Multi-band neurofeedback method and system combined with sports training

Abstract

The present invention relates to a multi-band neural feedback method and system combined with sports training. The method includes: configuring a collection template of a near-infrared cerebral blood oxygen collection module and basic training parameters of a sports training module; calculating the mean and standard deviation of the heart rate and respiratory rate in a resting state, and calculating the mean and standard deviation of the cerebral blood oxygen concentration from the neural signal source; real-time collection of near-infrared cerebral blood oxygen signals during sports training, real-time calculation of the heart rate and respiratory rate, and a comparison and analysis module to judge the heartbeat and respiratory information in real time and transmit the judgment information to the sports training module, adjust the training parameters in real time, and at the same time, real-time analysis of the near-infrared cerebral blood oxygen signals collected by the near-infrared cerebral blood oxygen collection module, calculate the real-time cerebral blood oxygen concentration from the neural signal source, and provide feedback in the form of animation through an animation feedback module.

Description

Technical Field

[0001] The present invention belongs to the field of rehabilitation technology and is aimed at people with motor dysfunction caused by cranial nerve damage, and specifically relates to a multi-band neural feedback method and system combined with sports training. Background Art

[0002] Neurofeedback is a form of biofeedback that intervenes in behavior-related neural activity by presenting measured neural activity to participants in real time in visual, auditory or other forms, so as to promote self-regulation of neural activity based on specific behaviors or pathologies, and then restore or enhance related functions. Biofeedback training technology presents physiological information that can be perceived by the human body, and uses this as a signal marker to continuously train the human body to learn and regulate its own inherent or special neural functions, so as to achieve the goal of self-regulating and training specific neural functions. In recent years, patients have completed biofeedback training under the guidance of physicians and relearned to regulate their lost control or cognitive abilities, which has become a hot method of active rehabilitation training. Among them, neurofeedback training, as an active intervention method that directly trains brain function, obtains neurophysiological signals in real time through sensing devices, and uses advanced information processing technology to complete rapid decoding of brain information, matching with multiple stimulation methods such as vision and touch, to form a brain-computer interactive movement training pathway, which can realize online feedback of brain working status and closed-loop training of operating performance.

[0003] Patent CN102319067A relates to a neurofeedback training instrument based on EEG signals, which can present EEG rhythmic waves representing memory levels to users, guide users to autonomously adjust EEG rhythmic waves, and achieve the purpose of improving memory levels. Patent CN101912255A relates to a neurofeedback system based on real-time functional magnetic resonance signals, which detects the activation state of the brain online through functional magnetic resonance and provides real-time feedback to trainees. Through repeated training, the cognitive activity level of the brain is regulated to improve or restore the corresponding cognitive functions of the trainees. Patent CN103040446A provides a neurofeedback training system and a neurofeedback training method based on optical brain imaging, which extracts the neural activity intensity index of a specific functional system of the brain from the optical brain imaging neural activity data and presents it to the trainees as feedback information, so that the neural activity of their specific functional systems can be trained.

[0004] The above invention provides a new way for brain function and neural plasticity training. The repetitive rehabilitation training of patients actively participating in self-regulating neural activities under neural feedback intervention is more effective for the reconstruction of the patient's nervous system and the recovery of motor function. However, the above invention only focuses on the neural-related components in brain signals, ignoring the complex multi-band information in brain signals. The spontaneous near-infrared brain blood oxygen fluctuation signal in the human body has different physiological sources in different frequency bands, including information such as heartbeat activity, respiratory activity and neural activity. Especially for the motor function training of patients with limb motor dysfunction, heartbeat and respiratory information can reflect the energy consumption and intensity of training, which is crucial for adjusting the patient's training strategy. Heart rate can directly reflect the current heart condition, which is the embodiment of the function of the brain and nervous system in the parasympathetic nervous system. The heart rate itself directly reflects the body's demand for oxygen. During exercise training, when the breathing rate is too high, it means that the heart rate is too fast, which means that the exercise intensity is too high, which is easy to cause excessive heart load in the long run; and, too fast breathing rate is generally accompanied by short and shallow breathing, which will lead to insufficient oxygen intake and lack of muscle energy. Therefore, heart rate and breathing rate are key variables for measuring energy consumption or training intensity during training. Summary of the invention

[0005] In view of this, the purpose of the present invention is to provide a multi-band neural feedback method and system combined with exercise training, which can feedback heart rate and respiratory rate information during exercise training, adaptively adjust the training parameters of the exercise training module, and extract neural activity information for neural feedback training, so that patients can actively participate in rehabilitation training more safely and more personalized, and achieve better rehabilitation effects.

[0006] In one aspect, the present invention provides a multi-band neurofeedback method combined with sports training, the method comprising the following steps:

[0007] Configure the acquisition template of the near-infrared cerebral blood oxygen acquisition module and the basic training parameters of the sports training module;

[0008] Keep the patient in a resting state for more than 5 minutes. The heart rate and respiratory rate extraction module will calculate the mean and standard deviation of the heart rate and respiratory rate in the resting state. The neural signal processing module will calculate the mean and standard deviation of the brain blood oxygen concentration from the neural signal.

[0009] The near-infrared cerebral blood oxygen acquisition module collects near-infrared cerebral blood oxygen signals during exercise training in real time, the heart rate and respiratory frequency extraction module calculates the heart rate and respiratory frequency in real time, and the heart rate and respiratory frequency comparison and analysis module judges the heartbeat and respiratory information in real time and transmits the judgment information to the exercise training module, and the exercise training module adjusts the training parameters in real time. At the same time, the neural signal processing module analyzes the near-infrared cerebral blood oxygen signals collected by the near-infrared cerebral blood oxygen acquisition module in real time, calculates the real-time cerebral blood oxygen concentration of the neural signal source, and provides feedback in the form of animation through the animation feedback module.

[0010] Among them, the near-infrared cerebral blood oxygen acquisition module can at least collect infrared cerebral blood oxygen signals from the prefrontal lobe and motor cortex; the basic training parameters of the motion training module include but are not limited to motion speed, motion damping and motion range.

[0011] In some embodiments, the heart rate at rest or during exercise training is calculated in real time by:

[0012] The near-infrared acquisition channel signal of the frontal lobe is used as the basic brain blood oxygen signal for extracting heart rate and respiratory rate. The heart rate and respiratory rate extraction module performs principal component analysis on all basic brain blood oxygen signals, and uses the largest principal component corresponding to the largest characteristic root as the time series for further analyzing heart rate and respiratory rate.

[0013] After obtaining the maximum principal component time series of all channels of the prefrontal lobe, a fast Fourier transform of a sliding window with a time window of 25.6 seconds is used to convert the maximum principal component from the time domain into a frequency domain signal, and the frequency corresponding to the maximum value in the range of 0.6-2Hz of each sliding window frequency domain signal is found, that is, the heart rate time series of the trainee at rest or during exercise training is obtained.

[0014] Among them, the maximum principal component time series is obtained by the following steps:

[0015] The cerebral blood oxygen signal of all channels of the prefrontal lobe is standardized to obtain a standardized matrix (X1, X2, ..., Xn), where n is the total number of prefrontal channels. The standardization method is to subtract the mean of each channel and divide it by the standard deviation of each channel; the covariance matrix of the standardized cerebral blood oxygen signal matrix is ​​calculated, and the eigenvalues ​​and corresponding eigenvectors of the covariance matrix are calculated; the eigenvector a (column vector) corresponding to the maximum eigenvalue is the coefficient corresponding to the maximum principal component, and finally the maximum principal component is PCmax = a (X1, X2, ..., Xn).

[0016] In some embodiments, the breathing rate at rest or the breathing rate of the user during exercise training is calculated in real time by:

[0017] Calculate the upper envelope parameter sequence, lower envelope parameter sequence, interval parameter sequence, amplitude parameter sequence and area parameter sequence of the maximum principal component time series;

[0018] For the upper envelope parameter sequence, lower envelope parameter sequence, interval parameter sequence, amplitude parameter sequence and area parameter sequence of the maximum principal component time series, a 25.6-second time window is used to perform a sliding window fast Fourier transform, and the frequency corresponding to the maximum value in the range of 0.145-0.6Hz of the frequency domain signal of each sliding window is found, and the respiratory frequency time series calculated by the upper envelope parameter sequence, lower envelope parameter sequence, interval parameter sequence, amplitude parameter sequence and area parameter sequence can be obtained respectively; the five respiratory frequency time series are averaged to obtain the respiratory frequency time series of the trainee at rest or during exercise training.

[0019] In some embodiments, the peak point and the valley point of each fluctuation period of the maximum principal component time series are identified, and the upper envelope parameter sequence and the lower envelope parameter sequence are obtained by performing cubic spline interpolation on the peak point and the valley point time series;

[0020] Calculate the time interval between every two peak points, and obtain the interval parameter sequence from the time interval time series through cubic spline interpolation;

[0021] Calculate the concentration difference between the peak and valley values ​​of each fluctuation cycle, and obtain the amplitude parameter sequence by cubic spline interpolation of the concentration difference time series; and

[0022] The signal enclosing area of ​​each fluctuation period is calculated, and the area parameter sequence is obtained by cubic spline interpolation of the enclosing area time series.

[0023] In some embodiments, the brain blood oxygen concentration of the neural signal source is calculated by:

[0024] For single-channel cerebral blood oxygen signal, the cerebral blood oxygen signal from the noise channel is first excluded; then the time derivative distribution repair algorithm is used to remove motion artifact interference; signal spikes are removed; low-frequency drift and high-frequency instrument noise are removed; the hemodynamic separation algorithm is used to purify the neural source components in the cerebral blood oxygen signal; after purification by the hemodynamic separation algorithm, the cerebral blood oxygen concentration of the neural signal source is filtered using a 0.01-0.08Hz second-order bandpass Butterworth filter.

[0025] In some embodiments, the heart rate and respiratory frequency comparison and analysis module records the mean and standard deviation of the heart rate and respiratory frequency in the resting state, and compares the heart rate and respiratory frequency calculated in real time with the mean and standard deviation of the heart rate and respiratory frequency in the resting state;

[0026] If the real-time heart rate and respiratory rate are within the range of the mean of the resting heart rate and respiratory rate plus one standard deviation, gradually increase the training parameters of the exercise training module, including but not limited to increasing the exercise speed, increasing the exercise damping and expanding the exercise range, so as to provide effective training intensity;

[0027] If the real-time heart rate and respiratory rate are within the range of one to two standard deviations of the mean of the resting heart rate and respiratory rate, the training parameters of the exercise training module are not changed;

[0028] If the real-time heart rate and respiratory rate are within the range of the mean of the resting heart rate and respiratory rate plus two to three times the standard deviation, gradually reduce the training parameters of the exercise training module, including but not limited to reducing the exercise speed, reducing the exercise damping and narrowing the exercise range, so as to provide appropriate training intensity;

[0029] If the real-time heart rate and respiratory rate exceed the mean of the resting heart rate and respiratory rate plus three times the standard deviation, the training parameters are gradually reduced until the exercise training module stops.

[0030] On the other hand, the present invention also provides a multi-band neural feedback system combined with sports training, the system comprising:

[0031] A near-infrared cerebral blood oxygen collection module, which is used to collect near-infrared cerebral blood oxygen signals including at least the prefrontal lobe and motor area;

[0032] A heart rate and respiratory frequency extraction module, which is used to extract the heart rate and respiratory frequency at rest, as well as the real-time heart rate and respiratory frequency during exercise training, based on the maximum principal component of the near-infrared cerebral blood oxygen signal of the frontal lobe;

[0033] A heart rate and respiratory frequency comparison and analysis module, which is used to compare the real-time heart rate and respiratory frequency with the resting heart rate and respiratory frequency during exercise training to generate control instructions for training parameters;

[0034] A sports training module, which is used to provide sports training support for the user and adjust the training parameters according to the control instructions of the training parameters generated by the heart rate and respiratory frequency comparison and analysis module;

[0035] A neural signal processing module, which is used to extract the brain blood oxygen concentration of the neural signal source in the resting state and the brain blood oxygen concentration of the neural signal source during exercise training based on the near-infrared brain blood oxygen signal of the prefrontal lobe and the motor area;

[0036] The animation feedback module is used to feed back the brain blood oxygen concentration of the neural signal to the user in the form of animation based on the brain blood oxygen concentration of the neural signal source in the resting state and the brain blood oxygen concentration of the neural signal source during exercise training.

[0037] In some embodiments, the heart rate and respiratory frequency extraction module calculates the heart rate at rest or calculates the heart rate during exercise training in real time by:

[0038] The near-infrared acquisition channel signal of the frontal lobe is used as the basic brain blood oxygen signal for extracting heart rate and respiratory rate. The heart rate and respiratory rate extraction module performs principal component analysis on all basic brain blood oxygen signals, and uses the largest principal component corresponding to the largest characteristic root as the time series for further analyzing heart rate and respiratory rate.

[0039] After obtaining the maximum principal component time series of all channels of the prefrontal lobe, a fast Fourier transform of a sliding window with a time window of 25.6 seconds is used to convert the maximum principal component from the time domain into a frequency domain signal, and the frequency corresponding to the maximum value in the range of 0.6-2Hz of each sliding window frequency domain signal is found, that is, the heart rate time series of the trainee at rest or during exercise training is obtained.

[0040] Among them, the maximum principal component time series is obtained by the following steps:

[0041] The cerebral blood oxygen signal of all channels of the prefrontal lobe is standardized to obtain a standardized matrix (X1, X2, ..., Xn), where n is the total number of prefrontal channels. The standardization method is to subtract the mean of each channel and divide it by the standard deviation of each channel; the covariance matrix of the standardized cerebral blood oxygen signal matrix is ​​calculated, and the eigenvalues ​​and corresponding eigenvectors of the covariance matrix are calculated; the eigenvector a (column vector) corresponding to the maximum eigenvalue is the coefficient corresponding to the maximum principal component, and finally the maximum principal component is PCmax = a (X1, X2, ..., Xn).

[0042] In some embodiments, the heart rate respiratory frequency extraction module calculates the respiratory frequency in a resting state or calculates the respiratory frequency during exercise training in real time by:

[0043] Calculate the upper envelope parameter sequence, lower envelope parameter sequence, interval parameter sequence, amplitude parameter sequence and area parameter sequence of the maximum principal component time series;

[0044] For the upper envelope parameter sequence, lower envelope parameter sequence, interval parameter sequence, amplitude parameter sequence and area parameter sequence of the maximum principal component time series, a 25.6-second time window is used to perform a sliding window fast Fourier transform, and the frequency corresponding to the maximum value in the range of 0.145-0.6Hz of the frequency domain signal of each sliding window is found, and the respiratory frequency time series calculated by the upper envelope parameter sequence, lower envelope parameter sequence, interval parameter sequence, amplitude parameter sequence and area parameter sequence can be obtained respectively; the five respiratory frequency time series are averaged to obtain the respiratory frequency time series of the trainee at rest or during exercise training.

[0045] The neural signal processing module calculates the brain blood oxygen concentration of the neural signal source in the following way:

[0046] For single-channel cerebral blood oxygen signal, the cerebral blood oxygen signal from the noise channel is first excluded; then the time derivative distribution repair algorithm is used to remove motion artifact interference; signal spikes are removed; low-frequency drift and high-frequency instrument noise are removed; the hemodynamic separation algorithm is used to purify the neural source components in the cerebral blood oxygen signal; after purification by the hemodynamic separation algorithm, the cerebral blood oxygen concentration of the neural signal source is filtered using a 0.01-0.08Hz second-order bandpass Butterworth filter.

[0047] The beneficial effects of the present invention are at least: a multi-band neural feedback method and system combined with sports training can extract and feedback heart rate and respiratory rate information during sports training through the collection of near-infrared spectral signals of the cerebral cortex, characterize the intensity and energy consumption of the user's sports training, and adaptively adjust the training parameters of the sports training module; it can extract neural activity information, reflect the intensity of the user's brain nerve activity induced by sports training, and perform neural feedback training in the form of animation; it enables users to actively participate in rehabilitation training in a safer and more personalized manner, and achieve better rehabilitation effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The drawings described herein are used to provide a further understanding of the present invention, constitute a part of this application, and do not constitute a limitation of the present invention. The components in the drawings are not drawn to scale, but are only for illustrating the principles of the present invention. In order to facilitate the illustration and description of some parts of the present invention, the corresponding parts in the drawings may be enlarged, that is, they may become larger relative to other components in the exemplary device actually manufactured according to the present invention. In the drawings:

[0049] Figure 1 A schematic diagram of a multi-band neurofeedback system combined with sports training;

[0050] Figure 2 The diagram shows the probe arrangement in the near-infrared cerebral blood oxygen collection module and the schematic diagram of the maximum principal component of the frontal lobe;

[0051] Figure 3 This is a schematic diagram of the Fourier transform spectrum of the largest principal component of the prefrontal cerebral blood oxygen signal;

[0052] Figure 4 Schematic diagram of the five respiratory regulation parameters of the largest principal component of the prefrontal cerebral blood oxygen signal;

[0053] Figure 5 Schematic diagram of cubic spline interpolation time series of five respiratory regulation parameters;

[0054] Figure 6 This is a schematic diagram of the animation interface of the animation feedback module. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments and the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0056] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the accompanying drawings, the same reference numerals represent the same or similar components, or the same or similar steps.

[0057] like Figure 1 The figure shows a schematic diagram of a multi-band neural feedback system combined with sports training according to an embodiment of the present invention. The system includes a near-infrared cerebral blood oxygen acquisition module, a heart rate and respiratory frequency extraction module, a heart rate and respiratory frequency comparison and analysis module, a sports training module, a neural signal processing module and an animation feedback module.

[0058] The near-infrared cerebral blood oxygen acquisition module is used to acquire near-infrared cerebral blood oxygen signals including at least the prefrontal lobe and motor area.

[0059] The heart rate and respiratory rate extraction module is used to extract the resting heart rate and respiratory rate, as well as the real-time heart rate and respiratory rate during exercise training, based on the maximum principal component of the near-infrared cerebral blood oxygen signal of the frontal lobe.

[0060] The heart rate and respiratory frequency comparison and analysis module is used to compare the real-time heart rate and respiratory frequency with the resting heart rate and respiratory frequency during exercise training, and generate control instructions for training parameters.

[0061] The exercise training module is used to provide exercise training support for the user and adjust the training parameters according to the control instructions of the training parameters generated by the heart rate and respiratory frequency comparison and analysis module.

[0062] The neural signal processing module is used to extract the brain blood oxygen concentration of the neural signal source in the resting state and the brain blood oxygen concentration of the neural signal source during exercise training based on the near-infrared brain blood oxygen signal of the prefrontal lobe and motor area.

[0063] The animation feedback module is used to feed back the brain blood oxygen concentration of neural signals to the user in the form of animation based on the brain blood oxygen concentration of neural signals in the resting state and the brain blood oxygen concentration of neural signals during exercise training.

[0064] The following describes the multi-band neural feedback method combined with sports training of the present invention in combination with a multi-band neural feedback system. The method includes the following steps:

[0065] S1: Establishing a connection between the user and the near-infrared cerebral blood oxygen collection module and the sports training module, and configuring the collection template of the near-infrared cerebral blood oxygen collection module and the basic training parameters of the sports training module;

[0066] S2: Maintain a resting state for more than 5 minutes, and the heart rate and respiratory rate extraction module calculates the mean and standard deviation of the heart rate and respiratory rate in the resting state, and the neural signal processing module calculates the mean and standard deviation of the brain blood oxygen concentration from the neural signal;

[0067] S3: The user performs exercise training through the exercise training module, the near-infrared cerebral blood oxygen acquisition module collects near-infrared cerebral blood oxygen signals in real time, the heart rate and respiratory frequency extraction module calculates the heart rate and respiratory frequency in real time, the heart rate and respiratory frequency comparison and analysis module judges the heartbeat and respiratory conditions in real time, and transmits the judgment information to the exercise training module, which adjusts the training parameters in real time;

[0068] S4: While performing step S3, the neural signal processing module analyzes the near-infrared cerebral blood oxygen signal collected by the near-infrared cerebral blood oxygen collection module in real time, calculates the real-time cerebral blood oxygen concentration of the neural signal source, and interacts with the user in the form of animation through the animation feedback module.

[0069] Specifically, Figure 2 The diagram shows the probe arrangement position diagram in the near-infrared cerebral blood oxygen acquisition module. In step S1, the acquisition template of the near-infrared cerebral blood oxygen acquisition module should at least cover the prefrontal lobe and motor cortex; the basic training parameters of the movement training module include but are not limited to movement speed, movement damping and movement range.

[0070] In step S2, the heart rate in the resting state is calculated by:

[0071] For a single-channel cerebral blood oxygen signal, if the mean / standard deviation of the corresponding raw light intensity time series is <2, it is considered a noise channel and will be excluded from subsequent calculation analysis;

[0072] The time derivative distribution repair algorithm was used to remove motion artifacts. The Hample filter algorithm was used to remove signal spikes, with filter parameters k = 50 and nsigma = 4. An 8th-order bandpass Butterworth filter with a range of 0.01-2 Hz was used to remove low-frequency drift and high-frequency instrument noise.

[0073] The light intensity signal is converted into a light density signal, and the modified Beer-Lambert law is used to convert the light density signal into a brain blood oxygen signal; because the near-infrared signal of the frontal lobe is not blocked by hair, it is more stable in contact with the skin than the motor area channel, and is less susceptible to motion artifacts and clutter. Therefore, the near-infrared acquisition channel signal of the frontal lobe is used as the basic brain blood oxygen signal for extracting heart rate and respiratory rate. In order to reduce the computing load, the heart rate and respiratory rate extraction module performs principal component analysis on the brain blood oxygen signals of all frontal lobe channels, and uses the largest principal component corresponding to the largest characteristic root as the time series for further analysis of heart rate and respiratory rate.

[0074] like Figure 2 The maximum principal component of the prefrontal cerebral blood oxygen signal is shown in the lower middle part. The maximum principal component time series is obtained by the following steps: the cerebral blood oxygen signals of all channels of the prefrontal lobe are standardized to obtain a standardized matrix (X1, X2, ..., Xn), where n is the total number of prefrontal channels, and the standardization method is to subtract the mean of each channel and divide it by the standard deviation of each channel; calculate the covariance matrix of the standardized cerebral blood oxygen signal matrix, calculate the eigenvalues ​​and corresponding eigenvectors of the covariance matrix; the eigenvector a (column vector) corresponding to the maximum eigenvalue is the coefficient corresponding to the maximum principal component, and finally the maximum principal component is PCmax = a (X1, X2, ..., Xn).

[0075] like Figure 3 As shown in the figure, after obtaining the maximum principal component time series of all channels of the prefrontal lobe, a 25.6-second time window is used to perform a fast Fourier transform of the sliding window to convert it from the time domain to the frequency domain signal, and the frequency corresponding to the maximum value in the range of 0.6-2Hz of each sliding window frequency domain signal is found to obtain the trainee's resting heart rate time series.

[0076] like Figure 4 and Figure 5 As shown, in step S2, the respiratory frequency in the resting state is calculated by:

[0077] Calculate the upper envelope parameter sequence, lower envelope parameter sequence, interval parameter sequence, amplitude parameter sequence and area parameter sequence of the maximum principal component time series;

[0078] The respiratory signal will be modulated to the blood oxygen signal through amplitude modulation, that is, the amplitude of the blood oxygen signal is changed, and the envelope of the blood oxygen signal is shown on the blood oxygen signal, so that the relevant information of the respiratory signal is reflected through the blood oxygen signal. The upper envelope parameter sequence and the lower envelope parameter sequence need to identify the peak point and valley point of each fluctuation cycle of the maximum principal component time series, and the peak point and valley point time series are obtained by cubic spline interpolation;

[0079] When breathing, the heart rate will change slightly due to the action of the vagus nerve. When inhaling, the vagus nerve is inhibited and the heart rate increases. When exhaling, the vagus nerve inhibition is cancelled and the heart rate decreases. In this way, the respiratory signal transmits respiratory related information by changing the frequency of the pulse wave signal. By detecting the interval between adjacent peaks, accurate heart rate information can be obtained, and the respiratory frequency can be calculated from the changes in heart rate information. The interval parameter sequence needs to calculate the time interval between every two peak points, which is obtained by cubic spline interpolation of the time interval time series;

[0080] Breathing causes changes in blood vessels, which not only causes changes in the blood oxygen signal envelope, but also causes the amplitude of the collected blood oxygen signal to be affected by breathing. The amplitude is defined as the difference between the maximum and minimum values, and the frequency information related to breathing is obtained through the change of the amplitude. The amplitude parameter sequence needs to calculate the concentration difference between the peak and valley values ​​of each fluctuation cycle, which is obtained by cubic spline interpolation of the concentration difference time series;

[0081] The area enclosed by the blood oxygen signal is also closely related to the respiratory signal, and the respiratory condition will also be reflected in the change of the blood oxygen signal area. The area parameter sequence needs to calculate the signal enclosed area of ​​each fluctuation cycle, which is obtained by cubic spline interpolation of the enclosed area time series;

[0082] For the upper envelope parameter sequence, lower envelope parameter sequence, interval parameter sequence, amplitude parameter sequence and area parameter sequence of the maximum principal component time series, a 25.6-second time window is used to perform a sliding window fast Fourier transform, and the frequency corresponding to the maximum value in the range of 0.145-0.6Hz of the frequency domain signal of each sliding window is found, and the respiratory frequency time series calculated by the upper envelope parameter sequence, lower envelope parameter sequence, interval parameter sequence, amplitude parameter sequence and area parameter sequence can be obtained respectively; the five respiratory frequency time series are averaged to obtain the trainee's resting respiratory frequency time series.

[0083] In step S2, the brain blood oxygen concentration of the neural signal source is calculated by the following:

[0084] For a single-channel cerebral blood oxygen signal, if the corresponding raw light intensity is higher than 1000, or the raw light intensity is <0.5, or the mean / standard deviation of the raw light intensity time series is <2, it is considered to be a noise channel and will be excluded from subsequent calculation analysis;

[0085] The time derivative distribution repair algorithm was used to remove motion artifacts. The Hample filter algorithm was used to remove signal spikes, with filter parameters k = 50 and nsigma = 4. An 8th-order bandpass Butterworth filter with a range of 0.01-2 Hz was used to remove low-frequency drift and high-frequency instrument noise.

[0086] The HMS algorithm is used to purify the neural components in the brain blood oxygen signal. Under the HMS analysis framework, it is assumed that there is a positive correlation between oxyhemoglobin and deoxyhemoglobin in the physiological interference components, and a negative correlation between the changes in oxyhemoglobin and deoxyhemoglobin related to neural activity. The functional components and system noise components are separated by determining the coefficients of the corresponding correlations;

[0087] After being purified by the HMS algorithm, the cerebral blood oxygen concentration of the neural signal source was obtained by filtering with a 0.01-0.08 Hz second-order bandpass Butterworth filter.

[0088] In step S3, the method for real-time calculation of heart rate and respiratory rate by the heart rate respiratory rate extraction module is the same as that in step (2); preferably, a 25.6 second time window is also used for fast Fourier transform in the real-time calculation process.

[0089] In step S3, the heart rate and respiratory frequency comparison and analysis module records the mean and standard deviation of the heart rate and respiratory frequency in the resting state, and compares the heart rate and respiratory frequency calculated in real time with the mean and standard deviation of the heart rate and respiratory frequency in the resting state;

[0090] If the real-time heart rate and respiratory rate are within the range of the mean of the resting heart rate and respiratory rate plus 1 standard deviation, gradually increase the training parameters of the exercise training module, including but not limited to increasing the exercise speed, increasing the exercise damping and expanding the exercise range, so as to provide effective training intensity;

[0091] If the real-time heart rate and respiratory rate are within the range of 1 to 2 standard deviations of the mean of the resting heart rate and respiratory rate, the training parameters of the exercise training module will not be changed;

[0092] If the real-time heart rate and respiratory rate are within the range of the mean of the resting heart rate and respiratory rate plus 2 to 3 times the standard deviation, gradually reduce the training parameters of the exercise training module, including but not limited to reducing the exercise speed, reducing the exercise damping and narrowing the exercise range, so as to provide appropriate training intensity;

[0093] If the real-time heart rate and respiratory rate exceed the mean of the resting heart rate and respiratory rate plus 3 times the standard deviation, the training parameters will be gradually reduced within 5 seconds until the exercise training module stops.

[0094] In step S4, the method in which the neural signal processing module calculates the brain blood oxygen concentration of the neural signal source in real time is the same as that in step S2; preferably, a 5-second time window is also used to calculate the brain blood oxygen concentration of the sliding window during the real-time calculation process.

[0095] like Figure 5As shown, in this embodiment, the movement of a small fish in different directions reflects the intensity of neural activity in the prefrontal lobe and motor area in the animation feedback module. The up and down swimming of the small fish uses the mean of the brain blood oxygen concentration of the resting neural signal source of the prefrontal lobe as the baseline value, and the brain blood oxygen concentration of the real-time neural signal source of the prefrontal lobe rises or falls according to the movement training; the left and right swimming of the small fish uses the mean of the brain blood oxygen concentration of the resting neural signal source of the motor area as the baseline value, and the brain blood oxygen concentration of the real-time neural signal source of the motor area moves forward or backward according to the movement training. Among them, when the brain blood oxygen concentration of the real-time neural signal source of the motor area exceeds the mean of the brain blood oxygen concentration of the resting neural signal source plus 0.2 times the standard deviation, the score is 20 points, and when it exceeds the mean plus 0.4 times the standard deviation, the score is 40 points, and so on. The cumulative score is directly displayed in the upper right corner of the animation interface of the animation feedback module.

[0096] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A multi-band neurofeedback method combined with sports training, characterized in that: The method comprises the following steps: Configure the acquisition template of the near-infrared cerebral blood oxygen acquisition module and the basic training parameters of the sports training module; Keep the patient in a resting state for more than 5 minutes. The heart rate and respiratory rate extraction module will calculate the mean and standard deviation of the heart rate and respiratory rate in the resting state. The neural signal processing module will calculate the mean and standard deviation of the brain blood oxygen concentration from the neural signal. The near-infrared cerebral blood oxygen acquisition module collects near-infrared cerebral blood oxygen signals during exercise training in real time, the heart rate and respiratory rate extraction module calculates the heart rate and respiratory rate in real time based on the near-infrared cerebral blood oxygen signals, and the heart rate and respiratory rate comparison and analysis module judges the heartbeat and respiratory information in real time and transmits the judgment information to the exercise training module, and the exercise training module adjusts the training parameters in real time. At the same time, the neural signal processing module analyzes the near-infrared cerebral blood oxygen signals collected by the near-infrared cerebral blood oxygen acquisition module in real time, calculates the real-time cerebral blood oxygen concentration of the neural signal source, and provides feedback in the form of animation through the animation feedback module. Among them, the near-infrared cerebral blood oxygen acquisition module can at least collect infrared cerebral blood oxygen signals from the prefrontal lobe and motor cortex; the basic training parameters of the sports training module include movement speed, movement damping and movement range. It is characterized in that the heart rate in a resting state or the heart rate during exercise training is calculated in real time by the following methods: The near-infrared acquisition channel signal of the frontal lobe is used as the basic brain blood oxygen signal for extracting heart rate and respiratory rate. The heart rate and respiratory rate extraction module performs principal component analysis on all basic brain blood oxygen signals, and uses the largest principal component corresponding to the largest characteristic root as the time series for further analyzing heart rate and respiratory rate. After obtaining the maximum principal component time series of all channels of the prefrontal lobe, a fast Fourier transform of a sliding window with a time window of 25.6 seconds is used to convert the maximum principal component from the time domain into a frequency domain signal, and the frequency corresponding to the maximum value in the range of 0.6-2Hz of each sliding window frequency domain signal is found, that is, the heart rate time series of the trainee at rest or during exercise training is obtained. Among them, the maximum principal component time series is obtained by the following steps: Normalize the cerebral blood oxygen signals of all channels in the frontal lobe to obtain the standardized matrix , n is the total number of prefrontal channels, and the standardization method is to subtract the mean of each channel and divide it by the standard deviation of each channel; calculate the covariance matrix of the standardized brain blood oxygen signal matrix, calculate the eigenvalues ​​and corresponding eigenvectors of the covariance matrix; the eigenvector a corresponding to the maximum eigenvalue is the coefficient corresponding to the maximum principal component, and finally the maximum principal component is PCmax=a , where the eigenvector a is a column vector.

2. The multi-band neurofeedback method combined with sports training according to claim 1, characterized in that: Calculate your breathing rate at rest or in real time during exercise training using the following methods: Calculate the upper envelope parameter sequence, lower envelope parameter sequence, interval parameter sequence, amplitude parameter sequence and area parameter sequence of the maximum principal component time series; For the upper envelope parameter sequence, lower envelope parameter sequence, interval parameter sequence, amplitude parameter sequence and area parameter sequence of the maximum principal component time series, a 25.6-second time window is used to perform a sliding window fast Fourier transform, and the frequency corresponding to the maximum value in the range of 0.145-0.6Hz of the frequency domain signal of each sliding window is found, and the respiratory frequency time series calculated by the upper envelope parameter sequence, lower envelope parameter sequence, interval parameter sequence, amplitude parameter sequence and area parameter sequence can be obtained respectively; the five respiratory frequency time series are averaged to obtain the respiratory frequency time series of the trainee at rest or during exercise training.

3. The multi-band neurofeedback method combined with sports training according to claim 2, characterized in that: Identify the peak and valley points of each fluctuation cycle of the maximum principal component time series, and obtain the upper envelope parameter sequence and the lower envelope parameter sequence by interpolating the peak and valley point time series through cubic spline; Calculate the time interval between every two peak points, and obtain the interval parameter sequence from the time interval time series through cubic spline interpolation; The concentration difference between the peak and valley values ​​of each fluctuation cycle is calculated, and the amplitude parameter sequence is obtained by cubic spline interpolation of the concentration difference time series; and The signal enclosing area of ​​each fluctuation period is calculated, and the area parameter sequence is obtained by cubic spline interpolation of the enclosing area time series.

4. The multi-band neurofeedback method combined with sports training according to claim 1, characterized in that: The brain blood oxygen concentration of the neural signal source is calculated by: For single-channel cerebral blood oxygen signal, the cerebral blood oxygen signal from the noise channel is first excluded; then the time derivative distribution repair algorithm is used to remove motion artifact interference; signal spikes are removed; low-frequency drift and high-frequency instrument noise are removed; the hemodynamic separation algorithm is used to purify the neural source components in the cerebral blood oxygen signal; after purification by the hemodynamic separation algorithm, the cerebral blood oxygen concentration of the neural signal source is filtered using a 0.01-0.08Hz second-order bandpass Butterworth filter.

5. The multi-band neurofeedback method combined with sports training according to claim 1, characterized in that: The heart rate and respiratory frequency comparison and analysis module records the mean and standard deviation of the heart rate and respiratory frequency in the resting state, and compares the heart rate and respiratory frequency calculated in real time with the mean and standard deviation of the heart rate and respiratory frequency in the resting state; If the real-time heart rate and respiratory rate are within the range of the mean of the resting heart rate and respiratory rate plus one standard deviation, the training parameters of the exercise training module are gradually increased to provide effective training intensity; If the real-time heart rate and respiratory rate are within the range of one to two standard deviations of the mean of the resting heart rate and respiratory rate, the training parameters of the exercise training module are not changed; If the real-time heart rate and respiratory rate are within the range of the mean of the resting heart rate and respiratory rate plus two to three standard deviations, the training parameters of the exercise training module are gradually reduced to provide an appropriate training intensity; If the real-time heart rate and respiratory rate exceed the mean of the resting heart rate and respiratory rate plus three times the standard deviation, the training parameters are gradually reduced until the exercise training module stops.

6. A multi-band neurofeedback system combined with sports training, characterized in that: The system includes: A near-infrared cerebral blood oxygen collection module, which is used to collect near-infrared cerebral blood oxygen signals including at least the prefrontal lobe and motor area; A heart rate and respiratory frequency extraction module, which is used to extract the heart rate and respiratory frequency at rest, as well as the real-time heart rate and respiratory frequency during exercise training, based on the maximum principal component of the near-infrared cerebral blood oxygen signal of the frontal lobe; A heart rate and respiratory frequency comparison and analysis module, which is used to compare the real-time heart rate and respiratory frequency with the resting heart rate and respiratory frequency during exercise training to generate control instructions for training parameters; A sports training module, which is used to provide sports training support for the user and adjust the training parameters according to the control instructions of the training parameters generated by the heart rate and respiratory frequency comparison and analysis module; A neural signal processing module, which is used to extract the brain blood oxygen concentration of the neural signal source in the resting state and the brain blood oxygen concentration of the neural signal source during exercise training based on the near-infrared brain blood oxygen signal of the prefrontal lobe and the motor area; The animation feedback module is used to feed back the brain blood oxygen concentration of the neural signal to the user in the form of animation according to the brain blood oxygen concentration of the neural signal source in the resting state and the brain blood oxygen concentration of the neural signal source during exercise training. The invention is characterized in that the heart rate and respiratory frequency extraction module calculates the heart rate in a resting state or calculates the heart rate during exercise training in real time by the following methods: The near-infrared acquisition channel signal of the frontal lobe is used as the basic brain blood oxygen signal for extracting heart rate and respiratory rate. The heart rate and respiratory rate extraction module performs principal component analysis on all basic brain blood oxygen signals, and uses the largest principal component corresponding to the largest characteristic root as the time series for further analyzing heart rate and respiratory rate. After obtaining the maximum principal component time series of all channels of the prefrontal lobe, a fast Fourier transform of a sliding window with a time window of 25.6 seconds is used to convert the maximum principal component from the time domain into a frequency domain signal, and the frequency corresponding to the maximum value in the range of 0.6-2Hz of each sliding window frequency domain signal is found, that is, the heart rate time series of the trainee at rest or during exercise training is obtained. Among them, the maximum principal component time series is obtained by the following steps: Normalize the cerebral blood oxygen signals of all channels in the frontal lobe to obtain the standardized matrix , n is the total number of prefrontal channels, and the standardization method is to subtract the mean of each channel and divide it by the standard deviation of each channel; calculate the covariance matrix of the standardized brain blood oxygen signal matrix, calculate the eigenvalues ​​and corresponding eigenvectors of the covariance matrix; the eigenvector a corresponding to the maximum eigenvalue is the coefficient corresponding to the maximum principal component, and finally the maximum principal component is PCmax=a , where the eigenvector a is a column vector.

7. The multi-band neural feedback system combined with sports training according to claim 6, characterized in that: The heart rate respiratory rate extraction module calculates the respiratory rate at rest or in real time during exercise training in the following ways: Calculate the upper envelope parameter sequence, lower envelope parameter sequence, interval parameter sequence, amplitude parameter sequence and area parameter sequence of the maximum principal component time series; For the upper envelope parameter sequence, lower envelope parameter sequence, interval parameter sequence, amplitude parameter sequence and area parameter sequence of the maximum principal component time series, a 25.6-second time window is used to perform a sliding window fast Fourier transform, and the frequency corresponding to the maximum value in the range of 0.145-0.6Hz of the frequency domain signal of each sliding window is found, and the respiratory frequency time series calculated by the upper envelope parameter sequence, lower envelope parameter sequence, interval parameter sequence, amplitude parameter sequence and area parameter sequence can be obtained respectively; the five respiratory frequency time series are averaged to obtain the respiratory frequency time series of the trainee at rest or during exercise training.

8. The multi-band neural feedback system combined with sports training according to claim 6, characterized in that: The neural signal processing module calculates the brain blood oxygen concentration of the neural signal source in the following way: For single-channel cerebral blood oxygen signal, the cerebral blood oxygen signal from the noise channel is first excluded; then the time derivative distribution repair algorithm is used to remove motion artifact interference; signal spikes are removed; low-frequency drift and high-frequency instrument noise are removed; the hemodynamic separation algorithm is used to purify the neural source components in the cerebral blood oxygen signal; after purification by the hemodynamic separation algorithm, the cerebral blood oxygen concentration of the neural signal source is filtered using a 0.01-0.08Hz second-order bandpass Butterworth filter.

Images