three pre-trained models to obtain three predicted ages, namely predicted age I, predicted age II, and predicted age III;

[0011] When the three predicted ages belong to the same age group, the data processing method I is used to obtain the final predicted value;

[0012] When only two of the three predicted ages belong to the same age group, the data processing method II is used to obtain the brain age prediction value;

[0013] The data processing method I comprises:

[0014] When the difference between the maximum value and the minimum value in the three predicted ages is less than or equal to the age difference threshold I, the final predicted age is calculated using the formula: A r =(A I +A II +A III ) / 3;

[0015] When the difference between the maximum value and the minimum value in the three predicted ages is greater than the age difference threshold I, the reliability of the three predicted ages is calculated, and the three predicted ages are given weights according to the reliability;

[0016] The final predicted age is calculated using the formula: A r =k1*A I +k2A II +k3*A III ; and

[0017] The data processing method II comprises:

[0018] When the difference between the maximum value and the minimum value in the three predicted ages is less than or equal to the age difference threshold II, the final predicted age is calculated using the formula: A r =(A I +A II +A III ) / 3; otherwise, the EEG data of the subject in the resting state is re-collected for re-prediction; wherein A r is the final predicted value, A I , A II , and A III are the predicted age I, the predicted age II, and the predicted age III, respectively, k1, k2, and k3 are the weights of the three predicted ages, and the age difference threshold II is less than the age difference threshold I.

[0019] As an improvement, the method for calculating the reliability of the predicted age comprises using the formula: d=(s t -s p ) / s t, calculate the reliability of the predicted age; wherein, d is the reliability of the current predicted age, s t is the sum of the absolute values of the differences between the three predicted ages, s p is the sum of the absolute values of the differences between the current predicted age and the other two predicted ages.

[0020] As an improvement, the method of assigning weights to the predicted age using reliability includes using the formula: k = d / d t , calculate the weight of the predicted age, wherein, k is the weight of the current predicted age, d is the reliability of the current predicted age, d t is the sum of the reliabilities of the three predicted ages.

[0021] As an improvement, when collecting the electroencephalogram data of the subject in a resting state, a mobile wireless electroencephalogram headset is used for collection;The mobile wireless electroencephalogram headset uses 14 salted felt sensors and 2 reference sensors for electroencephalogram data collection, and uses 0.2-45.0Hz band-pass filter and 50Hz or 60Hz notch filter for recording.

[0022] As an improvement, the positions of the 14 salted felt sensors are AF3 / 4, F3 / 4, F7 / 8, FC5 / 6, T7 / 8, P7 / 8 and O1 / 2;The positions of the two reference sensors are P3 / 4.

[0023] As an improvement, when collecting electroencephalogram data, the subject alternates between open eyes and closed eyes at 1 minute intervals;And collect three groups with 4 minutes as a group.

[0024] As an improvement, the three models are radial basis kernel support vector regression model, elastic net regression model and radial basis function kernel Gaussian process model.

[0025] As an improvement, when training the model, the electroencephalogram data of children under 12 years old, patients with psychiatric and brain diseases and people with bad habits are excluded from the data set.

[0026] As an improvement, the age range is divided into 13-20 years old, 21-30 years old, 31-60 years old, 61 years old and above.

[0027] The present application also provides a brain age prediction system, comprising:

[0028] The electroencephalogram data acquisition module is used for collecting the electroencephalogram data of the subject in a resting state;

[0029] The prediction module is used for inputting the electroencephalogram data into the three pre-trained models to obtain three predicted ages, namely predicted age I, predicted age II and predicted age III;

[0030] an ultimate predicted age acquisition module, configured to, in the case that all of the three predicted ages belong to the same age group, obtain an ultimate predicted value by using a data processing method I;

[0031] in the case that only two of the three predicted ages belong to the same age group, obtain a brain age predicted value by using a data processing method II;

[0032] the data processing method I comprises: in the case that the difference between the maximum value and the minimum value of the three predicted ages is less than or equal to an age difference threshold I, calculating an ultimate predicted age by using a formula: A r =(A I +A II +A III ) / 3;

[0033] in the case that the difference between the maximum value and the minimum value of the three predicted ages is greater than the age difference threshold I, calculating the reliability of the three predicted ages and assigning weights to the three predicted ages according to the reliability;

[0034] calculating an ultimate predicted age by using a formula: A r =k1*A I +k2A II +k3*A III ;

[0035] the data processing method II comprises: in the case that the difference between the maximum value and the minimum value of the three predicted ages is less than or equal to an age difference threshold II, calculating an ultimate predicted age by using a formula: A r =(A I +A II +A III ) / 3; otherwise, re-collecting the brain electrical data of the subject in a resting state to perform re-prediction; wherein, A r is an ultimate predicted value, A I , A II and A III are predicted age I, predicted age II and predicted age III respectively, k1, k2 and k3 are weights of the three predicted ages respectively, and the age difference threshold II is less than the age difference threshold I.

[0036] The application further provides a computer readable storage medium, which has a computer program stored thereon, the program being executed by a processor to implement the brain age prediction method.

[0037] The brain age measurement based on electroencephalogram has the advantages that (1) the brain age based on electroencephalogram can reflect functional changes rather than structural changes; (2) the data is more objective and more convincing; (3) electroencephalogram is more friendly to participants, has less cost and contraindications, and can be measured by a home device in principle; (4) the brain age based on electroencephalogram can facilitate repeated measurements of participants to evaluate the effectiveness of interventions, such as drugs or brain stimulation aimed at maintaining and improving brain function and longevity. Therefore, in measuring brain age, machine learning can be used to provide an easily accessible and low-cost biomarker of brain health.

[0038] On the basis described above, the present application combines the prediction results of the three models, and processes the three predicted ages by technical means to obtain a predicted age with higher reliability. First, it is judged whether the three predicted ages are in the same age range, and then the situation is handled. The brain age is divided into four stages, namely 13-20 years old, 21-30 years old, 31-60 years old, and 61 years old and above. The electroencephalogram data of the above four age ranges will have more obvious differences. By judging whether it is in an age range, a preliminary reliability judgment is made on the three predicted ages. For the case that the three predicted ages are not in the same age range, it indicates that there is a problem in the collection of electroencephalogram data, and the standby electroencephalogram data needs to be re-input or re-collected. For the case that the three predicted ages are in the same age range, it indicates that the predictions of the three models are roughly accurate, and only need to be processed according to the data processing method I. The data processing method I has two cases. In the case of relatively compact data, the mean value of the three is taken as the final result. For the case of relatively loose data, the reliability of the three predicted ages is calculated, and the three predicted ages are weighted according to the reliability. Finally, the proportion of the three predicted ages in the final result is adjusted by the weight, so that a more accurate prediction result is obtained more scientifically. For the case that two of the three predicted ages are in the same age range, since one of the predicted ages has crossed the age range, more caution should be taken for this case, and the data should be more compact before being adopted.

[0039] The present application comprehensively utilizes the three prediction results of the three models by judging the age range first and then processing the situation, and finally obtains a more accurate prediction result. BRIEF DESCRIPTION OF DRAWINGS

[0040] In order to describe the technical solutions of the embodiments of the present application or the prior art more clearly, the accompanying drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced. In all the drawings, similar elements or parts are generally identified by similar reference signs. In the drawings, each element or part is not necessarily drawn according to the actual proportion. Obviously, the drawings described below are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0041] Fig. 1 is a flow chart of the present application;

[0042] Fig. 2 is a schematic diagram of the position of the electroencephalogram data acquisition;

[0043] Fig. 3 is a structural diagram of the present application. DETAILED DESCRIPTION

[0044] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions of the embodiments of the present application will be described clearly and completely below with reference to the drawings of the embodiments of the present application. Obviously, the described embodiments are some of the embodiments of the present application, not all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0045] In this document, the suffixes such as "module", "part" or "unit" used to indicate elements are only for the convenience of describing the present application, and have no specific meaning. Therefore, "module", "part" or "unit" can be used interchangeably. In this document, the terms "upper", "lower", "inner", "outer", "front", "back", "one end", "the other end" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first", "second" are only for the purpose of description, and cannot be understood as indicating or implying relative importance.

[0046] In this article, unless otherwise specified and limited, the terms "mounting", "provided with", "connection" and the like should be understood broadly, for example, "connection" can be fixed connection, can also be detachable connection, or integrally connected; can be mechanical connection, can be directly connected, can also be indirectly connected through intermediate medium, can be the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances. In this article, "and / or" includes any and all combinations of one or more listed related items. In this article, "multiple" means two or more, that is, it includes two, three, four, five, etc.

[0047] Embodiment one: the prior art Chinese patent application CN201910701028.0 discloses an age classification method based on resting state electroencephalogram data, including the following steps: S1, collecting the original resting state electroencephalogram data corresponding to each age group; S2, preprocessing the original resting state electroencephalogram data to obtain the resting state electroencephalogram data without artifacts; S3, constructing a convolutional neural network, and inputting the resting state electroencephalogram data without artifacts to the convolutional neural network, training and testing the convolutional neural network to obtain the trained convolutional neural network; S4, classifying the actual resting state electroencephalogram data by the trained convolutional neural network according to age.

[0048] In the above-mentioned prior art, the resting state electroencephalogram data of each age group is collected as a training set to train a convolution model, and the convolution model is used to predict the test sample, so as to obtain the brain age of the patient. However, in actual use, due to the limitation of the model itself, the prediction result will deviate. On the other hand, as the age increases, the difference between the brain age and the actual age will become larger and larger, and the limitation of the single model will be further magnified.

[0049] In order to solve this problem, as shown in FIG. 1, the present application provides a brain age prediction method, which combines the prediction results of three models, and strives to combine the advantages of the three models to make the prediction result more reliable. The specific steps include:

[0050] S1, collecting the electroencephalogram data of the subject in the resting state.

[0051] FIG. 2 marks the position of the subject collecting the electroencephalogram data. In this embodiment, when collecting the electroencephalogram data of the subject in the resting state, a mobile wireless electroencephalogram headset is used for collection; the mobile wireless electroencephalogram headset uses 14 saline-soaked felt sensors (saline-soaked felt sensors, which are based on salt solution soaked felt pads attached to the brain to collect data) and 2 reference sensors to collect electroencephalogram data, and uses a 0.2-45.0 Hz band-pass filter and a 50 Hz or 60 Hz notch filter to record.

[0052] More specifically, the 14 salted felt sensor positions are AF3 / 4, F3 / 4, F7 / 8, FC5 / 6, T7 / 8, P7 / 8 and O1 / 2, respectively; and the 2 reference sensor positions are P3 / 4.

[0053] In addition, when collecting the electroencephalogram data, the subject alternates between open eyes and closed eyes at 1-minute intervals; and 3 groups of 4 minutes are collected in total, with 6 minutes of open eyes and 6 minutes of closed eyes. A short rest can be taken between two collections to reduce drowsiness.

[0054] Since the electroencephalogram data is severely affected by short-term external interference, in this embodiment, multiple groups are collected in advance to avoid re-collection.

[0055] S2 inputs the electroencephalogram data into the three pre-trained models to obtain three predicted ages, namely predicted age I, predicted age II and predicted age III.

[0056] In this embodiment, the three models used to provide predictions are a radial basis kernel support vector regression model, an elastic net regression model and a radial basis function kernel Gaussian process model.

[0057] The three models have advantages and disadvantages when modeling and predicting:

[0058] 1. Radial basis kernel support vector regression (SVR).

[0059] Advantages: capable of handling nonlinear relationships, suitable for complex data. Good performance for high-dimensional data. Generally robust to noisy data.

[0060] Disadvantages: selecting the appropriate kernel function and parameters (such as C and γ) is complex. Longer training time, especially on large-scale data sets.

[0061] 2. Elastic net regression.

[0062] Advantages: combines L1 (Lasso) and L2 (Ridge) regularization, which helps feature selection and reduces overfitting. Simple to use, high computational efficiency. Effective for multiple collinearity problems.

[0063] Disadvantages: suitable for linear relationships, limited ability to model nonlinear relationships. Cross-validation is required to select the regularization parameter, which is a complex process.

[0064] 3. Radial basis function kernel Gaussian process (GPR).

[0065] Advantages: good theoretical foundation, can provide uncertainty estimates. Suitable for small data sets and can model complex nonlinear relationships. Kernel functions can be selected flexibly.

[0066] Disadvantages: high computational complexity, especially when the data volume is large, the training and inference speed is slow. For large-scale data sets, the storage and computing overhead is very high.

[0067] In this embodiment, when training the above three models, the process of collecting electroencephalogram data as a data set is similar to step S1, which will not be described in detail in this step. It is worth noting that when collecting the data set to train the model, the electroencephalogram data of children under the age of 12, patients with psychiatric and brain diseases, and people with bad habits such as smoking, alcoholism, drug dependence, etc. should be excluded from the data set.

[0068] For normal people, we can consider that their brain age is basically consistent with their actual age. That is, the brain age of a normal person will be within one standard deviation of the actual age. According to existing data, the standard deviation is ±7.6 years. That is, a normal 30-year-old subject will have a brain age between 22.4 and 37.6.

[0069] For normal people, the electroencephalogram data of different brain ages in the resting state has certain characteristics. The three models in this application are based on this principle to predict brain age.

[0070] It can be understood that the three groups of people excluded from the data set will have brain ages that do not match their actual ages due to immature brain development, disease, drugs, etc. Therefore, they cannot be used as training sets and validation sets to train the model.

[0071] By inputting the electroencephalogram data of the subject into the above three models, three predicted ages, namely predicted age I, predicted age II, and predicted age III, are obtained. In an ideal case, the three predicted ages should be consistent. However, in reality, due to the advantages and disadvantages of the three models, the predicted brain ages will also have deviations. Therefore, in the subsequent steps, technical means are needed to process the three predicted ages to obtain a predicted age with higher reliability.

[0072] S3 in the case where the three predicted ages belong to the same age group, the data processing method I is used to obtain the final prediction value;

[0073] In the case where only two of the three predicted ages belong to the same age group, the data processing method II is used to obtain the brain age prediction value;

[0074] The data processing method I includes:

[0075] In the case where the difference between the maximum value and the minimum value of the three predicted ages is less than or equal to the age difference threshold I, the formula A r =(A I +A II +AIII ) / 3, calculate the final predicted age;

[0076] If the difference between the maximum and minimum values ​​of the three predicted ages is greater than the age difference threshold I, calculate the confidence of the three predicted ages and assign weights to the three predicted ages based on the confidence.

[0077] Using formula: A r =k1*A I +k2A II +k3*A III Calculate the final predicted age;

[0078] The data processing method II includes:

[0079] When the difference between the maximum and minimum values ​​of the three predicted ages is less than or equal to the age difference threshold II, the formula is used: A r =(A I +A II +A III ) / 3, calculate the final predicted age; otherwise, recollect the subject's resting EEG data for re-prediction; where A r For the final predicted value, A I A II A III The predicted ages are I, II, and III, respectively, with k1, k2, and k3 being the weights of the three predicted ages. The age difference threshold II is less than the age difference threshold I.

[0080] In this invention, brain age is first segmented into three groups: 13-20 years old, 21-30 years old, 31-60 years old, and 61 years old and above.

[0081] The EEG data for the four age groups mentioned above show significant differences.

[0082] Adolescence 13-20 years: The occipital rhythm is between 8-12 Hz, with an average of 10 Hz. The amplitude gradually decreases to 50-75 μV. The posterior slow wave activity gradually decreases, with low-moderate amplitude θ and δ band slow waves not exceeding 10-15%. In normal adolescents, slow waves are often more widespread, rather than being mainly in the posterior head. Fast wave activity in this age group is more and more widespread during the waking period, with the largest number in the frontal region, and also in the central and occipital regions. The μ rhythm is relatively common, but gradually decreases from 15 years of age. The occipital λ wave is also relatively common. The anterior head 6-7 Hz θ activity gradually decreases after 13-15 years of age. There is no low voltage in the normal EEG before 13 years of age. However, with the maturation of the EEG development after adolescence, the amplitude of the basic EEG activity decreases, with about 5% showing low voltage, mainly in the form of low-amplitude fast waves in the β band, more common in females, and possibly with a genetic factor.

[0083] Young adults 21-30 years: After 20 years of age, the EEG wave development is basically mature, with the occipital α rhythm stable at 9-10 Hz, with a normal full-head α wave frequency variation range of no more than 2 Hz, and a frequency difference of no more than 0.5 Hz in the corresponding regions of the two hemispheres. The amplitude further decreases, between 10-100 μV, with an average of 50-70 uV, and the amplitude shows a good amplitude modulation phenomenon. The individual difference in the α index is large, from 10% to 90%, generally 60%-70%. The low-amplitude β activity of 14-30 Hz during the waking period increases with age, mainly distributed in the frontal, central, and anterior and middle temporal regions, with only the occipital region appearing when the α rhythm is blocked. The low-amplitude θ activity of 4-7 Hz is mainly distributed in the frontal, central, and in addition to the temporal region, a small amount of θ activity can also be seen, with an index of 4%-5%. The slow wave activity of the 8 band is very small, with an amplitude of 20 μV or less, scattered in the frontal region, with an index of no more than 5%. The μ rhythm can be seen in the Rolandic region, mainly distributed in the central region, and sometimes also in the parietal region, seen in 19% of healthy adults. The k (Kappa) rhythm is seen in 30% of normal adults, with a frequency of 6-12 Hz and an amplitude of 10-40 μV, mainly appearing in the frontal and temporal regions during mental activity. The occipital λ wave can be seen when scanning complex patterns.

[0084] Middle-aged people 31-60 years old: After entering middle age, the brain waves are still in a relatively stable state. However, most middle-aged people are in different degrees of "sub-health" state, and this feature is also reflected in different degrees in the brain waves. In the basic healthy state, without obvious nervous system diseases or systemic diseases, the brain waves of middle-aged people basically maintain the characteristics of normal brain waves of young people. However, when there are other organic or functional lesions other than the nervous system, some marginal or mild abnormal brain wave manifestations may occur, such as decreased α index, slightly increased slow wave activity, poor regulation of α rhythm, poor amplitude modulation, and slightly increased temporal region 0 activity. A small number of middle-aged and elderly people over 50 years old may have sharp θ rhythm in the temporal region, which is more common in the drowsy period and light sleep period. A small number of adults may see widely distributed 5-7 Hz sharp or sinusoidal waves in the clear period and drowsy period, which are most prominent in the parietal and temporal regions, often bilateral, but can be asymmetric, lasting for several seconds to several minutes, and can suddenly appear and disappear, or gradually appear and disappear. The incidence of this subclinical rhythmicity in people over 50 years old is 0.5%, which belongs to a benign electroencephalogram pattern.

[0085] Old people 60 years old and above: Even if the basic health of the elderly, the nervous system will show different degrees of degenerative changes with age, including selective reduction in the number of neurons, enlargement of the ventricular system, reduction in brain weight, and visible neuronal shrinkage, lipofuscin or amyloid deposition, etc. These changes are most obvious in the frontal and temporal lobes. In addition, some hidden diseases may exist in generally healthy elderly people, such as cerebral arteriosclerosis or hidden diabetes, etc. These changes will affect the changes in cerebral blood flow and brain metabolism, thereby affecting the brain waves to varying degrees.

[0086] Changes in α rhythm: Old people often show slow α rhythm, reduced amplitude, and decreased α index, among which the decrease in fast α of 10-13 Hz is more obvious than the decrease in slow α of 8-9 Hz. The α rhythm of people over 60 years old is usually 8.5-9.5 Hz, with a decrease in fast α of 10-13 Hz, and the basic α rhythm is 1 Hz slower than that of young people on average. With increasing age, α waves further slow down after entering old age. According to a continuous follow-up study, the frequency of α rhythm decreases by 0.08 Hz per year on average after the age of 60. If the occipital α rhythm of the elderly is lower than 8.5 Hz in a completely clear state, it should be considered abnormal. In addition, with aging, the amplitude of α rhythm decreases, mostly around 30 μV, and becomes more fragmented and discontinuous. The distribution range of α waves in the elderly expands, which can extend to the frontal and temporal regions, making the difference between the front and back of the head disappear. Sometimes, the synchronization trend of the whole head can be seen to increase. Some people believe that this generalization of α rhythm may also be related to the decrease in alertness or mild drowsiness during recording.

[0087] The change of β activity: normally, β activity increases gradually with age from birth to adulthood. Before 50 years old, β activity accounts for about 20% in the wakeful brain waves, mainly in the frontal and central regions. After 50 years old, it accounts for about 30%. After 60 years old, β activity still increases slowly, but tends to decrease after 80 years old. The β / α ratio of the old people increases with age, which can be used as an index to evaluate the aging.

[0088] Therefore, in the present application, it is firstly judged whether the predicted ages predicted by the three models are in the same age stage. Therefore, there are three cases, i.e. the three predicted ages are all in the same age stage, only two predicted ages are in the same age stage, and the three predicted ages are not in the same age stage.

[0089] For the case that the three predicted ages are not in the same age stage, it indicates that there is a problem in collecting the electroencephalogram data, and the standby electroencephalogram data needs to be input again or be collected again.

[0090] For the case that the three predicted ages are all in the same age stage, it indicates that the predictions of the three models are roughly accurate, and only the data processing method I needs to be performed.

[0091] In the data processing method I, there are two cases:

[0092] One is that the difference between the maximum value and the minimum value of the three predicted ages is less than or equal to the age difference threshold I. In the present embodiment, the age difference threshold I is preset to be 6, that is, the difference between the maximum value and the minimum value of the predicted ages of the three models does not exceed 6, for example, 20, 21 and 25.

[0093] In the above case, the three predicted ages are relatively compact, and it can be considered that the credibility of the three predicted ages is relatively high, and therefore the average value of the three predicted ages is taken as the final predicted age, i.e. A r =(A I +A II +A III ) / 3.

[0094] Two is that the difference between the maximum value and the minimum value of the three predicted ages is greater than the age difference threshold I. That is, the difference between the maximum value and the minimum value of the predicted ages of the three models has exceeded 6, for example, 20, 23 and 28.

[0095] In the above case, the three predicted ages are relatively loose, and therefore the credibility of the three predicted ages needs to be calculated, and the three predicted ages are weighted by the credibility. The higher the credibility, the higher the weight, and the lower the credibility, the lower the weight. Specifically, in the present embodiment, the formula A r =k1*A I +k2A II +k3*A III; calculate the final predicted age; the method for calculating the predicted age reliability comprises using the formula: d = (s t -s p ) / s t ; calculate the reliability of the predicted age; wherein d is the reliability of the current predicted age, s t is the sum of the absolute values of the differences between the three predicted ages, and s p is the sum of the absolute values of the differences between the current predicted age and the other two predicted ages.

[0096] For example, the predicted age I, the predicted age II, and the predicted age III are 20, 23, and 28 respectively. Then the absolute values of the differences between the predicted age I and the other predicted ages are 3 and 8 respectively, the absolute values of the differences between the predicted age II and the other predicted ages are 3 and 5 respectively, and the absolute values of the differences between the predicted age III and the other predicted ages are 8 and 5 respectively. That is, the sum s t of the absolute values of the differences between the three predicted ages is 3+5+8=16. Then the reliability of the predicted age I is (16-11) / 16=5 / 16, the reliability of the predicted age II is 8 / 16, and the reliability of the predicted age III is 3 / 16.

[0097] Then use the formula: k = d / d t ; calculate the weight of the predicted age, wherein k is the weight of the current predicted age, d is the reliability of the current predicted age, and d t is the sum of the reliabilities of the three predicted ages.

[0098] The final predicted value can be calculated by substituting the calculated weights of the three predicted ages into the above formula for calculating the final predicted value.

[0099] For the case where two of the three predicted ages are in the same age range, since one of the predicted ages has crossed the age range at this time, more caution should be taken, and the data should be more compact. That is, the difference between the maximum value and the minimum value of the three predicted ages is less than or equal to the age difference threshold II, which is preset to be 4, which is smaller than the age difference threshold I, so as to ensure that the three predicted ages are in a smaller range.

[0100] If the condition is met, the average of the three predicted ages is taken as the final predicted value. Otherwise, it is considered that the collected electroencephalogram data is incorrect, and the backup data needs to be input for prediction again or the electroencephalogram data needs to be re-collected for prediction.

[0101] Embodiment 2: As shown in FIG. 3, the present application also provides a brain age prediction system, comprising:

[0102] an electroencephalogram data collection module for collecting electroencephalogram data of a subject in a resting state;

[0103] a prediction module, configured to input the electroencephalogram data into three pre-trained models to obtain three predicted ages, i.e., a predicted age I, a predicted age II, and a predicted age III;

[0104] a final predicted age acquisition module, configured to, in a case where the three predicted ages belong to the same age group, obtain a final predicted value by using a data processing method I;

[0105] in a case where only two of the three predicted ages belong to the same age group, obtain a brain age prediction value by using a data processing method II;

[0106] The data processing method I comprises:

[0107] in a case where the difference between the maximum value and the minimum value of the three predicted ages is less than or equal to an age difference threshold I, calculating a final predicted age by using a formula: A r =(A I +A II +A III ) / 3;

[0108] in a case where the difference between the maximum value and the minimum value of the three predicted ages is greater than the age difference threshold I, calculating the reliability of the three predicted ages and assigning weights to the three predicted ages according to the reliability; and calculating a final predicted age by using a formula: A r =k1*A I +k2A II +k3*A III .

[0109] The data processing method II comprises:

[0110] in a case where the difference between the maximum value and the minimum value of the three predicted ages is less than or equal to an age difference threshold II, calculating a final predicted age by using a formula: A r =(A I +A II +A III ) / 3; otherwise, re-collecting electroencephalogram data of the subject in a resting state to perform re-prediction; wherein A r is a final predicted value, A I , A II , and A III are the predicted age I, the predicted age II, and the predicted age III respectively, k1, k2, and k3 are weights of the three predicted ages respectively, and the age difference threshold II is less than the age difference threshold I.

[0111] It should be noted that, in this document, 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 can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a" does not, without more constraints, exclude the presence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0112] Those skilled in the art can clearly understand the above-mentioned embodiment method can be realized by means of software and necessary general hardware platform, of course, also can be realized by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of software product, and the computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), including a plurality of instructions to make a computer terminal (may be mobile phone, computer, server, or network equipment, etc.) execute the method described in various embodiments of the present application.

[0113] The embodiments of the present application are described above in conjunction with the drawings, but the present application is not limited to the above-mentioned specific embodiments, and the above-mentioned specific embodiments are only illustrative, not restrictive. Those skilled in the art can make many forms under the inspiration of the present application without departing from the purpose of the present application and the scope protected by the claims.

Claims

1. A brain age prediction method, characterized by, The application comprises the following steps: Collecting electroencephalogram data of a subject in a resting state; Inputting the electroencephalogram data into three pre-trained models to obtain three predicted ages, namely predicted age I, predicted age II and predicted age III; the three models are a radial basis kernel support vector regression model, an elastic net regression model and a radial basis function kernel Gaussian process model; In the case that the three predicted ages belong to the same age group, a final predicted value is obtained by using data processing method I; the age groups are divided into 13-20 years old, 21-30 years old, 31-60 years old and 61 years old and above; In the case that only two of the three predicted ages belong to the same age group, a brain age predicted value is obtained by using data processing method II; The data processing method I comprises the following steps: In the case that the difference between the maximum value and the minimum value of the three predicted ages is less than or equal to an age difference threshold I, a formula is used: A r = (A I + A II + A III ) / 3, calculate the final predicted age; In the case that the difference between the maximum and minimum of the three predicted ages is greater than the age difference threshold I, the reliability d of the three predicted ages is calculated, wherein the reliability d of the current predicted age = (s t - p ) / s t , s t is the sum of the absolute values of the differences between the three predicted ages two by two, and s p is the sum of the absolute values of the differences between the current predicted age and the other two predicted ages. According to the reliability, weights are assigned to the three predicted ages, and a formula is used: A r = k1 * A I+ k2A II + k3 * A III , calculating a final predicted age; wherein the method of assigning a weight to the predicted age using the trustworthiness comprises using the formula: k = d / d t , where k is the weight of the current predicted age, d is the trustworthiness of the current predicted age, and d t is the sum of the trustworthiness of the three predicted ages. The data processing method II comprises the following steps: In the case that the difference between the maximum value and the minimum value of the three predicted ages is less than or equal to an age difference threshold II, a formula is used: A r = (A I + A II + A III ) / 3, calculate the final predicted age; otherwise, re-collect the resting-state EEG data of the subject for re-prediction; wherein A r is the final prediction value, A I , A II , A III are the predicted age I, predicted age II, predicted age III, respectively, and k1, k2, k3 are the weights of the three predicted ages, respectively, and age difference threshold II < age difference threshold I.

2. The brain age prediction method of claim 1, wherein, When collecting electroencephalogram data of a subject in a resting state, a mobile wireless electroencephalogram headset is used for collection; the mobile wireless electroencephalogram headset uses 14 salted felt sensors and 2 reference sensors for electroencephalogram data collection, and uses a 0.2-45.0 Hz band-pass filter and a 50 Hz or 60 Hz notch filter for recording.

3. The brain age prediction method of claim 2, wherein, The positions of the 14 salted felt sensors are AF3 / 4, F3 / 4, F7 / 8, FC5 / 6, T7 / 8, P7 / 8 and O1 / 2; the positions of the 2 reference sensors are P3 / 4.

4. The brain age prediction method of claim 2, wherein, When collecting electroencephalogram data, the subject alternates between opening and closing eyes at an interval of 1 minute; and 4 minutes are taken as a group, and three groups are collected.

5. The brain age prediction method according to claim 1, characterized by, When training the model, the electroencephalogram data of children under 12 years old, patients with psychiatric and brain diseases and people with bad habits are excluded from the data set.

6. A brain age prediction system characterized by, The application comprises the following steps: An electroencephalogram data collection module is used to collect electroencephalogram data of a subject in a resting state; A prediction module is used to input the electroencephalogram data into three pre-trained models to obtain three predicted ages, namely predicted age I, predicted age II and predicted age III; the three models are a radial basis kernel support vector regression model, an elastic net regression model and a radial basis function kernel Gaussian process model; A final predicted age acquisition module is used to obtain a final predicted value by using data processing method I in the case that the three predicted ages belong to the same age group; the age groups are divided into 13-20 years old, 21-30 years old, 31-60 years old and 61 years old and above; In the case that only two of the three predicted ages belong to the same age group, a brain age predicted value is obtained by using data processing method II; The data processing method I comprises the following steps: In the case that the difference between the maximum value and the minimum value of the three predicted ages is less than or equal to an age difference threshold I, a formula is used: A r = (AI + A II + A III ) / 3, calculate the final predicted age; In the case that the difference between the maximum and minimum of the three predicted ages is greater than the age difference threshold I, the reliability d of the three predicted ages is calculated, wherein the reliability d of the current predicted age = (s t -s p ) / s t , s t is the sum of the absolute values of the differences between the three predicted ages two by two, and s p is the sum of the absolute values of the differences between the current predicted age and the other two predicted ages. According to the reliability, weights are assigned to the three predicted ages, and a formula is used: A r = k1 * A I + k2 * A II + k3 * A III , to calculate the final predicted age; wherein the method of assigning a weight to the predicted age using the credibility comprises using the formula: k = d / d t , to calculate the weight of the predicted age, wherein k is the weight of the current predicted age, d is the credibility of the current predicted age, d t is the sum of the credibility of the three predicted ages; The data processing method II comprises the following steps: In case the difference between the maximum and the minimum of the three predicted ages is less than or equal to the age difference threshold II, the formula: A r = (A I + A II + A III ) / 3, calculate the final predicted age; otherwise reacquire the resting state EEG data of the subject for re-prediction; wherein A r is the final prediction value, A I , A II , A III are the predicted age I, predicted age II, predicted age III, respectively, and k1, k2, k3 are the weights of the three predicted ages, respectively, and age difference threshold II < age difference threshold I.

7. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program, when executed by the processor, implements the brain age prediction method according to any one of claims 1-5. The program, when executed by the processor, implements the brain age prediction method according to any one of claims 1-5.

Images