Multi-modal fracture closed-loop rehabilitation method and system for neural remodeling

Abstract

The application discloses a kind of multi-modal fracture closed-loop rehabilitation methods and systems for nerve remodeling, comprising: collecting the multi-modal data when executing rehabilitation gait action under the action of phase-specific electrical stimulation, at least including electroencephalogram signal, near-infrared blood oxygen signal, electromyogram signal, joint angular velocity signal and foot sole pressure signal;After the multi-modal physiological parameter data is preprocessed, fusion analysis is carried out, specifically, the motion intention and concentration state evaluation score and the first partial current are estimated, the action category and completion degree score are estimated, the second partial current is output based on the phase-specific site scheme, the total current is set with the action category and completion degree score as weight and combined with the first partial current and the second partial current, and the specific treatment electrode site and electrical stimulation treatment parameter are comprehensively output and fed back to control corresponding electrode to apply phase-specific distributed electrical stimulation. Thus multi-modal information fusion evaluation and on-demand auxiliary regulation can be realized, nerve remodeling is effectively promoted and rehabilitation training effect is improved.

Description

Technical Field

[0001] This invention belongs to the field of rehabilitation medicine technology, specifically relating to a multimodal closed-loop rehabilitation method and system for fractures oriented towards neural remodeling. Background Technology

[0002] Fractures and postoperative functional impairments are among the most common and prevalent types of motor function injuries in clinical rehabilitation medicine, especially in patients with lower limb fractures, often accompanied by gait abnormalities, muscle weakness, and neurological control imbalances. Numerous studies have shown that post-fracture motor function recovery depends not only on the bone healing process but also heavily on the plastic remodeling of motor control pathways by the central nervous system, i.e., neural remodeling. The degree of neural remodeling directly affects the quality of gait reconstruction, rehabilitation efficiency, and long-term functional prognosis.

[0003] Existing fracture rehabilitation methods mainly include passive joint mobility training, walking aids or exoskeleton-assisted training, and experience-based electrical stimulation therapy. For example, the invention patent application with publication number CN113905781A discloses a digital twin neurorehabilitation system. These methods can improve patients' motor abilities to a certain extent, but they still have certain defects: (1) The system as a whole is still mainly driven by mechanical or fixed parameters, making it difficult to achieve a continuous closed-loop operation of "assessment-decision-intervention-reassessment" in real rehabilitation training. It lacks a synchronous perception and dynamic adjustment mechanism for the patient's central nervous system status, motor intention, and peripheral execution, making it difficult to conduct refined intervention for the neural remodeling process. (2) In terms of rehabilitation control strategies, existing systems mostly adopt uniform stimulation parameters or simple staged stimulation methods. The stimulation points and intensity lack a precise correspondence with specific gait stages and key muscle group functions, making it difficult to achieve phase-specific neuromuscular regulation. This coarse-grained stimulation method is prone to introducing unnecessary muscle synergy interference in complex gait rehabilitation scenarios, affecting the pertinence and safety of rehabilitation training. (3) Existing technologies are mostly focused on parallel acquisition of single-modality or simple multimodality. However, there are many differences in patient information. For example, patients need to concentrate more when performing actions. Therefore, changes in near-infrared blood oxygenation signals are strongly correlated with muscle movements. At present, there is a lack of in-depth fusion analysis of multimodal information such as EEG, near-infrared blood oxygenation signals, electromyography, joint kinematics and plantar pressure in a unified spatiotemporal framework. It is difficult to fully depict the coupling relationship between central intention, neural compensation mechanism and peripheral motor execution. Therefore, the robustness of the system is poor.

[0004] Therefore, there is an urgent need for a closed-loop rehabilitation method and system that can simultaneously collect and integrate multimodal physiological and motor information during fracture rehabilitation training, accurately identify gait movements through electromyography, pressure, and accelerometers, accurately identify changes in neural state through electroencephalography and near-infrared spectroscopy, and on this basis, realize on-demand assistance and phase-specific electrical stimulation, so as to more effectively promote the neural remodeling process and improve the quality of motor function recovery in fracture patients. Summary of the Invention

[0005] In view of the above, the purpose of this invention is to provide a multimodal fracture closed-loop rehabilitation method and system for neural remodeling. Based on the analysis of multimodal data from real-time brain-computer interaction and the real-time feedback closed-loop electrical stimulation modulation with different weights, the invention achieves synchronization of central motor commands and peripheral dynamic responses at the physical level, thereby promoting functional recovery based on the Hebbian plasticity principle.

[0006] To achieve the above-mentioned objectives, an embodiment provides a multimodal closed-loop rehabilitation system for fractures oriented towards neural remodeling, comprising: A multimodal closed-loop rehabilitation equipment system is used to collect multimodal data when performing rehabilitation gait movements under phase-specific electrical stimulation, including at least electroencephalogram (EEG) signals, near-infrared blood oxygenation signals, electromyography (EMG) signals, joint angular velocity signals, and foot pressure signals. The software algorithm module is used to preprocess and fuse multimodal physiological parameter data. Specifically, it estimates the motor intention and attention state assessment scores based on EEG signals, near-infrared blood oxygenation signals, and joint angular velocity signals. It estimates the first component current based on the motor intention and attention state assessment scores, and estimates the movement category and completion score based on EMG signals, joint angular velocity signals, and foot pressure signals. It outputs the second component current based on a phase-specific site scheme. The total current is set by using the movement category and completion score as weights and combining the first and second component currents. It also outputs the specific treatment electrode sites and electrical stimulation parameters corresponding to the current rehabilitation gait movement and feeds them back to the multimodal closed-loop rehabilitation system equipment to control the application of phase-specific electrical stimulation to the corresponding electrodes.

[0007] Preferably, the step of estimating motor intention and attention state assessment scores based on electroencephalogram (EEG) signals, near-infrared blood oxygenation signals, and joint angular velocity signals includes: S attn =K IMU ×(1-X IMU )+K EEG ×X EEG +K fNIRS ×X fNIRS Among them, K IMU K EEG K fNIRS X represents the weights of each feature.IMU X EEG and X fNIRS These are the normalized characteristics of joint angular velocity signals, electroencephalogram (EEG) signals, and near-infrared blood oxygenation signals, respectively, ranging from 0 to 1; S attn The assessment scores the participants’ intention to move and their level of focus, ranging from 0 to 1.

[0008] Preferably, estimating the first fractional current based on the assessment score of motor intention and focus status includes:

[0009] in, The score indicates the assessment of movement intention and focus. The maximum current is 1.5au, where au is the threshold value ranging from 0.1 mA to 30 mA. γ determines the smoothness and steepness of the electrical stimulation, ranging from 0.5 to 2.5. This is the first component current.

[0010] Preferably, the step of estimating the movement category and completion score based on electromyographic signals, joint angular velocity signals, and foot pressure signals includes: ω=K EMG ×X EMG +K IMU2 ×X IMU2 +K PRESSURE ×X PRESSURE Among them, K EMG K IMU2 K PRESSURE These are the weights of each feature, X. EMG X IMU2 X PRESSURE These are the normalized features of electromyography (EMG), joint angular velocity (JV), and foot pressure signals, respectively, ranging from 0 to 1. ω represents the action category and completion score, with a value between 0 and 1.

[0011] Preferably, the second current output based on the specific site scheme includes: The rehabilitation gait movements are cyclical movements, including the propulsion phase, the swing phase, and the weight-bearing support phase. The propulsion phase includes the pre-push-off phase and the push-off phase, while the weight-bearing support phase includes double support and single-leg support. The system implements on-demand auxiliary logic, and the stimulation current exhibits a phase-specific distribution. After determining the gait classification, the gastrocnemius channel is given priority for explosive stimulation during the propulsion phase, with a current intensity I2 ranging from 0.5 to 1.5 au; the tibialis anterior channel is mainly stimulated during the swing phase, with a current intensity I2 ranging from 0.6 to 0.8 au; and the quadriceps channel is mainly stimulated during the weight-bearing support phase, with a current intensity I2 ranging from 0.5 to 1.2 au, where au is a preset threshold ranging from 0.1 mA to 30 mA.

[0012] Preferably, the total current is set by using the action category and completion score as weights, combined with the first and second sub-currents, including: I = (1-ω)×I1 + ω×I2 Where ω represents the action category and completion score, I1 represents the first current component, I2 represents the second current component, and I represents the total current.

[0013] Preferably, the electrical stimulation therapy parameters include the amplitude, frequency, and duty cycle of the current, wherein the amplitude is determined by the total current, the frequency is adjustable from 20Hz to 40Hz, and the duty cycle is adjustable from 10% to 60%, which are set during initialization.

[0014] Preferably, the multimodal physiological parameter data are preprocessed, including: Multimodal data is segmented according to a preset time window, and noise removal and filtering preprocessing operations are performed on the segmented data.

[0015] To achieve the above-mentioned objectives, embodiments of the present invention also provide a multimodal closed-loop rehabilitation method for fractures oriented towards neural remodeling. The method employs the aforementioned system and includes the following steps: Step 1: Collect multimodal data of the patient performing rehabilitation gait movements under phase-specific electrical stimulation through the multimodal closed-loop rehabilitation equipment system and transmit it to the software algorithm module. The multimodal data includes at least EEG signals, near-infrared blood oxygenation signals, electromyography signals, joint angular velocity signals, and foot pressure signals. Step 2: The software algorithm module preprocesses the multimodal physiological parameter data and then performs fusion analysis. Specifically, it estimates the motor intention and attention state assessment scores based on EEG signals, near-infrared blood oxygenation signals, and joint angular velocity signals. It estimates the first component current based on the motor intention and attention state assessment scores, and estimates the movement category and completion score based on EMG signals, joint angular velocity signals, and foot pressure signals. It outputs the second component current based on the phase-specific site scheme. The total current is set by using the movement category and completion score as weights and combining the first and second component currents. The system then outputs the specific treatment electrode sites and electrical stimulation treatment parameters corresponding to the current rehabilitation gait movement and feeds them back to the multimodal closed-loop rehabilitation equipment system. Step 3: The multimodal closed-loop rehabilitation equipment system applies phase-specific electrical stimulation to the corresponding electrodes at the corresponding gait stage based on the specific treatment electrode sites and electrical stimulation treatment parameters. Repeat steps 1-3 to form a closed-loop rehabilitation training process.

[0016] Compared with the prior art, the beneficial effects of the present invention include at least the following: This invention achieves a synergistic assessment of a patient's central motor intention, attention state, peripheral muscle activation, and gait execution by simultaneously collecting and fusing electroencephalogram (EEG), near-infrared blood oxygenation, electromyography (EMG), joint angular velocity, and plantar pressure signals within a unified spatiotemporal framework. This comprehensively characterizes the central-peripheral neural control state during fracture rehabilitation, overcoming the problem in existing technologies where single physiological signals or mechanical parameters are insufficient to reflect the neural remodeling process.

[0017] Based on the results of multimodal fusion, a closed-loop rehabilitation control mechanism is constructed, which combines motor intention assessment, motor completion score and adaptive adjustment of electrical stimulation parameters. This allows electrical stimulation to intervene only when central motor commands are detected but peripheral execution is insufficient, achieving real-time closed-loop operation of "assessment-decision-intervention-reassessment". This effectively avoids over-assistance and promotes neural plasticity remodeling and functional recovery.

[0018] By introducing a phase-specific electrical stimulation strategy oriented towards the gait cycle, precise timing and intensity matching stimulation control is implemented for key muscle groups at different gait stages. This highly couples neural stimulation with actual motor function needs, improving the pertinence, safety, and efficiency of rehabilitation training, and is beneficial for improving the quality of gait reconstruction and long-term motor function prognosis in fracture patients. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the structure of the multimodal fracture closed-loop rehabilitation system for neural remodeling provided in the embodiment; Figure 2 This is a physical image of the multimodal closed-loop rehabilitation equipment system provided in the embodiment; Figure 3 This is a physical image of the acquisition / stimulation multiplexing electrode provided in the embodiment; Figure 4 This is a schematic diagram of gait movements provided in the embodiment; Figure 5This is a multimodal data acquisition diagram provided in the embodiment; Figure 6 This is a flowchart of the multimodal fracture closed-loop rehabilitation method for neural remodeling provided in the embodiment; Figure 7 This is a comparison chart of the difference in thigh circumference between the affected and healthy sides after training, provided in the embodiment. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not limit the scope of protection of this invention.

[0022] like Figure 1 As shown in the embodiment, the multimodal fracture closed-loop rehabilitation system for neural remodeling provided includes a multimodal closed-loop rehabilitation equipment system and a software algorithm module located on a host computer.

[0023] In the embodiments, such as Figure 2 As shown, the multimodal closed-loop rehabilitation equipment system includes at least an EEG-near-infrared blood oxygen signal acquisition module, a plantar pressure sensing module 204, an electromyography acquisition module 205, a joint angular velocity acquisition module 206, an electrical stimulation module, and a communication module 207, and completes system initialization and basic calibration.

[0024] The EEG-Near-Infrared Oxygen Signal Acquisition Module is a headband consisting of a near-infrared emitting area 201, EEG electrodes 202, and a near-infrared receiving area 203, which the patient wears on the frontal lobe. The EEG electrodes 202 are made of one of the following metals: stainless steel, copper, gold, silver, aluminum, and titanium, or an alloy of at least two metals, preferably stainless steel. The near-infrared emitting area 201 is a light-emitting diode, and the near-infrared receiving area 203 is a photodiode. The number of EEG signals is 2 to 10 channels, for example, 2 channels, and the number of near-infrared signals is 2 to 10 channels, for example, 8 channels.

[0025] The electromyography (EMG) acquisition module corresponds to an array of paired small electrodes, numbering 8 to 200; the electrical stimulation module corresponds to an array of large electrodes surrounding the EMG acquisition sites, numbering 4 to 100. These electrical stimulation points target key muscle groups from the quadriceps and hamstrings on the thigh side to the tibialis anterior and gastrocnemius on the calf side.

[0026] The electrode array is structurally designed to include a first insulating layer, a first conductive layer disposed on the first insulating layer, a second insulating layer disposed above the first conductive layer, and a second conductive layer disposed above the second insulating layer. The second insulating layer does not cover the electrode disk area, and the second conductive layer is directly connected to the first conductive layer, covering only the disk area. The first and second insulating layers are made of non-stretchable polymer materials with a thickness of 1 μm to 1 mm. The first conductive layer is made of one of copper, gold, silver, aluminum, or titanium, or an alloy of at least two metals, with a thickness of 10 nm to 2 mm. The second conductive layer is made of one of conductive gel, silver chloride gel, or conductive carbon black, or a mixture of at least two of these materials, with a thickness of 0.1 to 10 mm.

[0027] Specifically, Figure 3 Detailed diagrams of the integrated electromyography-electrical stimulation electrode provided in the embodiment are given. The first and second insulating layers are made of PET with a thickness of 100 μm; the first conductive layer is made of gold-plated copper with a thickness of 20 μm; and the second conductive layer is made of conductive gel with a thickness of 1 mm. The electromyography acquisition sites are paired small electrode arrays 308, totaling 8; the electrical stimulation sites are large electrode arrays 309 surrounding the electromyography acquisition sites, totaling 4.

[0028] The joint angular velocity acquisition modules 206 are located on the left and right sides of the thigh and calf, respectively, with a quantity of 2 to 4, for example, 4 for Weiwei.

[0029] The aforementioned multimodal closed-loop rehabilitation equipment system is used to collect multimodal data of patients performing rehabilitation gait movements under phase-specific electrical stimulation, such as... Figure 5 As shown, it includes at least five categories: electroencephalogram (EEG) signals, near-infrared blood oxygenation signals, electromyography (EMG) signals, joint angular velocity signals, and foot pressure signals. The number of sub-parameters for each category ranges from 10 to 200. Specifically, each category has 10 sub-parameters: EEG_PSD; near-infrared: ΔHbO, ΔHbT; EMG: RMS, MDF, MPF; accelerometer: ACC_X, ACC_Y, ACC_Z; and pressure. The sub-parameters are input into the software algorithm module 208 after segmentation. The software algorithm module 208 outputs the completion score of each movement in real time based on the multimodal data of the healthy side and adjusts the electrical stimulation treatment parameters according to the score.

[0030] In this embodiment, the patient alternates between wearing the multimodal closed-loop rehabilitation equipment system to perform rehabilitation gait movements. These gait movements are periodic, including a propulsion phase, a swing phase, and a weight-bearing support phase. The propulsion phase includes the pre-push-off phase and the push-off phase, while the weight-bearing support phase includes double support and single-leg support. Figure 4 As shown.

[0031] In this embodiment, the software algorithm module 208 is used to preprocess the multimodal physiological parameter data. Specifically, the received multimodal data is segmented according to a preset time window, and preprocessing operations such as noise removal and filtering are performed on the segmented data. Then, the preprocessed data is fused and analyzed, specifically including: (1) Estimate the motor intention and attention state assessment score based on EEG signal, near-infrared blood oxygen signal and joint angular velocity signal, and estimate the first current based on the motor intention and attention state assessment score.

[0032] In this embodiment, the assessment score for motor intention and attention state is estimated based on electroencephalogram (EEG) signals, near-infrared blood oxygenation signals, and joint angular velocity signals, including: S attn =K IMU ×(1-X IMU )+K EEG ×X EEG +K fNIRS ×X fNIRS Among them, K IMU =0.5, K EEG =0.25, K fNIRS =0.25 is the weight of each feature, X IMU X EEG and X fNIRS These are the normalized characteristics of joint angular velocity signals, electroencephalogram (EEG) signals, and near-infrared blood oxygenation signals, respectively, ranging from 0 to 1; S attn The assessment scores the participants’ intention to move and their level of focus, ranging from 0 to 1.

[0033] The IMU parameters are triaxial angular acceleration, X... IMU = min (max (X IMU,norm , 0), 1), , , , where N IMU The number of sampling points within a 1-second window, ranging from 100 to 10000; EEG parameters are the spectral fluctuations of alpha, beta, and gamma waves, X. EEG = min (max (X EEG,norm , 0), 1), X EEG,raw = std(PSD(f)), X EEG,max = std(PSD(8)), PSD(f) = Welch(EEG fltered, winlen = 256, noverlap = 128, f s= 250), where Welch() is a classic method for calculating power spectral density (PSD), EEG fltered, winlen is the window function length (number of window points), noverlap is the number of overlapping points between two adjacent windows, f is the data sampling frequency, and the sampling EEG frequency range is 4-8Hz; Near-infrared parameters are percentages of HbO2 concentration, X. fNIRS = min (max (X fNIRS,norm , 0), 1), , , , where N fNIRS The number of sampling points per second, ranging from 100 to 10000. This represents the percentage of HbO2 concentration at the k-th sampling point.

[0034] The first fractional current was also estimated based on the assessment scores of motor intention and focus. ,include:

[0035] in, The score indicates the assessment of movement intention and focus. The maximum current is 1.5au, where au is the threshold value, ranging from 0.1 mA to 30 mA. γ determines the smoothness or steepness of the electrical stimulation, ranging from 0.5 to 2.5.

[0036] (2) Estimate the movement category and completion score based on electromyographic signals, joint angular velocity signals and foot pressure signals, and output the second current based on the phase-specific site scheme.

[0037] The embodiment estimates the movement category and completion score based on electromyographic signals, joint angular velocity signals, and foot pressure signals, including: ω=K EMG ×X EMG +K IMU2 ×X IMU2 +K PRESSURE ×X PRESSURE Among them, K EMG =0.6, K IMU2 =0.2, K PRESSURE =0.2 represents the weights of each feature, X EMG X IMU2 X PRESSURE These are the normalized features of electromyography (EMG), joint angular velocity (JV), and foot pressure signals, respectively, ranging from 0 to 1. ω represents the action category and completion score, with a value between 0 and 1.

[0038] The EMG parameter is the RMS value, X EMG = min (max (X EMG,norm , 0), 1), , , , where N EMG The number of sampling points within a 1-second window, ranging from 500 to 160,000; Gait IMU parameters are for stability, X IMU2 = min (max (X IMU2,norm , 0), 1), , ,in, , ,in, The gait angle value at the k-th sampling point. This is the average value of 200 angle sampling points.

[0039] The pressure parameter is the pressure value, X. PRESSURE = min (max (X PRESSURE,norm , 0), 1), , , ,in Human body mass value, N PRESSURE The number of sampling points per second window, ranging from 10 to 1000. This represents the pressure value at the k-th sampling point.

[0040] The embodiment also provides a phase-specific site scheme that combines the actions at different stages, specifically including: the system implements on-demand auxiliary logic, the stimulation current exhibits a phase-specific distribution, after determining the classification of each gait, during the propulsion phase, the gastrocnemius channel is given priority for explosive stimulation, with a current intensity I2 ranging from 0.5 to 1.5 au; during the swing phase, the tibialis anterior channel is mainly stimulated, with a current intensity I2 ranging from 0.6 to 0.8 au; during the weight-bearing support phase, the quadriceps channel is mainly stimulated, with a current intensity I2 ranging from 0.5 to 1.2 au.

[0041] (3) The total current is set by using the action category and completion score as weights and combining the first and second sub-currents. The specific treatment electrode sites and electrical stimulation treatment parameters corresponding to the current rehabilitation gait are fed back to the multimodal closed-loop rehabilitation system equipment to control the corresponding electrodes to apply phase-specific electrical stimulation.

[0042] The embodiment uses the action category and completion score ω as weights to set the total current I, which includes the first component current I1 and the second component current I2. I = (1-ω)×I1 + ω×I2 Then, electrical stimulation therapy parameters are output based on the total current. These parameters include the current amplitude, frequency, and duty cycle. The amplitude is determined by the total current, the frequency is adjustable from 20Hz to 40Hz, and the duty cycle is adjustable from 10% to 60%, set during initialization. It should also be noted that the initial electrical stimulation parameters for the healthy side are adjusted based on the patient's sensory threshold, while the initial electrical stimulation parameters for the affected side are adjusted based on the action trigger threshold. Both thresholds are preset.

[0043] While outputting electrical stimulation treatment parameters, it also outputs specific treatment electrode sites corresponding to the current rehabilitation gait movement. These treatment electrode sites are determined according to the phase-specific site scheme.

[0044] like Figure 6 As shown, the embodiment also provides a multimodal closed-loop rehabilitation method for fractures oriented towards neural remodeling. The method uses the above-described system and includes the following steps: S1, the multimodal data of the patient performing rehabilitation gait movements under phase-specific electrical stimulation is collected by the multimodal closed-loop rehabilitation equipment system and transmitted to the software algorithm module. The multimodal data includes at least electroencephalogram (EEG) signals, near-infrared blood oxygenation signals, electromyography (EMG) signals, joint angular velocity signals, and foot pressure signals. S2, the software algorithm module preprocesses the multimodal physiological parameter data and then performs fusion analysis, comprehensively outputting the specific treatment electrode sites and electrical stimulation treatment parameters corresponding to the current rehabilitation gait movement and feeding them back to the multimodal closed-loop rehabilitation equipment system; S3, the multimodal closed-loop rehabilitation equipment system applies phase-specific electrical stimulation to the corresponding electrodes at the corresponding gait stage based on the specific therapeutic electrode sites and electrical stimulation parameters.

[0045] Under phase-specific electrical stimulation, patients undergo active rehabilitation gait training. During the active rehabilitation training, multimodal physiological and motor data are collected simultaneously. Steps S1-S3 are repeated to form a closed-loop rehabilitation training process.

[0046] In the specific experiment, the result was determined as follows: during the propulsion period, S attn =0.5, γ=1, ω=0.8, au =30mA, I max=45mA, I1=22.5mA, I2=30mA; therefore, for this movement, the electrical stimulation points are targeted at the key muscle groups of the gastrocnemius muscle; the initial electrical stimulation parameters are set as follows: healthy side: amplitude: 5mA, frequency: 15Hz, duty cycle: 50%; affected side: amplitude: 28.5mA, frequency: 15Hz, duty cycle: 60%. Here, the healthy side is the side with the disease, and the affected side is the side with the disease. The initial electrical stimulation parameters of the healthy side are adjusted according to the patient's sensory threshold, and the initial electrical stimulation parameters of the affected side are adjusted according to the movement trigger threshold. Figure 7 The illustration shows a comparison of the thigh circumference difference between the affected and unaffected sides after training with the system provided in the embodiment. Initially (before training), the thigh circumferences of the two patients were similar on both sides. After two months, the patient trained using the method described in the embodiment showed a significant reduction in the thigh circumference difference between the affected and unaffected sides, demonstrating the rehabilitation effect of this method.

[0047] The above method integrates real-time analysis of multimodal brain-computer interface data, sets key stimulation sites based on the real-time status of rehabilitation movements at different stages, and adjusts the intensity of electrical stimulation based on the patient's movement intentions, in order to achieve relevant neural remodeling and improve the rehabilitation effect of patients after fracture.

[0048] The specific embodiments described above illustrate the technical solution and beneficial effects of the present invention in detail. It should be understood that the above description is only the most preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, additions, and equivalent substitutions made within the scope of the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multimodal closed-loop rehabilitation system for fractures oriented towards neural remodeling, characterized in that, include: A multimodal closed-loop rehabilitation equipment system is used to collect multimodal data when performing rehabilitation gait movements under phase-specific electrical stimulation, including at least electroencephalogram (EEG) signals, near-infrared blood oxygenation signals, electromyography (EMG) signals, joint angular velocity signals, and foot pressure signals. The software algorithm module is used to preprocess and fuse multimodal physiological parameter data. Specifically, it estimates the motor intention and attention state assessment scores based on EEG signals, near-infrared blood oxygenation signals, and joint angular velocity signals. It estimates the first component current based on the motor intention and attention state assessment scores, and estimates the movement category and completion score based on EMG signals, joint angular velocity signals, and foot pressure signals. It outputs the second component current based on a phase-specific site scheme. The total current is set by using the movement category and completion score as weights and combining the first and second component currents. It also outputs the specific treatment electrode sites and electrical stimulation parameters corresponding to the current rehabilitation gait movement and feeds them back to the multimodal closed-loop rehabilitation system equipment to control the application of phase-specific electrical stimulation to the corresponding electrodes.

2. The multimodal closed-loop rehabilitation system for neural remodeling according to claim 1, characterized in that, The assessment score for estimating motor intention and attention status based on EEG signals, near-infrared blood oxygenation signals, and joint angular velocity signals includes: S attn =K IMU ×(1-X IMU )+K EEG ×X EEG +K fNIRS ×X fNIRS Among them, K IMU K EEG K fNIRS X represents the weights of each feature. IMU X EEG and X fNIRS These are the normalized characteristics of joint angular velocity signals, electroencephalogram (EEG) signals, and near-infrared blood oxygenation signals, respectively, ranging from 0 to 1; S attn The assessment scores the participants’ intention to move and their level of focus, ranging from 0 to 1.

3. The multimodal closed-loop rehabilitation system for neural remodeling according to claim 1, characterized in that, The estimation of the first current based on the assessment score of motor intention and focus status includes: in, The score indicates the assessment of movement intention and focus. The maximum current is 1.5 au, where au is the threshold value ranging from 0.1 mA to 30 mA. γ determines the smoothness and steepness of the electrical stimulation, ranging from 0.5 to 2.

5. This is the first component current.

4. The multimodal closed-loop rehabilitation system for neural remodeling according to claim 1, characterized in that, The estimation of movement category and completion score based on electromyographic signals, joint angular velocity signals, and foot pressure signals includes: ω=K EMG ×X EMG +K IMU2 ×X IMU2 +K PRESSURE ×X PRESSURE Among them, K EMG K IMU2 K PRESSURE These are the weights of each feature, X. EMG X IMU2 X PRESSURE These are the normalized features of electromyography (EMG), joint angular velocity (JV), and foot pressure signals, respectively, ranging from 0 to 1. ω represents the action category and completion score, with a value between 0 and 1.

5. The multimodal closed-loop rehabilitation system for neural remodeling according to claim 1, characterized in that, The second current branch output based on the specific site scheme includes: The rehabilitation gait movements are cyclical movements, including the propulsion phase, the swing phase, and the weight-bearing support phase. The propulsion phase includes the pre-push-off phase and the push-off phase, while the weight-bearing support phase includes double support and single-leg support. The system implements on-demand auxiliary logic, and the stimulation current exhibits a phase-specific distribution. After determining the gait classification, the gastrocnemius channel is given priority for explosive stimulation during the propulsion phase, with a current intensity I2 ranging from 0.5 to 1.5 au; the tibialis anterior channel is mainly stimulated during the swing phase, with a current intensity I2 ranging from 0.6 to 0.8 au; and the quadriceps channel is mainly stimulated during the weight-bearing support phase, with a current intensity I2 ranging from 0.5 to 1.2 au, where au is a preset threshold ranging from 0.1 mA to 30 mA.

6. The multimodal closed-loop rehabilitation system for neural remodeling according to claim 1, characterized in that, The total current is set by weighting the action category and completion score, and combining the first and second current components, including: I = (1-ω)×I1 + ω×I2 Where ω represents the action category and completion score, I1 represents the first current component, I2 represents the second current component, and I represents the total current.

7. The multimodal closed-loop rehabilitation system for neural remodeling according to claim 1, characterized in that, The electrical stimulation therapy parameters include the amplitude, frequency, and duty cycle of the current. The amplitude is determined by the total current, the frequency is adjustable from 20Hz to 40Hz, and the duty cycle is adjustable from 10% to 60%, which are set during initialization.

8. The multimodal closed-loop rehabilitation system for neural remodeling according to claim 1, characterized in that, Preprocessing of multimodal physiological parameter data includes: Multimodal data is segmented according to a preset time window, and noise removal and filtering preprocessing operations are performed on the segmented data.

9. A multimodal closed-loop rehabilitation method for fractures oriented towards neural remodeling, characterized in that, The method employs the system described in any one of claims 1-8 and includes the following steps: Step 1: Collect multimodal data of the patient performing rehabilitation gait movements under phase-specific electrical stimulation through the multimodal closed-loop rehabilitation equipment system and transmit it to the software algorithm module. The multimodal data includes at least EEG signals, near-infrared blood oxygenation signals, electromyography signals, joint angular velocity signals, and foot pressure signals. Step 2: The software algorithm module preprocesses the multimodal physiological parameter data and then performs fusion analysis. Specifically, it estimates the motor intention and attention state assessment scores based on EEG signals, near-infrared blood oxygenation signals, and joint angular velocity signals. It estimates the first component current based on the motor intention and attention state assessment scores, and estimates the movement category and completion score based on EMG signals, joint angular velocity signals, and foot pressure signals. It outputs the second component current based on the phase-specific site scheme. The total current is set by using the movement category and completion score as weights and combining the first and second component currents. The system then outputs the specific treatment electrode sites and electrical stimulation treatment parameters corresponding to the current rehabilitation gait movement and feeds them back to the multimodal closed-loop rehabilitation equipment system. Step 3: The multimodal closed-loop rehabilitation equipment system applies phase-specific electrical stimulation to the corresponding electrodes at the corresponding gait stage based on the specific treatment electrode sites and electrical stimulation treatment parameters. Repeat steps 1-3 to form a closed-loop rehabilitation training process.

Images