A method for calculating a sleep memory consolidation index based on electroencephalogram signals

Abstract

This invention provides a method for calculating the sleep memory consolidation index based on electroencephalogram (EEG) signals, comprising the following specific steps: Step 1: Testing the subject's pre-sleep memory test value and post-sleep memory test value, and collecting single-lead EEG data and PSG multi-lead EEG data during the subject's sleep; Step 2: Segmenting the single-lead EEG data; Step 3: Performing sleep staging on the single-lead EEG data; Step 4: Calculating the total number of spindle waves N. sp Spindle wave ratio R sp Average amplitude of spindle wave (Amp) avg Step 5: Establish the calculation formula for the sleep memory consolidation index; Step 6: Determine the characteristic coefficients in the calculation formula through multiple linear regression. This invention provides a method for calculating the sleep memory consolidation index, which can intuitively characterize the level of memory consolidation after sleep in subjects.

Description

Technical Field

[0001] This invention relates to the field of electroencephalogram (EEG) signal analysis technology, and in particular to a method for calculating the sleep memory consolidation index based on EEG signals. Background Technology

[0002] Sleep is an important physiological phenomenon for the human body and is crucial for health. High-quality sleep can make people energetic and improve the brain's memory.

[0003] Further analysis of human sleep has revealed numerous characteristics in brain signals during sleep, such as: spindle waves, slow waves, K-complexes, delta waves, and theta waves. Spindle waves, beta waves, etc. Studies have shown that an increase in the amount and duration of spindle waves during sleep can increase and improve sleep quality, and spindle waves in sleep EEG signals are associated with memory consolidation.

[0004] Sleep spindles are an event in electroencephalography (EEG) that characterizes stage II sleep. At the electrophysiological level, sleep spindles reflect an ideal mechanism for inducing long-term synaptic changes in the neocortex. Studies have shown that spindle density spectral power is closely related to human intelligence and memory consolidation.

[0005] However, in the home setting, there is still a lack of data indicators and calculation and analysis methods that can intuitively represent the level of memory consolidation after sleep in subjects. Summary of the Invention

[0006] The purpose of this invention is to provide a method for calculating the sleep memory consolidation index based on electroencephalogram (EEG) signals, which can calculate the sleep memory consolidation index of a subject in a home setting. This sleep memory consolidation index can intuitively reflect the level of memory consolidation of the subject after sleep.

[0007] The objective of this invention is achieved through the following technical solution: a method for calculating the sleep memory consolidation index based on electroencephalogram (EEG) signals, comprising the following specific steps:

[0008] Step 1: Before sleep, the subject's pre-sleep memory test score is measured using a memory test method; during sleep, single-lead EEG data and PSG multi-lead EEG data are collected from the subject, with the PSG multi-lead EEG data having built-in staging labels; after sleep, the subject's post-sleep memory test score is measured using a memory test method.

[0009] Step 2: Divide the single-lead EEG data into several EEG signal segments according to a specific interval. i ;

[0010] Step 3: Use a deep learning model to perform sleep staging on the segmented single-lead EEG data, dividing it into Wake, REM, N1, N2, and N3 stages;

[0011] Step 4: Calculate all EEG signal segments s in N1 and N2 phases. i The total number of spindle waves N in sp Spindle wave ratio R sp Average amplitude of spindle wave (Amp) avg The ratio of high to low frequencies of the spindle wave, R;

[0012] Step 5: Establish the total number N of spindle waves sp Spindle wave ratio R sp Average amplitude of spindle wave (Amp) avg The formula for calculating the sleep memory consolidation index based on the ratio of high to low frequencies of spindle waves (R):

[0013] M = a * N sp +b*R sp +c*Amp avg +d*R+e

[0014] Where M is the sleep memory consolidation index, and a, b, c, d, and e are the characteristic coefficients, respectively.

[0015] Step Six: Obtain the sleep memory consolidation test value through the pre-sleep memory test value and the post-sleep memory test value; collect multiple sets of test data, and use the sleep memory consolidation test value as the sleep memory consolidation index M. Perform multiple linear regression on the sleep memory consolidation index calculation formula to determine the characteristic coefficients a, b, c, d, and e in the sleep memory consolidation index calculation formula; calculate the sleep memory consolidation index M of the subject using the sleep memory consolidation index calculation formula.

[0016] Preferably, in step one, the obtained single-lead EEG data is preprocessed to remove interference and baseline drift.

[0017] Preferably, in step two, the interval duration is 20-60 seconds.

[0018] Preferably, in step three, the deep learning model is a CNN-LSTM model.

[0019] As a preferred option, in step four, the individual EEG signal segments s are first determined. i Whether it is in stage N1 or N2, and obtain all EEG signal segments in stages N1 and N2. i Where i = 1, 2, 3, ..., n, and n is the EEG signal segment s in phase N1 and N2. i The total number;

[0020] Calculate the total number of spindle waves N sp :

[0021]

[0022] Where, N i The number of spindle waves in the i-th segment of the EEG signal;

[0023] Calculate the spindle wave ratio R sp :

[0024] R sp =N sp / n

[0025] Calculate the average amplitude Amp of the spindle wave avg :

[0026]

[0027] Among them, Amp i This represents the amplitude value of the i-th spindle wave.

[0028] Calculate the high-frequency ratio R of the spindle wave:

[0029] R = P high / P low

[0030] Among them, P high For high-frequency spindle wave power, P low The power of a spindle wave is defined as low frequency; spindle waves with frequencies below 13.5 Hz are considered low frequency spindle waves, while those with frequencies above 13.5 Hz are considered high frequency spindle waves.

[0031] As a preferred method, the sleep memory consolidation test value is calculated as follows: the sleep memory consolidation test value is obtained by subtracting the pre-sleep memory test value from the post-sleep memory test value.

[0032] The beneficial effects of this invention are: This invention provides a method for calculating the sleep memory consolidation index. It only requires collecting the subject's EEG data using a single-channel EEG data acquisition device to finally calculate the sleep memory consolidation index. The sleep memory consolidation index M can intuitively represent the subject's memory consolidation level after sleep. Furthermore, the single-channel EEG data acquisition device has a small number of electrodes, low equipment cost, and low user operation requirements, which can fully meet the needs of home use. Attached Figure Description

[0033] Figure 1 This is a flowchart of the present invention.

[0034] Figure 2 This is the network diagram of a deep learning model.

[0035] Figure 3 This is a schematic diagram of spindle wave extraction. Detailed Implementation

[0036] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the scope of protection of the present invention.

[0037] Those skilled in the art should understand that, in the disclosure of this invention, the terms "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the above terms should not be construed as limiting this invention.

[0038] It is understood that the term "a" should be understood as "at least one" or "one or more", that is, in one embodiment, the number of an element can be one, while in another embodiment, the number of the element can be multiple, and the term "a" should not be understood as a limitation on the number.

[0039] like Figure 1 As shown, a method for calculating the sleep memory consolidation index based on electroencephalogram (EEG) signals includes the following specific steps:

[0040] Step 1: Before sleep, the subject's pre-sleep memory test value is tested using a memory test method; during sleep, single-lead EEG data and PSG multi-lead EEG data are collected from the subject according to the preset sampling frequency and sampling duration. The PSG multi-lead EEG data comes with a staging label; after sleep, the subject's post-sleep memory test value is tested using a memory test method.

[0041] In this step, the memory test method is an existing technology. The memory test method combines number memory, phrase memory and image memory to test the memory test value of the subject. The evaluation score of the memory test value is between 0 and 100.

[0042] After obtaining single-lead EEG data, it is necessary to preprocess the data to remove interference and baseline drift.

[0043] Among them, single-lead EEG data were acquired using a single-channel EEG acquisition device, while multi-lead PSG EEG data were acquired using a multi-lead EEG acquisition device.

[0044] Step 2: Divide the single-lead EEG data into several EEG signal segments according to a specific interval. i .

[0045] The separation duration can be determined according to actual needs; in this invention, the separation duration is 20-60 seconds.

[0046] Step 3: Use a deep learning model to segment the segmented single-lead EEG data into sleep stages, namely Wake, REM, N1, N2, and N3.

[0047] The deep learning model used in this step is the CNN-LSTM model. The network diagram of the deep learning model is as follows: Figure 2 As shown.

[0048] The specific method for using deep learning models for sleep staging is as follows:

[0049] S1: Simultaneously collect multiple sets of single-lead EEG data and PSG multi-lead EEG data. The PSG multi-lead EEG data comes with staging labels.

[0050] S2: Construct a deep learning model (CNN-LSTM model);

[0051] S3: Use PSG multi-lead EEG data with built-in stage labels to train and optimize deep learning models;

[0052] S4: Use a trained and optimized deep learning model to perform sleep staging on single-lead EEG data.

[0053] Step 4: Calculate all EEG signal segments s in N1 and N2 phases. i The total number of spindle waves N in sp Spindle wave ratio R sp Average amplitude of spindle wave (Amp) avg The high-frequency ratio R of the spindle wave; details are as follows:

[0054] First, determine the individual EEG signal segments s i Whether it is in stage N1 or N2: Select one of the EEG signal segments s i Determine the EEG signal segment s i Is it in stage N1 or N2? If so, then the EEG signal segment s i Selected; otherwise, skip that EEG signal segment. iAnd proceed to the next EEG brain signal segment. i The judgment is made until all EEG signal segments in N1 and N2 stages are identified. i All were filtered out, and all EEG signal segments in N1 and N2 phases were obtained. i Where i = 1, 2, 3, ..., n, and n is the EEG signal segment s in phase N1 and N2. i The total number;

[0055] Total number of spindle waves N sp For each EEG signal segment in N1 and N2 stages i The total number of spindle waves in S, and the total number of spindle waves N sp The calculation formula is as follows:

[0056]

[0057] Where, N i The number of spindle waves in the i-th segment of the EEG signal;

[0058] Spindle wave ratio R sp The total number of spindle waves and the EEG signal segments in phases N1 and N2. i The ratio of the total number of spindle waves, R sp The calculation formula is as follows:

[0059] R sp =N sp / n

[0060] Spindle wave average amplitude (Amp) avg The average amplitude of the spindle waves is Amp, which represents the average amplitude of all spindle waves in periods N1 and N2. avg The calculation formula is as follows:

[0061]

[0062] Among them, Amp i This represents the amplitude value of the i-th spindle wave.

[0063] Calculate the high-frequency ratio R of the spindle wave:

[0064] R = P high / P low

[0065] Among them, P high For high-frequency spindle wave power, P low The power of a spindle wave is defined as low frequency; spindle waves with frequencies below 13.5 Hz are considered low frequency spindle waves, while those with frequencies above 13.5 Hz are considered high frequency spindle waves.

[0066] like Figure 2 As shown, Figure 2 This is a schematic diagram of spindle wave extraction. In this diagram, the waveform curve between two adjacent dashed lines represents a spindle wave. Figure 2 As can be seen from the waveform curve, a total of three spindle waves can be extracted.

[0067] Step 5: Establish the total number N of spindle waves sp Spindle wave ratio R sp Average amplitude of spindle wave (Amp) avg The formula for calculating the sleep memory consolidation index based on the ratio of high to low frequencies of spindle waves (R) is as follows:

[0068] M = a * N sp +b*R sp +c*Amp avg +d*R+e

[0069] Where M is the sleep memory consolidation index, and a, b, c, d, and e are characteristic coefficients, with a, b, c, d, and e being fixed values. The values ​​of a, b, c, d, and e can be obtained in subsequent steps. This formula shows that the sleep memory consolidation index is determined by the total number of spindle waves N. sp Spindle wave ratio R sp Average amplitude of spindle wave (Amp) avg The four parameters that determine this are: the ratio of high to low frequency of the spindle wave, and R.

[0070] Step Six: Obtain the sleep memory consolidation test value using the pre-sleep memory test value and the post-sleep memory test value. The sleep memory consolidation test value is calculated as follows: subtract the pre-sleep memory test value from the post-sleep memory test value. Collect multiple sets of test data, and use the sleep memory consolidation test value as the sleep memory consolidation index M. Perform multiple linear regression on the sleep memory consolidation index calculation formula to determine the characteristic coefficients a, b, c, d, and e in the sleep memory consolidation index calculation formula. Calculate the subject's sleep memory consolidation index M using the sleep memory consolidation index calculation formula.

[0071] In this step, when collecting multiple sets of test data, each set of test data includes the subject's pre-sleep memory test value, post-sleep memory test value, and the total number of spindle waves N. sp Spindle wave ratio R sp Average amplitude of spindle wave (Amp) avg The ratio of high to low frequency of the spindle wave, R, can be determined by multiple linear regression of multiple sets of data to determine the specific values ​​of each characteristic coefficient a, b, c, d, and e in the formula.

[0072] After determining the characteristic coefficients a, b, c, d, and e in the sleep memory consolidation index calculation formula, the sleep memory consolidation index M of the subject can be calculated using the formula. During calculation, only single-lead EEG data of the subject needs to be collected using a single-channel EEG acquisition device; the total number of spindle waves N can be obtained from the single-lead EEG data. sp Spindle wave ratio R sp Average amplitude of spindle wave (Amp) avg The four parameters, namely the ratio of high to low frequency of spindle waves (R), are used to calculate the sleep memory consolidation index (M) using the sleep memory consolidation index calculation formula.

[0073] After obtaining the formula for calculating the sleep memory consolidation index, the sleep memory consolidation index M can be calculated by simply collecting the subject's EEG data using a single-channel EEG data acquisition device. The single-channel EEG data acquisition device has a small number of electrodes, low equipment cost, and low user operation requirements, making it fully suitable for home use.

[0074] The Sleep Memory Consolidation Index (M) ranges from 0 to 100. It can intuitively represent the level of memory consolidation after sleep. The higher the value of the Sleep Memory Consolidation Index (M), the better the memory consolidation after sleep and the less forgetting.

[0075] Table 1 below records the experimental test values ​​and formula-calculated values ​​of the sleep memory consolidation index for five test samples. As can be seen from the table, the sleep memory consolidation index calculated by the formula is very close to the experimental test values, and the calculation results are relatively accurate and reliable.

[0076]

[0077] Table 1

[0078] This invention is not limited to the preferred embodiments described above. Anyone can derive other products in various forms under the guidance of this invention. However, regardless of any changes in shape or structure, any technical solution that is the same as or similar to this application falls within the protection scope of this invention.

Claims

1. A method for calculating the sleep memory consolidation index based on electroencephalogram (EEG) signals, characterized in that, The specific steps include the following: Step 1: Before sleep, the subject's pre-sleep memory test score is measured using a memory test method; during sleep, single-lead EEG and PSG multi-lead EEG data are collected from the subject, with the PSG multi-lead EEG data having built-in staging labels; after sleep, the subject's post-sleep memory test score is measured using a memory test method. Step 2: Divide the single-lead EEG data into several EEG signal segments according to a specific interval. i ; Step 3: Use a deep learning model to perform sleep staging on the segmented single-lead EEG data, dividing it into Wake, REM, N1, N2, and N3 stages; Step 4: Calculate all EEG signal segments s in N1 and N2 phases. i The total number of spindle waves N in sp Spindle wave ratio R sp Average amplitude of spindle wave (Amp) avg The ratio of high to low frequencies of the spindle wave, R; Step 5: Establish the total number N of spindle waves sp Spindle wave ratio R sp Average amplitude of spindle wave (Amp) avg The formula for calculating the sleep memory consolidation index based on the ratio of high to low frequencies of spindle waves (R): M=a*N sp +b*R sp +c*Amp avg +d*R+e Where M is the sleep memory consolidation index, and a, b, c, d, and e are the characteristic coefficients, respectively. Step Six: Obtain the sleep memory consolidation test value through the pre-sleep memory test value and the post-sleep memory test value; collect multiple sets of test data, and use the sleep memory consolidation test value as the sleep memory consolidation index M. Perform multiple linear regression on the sleep memory consolidation index calculation formula to determine the characteristic coefficients a, b, c, d, and e in the sleep memory consolidation index calculation formula; calculate the sleep memory consolidation index M of the subject using the sleep memory consolidation index calculation formula.

2. The method for calculating the sleep memory consolidation index based on electroencephalogram (EEG) signals according to claim 1, characterized in that, In step one, the obtained single-lead EEG data is preprocessed to remove interference and baseline drift.

3. The method for calculating the sleep memory consolidation index based on electroencephalogram (EEG) signals according to claim 1, characterized in that, In step two, the interval duration is 20-60 seconds.

4. The method for calculating the sleep memory consolidation index based on electroencephalogram (EEG) signals according to claim 1, characterized in that, In step three, the deep learning model is a CNN-LSTM model.

5. The method for calculating the sleep memory consolidation index based on electroencephalogram (EEG) signals according to claim 1, characterized in that, In step four, first determine each EEG signal segment s i Whether it is in stage N1 or N2, and obtain all EEG signal segments in stages N1 and N2. i Where i = 1, 2, 3, ..., n, n is the EEG signal segment s in phase N1 and N2. i The total number; Calculate the total number of spindle waves N sp : Where, N i The number of spindle waves in the i-th segment of the EEG signal; Calculate the spindle wave ratio R sp : R sp =N sp / n Calculate the average amplitude Amp of the spindle wave avg : Among them, Amp i This represents the amplitude value of the i-th spindle wave. Calculate the high-frequency ratio R of the spindle wave: R=P high / P low Among them, P high For high-frequency spindle wave power, P low The power of a spindle wave is defined as low frequency; spindle waves with frequencies below 13.5 Hz are considered low frequency spindle waves, while those with frequencies above 13.5 Hz are considered high frequency spindle waves.

6. The method for calculating the sleep memory consolidation index based on electroencephalogram (EEG) signals according to claim 1, characterized in that, The sleep memory consolidation test value is calculated as follows: the sleep memory consolidation test value is obtained by subtracting the pre-sleep memory test value from the post-sleep memory test value.

Images