Cerebral stroke exercise rehabilitation system and method based on cortical hippocampus network neural oscillation regulation and control, electronic equipment and storage medium

Abstract

The invention discloses a cerebral apoplexy exercise rehabilitation system and method based on cortical hippocampus network neural oscillation regulation and control, electronic equipment and a storage medium, and the system comprises a cerebral apoplexy patient comprehensive evaluation module which is used for carrying out the comprehensive and systematic evaluation of a cerebral apoplexy patient; the individualized model building module is used for customizing an individualized model for a set patient; the transcranial alternating current stimulation accurate intervention module is used for accurately intervening transcranial alternating current stimulation; and the multi-modal joint evaluation and dynamic monitoring module is used for carrying out multi-modal joint evaluation and dynamic monitoring. According to the cerebral apoplexy exercise rehabilitation system and method based on cortical hippocampus network neural oscillation regulation and control, the electronic equipment and the storage medium, exercise function recovery can be promoted.

Description

Technical Field

[0001] This invention belongs to the field of sports rehabilitation technology, and relates to a sports rehabilitation system, and more particularly to a stroke sports rehabilitation system, method, electronic device and storage medium based on cortical hippocampal network neural oscillation regulation. Background Technology

[0002] The core of motor function rehabilitation after stroke lies in the reconstruction of neural plasticity compensatory circuits. However, current clinical strategies mostly focus on motor enhancement training, lacking in-depth understanding of the circuit mechanisms of learning and memory consolidation. Existing rehabilitation systems neglect memory consolidation regulation, leading to unstable motor memory encoding. Although transcranial alternating current stimulation (tACS) has achieved some success in regulating high-frequency oscillations to enhance motor recovery or waking-time memory, its intervention effect on slow-wave oscillations and its mechanism of action in sleep-dependent memory consolidation remain unclear.

[0003] S0, as a key rhythm during non-rapid eye movement (NREM) sleep, can coordinate the strength of synaptic connections in the cortical hippocampal network and promote the long-term consolidation of motor memory. However, current research has not yet clarified whether S0-tACS can effectively synchronize slow-wave rhythms during sleep to generate resonance, thereby enhancing synaptic plasticity to assist in post-stroke motor rehabilitation. This is a key problem that urgently needs to be solved in the field of stroke rehabilitation.

[0004] In view of this, there is an urgent need to design a new stroke motor rehabilitation system in order to overcome at least some of the aforementioned shortcomings of existing stroke motor rehabilitation systems. Summary of the Invention

[0005] This invention provides a stroke motor rehabilitation system, method, electronic device, and storage medium based on cortical hippocampal network neural oscillation modulation, which can help promote the recovery of motor function.

[0006] To solve the above-mentioned technical problems, according to one aspect of the present invention, the following technical solution is adopted:

[0007] A stroke motor rehabilitation system based on cortical hippocampal network neural oscillation modulation, the motor rehabilitation system comprising:

[0008] The comprehensive assessment module for stroke patients is used to assess stroke patients.

[0009] The individualized model building module is used to create customized individualized models for patients.

[0010] Transcranial alternating current stimulation (TAC) precision intervention module, used for precise intervention of transcranial alternating current stimulation;

[0011] The multimodal joint assessment and dynamic monitoring module is used to perform multimodal joint assessment and dynamic monitoring.

[0012] As one embodiment of the present invention, the comprehensive assessment module for stroke patients is used to conduct a comprehensive and systematic assessment of stroke patients, including: (1) detailed medical history collection; (2) neurological examination; (3) motor function assessment, using the Fugl-Meyer Assessment Scale (FMA) combined with a quantitative system for motor performance (such as surface electromyography sEMG); (4) scientific assessment of cognitive function, combining the Montreal Cognitive Assessment Scale (MoCA) and computerized neuropsychological tests (such as CANTAB); (5) comprehensive assessment of sleep quality, integrating Pittsburgh Sleep Quality Index (PSQI) and polysomnography (PSG) data; all data are input into the assessment algorithm, and a multi-factor weighted decision algorithm is used for quantitative scoring, and a multi-dimensional patient status profile is constructed.

[0013] The individualized model construction module, based on the severity of the patient's condition, physical function, and MRI / CT imaging results, selects patients suitable for treatment plans and tailors individualized clinical models for them, laying the foundation for subsequent precision treatment. The individualized model construction module specifically includes:

[0014] The feature extraction unit is used to extract key indicators based on structural images (MRI / CT), functional images (fMRI, DTI), and electroencephalogram (EEG) data obtained by the comprehensive assessment module, using feature selection algorithms in machine learning (such as LASSO regression and principal component analysis PCA), including: cortical hippocampal network functional connectivity strength, slow wave power spectral density, cross-frequency coupling index, Fugl-Meyer score change rate, etc.

[0015] The adaptive screening unit combines the extracted indicators and uses classification algorithms such as support vector machine (SVM) or random forest to classify and predict whether patients are suitable for S0-tACS treatment, and outputs "recommended / not recommended" labels.

[0016] The individual parameter modeling unit is used to group recommended patients based on their feature inputs using K-means clustering or hierarchical clustering algorithms, extract the optimal tACS stimulation parameters (including stimulation frequency, intensity, and duration) and placement location for each group, and finally construct an individualized treatment parameter matrix model based on this.

[0017] The model visualization feedback interface unit is used to output individual treatment models through the system interface, including patient neurological state atlases, recommended stimulation parameters, expected efficacy scores, and intervention risk assessment results, for clinicians' decision-making reference.

[0018] As one embodiment of the present invention, the transcranial alternating current stimulation precision intervention module includes:

[0019] The stimulation parameter optimization setting unit is used to dynamically identify key intervention windows in slow-wave sleep (SWS) based on the neurophysiological feature extraction algorithm, and automatically set individualized tACS stimulation parameters.

[0020] The stimulation site precise positioning unit is used to fuse multimodal image information and automatically recommend the tACS electrode placement position to achieve precise targeting;

[0021] The Stimulation Timing Scientific Planning Unit is used to develop individualized treatment schedules and dynamically adjust the intervention process;

[0022] The stimulation parameter optimization setting unit specifically includes:

[0023] The time window identification unit is used to perform real-time analysis of EEG signals using short-time Fourier transform (STFT) or wavelet transform, extract the energy trend in the 0.5–4Hz frequency band, and identify the S0 peak period by combining the slow wave rhythm amplitude threshold and event detection window as the intervention starting point.

[0024] The individual tolerance adaptive regulation unit is used to automatically adjust the stimulus intensity using an incremental load trial method based on baseline stimulus response assessment, within a range not exceeding 2 mA / cm². 2 Maximize neural plasticity within a safe threshold.

[0025] The frequency matching optimization model building unit is used to perform autocorrelation analysis on the patient's EEG data, extract its individualized S0 dominant frequency, and lock the tACS frequency within the range of ±0.2Hz of the dominant frequency to achieve resonant coupling of the endogenous rhythm.

[0026] The waveform generation unit is used to generate a phase-continuous stimulation waveform based on the set parameters using a sine wave function generator.

[0027] The precise positioning unit for the stimulation site specifically includes:

[0028] The brain region target identification unit is used to combine fMRI functional connectivity analysis with DTI reconstructed white matter fiber trajectories to identify key hubs in the S0–SWR pathway as candidate targets.

[0029] The electrode positioning optimization model construction unit is used to register the standard head model with the patient's individual MRI image, and combine the international 10-20 electrode system and nerve electric field simulation to output the optimal electrode combination and spatial coordinates.

[0030] The recommended interface for electrode patches is used to convert electrode position information into actual operating coordinates to assist in precise clinical electrode placement.

[0031] The stimulation timing scientific planning unit specifically includes:

[0032] The sleep stimulation scheduling unit is used to estimate the entry into the SWS period in advance based on the patient's daily PSG data and biorhythm prediction model, and to initiate EEG monitoring to confirm the stage status, so as to achieve precise intervention triggering.

[0033] The treatment frequency and cycle optimization model building unit is used to combine the patient's motor function recovery rate, Fugl-Meyer score improvement rate and PAC coupling strength change trend, and dynamically recommend treatment frequency and cycle using Bayesian optimization algorithm;

[0034] The treatment interruption automatic detection unit is used to automatically stop stimulation and adjust the subsequent plan when an EEG signal interruption or non-target rhythm is detected.

[0035] As one embodiment of the present invention, the multimodal joint evaluation and dynamic monitoring module includes:

[0036] The multimodal magnetic resonance imaging depth assessment unit is used to comprehensively assess the connectivity status and network plasticity evolution of brain regions before, during, and after treatment.

[0037] The in vivo electrophysiological real-time monitoring unit is used to monitor cortical-hippocampal network neural oscillations in real time during tACS intervention using high-density EEG, and to assess the intervention effect.

[0038] The multimodal magnetic resonance depth assessment unit includes:

[0039] The brain region connectivity pattern recognition unit is used to construct a structural connectivity map based on DTI data using the shortest path and trajectory tracking algorithm; combined with the Pearson correlation coefficient matrix extracted from functional magnetic resonance imaging, it generates a functional connectivity strength map.

[0040] The brain network topology attribute calculation unit is used to monitor the topological evolution of central nodes in the cortical-hippocampal motor memory network at different treatment stages using graph neural network analysis or traditional graph theory indicators.

[0041] The plasticity evolution trend determination unit is used to quantify the evolution of synergistic activation levels and assess neural plasticity and rehabilitation effects by using dynamic factor analysis or principal component rate of change analysis of multi-time point network diagrams.

[0042] The in vivo electrophysiological real-time monitoring unit includes:

[0043] The slow wave rhythm identification unit is used to extract EEG activity in the 0.5–4 Hz frequency band in real time. It uses wavelet packet decomposition or multi-scale entropy method to identify S0 rhythm features and evaluate their stability and amplitude fluctuations.

[0044] The S0–SWR cross-frequency coupling detection unit is used to quantify the coupling strength between the S0 band phase and the SWR high-frequency amplitude envelope using PAC calculation methods (such as the Tort or Canolty algorithm).

[0045] The intervention effect trend modeling unit is used to perform time series modeling of the coupling strength (MI value) before and after the intervention with behavioral assessment data. It uses multiple linear regression or random forest regression analysis to restore the association with neural synchronization and provide a basis for treatment optimization.

[0046] As one embodiment of the present invention, the stroke motor rehabilitation system further includes: a clinical function periodic assessment module, which is used to periodically assess the patient's motor function (using the Fugl-Meyer assessment and Wolf motor function test), cognitive function (using the MoCA), and activities of daily living (using the Barthel Index) weekly, dynamically monitor the patient's motor learning and memory abilities and the recovery process of motor function, and provide timely feedback on the treatment effect; the clinical function periodic assessment module includes:

[0047] A multidimensional clinical scoring interface was set up to collect patients' motor function scores (using the Fugl-Meyer Assessment Scale and Wolf Motor Function Test), cognitive function scores (using the MoCA Scale), and activities of daily living scores (using the Barthel Index) on a weekly basis.

[0048] The embedded data fusion algorithm unit is used to normalize the collected scale data by employing a multi-index scoring fusion algorithm based on weighted feature aggregation, and extract core dimension indicators through principal component analysis (PCA) or factor analysis to form the Composite Rehabilitation Progress Index (CRPI).

[0049] The dynamic trend modeling and anomaly detection algorithm unit is used to model the trend of patient rehabilitation indicators based on linear regression model, locally weighted regression (LOWESS) or time series prediction model (such as ARIMA, LSTM), determine the rate of change and inflection point of rehabilitation progress, and help identify the fluctuation and stagnation risk of efficacy.

[0050] The adaptive feedback adjustment unit links rehabilitation assessment results with neuromodulation parameters and, based on a set rule base (such as extending the tACS intervention time or adjusting the S0 rhythm frequency when rehabilitation progress is slow), achieves individualized program adjustment and closed-loop optimization.

[0051] According to another aspect of the present invention, the following technical solution is adopted: a method for motor rehabilitation of stroke based on cortical-hippocampal network neural oscillation modulation, the method comprising:

[0052] Comprehensive assessment steps for stroke patients: Stroke patients are assessed using the comprehensive assessment module for stroke patients;

[0053] Personalized model building steps: Use the personalized model building module to customize a personalized model for the specified patient;

[0054] Precise intervention steps for transcranial alternating current stimulation: Precise intervention using the transcranial alternating current stimulation precision intervention module;

[0055] Multimodal joint assessment and dynamic monitoring steps: Use the multimodal joint assessment and dynamic monitoring module to perform multimodal joint assessment and dynamic monitoring.

[0056] As one embodiment of the present invention, the comprehensive assessment module for stroke patients is used to conduct a comprehensive and systematic assessment of stroke patients, including: (1) detailed medical history collection; (2) neurological examination; (3) motor function assessment, using the Fugl-Meyer Assessment Scale (FMA) combined with a quantitative system for motor performance (such as surface electromyography sEMG); (4) scientific assessment of cognitive function, combining the Montreal Cognitive Assessment Scale (MoCA) and computerized neuropsychological tests (such as CANTAB); (5) comprehensive assessment of sleep quality, integrating Pittsburgh Sleep Quality Index (PSQI) and polysomnography (PSG) data; all data are input into the assessment algorithm, and a multi-factor weighted decision algorithm is used for quantitative scoring, and a multi-dimensional patient status profile is constructed.

[0057] The individualized model construction module, based on the severity of the patient's condition, physical function, and MRI / CT imaging results, selects patients suitable for treatment plans and tailors individualized clinical models for them, laying the foundation for subsequent precision treatment. The individualized model construction module specifically includes:

[0058] The feature extraction unit is used to extract key indicators based on structural images (MRI / CT), functional images (fMRI, DTI), and electroencephalogram (EEG) data obtained by the comprehensive assessment module, using feature selection algorithms in machine learning (such as LASSO regression and principal component analysis PCA), including: cortical hippocampal network functional connectivity strength, slow wave power spectral density, cross-frequency coupling index, Fugl-Meyer score change rate, etc.

[0059] The adaptive screening unit combines the extracted indicators and uses classification algorithms such as support vector machine (SVM) or random forest to classify and predict whether patients are suitable for S0-tACS treatment, and outputs "recommended / not recommended" labels.

[0060] The individual parameter modeling unit is used to group recommended patients based on their feature inputs using K-means clustering or hierarchical clustering algorithms, extract the optimal tACS stimulation parameters (including stimulation frequency, intensity, and duration) and placement location for each group, and finally construct an individualized treatment parameter matrix model based on this.

[0061] The model visualization feedback interface unit is used to output individual treatment models through the system interface, including patient neurological state atlases, recommended stimulation parameters, expected efficacy scores, and intervention risk assessment results, for clinicians' decision-making reference.

[0062] As one embodiment of the present invention, the transcranial alternating current stimulation precision intervention module includes:

[0063] Stimulation parameter optimization setting steps: Based on the neurophysiological feature extraction algorithm, the key intervention window of SWS during slow wave sleep is dynamically identified, and individualized tACS stimulation parameters are automatically set;

[0064] Precise stimulation site localization steps: By fusing multimodal imaging information, the placement of tACS electrodes is automatically recommended to achieve precise targeting;

[0065] The scientific planning steps for stimulation timing include: developing an individualized treatment schedule and dynamically adjusting the intervention process;

[0066] The stimulation parameter optimization setting step specifically includes:

[0067] Time window identification steps: Short-time Fourier transform (STFT) or wavelet transform is used to analyze the EEG signal in real time, extract the energy trend of the 0.5–4 Hz frequency band, and combine the slow wave rhythm amplitude threshold and event detection window to identify the S0 peak period as the intervention starting point;

[0068] Individual tolerance adaptive regulation steps: Based on baseline stimulus response assessment, the stimulus intensity is automatically adjusted using an incremental load trial-and-error method, not exceeding 2 mA / cm². 2 Maximize neural plasticity within a safe threshold.

[0069] The steps for constructing the frequency matching optimization model are as follows: perform autocorrelation analysis on the patient's EEG data, extract the individualized S0 dominant frequency, and lock the tACS frequency within the range of ±0.2Hz of the dominant frequency to achieve resonant coupling of the endogenous rhythm;

[0070] Waveform generation steps: A phase-continuous stimulation waveform is generated using a sine wave function generator based on the set parameters;

[0071] The precise positioning unit for the stimulation site specifically includes:

[0072] Brain region target identification steps: Combine fMRI functional connectivity analysis and DTI reconstructed white matter fiber trajectories to identify key hubs in the S0–SWR pathway as candidate targets.

[0073] Electrode localization optimization model construction steps: Based on the standard head model and the patient's individual MRI image registration, combined with the international 10-20 electrode system and nerve electric field simulation, the optimal electrode combination and spatial coordinates are output;

[0074] Recommended interface steps for electrode patching: Convert electrode position information into actual operational coordinates to assist in precise clinical electrode placement;

[0075] The stimulation timing scientific planning unit specifically includes:

[0076] Sleep stimulation scheduling steps: Based on the patient's daily PSG data and biorhythm prediction model, the entry into the SWS period is estimated in advance, and EEG monitoring is initiated to confirm the stage status, so as to achieve precise intervention triggering;

[0077] Steps for constructing the treatment frequency and cycle optimization model: Combining the patient's motor function recovery rate, Fugl-Meyer score improvement rate and PAC coupling strength change trend, a Bayesian optimization algorithm is used to dynamically recommend the treatment frequency and cycle;

[0078] Automatic treatment interruption detection procedure: When an interruption in the EEG signal or a non-target rhythm is detected, stimulation is automatically stopped and subsequent plans are adjusted.

[0079] As one embodiment of the present invention, the multimodal joint evaluation and dynamic monitoring steps include:

[0080] Multimodal magnetic resonance imaging (MRI) depth assessment steps: Comprehensive assessment of brain inter-brain connectivity and network plasticity evolution before, during, and after treatment;

[0081] In vivo electrophysiological real-time monitoring procedure: During tACS intervention, high-density EEG was used to monitor the cortico-hippocampal network neural oscillation activity in real time to assess the intervention effect.

[0082] The multimodal magnetic resonance depth assessment steps include:

[0083] Brain region connectivity pattern recognition steps: Based on DTI data, construct a structural connectivity map using the shortest path and trajectory tracking algorithm; combine the Pearson correlation coefficient matrix extracted from functional magnetic resonance imaging to generate a functional connectivity strength map;

[0084] Steps for calculating brain network topological properties: Use graph neural network analysis or traditional graph theory indicators to monitor the topological evolution of central nodes in the cortical-hippocampal motor memory network at different treatment stages;

[0085] Steps for determining the trend of plasticity evolution: Through dynamic factor analysis or principal component rate of change analysis of multi-time point network diagrams, the evolution of synergistic activation level is quantified to assess neural plasticity and rehabilitation effect.

[0086] In vivo electrophysiological real-time monitoring steps include:

[0087] Slow wave rhythm identification steps: extract EEG activity in the 0.5–4 Hz frequency band in real time, identify S0 rhythm features using wavelet packet decomposition or multi-scale entropy method, and evaluate its stability and amplitude fluctuations;

[0088] S0–SWR cross-frequency coupling detection steps: The coupling strength between the S0 band phase and the SWR high-frequency amplitude envelope is quantified using the PAC calculation method;

[0089] Intervention effect trend modeling steps: The coupling strength MI value before and after the intervention is modeled with behavioral assessment data in time series. Multiple linear regression or random forest regression analysis is used to restore the association with neural synchronization, providing a basis for treatment optimization.

[0090] The stroke rehabilitation method further includes a periodic clinical functional assessment step: weekly assessments of motor function (using the Fugl-Meyer assessment and Wolf motor function test), cognitive function (using the MoCA), and activities of daily living (using the Barthel Index) are conducted to dynamically monitor the patient's motor learning and memory abilities and the recovery progress of motor function, providing timely feedback on treatment effectiveness. This periodic clinical functional assessment step is implemented through a periodic clinical functional assessment module, which includes:

[0091] A multidimensional clinical scoring interface was set up to collect patients' motor function scores (using the Fugl-Meyer Assessment Scale and Wolf Motor Function Test), cognitive function scores (using the MoCA Scale), and activities of daily living scores (using the Barthel Index) on a weekly basis.

[0092] The embedded data fusion algorithm unit is used to normalize the collected scale data by employing a multi-index scoring fusion algorithm based on weighted feature aggregation, and extract core dimension indicators through principal component analysis (PCA) or factor analysis to form the Composite Rehabilitation Progress Index (CRPI).

[0093] The dynamic trend modeling and anomaly detection algorithm unit is used to model the trend of patient rehabilitation indicators based on linear regression model, locally weighted regression (LOWESS) or time series prediction model (such as ARIMA, LSTM), determine the rate of change and inflection point of rehabilitation progress, and help identify the fluctuation and stagnation risk of efficacy.

[0094] The adaptive feedback adjustment unit links rehabilitation assessment results with neuromodulation parameters and, based on a set rule base (such as extending the tACS intervention time or adjusting the S0 rhythm frequency when rehabilitation progress is slow), achieves individualized program adjustment and closed-loop optimization.

[0095] According to another aspect of the present invention, the following technical solution is adopted: an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above method.

[0096] According to another aspect of the present invention, the following technical solution is adopted: a storage medium storing computer program instructions thereon, which, when executed by a processor, implement the steps of the above-described method.

[0097] The beneficial effects of this invention are as follows: the stroke motor rehabilitation system, method, electronic device and storage medium based on cortical hippocampal network neural oscillation regulation proposed in this invention can help promote the recovery of motor function.

[0098] In one application scenario of the present invention, the stroke motor rehabilitation system and method based on cortical hippocampal network neural oscillation modulation has the following advantages:

[0099] Innovative Rehabilitation Theories and Mechanisms: Breaking through the limitations of traditional peripheral compensation models, this research pioneers a rehabilitation strategy based on a multi-scale theoretical framework of "oscillation-plasticity-network." By targeting and modulating the S0-SWR neural oscillations of the cortical hippocampal network, it effectively promotes the directional transfer and enhancement of motor memory, providing new theoretical support and technical pathways for stroke rehabilitation and leading technological innovation in the field of rehabilitation.

[0100] Achieving precise and personalized treatment: Utilizing multimodal neuroimaging technology, precise localization of stimulation sites is achieved, and tACS stimulation parameters are flexibly set according to individual patient differences, truly realizing a precise and personalized treatment model of "one policy for one person." This significantly improves treatment efficacy while greatly reducing the risk of adverse reactions, enhancing the patient's treatment experience and safety.

[0101] Multi-dimensional and comprehensive assessment and monitoring: Integrating multiple methods such as multimodal magnetic resonance imaging, in vivo electrophysiology, and clinical functional assessment, this approach comprehensively and dynamically monitors treatment effects and changes in neural plasticity from multiple dimensions, including macroscopic brain network structure, mesoscopic neural oscillation mechanisms, and actual clinical function. This provides rich and accurate scientific evidence for real-time optimization and adjustment of treatment plans, powerfully promoting the rapid translation of this technology from laboratory research to widespread clinical application.

[0102] Significant clinical and social value: By effectively regulating the neural mechanisms of sleep-dependent motor memory consolidation, it can significantly improve the motor learning and memory abilities of stroke patients, accelerate the recovery of motor function, and greatly improve patients' self-care ability and quality of life. At the same time, it reduces the burden of long-term care on families and the pressure on social medical resources, possessing immeasurable scientific value and far-reaching social significance. Attached Figure Description

[0103] Figure 1 This is a schematic diagram of the composition of a stroke motor rehabilitation system based on cortical hippocampal network neural oscillation regulation in one embodiment of the present invention.

[0104] Figure 2 This is a flowchart of a stroke motor rehabilitation method based on cortical hippocampal network neural oscillation modulation in one embodiment of the present invention.

[0105] Figure 3 This is a schematic diagram of the composition of a stroke rehabilitation system according to an embodiment of the present invention.

[0106] Figure 4 This is a schematic diagram of the composition of the individualized model construction module in one embodiment of the present invention.

[0107] Figure 5 This is a schematic diagram of the transcranial alternating current stimulation precision intervention module in one embodiment of the present invention.

[0108] Figure 6This is a schematic diagram of the composition of the multimodal joint evaluation and dynamic monitoring module in one embodiment of the present invention.

[0109] Figure 7 This is a schematic diagram of the composition of a clinical function periodic assessment module in one embodiment of the present invention.

[0110] Figure 8 This is a flowchart of a stroke rehabilitation method according to an embodiment of the present invention.

[0111] Figure 9 This is a schematic diagram of the composition of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0112] The preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0113] To further understand the present invention, preferred embodiments of the present invention are described below in conjunction with examples. However, it should be understood that these descriptions are only for further illustrating the features and advantages of the present invention, and not for limiting the scope of the claims of the present invention.

[0114] The description in this section pertains to only a few typical embodiments, and the present invention is not limited to the scope of the embodiments described. Substitution of identical or similar prior art methods with some technical features in the embodiments is also within the scope of the description and protection of this invention.

[0115] The steps described in the various embodiments in the specification are for illustrative purposes only, and the implementation of this application is not limited by the order of the steps.

[0116] The term "connection" in the instruction manual includes both direct and indirect connections.

[0117] This invention discloses a stroke motor rehabilitation system based on cortical hippocampal network neural oscillation modulation. Figure 1 , Figure 3 This is a schematic diagram of the composition of a stroke motor rehabilitation system based on cortical hippocampal network neural oscillation modulation in one embodiment of the present invention; please refer to [link / reference]. Figure 1 , Figure 3 The exercise rehabilitation system includes: a comprehensive assessment module 1 for stroke patients, an individualized model construction module 2, a transcranial alternating current stimulation precision intervention module 3, and a multimodal joint assessment and dynamic monitoring module 4.

[0118] The comprehensive assessment module 1 for stroke patients is used to assess stroke patients; the individualized model construction module 2 is used to set up a customized individualized model for patients; the transcranial alternating current stimulation precision intervention module 3 is used to precisely intervene in transcranial alternating current stimulation; and the multimodal joint assessment and dynamic monitoring module 4 is used to perform multimodal joint assessment and dynamic monitoring.

[0119] In one embodiment of the present invention, the comprehensive assessment module 1 for stroke patients is used to conduct a comprehensive and systematic assessment of stroke patients, covering multiple dimensions such as detailed medical history collection, professional neurological examination, accurate assessment of motor function (using the Fugl-Meyer assessment scale), scientific assessment of cognitive function (using the Montreal Cognitive Assessment Scale MoCA), and comprehensive assessment of sleep quality (using the Pittsburgh Sleep Quality Index PSQI).

[0120] In one embodiment, the comprehensive assessment module 1 for stroke patients is used to conduct a comprehensive and systematic assessment of stroke patients, including: (1) detailed medical history collection; (2) neurological examination; (3) motor function assessment, using the Fugl-Meyer Assessment Scale (FMA) combined with a quantitative system for motor performance (such as surface electromyography, sEMG); (4) scientific assessment of cognitive function, combining the Montreal Cognitive Assessment Scale (MoCA) and computerized neuropsychological tests (such as CANTAB); (5) comprehensive assessment of sleep quality, integrating Pittsburgh Sleep Quality Index (PSQI) and polysomnography (PSG) data; all data are input into the assessment algorithm, and a multi-factor weighted decision algorithm is used for quantitative scoring, and a multi-dimensional patient status profile is constructed.

[0121] The individualized model construction module 2 selects patients suitable for treatment plans based on the severity of their condition, their physical function, and the results of MRI and CT imaging examinations. It then tailors individualized clinical models for these patients, laying the foundation for subsequent precision treatment.

[0122] Figure 4 This is a schematic diagram illustrating the composition of the individualized model construction module in one embodiment of the present invention; please refer to [link / reference]. Figure 4 In one embodiment of the present invention, the individualized model construction module 2 specifically includes: a feature extraction unit 21, an adaptive screening unit 22, an individual parameter modeling unit 23, and a model visualization feedback interface unit 24.

[0123] The feature extraction unit 21 is used to extract key indicators based on structural images (MRI / CT), functional images (fMRI, DTI), and electroencephalogram (EEG) data obtained by the comprehensive evaluation module, using feature selection algorithms in machine learning (such as LASSO regression and principal component analysis PCA), including: cortical hippocampal network functional connectivity strength, slow wave power spectral density, cross-frequency coupling index, Fugl-Meyer score change rate, etc.

[0124] The adaptive screening unit 22 is used to combine the extracted indicators and use classification algorithms such as support vector machine (SVM) or random forest to classify and predict whether the patient is suitable for S0-tACS treatment, and output "recommended / not recommended" labels.

[0125] The individual parameter modeling unit 23 is used to group recommended patients based on their feature inputs using K-means clustering or hierarchical clustering algorithms, extract the best tACS stimulation parameters (including stimulation frequency, intensity, duration) and placement position in each group, and finally construct an individualized treatment parameter matrix model based on this.

[0126] The model visualization feedback interface unit 24 is used to output individual treatment models through the system interface, including patient neurological state atlas, recommended stimulation parameters, expected efficacy scores and intervention risk assessment results, for clinicians to use as a reference for decision-making.

[0127] Figure 5 This is a schematic diagram of the transcranial alternating current stimulation precision intervention module in one embodiment of the present invention; please refer to [link / reference]. Figure 5 In one embodiment of the present invention, the transcranial alternating current stimulation precision intervention module 3 includes: a stimulation parameter optimization setting unit 31, a stimulation site precise positioning unit 32, and a stimulation time scientific planning unit 33.

[0128] The stimulation parameter optimization setting unit 31 dynamically identifies the key intervention window of slow-wave sleep (SWS) based on the neurophysiological feature extraction algorithm and automatically sets individualized tACS stimulation parameters; the stimulation site precise positioning unit 32 integrates multimodal imaging information and automatically recommends the placement position of tACS electrodes to achieve precise targeting; the stimulation time scientific planning unit 33 formulates an individualized treatment schedule and dynamically adjusts the intervention process.

[0129] The stimulation parameter optimization setting unit 31 is used to accurately grasp the specific time window of slow-wave sleep (SWS) and use tACS to intervene in the S0-SWR neural oscillation of the patient's cortical hippocampal network; the stimulation frequency is set in the 0.5-4Hz band closely related to the S0 rhythm, and the stimulation intensity is flexibly adjusted according to the individual patient's tolerance, strictly controlled not to exceed 2mA / cm 2 Within the safe range, a sine wave is used as the stimulation waveform to ensure the effectiveness and safety of the stimulation;

[0130] The precise stimulation site localization unit 32 utilizes multimodal neuroimaging technologies such as functional magnetic resonance imaging (fMRI) and electroencephalography (EEG) to precisely locate key areas of the patient's cortical hippocampal network. The tACS electrode pads are accurately placed on the corresponding scalp location to ensure that stimulation is precisely applied to the target nerve oscillation area, achieving targeted intervention.

[0131] The stimulation time scientific planning unit 33 is used to immediately initiate tACS intervention after the patient enters the SWS stage. Each intervention lasts for 30-60 minutes, with 3-5 treatments per week, and continuous systemic treatment for 4-8 weeks to achieve the best therapeutic effect.

[0132] In one embodiment of the present invention, the stimulation parameter optimization setting unit 31 includes: a time window identification unit 311, an individual tolerance adaptive control unit 312, a frequency matching optimization model construction unit 313, and a waveform generation unit 314.

[0133] The time window identification unit 311 is used to analyze EEG signals in real time using short-time Fourier transform (STFT) or wavelet transform to extract the energy change trend in the 0.5–4Hz frequency band; and to identify the S0 peak period as the intervention starting point by combining the slow wave rhythm amplitude threshold and the event detection window.

[0134] The individual tolerance adaptive control unit 312 is used to automatically adjust the stimulus intensity using the patient's baseline stimulus response assessment (such as subjective discomfort level, EEG evoked potential analysis) and the staircase algorithm, ensuring maximum neuroplasticity while not exceeding 2 mA / cm². 2 Maximize neural plasticity within a safe threshold;

[0135] The frequency matching optimization model construction unit 313 is used to extract the individualized S0 peak frequency by performing autocorrelation analysis on the EEG data of each patient, and lock the tACS frequency within the range of ±0.2Hz of the peak frequency to ensure that the stimulus and the endogenous oscillation rhythm resonate and couple.

[0136] The waveform generation unit 314 is used to generate a phase-continuous waveform based on the above-mentioned tACS stimulation parameters using a sine wave function generator.

[0137] The precise location unit 32 for stimulation sites includes: a brain region target identification unit 321, an electrode location optimization model construction unit 322, and an automatic electrode patch recommendation interface 323.

[0138] The brain region target identification unit 321 is used to combine fMRI functional connectivity analysis (such as the resting state functional connectivity strength between the cortex and hippocampus) with white matter fiber trajectories constructed by DTI to screen key hub regions in the S0–SWR pathway (such as the parahippocampal gyrus, medial prefrontal cortex, etc.) as candidate stimulation targets.

[0139] The electrode positioning optimization model construction unit 322 is used to register the standard head model (MNI space) with the patient's individual MRI, and output the optimal electrode combination and coordinate information by combining the international 10-20 electrode placement system and the results of nerve electric field simulation (using FEM finite element method modeling).

[0140] The electrode patch automatic recommendation interface 323 is used to convert the final coordinates into the actual operating coordinate system, assisting medical staff in accurately placing the electrodes.

[0141] The stimulation time scientific planning unit 33 includes: a sleep period stimulation scheduling unit 331, a treatment frequency and cycle optimization model construction unit 332, and a treatment interruption automatic detection unit 333.

[0142] The sleep stimulation scheduling unit 331 is used to combine the patient's daily PSG data with a biological rhythm prediction model (such as sleep stage prediction based on a long short-term memory LSTM network) to estimate the entry into the SWS period in advance, and to initiate EEG monitoring to confirm the stage status and accurately trigger intervention.

[0143] The treatment frequency and cycle optimization model building unit 332 is used to dynamically recommend treatment frequency (e.g., 3–5 times per week) and cycle (e.g., 4–8 weeks) based on the patient's motor function recovery speed, Fugl-Meyer score improvement rate and PAC coupling strength change trend, using Bayesian optimization algorithm.

[0144] The automatic treatment interruption detection unit 333 is used to stop stimulation and adjust subsequent plans when EEG interruption or non-target rhythm dominance is detected during monitoring.

[0145] Figure 6 This is a schematic diagram illustrating the composition of the multimodal joint evaluation and dynamic monitoring module in one embodiment of the present invention; please refer to [link / reference]. Figure 6 In one embodiment of the present invention, the multimodal joint assessment and dynamic monitoring module 4 includes: a multimodal magnetic resonance imaging depth assessment unit 41 and an in vivo electrophysiological real-time monitoring unit 42. The multimodal magnetic resonance imaging depth assessment unit 41 comprehensively assesses the connectivity status and network plasticity evolution of brain regions before, during, and after treatment; the in vivo electrophysiological real-time monitoring unit 42 is used to monitor the cortico-hippocampal network neural oscillation activity in real time using high-density EEG during tACS intervention to assess the intervention effect.

[0146] In one embodiment, the multimodal magnetic resonance depth assessment unit 41 is used to analyze the brain region connectivity patterns, network topology changes, and neural plasticity development of the motor learning memory consolidation network from a macroscopic perspective before treatment, during treatment (every 2 weeks), and after treatment, using multimodal magnetic resonance techniques such as diffusion tensor imaging (DTI) and functional connectivity magnetic resonance imaging (fcMRI), so as to comprehensively grasp the dynamic evolution of brain network structure and function.

[0147] The in vivo electrophysiological real-time monitoring unit 42 is used to monitor the neural oscillation activity of the cortical hippocampal network in real time during tACS intervention by placing high-density EEG electrodes on the scalp. Close observation of the S0 rhythm entrainment effect and changes in S0-SWR cross-frequency phase-amplitude coupling (PAC) allows for precise interpretation of the neural oscillation regulation mechanism at the mesoscopic level, providing crucial evidence for optimizing treatment plans.

[0148] The multimodal magnetic resonance depth assessment unit 41 specifically includes: a brain region connectivity pattern recognition unit 411, a brain network topology attribute calculation unit 412, and a plasticity evolution trend determination unit 413.

[0149] The brain region connectivity pattern recognition unit 411 is used to construct a structural connectivity map of the cortical hippocampal network using diffusion tensor imaging (DTI) data and shortest path algorithm (Dijkstra's algorithm) and trajectory tracking algorithm (FACT or iFOD2); and generate a functional connectivity intensity map by combining the Pearson correlation coefficient matrix extracted from functional connectivity magnetic resonance imaging (fcMRI).

[0150] The brain network topology attribute calculation unit 412 is used to use graph neural network analysis or traditional graph theory index calculation (such as node degree, local clustering coefficient, path length, modularity, etc.) to monitor the topological evolution of the central nodes of the cortical-hippocampal motor memory network at different treatment stages and assess the development trend of neural plasticity.

[0151] The plasticity evolution trend determination unit 413 is used to perform dynamic factor analysis or principal component change rate analysis of multi-time point network diagrams, quantify the evolution of synergistic activation levels, and evaluate neural plasticity and rehabilitation effects.

[0152] The in vivo electrophysiological real-time monitoring unit 42 specifically includes: a slow wave rhythm recognition unit 421, an S0-SWR cross-frequency coupling detection unit 422, and an intervention effect trend modeling unit 423.

[0153] The slow wave rhythm identification unit 421 extracts EEG activity in the 0.5–4 Hz frequency band in real time, and uses wavelet packet decomposition or multi-scale entropy method to identify S0 rhythm features and evaluate its stability and amplitude fluctuations.

[0154] The S0–SWR cross-frequency coupling detection unit 422 uses the PAC calculation method (such as the Tort or Canolty algorithm) to quantify the coupling strength between the S0 band phase and the SWR high-frequency amplitude envelope.

[0155] The intervention effect trend modeling unit 423 performs time series modeling of the coupling strength (MI value) before and after the intervention with behavioral assessment data, and uses multiple linear regression or random forest regression analysis to restore the association with neural synchronization, providing a basis for treatment optimization.

[0156] The stroke rehabilitation system may further include a periodic clinical function assessment module 5, which is used to periodically assess the patient's motor function (using Fugl-Meyer assessment and Wolf motor function test), cognitive function (using MoCA), and activities of daily living (using Barthel index) on a weekly basis, dynamically monitor the patient's motor learning and memory abilities and the recovery process of motor function, and provide timely feedback on the treatment effect.

[0157] Figure 7 This is a schematic diagram of the composition of a clinical functional periodic assessment module in one embodiment of the present invention; please refer to [link / reference]. Figure 7 In one embodiment of the present invention, the periodic clinical function assessment module 5 includes: a multidimensional clinical score acquisition interface 51, an embedded data fusion algorithm unit 52, a dynamic trend modeling and anomaly detection algorithm unit 53, and an adaptive feedback adjustment unit 54.

[0158] The multidimensional clinical score acquisition interface 51 is used to collect patients' motor function scores (using the Fugl-Meyer assessment scale and Wolf motor function test), cognitive function scores (using the MoCA scale), and activities of daily living scores (using the Barthel index) on a weekly basis.

[0159] The embedded data fusion algorithm unit 52 is used to normalize the collected scale data by employing a multi-index scoring fusion algorithm based on weighted feature aggregation, and extract core dimension indicators through principal component analysis (PCA) or factor analysis to form the Composite Rehabilitation Progress Index (CRPI).

[0160] The dynamic trend modeling and anomaly detection algorithm unit 53 is used to model the trend of patient rehabilitation indicators based on linear regression model, local weighted regression (LOWESS) or time series prediction model (such as ARIMA, LSTM), determine the rate of change and inflection point of rehabilitation progress, and help identify the fluctuation and stagnation risk of efficacy.

[0161] The adaptive feedback adjustment unit 54 links the rehabilitation assessment results with the neuromodulation parameters and, based on the set rule base (such as extending the tACS intervention time or adjusting the S0 rhythm frequency when rehabilitation progress is slow), realizes individualized plan adjustment and closed-loop optimization.

[0162] This invention further discloses a method for motor rehabilitation of stroke based on the neural oscillation regulation of the cortical hippocampal network. Figure 2 and Figure 8 This is a flowchart of a stroke rehabilitation method according to an embodiment of the present invention; please refer to [link / reference]. Figure 2 , Figure 8 The sports rehabilitation method includes:

[0163]

Step S1

[0164]

Step S2

[0165]

Step S3

[0166]

Step S4

[0167] In one embodiment of the present invention, the comprehensive assessment of stroke patients involves a comprehensive and systematic assessment of stroke patients, covering multiple dimensions such as detailed medical history collection, professional neurological examination, accurate assessment of motor function (using the Fugl-Meyer Assessment Scale), scientific assessment of cognitive function (using the Montreal Cognitive Assessment Scale MoCA), and comprehensive assessment of sleep quality (using the Pittsburgh Sleep Quality Index PSQI).

[0168] In one embodiment, the comprehensive assessment of stroke patients involves a comprehensive and systematic evaluation of stroke patients, including: (1) detailed medical history collection; (2) neurological examination; (3) motor function assessment, using the Fugl-Meyer Assessment Scale (FMA) combined with a quantitative system for motor performance (such as surface electromyography, sEMG); (4) scientific assessment of cognitive function, combining the Montreal Cognitive Assessment Scale (MoCA) and computerized neuropsychological tests (such as CANTAB); and (5) comprehensive assessment of sleep quality, integrating Pittsburgh Sleep Quality Index (PSQI) and polysomnography (PSG) data. All data are input into the assessment algorithm, and a multi-factor weighted decision algorithm is used for quantitative scoring, and a multi-dimensional patient status profile is constructed.

[0169] The individualized model building module selects patients suitable for treatment plans based on the severity of their condition, their physical function, and MRI and CT imaging results, and tailors individualized clinical models for them, laying the foundation for subsequent precision treatment.

[0170] The precise intervention steps of transcranial alternating current stimulation include:

[0171] Stimulation parameter optimization steps: Accurately grasp the specific time window of slow-wave sleep (SWS) and use tACS to intervene in the S0-SWR neural oscillations of the patient's cortical hippocampal network; set the stimulation frequency in the 0.5-4Hz band closely related to the S0 rhythm, and flexibly adjust the stimulation intensity according to the individual patient's tolerance, strictly controlling it to not exceed 2mA / cm. 2 Within the safe range, a sine wave is used as the stimulation waveform to ensure the effectiveness and safety of the stimulation;

[0172] Precise stimulation site localization steps: Utilizing multimodal neuroimaging techniques such as functional magnetic resonance imaging (fMRI) and electroencephalography (EEG), key areas of the patient's cortical hippocampal network are precisely located. The tACS electrode pads are accurately placed on the corresponding scalp location to ensure that stimulation is precisely applied to the target nerve oscillation area, achieving targeted intervention.

[0173] Scientific planning steps for stimulation time: After the patient enters the SWS stage, tACS intervention is initiated immediately. Each intervention lasts 30-60 minutes, with 3-5 treatments per week, and the systemic treatment is carried out continuously for 4-8 weeks to achieve the best treatment effect.

[0174] The precise intervention steps of transcranial alternating current stimulation include:

[0175] Stimulation parameter optimization setting steps: Based on the neurophysiological feature extraction algorithm, the specific time window of slow wave sleep (SWS) is dynamically identified, and individualized tACS stimulation parameters are automatically set;

[0176] Precise localization of stimulation sites: Based on multimodal imaging information, the placement of tACS electrodes is automatically recommended;

[0177] The scientific planning steps for stimulation timing include: developing individualized treatment plans and dynamic treatment scheduling.

[0178] The steps for optimizing the stimulation parameters include:

[0179] Time window identification steps: Short-time Fourier transform (STFT) or wavelet transform is used to analyze EEG signals in real time and extract the energy change trend in the 0.5–4Hz frequency band; combined with the slow wave rhythm amplitude threshold and event detection window, the S0 peak period is identified as the intervention starting point;

[0180] Individual tolerance adaptive regulation steps: Utilizing patient baseline stimulus response assessments (such as subjective discomfort level and EEG evoked potential analysis), the stimulus intensity is automatically adjusted using the staircase algorithm to ensure maximum neural plasticity without exceeding 2 mA / cm². 2 Safety threshold;

[0181] Steps for constructing the frequency matching optimization model: By performing autocorrelation analysis on the EEG data of each patient, the individualized S0 dominant frequency (peak frequency) is extracted, and the tACS frequency is locked within the range of ±0.2Hz of the dominant frequency to ensure that the stimulus and the intrinsic oscillatory rhythm resonate and couple.

[0182] Waveform generation steps: A sine wave function generator is used to generate a phase-continuous waveform based on the above tACS stimulation parameters.

[0183] The steps for precisely locating the stimulation site include:

[0184] Brain region target identification steps: Integrate fMRI functional connectivity analysis (such as the resting state functional connectivity strength between the cortex and hippocampus) with white matter fiber trajectories constructed by DTI to screen key hub regions in the S0–SWR pathway (such as the parahippocampal gyrus, medial prefrontal cortex, etc.) as candidate stimulation targets.

[0185] Electrode positioning optimization model construction steps: Based on the standard head model (MNI space) and the patient's individual MRI, the optimal electrode combination and coordinate information are output by combining the international 10-20 electrode placement system and the results of nerve electric field simulation (using FEM finite element method modeling).

[0186] The automatic electrode patch recommendation interface converts the final coordinates into the actual operating coordinate system, assisting medical staff in accurately placing the electrodes.

[0187] The steps for scientifically planning the timing of stimulation include:

[0188] Sleep stimulation scheduling steps: Combine the patient's daily PSG data with a biological rhythm prediction model (such as sleep stage prediction based on long short-term memory LSTM network) to estimate the time of entering SWS in advance, and perform EEG monitoring 30 minutes in advance to confirm the entry into the stage and accurately trigger the intervention.

[0189] Steps for constructing the treatment frequency and cycle optimization model: Based on the patient's motor function recovery speed, Fugl-Meyer score improvement rate and PAC coupling strength change trend, the Bayesian optimization algorithm is used to dynamically recommend the treatment frequency (3-5 times per week) and cycle length (4-8 weeks).

[0190] Automatic treatment interruption detection procedure: When EEG interruption or non-target rhythm dominance is detected during monitoring, stimulation is stopped and subsequent plans are adjusted.

[0191] The multimodal joint evaluation and dynamic monitoring steps include:

[0192] Multimodal magnetic resonance imaging (MRI) depth assessment steps: Before treatment, during treatment (every 2 weeks), and after treatment, comprehensively analyze the brain region connectivity and network plasticity evolution;

[0193] In vivo electrophysiological real-time monitoring steps: During tACS intervention, neural oscillation activity of the cortical hippocampal network is monitored in real time using high-density electroencephalography (EEG) data, and the intervention and modulation effects are analyzed.

[0194] The multimodal magnetic resonance depth assessment step specifically includes:

[0195] Brain region connectivity pattern recognition steps: Using diffusion tensor imaging (DTI) data, construct a structural connectivity map of the cortical hippocampal network using the shortest path algorithm (Dijkstra's algorithm) and trajectory tracking algorithm (FACT or iFOD2) in graph theory; combine the Pearson correlation coefficient matrix extracted from functional connectivity magnetic resonance imaging (fcMRI) to generate a functional connectivity intensity map.

[0196] Steps for calculating brain network topology attributes: Using graph neural network analysis methods or traditional graph theory indicators (such as node degree, local clustering coefficient, path length, modularity, etc.), monitor the network topology changes of the central hub regions of the motor memory network (such as the premotor area and parahippocampal gyrus) at different treatment stages after stroke, and assess the development trend of neural plasticity.

[0197] Steps for determining the trend of plasticity evolution: By comparing the dynamic attributes of network graphs at multiple time points (such as small-world index and global efficiency), dynamic factor analysis (DFA) or principal component rate of change analysis (PCVA) is used to determine whether the degree of synergistic activation between brain regions is enhanced, thereby quantifying the rehabilitation effect.

[0198] The in vivo electrophysiological real-time monitoring steps specifically include:

[0199] Slow wave rhythm identification steps: Extract the activity in the 0.5–4 Hz frequency band from the EEG signal in real time, and use wavelet packet decomposition (WPD) or multiscale entropy (MSE) to identify S0 rhythm features and evaluate its stability and amplitude fluctuations.

[0200] S0-SWR cross-frequency coupling detection steps: Using the phase-amplitude coupling (PAC) calculation method, based on classic algorithms such as Tort or Canolty, the modulation index (MI) between the phase information of the S0 band and the amplitude envelope of the SWR high frequency (~80–120Hz) is extracted, and the coupling strength is quantified in real time.

[0201] The steps for modeling the trend of intervention effects are as follows: Time series regression analysis is performed on the MI value and behavioral indicators (such as Fugl-Meyer scores) before and after each intervention. Multiple linear regression or random forest regression models are used to evaluate the correlation between oscillatory synchronous changes and functional recovery, providing data basis for real-time optimization of intervention programs.

[0202] The stroke rehabilitation method further includes a periodic clinical functional assessment step: weekly assessments of motor function (using the Fugl-Meyer assessment and Wolf motor function test), cognitive function (using the MoCA), and activities of daily living (using the Barthel Index) are conducted to dynamically monitor the patient's motor learning and memory abilities and the recovery progress of motor function, providing timely feedback on treatment effectiveness. The periodic clinical functional assessment step includes:

[0203] A multidimensional clinical scoring interface was set up to collect patients' motor function scores (using the Fugl-Meyer Assessment Scale and Wolf Motor Function Test), cognitive function scores (using the MoCA Scale), and activities of daily living scores (using the Barthel Index) on a weekly basis.

[0204] Embedded data fusion steps: The embedded data fusion algorithm adopts a multi-index scoring fusion algorithm based on weighted feature aggregation, normalizes the collected scale data, and extracts core dimension indicators through principal component analysis (PCA) or factor analysis to form the Composite Rehabilitation Progress Index (CRPI).

[0205] Dynamic trend modeling and anomaly detection steps: Using dynamic trend modeling and anomaly detection algorithms, based on linear regression models, locally weighted regression (LOWESS), or time series prediction models (such as ARIMA, LSTM), the trends of patient rehabilitation indicators are modeled to determine the rate of change and inflection point of rehabilitation progress, and to help identify fluctuations and stagnation risks in treatment efficacy.

[0206] Adaptive feedback adjustment steps: Link rehabilitation assessment results with neuromodulation parameters, and achieve individualized program adjustment and closed-loop optimization based on a set rule base (such as extending tACS intervention time or adjusting S0 rhythm frequency when rehabilitation progress is slow).

[0207] This invention also discloses an electronic device, Figure 9 This is a schematic diagram of the composition of an electronic device according to an embodiment of the present invention; please refer to [link / reference]. Figure 9 At the hardware level, the electronic device includes a memory, a processor, and at least one communication interface; the processor may be a microprocessor, and the memory may include main memory, such as random access memory (RAM) or non-volatile memory. Of course, the electronic device may also include other hardware as needed.

[0208] The processor, communication interface, and memory can be interconnected via an internal bus. The memory stores programs (including operating system programs and application programs); the programs may include program code, which may include computer operation instructions. The memory may include main memory and non-volatile memory, and provides instructions and data to the processor.

[0209] In one embodiment, the processor can read the corresponding program from non-volatile memory into memory and then run it; the processor can execute the program stored in memory and specifically perform the following operations (e.g. Figure 8 As shown):

[0210]

Step S1

[0211]

Step S2

[0212]

Step S3

[0213]

Step S4

[0214] This invention further discloses a storage medium storing computer program instructions, which, when executed by a processor, implement the following steps of the method of this invention (e.g. Figure 8 As shown):

[0215]

Step S1

[0216]

Step S2

[0217]

Step S3

[0218]

Step S4

[0219] In summary, the stroke motor rehabilitation system, method, electronic device, and storage medium based on cortical hippocampal network neural oscillation modulation proposed in this invention can help promote the recovery of motor function.

[0220] In one application scenario of the present invention, the stroke motor rehabilitation system and method based on cortical hippocampal network neural oscillation modulation has the following advantages:

[0221] Innovative Rehabilitation Theories and Mechanisms: Breaking through the limitations of traditional peripheral compensation models, this research pioneers a rehabilitation strategy based on a multi-scale theoretical framework of "oscillation-plasticity-network." By targeting and modulating the S0-SWR neural oscillations of the cortical hippocampal network, it effectively promotes the directional transfer and enhancement of motor memory, providing new theoretical support and technical pathways for stroke rehabilitation and leading technological innovation in the field of rehabilitation.

[0222] Achieving precise and personalized treatment: Utilizing multimodal neuroimaging technology, precise localization of stimulation sites is achieved, and tACS stimulation parameters are flexibly set according to individual patient differences, truly realizing a precise and personalized treatment model of "one policy for one person." This significantly improves treatment efficacy while greatly reducing the risk of adverse reactions, enhancing the patient's treatment experience and safety.

[0223] Multi-dimensional and comprehensive assessment and monitoring: Integrating multiple methods such as multimodal magnetic resonance imaging, in vivo electrophysiology, and clinical functional assessment, this approach comprehensively and dynamically monitors treatment effects and changes in neural plasticity from multiple dimensions, including macroscopic brain network structure, mesoscopic neural oscillation mechanisms, and actual clinical function. This provides rich and accurate scientific evidence for real-time optimization and adjustment of treatment plans, powerfully promoting the rapid translation of this technology from laboratory research to widespread clinical application.

[0224] Significant clinical and social value: By effectively regulating the neural mechanisms of sleep-dependent motor memory consolidation, it can significantly improve the motor learning and memory abilities of stroke patients, accelerate the recovery of motor function, and greatly improve patients' self-care ability and quality of life. At the same time, it reduces the burden of long-term care on families and the pressure on social medical resources, possessing immeasurable scientific value and far-reaching social significance.

[0225] It should be noted that this application can be implemented in software and / or a combination of software and hardware; for example, it can be implemented using an application-specific integrated circuit (ASIC), a general-purpose computer, or any other similar hardware device. In some embodiments, the software program of this application can be executed by a processor to implement the steps or functions described above. Similarly, the software program of this application (including related data structures) can be stored in a computer-readable recording medium; for example, RAM memory, magnetic or optical drives, floppy disks, and similar devices. In addition, some steps or functions of this application can be implemented in hardware; for example, as circuitry that cooperates with a processor to perform the various steps or functions.

[0226] 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.

[0227] The description and application of the present invention herein are illustrative and not intended to limit the scope of the invention to the embodiments described above. Effects or advantages involved in the embodiments may not be apparent due to various factors, and the description of effects or advantages is not intended to limit the embodiments. Variations and modifications of the embodiments disclosed herein are possible, and various substitutions and equivalents of the components in the embodiments are well known to those skilled in the art. It should be apparent to those skilled in the art that the invention can be implemented in other forms, structures, arrangements, proportions, and with other components, materials, and parts without departing from the spirit or essential characteristics of the invention. Other variations and modifications can be made to the embodiments disclosed herein without departing from the scope and spirit of the invention.

Claims

1. A stroke motor rehabilitation system based on cortical-hippocampal network neural oscillation modulation, characterized in that, The sports rehabilitation system includes: The comprehensive assessment module for stroke patients is used to assess stroke patients. The individualized model building module is used to create customized individualized models for patients. Transcranial alternating current stimulation (TAC) precision intervention module, used for precise intervention of transcranial alternating current stimulation; The multimodal joint assessment and dynamic monitoring module is used to perform multimodal joint assessment and dynamic monitoring.

2. The stroke motor rehabilitation system based on cortical hippocampal network neural oscillation modulation according to claim 1, characterized in that: The comprehensive assessment module for stroke patients is used to conduct a comprehensive and systematic assessment of stroke patients, covering: (1) detailed medical history collection; (2) neurological examination; (3) motor function assessment, using the Fugl-Meyer assessment scale combined with a motor performance quantification system; (4) scientific assessment of cognitive function, combining the Montreal Cognitive Assessment Scale and computerized neuropsychological testing; (5) comprehensive assessment of sleep quality, integrating the Pittsburgh Sleep Quality Index (PSQI) and polysomnography (PSG) data; all data are input into the assessment algorithm, which uses a multi-factor weighted decision algorithm for quantitative scoring and constructs a multi-dimensional patient status profile; The individualized model building module selects patients suitable for treatment plans based on the severity of their condition, their physical function, and MRI and CT imaging results, and tailors individualized clinical models for them, laying the foundation for subsequent precision treatment. The individualized model construction module specifically includes: The feature extraction unit is used to extract key indicators based on the structural images, functional images and EEG data obtained by the comprehensive evaluation module, using feature selection algorithms in machine learning, including: cortical-hippocampal network functional connectivity strength, slow wave power spectral density, cross-frequency coupling index, and Fugl-Meyer score change rate. The adaptive screening unit combines the extracted indicators and uses support vector machine (SVM) or random forest classification algorithms to classify and predict whether patients are suitable for S0-tACS treatment, and outputs "recommended / not recommended" labels. The individual parameter modeling unit is used to group recommended patients based on their feature inputs using K-means clustering or hierarchical clustering algorithms, extract the optimal tACS stimulation parameters and placement positions in each group, and finally construct an individualized treatment parameter matrix model based on this. The model visualization feedback interface unit is used to output individual treatment models through the system interface, including patient neurological state atlases, recommended stimulation parameters, expected efficacy scores, and intervention risk assessment results, for clinicians' decision-making reference.

3. The stroke motor rehabilitation system based on cortical hippocampal network neural oscillation modulation according to claim 1, characterized in that: The transcranial alternating current stimulation precision intervention module includes: The stimulation parameter optimization setting unit is used to dynamically identify key intervention windows in slow-wave sleep (SWS) based on the neurophysiological feature extraction algorithm, and automatically set individualized tACS stimulation parameters. The stimulation site precise positioning unit is used to fuse multimodal image information and automatically recommend the tACS electrode placement position to achieve precise targeting; The Stimulation Timing Scientific Planning Unit is used to develop individualized treatment schedules and dynamically adjust the intervention process; The stimulation parameter optimization setting unit specifically includes: The time window identification unit is used to perform real-time analysis of EEG signals using short-time Fourier transform (STFT) or wavelet transform, extract the energy trend in the 0.5–4Hz frequency band, and identify the S0 peak period by combining the slow wave rhythm amplitude threshold and event detection window as the intervention starting point. The individual tolerance adaptive regulation unit is used to automatically adjust the stimulus intensity using an incremental load trial method based on baseline stimulus response assessment, within a range not exceeding 2 mA / cm². 2 Maximize neural plasticity within a safe threshold. The frequency matching optimization model building unit is used to perform autocorrelation analysis on the patient's EEG data, extract its individualized S0 dominant frequency, and lock the tACS frequency within the range of ±0.2Hz of the dominant frequency to achieve resonant coupling of the endogenous rhythm. The waveform generation unit is used to generate a phase-continuous stimulation waveform based on the set parameters using a sine wave function generator. The precise positioning unit for the stimulation site specifically includes: The brain region target identification unit is used to combine fMRI functional connectivity analysis with DTI reconstructed white matter fiber trajectories to identify key hubs in the S0–SWR pathway as candidate targets. The electrode positioning optimization model construction unit is used to register the standard head model with the patient's individual MRI image, and combine the international 10-20 electrode system and nerve electric field simulation to output the optimal electrode combination and spatial coordinates. The recommended interface for electrode patches is used to convert electrode position information into actual operating coordinates to assist in precise clinical electrode placement. The stimulation timing scientific planning unit specifically includes: The sleep stimulation scheduling unit is used to estimate the entry into the SWS period in advance based on the patient's daily PSG data and biorhythm prediction model, and to initiate EEG monitoring to confirm the stage status, so as to achieve precise intervention triggering. The treatment frequency and cycle optimization model building unit is used to combine the patient's motor function recovery rate, Fugl-Meyer score improvement rate and PAC coupling strength change trend, and dynamically recommend treatment frequency and cycle using Bayesian optimization algorithm; The treatment interruption automatic detection unit is used to automatically stop stimulation and adjust the subsequent plan when an EEG signal interruption or non-target rhythm is detected.

4. The stroke motor rehabilitation system based on cortical hippocampal network neural oscillation modulation according to claim 1, characterized in that: The multimodal joint evaluation and dynamic monitoring module includes: The multimodal magnetic resonance imaging depth assessment unit is used to comprehensively assess the connectivity status and network plasticity evolution of brain regions before, during, and after treatment. The in vivo electrophysiological real-time monitoring unit is used to monitor cortico-hippocampal network neural oscillation activity in real time during tACS intervention using high-density EEG to assess the intervention effect. The multimodal magnetic resonance depth assessment unit includes: The brain region connectivity pattern recognition unit is used to construct a structural connectivity map based on DTI data using the shortest path and trajectory tracking algorithm; combined with the Pearson correlation coefficient matrix extracted from functional magnetic resonance imaging, it generates a functional connectivity strength map. The brain network topology attribute calculation unit is used to monitor the topological evolution of central nodes in the cortical-hippocampal motor memory network at different treatment stages using graph neural network analysis or traditional graph theory indicators. The plasticity evolution trend determination unit is used to quantify the evolution of synergistic activation level and evaluate neural plasticity and rehabilitation effect through dynamic factor analysis or principal component rate of change analysis of multi-time point network diagrams. The in vivo electrophysiological real-time monitoring unit includes: The slow wave rhythm identification unit is used to extract EEG activity in the 0.5–4 Hz frequency band in real time. It uses wavelet packet decomposition or multi-scale entropy method to identify S0 rhythm features and evaluate their stability and amplitude fluctuations. The S0–SWR cross-frequency coupling detection unit is used to quantify the coupling strength between the S0 band phase and the SWR high-frequency amplitude envelope using the PAC calculation method. The intervention effect trend modeling unit is used to perform time series modeling of the coupling strength MI value before and after the intervention with behavioral assessment data. It uses multiple linear regression or random forest regression analysis to restore the association with neural synchronization and provide a basis for treatment optimization.

5. The stroke motor rehabilitation system based on cortical-hippocampal network neural oscillation modulation according to claim 1, characterized in that: The stroke rehabilitation system further includes: a periodic clinical function assessment module, which is used to periodically assess the patient's motor function, cognitive function, and activities of daily living on a weekly basis, dynamically monitor the patient's motor learning and memory abilities and the recovery progress of motor function, and provide timely feedback on treatment effectiveness; the periodic clinical function assessment module includes: A multidimensional clinical score collection interface was set up to collect patients' motor function scores, cognitive function scores, and daily living activities scores on a weekly basis. The embedded data fusion algorithm unit is used to normalize the collected scale data by employing a multi-index scoring fusion algorithm based on weighted feature aggregation, and extract core dimension indicators through principal component analysis or factor analysis to form a comprehensive rehabilitation progress index. The dynamic trend modeling and anomaly detection algorithm unit is used to model the trend of patient rehabilitation indicators based on linear regression model, local weighted regression or time series prediction model, determine the rate of change and inflection point of rehabilitation progress, and help identify the fluctuation and stagnation risk of efficacy. The adaptive feedback adjustment unit links rehabilitation assessment results with neuromodulation parameters and, based on a set rule base, achieves individualized plan adjustments and closed-loop optimization.

6. A method for motor rehabilitation of stroke based on cortical-hippocampal network neural oscillation modulation, characterized in that, The sports rehabilitation methods include: Comprehensive assessment steps for stroke patients: Stroke patients are assessed using the comprehensive assessment module for stroke patients; Personalized model building steps: Use the personalized model building module to customize a personalized model for the specified patient; Precise intervention steps for transcranial alternating current stimulation: Precise intervention using the transcranial alternating current stimulation precision intervention module; Multimodal joint assessment and dynamic monitoring steps: Use the multimodal joint assessment and dynamic monitoring module to perform multimodal joint assessment and dynamic monitoring.

7. The stroke motor rehabilitation method based on cortical-hippocampal network neural oscillation modulation according to claim 6, characterized in that: The comprehensive assessment steps for stroke patients include a comprehensive and systematic assessment covering: (1) detailed medical history collection; (2) neurological examination; (3) motor function assessment using the Fugl-Meyer Assessment Scale (FMA) combined with a motor performance quantification system; (4) scientific assessment of cognitive function using the Montreal Cognitive Assessment Scale and computerized neuropsychological testing; and (5) comprehensive assessment of sleep quality using the Pittsburgh Sleep Quality Index and polysomnography data. All data are input into the assessment algorithm, which uses a multi-factor weighted decision algorithm for quantitative scoring and constructs a multi-dimensional patient status profile. The individualized model building module selects patients suitable for treatment plans based on the severity of their condition, their physical function, and MRI and CT imaging results, and tailors individualized clinical models for them, laying the foundation for subsequent precision treatment. The individualized model construction steps specifically include: Feature extraction steps: Based on the structural images, functional images and EEG data obtained by the comprehensive evaluation module, key indicators are extracted using feature selection algorithms in machine learning, including: cortical-hippocampal network functional connectivity strength, slow wave power spectral density, cross-frequency coupling index, and Fugl-Meyer score change rate. Adaptive screening steps: Combining the extracted indicators, support vector machine or random forest classification algorithms are used to classify and predict whether patients are suitable for S0-tACS treatment, and output "recommended / not recommended" labels; Individual parameter modeling steps: For the recommended patients, K-means clustering or hierarchical clustering algorithms are used to group the patients based on their feature inputs, and the optimal tACS stimulation parameters and placement positions in each group are extracted; finally, an individualized treatment parameter matrix model is constructed based on this. Model visualization feedback steps: The system interface outputs an individual treatment model, including the patient's neurological state atlas, recommended stimulation parameters, expected efficacy score, and intervention risk assessment results, for clinicians' decision-making reference; The precise intervention steps of transcranial alternating current stimulation include: Stimulation parameter optimization setting steps: Based on the neurophysiological feature extraction algorithm, the key intervention window of SWS during slow wave sleep is dynamically identified, and individualized tACS stimulation parameters are automatically set; Precise stimulation site localization steps: By fusing multimodal imaging information, the placement of tACS electrodes is automatically recommended to achieve precise targeting; The scientific planning steps for stimulation timing include: developing an individualized treatment schedule and dynamically adjusting the intervention process; The stimulation parameter optimization setting step specifically includes: Time window identification steps: Short-time Fourier transform (STFT) or wavelet transform is used to analyze the EEG signal in real time, extract the energy trend of the 0.5–4 Hz frequency band, and combine the slow wave rhythm amplitude threshold and event detection window to identify the S0 peak period as the intervention starting point; Individual tolerance adaptive regulation steps: Based on baseline stimulus response assessment, the stimulus intensity is automatically adjusted using an incremental load trial-and-error method, not exceeding 2 mA / cm². 2 Maximize neural plasticity within a safe threshold. Frequency matching optimization model construction steps: perform autocorrelation analysis on patient EEG data, extract their individualized S0 dominant frequency, and lock the tACS frequency within the range of ±0.2Hz of the dominant frequency to achieve resonant coupling of endogenous rhythm; waveform generation steps: generate a phase-continuous stimulation waveform according to the set parameters using a sine wave function generator; The precise positioning unit for the stimulation site specifically includes: Brain region target identification steps: Combine fMRI functional connectivity analysis and DTI reconstructed white matter fiber trajectories to identify key hubs in the S0–SWR pathway as candidate targets. Electrode localization optimization model construction steps: Based on the standard head model and the patient's individual MRI image registration, combined with the international 10-20 electrode system and nerve electric field simulation, the optimal electrode combination and spatial coordinates are output; Recommended interface steps for electrode patching: Convert electrode position information into actual operational coordinates to assist in precise clinical electrode placement; The stimulation timing scientific planning unit specifically includes: Sleep stimulation scheduling steps: Based on the patient's daily PSG data and biorhythm prediction model, the entry into the SWS period is estimated in advance, and EEG monitoring is initiated to confirm the stage status, so as to achieve precise intervention triggering; Steps for constructing the treatment frequency and cycle optimization model: Combining the patient's motor function recovery rate, Fugl-Meyer score improvement rate and PAC coupling strength change trend, a Bayesian optimization algorithm is used to dynamically recommend the treatment frequency and cycle; Automatic treatment interruption detection procedure: When an interruption in the EEG signal or a non-target rhythm is detected, stimulation is automatically stopped and subsequent plans are adjusted.

8. The stroke motor rehabilitation method based on cortical-hippocampal network neural oscillation modulation according to claim 6, characterized in that: The multimodal joint evaluation and dynamic monitoring steps include: Multimodal magnetic resonance imaging (MRI) depth assessment steps: Comprehensive assessment of brain inter-brain connectivity and network plasticity evolution before, during, and after treatment; In vivo electrophysiological real-time monitoring steps: During tACS intervention, high-density EEG is used to monitor cortical-hippocampal network neural oscillation activity in real time to assess the intervention effect; The multimodal magnetic resonance depth assessment steps include: Brain region connectivity pattern recognition steps: Based on DTI data, construct a structural connectivity map using the shortest path and trajectory tracking algorithm; combine the Pearson correlation coefficient matrix extracted from functional magnetic resonance imaging to generate a functional connectivity strength map; Steps for calculating brain network topological properties: Use graph neural network analysis or traditional graph theory indicators to monitor the topological evolution of central nodes in the cortical-hippocampal motor memory network at different treatment stages; Steps for determining the trend of plasticity evolution: Quantify the evolution of synergistic activation level and evaluate neural plasticity and rehabilitation effect by using dynamic factor analysis or principal component rate of change analysis of multi-time point network diagrams. In vivo electrophysiological real-time monitoring steps include: Slow wave rhythm identification steps: extract EEG activity in the 0.5–4 Hz frequency band in real time, identify S0 rhythm features using wavelet packet decomposition or multi-scale entropy method, and evaluate its stability and amplitude fluctuations; S0–SWR cross-frequency coupling detection steps: The coupling strength between the S0 band phase and the SWR high-frequency amplitude envelope is quantified using the PAC calculation method; Intervention effect trend modeling steps: The coupling strength MI value before and after the intervention is modeled with behavioral assessment data in time series. Multiple linear regression or random forest regression analysis is used to restore the association with neural synchronization, providing a basis for treatment optimization. The stroke rehabilitation method further includes a periodic clinical function assessment step: weekly assessments of motor function, cognitive function, and activities of daily living (ADL) are conducted to dynamically monitor the recovery progress of the patient's motor learning and memory abilities and motor function, providing timely feedback on treatment effectiveness. This periodic clinical function assessment step is implemented through a periodic clinical function assessment module, which includes: A multidimensional clinical score collection interface was set up to collect patients' motor function scores, cognitive function scores, and daily living activities scores on a weekly basis. The embedded data fusion algorithm unit is used to normalize the collected scale data by employing a multi-index scoring fusion algorithm based on weighted feature aggregation, and extract core dimension indicators through principal component analysis (PCA) or factor analysis to form a comprehensive rehabilitation progress index. The dynamic trend modeling and anomaly detection algorithm unit is used to model the trend of patient rehabilitation indicators based on linear regression model, local weighted regression or time series prediction model, determine the rate of change and inflection point of rehabilitation progress, and help identify the fluctuation and stagnation risk of efficacy. The adaptive feedback adjustment unit links rehabilitation assessment results with neuromodulation parameters and, based on a set rule base, achieves individualized plan adjustments and closed-loop optimization.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 6 to 8.

10. A storage medium storing computer program instructions thereon, characterized in that, When executed by a processor, the computer program instructions implement the steps of the method according to any one of claims 6 to 8.

Images