Human body balance ability assessment method based on wavelet packet and transfer entropy

Abstract

The invention discloses a human body balance ability evaluation method based on a wavelet packet and transfer entropy. The method comprises the following steps: step 1, synchronously collecting six surface EMG signals and 39 channels of EEG signals in real time during heel raising of a single foot and heel raising of double feet; step 2, preprocessing the sampled EMG and EEG; step 3, decomposing the processed EMG and EEG into various frequency bands by using a wavelet packet; 4, solving a coupling value of the EEG and the EMG which are subjected to wavelet packet decomposition through transfer entropy; 5, the coupling values obtained under different balance tasks are presented on the brain topographic map, and balance capability evaluation is carried out in combination with statistical analysis, the balance tasks and the coupling values. According to the method, in combination with the directivity of information transmission and the brain muscle coupling coherence law, a gamma frequency band is selected as an analysis frequency band, and the difference of balance capabilities reflected by lifting and lowering of left and right heels and lifting and lowering of a single heel and double heels is compared.

Description

Technical Field

[0001] The invention relates to the technical field of nervous system movement control mechanism, and in particular to a method for evaluating human balance ability based on wavelet packets and transfer entropy. Background Art

[0002] Maintaining body balance is essential in daily activities and sports, helping to prevent falls and reduce the risk of injury. The calf muscles, including the medial and lateral gastrocnemius and tibialis anterior, play a key role in maintaining body balance, especially for stability in a stationary state and propulsion and control when walking. Studies have shown that regular bilateral and unilateral heel raise exercises can enhance postural control in the elderly, thereby improving balance ability. However, the process of maintaining balance is complex and involves not only the participation of relevant muscles, but also the coordination of the central nervous system and cerebral cortex.

[0003] The current assessment methods in the medical field are based on the external manifestations of human balance, but rarely consider the internal neurophysiological regulation mechanism. The most commonly used medical scale assessment method is highly subjective; although the balance ability assessment based on the center of pressure (COP) signal is based on objective test data, it is not comprehensive and does not fully reflect the internal regulation mechanism of human balance.

[0004] For this purpose, the Chinese patent discloses a method for evaluating the balance ability of the human body based on binocular images. The method obtains binocular images of a preset time length in each prescribed posture for P prescribed postures of the subject in the closed-eye state; inputs the first image and / or the second image into the preset open-closed-eye recognition model and the preset posture recognition model to determine whether the subject is in the closed-eye state and makes P prescribed postures in a prescribed order; calculates the 3D key points of the human body of the subject at each shooting moment; generates a time series according to the 3D key points at different shooting moments in each prescribed posture, thereby obtaining P time series, and then inputs them into P preset balance evaluation models respectively, obtains P balance grades, calculates the average value and rounds, and obtains the balance grade. It can be seen that this method is also based on the external performance of the human body. Although it improves the evaluation efficiency and evaluation accuracy, it is based on the external performance of the human body. When obtaining the key point features of the human body, it is easy to be interfered and affected by the outside world, or it will also have a certain impact on its own state. Therefore, there are still certain limitations on the accuracy of the evaluation results.

[0005] The literature has emphasized that the central nervous system integrates and processes inputs from the vestibular, visual, and somatosensory systems to regulate the coordination of bones and muscles to maintain a stable body posture. And a large number of studies using non-invasive scalp electroencephalogram (EEG) acquisition technology have confirmed the key role of the cerebral cortex in balance regulation, but focusing only on the activation of the cerebral cortex during balance tasks is not enough to elucidate the neural mechanisms of balance control. Therefore, a comprehensive examination of the cortical-muscle interactions related to balance will further clarify the neural mechanisms that control balance, thereby effectively improving the accuracy of assessing human balance. Summary of the invention

[0006] In view of the shortcomings of the prior art, the present invention proposes a method for evaluating human balance ability based on wavelet packets and transfer entropy, combines the directionality of information transmission and the coherence law of brain-muscle coupling, selects the gamma frequency band as the analysis frequency band, and compares the differences in balance ability reflected by lifting and lowering the left and right heels and between lifting and lowering the heels of one foot and two feet.

[0007] In order to solve the above technical problems, the technical solution of the present invention is:

[0008] A method for evaluating human balance ability based on wavelet packets and transfer entropy comprises the following steps:

[0009] Step 1, synchronously and real-timely collect 6 surface EMG signals and 39 channels of EEG signals during single-leg heel raise and double-leg heel raise;

[0010] Step 2: preprocess the sampled EMG and EEG;

[0011] Step 3, perform wavelet packet decomposition on the processed EMG and EEG into various frequency bands;

[0012] Step 4, calculate the coupling value of EEG and EMG after wavelet packet decomposition by transferring entropy;

[0013] Step 5: Present the coupling values ​​obtained under different balance tasks on the brain topography map, and evaluate the balance ability based on the balance tasks and coupling values ​​on the brain topography map.

[0014] Preferably, in step 1, EMG is sampled at a sampling frequency of 2148 Hz by an electromyography sensor, and EEG is sampled at a sampling frequency of 1000 Hz by an EEG collector.

[0015] Preferably, the EEG processing method includes: 0.1-60 Hz bandpass filtering, 50 Hz power frequency interference bandstop filtering, baseline drift removal, and independent component analysis to remove artifacts caused by electrocardiographic signals, electrooculographic signals, and head movement.

[0016] Preferably, the EMG processing method includes: 0.1-60 Hz band-pass filtering, resampling, and signal segment cutting.

[0017] Preferably, the wavelet packet decomposition method is:

[0018] Assume that the EMG sequence is X = {x(n)} and the EEG sequence is Y = {y(n)}, (n = 1, 2, ..., T, T ≥ 1), and use wavelet packets to decompose the sequences X and Y into dimensions i and j; X and Y can be expressed as:

[0019]

[0020] where i, j, k∈Z, N=1, 2, ..., 2 i , M = 1, 2, ..., 2 j . and is the wavelet packet coefficient of the i-th subspace N and the j-th subspace M, Xφ N，k (t) and Yφ M，k (t) is the wavelet function, and the highest frequencies of X and Y are defined as f L and f H , then the bandwidth range of sequence X in N subspace is The bandwidth range of sequence Y in subspace M is

[0021] Preferably, the coupling value solved by the transfer entropy includes the downstream information flow of EMG->EEG and the upstream information flow of EEG->EMG.

[0022] The solution method for the downstream information flow of EMG->EEG is:

[0023] The sequence of N subspaces To the sequence of M subspace The WPTE is defined as

[0024]

[0025] Among them, t represents the discrete time value, and u represents the predicted time value;

[0026] and

[0027] are the time delay vectors of m and n dimensions;

[0028] It can be calculated as:

[0029]

[0030] It can be calculated as:

[0031]

[0032] According to the above formula, we can get:

[0033]

[0034] Therefore, the WPTE from sequence X to sequence Y can be written as follows, where and Indicates the frequency resolution, f l1 and f l2 is the minimum and maximum frequency range of sequence X, f h1 and f h2 is the minimum and maximum value of the frequency range of sequence Y, 0≤f l1 ≤f k2 ≤f L , 0≤f h1 ≤f h2 ≤f H

[0035]

[0036] Preferably, the method for solving the EEG->EMG upstream information flow is the same as the method for solving the EMG->EEG downstream information flow.

[0037] Preferably, step 5 also includes selecting 10 repeated data from each task for statistical analysis: comparing the difference between the uplink information flow and the downlink information flow in the gamma frequency band of each task; if the uplink information flow is significantly greater than the downlink information flow, the amount of information transmitted to the brain by the muscles and sensory receptors in a state of maintaining balance is greater than the amount of information transmitted from the brain to the muscles, indicating that the brain adjusts the downlink information flow according to the feedback of the uplink information flow, thereby optimizing the balance control ability. Therefore, the large uplink information flow and the downlink information flow are used as an indicator for detecting the balance ability.

[0038] The present invention has the following characteristics and beneficial effects:

[0039] By adopting the above technical solution, both EEG and EMG are non-invasive detection methods, which can be measured without harming the human body. The coherence between EEG and EMG provides objective quantitative data, which is conducive to establishing an evaluation index for balance ability. In addition, the paradigm adopted by the present invention is also relatively easy to implement, and does not require additional tools for assistance. EEG and EMG equipment are also very portable, which can save the cost of evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0041] Figure 1 This is a schematic diagram of the experimental process in an embodiment of the present invention;

[0042] Figure 2 A schematic diagram of the working process of an embodiment of the present invention;

[0043] Figure 3 To establish brain topography;

[0044] Figure 4 A comparison diagram of the established upstream and downstream information flows;

[0045] Figure 5 A comparison chart of the coupling values ​​of the left and right feet is established. DETAILED DESCRIPTION

[0046] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0047] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings.

[0048] On the contrary, the present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention as defined by the claims. Further, in order to make the public have a better understanding of the present invention, some specific details are described in detail in the detailed description of the present invention below. Those skilled in the art can fully understand the present invention without the description of these details.

[0049] Implementation paradigm and experimental environment description:

[0050] Paradigm: The subject stood on one foot, holding the chair in front of him with his hands, facing the screen. The screen displayed instructions to tell the subject what to do. First, the subject tested the right foot, first lifting the heel in a controlled manner for 2 seconds, then lowering the heel in a controlled manner for 2 seconds. In order to prevent muscle fatigue, rest for 5 seconds between sets, and repeat ten sets in total. Then do the left foot test, the same process as the right foot.

[0051] After resting for 10 minutes, perform the test of lifting both heels at the same time. Same as single-leg test, first lift both heels at the same time in a controlled manner for 2 seconds, then lower the heels in a controlled manner for 2 seconds. To prevent muscle fatigue, rest for 5 seconds between sets. Repeat ten sets in total.

[0052] Subject composition: 10 people in the youth group, aged 22±3, 10 male subjects.

[0053] In order to reduce the impact of environmental factors on the accuracy of experimental equipment, the data collection period for this experiment was selected in the afternoon, and the experimental collection environment temperature was 25±4℃. Figure 1 Schematic diagram of the experimental process.

[0054] The present invention provides a method for evaluating human balance ability based on wavelet packets and transfer entropy. Figure 2 As shown, the following steps are included:

[0055] Step 1: synchronously and real-timely collect 6 surface EMG (electromyography) signals and 39 channels of EEG (electroencephalography) signals during single-leg heel raise and double-leg heel raise.

[0056] Specifically, surface electromyographic signals (including the tibialis anterior and medial and lateral gastrocnemius muscles on both legs) and 39-channel EEG signals during single-leg and double-leg heel raises were synchronously collected in real time using an electromyographic sensor with a sampling frequency of 2148 Hz and an EEG signal collector with a sampling frequency of 1000 Hz. The EEG electrodes were placed according to the international 10-20 system.

[0057] In this embodiment, EEG and EMG signals under this paradigm are collected. Each subject is collected ten times for 4s left foot data, 4s right foot data, and both feet data. During the test, avoid the subject's head movement, eye movement, talking, and other events that affect the data quality. If such an event occurs, the test should be terminated immediately, and a new round of testing should be restarted after a proper rest.

[0058] There are 39 EEG channels in total, namely:

[0059] AF3,AF4,FZ,F1-F6,FCZ,FC1-FC6,CZ,C1-C6,CP1-CP6,PZ,P3-P6,POZ,PO3-P O6.

[0060] Step 2: Preprocess the sampled EMG and EEG.

[0061] Specifically, the original EEG and EMG data were preprocessed. EEG signal processing included: 0.1-60Hz bandpass filtering, 50Hz power frequency interference bandstop filtering, baseline drift removal, and independent component analysis (ICA) to remove ECG signals, eye signals, and artifacts caused by head movement. Myoelectric signal processing included: 0.1-60Hz bandpass filtering, resampling, and signal segment cutting. Then relatively high-quality data was obtained for subsequent analysis.

[0062] Step 3: Decompose the processed EMG and EEG into various frequency bands using wavelet packets, namely, alpha band (8-13 Hz), beta1 band (13-19 Hz), beta2 band (20-30 Hz), gamma1 band (30-45 Hz), and gamma2 band (45-60 Hz).

[0063] Specifically, the wavelet packet decomposition method is:

[0064] Assume that the EMG sequence is X = {x(n)} and the EEG sequence is Y = {y(n)}, (n = 1, 2, ..., T, T ≥ 1), and use wavelet packets to decompose the sequences X and Y into dimensions i and j; X and Y can be expressed as:

[0065]

[0066] where i, j, k∈Z, N=1, 2, ..., 2 i , M = 1, 2, ..., 2 j . and is the wavelet packet coefficient of the i-th subspace N and the j-th subspace M, Xφ N，k (t) and Yφ M，k (t) is the wavelet function, and the highest frequencies of X and Y are defined as f L and f H , then the bandwidth range of sequence X in N subspace is The bandwidth range of sequence Y in subspace M is

[0067] Step 4: Calculate the coupling value of EEG and EMG after wavelet packet decomposition through transfer entropy.

[0068] Specifically, the method for calculating transfer entropy is as follows:

[0069] The sequence of N subspaces To the sequence of M subspace The WPTE is defined as

[0070]

[0071] Among them, t represents the discrete time value, and u represents the predicted time value;

[0072] and

[0073] are the time delay vectors of m and n dimensions;

[0074] It can be calculated as:

[0075]

[0076] It can be calculated as:

[0077]

[0078] According to the above formula, we can get:

[0079]

[0080] Therefore, the WPTE from sequence X to sequence Y can be written as follows, where and Indicates the frequency resolution, f l1 and f l2 is the minimum and maximum frequency range of sequence X, f h1 and f h2 is the minimum and maximum value of the frequency range of sequence Y, 0≤f l1 ≤f l2 ≤f L , 0≤f h1 ≤f h2 ≤f H

[0081]

[0082] It can be understood that the EEG of each channel will be coupled with the EMG of the corresponding six muscles on the leg (left tibialis anterior, left gastrocnemius medial and lateral, right tibialis anterior, right gastrocnemius medial and lateral). Therefore, in this embodiment, the coupling values ​​of two actions (heel lifting and heel lowering) under three different tasks and the coupling values ​​of different directions (upward information flow in the EEG->EMG direction and downward information flow in the EMG->EEG direction) are obtained.

[0083] Step 5: Present the coupling values ​​obtained under different balance tasks on the brain topography map, and evaluate the balance ability based on the balance tasks and coupling values ​​on the brain topography map.

[0084] Specifically, Figure 3 As shown, the wavelet packet-transfer entropy coupling values ​​of each frequency band, all EEG channels and each muscle under left-foot, right-foot and double-foot balance tasks are presented in the brain topography map.

[0085] Compare brain maps visually and select appropriate channels, muscles, and frequency bands for statistical analysis. Figure 3It can be seen that in this embodiment, the coupling value between the C3 channel and the right tibialis anterior muscle and the coupling value between the C4 channel and the left tibialis anterior muscle in the brain topography are higher than other coupling values ​​in single heel lifting and lowering and double heel lifting and lowering, and the gamma frequency band has a higher coupling value than the other four frequency bands, so it is used as a judgment criterion for subsequent analysis.

[0086] Furthermore, the 10 repeated data from each task were selected for statistical analysis. First, the difference between the upward information flow (EMG->EEG) and the downward information flow (EEG->EMG) was compared. The difference between the upward information flow and the downward information flow in the gamma frequency band in the single heel lifting task, the single heel lowering task, the double heel lifting task, and the double heel lowering task was compared, such as Figure 4 As shown. If the upward information flow is significantly greater than the downward information flow, the amount of information transmitted to the brain by the muscles and sensory receptors in the state of maintaining balance is greater than the amount of information transmitted from the brain to the muscles, indicating that the brain adjusts the downward information flow according to the feedback of the upward information flow, thereby optimizing the balance control ability. Therefore, the upward information flow that is greater than the downward information flow is used as an indicator of balance ability.

[0087] Finally, in this embodiment, the difference in coupling values ​​between the left and right feet in the uplink information flow of the gamma frequency band is compared. 10 data of coupling values ​​between the right tibialis anterior muscle and the C3 channel and the left tibialis anterior muscle and the C4 channel are selected for statistical analysis, as shown in FIG. Figure 5 If the coupling values ​​of the left foot and the right foot are significantly different, it means that there is a difference in the balance ability of the left and right legs, and the side with a higher coupling value has a stronger balance ability. If there is no significant difference in the coupling values ​​of the two legs, it means that there is no difference in the balance ability of the left and right legs.

[0088] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the described embodiments. For those skilled in the art, various changes, modifications, substitutions and variations of these embodiments including components are made without departing from the principles and spirit of the present invention, and still fall within the scope of protection of the present invention.

Claims

1. A method for evaluating human balance ability based on wavelet packets and transfer entropy, characterized in that: The steps include: Step 1, synchronously and real-timely collect 6 surface EMG signals and 39 channels of EEG signals during single-leg heel raise and double-leg heel raise; Step 2: preprocess the sampled EMG and EEG; Step 3, perform wavelet packet decomposition on the processed EMG and EEG into various frequency bands; Step 4, calculate the coupling value of EEG and EMG after wavelet packet decomposition by transferring entropy; Step 5: Present the coupling values ​​obtained under different balance tasks on the brain topography map, and evaluate the balance ability based on the balance tasks and coupling values ​​on the brain topography map.

2. The method for evaluating human balance ability based on wavelet packets and transfer entropy according to claim 1, characterized in that: In the step 1, EMG is sampled at a sampling frequency of 2148 Hz by an electromyography sensor, and EEG is sampled at a sampling frequency of 1000 Hz by an EEG collector.

3. The method for evaluating human balance ability based on wavelet packets and transfer entropy according to claim 1, characterized in that: The EEG processing method includes: 0.1-60Hz bandpass filtering, 50Hz power frequency interference bandstop filtering, baseline drift removal, and independent component analysis to remove electrocardiographic signals, electrooculographic signals, and artifacts caused by head movement.

4. The method for evaluating human balance ability based on wavelet packets and transfer entropy according to claim 1, characterized in that: The EMG processing method includes: 0.1-60 Hz band-pass filtering, resampling, and signal segment cutting.

5. The method for evaluating human balance ability based on wavelet packets and transfer entropy according to claim 1, characterized in that: The wavelet packet decomposition method is: Assume that the EMG sequence is X = {x(n)} and the EEG sequence is Y = {y(n)}, (n = 1, 2, ..., T, T ≥ 1), and use wavelet packets to decompose the sequences X and Y into dimensions i and j; X and Y can be expressed as: in and is the wavelet packet coefficient of the i-th subspace N and the j-th subspace M, Xφ N，k (t) and Yφ M，k (t) is the wavelet function, and the highest frequencies of X and Y are defined as f L and f H , then the bandwidth range of sequence X in N subspace is The bandwidth range of sequence Y in subspace M is 6. The method for evaluating human balance ability based on wavelet packets and transfer entropy according to claim 5, characterized in that: The coupling value solved by the transfer entropy includes the downstream information flow of EMG->EEG and the upstream information flow of EEG->EMG.

7. The method for evaluating human balance ability based on wavelet packets and transfer entropy according to claim 6, characterized in that: The solution method for the downstream information flow of EMG->EEG is: The sequence of N subspaces To the sequence of M subspace The WPTE is defined as Among them, t represents the discrete time value, and u represents the predicted time value; and are the time delay vectors of m and n dimensions; It can be calculated as: It can be calculated as: According to the above formula, we can get: Therefore, the WPTE from sequence X to sequence Y can be written as follows, where and Indicates the frequency resolution, f l1 and f l2 is the minimum and maximum frequency range of sequence X, f h1 and f h2 is the minimum and maximum value of the frequency range of sequence Y, 0≤f l1 ≤fl2≤f L , 0≤f h1 ≤f h2 ≤f H 8. The method for evaluating human balance ability based on wavelet packets and transfer entropy according to claim 7, characterized in that: The method for solving the EEG->EMG upstream information flow is the same as the method for solving the EMG->EEG downstream information flow.

9. The method for evaluating human balance ability based on wavelet packets and transfer entropy according to claim 8, characterized in that: The step 5 also includes selecting 10 repeated data from each task for statistical analysis: comparing the difference between the upward information flow and the downward information flow of each task through brain topography. If the upward information flow is significantly greater than the downward information flow, the amount of information transmitted to the brain by the muscles and sensory receptors in a balanced state is greater than the amount of information transmitted from the brain to the muscles, indicating that the brain adjusts the downward information flow based on the feedback of the upward information flow, so the upward information flow is used as an indicator for detecting balance ability.

10. A method for evaluating human balance ability based on wavelet packets and transfer entropy according to any one of claims 1 to 9, characterized in that: In step 5, the gamma band is selected as the analysis band.

Images