A personalized brain development training method based on electroencephalogram signals

Abstract

The application relates to the cross field of biomedical engineering and artificial intelligence, and discloses a personalized brain power development training method based on electroencephalogram signals. The method comprises the following steps: collecting resting state and task state multi-channel electroencephalogram signals of a subject, constructing a functional connection matrix after pretreatment, identifying individualized weak connection target points through difference operation and cluster analysis; matching a neural feedback training protocol from a preset paradigm library based on the target points, and generating a feedback signal by extracting a target point synchronicity feature in real time during training to guide the subject to actively enhance the weak connection; updating the model after each training and dynamically optimizing subsequent parameters to form a closed-loop regulation. The application improves working memory and attention through individualized targeted training, induces neural plasticity, and realizes efficient and accurate brain power development.

Description

Technical Field

[0001] This invention belongs to the interdisciplinary field of biomedical engineering and artificial intelligence, and specifically relates to a personalized brainpower development and training method based on electroencephalogram (EEG) signals. Background Technology

[0002] With the rapid development of cognitive neuroscience and brain-computer interface technology, brain training methods based on electroencephalogram (EEG) signals have gradually become an important means of improving higher cognitive abilities such as attention, memory, and executive function. Current mainstream brain development programs mostly rely on standardized cognitive tasks or general neurofeedback paradigms. Their design logic is based on the average effect of the group, ignoring the differences in brain functional organization and structure between individuals. The human brain, as a highly heterogeneous and complex network system, exhibits strong individual specificity in its functional connectivity patterns in both resting and task states. This difference directly determines the response of different individuals to the same training stimuli. Using a uniform training strategy makes it difficult to activate or reshape the key neural pathways of a specific individual, leading to low training efficiency and even adaptive inhibition.

[0003] Functional connectomics analysis based on high-density electroencephalography (EEG) offers a new technical approach for personalized cognitive intervention. This research aims to quantify the intensity and topological characteristics of information interaction between brain regions by constructing a whole-brain functional connectome for an individual under resting or task-oriented conditions. Graph theory, as a core mathematical tool for analyzing complex networks, can characterize the organizational principles of brain networks from multiple dimensions, such as global efficiency, local clustering, node centrality, and modular structure, thereby identifying specific brain region pairs or sub-network modules that are closely related to target cognitive functions but exhibit abnormal connectivity. This analytical framework provides objective neural indicators for accurately locating individual cognitive problems.

[0004] Existing technologies for transforming brain network features into personalized training programs still suffer from the following problems: most systems rely solely on single-band power or simple coherence metrics for feedback regulation, lacking systematic modeling of the dynamic reorganization capabilities of the whole brain network; the selection of training paradigms often depends on expert experience or fixed rule bases, failing to establish a data-driven mapping relationship between "individual brain network features and effective intervention strategies." When faced with diverse cognitive enhancement goals, existing methods cannot intelligently recommend the most suitable neurofeedback patterns or cognitive task types based on the user's real-time or baseline brain network state, resulting in a lack of targeting and adaptability in interventions. A personalized brain development training method integrating brain connectomics, graph theory analysis, and intelligent recommendation mechanisms is needed to achieve precise identification and efficient regulation of weak points in individual brain function. Summary of the Invention

[0005] This invention provides a personalized brain development training method based on electroencephalogram (EEG) signals, aiming to solve the technical problem that traditional brain training programs use a uniform model and cannot accurately intervene in weak links of individual brain functional connectivity. This method collects multi-channel EEG signals from individuals in resting and task states to construct an individualized whole-brain functional connectivity map, identifies specific weak connectivity regions, and generates a matching neurofeedback training paradigm based on this map, achieving targeted enhancement of brain network efficiency.

[0006] This invention provides a personalized brainpower development training method based on electroencephalogram (EEG) signals, comprising: The first EEG signal sequence of the subject in a resting, closed-eye state was obtained using an EEG acquisition device; The EEG acquisition device is used to acquire the second EEG signal sequence of the subject during the performance of a standard cognitive task; The first EEG signal sequence and the second EEG signal sequence are preprocessed respectively. The preprocessing includes filtering, artifact removal, rereference, and segmentation to obtain clean time-domain EEG data. Based on the clean time-domain EEG data, the phase-locked value algorithm is used to calculate the functional connectivity strength between electrode pairs in the whole brain, generating a resting-state functional connectivity matrix and a task-state functional connectivity matrix. The resting state functional connection matrix and the task state functional connection matrix are compared to obtain the task-induced functional connection change matrix. Cluster analysis was performed on the task-induced functional connectivity change matrix to identify brain region pairs with insufficient or weakened functional connectivity enhancement, which were then used as a set of individualized weak connectivity targets. Based on the individualized set of weak connection targets, a corresponding training protocol is matched from a preset neurofeedback training paradigm library. The training protocol includes the target frequency band, feedback signal type, visual or auditory stimulation pattern, and training duration. During the training phase, the subjects' EEG signals are collected in real time, the synchronization characteristics of the electrode pairs corresponding to the individualized weak connection targets are extracted, and the signals are converted into feedback signals to guide the subjects to actively regulate their brain activity to enhance individualized weak connections. After each training session, the functional connectivity matrix is ​​updated, and the target points and paradigm parameters for the next training session are dynamically adjusted to form a closed-loop optimization mechanism.

[0007] Preferably, the first EEG signal sequence and the second EEG signal sequence are preprocessed respectively, including: Bandpass filtering is performed using a zero-phase fourth-order Butterworth filter. Independent component analysis was used to separate and remove interference components from eye movement, electromyography, and electrocardiography. Rereference was performed using a whole-brain average reference method; The time window, from one second before the event to three seconds after the event, is segmented based on the task event as the trigger point.

[0008] Preferably, based on the clean time-domain EEG data, a phase-locked value algorithm is used to calculate the functional connectivity strength between electrode pairs in the whole brain, including: For any two channels and In time The instantaneous phases at the points are respectively and Calculate the phase difference ; Calculate the phase lock value ,in Indicates time average. The imaginary unit; The value ranges from 0 to 1, and the closer the value is to 1, the stronger the phase synchronization between the two channels.

[0009] Preferably, the resting-state functional connectivity matrix and the task-state functional connectivity matrix are subtracted to obtain the task-induced functional connectivity change matrix, including: Define the task-induced function to connect each element of the change matrix. For task state Subtract resting state ; like If the value is less than 0.10.15, then the electrode pair is considered to be... Targets for weakening functional connectivity; like If the value is greater than +0.1 but less than the mean plus standard deviation of the population norm, it is determined to be a target with insufficient functional connectivity enhancement.

[0010] Preferably, cluster analysis is performed on the task-induced functional connectivity change matrix to identify brain region pairs with insufficient or weakened functional connectivity enhancement, including: Hierarchical clustering algorithm is used, Euclidean distance is used as the distance metric, and Ward's method is used as the connection criterion; The number of clusters is determined by maximizing the silhouette coefficient. Output several spatially adjacent clusters of weakly connected brain regions with low functional synergy as the individualized set of weakly connected target points.

[0011] Preferably, based on the individualized set of weak connection targets, a corresponding training protocol is matched from a preset neural feedback training paradigm library, including: For the weak connection between the prefrontal and parietal lobes, the feedback signal is a geometric pattern whose brightness increases with the increase of synchronicity; For the weak temporal-occipital connection, the feedback signal is an audio sequence in which the pitch increases with the degree of desynchronization.

[0012] Preferably, during the training phase, the subject's electroencephalogram (EEG) signals are collected in real time, and the synchronization characteristics of electrode pairs corresponding to weakly connected target points are extracted, including: Perform a short-time Fourier transform on the currently acquired EEG signals; Calculate the instantaneous phase of the target electrode pair within a specified frequency band; Calculate the phase lock value within the sliding window; The obtained phase-locked values ​​are linearly mapped to feedback strength values ​​from zero to 100.

[0013] Preferably, after each training iteration, the functional connectivity matrix is ​​updated, and the target points and paradigm parameters for the next training iteration are dynamically adjusted, including: If the phase lock value of a weakly connected target increases by more than 20% cumulatively in three consecutive training sessions, it will be removed from the target set. Reassess the relative weakness of the remaining connections and generate a new priority ranking list; The training protocol for the next stage is retrieved from the neural feedback training paradigm library according to the priority sorting list.

[0014] Preferably, the EEG acquisition device includes a dry electrode or wet electrode array with more than sixty-four channels.

[0015] Preferably, the standard cognitive tasks include working memory tasks, attention maintenance tasks, and executive control tasks.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention abandons the non-personalized model of fixed content and uniform intensity in traditional brain training, and for the first time realizes a paradigm shift from universal group application to individual precision.

[0017] 2. By constructing an individual-specific task-induced functional connectivity change map, we can objectively quantify the dynamic response of the brain network under cognitive load, thereby locating weak links in functional connectivity.

[0018] 3. Match the preset neurofeedback training paradigm, and ensure that the training always focuses on the neural pathways that need to be strengthened through real-time synchronization monitoring and closed-loop optimization mechanism.

[0019] 4. This method not only improves the targeting and effectiveness of brain training, but also avoids the waste of brain resources caused by ineffective stimulation. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the overall technical solution architecture of the present invention; Figure 2 This is a schematic diagram of the core principle framework for constructing the task-induced functional connectivity change matrix and identifying individualized weak connectivity targets in this invention. Figure 3 This is a flowchart illustrating the logical process of acquiring and preprocessing EEG signals in resting and task states in this invention. Figure 4 This is a logical flowchart of the functional connection modeling and cluster analysis based on phase-locked values ​​in this invention. Figure 5 This is a flowchart illustrating the logical flow of neural feedback training paradigm matching and real-time synchronous feedback execution in this invention. Figure 6 This is a schematic diagram of the multi-level interaction relationship and data flow of the post-training functional connection update and closed-loop optimization mechanism in this invention. Detailed Implementation

[0021] refer to Figures 1 to 6 This invention provides a personalized brain development training method based on electroencephalogram (EEG) signals. It acquires a first EEG signal sequence from a subject in a resting, eyes-closed state and a second EEG signal sequence during the execution of a standard cognitive task using an EEG acquisition device. The two types of signals are then systematically preprocessed, functional connectivity modeled, and weak connectivity target identified. Based on this, a neurofeedback training paradigm is matched, and finally, precise enhancement of brain network performance is achieved through real-time synchronization monitoring and a closed-loop optimization mechanism. The following will describe each step of this method in detail.

[0022] The method first involves acquiring the first EEG signal sequence of the subject in a resting, eyes-closed state using a 50-channel EEG acquisition device. This EEG acquisition device is equipped with an electrode array of no less than 64 channels, and can be in the form of dry or wet electrodes, with a sampling frequency of no less than 500 Hz, a common-mode rejection ratio of no less than 100 dB, and an input impedance of no more than megaohms. During the resting-state data acquisition process, the subject is in a quiet environment free from external interference, with eyes closed, remaining awake but not engaging in any active thought activity, for a duration of 5 minutes. This stage aims to capture the subject's baseline EEG activity patterns under no-task load, providing a reference benchmark for subsequent functional connectivity analysis.

[0023] Secondary EEG signal sequences were acquired from subjects performing standard cognitive tasks using the same EEG acquisition device. These standard cognitive tasks included three categories: working memory tasks, attention maintenance tasks, and executive control tasks. Each task lasted at least 3 minutes, with rest intervals of at least 2 minutes between tasks to prevent fatigue accumulation. The working memory task employed an N-back paradigm, requiring subjects to determine whether the currently presented stimulus was the same as the previous N stimuli. The attention maintenance task used a continuous performance task, requiring subjects to respond quickly to randomly presented target stimuli. The executive control task used a Stroop interference task, requiring subjects to correctly name the font color when a color word conflicted with the font color. All tasks presented visual stimuli on a computer screen, and the subjects' keystroke response time and accuracy were recorded for subsequent behavioral-neural coupling analysis.

[0024] After obtaining the first and second EEG signal sequences, the signal preprocessing stage begins. This stage performs filtering, artifact removal, rereference, and segmentation operations on the two types of raw EEG signals to generate clean time-domain EEG data. The filtering operation uses a zero-phase fourth-order Butterworth filter for bandpass filtering, with a passband range set from 0.5 Hz to 45 Hz, to remove DC drift and high-frequency electromyographic interference.

[0025] Artifact removal employs independent component analysis (ICA) to decompose the original multichannel signal into several independent components. Components related to eye movement, electromyography (EMG), and electrocardiography (ECG) are manually or automatically removed, while neurogenic components are preserved. Rereference processing uses a whole-brain average reference method, subtracting the instantaneous average of all channels from the signal of each channel to eliminate common-mode noise and enhance spatial resolution. Segmentation is performed using the task event as a trigger point, capturing a time window from one second before the event to three seconds after, with each data segment being four seconds long to ensure coverage of the complete neural response induced by the task. All preprocessing steps are executed in parallel on a dedicated digital signal processor to ensure real-time performance and consistency in data processing.

[0026] Based on the clean temporal EEG data described above, the functional connectivity modeling stage begins. This stage employs a phase-locked value algorithm to calculate the functional connectivity strength between electrode pairs throughout the brain, generating resting-state and task-state functional connectivity matrices, respectively.

[0027] The calculation process of the phase-locked value algorithm is as follows: For any two channels and In time The instantaneous phases at the points are respectively and Calculate the phase difference ; Calculate the phase-locked value , Indicates time average. The imaginary unit; The value ranges from 0 to 1. The closer the value is to 1, the stronger the phase synchronization between the two channels, reflecting a tighter functional connection. The calculation process is completed by a dedicated hardware accelerator within the functional connection modeling unit, supporting parallel calculation of all electrode pairs and ensuring the output of a complete connection matrix within milliseconds.

[0028] After obtaining the resting-state functional connectivity matrix and the task-state functional connectivity matrix, an interpolation operation is performed to generate the task-induced functional connectivity change matrix. Each element of this task-induced functional connectivity change matrix... Defined as task state Subtract resting state .like If the value is less than the preset threshold -0.15, then the electrode pair is determined to be faulty. The fact that this is a target for weakened functional connectivity indicates that this functional connectivity was not enhanced but weakened during cognitive tasks, possibly indicating an imbalance in the allocation of neural resources; if If the value is greater than the preset threshold + 0.1 but less than the mean plus standard deviation of the population norm, it is identified as a target with insufficient functional connectivity enhancement. This indicates that although the functional connectivity shows an enhancing trend, it has not reached the average level of the normal population and is considered a potential weak link. The population norm mean was pre-established using a large-scale database of healthy subjects, covering stratified statistical results of different ages, genders, and educational backgrounds.

[0029] Cluster analysis was performed on the task-induced functional connectivity change matrix to identify brain region pairs with insufficient or weakened functional connectivity, serving as an individualized set of weak connectivity targets. Hierarchical clustering was employed, with Euclidean distance as the distance metric and Ward's method as the connectivity criterion, aiming to minimize intra-cluster variance. The number of clusters was automatically determined by maximizing the silhouette coefficient, avoiding subjective setting. The clustering results output several spatially proximal weak connectivity brain region clusters with low functional synergy. Each cluster contains a set of electrode pairs that are anatomically adjacent but lack synchronicity in functional response. For example, the F3-F4 electrode pair in the prefrontal cortex and the P3-P4 electrode pair in the parietal cortex, if both exhibiting insufficient synchronicity enhancement during the task and being spatially close, may be clustered into the same cluster and labeled as "prefrontal-parietal synergy deficiency target areas".

[0030] After obtaining the individualized set of weak connection targets, the training paradigm matching stage begins. This stage matches the corresponding training protocol from a pre-defined neural feedback training paradigm library based on the set of weak connection targets. The paradigm library contains at least 12 basic paradigms, each corresponding to a specific combination of frequency bands, feedback logic, and stimulation pattern.

[0031] For the weak connectivity between the prefrontal and parietal lobes, a gamma-band synchronous enhancement paradigm is employed. The feedback signal is a geometric shape, such as a circle or cube, whose brightness increases with increasing synchronicity, displayed in the center of a high refresh rate display. For the weak connectivity between the temporal and occipital lobes, an alpha-band desynchronization inhibition paradigm is employed. The feedback signal is an audio sequence whose pitch increases with increasing desynchronization, output through high-fidelity headphones. Each paradigm includes parameters such as the target frequency band, feedback signal type, visual or auditory stimulation pattern, and training duration, stored in a structured database, supporting rapid retrieval by target brain region, indication tags, and other fields.

[0032] During the training phase, the subjects' electroencephalogram (EEG) signals are acquired in real time, and the synchronicity characteristics of the electrode pairs corresponding to the weak connectivity target points are extracted and converted into perceptible feedback signals to guide the subjects to actively regulate their brain activity to enhance the weak connectivity. The specific process of real-time extraction of synchronicity characteristics includes: performing a short-time Fourier transform on the currently acquired EEG signals, calculating the instantaneous phase of the target electrode pairs in a specified frequency band, and then calculating the phase lock value within a sliding window.

[0033] The window length is set to 500 milliseconds, with a step size of 100 milliseconds, ensuring sufficient time resolution and stability for the feedback signal. The obtained phase-locked value is linearly mapped to a feedback intensity value from 0 to 100, used to control the brightness of the visual graphic or the frequency of the audio signal. For example, when the phase-locked value is 0.6, the mapping is 60, corresponding to a graphic brightness of 60% of the maximum brightness, or an audio frequency offset of 60% of the reference frequency. This process is completed by a real-time feedback execution unit, which internally includes a signal feature extraction submodule and a feedback signal generation submodule, connected via a low-latency communication bus to ensure that the total delay from signal acquisition to feedback output does not exceed 300 milliseconds.

[0034] After each training session, a closed-loop optimization mechanism is initiated. This mechanism includes updating the functional connectivity matrix and dynamically adjusting the target points and paradigm parameters for the next training iteration. Specifically, if the phase-locked value of a weak connection target point increases by more than 20% cumulatively over three consecutive training sessions, it is removed from the target point set, indicating that the connection has been strengthened. Simultaneously, the system reassesses the relative weakness of the remaining connections and generates a new priority ranking list based on the latest task-induced functional connectivity change matrix. Priority ranking is based on... The absolute deviation of the value indicates its priority; the greater the deviation, the higher the priority. The paradigm scheduler retrieves the next stage of training protocol from the paradigm library based on the new list, which may involve adding new targets, adjusting target frequency bands, or switching feedback modalities. For example, if the prefrontal-parietal connection has reached the target, but a new temporal-parietal connection weakens, it automatically switches to a beta band synchronous enhancement paradigm targeting that region, and the feedback signal changes from visual to auditory to maintain the subject's interest and engagement.

[0035] Throughout the entire methodology, all data processing modules run on an embedded computing platform, featuring high throughput and low power consumption. The EEG signal acquisition unit is connected to the preprocessing unit via a shielded cable to ensure signal integrity; the functional connectivity modeling unit and the weak connectivity identification unit share a memory pool to avoid data copying overhead; the training paradigm matching unit and the real-time feedback execution unit work together through an event-driven mechanism, triggering feedback updates only when valid neural activity is detected. Furthermore, the system incorporates an anomaly handling mechanism: if a sudden change occurs in the real-time phase lock value, such as a single jump greater than 0.3, a data verification process is triggered, retracing the original signal from the previous 5 seconds to eliminate artifact interference; if the feedback intensity is less than the threshold of 20 for 10 consecutive seconds, the subject's attention is alerted, training is automatically paused, and guiding audio is played.

[0036] In summary, this method constructs an individual-specific task-induced functional connectivity change map to objectively quantify the dynamic response of the brain network under cognitive load, accurately locate weak links in functional connectivity, and match targeted neurofeedback training paradigms. Through real-time synchronous monitoring and closed-loop optimization mechanisms, it ensures that training always focuses on the neural pathways most in need of strengthening, achieving a paradigm shift from generalized approaches to individualized, precise brain development.

[0037] At the system level, this invention also provides a personalized brainpower development training system based on electroencephalogram (EEG) signals, which includes an EEG signal acquisition unit, a signal preprocessing unit, a functional connectivity modeling unit, a weak connectivity identification unit, a training paradigm matching unit, a real-time feedback execution unit, and a dynamic optimization unit. The EEG signal acquisition unit is equipped with an electrode array of 64 or more channels, supports dry and wet electrode compatibility, and features high sampling rate and low noise characteristics.

[0038] The signal preprocessing unit integrates filtering, artifact removal, rereference, and segmentation modules, employing a pipelined architecture to achieve millisecond-level processing latency. The functional connectivity modeling unit has a built-in phase-locked value calculation module, configured with a dedicated digital signal processor, supporting parallel calculation of the phase synchronization index for all electrode pairs. The weak connectivity identification unit integrates a clustering analysis engine, using contour coefficients to adaptively determine the optimal number of clusters, outputting spatially proximate weakly connected brain region clusters. The training paradigm matching unit stores a structured paradigm database; each record includes the target brain region, target frequency band, feedback modality, stimulus parameters, and indication label, supporting multi-condition joint queries.

[0039] The real-time feedback execution unit comprises a signal feature extraction submodule and a feedback signal generation submodule. The former calculates the phase-lock value in real time using a 500-millisecond window, while the latter maps the calculated phase-lock value to visual brightness or audio frequency and outputs it through a high refresh rate display or high-fidelity headphones. The dynamic optimization unit is equipped with a connection strength tracker and a paradigm scheduler. The former records the rate of change of the phase-lock value of each target point before and after each training session, while the latter determines whether the target point meets the standard based on the rate of change threshold and reallocates training resources to unmet or newly emerging weak connection regions. All units are interconnected via a high-speed internal bus, forming an integrated closed-loop control system to ensure efficient collaboration throughout the entire process from signal acquisition to training optimization.

[0040] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0041] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A personalized brainpower development training method based on electroencephalogram (EEG) signals, characterized in that, include: The first EEG signal sequence of the subject in a resting, closed-eye state was obtained using an EEG acquisition device; The EEG acquisition device is used to acquire the second EEG signal sequence of the subject during the performance of a standard cognitive task; The first EEG signal sequence and the second EEG signal sequence are preprocessed respectively. The preprocessing includes filtering, artifact removal, rereference, and segmentation to obtain clean time-domain EEG data. Based on the clean time-domain EEG data, the phase-locked value algorithm is used to calculate the functional connectivity strength between electrode pairs in the whole brain, generating a resting-state functional connectivity matrix and a task-state functional connectivity matrix. The resting state functional connection matrix and the task state functional connection matrix are compared to obtain the task-induced functional connection change matrix. Cluster analysis was performed on the task-induced functional connectivity change matrix to identify brain region pairs with insufficient or weakened functional connectivity enhancement, which were then used as a set of individualized weak connectivity targets. Based on the individualized set of weak connection targets, a corresponding training protocol is matched from a preset neurofeedback training paradigm library. The training protocol includes the target frequency band, feedback signal type, visual or auditory stimulation pattern, and training duration. During the training phase, the subjects' EEG signals are collected in real time, the synchronization characteristics of the electrode pairs corresponding to the individualized weak connection targets are extracted, and the signals are converted into feedback signals to guide the subjects to actively regulate their brain activity to enhance individualized weak connections. After each training session, the functional connectivity matrix is ​​updated, and the target points and paradigm parameters for the next training session are dynamically adjusted to form a closed-loop optimization mechanism.

2. The personalized brainpower development training method based on electroencephalogram (EEG) signals according to claim 1, characterized in that, Preprocessing is performed on the first EEG signal sequence and the second EEG signal sequence, including: Bandpass filtering is performed using a zero-phase fourth-order Butterworth filter. Independent component analysis was used to separate and remove interference components from eye movement, electromyography, and electrocardiography. Rereference was performed using a whole-brain average reference method; The time window, from one second before the event to three seconds after the event, is segmented based on the task event as the trigger point.

3. The personalized brainpower development training method based on electroencephalogram (EEG) signals according to claim 2, characterized in that, Based on the clean time-domain EEG data, the phase-locked value algorithm is used to calculate the functional connectivity strength between electrode pairs in the whole brain, including: For any two channels and In time The instantaneous phases at the points are respectively and Calculate the phase difference ; Calculate the phase lock value , Indicates time average. The imaginary unit; The value ranges from 0 to 1, and the closer the value is to 1, the stronger the phase synchronization between the two channels.

4. The personalized brainpower development training method based on electroencephalogram (EEG) signals according to claim 3, characterized in that, The resting-state functional connectivity matrix and the task-state functional connectivity matrix are subtracted to obtain the task-induced functional connectivity change matrix, including: Define the task-induced function to connect each element of the change matrix. For task state Subtract resting state ; like If the value is less than 0.10.15, then the electrode pair is considered to be... Targets for weakening functional connectivity; like If the value is greater than +0.1 but less than the mean plus standard deviation of the population norm, it is determined to be a target with insufficient functional connectivity enhancement.

5. The personalized brainpower development training method based on electroencephalogram (EEG) signals according to claim 4, characterized in that, Cluster analysis was performed on the task-induced functional connectivity change matrix to identify brain region pairs with insufficient or weakened functional connectivity, including: Hierarchical clustering algorithm is used, Euclidean distance is used as the distance metric, and Ward's method is used as the connection criterion; The number of clusters is determined by maximizing the silhouette coefficient. Output several spatially adjacent clusters of weakly connected brain regions with low functional synergy as the individualized set of weakly connected target points.

6. The personalized brainpower development training method based on electroencephalogram (EEG) signals according to claim 5, characterized in that, Based on the individualized set of weak connection targets, a corresponding training protocol is matched from a pre-defined neural feedback training paradigm library, including: For the weak connection between the prefrontal and parietal lobes, the feedback signal is a geometric pattern whose brightness increases with the increase of synchronicity; For the weak temporal-occipital connection, the feedback signal is an audio sequence in which the pitch increases with the degree of desynchronization.

7. The personalized brainpower development training method based on electroencephalogram (EEG) signals according to claim 6, characterized in that, During the training phase, the subjects' electroencephalogram (EEG) signals were collected in real time, and the synchronization characteristics of electrode pairs corresponding to weakly connected target points were extracted, including: Perform a short-time Fourier transform on the currently acquired EEG signals; Calculate the instantaneous phase of the target electrode pair within a specified frequency band; Calculate the phase lock value within the sliding window; The obtained phase-locked values ​​are linearly mapped to feedback strength values ​​from 0 to 100.

8. The personalized brainpower development training method based on electroencephalogram (EEG) signals according to claim 7, characterized in that, After each training session, the functional connectivity matrix is ​​updated, and the target points and paradigm parameters for the next training session are dynamically adjusted, including: If the phase lock value of a weakly connected target increases by more than 20% cumulatively in three consecutive training sessions, it will be removed from the target set. Reassess the relative weakness of the remaining connections and generate a new priority ranking list; The training protocol for the next stage is retrieved from the neural feedback training paradigm library according to the priority sorting list.

9. The personalized brainpower development training method based on electroencephalogram (EEG) signals according to claim 8, characterized in that, The EEG acquisition device includes an array of dry or wet electrodes with more than sixty-four channels.

10. The personalized brainpower development training method based on electroencephalogram (EEG) signals according to claim 9, characterized in that, The standard cognitive tasks include working memory tasks, attention maintenance tasks, and executive control tasks.

Images