Sequential rehabilitation training brain network dynamic monitoring method and device based on self-calibration initial space

Abstract

The application discloses a sequential rehabilitation training brain network dynamic monitoring method and device based on a self-calibration initial space, relates to the sequential rehabilitation training brain network monitoring field, and comprises the following steps: acquiring resting-state multi-channel brain blood oxygen data of a subject in a preset time period before sequential rehabilitation training starts, and constructing a baseline dynamic functional connection matrix sequence, and constructing a self-calibration initial space based on the same; acquiring real-time multi-channel brain blood oxygen data during peripheral electric stimulation and during execution of bilateral upper limb movement of the subject, and constructing real-time functional connection matrices among channels; and calibrating the real-time functional connection matrices of each stage based on the self-calibration initial space to obtain brain network deviation indexes of each stage. Through construction of the self-calibration initial space and real-time calibration in stages, the application solves the problem that individual background differences are difficult to eliminate in the prior art.

Description

Technical Field

[0001] This application relates to the field of sequential rehabilitation training brain network monitoring, and in particular to a method and device for dynamic monitoring of sequential rehabilitation training brain networks based on a self-calibrated initial space. Background Technology

[0002] Functional near-infrared spectroscopy (fNIRS) can reflect the brain's oxygenation dynamics under different task conditions by measuring changes in the concentrations of oxyhemoglobin and deoxyhemoglobin in brain tissue. Due to its non-invasive, portable, and usable characteristics in sports training scenarios, fNIRS is increasingly being applied to brain function monitoring during rehabilitation training.

[0003] In upper limb rehabilitation training, a sequential training paradigm combining peripheral electrical stimulation and bilateral upper limb movement has significant application value. This paradigm typically begins by providing sensory input and neural arousal through peripheral electrical stimulation, followed by bilateral upper limb movement to promote motor execution and bilateral hemispheric coordination. However, current brain function monitoring technologies still have the following limitations: Existing methods often use fixed thresholds, population averages, or single-time-point data as references, which are difficult to adapt to the differences in baseline cerebral oxygenation status among different subjects on different training days. The cerebral oxygenation signal of subjects may be affected by fatigue, emotions, attention levels, drug status, and diurnal fluctuations. Without individualized real-time self-calibration, the monitoring results are easily distorted. Summary of the Invention

[0004] The purpose of this application is to provide a method and device for dynamic monitoring of brain networks in sequential rehabilitation training based on self-calibrated initial space, which solves the problem that individual background differences are difficult to eliminate in the prior art.

[0005] To achieve the above objectives, this application provides the following solution: Firstly, this application provides a method for dynamic monitoring of brain networks in sequential rehabilitation training based on a self-calibrated initial space, including: Obtain resting-state multichannel cerebral blood oxygenation data of subjects during a preset time period before the start of sequential rehabilitation training; Based on resting-state multichannel cerebral blood oxygenation data, a baseline dynamic functional connectivity matrix sequence is constructed using a sliding window algorithm. The time-domain mean spatial matrix and the time-domain standard deviation spatial matrix are then calculated based on the baseline dynamic functional connectivity matrix sequence as the initial space for self-calibration. During the period when the subject receives peripheral electrical stimulation, real-time multi-channel cerebral blood oxygenation data of the subject's targeted sensorimotor brain regions are acquired, and a first real-time functional connectivity matrix between each channel is constructed; based on the self-calibrated initial space, the first real-time functional connectivity matrix is ​​normalized and aligned to obtain a first brain network deviation index. During the subjects' bilateral upper limb movements, real-time multi-channel cerebral blood oxygenation data of the subject's bilateral upper limb movement brain regions were acquired, and a second real-time functional connectivity matrix between each channel was constructed. Based on the self-calibrated initial space, the second real-time functional connectivity matrix was calibrated using a calibration process based on homologous connectivity difference constraints to obtain a second brain network deviation index.

[0006] Secondly, this application provides a sequential rehabilitation training brain network dynamic monitoring system based on self-calibrated initial space, including: acquisition equipment, central matrix processing module and interactive feedback display terminal; The data acquisition device is connected to the central matrix processing module and is used to collect the subject's resting state multi-channel cerebral blood oxygen data during a preset period before the start of sequential rehabilitation training, the subject's real-time multi-channel cerebral blood oxygen data of the targeted sensorimotor brain regions during the subject's peripheral electrical stimulation, and the subject's real-time multi-channel cerebral blood oxygen data of the bilateral upper limb motor brain regions during the subject's bilateral upper limb movement. The central matrix processing module is used to execute the above-mentioned sequential rehabilitation training brain network dynamic monitoring method based on self-calibrated initial space. The interactive feedback display terminal is used to display the dynamic monitoring results output by the central matrix processing module.

[0007] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method for dynamic monitoring of sequential rehabilitation training brain networks based on a self-calibrated initial space.

[0008] According to the specific embodiments provided in this application, this application has the following technical effects: This application provides a method and device for dynamic monitoring of brain networks in sequential rehabilitation training based on a self-calibrated initial space. A self-calibrated initial space is constructed using resting-state cerebral blood oxygenation data before training. Dynamic monitoring of the brain network is performed in stages during sequential training. During training, the self-calibrated initial space is used to normalize and align the first real-time functional connectivity matrix during peripheral electrical stimulation. The self-calibrated initial space is also used to calibrate the second real-time functional connectivity matrix during bilateral upper limb movement based on homologous connectivity difference constraints. The real-time monitoring results use the subject's current state as a reference. Individual background calibration is achieved based on the self-calibrated initial space, reducing the impact of individual differences, diurnal fluctuations, and background noise on the monitoring results. This solves the problem of difficulty in eliminating individual background differences in existing technologies. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is an application environment diagram of a sequential rehabilitation training brain network dynamic monitoring method based on self-calibrated initial space according to an embodiment of this application. Figure 2 A flowchart illustrating a sequential rehabilitation training brain network dynamic monitoring method based on a self-calibrated initial space, provided as an embodiment of this application; Figure 3 A schematic diagram illustrating the technical concept of a sequential rehabilitation training brain network dynamic monitoring method based on a self-calibrated initial space, provided in an embodiment of this application. Figure 4 A schematic diagram of self-calibrated initial space and real-time matrix alignment calibration provided in an embodiment of this application; Figure 5 A sequence diagram of sequential rehabilitation training provided in an embodiment of this application; Figure 6 A logical module block diagram of a sequential rehabilitation training brain network dynamic monitoring system based on a self-calibrated initial space provided in an embodiment of this application; Figure 7 A schematic diagram of the display interface of an interactive feedback display terminal provided in an embodiment of this application; Figure 8 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application.

[0011] Figure label: 100: fNIRS acquisition device; 200: Central matrix processing module; 201: Initial spatial self-calibration unit; 202: Sequential state temporal control unit; 203: Local targeted algebraic alignment calibration unit; 204: Homologous difference calibration unit; 205: Global whole-brain temporal history reconstruction and multi-parameter evaluation unit; 2051: Wavelet amplitude calculation unit; 2052: Lateral bias index calculation unit; 2053: Dynamic functional connectivity variability mean calculation unit; 300: Interactive feedback display terminal; 301: Bidirectional coordinate real-time renderer; 302: Post-meeting whole-brain report generator. Detailed Implementation

[0012] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0013] To make the above-mentioned objectives, features and advantages of this application more readily understood, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0014] The sequential rehabilitation training brain network dynamic monitoring method based on self-calibrated initial space provided in this application embodiment can be applied to, for example... Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be set up independently, integrated into server 104, or placed in the cloud or on another server. Terminal 102 can send the subject's resting-state multi-channel cerebral oxygenation data during a preset period before the start of sequential rehabilitation training, the subject's real-time multi-channel cerebral oxygenation data of the targeted sensorimotor brain regions during peripheral electrical stimulation, and the subject's real-time multi-channel cerebral oxygenation data of the bilateral upper limb motor brain regions during bilateral upper limb movement to server 104. After receiving the data, server 104 constructs a baseline dynamic functional connectivity matrix sequence based on the resting-state multi-channel cerebral oxygenation data using a sliding window algorithm, and calculates the time-domain mean spatial matrix and the time-domain standard deviation spatial matrix based on the baseline dynamic functional connectivity matrix sequence, which serve as the initial space for self-calibration. During peripheral electrical stimulation of the subject, a first real-time functional connectivity matrix is ​​constructed based on real-time multi-channel cerebral oxygenation data of the subject's targeted sensorimotor brain regions. The first real-time functional connectivity matrix is ​​then normalized and aligned using the self-calibrated initial space to obtain a first brain network deviation index. During bilateral upper limb movement of the subject, a second real-time functional connectivity matrix is ​​constructed based on real-time multi-channel cerebral oxygenation data of the subject's bilateral upper limb motor brain regions. The second real-time functional connectivity matrix is ​​then calibrated using a homologous connectivity difference constraint based on the self-calibrated initial space to obtain a second brain network deviation index. The server 104 can feed back the obtained first and second brain network deviation indices to the terminal 102.

[0015] Among them, terminal 102 can be, but is not limited to, various desktop computers, laptops, smartphones, tablets, IoT devices and portable wearable devices, and server 104 can be implemented by independent servers or server clusters composed of multiple servers, or it can be a cloud server.

[0016] In one exemplary embodiment, such as Figure 2 , Figure 3 and Figure 4 As shown, a method for dynamic monitoring of brain networks in sequential rehabilitation training based on a self-calibrated initial space is provided. This method is executed by a computer device, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the method is applied to... Figure 1 Taking server 104 as an example, the explanation includes the following steps 201 to 204. Wherein: Step 201: Obtain resting-state multichannel cerebral blood oxygenation data of the subject during a preset time period before the start of sequential rehabilitation training.

[0017] The preset time period is the period during which the subject is in a resting state before the peripheral electrical stimulation and bilateral upper limb movement begin, such as... Figure 5 The 10-minute resting state shown refers to a state of fixed gaze with eyes open, or a state of sitting still with eyes closed. Before the sequential rehabilitation training begins, the subject is in a resting state, and multi-channel cerebral oxygenation data can be acquired using an fNIRS acquisition device. This cerebral oxygenation data includes changes in oxyhemoglobin concentration (…). ), change in deoxyhemoglobin concentration ( At least one of the following: changes in total hemoglobin concentration ).

[0018] Furthermore, preprocessing of the resting-state multichannel cerebral oxygenation data is required. Preprocessing includes at least one of the following: detrending, bandpass filtering, motion artifact correction, bad channel removal, short channel regression, and standardization.

[0019] Step 202: Based on resting-state multichannel cerebral blood oxygenation data, a baseline dynamic functional connectivity matrix sequence is constructed using the sliding window algorithm, and the time-domain average spatial matrix is ​​calculated based on the baseline dynamic functional connectivity matrix sequence. and time-domain standard deviation space matrix , as the self-calibrated initial space.

[0020] The baseline dynamic functional connectivity matrix sequence was obtained by performing correlation analysis within a sliding window on multi-channel cerebral oxygenation data. The correlation analysis included Pearson correlation, partial correlation, coherence analysis, mutual information analysis, or a combination thereof.

[0021] Let the sequence of baseline dynamic functional connectivity matrices obtained in the resting phase be: in, This indicates the number of sliding windows in the resting state phase.

[0022] The time-domain averaged spatial matrix is ​​obtained by averaging the baseline dynamic functional connectivity matrix sequence element by element along the time dimension. : The standard deviation of each element in the time dimension of the matrix sequence is calculated to obtain the time-domain standard deviation space matrix. : .

[0023] Time-domain average spatial matrix Used to describe the average brain network connectivity background of the subject in the current resting state; temporal standard deviation spatial matrix This describes the temporal discreteness of the corresponding connected elements. Together, they constitute the self-calibrated initial space.

[0024] The self-calibrated initial space is a matrix space generated by a pure software algorithm, which does not depend on the specific hardware calibration structure of the fNIRS acquisition device.

[0025] Step 203: During the period when the subject receives peripheral electrical stimulation, acquire real-time multi-channel cerebral blood oxygenation data of the subject's targeted sensorimotor brain regions, and construct a first real-time functional connectivity matrix between each channel; perform algebraic alignment calibration (i.e., functional connectivity matrix normalization alignment) on the first real-time functional connectivity matrix based on the self-calibrated initial space to obtain a first brain network deviation index. First Brain Network Deviation Index This is used to characterize the brain network response changes of targeted sensorimotor brain regions relative to resting background during peripheral electrical stimulation.

[0026] The targeted sensorimotor brain regions include at least one of the subject's sensorimotor cortex, primary motor cortex, primary sensory cortex, supplementary motor area, premotor area, and prefrontal cortex.

[0027] In the first phase of sequential rehabilitation training, peripheral electrical stimulation (e.g., ...) was applied to the subjects. Figure 5 The 10-minute electrical stimulation shown extracts channel data targeting sensorimotor brain regions from real-time cerebral oxygenation data to construct a first real-time functional connectivity matrix. Peripheral electrical stimulation can be applied to relevant muscle groups in the upper limbs or peripheral nerve regions of the subject to provide sensory afferent or motor evoked stimuli.

[0028] Through the above calculations, the brain network response during the peripheral electrical stimulation phase is converted into a standardized positive and negative bidirectional exponent relative to the self-calibrated initial space. If A reading above the zero baseline indicates enhanced connectivity of the target brain region relative to the resting background; a reading below the zero baseline indicates reduced or inhibited connectivity of the target brain region relative to the resting background.

[0029] Step 204: During the subject's bilateral upper limb movement, real-time multi-channel cerebral blood oxygenation data of the subject's bilateral upper limb movement brain regions are acquired, and a second real-time functional connectivity matrix between each channel is constructed; based on the self-calibrated initial space, the second real-time functional connectivity matrix is ​​subjected to homology difference calibration (i.e., a calibration process based on homology connection difference constraints) to obtain the second brain network deviation index. Second Brain Network Deviation Index This is used to characterize the synergy of the motor network and changes in interhemispheric balance during bilateral upper limb movement. BNDImove = ||Ccal - Crest||; where... This represents the standardized motor state functional connectivity matrix obtained after performing homologous difference calibration on the second real-time functional connectivity matrix of the bilateral upper limb motor phase based on the resting state self-calibrated initial space. This represents the individualized baseline functional connectivity matrix constructed from the resting-state average functional connectivity matrix.

[0030] The bilateral upper limb motor brain regions include the homologous brain regions of the left and right motor cortices. The homologous differential calibration refers to calculating the changes in the correlation of homologous channels (functional connectivity changes) between the left and right motor-related brain regions and subtracting the background correlation (local connectivity strength) of the corresponding channel pairs in the self-calibrated initial space to reduce the influence of the subject's inherent hemispheric asymmetry.

[0031] In the second phase of sequential rehabilitation training, subjects performed bilateral upper limb motor tasks (such as... Figure 5 The 10-minute bilateral upper limb movement data shown is used to extract channel data from bilateral motor-related brain regions, especially homologous motor brain regions in the left and right hemispheres, from real-time cerebral blood oxygenation data, and to construct a second real-time functional connectivity matrix. Based on the self-calibrated initial space, homology difference calibration is performed to obtain the second brain network deviation index. Bilateral upper limb movements can include bilateral synchronous push-pull, bilateral grasping, bilateral flexion and extension, bilateral mirror movements, or other movement tasks that require the coordinated participation of the left and right upper limbs.

[0032] The peripheral electrical stimulation phase and the bilateral upper limb motor phase can be separated by timestamps, external trigger signals, operator input commands, or training device status signals.

[0033] Implementing steps 201 to 204 above, a self-calibrated initial space is constructed using resting-state cerebral blood oxygenation data before training. This allows real-time monitoring results to be referenced to the subject's current state, reducing the impact of individual differences, diurnal fluctuations, and background noise on the monitoring results. Furthermore, existing methods typically analyze the training process as a whole, failing to fully consider the differences between the sequential stages of "peripheral electrical stimulation followed by bilateral upper limb movement." The peripheral electrical stimulation stage primarily reflects sensory input response and cortical arousal, while the bilateral upper limb movement stage primarily reflects motor execution, bilateral hemispheric coordination, and network dynamic changes. The target brain regions and analytical parameters for these two stages are not entirely identical. Therefore, this application, for the sequential training process of "peripheral electrical stimulation—bilateral upper limb movement," sets up targeted sensorimotor brain region monitoring for the peripheral electrical stimulation stage and homologous differential monitoring for the bilateral upper limb movement stage, which can more accurately reflect the brain network response characteristics of different training stages.

[0034] In another exemplary embodiment of this application, in step 203, the expression for the deviation index of the first brain network is: In the formula, The deviation index of the first brain network at time t; Represents the first real-time function connection matrix The Middle Line number Column elements; Represents the time-domain average spatial matrix The Middle Line number Column elements; Represents the spatial matrix of time-domain standard deviation The Middle Line number Column elements; To prevent division by zero of constants; This represents the set of target channel pairs corresponding to the peripheral electrical stimulation phase; In another exemplary embodiment of this application, in step 204, the expression for the second brain network deviation index is: In the formula, The second brain network deviation index at time t; Represents the second real-time function connection matrix The Middle Line number Column elements; Represents the time-domain average spatial matrix The Middle Line number Column elements; Represents the spatial matrix of time-domain standard deviation The Middle Line number Column elements; To prevent division by zero of constants; This represents the set of target channel pairs corresponding to the movement phases of both upper limbs. It can include pathway pairs between homologous motor brain regions in the left and right hemispheres, or it can include motor brain regions on the affected side, motor brain regions on the healthy side, and their cross-hemispheric connectivity pathway pairs.

[0035] In another exemplary embodiment of this application, considering that existing display methods often focus on the absolute change in oxyhemoglobin or the intensity of local activation, it is difficult to present the positive and negative bidirectional shifts in the brain network state relative to its own baseline state, thus hindering the operator from intuitively distinguishing between enhanced brain network synergy, functional suppression, or fatigue-related changes. Therefore, the sequential rehabilitation training brain network dynamic monitoring method based on self-calibrated initial space further includes: Step 205: In a bidirectional coordinate system centered on the zero baseline, render the positive and negative bidirectional fluctuation curves of the first brain network deviation index and / or the second brain network deviation index in real time to achieve real-time visualization output.

[0036] In this application, the real-time functional connectivity matrix (first real-time functional connectivity matrix and second real-time functional connectivity matrix) is automatically calibrated using the time-domain average spatial matrix and the time-domain standard deviation spatial matrix. This allows the real-time functional connectivity matrix to be mapped to a zero-centered standardized space, thereby generating a brain network deviation index with positive and negative polarities. This enables a more accurate and intuitive distinction between changes related to enhanced brain network synergy, functional suppression, or fatigue.

[0037] In the bidirectional fluctuation curve, positive fluctuations above the zero baseline are used to characterize enhanced brain network synergy or enhanced activation of targeted brain regions, while negative fluctuations below the zero baseline are used to characterize reduced brain network synergy, functional suppression, or fatigue-related states.

[0038] In another exemplary embodiment of this application, considering that existing monitoring systems often struggle to simultaneously handle real-time monitoring and whole-brain assessment, real-time calculation of all parameters of the whole-brain network is computationally complex and may affect display latency; conversely, if only local real-time monitoring is performed, comprehensive information on changes in the whole-brain network cannot be obtained after training. Therefore, the sequential rehabilitation training brain network dynamic monitoring method based on a self-calibrated initial space further includes: Step 206: After the sequential rehabilitation training is completed, calculate the whole brain network change parameters based on the whole brain blood oxygenation data throughout the training period, generate multi-parameter evaluation results based on the whole brain network change parameters, and display them.

[0039] The parameters of change in the whole brain network include at least one of the following: wavelet amplitude (WA), laterality index (LI), and mean dynamic functional connectivity variability.

[0040] The wavelet amplitude is obtained by performing wavelet transform on the cerebral blood oxygenation data throughout the training period. It is used to characterize the change in the intensity of cerebral blood oxygenation oscillations within a preset frequency band, and its calculation formula is as follows: in, Indicates the frequency of cerebral blood oxygenation signal and time Wavelet coefficients below; This indicates the number of time sampling points. Wavelet amplitude values ​​can be calculated by single channel, target brain region, hemispherical region, or whole brain averaging.

[0041] The lateral bias index is used to characterize the functional asymmetry between bilateral brain regions and is calculated according to the following formula: in, Indicates the intensity of features of brain regions on the same side as the affected limb, or the intensity of features of predefined brain regions on the affected side; This indicates the intensity of features in the corresponding brain regions on the opposite side of the affected limb or the intensity of features in predefined, healthy-side related brain regions. This is a constant used to prevent the denominator from being zero. Brain region characteristic intensities include the mean amplitude of cerebral blood oxygenation, wavelet amplitude, functional connectivity strength, mean brain network deviation index, dynamic functional connectivity strength, or any combination thereof.

[0042] Among them, the mean variability of dynamic functional connectivity The standard deviation of the functional connectivity matrix elements over time is calculated using multiple sliding windows, and then averaged over the target channel pair set. The calculation includes: in, Indicates channel With channel In the The functional connection strength within each sliding window; This indicates that the standard deviation is calculated along the time dimension; This refers to a set of pathway pairs for the whole brain or a target brain region.

[0043] Functional connectivity strength is represented by wavelet phase coherence (WPCO), which identifies possible relationships by evaluating the matching between the instantaneous phases of two signals. It is relatively unaffected by noise and allows the detection of coordination between two dissimilar signals.

[0044] Assuming an oscillating signal exists and After wavelet transform, their sampling times are obtained respectively. scale point instantaneous phase at and Then their relative phase difference is The sine of the phase difference Sum and cosine components The wavelet phase coherence value is obtained by performing time-domain averaging over the entire signal length: Wavelet phase coherence is an indicator of the phase synchronization between two signals, reflecting the degree of correlation between their phases at a specific frequency. The wavelet phase coherence value ranges from 0 to 1. The closer the value is to 1, the smaller the phase difference between the two oscillating signal time series, indicating higher phase synchronization and a stronger correlation between the two oscillations. When the two oscillations are uncorrelated, the value approaches zero.

[0045] This application focuses on localized target brain regions or bilateral homologous brain regions during the real-time phase to reduce computational complexity and display latency; a whole-brain retrospective analysis is then performed after training, thus balancing real-time performance with the completeness of whole-brain assessment. Furthermore, this application incorporates wavelet amplitude, lateralization index, and mean dynamic functional connectivity variability in the post-assessment, enabling the description of brain network changes during training from three dimensions: frequency domain oscillations, hemispheric lateralization, and dynamic connectivity fluctuations, thereby enhancing the multi-dimensional interpretability of the assessment results.

[0046] In this application, before the start of sequential rehabilitation training, resting-state cerebral blood oxygenation data of the subjects are acquired, and a self-calibrated initial space including a time-domain mean spatial matrix and a time-domain standard deviation spatial matrix is ​​constructed using a pure software matrix algorithm. During the peripheral electrical stimulation phase, a real-time functional connectivity matrix is ​​constructed for the targeted sensorimotor brain regions, and the first brain network deviation index is calculated based on the self-calibrated initial space. During the bilateral upper limb motor phase, a real-time functional connectivity matrix is ​​constructed for the homologous channels of the bilateral motor cortex, and the second brain network deviation index is calculated based on the self-calibrated initial space. A positive and negative bidirectional fluctuation curve centered on the zero baseline is displayed in real time on the monitoring terminal. After the training is completed, the whole-brain cerebral blood oxygenation data is retrospectively analyzed to generate multi-parameter evaluation results including wavelet amplitude, lateral bias index, and mean dynamic functional connectivity variability. This application achieves individual baseline calibration through self-calibrated initial space, realizes phased monitoring of electrical stimulation response and bilateral motor response through sequential stage identification, and improves the real-time performance, stability and clinical interpretability of monitoring brain network status during rehabilitation training by combining local real-time display with post-meeting whole-brain assessment. It solves the problems of existing technologies, such as difficulty in eliminating individual baseline differences, inability to distinguish sequential training stage characteristics, lack of positive and negative polarity in real-time display, and insufficient parameters in post-meeting whole-brain assessment.

[0047] Based on the same inventive concept, this application also provides a system for implementing the above-described sequential rehabilitation training brain network dynamic monitoring method based on self-calibrated initial space. The solution provided by this device is similar to the implementation described in the above method. Therefore, the specific limitations of one or more embodiments of the sequential rehabilitation training brain network dynamic monitoring system based on self-calibrated initial space provided below can be found in the limitations of the sequential rehabilitation training brain network dynamic monitoring method based on self-calibrated initial space described above, and will not be repeated here.

[0048] In one exemplary embodiment, such as Figure 6 As shown, a sequential rehabilitation training brain network dynamic monitoring system based on self-calibrated initial space is provided, including: a data acquisition device (e.g., fNIRS data acquisition device 100), a central matrix processing module 200, and an interactive feedback display terminal 300.

[0049] The data acquisition device (e.g., fNIRS acquisition device 100) is communicatively connected to the central matrix processing module 200. It is used to acquire resting-state multi-channel cerebral oxygenation data of the subject during a preset time period before the start of sequential rehabilitation training; real-time multi-channel cerebral oxygenation data of the targeted sensorimotor brain regions during the subject's peripheral electrical stimulation; and real-time multi-channel cerebral oxygenation data of the bilateral upper limb motor brain regions during the subject's bilateral upper limb movements. This application does not limit the specific light source wavelength, probe arrangement, number of channels, or hardware circuit structure of the fNIRS acquisition device 100; these can be set according to actual needs.

[0050] Before training begins, subjects wear the fNIRS acquisition device 100 and remain in a resting state. The resting state can be either a fixed gaze with eyes open or a seated position with eyes closed. The fNIRS acquisition device 100 continuously acquires cerebral blood oxygenation data for a preset duration. The preset duration can be 3 to 10 minutes, preferably 5 to 6 minutes.

[0051] The central matrix processing module 200 is used to execute the aforementioned sequential rehabilitation training brain network dynamic monitoring method based on a self-calibrated initial space, to receive cerebral blood oxygenation data and perform brain network matrix calculation, stage identification, index calculation, and post-session evaluation. The central matrix processing module 200 can run on a local computer, embedded computing device, server, or cloud computing platform.

[0052] The interactive feedback display terminal 300 is used to display the dynamic monitoring results output by the central matrix processing module 200, such as... Figure 7 As shown.

[0053] In another exemplary embodiment of this application, the central matrix processing module 200 includes: an initial spatial self-calibration unit 201, a sequential state timing control unit 202, a local targeted algebraic alignment calibration unit 203, and a homogeneous differential calibration unit 204.

[0054] The initial spatial self-calibration unit 201 is used to acquire resting-state multi-channel cerebral oxygenation data of the subject within a preset time period before the start of sequential rehabilitation training. Based on the resting-state multi-channel cerebral oxygenation data, a baseline dynamic functional connectivity matrix sequence is constructed using a sliding window algorithm. The time-domain mean spatial matrix and the time-domain standard deviation spatial matrix are calculated based on this baseline dynamic functional connectivity matrix sequence, serving as the initial space for self-calibration. For each sliding window, the baseline dynamic functional connectivity matrix between channels is calculated. Functional connectivity calculations can use Pearson correlation, partial correlation, coherence, mutual information, or other calculation methods that can characterize the relationships between channels.

[0055] The sequential state timing control unit 202 is used to identify the peripheral electrical stimulation phase and the bilateral upper limb motor phase, and switch the target channel pair set according to the phase type. The identification method can be reading the trigger signal from the peripheral electrical stimulation device, reading the trigger signal from the motor training device, receiving manual markings from the operator, or automatically switching according to a preset training process. The sequential state timing control unit 202 receives the peripheral electrical stimulation start signal and marks the current data stream as the peripheral electrical stimulation phase.

[0056] The local targeted algebraic alignment calibration unit 203 is used to acquire real-time multi-channel cerebral blood oxygenation data of the subject's targeted sensorimotor brain region during the subject's peripheral electrical stimulation, and to construct a first real-time functional connectivity matrix between each channel; based on the self-calibrated initial space, the first real-time functional connectivity matrix is ​​normalized and aligned to obtain a first brain network deviation index.

[0057] The local targeted algebraic alignment calibration unit 203 selects the set of channels corresponding to the targeted sensorimotor brain regions from real-time cerebral blood oxygenation data based on preset brain region templates or channel markers. Within each real-time sliding window, the first real-time functional connectivity matrix is ​​calculated. And based on the target channel corresponding to the current stage, the set Calculate the deviation index of the first brain network.

[0058] The sequential state timing control unit 202 is used to receive the bilateral upper limb movement start signal and mark the current data stream as the bilateral upper limb movement stage.

[0059] The homologous differential calibration unit 204 is used to acquire real-time multi-channel cerebral blood oxygenation data of the subject's bilateral upper limb movement brain region during the subject's bilateral upper limb movement, and construct a second real-time functional connectivity matrix between each channel; based on the self-calibrated initial space, homologous differential calibration is performed on the second real-time functional connectivity matrix to obtain a second brain network deviation index.

[0060] The homologous differential calibration unit 204 extracts homologous channel data between the left and right motor-related brain regions from real-time cerebral blood oxygenation data. Within each real-time sliding window, the second real-time functional connectivity matrix is ​​calculated. And based on the target channel corresponding to the current stage, the set Calculate the deviation index of the second brain network.

[0061] In another exemplary embodiment of this application, the central matrix processing module 200 further includes: The global brain temporal reconstruction and multi-parameter assessment unit 205 is used to perform retrospective analysis based on whole-brain blood oxygenation data throughout the training period after the sequential rehabilitation training is completed, and to calculate the changes in whole-brain network parameters. These parameters include wavelet amplitude, lateralization index, and mean dynamic functional connectivity variability.

[0062] The wavelet amplitude is used to characterize the change in the oscillation intensity of the cerebral blood oxygenation signal within a preset frequency band. The global whole-brain temporal reconstruction and multi-parameter evaluation unit 205 performs wavelet transform on the cerebral blood oxygenation signal throughout the training period to obtain wavelet coefficients (W(f,t)) and calculates the wavelet amplitude. .

[0063] Laterality Index This method is used to characterize functional asymmetry between bilateral brain regions in a subject. It can be calculated based on the characteristic intensity of brain regions in the left and right hemispheres, or on the characteristic intensity of brain regions on the same side as and opposite to the affected limb. It can also be calculated based on the characteristic intensity of predefined brain regions related to the affected side and related brain regions on the healthy side. Brain region characteristic intensities include the average amplitude of cerebral oxygenation, wavelet amplitude, functional connectivity strength, mean brain network deviation index, dynamic functional connectivity strength, or any combination thereof.

[0064] Mean variability of dynamic functional connections The mean dynamic functional connectivity variability is used to characterize the degree of fluctuation in the state of brain network connectivity over time during training. It can be used to describe the flexibility, stability, or volatility of the state of brain network connectivity.

[0065] In another exemplary embodiment of this application, the interactive feedback display terminal 300 includes: A bidirectional coordinate real-time renderer 301 is used to establish a bidirectional coordinate system centered on the zero baseline on the screen, and to render in real time the positive and negative bidirectional fluctuation curves of the deviation index of the first brain network and / or the positive and negative bidirectional fluctuation curves of the deviation index of the second brain network. For example... Figure 7 As shown.

[0066] During the peripheral electrical stimulation phase, the curves displayed by the bidirectional coordinate real-time renderer 301 correspond to... It is used to reflect the brain network response of targeted sensorimotor brain regions to peripheral electrical stimulation.

[0067] During the bilateral upper limb movement phase, the curves displayed by the bidirectional coordinate real-time renderer 301 correspond to... It is used to reflect the coordinated state of bilateral motor-related brain regions during bilateral upper limb movement.

[0068] In one embodiment, the curve is displayed in a first display format when it is above the zero baseline, and in a second display format when it is below the zero baseline. The first and second display formats may use different colors, line styles, filled areas, or prompts, but the present invention does not limit the specific display style.

[0069] This display method transforms complex brain network matrix changes into easily understandable bidirectional fluctuation curves, allowing operators to observe changes in brain network states during training. Positive fluctuations can indicate enhanced brain network synergy or increased activation of targeted brain regions, while negative fluctuations can indicate reduced brain network synergy, functional suppression, or fatigue-related states.

[0070] The post-meeting whole-brain report generator 302 is used to generate a multi-parameter evaluation report based on changes in whole-brain network parameters and display it on the screen, such as... Figure 7 As shown. Multi-parameter assessment reports may include numerical tables, trend curves, brain region heatmaps, lateralization maps, dynamic functional connectivity matrices, or radar charts. Multi-parameter assessment results include at least one of the following: wavelet amplitude, lateralization index, mean dynamic functional connectivity variability, whole-brain average functional connectivity strength, whole-brain connectivity density, network efficiency, and clustering coefficient.

[0071] Among them, the mean functional connectivity strength (MFCS) of the whole brain is: The average functional connectivity strength of the whole brain is used to describe the overall functional synchronization level between different brain regions, reflecting the synergy of neural activities across the whole brain.

[0072] Where N represents the number of brain regions or channels; This represents the functional connectivity coefficient between brain region i and brain region j. An increased MFCS indicates enhanced overall synchronization between brain regions; a decreased MFCS indicates weakened functional integration of the brain network.

[0073] Among them, the network density (ND) of the whole brain is: Connection density measures the proportion of actual connections in a brain network to the total number of possible connections, reflecting the richness of network connectivity.

[0074] Where E represents the actual number of functional connections; N represents the number of nodes. High connection density indicates richer connections between brain regions; low connection density suggests reduced network connections or functional decoupling.

[0075] Among them, network efficiency: Network efficiency is used to evaluate the information transmission capacity of brain networks and to measure the level of network integration.

[0076] in, This represents the shortest path length between node i and node j. High network efficiency indicates faster and more efficient information transmission between brain regions; lower network efficiency indicates a decline in the integration of prompting functions. Here, "node" refers to a brain region or pathway.

[0077] Among them, the clustering coefficient is: Clustering coefficients reflect the degree to which neighboring nodes in a brain region form a local network structure, representing the local information processing capacity.

[0078] in, This represents the actual number of connections between node i's neighbors; This represents the number of connections (node ​​degree) of node i.

[0079] A multi-parameter evaluation report may include the following: (1) Peripheral electrical stimulation phase Mean, peak value, range of fluctuation, and duration; (2) Bilateral upper limb motor phase Mean, peak value, range of fluctuation, and duration; (3) Results of the changes in wavelet amplitude throughout the training period; (4) Lateral deviation index of the left and right hemispheres or the affected and healthy sides; (5) Mean value of dynamic functional connectivity variability in the whole brain or target brain region; (6) Heat map of whole brain functional connectivity matrix, distribution map of brain region parameters or trend curve.

[0080] The aforementioned multi-parameter assessment report is used to characterize the dynamic changes in brain networks during sequential rehabilitation training, and is not limited to disease diagnosis conclusions or treatment prescriptions.

[0081] This application can be applied to rehabilitation training status monitoring, brain function training feedback, rehabilitation equipment supporting software, fNIRS data processing platforms, and quantitative assessment systems for the rehabilitation process. Since the self-calibration initial space and brain network deviation index of this application are both implemented by software algorithms, it is compatible with different models of fNIRS acquisition devices 100, exhibiting good portability and industrial application value.

[0082] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 8 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores dynamic monitoring data of a sequential rehabilitation training brain network based on a self-calibrated initial space. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for dynamic monitoring of a sequential rehabilitation training brain network based on a self-calibrated initial space.

[0083] Figure 8 The structures shown are merely block diagrams of some structures related to the present application and do not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than shown in the figures, or combine certain components, or have different component arrangements. In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0084] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0085] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0086] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with relevant regulations and be authorized by the owner of the corresponding device.

[0087] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0088] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchain. The processors involved in the embodiments provided in this application may be, but are not limited to, general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc.

[0089] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0090] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for dynamic monitoring of brain networks in sequential rehabilitation training based on a self-calibrated initial space, characterized in that, include: Obtain resting-state multichannel cerebral blood oxygenation data of subjects during a preset time period before the start of sequential rehabilitation training; Based on resting-state multichannel cerebral blood oxygenation data, a baseline dynamic functional connectivity matrix sequence is constructed using a sliding window algorithm. The time-domain mean spatial matrix and the time-domain standard deviation spatial matrix are then calculated based on the baseline dynamic functional connectivity matrix sequence as the initial space for self-calibration. During the period when the subject receives peripheral electrical stimulation, real-time multi-channel cerebral blood oxygenation data of the subject's targeted sensorimotor brain regions are acquired, and a first real-time functional connectivity matrix between each channel is constructed; based on the self-calibrated initial space, the first real-time functional connectivity matrix is ​​normalized and aligned to obtain a first brain network deviation index. During the subjects' bilateral upper limb movements, real-time multi-channel cerebral blood oxygenation data of the subject's bilateral upper limb movement brain regions were acquired, and a second real-time functional connectivity matrix between each channel was constructed. Based on the self-calibrated initial space, the second real-time functional connectivity matrix was calibrated using a calibration process based on homologous connectivity difference constraints to obtain a second brain network deviation index.

2. The method for dynamic monitoring of brain networks in sequential rehabilitation training based on a self-calibrated initial space according to claim 1, characterized in that, The expression for the deviation index of the first brain network is: In the formula, The deviation index of the first brain network at time t; Represents the first real-time function connection matrix The Middle Line number Column elements; Represents the time-domain average spatial matrix The Middle Line number Column elements; Represents the spatial matrix of time-domain standard deviation The Middle Line number Column elements; To prevent division by zero of constants; This represents the set of target channel pairs corresponding to the peripheral electrical stimulation phase.

3. The method for dynamic monitoring of brain networks in sequential rehabilitation training based on a self-calibrated initial space according to claim 1, characterized in that, The expression for the deviation index of the second brain network is: In the formula, The second brain network deviation index at time t; Represents the second real-time function connection matrix The Middle Line number Column elements; Represents the time-domain average spatial matrix The Middle Line number Column elements; Represents the spatial matrix of time-domain standard deviation The Middle Line number Column elements; To prevent division by zero of constants; This represents the set of target channel pairs corresponding to the movement phases of both upper limbs.

4. The method for dynamic monitoring of brain networks in sequential rehabilitation training based on a self-calibrated initial space according to claim 1, characterized in that, The sequential rehabilitation training brain network dynamic monitoring method based on self-calibrated initial space also includes: In a bidirectional coordinate system centered on the zero baseline, the positive and negative bidirectional fluctuation curves of the first brain network deviation index and / or the second brain network deviation index are rendered in real time.

5. The method for dynamic monitoring of sequential rehabilitation training brain networks based on self-calibrated initial space according to claim 1 or 4, characterized in that, The sequential rehabilitation training brain network dynamic monitoring method based on self-calibrated initial space also includes: After the sequential rehabilitation training is completed, the whole brain blood oxygenation data throughout the training period are used to calculate the whole brain network change parameters, generate multi-parameter evaluation results based on the whole brain network change parameters, and display them.

6. A sequential rehabilitation training brain network dynamic monitoring system based on a self-calibrated initial space, characterized in that, include: Data acquisition equipment, central matrix processing module, and interactive feedback display terminal; The data acquisition device is connected to the central matrix processing module and is used to collect the subject's resting state multi-channel cerebral blood oxygen data during a preset period before the start of sequential rehabilitation training, the subject's real-time multi-channel cerebral blood oxygen data of the targeted sensorimotor brain regions during the subject's peripheral electrical stimulation, and the subject's real-time multi-channel cerebral blood oxygen data of the bilateral upper limb motor brain regions during the subject's bilateral upper limb movement. The central matrix processing module is used to execute the sequential rehabilitation training brain network dynamic monitoring method based on self-calibrated initial space as described in any one of claims 1 to 5. The interactive feedback display terminal is used to display the dynamic monitoring results output by the central matrix processing module.

7. The sequential rehabilitation training brain network dynamic monitoring system based on self-calibrated initial space according to claim 6, characterized in that, The central matrix processing module includes: The initial spatial self-calibration unit is used to acquire the resting state multi-channel cerebral blood oxygenation data of the subject within a preset time period before the start of sequential rehabilitation training; based on the resting state multi-channel cerebral blood oxygenation data, a baseline dynamic functional connectivity matrix sequence is constructed using a sliding window algorithm, and the time-domain mean spatial matrix and the time-domain standard deviation spatial matrix are calculated based on the baseline dynamic functional connectivity matrix sequence as the initial space for self-calibration; A sequential state timing control unit is used to identify the peripheral electrical stimulation phase and the bilateral upper limb motor phase; The local targeted functional connectivity matrix normalization and alignment unit is used to acquire real-time multi-channel cerebral blood oxygenation data of the subject's targeted sensorimotor brain regions during the subject's peripheral electrical stimulation, and to construct the first real-time functional connectivity matrix between each channel; based on the self-calibrated initial space, the first real-time functional connectivity matrix is ​​normalized and aligned to obtain the first brain network deviation index. The homologous differential calibration unit is used to acquire real-time multi-channel cerebral blood oxygenation data of the subject's bilateral upper limb movement brain regions during the subject's bilateral upper limb movement, and to construct a second real-time functional connectivity matrix between each channel; based on the self-calibrated initial space, the second real-time functional connectivity matrix is ​​calibrated using a homologous connectivity difference constraint to obtain a second brain network deviation index.

8. The sequential rehabilitation training brain network dynamic monitoring system based on self-calibrated initial space according to claim 7, characterized in that, The central matrix processing module also includes: The global whole-brain temporal reconstruction and multi-parameter assessment unit is used to calculate the changes in whole-brain network parameters based on whole-brain blood oxygenation data throughout the training period after the sequential rehabilitation training is completed.

9. The sequential rehabilitation training brain network dynamic monitoring system based on self-calibrated initial space according to claim 8, characterized in that, Interactive feedback display terminals include: A bidirectional coordinate real-time renderer is used to render the positive and negative bidirectional fluctuation curves of the first brain network deviation index and / or the positive and negative bidirectional fluctuation curves of the second brain network deviation index in real time in a bidirectional coordinate system centered on the zero baseline, and display them on the display screen. The post-meeting whole-brain report generator is used to generate multi-parameter evaluation reports based on changes in whole-brain network parameters and display them on the screen.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the sequential rehabilitation training brain network dynamic monitoring method based on a self-calibrated initial space as described in any one of claims 1-5.

Images