Cerebral small vascular disease lower limb dysfunction self-adaptive electrical stimulation auxiliary rehabilitation system

Abstract

The invention discloses a brain small vascular disease lower limb dysfunction self-adaptive electrical stimulation auxiliary rehabilitation system in the technical field of medical instruments, and the system comprises a data collection module which is used for obtaining the multi-modal data of a patient in real time; the decision regulation and control module is used for generating dynamically optimized electrical stimulation parameters and rehabilitation task instructions based on a deep reinforcement learning algorithm and patient individualized pathological characteristics according to the multi-modal data; the electrical stimulation execution module is used for synchronously regulating and stimulating nerve targets according to the optimized electrical stimulation parameters; and the interactive virtual reality module is used for combining immersive gait training and cognitive training. By detecting the multi-modal physiological data of the patient, combining the deep reinforcement learning model to dynamically optimize the electrical stimulation parameters, and coupling VR exercise training and cognitive training, the three-in-one collaborative rehabilitation mode of electrical stimulation-physical treatment-cognitive intervention is realized, the rehabilitation training efficiency is improved, the relapse risk of the patient is reduced, and the rehabilitation training experience of the patient is enhanced.

Description

Technical Field

[0001] The present invention belongs to the technical field of medical devices, and in particular is an adaptive electrical stimulation-assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease. Background Art

[0002] Cerebral small vessel disease (CSVD) is a major cause of disability in the elderly, accounting for 25%-30% of ischemic strokes and detected by MRI in over 80% of individuals over 65 years of age. Its pathological features include microinfarctions, white matter lesions, and enlarged perivascular spaces. It directly damages core motor control circuits, such as the corticospinal tract and cerebellar-thalamic pathway, often leading to complex symptoms such as lower limb motor dysfunction combined with cognitive impairment and autonomic dysfunction. Traditional rehabilitation methods, such as gait training and muscle strengthening, are time-consuming, but only a small percentage of patients can regain independent walking ability.

[0003] At present, clinical rehabilitation treatment for CSVD lower limb dysfunction mainly relies on functional electrical stimulation (FES) technology to stimulate neuromuscular system through low-frequency current to induce active movement. However, existing technologies have certain defects: traditional electrical stimulation systems only focus on single symptom intervention of lower limb motor function, such as the YSL02 lower limb functional electrical stimulation system of IELTS, which operates independently from rehabilitation methods such as physical therapy (such as gait training), cognitive intervention (such as balance training) and drug management, and lacks feedback data interaction and coordination of rehabilitation control. For example, the setting of electrical stimulation parameters is not linked to the joint range of motion data in physical therapy, nor is the effect of antidepressant drugs on neural excitability integrated, resulting in the inability to form a therapeutic synergy between motor function recovery, cognitive improvement and autonomic nervous system regulation.

[0004] Secondly, parameter adjustment in existing electrical stimulation systems relies primarily on manual tuning based on the experience of rehabilitation physicians, lacking intelligent optimization mechanisms based on patient-specific pathological characteristics. CSVD patients exhibit significant individual variability in the extent of white matter lesions, distribution of microinfarcts, and nerve conduction velocity. However, traditional systems lack access to MRI imaging data and real-time physiological feedback, resulting in an inability to dynamically adapt stimulation protocols to changes in neuromuscular function, leading to ineffective stimulation and excessive muscle fatigue.

[0005] Thirdly, existing systems only address lower limb motor dysfunction, ignoring the need for intervention for the common non-motor symptoms of CSVD patients. For example, for urinary incontinence caused by bladder detrusor dysfunction, traditional electrical stimulation lacks integrated pelvic floor electromyography (EMG) signal monitoring and targeted stimulation. Furthermore, while depressive symptoms are linked to decreased motor cortex excitability through neural circuitry, existing technologies fail to incorporate cognitive assessment data into stimulation protocol adjustment models, limiting rehabilitation intervention to improving motor function and failing to achieve comprehensive neurological restoration.

[0006] Therefore, it is necessary to propose an adaptive electrical stimulation-assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease that can realize the three-in-one collaborative rehabilitation model of "electrical stimulation-physical therapy-cognitive intervention", improve the efficiency of neural remodeling, reduce the risk of repeated microbleeding, and improve the patient's rehabilitation training experience. Summary of the Invention

[0007] In order to solve the above problems, the purpose of the present invention is to provide an adaptive electrical stimulation assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease. By detecting the patient's multimodal physiological data, identifying white matter lesions and microbleeding foci based on the patient's MRI image segmentation, and dynamically optimizing the electrical stimulation parameters in combination with a deep reinforcement learning model, the three-in-one collaborative rehabilitation model of "electrical stimulation-physical therapy-cognitive intervention" is realized through dynamic coupling of motion and cognitive training through VR tasks, thereby improving the efficiency of rehabilitation training, reducing the risk of recurrence of patients, and enhancing the patient's rehabilitation training experience.

[0008] To achieve the above objectives, the technical solution of the present invention is as follows: an adaptive electrical stimulation-assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease, comprising:

[0009] The data acquisition module is used to obtain multimodal data of patients in real time. The multimodal data includes motor function data, cognitive status data, autonomic nervous activity data and emotional index data;

[0010] The decision-making and control module generates dynamically optimized electrical stimulation parameters and rehabilitation task instructions based on multimodal data, deep reinforcement learning algorithms, and the patient's individualized pathological characteristics;

[0011] The electrical stimulation execution module is used to synchronously control and stimulate the targets corresponding to motor, cognitive, autonomic and emotional functions according to the optimized electrical stimulation parameters;

[0012] An interactive virtual reality module is used to combine immersive gait training and cognitive training to simultaneously assist in the rehabilitation of cognitive impairment.

[0013] Furthermore, the data acquisition module includes:

[0014] A flexible electrode array and a surface electromyography sensor are used to collect lower limb muscle activation signals. Both the flexible electrode array and the surface electromyography sensor are fixedly connected to the patient's lower limb skin;

[0015] Inertial sensors and plantar pressure sensors are used to monitor gait symmetry and balance. The inertial sensors are fixedly connected to the patient's lower limbs.

[0016] Near-infrared functional brain imaging device, used to detect activation states of the motor cortex and prefrontal cortex;

[0017] Biosignal collector, used to detect real-time physiological data of patients, including heart rate variability sensor, skin galvanic response sensor and bladder pressure sensor.

[0018] Furthermore, the decision-making and control module includes:

[0019] The pathological feature profiling unit uses a convolutional neural network to analyze the patient's brain MRI images, segment white matter high signal areas and microbleeds, and generate CSVD subtype labels;

[0020] A reinforcement learning model unit dynamically adjusts electrical stimulation parameters, including frequency, intensity, and target location, based on real-time physiological data and CSVD subtype labels;

[0021] The drug-electrical stimulation interaction database links patient medication records with electrical stimulation effects to avoid the risk of drug side effects.

[0022] Furthermore, the workflow of the reinforcement learning model unit is:

[0023] Receive gait cycle data from inertial sensors, muscle activation latency from surface electromyography sensors, and cortical activation signals from near-infrared brain functional imaging;

[0024] The electrical stimulation parameters were updated in a cycle of 0.1 seconds. When the electromyographic signal of the affected side was detected to be 80% lower than that of the healthy side, the electrical stimulation intensity was increased by 0.2 mA.

[0025] The initial parameter combination was selected based on the CSVD subtype label. Amyloidosis patients were treated with a low-frequency mode by default, with a stimulation frequency of ≤30 Hz.

[0026] Furthermore, the electrical stimulation execution module includes a lower limb muscle surface electrical stimulation unit, a transcranial direct current stimulation unit, a sacral nerve surface electrical stimulation unit and a vagus nerve stimulation unit.

[0027] Furthermore, in the electrical stimulation execution module:

[0028] A transcranial direct current stimulation unit applies 1-2 mA current to the dorsolateral prefrontal cortex to enhance cognitive function;

[0029] The sacral nerve surface electrical stimulation unit inhibits detrusor overactivity in a 50 Hz high-frequency mode via perineal electrodes;

[0030] The vagus nerve stimulation unit modulates parasympathetic nerve activity in a 1 Hz low-frequency mode via ear electrodes.

[0031] Furthermore, the interactive virtual reality module includes a virtual gait training unit, a cognitive task unit, and a feedback display interface;

[0032] A virtual gait training unit for dynamically adjusting virtual obstacle density and path complexity;

[0033] Cognitive task unit, used to generate mental arithmetic problems or spatial memory tasks in real time;

[0034] Feedback display interface, used to display real-time gait symmetry index and emotion score.

[0035] Furthermore, the virtual gait training unit includes a VR walking platform and a VR helmet.

[0036] The VR walking platform is equipped with a circular guardrail, the top of which is detachably connected to an underarm support, which is used to assist patients who are unable to walk upright in rehabilitation training.

[0037] The VR helmet is fixedly connected to a near-infrared brain function imager, a transcranial direct current stimulation unit and a vagus nerve stimulation unit. The VR helmet is used to provide integrated display of virtual gait training scenes, brain cortex activation state detection and electrical stimulation functions.

[0038] The basic solution works as follows: The system uses the data acquisition module to acquire multimodal patient data in real time, including motor function, cognitive status, autonomic nervous activity, and emotional indicators. The decision-making and control module analyzes and processes this data based on a deep reinforcement learning algorithm and the patient's individualized pathological characteristics, generating dynamically optimized electrical stimulation parameters and rehabilitation task instructions. The electrical stimulation execution module synchronously regulates and stimulates targets corresponding to motor, cognitive, autonomic, and emotional functions based on the optimized electrical stimulation parameters. The interactive virtual reality module combines immersive gait training with cognitive training, allowing patients to undergo rehabilitation training in a virtual reality environment while also assisting in the rehabilitation of cognitive impairments.

[0039] The beneficial effects of the basic program are: 1. Through multimodal data acquisition and deep reinforcement learning algorithms, the system can dynamically adjust electrical stimulation parameters according to the patient's real-time status to achieve personalized rehabilitation treatment plans. Compared with traditional rehabilitation methods, it can more efficiently promote the recovery of patients' lower limb function.

[0040] 2. It has realized the three-in-one collaborative rehabilitation model of "electrical stimulation-physical therapy-cognitive intervention", which not only focuses on lower limb motor dysfunction, but also takes into account the intervention of non-motor symptoms such as cognitive impairment and autonomic dysfunction, which helps to comprehensively improve the patient's neurological function and improve overall health.

[0041] 3. Based on the patient's MRI image segmentation to identify pathological features such as white matter lesions and microbleeds, combined with the reinforcement learning model to dynamically optimize the stimulation plan, it can better adapt to changes in the patient's neuromuscular function and reduce the risk of recurrence caused by ineffective stimulation or excessive stimulation.

[0042] 4. With the help of interactive virtual reality modules, patients are provided with an immersive gait training and cognitive training environment, which increases the fun and enthusiasm of rehabilitation training. At the same time, information such as gait symmetry index and emotional score are displayed in real time through the feedback display interface, allowing patients to understand their rehabilitation progress more intuitively, enhance their confidence in rehabilitation, and further strengthen their rehabilitation training experience.

[0043] Furthermore, it also includes:

[0044] The urinary incontinence management unit predicts abnormal detrusor contraction through the bladder pressure sensor, triggering high-frequency stimulation of the sacral nerve surface electrical stimulation unit and VR pelvic floor muscle training tasks;

[0045] The emotion regulation unit assesses depression based on heart rate variability sensor data, galvanic skin response sensor data, and a voice emotion recognition algorithm, automatically switching to low-frequency stimulation of the vagus nerve stimulation unit and a natural landscape VR scene.

[0046] The beneficial effects of the basic program are: 1. By adding new units, the system expands from "motor function + cognitive function" to "autonomic nervous function + emotional management", achieving full-dimensional intervention for lower limb dysfunction caused by cerebral small vessel disease (CSVD) combined with complex symptoms such as urinary incontinence and depression, enhancing rehabilitation effects, improving patients' experience, and reducing the difficulty of rehabilitation care.

[0047] 2. The emotion regulation unit assesses the depressive state based on heart rate variability sensor data, skin electrical response sensor data and voice emotion recognition algorithm, and automatically switches to low-frequency stimulation of the vagus nerve stimulation unit and natural landscape VR scene, which helps to relieve the patient's anxiety and depression, improve psychological resilience during the rehabilitation process, and enable patients to participate more actively in rehabilitation training.

[0048] 3. The system integrates urinary incontinence management and emotion regulation functions into the same system. Combined with multimodal data acquisition and deep reinforcement learning algorithms, it can dynamically adjust electrical stimulation parameters and virtual reality scenarios according to the patient's real-time physiological and psychological state, providing a more personalized and comprehensive rehabilitation treatment plan.

[0049] 4. By integrating urinary incontinence management and mood regulation functions into the same system, patients no longer need to switch between multiple devices or treatment methods during rehabilitation treatment, providing a smoother and more integrated rehabilitation experience and reducing the operational complexity and burden on patients during the treatment process.

[0050] Furthermore, the data acquisition module, decision-making and control module, electrical stimulation execution module, interactive virtual reality module, urinary incontinence management unit, and emotion regulation unit achieve multi-center data collaboration through federated learning technology:

[0051] The patient's physiological data is processed locally, and only the feature vectors are uploaded to the cloud;

[0052] Cross-institutional joint training of efficacy prediction models to optimize individualized parameter recommendations.

[0053] The beneficial effects of the basic solution are: 1. Through federated learning technology, the patient's physiological data is processed locally, and only the feature vectors are uploaded to the cloud, avoiding the direct transmission of raw data and effectively protecting the patient's privacy and data security.

[0054] 2. It achieves cross-institutional data collaboration. Data from different medical institutions can be used together to train efficacy prediction models, breaking down data silos, making full use of the rich data resources of multiple centers, and improving the accuracy and generalization ability of the model.

[0055] 3. The efficacy prediction model based on federated learning training can more accurately recommend personalized electrical stimulation parameters for different patients, improve the effectiveness and targetedness of rehabilitation treatment, and make the treatment plan more in line with the individual characteristics of the patient.

[0056] 4. The multi-center data diversity enhances the robustness of the efficacy prediction model, enabling it to maintain good performance and stability when facing patients from different populations and with different conditions, and better adapt to the complex and changing clinical reality. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 Schematic diagram of an adaptive electrical stimulation-assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease in an embodiment of the present invention.

[0058] Figure 2 Schematic diagram of a data acquisition module in an embodiment of the present invention.

[0059] Figure 3 Schematic diagram of a decision-making and control module in an embodiment of the present invention.

[0060] Figure 4 Schematic diagram of an electrical stimulation execution module in an embodiment of the present invention.

[0061] Figure 5 Schematic diagram of an interactive virtual reality module in an embodiment of the present invention.

[0062] Figure 6 Schematic diagram of a biological signal collector in an embodiment of the present invention.

[0063] Figure 7 Schematic diagram of a virtual gait training unit in an embodiment of the present invention.

[0064] Figure 8 Schematic diagram of the operation of the adaptive electrical stimulation-assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease in an embodiment of the present invention.

[0065] Figure 9 This is an axonometric view of the VR walking platform in an embodiment of the present invention.

[0066] The reference numerals in the drawings of the specification include: 1. VR walking platform; 2. annular guardrail; 3. underarm support. DETAILED DESCRIPTION

[0067] The following is further described in detail through specific implementation methods:

[0068] Example 1

[0069] An adaptive electrical stimulation-assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease, characterized by comprising:

[0070] The data acquisition module is used to acquire multimodal data from patients in real time, including motor function data, cognitive status data, autonomic nervous system activity data, and emotional indicators. The data acquisition module includes: a flexible electrode array and surface electromyography sensor for collecting lower limb muscle activation signals, both of which are fixedly connected to the patient's lower limb skin; an inertial sensor and plantar pressure sensor for monitoring gait symmetry and balance ability, with the inertial sensor fixedly connected to the patient's lower limb; a near-infrared brain function imager for detecting the activation status of the motor cortex and prefrontal cortex; and a biosignal collector for detecting the patient's real-time physiological data, including a heart rate variability sensor, a galvanic skin response sensor, and a bladder pressure sensor.

[0071] The decision-making and control module generates dynamically optimized electrical stimulation parameters and rehabilitation task instructions based on multimodal data, deep reinforcement learning algorithms, and the patient's individualized pathological characteristics. The module includes a pathological feature profiling unit that analyzes the patient's brain MRI images using a convolutional neural network to segment white matter hyperintensity areas and microbleeds, generating CSVD subtype labels; a reinforcement learning model unit that dynamically adjusts electrical stimulation parameters, including frequency, intensity, and target location, based on real-time physiological data and CSVD subtype labels; and a drug-electrical stimulation interaction database that links patient medication records with electrical stimulation effects to mitigate the risk of drug side effects.

[0072] The workflow of the reinforcement learning model unit is as follows: receiving gait cycle data from inertial sensors, muscle activation delay time from surface electromyography sensors, and cortical activation signals from near-infrared brain functional imaging; updating electrical stimulation parameters with a cycle of 0.1 seconds, and increasing the electrical stimulation intensity by a gradient of 0.2 mA when the electromyographic signal on the affected side is detected to be lower than 80% of the healthy side; selecting the initial parameter combination based on the CSVD subtype label; the low-frequency mode is used by default for patients with amyloidosis, with a stimulation frequency of ≤30 Hz.

[0073] The electrical stimulation execution module is used to synchronously control and stimulate targets corresponding to motor, cognitive, autonomic, and emotional functions based on optimized electrical stimulation parameters. The electrical stimulation execution module includes a lower limb muscle surface electrical stimulation unit, a transcranial direct current stimulation unit, a sacral nerve surface electrical stimulation unit, and a vagus nerve stimulation unit. In the electrical stimulation execution module, the transcranial direct current stimulation unit applies a 1-2 mA current to the dorsolateral prefrontal cortex to enhance cognitive function; the sacral nerve surface electrical stimulation unit suppresses detrusor overactivity in a 50 Hz high-frequency mode via perineal electrodes; and the vagus nerve stimulation unit regulates parasympathetic nerve activity in a 1 Hz low-frequency mode via ear electrodes.

[0074] The interactive virtual reality module combines immersive gait training with cognitive training to simultaneously assist in the rehabilitation of cognitive impairment. It includes a virtual gait training unit, a cognitive task unit, and a feedback display interface. The virtual gait training unit dynamically adjusts virtual obstacle density and path complexity; the cognitive task unit generates mental arithmetic problems or spatial memory tasks in real time; and the feedback display interface displays real-time gait symmetry index and emotion scores.

[0075] The virtual gait training unit includes a VR walking platform and a VR helmet. The VR walking platform is equipped with a circular guardrail. The top of the circular guardrail is detachably connected to an underarm support, which is used to assist patients who are unable to walk upright in rehabilitation training. The VR helmet is fixedly connected to a near-infrared brain function imager, a transcranial direct current stimulation unit and a vagus nerve stimulation unit. The VR helmet is used to provide an integrated display of virtual gait training scenes, cerebral cortex activation state detection and electrical stimulation functions.

[0076] It also includes: a urinary incontinence management unit, which predicts abnormal detrusor contraction through a bladder pressure sensor, triggering high-frequency stimulation of the sacral nerve surface electrical stimulation unit and VR pelvic floor muscle training tasks; an emotion regulation unit, which assesses depression based on heart rate variability sensor data, skin electrical response sensor data and voice emotion recognition algorithm, and automatically switches to low-frequency stimulation of the vagus nerve stimulation unit and natural landscape VR scenes.

[0077] The specific implementation process is as follows: After the patient wears the suit with distributed flexible electrode arrays and surface electromyography sensors, the system begins to capture the activation signals of the lower limb muscles in real time. Inertial sensors and plantar pressure sensors track the patient's gait symmetry and balance ability. For example, during walking, if the left leg stride is significantly shorter than the right leg, the inertial data and plantar pressure data will be different, and the sensor will immediately detect this asymmetry. At the same time, the near-infrared brain functional imager scans the patient's motor cortex and prefrontal cortex to monitor their activation status. When the patient tries to lift his legs, the changes in blood oxygen concentration in the motor cortex will be accurately captured. The biosignal collector continuously records heart rate variability, skin electrical response and bladder pressure. For example, a sudden increase in bladder pressure may indicate the risk of urinary incontinence, as shown in the attached Figure 6 These data are wirelessly transmitted to the central processor to form a real-time digital portrait of the patient's physiological and neurological functions, as shown in the attached Figure 2 shown.

[0078] Then the deep reinforcement learning model in the decision-making and control module receives the above digital portrait and starts to operate. First, the pathological feature portrait unit analyzes the patient's brain MRI image to identify the white matter high signal area and the location of microbleeding foci to determine whether it is hypertensive or amyloidosis CSVD subtype. For example, the initial electrical stimulation frequency of patients with amyloidosis will be automatically limited to below 30Hz to reduce the risk of microbleeding. Next, the reinforcement learning model combines real-time gait data, such as the 100ms delay in muscle activation on the affected side and the cortical activation state, such as insufficient prefrontal lobe activity, to dynamically adjust the parameters. If the electromyographic signal strength of the quadriceps femoris muscle of the right leg is only 75% of that of the healthy side, the system will increase the electrical stimulation intensity from 1.5mA to 1.7mA in a short time, and at the same time apply 1.2mA current to the prefrontal lobe through transcranial direct current stimulation (tDCS) to enhance attention, as shown in the attached Figure 3 As shown in the figure, the drug-electrical stimulation interaction database is involved simultaneously. If the patient is taking anticoagulants, the system will avoid high-frequency stimulation mode to prevent complications caused by increased vascular fragility.

[0079] The lower limb muscle surface electrical stimulation unit sends a current of a specific frequency to the quadriceps and gastrocnemius muscles according to the adjusted parameters, inducing muscle contraction to simulate natural gait, as shown in the attached figure. Figure 4As shown. For example, when the patient tries to cross a virtual obstacle, the quadriceps electrical stimulation intensity is temporarily increased to 2.0 mA to assist in completing the leg-lifting movement. At the same time, the VR helmet continuously stimulates the dorsolateral prefrontal lobe to help patients complete mental arithmetic tasks, such as 23+58, while walking on the VR walking platform, to avoid cognitive degeneration. The sacral nerve surface electrical stimulation unit monitors bladder pressure through perineal electrodes. If an abnormal contraction of the detrusor muscle is detected, such as a pressure value exceeding 40cm H2O, 50Hz high-frequency stimulation is immediately initiated to suppress the urge to urinate; the ear vagus nerve stimulation unit is based on the emotional score, such as a 20% increase in skin conductance, and switches to low-frequency mode to relieve anxiety in conjunction with the forest scene in VR, as shown in the attached figure. Figure 8 shown.

[0080] After the patient puts on the VR helmet, he enters the virtual shopping mall scene and needs to calculate the total price of the goods (cognitive training) while avoiding the moving shopping cart (gait training) as shown in the attached figure. Figure 7 The VR walking platform's circular guardrails and underarm supports provide physical support for patients who are unable to stand. The inertial sensor adjusts the scene difficulty in real time. If the patient successfully avoids obstacles five times in a row, the path complexity will automatically increase, such as adding ramps and revolving doors. Figure 9 The feedback interface displays the gait symmetry index and emotional rating in real time. After the training, the system generates a report on the 10-meter walking speed, the number of urinary incontinence episodes, and the depression scale score. This allows patients to understand their condition and the progress of rehabilitation training in real time, helping them build confidence in rehabilitation and improve their enthusiasm. Figure 5 shown.

[0081] Example 2

[0082] The difference from the above embodiment is that the data acquisition module, decision-making and control module, electrical stimulation execution module, interactive virtual reality module, urinary incontinence management unit and emotion regulation unit realize multi-center data collaboration through federated learning technology, the patient's physiological data is processed locally, and only the feature vector is uploaded to the cloud; cross-institutional joint training of efficacy prediction model optimizes individualized parameter recommendations.

[0083] The specific experimental process is as follows: Through federated learning technology, the patient's desensitized feature data, such as gait patterns and stimulus response curves, are uploaded to the cloud and used together with anonymous data from global medical institutions to train the efficacy prediction model, as shown in the attached figure. Figure 8 A primary care physician can use this model to recommend initial parameters for a newly diagnosed CSVD patient, such as 1.0mA low-frequency tDCS combined with gait training at difficulty level 2. Homebound patients can use this system to complete daily training, with data synced to a family member's mobile app. If the system detects a decrease in gait stability for three consecutive days, a follow-up visit reminder will be automatically sent.

[0084] 1. Experimental Design

[0085] Objective: To verify the effect of adaptive electrical stimulation-assisted rehabilitation system on improving lower limb dysfunction and non-motor symptoms in patients with cerebral small vessel disease (CSVD).

[0086] Participants: 100 patients diagnosed with CSVD (aged 65-80 years) were recruited and randomly divided into two groups:

[0087] The experimental group (50 cases) received rehabilitation training using the system, once a day for 45 minutes each time, for 12 weeks. The control group (50 cases) received traditional functional electrical stimulation (FES) plus conventional physical therapy, for the same duration and period.

[0088] Evaluation Metrics:

[0089] 1. Motor function: 10-meter walking speed (m / s), gait symmetry index (0-1), number of falls (times / month).

[0090] 2. Cognitive status: Montreal Cognitive Assessment (MoCA, total score 30 points).

[0091] 3. Autonomic nervous system symptoms: frequency of urinary incontinence (times / day), number of nocturnal urinations (times / night).

[0092] 4. Emotional management: Hamilton Depression Rating Scale (HAMD, total score 56 points).

[0093] 5. Treatment compliance: patient dropout rate (%).

[0094] 2. Experimental steps

[0095] 1. Baseline Assessment

[0096] All patients underwent MRI (to assess the extent of white matter lesions and the distribution of microbleeds), gait analysis, cognitive testing, and urodynamic testing.

[0097] 2. System intervention:

[0098] Experimental group:

[0099] Wearing flexible electrodes, a VR helmet, and biosensors, participants entered a virtual shopping mall scene for gait training (obstacle avoidance and mental arithmetic tasks). The system adjusted electrical stimulation parameters in real time (e.g., automatically increasing stimulation intensity on the affected side when gait asymmetry was detected). The urinary incontinence management unit triggered sacral nerve stimulation based on bladder pressure, and the emotion regulation unit switched to relaxation scenes through voice recognition.

[0100] Lower limb muscle stimulation was performed using a conventional FES device, combined with conventional gait training (without VR or cognitive tasks). Urinary incontinence and depressive symptoms were managed with oral medications (eg, tolterodine, sertraline).

[0101] 3. Data Collection

[0102] Mid-term assessments (gait speed, MoCA score) were performed every 4 weeks.

[0103] A comprehensive retest was conducted after 12 weeks to compare the changes before and after the intervention.

[0104] 4. Data Analysis

[0105] SPSS 26.0 was used to perform paired t-test and chi-square test, and the significance level was set at p < 0.05.

[0106] 3. Experimental Results

[0107] As shown in the following table:

[0108] Table 1. Baseline assessment data

[0109]

[0110] Table 2. Data after 12 weeks of systemic intervention

[0111]

[0112]

[0113] Note: Indicates that the comparison between the experimental group before and after the intervention or between groups is statistically significant (P < 0.05). *

[0114] From the data in Table 1 and Table 2, we can see that after 12 weeks of systematic intervention, the gait speed of the experimental group increased by 41%, significantly higher than the 21% of the control group. Thanks to the algorithm's dynamic optimization of stimulation parameters and VR task coupling training, the gait symmetry index increased from 0.68 to 0.82, indicating that the muscle strength recovery on the affected side was more balanced.

[0115] MoCA scores increased by 21% due to enhanced prefrontal cortex activation and cognitive task stimulation from tDCS, while the control group saw a 7% improvement. Depression scores decreased by 35% in the experimental group, compared to a 12% decrease in the control group, driven by the beneficial effects of vagus nerve stimulation and VR emotion regulation scenarios.

[0116] The experimental group experienced a 67% reduction in urinary incontinence frequency, driven by a synergistic effect of high-frequency sacral nerve inhibition and pelvic floor muscle training. The control group experienced a 23% reduction. The dropout rate in the experimental group was only 8%, significantly lower than the 28% in the control group, due to the engaging nature of VR training and its non-invasive design, which reduced discomfort. No serious adverse events were observed, confirming the safety of the system.

[0117] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0118] The above is only an embodiment of the present invention. Common knowledge such as the known specific structures and characteristics in the scheme is not described in detail here. Ordinary technicians in the field are aware of all common technical knowledge in the technical field of the invention before the application date or priority date, can obtain all existing technologies in the field, and have the ability to apply conventional experimental means before that date. Ordinary technicians in the field can improve and implement this scheme in combination with their own abilities under the inspiration given by this application. Some typical known structures or known methods should not become obstacles for ordinary technicians in the field to implement this application. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several variations and improvements can be made, which should also be regarded as the scope of protection of the present invention. These will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. An adaptive electrical stimulation-assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease, characterized by: include: The data acquisition module is used to obtain multimodal data of patients in real time. The multimodal data includes motor function data, cognitive status data, autonomic nervous activity data and emotional index data; The decision-making and control module generates dynamically optimized electrical stimulation parameters and rehabilitation task instructions based on multimodal data, deep reinforcement learning algorithms, and the patient's individualized pathological characteristics; The electrical stimulation execution module is used to synchronously control and stimulate the targets corresponding to motor, cognitive, autonomic and emotional functions according to the optimized electrical stimulation parameters; An interactive virtual reality module is used to combine immersive gait training and cognitive training to simultaneously assist in the rehabilitation of cognitive impairment.

2. The adaptive electrical stimulation-assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease according to claim 1, characterized in that: The data acquisition module includes: A flexible electrode array and a surface electromyography sensor are used to collect lower limb muscle activation signals. Both the flexible electrode array and the surface electromyography sensor are fixedly connected to the patient's lower limb skin; Inertial sensors and plantar pressure sensors are used to monitor gait symmetry and balance. The inertial sensors are fixedly connected to the patient's lower limbs. Near-infrared functional brain imaging device, used to detect activation states of the motor cortex and prefrontal cortex; Biosignal collector, used to detect real-time physiological data of patients, including heart rate variability sensor, skin galvanic response sensor and bladder pressure sensor.

3. The adaptive electrical stimulation-assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease according to claim 1, characterized in that: The decision-making and control modules include: The pathological feature profiling unit uses a convolutional neural network to analyze the patient's brain MRI images, segment the white matter high signal areas, and generate CSVD subtype labels; A reinforcement learning model unit dynamically adjusts electrical stimulation parameters, including frequency, intensity, and target location, based on real-time physiological data and CSVD subtype labels; The drug-electrical stimulation interaction database links patient medication records with electrical stimulation effects to avoid the risk of drug side effects.

4. The adaptive electrical stimulation-assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease according to claim 3, characterized in that: The workflow of the reinforcement learning model unit is: Receive gait cycle data from inertial sensors, muscle activation latency from surface electromyography sensors, and cortical activation signals from near-infrared brain functional imaging; The electrical stimulation parameters were updated in a cycle of 0.1 seconds. When the electromyographic signal of the affected side was detected to be 80% lower than that of the healthy side, the electrical stimulation intensity was increased by 0.2 mA. The initial parameter combination was selected based on the CSVD subtype label. Amyloidosis patients were treated with a low-frequency mode by default, with a stimulation frequency of ≤30 Hz.

5. The adaptive electrical stimulation-assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease according to claim 1, characterized in that: The electrical stimulation execution module includes a lower limb muscle surface electrical stimulation unit, a transcranial direct current stimulation unit, a sacral nerve surface electrical stimulation unit and a vagus nerve stimulation unit.

6. The adaptive electrical stimulation-assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease according to claim 5, characterized in that: In the electrical stimulation execution module: A transcranial direct current stimulation unit applies 1-2 mA current to the dorsolateral prefrontal cortex to enhance cognitive function; The sacral nerve surface electrical stimulation unit inhibits detrusor overactivity in a 50 Hz high-frequency mode via perineal electrodes; The vagus nerve stimulation unit modulates parasympathetic nerve activity in a 1 Hz low-frequency mode via ear electrodes.

7. The adaptive electrical stimulation-assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease according to claim 1, characterized in that: The interactive virtual reality module includes a virtual gait training unit, a cognitive task unit, and a feedback display interface; A virtual gait training unit for dynamically adjusting virtual obstacle density and path complexity; Cognitive task unit, used to generate mental arithmetic problems or spatial memory tasks in real time; Feedback display interface, used to display real-time gait symmetry index and emotion score.

8. The adaptive electrical stimulation-assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease according to claim 7, characterized in that: The virtual gait training unit includes a VR walking platform and a VR helmet. The VR walking platform is equipped with a circular guardrail, the top of which is detachably connected to an underarm support, which is used to assist patients who are unable to walk upright in rehabilitation training. The VR helmet is fixedly connected to a near-infrared brain function imager, a transcranial direct current stimulation unit and a vagus nerve stimulation unit. The VR helmet is used to provide integrated display of virtual gait training scenes, brain cortex activation state detection and electrical stimulation functions.

9. The adaptive electrical stimulation-assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease according to claim 1, characterized in that: Also includes: The urinary incontinence management unit predicts abnormal detrusor contraction through the bladder pressure sensor, triggering high-frequency stimulation of the sacral nerve surface electrical stimulation unit and VR pelvic floor muscle training tasks; The emotion regulation unit assesses the depressive state based on heart rate variability sensor data, skin electrical response sensor data and voice emotion recognition algorithm, and automatically switches to low-frequency stimulation of the vagus nerve stimulation unit and natural landscape VR scene.

10. The adaptive electrical stimulation-assisted rehabilitation system for lower limb dysfunction caused by cerebral small vessel disease according to claim 1, characterized in that: The data acquisition module, decision-making and control module, electrical stimulation execution module, interactive virtual reality module, urinary incontinence management unit, and emotion regulation unit achieve multi-center data collaboration through federated learning technology: The patient's physiological data is processed locally, and only the feature vectors are uploaded to the cloud; Cross-institutional joint training of efficacy prediction models to optimize individualized parameter recommendations.

Images