Effect of brain computer interface on limb motor function after intracerebral hemorrhage in basal ganglia and its rehabilitation mechanism

Effect of brain computer interface on limb motor function after intracerebral hemorrhage in basal ganglia and its rehabilitation mechanism | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Effect of brain computer interface on limb motor function after intracerebral hemorrhage in basal ganglia and its rehabilitation mechanism Peili Cao, Hong Ling, Hao Guo, JunHao Wang, Qihang Jin, Shijie Guo, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6791097/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Objective: Objective to explore the degree of recovery of limb motor function and potential rehabilitation mechanism of stroke patients with subacute basal ganglia cerebral hemorrhage after receiving rehabilitation training based on brain computer interface (BCI) of motor imagination. Methods: During hospitalization from December 2022 to October 2024, stroke patients with subacute basal ganglia intracerebral hemorrhage accompanied with limb movement disorder were randomly divided into conventional rehabilitation treatment group (Group C) and BCI group (group B), who completed traditional rehabilitation training, BCI training combined with traditional rehabilitation training, and the training time was 5 weeks. Before and after the training, the rehabilitation status of stroke patients was evaluated by Fugl Meyer Assessment Scale (FMA), simple mental state examination scale, modified Barthel index, online classification accuracy (CA), Brunnstrom staging, etc. Then resting state fMRI scanning was performed to compare the functional connectivity changes between the regions of interest and the whole brain. The imaginary part coherence analysis method was used to calculate the connectivity changes between channels in the 8-30hz frequency band of EEG, and the changes of power spectral density (PSD) and brain topography in the two stages before and after training. Results: A total of 44 stroke patients were included in this study, 22 in group C and 22 in group B. There was no significant difference in clinical related scores between the two groups before training ( P >0.05); After training, compared with group C, the FMA score of group B stroke patients was significantly improved ( P <0.05). The CA values of most stroke patients in group B showed fluctuations and a downward trend, and only 6 stroke patients could maintain relatively high Ca values (more than 70% threshold standard).In group B, the functional connectivity between sensorimotor network and default mode, ventral attention, language network and other networks was significantly enhanced in 19 stroke patients ( P <0.05). Among them, the connection between the prefrontal side of the left precentral gyrus (L_PEF) and the anterior parietal side of the left postcentral gyrus (L_3b), and the connection between the anterior inferior cortex of the left precentral gyrus (L_6v) and the lateral side of the left postcentral gyrus (L_1) were positively correlated with clinical scores ( P <0.05). The connectivity between multiple channels within 4-30hz (αandβfrequency bands) of EEG was significantly weakened, and the enhanced connectivity between C4-T7 and T8-T7 in theαfrequency band during the execution of motor imagination by the affected hand was positively correlated with the change of lower limb clinical score. After training, the PSD curve in the range of 8-30hz in the central motor area of the affected side of the brain tended to be smoother than before, and the energy value decreased to a certain extent; Through the changes of brain topographic map before and after, it was found that during the repeated training of the affected hands of many stroke patients, the ipsilateral hemisphere cortex had obvious personal leave related desynchronization phenomenon or this phenomenon was further enhanced than before, and tended to converge to the central region of the brain and the surrounding areas. Conclusion： BCI rehabilitation training can effectively promote the recovery of limb motor function in stroke patients with subacute basal ganglia cerebral hemorrhage. The possible mechanism of rehabilitation lies in enhancing the motor imagination ability of stroke patients and promoting the improvement of brain network activity. Basal ganglia intracerebral hemorrhage Brain computer interface Connectomics Neuroplasticity recovered Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Cerebral hemorrhage in the basal ganglia region is a significant cause of severe disability on a global scale. Despite attempts to intervene during the acute phase of cerebral hemorrhage, approximately 40% of stroke survivors exhibit varying degrees of functional impairment. Among these, loss of motor and sensory function in the upper extremities, including the hands and wrists, and cognitive dysfunction have a significant impact on stroke patients' quality of life [1,2] . Intracerebral hemorrhagic foci frequently result in the compression or destruction of brain tissue, which in turn can lead to damage to the functional areas of the brain. It has been demonstrated that minor lesions in functional areas are more challenging to recuperate from than dysfunction caused by more substantial lesions that do not affect functional areas [3] . In the field of neuroscience, the recovery of upper extremity motor function after stroke is often based on neural remodeling or compensation, focusing on restoring function by reinforcing correct movement patterns. Traditional rehabilitation includes physical and occupational therapy, which facilitates recovery with the help of external physical stimulation or activities of daily living; or acupuncture, which stimulates acupuncture points to aid in the recovery of muscle function. These methods focus on training peripheral nerves or tissues to further promote central nervous reorganization and ultimately improve upper extremity motor function. In recent years, innovative interventions on motor rehabilitation have emerged with the goal of enhancing stroke patients' ability to perform daily living tasks, functional independence, and quality of life. These interventions include BCI training, transcranial direct current stimulation, functional electrical stimulation, deep brain stimulation, robot-assisted therapy, and virtual reality training, providing additional avenues for rehabilitation. Brain-computer interface (BCI) is the process of transforming acquired brain biological signals through feature extraction into specific commands to drive external devices and promote the remodeling of brain function for stroke patients through feedback training of human-computer interaction [4] . Currently, BCI training is an effective intervention to promote the improvement of limb motor function and increase brain activation, and is dominated by BCI (MI-BCI) training based on motor imagery (MI) paradigm. Most studies [5] have shown the statistical significance and clinical relevance of rehabilitative BCI for the improvement of upper limb motor function recovery. Currently, the brain performs complex activities through the integration and coordination of information between different brain regions. And the study of brain functional networks plays an important role in revealing the encoding, processing and integration of information during cognitive processes [6] . The study of these brain network patterns helps to deeply understand the mechanism of functional activities of the brain and promote the development of neuroscience. However, how to accurately identify and characterize changes in brain network connectivity patterns during rehabilitation remains a major challenge for current research. The key method to study brain function is to analyze the neural signal activation in functional brain regions, and the main techniques include electroencephalography (EEG), magnetoencephalography, functional near-infrared spectroscopic imaging, and rest-state functional magnetic resonance (rs-fMRI) imaging [7] . A preliminary study [8] confirmed that the improvement of the function of the affected limb during BCI training in subacute stroke patients is closely related to the resting EEG oscillatory activity of the ipsilateral cerebral hemisphere and its connectivity changes. Some scholars [9] have also used this novel BCI task model to map the enhancement of functional connectivity between multiple brain regions in the frontal, occipital, and parietal lobes of the subjects from a functional magnetic resonance perspective. The aim of this study was to evaluate the rehabilitation effect of BCI training on stroke patients with cerebral hemorrhage in the basal ganglia region accompanied by limb motor dysfunction, to develop a functional analysis based on MRI imaging and EEG data, and to explore in-depth the potential neuroplasticity mechanisms in the process of recovery of their limb motor function. Objects and Methods 1.1 Study subjects Stroke patients with cerebral hemorrhage in the basal ganglia region in the subacute stage who were hospitalized in Shanxi Provincial People's Hospital and Shanxi Provincial Acupuncture Hospital from December 2022 to October 2024, accompanied by motor dysfunction of the upper limbs, were selected for this study. Clinical indicators such as stroke patient's name, gender, age, past history, disease duration, hemorrhage site and treatment modality were recorded. The study was approved by the Ethics Committee of Shanxi Provincial People's Hospital with the approval number (2022) Provincial Medical Science Lun Audit No. 95, and all stroke patients participating in the trial signed an informed consent form. Inclusion criteria: (1) Diagnosed with cerebral hemorrhage in the basal ganglia region by MRI and CT, accompanied by limb motor dysfunction; (2) First onset of the disease within 1–6 months; (3) Defined as age ≥ 18, gender is not limited; (4) Simplified Mental State Score ≥ 18, can understand and implement the training of the basic prompts and commands; (5) Due to the dysfunction caused by the lesion of the functional area, need to carry out rehabilitation treatment. stroke patients; (6) the subjects themselves or their legal representatives agreed to participate in the trial and signed a written informed consent form. Exclusion criteria: (1) the presence of serious cognitive, mental and psychological disorders; (2) serious heart, liver, kidney, lung and other important organ failure; (3) with upper limb defects, open wounds, deformities, or with severe pain; (4) suffering from severe vision, hearing impairment, claustrophobia, etc.; (5) the body contains metals that can not be cooperated with the magnetic resonance examination; (6) the subject is participating in other clinical trials at the same time. 1.2 Experimental grouping and BCI training A total of 44 stroke patients with subacute stage basal ganglia cerebral hemorrhage accompanied by limb movement disorders were included in this study, and were randomly assigned to the conventional rehabilitation therapy group (Group C) and the BCI group (Group B) according to whether or not BCI training was used in the training content, with 22 stroke patients in each group. Group C: stroke patients received traditional rehabilitation training, including occupational therapy, acupuncture, exercise therapy and other treatments, 5 times per week for 5 weeks; Group B: stroke patients received traditional rehabilitation training and BCI training. In addition to the traditional rehabilitation training, a professional with 3 or more years of experience conducted 25 BCI training sessions of 25 minutes/day, about 5 times per week for 5 weeks for Group B stroke patients. BCI training uses non-invasive EEG to recognize MI-based motor task signals and convert them into commands, controlling the exoskeleton hand to guide the stroke patient to perform grasping and opening training of the affected hand. It consists of two main phases, basic and elevation training, during which the stroke patient is told to visualize the specific activities of the hand, and the exoskeleton hand provides support to help the stroke patient complete the grip/open hand task. The system analyzes and processes the stroke patient's EEG data during motor imagery collected by a 16-lead EEG cap, extracts EEG signal features as the basis for task recognition, and if the recognition is successful, the exoskeleton device connected to the computer guides the stroke patient's hand movement, while the computer monitor prompts for success to be given as feedback, and the cycle repeats until the end of the training. Each training completed 55 MI tasks, the screen randomly displays the video of holding the right and left hands and prompts, to avoid coughing, body shaking and other strenuous movements. The training scenario is shown in Fig. 1 . 1.3 Clinical rehabilitation assessment Fugl-Meyer assessment (FMA) scale, including upper and lower extremity motor function scores (FMA-UE, FMA-LE), could assess stroke patients' limb motor function; Mini-mental State Examination (MMSE) was used to check stroke patients' cognitive function (total score 30 points, 21–26 points for mild cognitive deficits, 10–20 points for moderate cognitive deficits, 0–9 points for severe cognitive deficits); modified Bartlett's scale was used to assess stroke patients' cognitive function. Stroke patients' cognitive function can be examined (total score of 30 points, 21–26 points are mild cognitive deficits, 10–20 points are moderate cognitive deficits, and 0–9 points are severe cognitive deficits); the modified Barthel index can assess stroke patients' ability in daily life (≥ 60 points: basically taking care of themselves; 41–59 points: moderate dysfunction, needing other people's help; 21–40 points: severe dysfunction, obvious dependence); brunnstroms index can assess stroke patients' ability in daily life. The brunnstrom staging assesses the different stages of rehabilitation of stroke patients and consists of 6 clinical stages. Specific criteria for the above scales are detailed in the Appendix. All rehabilitation assessments were performed by 2 trained clinicians, who independently evaluated the scales before and after the entire training phase, and the results were taken after agreement. Online classification accuracy (CA), which is the classification accuracy of the stroke patient when performing the task, is used to measure each training performance. The first stage is offline data collection without real-time feedback. After the collection, the current offline accuracy was obtained by using ten-fold cross-validation. The whole data of the first stage was used as the training set to obtain the optimal classifier to predict the EEG data of the second stage, and the CA value of the MI task of the second stage was also obtained. A larger CA value for this stroke patient indicates a higher degree of human-computer interaction during training [ 10 ] . 1.4 MRI data processing Before and after the training, a GE 3.0T MRI scanner was used to examine the head of stroke patients in group B. The magnetic field was 3 T. The stroke patients were placed in a supine position with their heads fixed, and the acquisition of high-resolution MR images, including T1 phase and rs-fMRI phase images, was accomplished using a standard 32-channel head coil. During the rs-fMRI scanning process, stroke patients were required to close their eyes and remain awake without specific thinking. Finally, the acquired MR data were further constructed for connectomics image analysis. Based on the multimodal segmentation atlas of the Human Connectome Project [ 11 ] , an “individualized re-segmentation” technique was used. Twenty-one brain regions related to the sensorimotor cortex were selected as regions of interest (ROIs), which encompassed the sensorimotor network and premotor regions, and also corresponded to Brodmann areas 1–8, and were mainly composed of primary sensory cortex (areas 1, 2, 3a, 3b), primary motor cortex (area 4), paracentral lobular area (areas 5L, 5m, 5mv), supplementary motor area (areas 5L, 5m, 5mv), and supplementary motor area (areas 5L, 5m, 5mv). 5L, 5m, 5mv), supplementary motor areas (areas 6ma, 6mp, SFL), premotor areas (areas 6a, 6d, FEF, 55b, PEF, 6r, 6v), lateral part of parietal lobe (area AIP), and part of anterior cingulate gyrus (areas 24dd, 24dv). In this study, we will analyze the changes in functional connectivity (FC) between the stroke patients' ROI and the whole brain cortex, specifically by calculating the FC coefficients between the ROI and each voxel of the whole brain, and then statistically analyzing them. Finally, the relationship between clinical scores and FC will be analyzed using NumPy 1.12.1 and Scipy 0.19.0 software. 1.5 EEG signal processing MI analysis is the most commonly used method to detect motor intentions based on the electrical activity of the motor cortex [ 12 ] . 1.5.1 Data preprocessing Each training contained two phases, the first phase contained 30 imagery trainings and the second phase contained 58 imagery trainings. The imagery data for the entire training was band-pass filtered from 5-35Hz, using an 8th order Butterworth filter. Subsequently, the filtered data were sliced, and the data segments (epochs) 0–4 seconds after cue appearance were selected for each imagery training. The standard deviation of each lead was calculated for each Epoch, and if the standard deviation of a single lead exceeded a threshold (less than 1 µV, or greater than 150 µV), that lead was removed. If more than 30% of the epoch in a single phase was removed, the data for the entire Session was discarded. 1.5.2 Connectivity analysis In this study, we used imaginary coherence (iCoh) as the EEG connectivity analysis method to analyze the preprocessed EEG signals and calculate the coherence between the leads in order to assess the brain connectivity of stroke patients in the early and late stages of treatment. iCoh method is based on spectral coherence, which employs the imaginary part in the result of the spectral coherence as a way of reducing the volumetric conduction effect generated by the artifacts and better reveal interactions between brain regions. $$\:\text{i}\text{C}\text{o}\text{h}\left(\text{e}\text{e}\text{g}\right)=\frac{{\left|\text{i}\text{m}\text{a}\text{g}\left({\text{S}}_{\text{x}\text{y}}^{}\right(\text{e}\text{e}\text{g}\left)\right)\right|}^{2}}{{\text{S}}_{\text{x}\text{x}}^{}\left(\text{e}\text{e}\text{g}\right){\text{S}}_{\text{y}\text{y}}^{}\left(\text{e}\text{e}\text{g}\right)}$$ where, is the cross-frequency spectral density on electrodes X and Y, and, are the self-frequency spectral densities of electrodes X and Y, respectively. The connectivity data of all stroke patients in the experimental group before and after 3 effective trainings, stage 2, were selected and averaged for significance analysis. If the effective training was less than 3 for the first or second 6 treatments in a stroke patient, the data for that stroke patient were discarded. Significance analysis was performed using a two-tailed paired t test. To test the relationship between EEG connectivity and clinical scores, Spearman correlation analysis was used in this study. 1.5.3 Power spectral density and brain topography calculation and analysis The training EEG signals of the stroke patients were collected, and the EEG signals of the central motor cortex region were extracted within 4 seconds of the motor imagery, and the average energy value in the range of 8-30Hz was calculated to obtain the power spectral density (PSD), which can reflect the distribution of the power of the EEG signals of the central motor region of the bilateral cerebral hemispheres in the frequency domain. At the same time, the 8-30Hz EEG signals of the whole leads were collected, and the average energy values of each brain region at different stages were calculated, and the brain topography was drawn to observe the energy changes in the C3 and C4 regions and to analyze whether there was event-related desynchronization (ERD) phenomenon. 1.6 Statistical processing SPSS 26.0 software was used for statistical analysis. Normally distributed measurements were expressed as mean ± standard deviation (x ± s), two-sample independent t-test was used for comparison between the two groups, and paired t-test was used before and after the training; rank information was expressed as quartiles [M(Q1, Q3)], and rank-sum test was used for comparison between the two groups; categorical information was expressed as relative numbers, and the X2 test or the Fisher's exact probability method; correlation analysis was performed by Spearman's analysis; two-sided test was used, and P < 0.05 was regarded as statistically significant difference. Results 2.1 Comparison of general information of stroke patients in two groups There is no statistically significant difference between the general information of stroke patients in the two groups (P>0.05). See Table 1. Table 1Stroke patients’ Characteristics Characteristic BCI（n=22） Contrl（n=22） T/X 2 /Z-value P-value Age, (year） 47.14±10.74 50.32±7.82 1.124 0.268 Duration of disease (months) 2.89±1.70 3.00±1.75 0.218 0.828 Hypertension 0.101 0.750 yes 15 14 no 7 8 Diabetes Mellitus 0.820 0.365 yes 13 10 no 9 12 Sex 0.121 0.728 male 17 16 female 5 6 Hemorrhage site 0.364 0.546 right basal ganglia hemorrhage 10 12 Left basal ganglia hemorrhage 12 10 Treatment modality 0.863 0.353 Surgical treatment 12 15 Conservative treatment 10 7 2.2 Comparison of stroke patients' clinical related scores before and after training There was no statistically significant difference in the comparison of clinical related scores between the two groups before training (P>0.05). After training, compared with group C, the FMA-UE and FMA-LE scores of stroke patients in group B were significantly higher than before (P0.05), as shown in Tables 2、3, and 4. Table 2 Comparison of clinical related scores between the two groups of stroke patients before training Variables BCI Contrl T/Z-value P-value FMA-UE 11.91±5.53 11.09±7.32 -0.419 0.677 FMA-LE 14.41±5.32 13.45±5.97 -0.560 0.579 MMSE 22.36±3.67 23.41±4.63 0.830 0.411 MBI 50.50±16.48 54.32±21.78 0.656 0.516 Brunnstrom classification of upper limb 2(1，2) 2(1.75，3) 0.479 0.489 Brunnstrom classification of hand 1(1，2) 1.5(1，2) 1.158 0.282 Brunnstrom classification of lower limb 3(2，3) 3(2，3) 0.144 0.704 Table 3 Comparison of clinical related scores between the two groups of stroke patients after training Variables BCI Contrl T/Z-value P-value FMA-UE 21.41±9.83 13.50±7.20 -3.044 0.004 FMA-LE 20.91±5.33 17.14±5.88 -2.230 0.031 MMSE 25.36±3.51 24.23±4.45 -0.940 0.353 MBI 66.36±13.85 64.09±20.62 -0.429 0.670 Brunnstrom classification of upper limb 2.5(2，3.25) 2.5(2，3) 0.342 0.559 Brunnstrom classification of hand 2(1.75，2.25) 2(1，2) 2.111 0.146 Brunnstrom classification of lower limb 3(3，4) 3(3，4) 1.238 0.266 Table 4 Comparison of clinically relevant scores before and after training in the BCI group of stroke patients Variables pre post t/Z值 P-value FMA-UE 11.91±5.53 21.41±9.83 -5.74 <0.001 FMA-LE 14.41±5.31 20.91±5.32 -6.97 <0.001 MMSE 22.36±3.67 25.36±3.51 -6.39 <0.001 MBI 50.50±16.48 66.36±13.85 -7.30 <0.001 Brunnstrom classification of upper limb 2(1，2) 2.5(2，3.25) -2.951 0.003 Brunnstrom classification of hand 1(1，2) 2(1.75，2.25) -3.276 0.001 Brunnstrom classification of lower limb 3(2，3) 3(3，4) -3.314 0.001 2.3 Analysis of training accuracy of stroke patients in Group B Only six stroke patients in Group B could maintain relatively high CA values, and their average CA values exceeded the 70% threshold criterion, and most of the remaining stroke patients showed more obvious fluctuations in CA values and even a downward trend in the course of BCI training, see Figure 2. 2.4 MRI-related analysis of group B stroke patients 2.4.1 Rehabilitation assessment and rs-MRI analysis Excluding 4 stroke patients with important data missing in group B, a total of 18 stroke patients were included in this study who successfully completed MRI scanning, and their FMA-UE scores were significantly improved compared with the previous ones (P<0.05), see Table 5; rs-MRI analysis showed 84 enhanced FCs of the stroke patients (P< 0.05), see Figure 3 for details. Table 5 Comparison of FMA scores before and after training in 18 stroke patients in the BCI group Variables pre post T/Z-value P-value FMA-UE 11.56±4.902 20.89±10.203 -4.982 <0.001 FMA-LE 14.17±5.544 21.00±5.099 -6.349 <0.001 MBI 50.61±17.493 65.72±14.323 -5.847 <0.001 2.4.2 Correlation analysis between clinical scores and FC The pre-post comparison of the FMA scores of the 18 stroke patients scanned by MRI in Group B remained statistically significant (P < 0.05), as shown in Tables 5. The correlation analysis of the FC with the FMA-UE scores showed that: the lateral aspect of the left postcentral gyrus (L_1) and the anterior-inferior cortex of the left precentral gyrus (L_6v), the anterior-parietal aspect of the left postcentral gyrus (L_3b) and the anterior-inferior cortex of the left precentral gyrus (L_6v) , FC between the anterior parietal side of the left postcentral gyrus (L_3b) and the prefrontal side of the left precentral gyrus (L_PEF), between the anterolateral side of the left precentral gyrus (L_55b) and the anterior subcortex of the left precentral gyrus (L_6v), between the anteroparietal side of the left precentral gyrus (L_PEF) and the posterolateral side of the left prefrontal lobe (L_p9-46v), and all of them were positively correlated with the clinical scores (P< 0.05), showing that the stroke patients' improvement in limb motor function was associated with elevated FC within the brain network, as shown in Tables 6. Tables 6 Correlation analysis between FC and clinical scores FC clinical scores r-value p-value L_1-L_6v FMA-UE 0.478 0.045 L_3b-L_6v FMA-UE 0.547 0.019 L_3b-L_PEF FMA-UE 0.576 0.012 L_55b-L_6v FMA-UE 0.514 0.029 L_PEF-L_p9-46v FMA-UE 0.527 0.025 2.5 Network analysis of EEG function in group B stroke patients 2.5.1 Correlation analysis between clinical scores and connectivity between EEG channels Data from 19 stroke patients in Group B were eligible, of which 8 had left-handedness on the affected side and 11 had right-handedness on the affected side. The EEG functional data of the left affected hand was mirror reversed (EEG viewed as symmetrical distribution of left and right brain functions), transformed to the right hand, and then analyzed statistically. The results showed that when the affected hand performed motor imagery, the connectivity between most of the channels in the EEG frequency band within 8-30 Hz (α and β bands) was significantly weakened (P< 0.05); in addition, the connectivity between C4-T7 and T8-T7 in the α band was enhanced, and it was positively correlated with the FMA-LE scores (P< 0.05), as shown in Figures 4 and Tables 7. Tables 7 Correlation analysis between connectivity and clinical scores connectivity clinical scores r-value p-value α bands’sC4-T7 FMA-LE 0.536 0.018 α bands’sT8-T7 FMA-LE 0.506 0.027 2.5.2 PSD and brain topography Before and after training, it was observed that 18 stroke patients at C3 and C4 electrodes showed differences in the corresponding energies in the 8-30 HZ band of the EEG of the bilateral hemispheres during MI, which indicated the different abilities of the bilateral hemispheres to perform MI. 15 stroke patients after BCI training showed a flattening of the PSD curves of the central motor area of the brain of the affected side in this band and a decrease in the energy values; the topography of the brain showed that the cortical cortex of the affected side showed obvious ERD phenomena or phenomena that were further enhanced than before and had a tendency to converge to the central and peripheral areas of the brain; the remaining stroke patients showed a relative enhancement of the affected side of the brain. The topography of the brain showed that the affected side of the hemispheric cortex showed obvious ERD phenomenon or the phenomenon was further enhanced compared with the previous one, and there was a tendency to converge towards the central brain area and its surroundings; the remaining stroke patients could see the ERD phenomenon of the affected side of the brain relatively enhanced, or the ERS phenomenon of the healthy side of the brain relatively weakened, and so on. For example, in stroke patient 07, the PSD map showed that there was an opposite difference in EEG energy at electrodes C3 and C4 before and after training; the topographic brain map showed that the ERD phenomenon of the left hemisphere became more concentrated, mainly around electrode C3, and obvious ERS phenomenon was observed in the right cerebral cortex, as shown in Figures 5. Discussion 3.1 Stroke patient characteristics and training and rehabilitation efficacy study Cerebral hemorrhagic stroke is one of the most destructive diseases in neurosurgery, which not only directly leads to local brain tissue damage and irreversible neurological deficits [13] , but also leaves many serious complications. Among them, cerebral hemorrhage in the basal ganglia region with limb motor function impairment is one of the core challenges in clinical rehabilitation, which is often manifested as hemiparesis, dystonia abnormalities, muscle atrophy and other motor deficits, or even permanent disability. Medication and rehabilitation are still the main clinical treatments for limb motor function impairment after stroke [14] , but the rehabilitation efficacy is still poor. And BCI technology is to build a direct communication bridge between the brain and the external environment, which can realize the information interaction between the brain and the outside world. Currently, MI-BCI-based combined Neurofeedback (NF) training, as an innovative rehabilitation treatment method, can help stroke patients to complete the recovery and remodeling of brain function by processing and converting brain biological signals into specific commands, driving external devices, and through the feedback mechanism of human-computer interaction [15] . The efficacy of MI-BCI-based training in limb rehabilitation for stroke patients with functional brain injury has now been demonstrated.1 meta-analysis in 2021 [16] indicated that neurofeedback intervention with BCI was significantly more effective than the control group for paralyzed upper limbs, with a standardized mean difference in FMA-UE scores of 0.48 (95% CI 0.16-0.80), demonstrating that BCI-based training is superior to traditional interventions in stroke patients' upper limb motor recovery is superior to traditional interventions. On this basis, some scholars have done further research, Nan et al [17] conducted BCI intervention for chronic stroke patients who had received poor results from traditional therapies, and the stroke patients' limb movement, language and cognitive function improved in various aspects; Yuan et al [18] designed a randomized controlled clinical trial to provide real-time feedback based on the attention index of stroke patients during BCI pedaling training. A randomized controlled clinical trial was designed by Yuan et al [18] to provide real-time feedback in multiple forms to enhance stroke patient participation in training, and the results showed that the lower limb motor function scores of stroke patients with BCI intervention were significantly improved. The MI-BCI technique mainly instructs stroke patients to perform repetitive reinforcement training to simulate their own brain's potential awareness of a certain movement to achieve functional recovery; performing MI is also a cognitive exercise process, which imagines a part of your body's movement without actually moving that part of your body to approach the actual movement execution; therefore, performing MI training has a slightly higher demand on the stroke patient's initial cognitive function, and a smaller demand on the stroke patient's residual motor function. function is less demanding. The stroke patients with cerebral hemorrhage in the basal ganglia region included in this study, all of whom had mild cognitive impairment, were able to perform some HCI training, which was manifested by different degrees of improvement in clinical scores, especially in FMA scores (group B was larger than group C), reflecting the significant Improvement effect of BCI training on the stroke patients' limb motor function. MMSE scores and modified Barthel index before and after training of stroke patients in group B. The stroke patients in group B were found to be in good condition before and after the training, and there was a significant difference in MMSE scores, modified Barthel index before and after the training, brunnstrom staging improved to a certain extent (P< 0.05), but the difference was not statistically significant compared with C. The difference between group B and group C was not statistically significant, and the difference between group B and group C was not statistically significant. This may be due to the fact that most of the stroke patients were at the lower level of the mild cognitive impairment range, and the BCI training in this study mainly targeted the distal upper limbs of the stroke patients, so it was difficult to improve the cognitive deficits; and the modified Barthel index, brunnstrom stage and other indicators for the stage of the rehabilitation process of the composite scores, and it is difficult to cross the short-term training to improve. In the future, we need to further explore the length of BCI training intervention and the scope of the optimal population [19] . 3.2 Changes in CA with BCI training Not all stroke patients in this study were able to maintain relatively high CA levels or significantly elevated CA values during BCI training, which is consistent with the findings of Tam et al [20] . In addition, changes in CA values were not completely consistent with changes in FMA scores [21] , so CA values do not accurately reflect the final BCI advancement of stroke rehabilitation, but only suggest a training status of stroke patients performing MI. For example, stroke patient 7 had a poor mean CA value performance (<60%), but his FMA score improved by 5 points. In contrast, Group B had reasons why most stroke patients did not make any progress in CA: some stroke patients initially trained with confidence and curiosity to complete the training, and felt fidgety and bored after several repetitions of the training, and were unable to focus on motor imagery (e.g., stroke patients 8, 9, and 10); the adverse effects of the change in condition after stroke that caused stroke patients to have a low emotional state (e.g., stroke patients 12 and 22); and the fact that some stroke patients did not trust and were not willingness to use new technologies that are not yet widely used, resulting in low human-machine cooperation, etc. 3.3 BCI training-related brain network research Current interventions for post-cerebral hemorrhage often focus on acute-phase hematoma management, blood pressure regulation, and post-acute rehabilitation [22,23] , but there is still a lack of research strategies on the multidimensional rehabilitation mechanisms of limb movement disorders (e.g., functionally-driven around aspects of neuroplasticity). rs-fMRI, as a non-invasive technique that can provide comprehensive characteristic information about the structural and functionally-related aspects of the brain, is one of the powerful tools for in-depth analysis of neuroplasticity mechanisms [24] . 3.3.1 FC changes in brain networks of stroke patients after BCI training In this study, we analyzed the correlation between ROI and time series of other brain regions to discover the possible connection between the recovery of motor function and FC changes in brain networks in stroke patients with cerebral hemorrhage. The results suggested that stroke patients had multiple FC enhancements within sensorimotor networks, including somatosensory cortex (postcentral gyrus, areas 1, 2, 3a, 3b, etc.), primary motor cortex (precentral gyrus, area 4, etc.), and premotor cortex (precentral gyrus anterior region, areas 55b, 6v, 6mp, etc.); and FC enhancements between sensorimotor and default isomodal networks such as anterior cingulate cortex (anterior side of the corpus callosum knee region a24, p32); FC enhancement between sensorimotor and language networks, e.g., premotor cortex (precentral anterior gyrus, region 55b), Broca's area, and memory-learning-related areas (lateral inferior frontal gyrus, regions 44, 45); FC enhancement between sensorimotor and visual networks, vision-related (ventral medial occipital lobe, region VMV2); FC enhancement between sensorimotor and ventral attention FC enhancement between sensory-motor and visual networks, such as eye-movement-related areas (medial superior frontal gyrus, area 8BM), anterior cingulate cortex (posterior inferior superior frontal gyrus, area a32pr); FC enhancement between sensory-motor and emotion-related areas, such as lateral orbitofrontal cortex (area 47m, etc.).In contrast, BCI training can integrate biological activities such as auditory, visual, proprioceptive, and higher cognition, making it a more comprehensive intervention strategy [25,26] . During the training period, stroke patients are required to maintain a high level of concentration, be constantly alert to handshake/open hand movements, and receive repetitive motor cues and image feedback (e.g., display of a smiling or crying face) upon completion of an MI task. With familiarity with these two forms of signaling, stroke patients no longer need to understand the specific semantics of the cues [27] and can automatically translate these signals into auditory or visual information that aids in decision-making and acts directly on the MI session. This process relies on the synergy of multiple aspects of higher cognition, sensorimotor integration, memory-learning transformation, the audiovisual system, and emotional stabilization, and reflects the compensatory mechanism of FC enhancement among brain networks. 3.3.2 Stroke patient limb motor function rehabilitation affects FC changes Based on the fact that MI-BCI can act on multiple areas related to sensorimotor synergy (e.g., anterior cingulate gyrus, parietal lobe, and lateral frontal cortex), it involves both direct contact fiber projections (e.g., superior longitudinal fasciculus, frontal-parietal fasciculus) and thalamic indirect projections. The premotor cortex, a key area in the frontal lobe of the brain responsible for motor planning, coordination, and sensory-motor integration, is located anterior to the primary motor cortex (M1) and is an important component of the motor control network. The precentral anterior gyrus anterior inferior cortex (area 6v) is part of the premotor cortex, and area 6v exhibits an abnormally active state when the stroke patient is controlling hand manipulation of an object to occur movement (e.g., grasping and lifting an object), or is involved in performing a specific motor task based on visual cues [28] . Area PEF belongs to the same premotor area and is located at the junction of the precentral sulcus and the inferior frontal sulcus. This area is associated with reflexive eye movements, such as reflexive sweeps. Area 1 is the posterior part of the primary sensory area, located at the postcentral gyrus, which extends to the midline, and its white matter often projects to the pyramidal fasciculus, thalamocortex, and parietal lobe, among others.Area 1 is adjacent to area 3b and participates in the processing of tactile stimuli (manifested as area 1 being the second activation point of area 3b after receiving tactile activation); area 1 may also collaborate with area 2 of the somatosensory cortex in the reception of tactile stimulus information in both hands. Area 3b is the anterior half of the primary sensory area, which is the initial area of cortical activation related to tactile stimulation, especially specific information such as tactile stimulation or injurious stimulation of the finger skin. In this study, the increased connections within the premotor cortex (L_55b-L_6v) and between the premotor cortex and several regions such as primary sensory areas or eye movement-related areas (e.g., L_PEF-L_3b, L_6v-L_1, and L_6d-L_8BM) may reflect the close association between the involvement of somatosensory and eye movement behaviors during the rehabilitation of the upper limb motor function in stroke patients. Among them, the FC between the prefrontal side of the left precentral gyrus (L_PEF) and the anterior parietal side of the left postcentral gyrus (L_3b), and between the anterior subcortex of the left precentral gyrus (L_6v) and the left postcentral gyrus (L_3b, L_1) showed a moderate positive correlation with the FMA-UE scores (P< 0.05). This reflected a significant increase in the activation of the sensorimotor network during the recovery of the stroke patient's limb motor function, which was manifested by the occurrence of unification of the premotor cortex with the sensory system. The left precentral gyrus anterolateral (L_55b) is located in the premotor cortex area and plays an important role in language processing in addition to coordinating motor tasks [28] . The present study showed that FC between area 55b and the anterior inferior cortex of the precentral gyrus (area 6v) was positively correlated with FMA-UE scores (p< 0.05). The left lateral frontal lobe (L_p9-46v), as a key node of the ventral attentional network, has an important role in goal-directed higher-order cognitive processes as well as conscious active control of planned behavior [29] . The present study showed that the FC between the left precentral gyrus prefrontalis (L_PEF) and area p9-46v showed a significant positive correlation with FMA-UE scores (p< 0.05). In a multicenter observational study [30] , researchers found that the recovery of motor function after stroke was closely related to higher cognitive function, especially in stroke patients with subacute stroke. In this clinical trial, stroke patients needed to filter key information among a large number of information prompts, master the motor imagery nodes, and focus on completing the MI, a process that suggests that, in addition to focusing on the recovery of stroke patients' motor function during the rehabilitation process, attention should also be paid to the enhancement of cognitive function in order to promote the comprehensive rehabilitation of functional impairment in the future. 3.4 EEG signal correlation analysis EEG signals are the most commonly employed means of neural recording when exploring MI. Due to its non-invasive characteristics and millisecond temporal resolution, the use of EEG signals in MI studies has been widely promoted in many fields, including neuroscience [31] . In particular, changes in the frequency band 8-30 HZ are closely related to MI or motor execution. In previous studies using EEG recordings of motor imagery (MI), when a subject performs an MI task, the energy corresponding to 8-30 HZ in the sensorimotor cortex of the contralateral brain undergoes a decrease, a phenomenon known as the ERD phenomenon. On the contrary, when the MI activity is terminated, the corresponding energy within this frequency band in the same cortical region increases significantly, and this phenomenon of increased energy is known as Event-relatedsynchronization (ERS) [32] . In this study, we utilized the interchannel connectivity relationship, i.e., the energy change between channels in a certain EEG frequency band in a group of stroke patients, to indirectly reflect the attenuation or enhancement of the ERD/ERS phenomenon between channels. The results showed that when stroke patients in group B performed the MI task with the affected hand, their EEG frequency band was significantly reduced in most of the inter-channel connectivity within 8-30 Hz (α and β bands), such as Pz-C3 in the α band, Pz-F3 in the α + β bands, P3-Fp1 in the α + β bands, and so on. This connectivity was mostly concentrated near the C1 channel in the left cerebral hemisphere, mostly involving channels P3, PO7, C3, F3, etc. Considering the EEG functional data processed by mirror image inversion, all the affected hands were considered as the right hand, and after a period of BCI training, the affected brain of stroke patients in group B would show enhanced ERD phenomenon when performing MI, which also verified the results of PSD analysis and brain topography, reflecting the enhancement of their MI ability. The enhancement of the connectivity of the C4-T7 in α-band suggests that ERS phenomenon is enhanced in the healthy brain of the stroke patient, and the stroke patient The interhemispheric EEG balance was altered and positively correlated with the FMA-LE score (P< 0.05). The following limitations exist in this study: (1) only the MRI and EEG levels were analyzed in Group B stroke patients in this study, and changes in neural remodeling brought about by routine rehabilitation training and their own recovery could not be excluded, which need to be verified by further controlled studies; (2) this study only focused on the FC enhancement aspect of the brain network, and the meaning of FC attenuation has not been explored; (3) the site of cerebral hemorrhage, the extent of the hemorrhage, and the degree of fiber bundle damage may be different in this study, which may affect the results of the analysis of the functional brain network characteristics; (4) some studies have shown that long-term BCI training can bring more objective benefits. In this study, we only recorded the changes in the rehabilitation of stroke patients with subacute cerebral hemorrhage for 1 month, which may underestimate the benefit of BCI training and produce ambiguous results in the analysis of the functional recovery mechanism, so it is still necessary to expand the sample size, and to develop a more reasonable training cycle and follow-up time for in-depth study. Conclusion BCI rehabilitation training can effectively promote the recovery of limb motor function in stroke patients with cerebral hemorrhage in the basal ganglia region in the subacute stage, but it has little effect on the enhancement of some indexes (e.g., cognitive function scores, modified Barthel index, and brunnstrom's staging).The probable mechanism of BCI rehabilitation training lies in the enhancement of the stroke patient's motor imagery ability as well as the promotion of the brain network activity. Declarations Ethics approval and consent to participate The study was approved by the Ethics Committee of Shanxi Provincial People's Hospital under the approval number (2022) Provincial Medical Science Lun Audit No. 95, and all patients participating in the trial signed an informed consent form. Consent for publication All authors have read and approved the final manuscript. We confirm that this work is original and has not been published elsewhere. Availability of data and materials All data supporting this study are included in the article. Competing interests The authors declare no competing interests. Funding This study was supported by the upper-level project of Natural Science Foundation of Shanxi Provincial Department of Education (No. 2022L184) and the Science and Technology Innovation Project for Higher Schools of Shanxi Provincial Department of Science and Technology (No. 202203021211060). Authors' contributions (I) Conception and design: Peili Cao, Hong Ling, Rui Cheng; (II) Administrative support: Rui Cheng, Gangli Zhang; (III) Provision of study materials or patients: JunHao Wang, Hao Guo; (IV) Collection and assembly of data: Xiang Zan, FuLong Zhang; (V) Data analysis and interpretation: Peili Cao, Hong Ling, Qihang Jin; (VI) Manuscript writing: All authors; (VII) Final approval of manuscript: All authors. Acknowledgements The authors would like to thank all volunteers who volunteered for the study. References Meng G, Huang Y, Yu Q, Ding Y, Wild D, Zhao Y, Liu X, Song M. Adopting Text Mining on Rehabilitation Therapy Repositioning for Stroke[J]. Front Neuroinformatics. 2019;13:17. Soleimani M, Ghazisaeedi M, Heydari S. The efficacy of virtual reality for upper limb rehabilitation in stroke stroke patients: a systematic review and meta-analysis[J]. BMC Med Inf Decis Mak. 2024;24(1):135. Wu Q, Ge Y, Ma D, Pang X, Cao Y, Zhang X, Pan Y, Zhang T, Dou W. Analysis of Prognostic Risk Factors Determining Poor Functional Recovery After Comprehensive Rehabilitation Including Motor-Imagery Brain-Computer Interface Training in Stroke Stroke patients: A Prospective Study[J]. Front Neurol. 2021;12:661816. Jiang Y, Yin J, Zhao B, Zhang Y, Peng T, Zhuang W, Wang S, Huang S, Zhong M, Zhang Y, Tang G, Shen B, Ou H, Zheng Y, Lin Q. Motor Imagery Brain-Computer Interface in Rehabilitation of Upper Limb Motor Dysfunction After Stroke[J]. J Visualized Experiments: JoVE, 2023(199). Yang S, Li R, Li H, Xu K, Shi Y, Wang Q, Yang T, Sun X. Exploring the Use of Brain-Computer Interfaces in Stroke Neurorehabilitation[J]. BioMed Research International, 2021, 2021: 9967348. Fan C, Yang B, Li X, Zan P. Temporal-frequency-phase feature classification using 3D-convolutional neural networks for motor imagery and movement[J]. Front NeuroSci. 2023;17:1250991. Luu TP, Brantley JA, Nakagome S, Zhu F, Contreras-Vidal JL. Electrocortical correlates of human level-ground, slope, and stair walking[J]. PLoS ONE. 2017;12(11):e0188500. Pichiorri F, Morone G, Petti M, Toppi J, Pisotta I, Molinari M, Paolucci S, Inghilleri M, Astolfi L, Cincotti F, Mattia D. Brain-computer interface boosts motor imagery practice during stroke recovery[J]. Ann Neurol. 2015;77(5):851–65. Nunes JD, Vourvopoulos A, Blanco-Mora DA, Jorge C, Fernandes JC, Bermudez I, Badia S, Figueiredo P. Brain activation by a VR-based motor imagery and observation task: An fMRI study[J]. PLoS ONE. 2023;18(9):e0291528. Zhang S, Xie L. Improving Attention Mechanism in Graph Neural Networks via Cardinality Preservation[J]. IJCAI: proceedings of the conference, 2020, 2020: 1395–1402. Ganepola T, Nagy Z, Ghosh A, Papadopoulo T, Alexander DC, Sereno MI. Using diffusion MRI to discriminate areas of cortical grey matter[J]. NeuroImage. 2018;182:456–68. Barria P, Pino A, Tovar N, Gomez-Vargas D, Baleta K, Díaz CAR, Múnera M, Cifuentes CA. BCI-Based Control for Ankle Exoskeleton T-FLEX: Comparison of Visual and Haptic Stimuli with Stroke Survivors[J]. Sensors. 2021;21(19):6431. Liaw N, Liebeskind D. Emerging therapies in acute ischemic stroke[J]. F1000Research, 2020, 9: F1000 Faculty Rev-546. Zhi JF, Liao QH, He YB, Xu WW, Zhu DW, Shao LH. Superior treatment efficacy of neuromodulation rehabilitation for upper limb recovery after stroke: a meta-analysis[J]. Expert Rev Neurother. 2022;22(10):875–88. Zhan G, Chen S, Ji Y, Xu Y, Song Z, Wang J, Niu L, Bin J, Kang X, Jia J. EEG-Based Brain Network Analysis of Chronic Stroke Stroke patients After BCI Rehabilitation Training[J]. Front Hum Neurosci. 2022;16:909610. Nojima I, Sugata H, Takeuchi H, Mima T. Brain-Computer Interface Training Based on Brain Activity Can Induce Motor Recovery in Stroke patients With Stroke: A Meta-Analysis[J]. Neurorehabilit Neural Repair. 2022;36(2):83–96. Nan W, Dias APB, Rosa AC. Neurofeedback Training for Cognitive and Motor Function Rehabilitation in Chronic Stroke: Two Case Reports[J]. Front Neurol. 2019;10:800. Yuan Z, Peng Y, Wang L, Song S, Chen S, Yang L, Liu H, Wang H, Shi G, Han C, Cammon JA, Zhang Y, Qiao J, Wang G. Effect of BCI-Controlled Pedaling Training System With Multiple Modalities of Feedback on Motor and Cognitive Function Rehabilitation of Early Subacute Stroke Stroke patients[J]. IEEE Trans neural Syst rehabilitation engineering: publication IEEE Eng Med Biology Soc. 2021;29:2569–77. Liu Y, Huang S, Xu W, Wang Z, Ming D. An fMRI study on the generalization of motor learning after brain actuated supernumerary robot training[J]. NPJ Sci Learn. 2024;9(1):80. Tam WK, Tong Kyu, Meng F, Gao S. A Minimal Set of Electrodes for Motor Imagery BCI to Control an Assistive Device in Chronic Stroke Subjects: A Multi-Session Study[J]. IEEE Trans Neural Syst Rehabil Eng. 2011;19(6):617–27. Miao Y, Chen S, Zhang X, Jin J, Xu R, Daly I, Jia J, Wang X, Cichocki A, Jung TP. BCI-Based Rehabilitation on the Stroke in Sequela Stage[J]. Neural Plasticity, 2020, 2020: 1–10. Pham LT, Chu SD, Hoang HV. Characteristic of 24-Hour Blood Pressure Dipping Patterns in Hypertensive Stroke Stroke patients[J]. Vasc Health Risk Manag. 2024;20:501–9. Wang J, Zhang J, Ye Y, Xu Q, Li Y, Feng S, Xiong X, Jian Z, Gu L. Peripheral Organ Injury After Stroke[J]. Front Immunol. 2022;13:901209. Ruan X, Song Z, Zhang J, Yu T, Chen J, Zhou T. Alterations of brain activity in stroke patients with alcohol use disorder: a resting-state fMRI study[J]. BMC Psychiatry. 2023;23(1):894. Phunruangsakao C, Achanccaray D, Bhattacharyya S, Izumi SI, Hayashibe M. Effects of visual-electrotactile stimulation feedback on brain functional connectivity during motor imagery practice[J]. Sci Rep. 2023;13(1):17752. Wu Q, Yue Z, Ge Y, Ma D, Yin H, Zhao H, Liu G, Wang J, Dou W, Pan Y. Brain Functional Networks Study of Subacute Stroke Stroke patients With Upper Limb Dysfunction After Comprehensive Rehabilitation Including BCI Training[J]. Front Neurol. 2019;10:1419. Kruschwitz JD, Waller L, List D, Wisniewski D, Ludwig VU, Korb F, Wolfensteller U, Goschke T, Walter H. Anticipating the good and the bad: A study on the neural correlates of bivalent emotion anticipation and their malleability via attentional deployment[J]. NeuroImage. 2018;183:553–64. Baker CM, Burks JD, Briggs RG, Sheets JR, Conner AK, Glenn CA, Sali G, McCoy TM, Battiste JD, O’Donoghue DL, Sughrue ME. A Connectomic Atlas of the Human Cerebrum-Chap. 3: The Motor, Premotor, and Sensory Cortices[J]. Operative Neurosurgery (Hagerstown, Md.), 2018, 15(suppl_1): S75-S121. Baker CM, Burks JD, Briggs RG, Conner AK, Glenn CA, Morgan JP, Stafford J, Sali G, McCoy TM, Battiste JD, O’Donoghue DL, Sughrue ME. A Connectomic Atlas of the Human Cerebrum-Chap. 2: The Lateral Frontal Lobe[J]. Operative Neurosurgery (Hagerstown, Md.), 2018, 15(suppl_1): S10-S74. Mancuso M, Iosa M, Abbruzzese L, Matano A, Coccia M, Baudo S, Benedetti A, Gambarelli C, Spaccavento S, Ambiveri G, Megna M, Tognetti P, Maietti A, Rinaldesi ML, Gamberini G, Varalta V, Morone G, Ciancarelli I, CogniReMo Study Group. The impact of cognitive function deficits and their recovery on functional outcome in subjects affected by ischemic subacute stroke: results from the Italian multicenter longitudinal study CogniReMo[J]. Eur J Phys Rehabil Med. 2023;59(3):284–93. Wei Z, Li H, Ma L, Li H. Emotion recognition based on microstate analysis from temporal and spatial patterns of electroencephalogram[J]. Front NeuroSci. 2024;18:1355512. Rimbert S, Lelarge J, Guerci P, Bidgoli SJ, Meistelman C, Cheron G, Cebolla Alvarez AM, Schmartz D. Detection of Motor Cerebral Activity After Median Nerve Stimulation During General Anesthesia (STIM-MOTANA): Protocol for a Prospective Interventional Study[J]. Volume 12. JMIR research protocols; 2023. p. e43870. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6791097","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":477254496,"identity":"b32ca64d-9416-412f-b9f5-7c84fcc460e5","order_by":0,"name":"Peili Cao","email":"","orcid":"","institution":"Shanxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Peili","middleName":"","lastName":"Cao","suffix":""},{"id":477254497,"identity":"4e939082-c660-40f3-b14b-6d18b6d98dfa","order_by":1,"name":"Hong Ling","email":"","orcid":"","institution":"Shanxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Hong","middleName":"","lastName":"Ling","suffix":""},{"id":477254498,"identity":"b2cfebee-5b4c-4071-b12b-da8ccf155b40","order_by":2,"name":"Hao Guo","email":"","orcid":"","institution":"Shanxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Hao","middleName":"","lastName":"Guo","suffix":""},{"id":477254499,"identity":"2d80fb8d-9d98-46a0-abb5-299df4eaa0da","order_by":3,"name":"JunHao Wang","email":"","orcid":"","institution":"Shanxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"JunHao","middleName":"","lastName":"Wang","suffix":""},{"id":477254500,"identity":"ba36b07c-2231-4e41-ac3c-9290c6d5d27c","order_by":4,"name":"Qihang Jin","email":"","orcid":"","institution":"University of Science and Technology of China","correspondingAuthor":false,"prefix":"","firstName":"Qihang","middleName":"","lastName":"Jin","suffix":""},{"id":477254501,"identity":"ca95cae9-5d39-48ec-ad07-99778874740a","order_by":5,"name":"Shijie Guo","email":"","orcid":"","institution":"Jinan University Affiliated First Hospital","correspondingAuthor":false,"prefix":"","firstName":"Shijie","middleName":"","lastName":"Guo","suffix":""},{"id":477254503,"identity":"7c49bc4e-b928-4497-97df-4623b7d2262a","order_by":6,"name":"Xiang Zan","email":"","orcid":"","institution":"Shanxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Xiang","middleName":"","lastName":"Zan","suffix":""},{"id":477254504,"identity":"b583b189-f48f-4b52-82d3-2146490e1b28","order_by":7,"name":"FuLong Zhang","email":"","orcid":"","institution":"Shanxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"FuLong","middleName":"","lastName":"Zhang","suffix":""},{"id":477254505,"identity":"116d032b-38cd-4324-b45b-f2d5ced21094","order_by":8,"name":"Rui Cheng","email":"","orcid":"","institution":"Xiang'an Hospital of Xiamen University, Xiamen University","correspondingAuthor":false,"prefix":"","firstName":"Rui","middleName":"","lastName":"Cheng","suffix":""},{"id":477254506,"identity":"e4b6e20a-b4bb-4ea9-841c-53026637ba00","order_by":9,"name":"Gangli Zhang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABEElEQVRIiWNgGAWjYBAC+xkwFjtjw4GP/2rk2NjbD+DVwgjXwszYeHAG2zFjPp4zCcRqYWA+zMPGnDhPwsEArxZm6eZnD7/8ssuTd2ZuOMzDw5beJsGQwPCjYhtOLWwyx8yNZfuSiw0PMzYcnCMhk9sm3XiAsefMbZxaeCQSzKQle5gTNzYDvf/GgC23TeZAAjNjG24tEhLp34Ba6iFaeBKY09kkEgzwajGQyDGT/PDjcOJ8ZqDDeA4wJxCjpUyaseF44gaQlpkNxwzbgIF8EJ9f7Gekb5P88ac6cX57++MPHxtq5OXb2w8++FGBWwsIMPO2Aa07gCRyAIdKOGD88YeBQb6BkLJRMApGwSgYsQAAZBRdMJZcVFsAAAAASUVORK5CYII=","orcid":"","institution":"Shanxi Medical University","correspondingAuthor":true,"prefix":"","firstName":"Gangli","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2025-05-31 13:08:09","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6791097/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6791097/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":85819368,"identity":"d4dc2dea-8c85-4084-a342-53665f7e885b","added_by":"auto","created_at":"2025-07-02 06:14:25","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":444383,"visible":true,"origin":"","legend":"\u003cp\u003eBCI Rehabilitation Training Actual Scene\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6791097/v1/eebcc649a84a722390574869.png"},{"id":85819367,"identity":"4059d7d5-759e-41cc-80bc-bc527143f4c4","added_by":"auto","created_at":"2025-07-02 06:14:25","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":115186,"visible":true,"origin":"","legend":"\u003cp\u003eThe dynamics of CA values are listed here for only four randomized stroke patients\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6791097/v1/9df75c3b2a117889526f5a3b.png"},{"id":85819382,"identity":"f58a900c-65a5-4a73-8b10-47338869c2d5","added_by":"auto","created_at":"2025-07-02 06:14:25","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":180322,"visible":true,"origin":"","legend":"\u003cp\u003eBrain regions with significant FC enhancement after training\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-6791097/v1/0e77efffb207ff79af13c254.png"},{"id":85819377,"identity":"2e5fad95-08ee-4e94-9f52-c784ae54a42b","added_by":"auto","created_at":"2025-07-02 06:14:25","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":160881,"visible":true,"origin":"","legend":"\u003cp\u003eConnectivity of inter-channel attenuation within the EEG 8-30 Hz (alpha and beta bands) after training (the connecting lines in the figure indicate the absolute value of the corresponding energy difference in the 8-30 Hz band between the two channels before and after training, the redder and thicker the connecting lines, the greater the difference; vice versa, the smaller the difference).\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-6791097/v1/eb09797517434411d80c9dc2.png"},{"id":85819369,"identity":"6e2bc427-cbb4-488e-927e-070f0b1e2815","added_by":"auto","created_at":"2025-07-02 06:14:25","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":141412,"visible":true,"origin":"","legend":"\u003cp\u003ePSD (blue lines represent right hand movement images, orange lines represent left hand movement images) and brain topography from electrodes C3 and C4 on the affected side of the brain during the basic training phase of Stroke patient 07's 1st (top 4 panels) and 25th (bottom 4 panels) session and from electrode C3 during the lifting training phase.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-6791097/v1/c1702e818a0c8ee9ef7aedcb.png"},{"id":96244510,"identity":"d8df92e5-667a-4a5e-b8f5-24d7619b02ff","added_by":"auto","created_at":"2025-11-19 07:18:43","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1542000,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6791097/v1/79371554-4a06-4042-bf57-63b70bbecbfb.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Effect of brain computer interface on limb motor function after intracerebral hemorrhage in basal ganglia and its rehabilitation mechanism","fulltext":[{"header":"Introduction","content":"\u003cp\u003eCerebral hemorrhage in the basal ganglia region is a significant cause of severe disability on a global scale. Despite attempts to intervene during the acute phase of cerebral hemorrhage, approximately 40% of stroke survivors exhibit varying degrees of functional impairment. Among these, loss of motor and sensory function in the upper extremities, including the hands and wrists, and cognitive dysfunction have a significant impact on stroke patients' quality of life \u003csup\u003e[1,2]\u003c/sup\u003e. Intracerebral hemorrhagic foci frequently result in the compression or destruction of brain tissue, which in turn can lead to damage to the functional areas of the brain. It has been demonstrated that minor lesions in functional areas are more challenging to recuperate from than dysfunction caused by more substantial lesions that do not affect functional areas\u003csup\u003e\u0026nbsp;[3]\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eIn the field of neuroscience, the recovery of upper extremity motor function after stroke is often based on neural remodeling or compensation, focusing on restoring function by reinforcing correct movement patterns. Traditional rehabilitation includes physical and occupational therapy, which facilitates recovery with the help of external physical stimulation or activities of daily living; or acupuncture, which stimulates acupuncture points to aid in the recovery of muscle function. These methods focus on training peripheral nerves or tissues to further promote central nervous reorganization and ultimately improve upper extremity motor function. In recent years, innovative interventions on motor rehabilitation have emerged with the goal of enhancing stroke patients' ability to perform daily living tasks, functional independence, and quality of life. These interventions include BCI training, transcranial direct current stimulation, functional electrical stimulation, deep brain stimulation, robot-assisted therapy, and virtual reality training, providing additional avenues for rehabilitation.\u003c/p\u003e\n\u003cp\u003eBrain-computer interface (BCI) is the process of transforming acquired brain biological signals through feature extraction into specific commands to drive external devices and promote the remodeling of brain function for stroke patients through feedback training of human-computer interaction \u003csup\u003e[4]\u003c/sup\u003e. Currently, BCI training is an effective intervention to promote the improvement of limb motor function and increase brain activation, and is dominated by BCI (MI-BCI) training based on motor imagery (MI) paradigm. Most studies \u003csup\u003e[5]\u003c/sup\u003e have shown the statistical significance and clinical relevance of rehabilitative BCI for the improvement of upper limb motor function recovery.\u003c/p\u003e\n\u003cp\u003eCurrently, the brain performs complex activities through the integration and coordination of information between different brain regions. And the study of brain functional networks plays an important role in revealing the encoding, processing and integration of information during cognitive processes \u003csup\u003e[6]\u003c/sup\u003e. The study of these brain network patterns helps to deeply understand the mechanism of functional activities of the brain and promote the development of neuroscience. However, how to accurately identify and characterize changes in brain network connectivity patterns during rehabilitation remains a major challenge for current research. The key method to study brain function is to analyze the neural signal activation in functional brain regions, and the main techniques include electroencephalography (EEG), magnetoencephalography, functional near-infrared spectroscopic imaging, and rest-state functional magnetic resonance (rs-fMRI) imaging \u003csup\u003e[7]\u003c/sup\u003e. A preliminary study \u003csup\u003e[8]\u003c/sup\u003e confirmed that the improvement of the function of the affected limb during BCI training in subacute stroke patients is closely related to the resting EEG oscillatory activity of the ipsilateral cerebral hemisphere and its connectivity changes. Some scholars \u003csup\u003e[9]\u003c/sup\u003e have also used this novel BCI task model to map the enhancement of functional connectivity between multiple brain regions in the frontal, occipital, and parietal lobes of the subjects from a functional magnetic resonance perspective.\u003c/p\u003e\n\u003cp\u003eThe aim of this study was to evaluate the rehabilitation effect of BCI training on stroke patients with cerebral hemorrhage in the basal ganglia region accompanied by limb motor dysfunction, to develop a functional analysis based on MRI imaging and EEG data, and to explore in-depth the potential neuroplasticity mechanisms in the process of recovery of their limb motor function.\u003c/p\u003e"},{"header":"Objects and Methods","content":"\u003cp\u003e1.1 Study subjects\u003c/p\u003e\n\u003cp\u003eStroke patients with cerebral hemorrhage in the basal ganglia region in the subacute stage who were hospitalized in Shanxi Provincial People's Hospital and Shanxi Provincial Acupuncture Hospital from December 2022 to October 2024, accompanied by motor dysfunction of the upper limbs, were selected for this study. Clinical indicators such as stroke patient's name, gender, age, past history, disease duration, hemorrhage site and treatment modality were recorded. The study was approved by the Ethics Committee of Shanxi Provincial People's Hospital with the approval number (2022) Provincial Medical Science Lun Audit No. 95, and all stroke patients participating in the trial signed an informed consent form.\u003c/p\u003e\n\u003cp\u003eInclusion criteria: (1) Diagnosed with cerebral hemorrhage in the basal ganglia region by MRI and CT, accompanied by limb motor dysfunction; (2) First onset of the disease within 1\u0026ndash;6 months; (3) Defined as age\u0026thinsp;\u0026ge;\u0026thinsp;18, gender is not limited; (4) Simplified Mental State Score\u0026thinsp;\u0026ge;\u0026thinsp;18, can understand and implement the training of the basic prompts and commands; (5) Due to the dysfunction caused by the lesion of the functional area, need to carry out rehabilitation treatment. stroke patients; (6) the subjects themselves or their legal representatives agreed to participate in the trial and signed a written informed consent form.\u003c/p\u003e\n\u003cp\u003eExclusion criteria: (1) the presence of serious cognitive, mental and psychological disorders; (2) serious heart, liver, kidney, lung and other important organ failure; (3) with upper limb defects, open wounds, deformities, or with severe pain; (4) suffering from severe vision, hearing impairment, claustrophobia, etc.; (5) the body contains metals that can not be cooperated with the magnetic resonance examination; (6) the subject is participating in other clinical trials at the same time.\u003c/p\u003e\n\u003cp\u003e1.2 Experimental grouping and BCI training\u003c/p\u003e\n\u003cp\u003eA total of 44 stroke patients with subacute stage basal ganglia cerebral hemorrhage accompanied by limb movement disorders were included in this study, and were randomly assigned to the conventional rehabilitation therapy group (Group C) and the BCI group (Group B) according to whether or not BCI training was used in the training content, with 22 stroke patients in each group.\u003c/p\u003e\n\u003cp\u003eGroup C: stroke patients received traditional rehabilitation training, including occupational therapy, acupuncture, exercise therapy and other treatments, 5 times per week for 5 weeks; Group B: stroke patients received traditional rehabilitation training and BCI training. In addition to the traditional rehabilitation training, a professional with 3 or more years of experience conducted 25 BCI training sessions of 25 minutes/day, about 5 times per week for 5 weeks for Group B stroke patients.\u003c/p\u003e\n\u003cp\u003eBCI training uses non-invasive EEG to recognize MI-based motor task signals and convert them into commands, controlling the exoskeleton hand to guide the stroke patient to perform grasping and opening training of the affected hand. It consists of two main phases, basic and elevation training, during which the stroke patient is told to visualize the specific activities of the hand, and the exoskeleton hand provides support to help the stroke patient complete the grip/open hand task. The system analyzes and processes the stroke patient's EEG data during motor imagery collected by a 16-lead EEG cap, extracts EEG signal features as the basis for task recognition, and if the recognition is successful, the exoskeleton device connected to the computer guides the stroke patient's hand movement, while the computer monitor prompts for success to be given as feedback, and the cycle repeats until the end of the training. Each training completed 55 MI tasks, the screen randomly displays the video of holding the right and left hands and prompts, to avoid coughing, body shaking and other strenuous movements. The training scenario is shown in Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e1.3 Clinical rehabilitation assessment\u003c/p\u003e\n\u003cp\u003eFugl-Meyer assessment (FMA) scale, including upper and lower extremity motor function scores (FMA-UE, FMA-LE), could assess stroke patients' limb motor function; Mini-mental State Examination (MMSE) was used to check stroke patients' cognitive function (total score 30 points, 21\u0026ndash;26 points for mild cognitive deficits, 10\u0026ndash;20 points for moderate cognitive deficits, 0\u0026ndash;9 points for severe cognitive deficits); modified Bartlett's scale was used to assess stroke patients' cognitive function. Stroke patients' cognitive function can be examined (total score of 30 points, 21\u0026ndash;26 points are mild cognitive deficits, 10\u0026ndash;20 points are moderate cognitive deficits, and 0\u0026ndash;9 points are severe cognitive deficits); the modified Barthel index can assess stroke patients' ability in daily life (\u0026ge;\u0026thinsp;60 points: basically taking care of themselves; 41\u0026ndash;59 points: moderate dysfunction, needing other people's help; 21\u0026ndash;40 points: severe dysfunction, obvious dependence); brunnstroms index can assess stroke patients' ability in daily life. The brunnstrom staging assesses the different stages of rehabilitation of stroke patients and consists of 6 clinical stages. Specific criteria for the above scales are detailed in the Appendix. All rehabilitation assessments were performed by 2 trained clinicians, who independently evaluated the scales before and after the entire training phase, and the results were taken after agreement.\u003c/p\u003e\n\u003cp\u003eOnline classification accuracy (CA), which is the classification accuracy of the stroke patient when performing the task, is used to measure each training performance. The first stage is offline data collection without real-time feedback. After the collection, the current offline accuracy was obtained by using ten-fold cross-validation. The whole data of the first stage was used as the training set to obtain the optimal classifier to predict the EEG data of the second stage, and the CA value of the MI task of the second stage was also obtained. A larger CA value for this stroke patient indicates a higher degree of human-computer interaction during training \u003csup\u003e[\u003cspan class=\"CitationRef\"\u003e10\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e1.4 MRI data processing\u003c/p\u003e\n\u003cp\u003eBefore and after the training, a GE 3.0T MRI scanner was used to examine the head of stroke patients in group B. The magnetic field was 3 T. The stroke patients were placed in a supine position with their heads fixed, and the acquisition of high-resolution MR images, including T1 phase and rs-fMRI phase images, was accomplished using a standard 32-channel head coil. During the rs-fMRI scanning process, stroke patients were required to close their eyes and remain awake without specific thinking. Finally, the acquired MR data were further constructed for connectomics image analysis.\u003c/p\u003e\n\u003cp\u003eBased on the multimodal segmentation atlas of the Human Connectome Project \u003csup\u003e[\u003cspan class=\"CitationRef\"\u003e11\u003c/span\u003e]\u003c/sup\u003e, an \u0026ldquo;individualized re-segmentation\u0026rdquo; technique was used. Twenty-one brain regions related to the sensorimotor cortex were selected as regions of interest (ROIs), which encompassed the sensorimotor network and premotor regions, and also corresponded to Brodmann areas 1\u0026ndash;8, and were mainly composed of primary sensory cortex (areas 1, 2, 3a, 3b), primary motor cortex (area 4), paracentral lobular area (areas 5L, 5m, 5mv), supplementary motor area (areas 5L, 5m, 5mv), and supplementary motor area (areas 5L, 5m, 5mv). 5L, 5m, 5mv), supplementary motor areas (areas 6ma, 6mp, SFL), premotor areas (areas 6a, 6d, FEF, 55b, PEF, 6r, 6v), lateral part of parietal lobe (area AIP), and part of anterior cingulate gyrus (areas 24dd, 24dv). In this study, we will analyze the changes in functional connectivity (FC) between the stroke patients' ROI and the whole brain cortex, specifically by calculating the FC coefficients between the ROI and each voxel of the whole brain, and then statistically analyzing them.\u003c/p\u003e\n\u003cp\u003eFinally, the relationship between clinical scores and FC will be analyzed using NumPy 1.12.1 and Scipy 0.19.0 software.\u003c/p\u003e\n\u003cp\u003e1.5 EEG signal processing\u003c/p\u003e\n\u003cp\u003eMI analysis is the most commonly used method to detect motor intentions based on the electrical activity of the motor cortex \u003csup\u003e[\u003cspan class=\"CitationRef\"\u003e12\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e1.5.1 Data preprocessing\u003c/p\u003e\n\u003cp\u003eEach training contained two phases, the first phase contained 30 imagery trainings and the second phase contained 58 imagery trainings. The imagery data for the entire training was band-pass filtered from 5-35Hz, using an 8th order Butterworth filter. Subsequently, the filtered data were sliced, and the data segments (epochs) 0\u0026ndash;4 seconds after cue appearance were selected for each imagery training. The standard deviation of each lead was calculated for each Epoch, and if the standard deviation of a single lead exceeded a threshold (less than 1 \u0026micro;V, or greater than 150 \u0026micro;V), that lead was removed. If more than 30% of the epoch in a single phase was removed, the data for the entire Session was discarded.\u003c/p\u003e\n\u003cp\u003e1.5.2 Connectivity analysis\u003c/p\u003e\n\u003cp\u003eIn this study, we used imaginary coherence (iCoh) as the EEG connectivity analysis method to analyze the preprocessed EEG signals and calculate the coherence between the leads in order to assess the brain connectivity of stroke patients in the early and late stages of treatment. iCoh method is based on spectral coherence, which employs the imaginary part in the result of the spectral coherence as a way of reducing the volumetric conduction effect generated by the artifacts and better reveal interactions between brain regions.\u003c/p\u003e\n\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\n\u003cdiv id=\"FileID_Equa\" class=\"mathdisplay\"\u003e$$\\:\\text{i}\\text{C}\\text{o}\\text{h}\\left(\\text{e}\\text{e}\\text{g}\\right)=\\frac{{\\left|\\text{i}\\text{m}\\text{a}\\text{g}\\left({\\text{S}}_{\\text{x}\\text{y}}^{}\\right(\\text{e}\\text{e}\\text{g}\\left)\\right)\\right|}^{2}}{{\\text{S}}_{\\text{x}\\text{x}}^{}\\left(\\text{e}\\text{e}\\text{g}\\right){\\text{S}}_{\\text{y}\\text{y}}^{}\\left(\\text{e}\\text{e}\\text{g}\\right)}$$\u003c/div\u003e\n\u003c/div\u003e\n\u003cp\u003ewhere, is the cross-frequency spectral density on electrodes X and Y, and, are the self-frequency spectral densities of electrodes X and Y, respectively.\u003c/p\u003e\n\u003cp\u003eThe connectivity data of all stroke patients in the experimental group before and after 3 effective trainings, stage 2, were selected and averaged for significance analysis. If the effective training was less than 3 for the first or second 6 treatments in a stroke patient, the data for that stroke patient were discarded. Significance analysis was performed using a two-tailed paired t test. To test the relationship between EEG connectivity and clinical scores, Spearman correlation analysis was used in this study.\u003c/p\u003e\n\u003cp\u003e1.5.3 Power spectral density and brain topography calculation and analysis\u003c/p\u003e\n\u003cp\u003eThe training EEG signals of the stroke patients were collected, and the EEG signals of the central motor cortex region were extracted within 4 seconds of the motor imagery, and the average energy value in the range of 8-30Hz was calculated to obtain the power spectral density (PSD), which can reflect the distribution of the power of the EEG signals of the central motor region of the bilateral cerebral hemispheres in the frequency domain. At the same time, the 8-30Hz EEG signals of the whole leads were collected, and the average energy values of each brain region at different stages were calculated, and the brain topography was drawn to observe the energy changes in the C3 and C4 regions and to analyze whether there was event-related desynchronization (ERD) phenomenon.\u003c/p\u003e\n\u003cp\u003e1.6 Statistical processing\u003c/p\u003e\n\u003cp\u003eSPSS 26.0 software was used for statistical analysis. Normally distributed measurements were expressed as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation (x\u0026thinsp;\u0026plusmn;\u0026thinsp;s), two-sample independent t-test was used for comparison between the two groups, and paired t-test was used before and after the training; rank information was expressed as quartiles [M(Q1, Q3)], and rank-sum test was used for comparison between the two groups; categorical information was expressed as relative numbers, and the X2 test or the Fisher's exact probability method; correlation analysis was performed by Spearman's analysis; two-sided test was used, and P\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was regarded as statistically significant difference.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e2.1 Comparison of general information of stroke patients in two groups\u003c/p\u003e\n\u003cp\u003eThere is no statistically significant difference between the general information of stroke patients in the two groups (P\u0026gt;0.05). See Table 1.\u003c/p\u003e\n\u003cp\u003eTable 1Stroke patients\u0026rsquo; Characteristics\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"578\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003eCharacteristic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003eBCI（n=22）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003eContrl（n=22）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 96px;\"\u003e\n \u003cp\u003eT/X\u003csup\u003e2\u003c/sup\u003e/Z-value\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003eP-value\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003eAge, (year）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e47.14\u0026plusmn;10.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e50.32\u0026plusmn;7.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 96px;\"\u003e\n \u003cp\u003e1.124\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.268\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003eDuration of disease (months)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e2.89\u0026plusmn;1.70\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e3.00\u0026plusmn;1.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 96px;\"\u003e\n \u003cp\u003e0.218\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.828\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003eHypertension\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 96px;\"\u003e\n \u003cp\u003e0.101\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.750\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003eyes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 75px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003eno\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003eDiabetes Mellitus\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 96px;\"\u003e\n \u003cp\u003e0.820\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.365\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003eyes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 75px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003eno\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003eSex\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 96px;\"\u003e\n \u003cp\u003e0.121\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.728\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003emale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 75px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003efemale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003eHemorrhage site\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 96px;\"\u003e\n \u003cp\u003e0.364\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.546\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003eright basal ganglia hemorrhage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 75px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003eLeft basal ganglia hemorrhage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003eTreatment modality\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 96px;\"\u003e\n \u003cp\u003e0.863\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.353\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003eSurgical treatment\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 75px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 192px;\"\u003e\n \u003cp\u003eConservative treatment\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2.2 Comparison of stroke patients\u0026apos; clinical related scores before and after training\u003c/p\u003e\n\u003cp\u003eThere was no statistically significant difference in the comparison of clinical related scores between the two groups before training (P\u0026gt;0.05). After training, compared with group C, the FMA-UE and FMA-LE scores of stroke patients in group B were significantly higher than before (P\u0026lt;0.05), and the differences between the two groups of stroke patients in comparison of modified Barthel index, MMSE score, and brunnstrom staging were still not statistically significant (P\u0026gt;0.05), as shown in Tables 2、3, and 4.\u003c/p\u003e\n\u003cp\u003eTable\u0026nbsp;2 Comparison of clinical related scores between the two groups of stroke patients before training\u003c/p\u003e\n\u003cdiv align=\"center\"\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"100%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 27px;\"\u003e\n \u003cp\u003eVariables\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20px;\"\u003e\n \u003cp\u003eBCI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 21px;\"\u003e\n \u003cp\u003eContrl\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17px;\"\u003e\n \u003cp\u003eT/Z-value\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 13px;\"\u003e\n \u003cp\u003eP-value\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 27px;\"\u003e\n \u003cp\u003eFMA-UE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20px;\"\u003e\n \u003cp\u003e11.91\u0026plusmn;5.53\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 21px;\"\u003e\n \u003cp\u003e11.09\u0026plusmn;7.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17px;\"\u003e\n \u003cp\u003e-0.419\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 13px;\"\u003e\n \u003cp\u003e0.677\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 27px;\"\u003e\n \u003cp\u003eFMA-LE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20px;\"\u003e\n \u003cp\u003e14.41\u0026plusmn;5.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 21px;\"\u003e\n \u003cp\u003e13.45\u0026plusmn;5.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17px;\"\u003e\n \u003cp\u003e-0.560\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 13px;\"\u003e\n \u003cp\u003e0.579\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 27px;\"\u003e\n \u003cp\u003eMMSE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20px;\"\u003e\n \u003cp\u003e22.36\u0026plusmn;3.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 21px;\"\u003e\n \u003cp\u003e23.41\u0026plusmn;4.63\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17px;\"\u003e\n \u003cp\u003e0.830\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 13px;\"\u003e\n \u003cp\u003e0.411\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 27px;\"\u003e\n \u003cp\u003eMBI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20px;\"\u003e\n \u003cp\u003e50.50\u0026plusmn;16.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 21px;\"\u003e\n \u003cp\u003e54.32\u0026plusmn;21.78\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17px;\"\u003e\n \u003cp\u003e0.656\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 13px;\"\u003e\n \u003cp\u003e0.516\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 27px;\"\u003e\n \u003cp\u003eBrunnstrom classification of upper limb\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20px;\"\u003e\n \u003cp\u003e2(1，2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 21px;\"\u003e\n \u003cp\u003e2(1.75，3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17px;\"\u003e\n \u003cp\u003e0.479\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 13px;\"\u003e\n \u003cp\u003e0.489\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 27px;\"\u003e\n \u003cp\u003eBrunnstrom classification of hand\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20px;\"\u003e\n \u003cp\u003e1(1，2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 21px;\"\u003e\n \u003cp\u003e1.5(1，2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17px;\"\u003e\n \u003cp\u003e1.158\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 13px;\"\u003e\n \u003cp\u003e0.282\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 27px;\"\u003e\n \u003cp\u003eBrunnstrom classification of lower limb\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20px;\"\u003e\n \u003cp\u003e3(2，3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 21px;\"\u003e\n \u003cp\u003e3(2，3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17px;\"\u003e\n \u003cp\u003e0.144\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 13px;\"\u003e\n \u003cp\u003e0.704\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003cp\u003eTable\u0026nbsp;3 Comparison of clinical related scores between the two groups of stroke patients after training\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"100%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 27px;\"\u003e\n \u003cp\u003eVariables\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22px;\"\u003e\n \u003cp\u003eBCI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20px;\"\u003e\n \u003cp\u003eContrl\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 14px;\"\u003e\n \u003cp\u003eT/Z-value\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 13px;\"\u003e\n \u003cp\u003eP-value\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 27px;\"\u003e\n \u003cp\u003eFMA-UE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22px;\"\u003e\n \u003cp\u003e21.41\u0026plusmn;9.83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20px;\"\u003e\n \u003cp\u003e13.50\u0026plusmn;7.20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 14px;\"\u003e\n \u003cp\u003e-3.044\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 13px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 27px;\"\u003e\n \u003cp\u003eFMA-LE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22px;\"\u003e\n \u003cp\u003e20.91\u0026plusmn;5.33\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20px;\"\u003e\n \u003cp\u003e17.14\u0026plusmn;5.88\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 14px;\"\u003e\n \u003cp\u003e-2.230\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 13px;\"\u003e\n \u003cp\u003e0.031\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 27px;\"\u003e\n \u003cp\u003eMMSE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22px;\"\u003e\n \u003cp\u003e25.36\u0026plusmn;3.51\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20px;\"\u003e\n \u003cp\u003e24.23\u0026plusmn;4.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 14px;\"\u003e\n \u003cp\u003e-0.940\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 13px;\"\u003e\n \u003cp\u003e0.353\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 27px;\"\u003e\n \u003cp\u003eMBI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22px;\"\u003e\n \u003cp\u003e66.36\u0026plusmn;13.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20px;\"\u003e\n \u003cp\u003e64.09\u0026plusmn;20.62\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 14px;\"\u003e\n \u003cp\u003e-0.429\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 13px;\"\u003e\n \u003cp\u003e0.670\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 27px;\"\u003e\n \u003cp\u003eBrunnstrom classification of upper limb\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22px;\"\u003e\n \u003cp\u003e2.5(2，3.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20px;\"\u003e\n \u003cp\u003e2.5(2，3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 14px;\"\u003e\n \u003cp\u003e0.342\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 13px;\"\u003e\n \u003cp\u003e0.559\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 27px;\"\u003e\n \u003cp\u003eBrunnstrom classification of hand\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22px;\"\u003e\n \u003cp\u003e2(1.75，2.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20px;\"\u003e\n \u003cp\u003e2(1，2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 14px;\"\u003e\n \u003cp\u003e2.111\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 13px;\"\u003e\n \u003cp\u003e0.146\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 27px;\"\u003e\n \u003cp\u003eBrunnstrom classification of lower limb\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22px;\"\u003e\n \u003cp\u003e3(3，4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20px;\"\u003e\n \u003cp\u003e3(3，4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 14px;\"\u003e\n \u003cp\u003e1.238\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 13px;\"\u003e\n \u003cp\u003e0.266\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable\u0026nbsp;4 Comparison of clinically relevant scores before and after training in the BCI group of stroke patients\u003c/p\u003e\n\u003cdiv align=\"center\"\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"99%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 27px;\"\u003e\n \u003cp\u003eVariables\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 20px;\"\u003e\n \u003cp\u003epre\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22px;\"\u003e\n \u003cp\u003epost\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 14px;\"\u003e\n \u003cp\u003et/Z值\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15px;\"\u003e\n \u003cp\u003eP-value\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 27px;\"\u003e\n \u003cp\u003eFMA-UE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 20px;\"\u003e\n \u003cp\u003e11.91\u0026plusmn;5.53\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22px;\"\u003e\n \u003cp\u003e21.41\u0026plusmn;9.83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 14px;\"\u003e\n \u003cp\u003e-5.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 27px;\"\u003e\n \u003cp\u003eFMA-LE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 20px;\"\u003e\n \u003cp\u003e14.41\u0026plusmn;5.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22px;\"\u003e\n \u003cp\u003e20.91\u0026plusmn;5.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 14px;\"\u003e\n \u003cp\u003e-6.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 27px;\"\u003e\n \u003cp\u003eMMSE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 20px;\"\u003e\n \u003cp\u003e22.36\u0026plusmn;3.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22px;\"\u003e\n \u003cp\u003e25.36\u0026plusmn;3.51\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 14px;\"\u003e\n \u003cp\u003e-6.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 27px;\"\u003e\n \u003cp\u003eMBI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 20px;\"\u003e\n \u003cp\u003e50.50\u0026plusmn;16.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22px;\"\u003e\n \u003cp\u003e66.36\u0026plusmn;13.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 14px;\"\u003e\n \u003cp\u003e-7.30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 27px;\"\u003e\n \u003cp\u003eBrunnstrom classification of upper limb\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 20px;\"\u003e\n \u003cp\u003e2(1，2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22px;\"\u003e\n \u003cp\u003e2.5(2，3.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 14px;\"\u003e\n \u003cp\u003e-2.951\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15px;\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 27px;\"\u003e\n \u003cp\u003eBrunnstrom classification of hand\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 20px;\"\u003e\n \u003cp\u003e1(1，2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22px;\"\u003e\n \u003cp\u003e2(1.75，2.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 14px;\"\u003e\n \u003cp\u003e-3.276\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15px;\"\u003e\n \u003cp\u003e0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 27px;\"\u003e\n \u003cp\u003eBrunnstrom classification of lower limb\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 20px;\"\u003e\n \u003cp\u003e3(2，3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22px;\"\u003e\n \u003cp\u003e3(3，4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 14px;\"\u003e\n \u003cp\u003e-3.314\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 15px;\"\u003e\n \u003cp\u003e0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e2.3 Analysis of training accuracy of stroke patients in Group B\u003c/p\u003e\n\u003cp\u003eOnly six stroke patients in Group B could maintain relatively high CA values, and their average CA values exceeded the 70% threshold criterion, and most of the remaining stroke patients showed more obvious fluctuations in CA values and even a downward trend in the course of BCI training, see Figure 2.\u003c/p\u003e\n\u003cp\u003e2.4 MRI-related analysis of group B stroke patients\u003c/p\u003e\n\u003cp\u003e2.4.1 Rehabilitation assessment and rs-MRI analysis\u003c/p\u003e\n\u003cp\u003eExcluding 4 stroke patients with important data missing in group B, a total of 18 stroke patients were included in this study who successfully completed MRI scanning, and their FMA-UE scores were significantly improved compared with the previous ones (P\u0026lt;0.05), see Table 5; rs-MRI analysis showed 84 enhanced FCs of the stroke patients (P\u0026lt; 0.05), see Figure 3 for details.\u003c/p\u003e\n\u003cp\u003eTable 5 Comparison of FMA scores before and after training in 18 stroke patients in the BCI group\u003c/p\u003e\n\u003cdiv align=\"center\"\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"103%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 18px;\"\u003e\n \u003cp\u003eVariables\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 23px;\"\u003e\n \u003cp\u003epre\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24px;\"\u003e\n \u003cp\u003epost\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16px;\"\u003e\n \u003cp\u003eT/Z-value\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16px;\"\u003e\n \u003cp\u003eP-value\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 18px;\"\u003e\n \u003cp\u003eFMA-UE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 23px;\"\u003e\n \u003cp\u003e11.56\u0026plusmn;4.902\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24px;\"\u003e\n \u003cp\u003e20.89\u0026plusmn;10.203\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16px;\"\u003e\n \u003cp\u003e-4.982\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 18px;\"\u003e\n \u003cp\u003eFMA-LE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 23px;\"\u003e\n \u003cp\u003e14.17\u0026plusmn;5.544\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24px;\"\u003e\n \u003cp\u003e21.00\u0026plusmn;5.099\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16px;\"\u003e\n \u003cp\u003e-6.349\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 18px;\"\u003e\n \u003cp\u003eMBI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 23px;\"\u003e\n \u003cp\u003e50.61\u0026plusmn;17.493\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24px;\"\u003e\n \u003cp\u003e65.72\u0026plusmn;14.323\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16px;\"\u003e\n \u003cp\u003e-5.847\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e2.4.2 Correlation analysis between clinical scores and FC\u003c/p\u003e\n\u003cp\u003eThe pre-post comparison of the FMA scores of the 18 stroke patients scanned by MRI in Group B remained statistically significant (P \u0026lt; 0.05), as shown in Tables 5. The correlation analysis of the FC with the FMA-UE scores showed that: the lateral aspect of the left postcentral gyrus (L_1) and the anterior-inferior cortex of the left precentral gyrus (L_6v), the anterior-parietal aspect of the left postcentral gyrus (L_3b) and the anterior-inferior cortex of the left precentral gyrus (L_6v) , FC between the anterior parietal side of the left postcentral gyrus (L_3b) and the prefrontal side of the left precentral gyrus (L_PEF), between the anterolateral side of the left precentral gyrus (L_55b) and the anterior subcortex of the left precentral gyrus (L_6v), between the anteroparietal side of the left precentral gyrus (L_PEF) and the posterolateral side of the left prefrontal lobe (L_p9-46v), and all of them were positively correlated with the clinical scores (P\u0026lt; 0.05), showing that the stroke patients\u0026apos; improvement in limb motor function was associated with elevated FC within the brain network, as shown in Tables 6.\u003c/p\u003e\n\u003cp\u003eTables 6\u0026nbsp;Correlation analysis between FC and clinical scores\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"568\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003eFC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003eclinical scores\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003er-value\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003eL_1-L_6v\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003eFMA-UE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003e0.478\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003e0.045\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003eL_3b-L_6v\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003eFMA-UE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003e0.547\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003e0.019\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003eL_3b-L_PEF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003eFMA-UE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003e0.576\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003e0.012\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003eL_55b-L_6v\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003eFMA-UE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003e0.514\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003e0.029\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003eL_PEF-L_p9-46v\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003eFMA-UE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003e0.527\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 142px;\"\u003e\n \u003cp\u003e0.025\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2.5 Network analysis of EEG function in group B stroke patients\u003c/p\u003e\n\u003cp\u003e2.5.1 Correlation analysis between clinical scores and connectivity between EEG channels\u003c/p\u003e\n\u003cp\u003eData from 19 stroke patients in Group B were eligible, of which 8 had left-handedness on the affected side and 11 had right-handedness on the affected side. The EEG functional data of the left affected hand was mirror reversed (EEG viewed as symmetrical distribution of left and right brain functions), transformed to the right hand, and then analyzed statistically. The results showed that when the affected hand performed motor imagery, the connectivity between most of the channels in the EEG frequency band within 8-30 Hz (\u0026alpha; and \u0026beta; bands) was significantly weakened (P\u0026lt; 0.05); in addition, the connectivity between C4-T7 and T8-T7 in the \u0026alpha; band was enhanced, and it was positively correlated with the FMA-LE scores (P\u0026lt; 0.05), as shown in Figures 4 and Tables 7.\u003c/p\u003e\n\u003cp\u003eTables 7\u0026nbsp;Correlation analysis between connectivity and clinical scores\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"574\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 156px;\"\u003e\n \u003cp\u003econnectivity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 128px;\"\u003e\n \u003cp\u003eclinical scores\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003er-value\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 156px;\"\u003e\n \u003cp\u003e\u0026alpha;\u0026nbsp;bands\u0026rsquo;sC4-T7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 128px;\"\u003e\n \u003cp\u003eFMA-LE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003e0.536\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 138px;\"\u003e\n \u003cp\u003e0.018\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 156px;\"\u003e\n \u003cp\u003e\u0026alpha;\u0026nbsp;bands\u0026rsquo;sT8-T7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 128px;\"\u003e\n \u003cp\u003eFMA-LE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 152px;\"\u003e\n \u003cp\u003e0.506\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 138px;\"\u003e\n \u003cp\u003e0.027\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\u0026nbsp;\u003cp\u003e2.5.2 PSD and brain topography\u003c/p\u003e\n\u003cp\u003eBefore and after training, it was observed that 18 stroke patients at C3 and C4 electrodes showed differences in the corresponding energies in the 8-30 HZ band of the EEG of the bilateral hemispheres during MI, which indicated the different abilities of the bilateral hemispheres to perform MI. 15 stroke patients after BCI training showed a flattening of the PSD curves of the central motor area of the brain of the affected side in this band and a decrease in the energy values; the topography of the brain showed that the cortical cortex of the affected side showed obvious ERD phenomena or phenomena that were further enhanced than before and had a tendency to converge to the central and peripheral areas of the brain; the remaining stroke patients showed a relative enhancement of the affected side of the brain. The topography of the brain showed that the affected side of the hemispheric cortex showed obvious ERD phenomenon or the phenomenon was further enhanced compared with the previous one, and there was a tendency to converge towards the central brain area and its surroundings; the remaining stroke patients could see the ERD phenomenon of the affected side of the brain relatively enhanced, or the ERS phenomenon of the healthy side of the brain relatively weakened, and so on. For example, in stroke patient 07, the PSD map showed that there was an opposite difference in EEG energy at electrodes C3 and C4 before and after training; the topographic brain map showed that the ERD phenomenon of the left hemisphere became more concentrated, mainly around electrode C3, and obvious ERS phenomenon was observed in the right cerebral cortex, as shown in Figures 5.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003e3.1 Stroke patient characteristics and training and rehabilitation efficacy study\u003c/p\u003e\n\u003cp\u003eCerebral hemorrhagic stroke is one of the most destructive diseases in neurosurgery, which not only directly leads to local brain tissue damage and irreversible neurological deficits\u003csup\u003e\u0026nbsp;[13]\u003c/sup\u003e, but also leaves many serious complications. Among them, cerebral hemorrhage in the basal ganglia region with limb motor function impairment is one of the core challenges in clinical rehabilitation, which is often manifested as hemiparesis, dystonia abnormalities, muscle atrophy and other motor deficits, or even permanent disability. Medication and rehabilitation are still the main clinical treatments for limb motor function impairment after stroke \u003csup\u003e[14]\u003c/sup\u003e, but the rehabilitation efficacy is still poor. And BCI technology is to build a direct communication bridge between the brain and the external environment, which can realize the information interaction between the brain and the outside world. Currently, MI-BCI-based combined Neurofeedback (NF) training, as an innovative rehabilitation treatment method, can help stroke patients to complete the recovery and remodeling of brain function by processing and converting brain biological signals into specific commands, driving external devices, and through the feedback mechanism of human-computer interaction \u003csup\u003e[15]\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eThe efficacy of MI-BCI-based training in limb rehabilitation for stroke patients with functional brain injury has now been demonstrated.1 meta-analysis in 2021 \u003csup\u003e[16]\u003c/sup\u003e indicated that neurofeedback intervention with BCI was significantly more effective than the control group for paralyzed upper limbs, with a standardized mean difference in FMA-UE scores of 0.48 (95% CI 0.16-0.80), demonstrating that BCI-based training is superior to traditional interventions in stroke patients' upper limb motor recovery is superior to traditional interventions. On this basis, some scholars have done further research, Nan et al \u003csup\u003e[17]\u003c/sup\u003e conducted BCI intervention for chronic stroke patients who had received poor results from traditional therapies, and the stroke patients' limb movement, language and cognitive function improved in various aspects; Yuan et al [18] designed a randomized controlled clinical trial to provide real-time feedback based on the attention index of stroke patients during BCI pedaling training. A randomized controlled clinical trial was designed by Yuan et al\u003csup\u003e[18]\u003c/sup\u003e to provide real-time feedback in multiple forms to enhance stroke patient participation in training, and the results showed that the lower limb motor function scores of stroke patients with BCI intervention were significantly improved.\u003c/p\u003e\n\u003cp\u003eThe MI-BCI technique mainly instructs stroke patients to perform repetitive reinforcement training to simulate their own brain's potential awareness of a certain movement to achieve functional recovery; performing MI is also a cognitive exercise process, which imagines a part of your body's movement without actually moving that part of your body to approach the actual movement execution; therefore, performing MI training has a slightly higher demand on the stroke patient's initial cognitive function, and a smaller demand on the stroke patient's residual motor function. function is less demanding. The stroke patients with cerebral hemorrhage in the basal ganglia region included in this study, all of whom had mild cognitive impairment, were able to perform some HCI training, which was manifested by different degrees of improvement in clinical scores, especially in FMA scores (group B was larger than group C), reflecting the significant Improvement effect of BCI training on the stroke patients' limb motor function. MMSE scores and modified Barthel index before and after training of stroke patients in group B. The stroke patients in group B were found to be in good condition before and after the training, and there was a significant difference in MMSE scores, modified Barthel index before and after the training, brunnstrom staging improved to a certain extent (P\u0026lt; 0.05), but the difference was not statistically significant compared with C. The difference between group B and group C was not statistically significant, and the difference between group B and group C was not statistically significant. This may be due to the fact that most of the stroke patients were at the lower level of the mild cognitive impairment range, and the BCI training in this study mainly targeted the distal upper limbs of the stroke patients, so it was difficult to improve the cognitive deficits; and the modified Barthel index, brunnstrom stage and other indicators for the stage of the rehabilitation process of the composite scores, and it is difficult to cross the short-term training to improve. In the future, we need to further explore the length of BCI training intervention and the scope of the optimal population \u003csup\u003e[19]\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e3.2 Changes in CA with BCI training\u003c/p\u003e\n\u003cp\u003eNot all stroke patients in this study were able to maintain relatively high CA levels or significantly elevated CA values during BCI training, which is consistent with the findings of Tam et al\u0026nbsp;\u003csup\u003e[20]\u003c/sup\u003e. In addition, changes in CA values were not completely consistent with changes in FMA scores \u003csup\u003e[21]\u003c/sup\u003e, so CA values do not accurately reflect the final BCI advancement of stroke rehabilitation, but only suggest a training status of stroke patients performing MI. For example, stroke patient 7 had a poor mean CA value performance (\u0026lt;60%), but his FMA score improved by 5 points. In contrast, Group B had reasons why most stroke patients did not make any progress in CA: some stroke patients initially trained with confidence and curiosity to complete the training, and felt fidgety and bored after several repetitions of the training, and were unable to focus on motor imagery (e.g., stroke patients 8, 9, and 10); the adverse effects of the change in condition after stroke that caused stroke patients to have a low emotional state (e.g., stroke patients 12 and 22); and the fact that some stroke patients did not trust and were not willingness to use new technologies that are not yet widely used, resulting in low human-machine cooperation, etc.\u003c/p\u003e\n\u003cp\u003e3.3 BCI training-related brain network research\u003c/p\u003e\n\u003cp\u003eCurrent interventions for post-cerebral hemorrhage often focus on acute-phase hematoma management, blood pressure regulation, and post-acute rehabilitation \u003csup\u003e[22,23]\u003c/sup\u003e, but there is still a lack of research strategies on the multidimensional rehabilitation mechanisms of limb movement disorders (e.g., functionally-driven around aspects of neuroplasticity). rs-fMRI, as a non-invasive technique that can provide comprehensive characteristic information about the structural and functionally-related aspects of the brain, is one of the powerful tools for in-depth analysis of neuroplasticity mechanisms \u003csup\u003e[24]\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e3.3.1 FC changes in brain networks of stroke patients after BCI training\u003c/p\u003e\n\u003cp\u003eIn this study, we analyzed the correlation between ROI and time series of other brain regions to discover the possible connection between the recovery of motor function and FC changes in brain networks in stroke patients with cerebral hemorrhage. The results suggested that stroke patients had multiple FC enhancements within sensorimotor networks, including somatosensory cortex (postcentral gyrus, areas 1, 2, 3a, 3b, etc.), primary motor cortex (precentral gyrus, area 4, etc.), and premotor cortex (precentral gyrus anterior region, areas 55b, 6v, 6mp, etc.); and FC enhancements between sensorimotor and default isomodal networks such as anterior cingulate cortex (anterior side of the corpus callosum knee region a24, p32); FC enhancement between sensorimotor and language networks, e.g., premotor cortex (precentral anterior gyrus, region 55b), Broca's area, and memory-learning-related areas (lateral inferior frontal gyrus, regions 44, 45); FC enhancement between sensorimotor and visual networks, vision-related (ventral medial occipital lobe, region VMV2); FC enhancement between sensorimotor and ventral attention FC enhancement between sensory-motor and visual networks, such as eye-movement-related areas (medial superior frontal gyrus, area 8BM), anterior cingulate cortex (posterior inferior superior frontal gyrus, area a32pr); FC enhancement between sensory-motor and emotion-related areas, such as lateral orbitofrontal cortex (area 47m, etc.).In contrast, BCI training can integrate biological activities such as auditory, visual, proprioceptive, and higher cognition, making it a more comprehensive intervention strategy \u003csup\u003e[25,26]\u003c/sup\u003e. During the training period, stroke patients are required to maintain a high level of concentration, be constantly alert to handshake/open hand movements, and receive repetitive motor cues and image feedback (e.g., display of a smiling or crying face) upon completion of an MI task. With familiarity with these two forms of signaling, stroke patients no longer need to understand the specific semantics of the cues \u003csup\u003e[27]\u003c/sup\u003e and can automatically translate these signals into auditory or visual information that aids in decision-making and acts directly on the MI session. This process relies on the synergy of multiple aspects of higher cognition, sensorimotor integration, memory-learning transformation, the audiovisual system, and emotional stabilization, and reflects the compensatory mechanism of FC enhancement among brain networks.\u003c/p\u003e\n\u003cp\u003e3.3.2 Stroke patient limb motor function rehabilitation affects FC changes\u003c/p\u003e\n\u003cp\u003eBased on the fact that MI-BCI can act on multiple areas related to sensorimotor synergy (e.g., anterior cingulate gyrus, parietal lobe, and lateral frontal cortex), it involves both direct contact fiber projections (e.g., superior longitudinal fasciculus, frontal-parietal fasciculus) and thalamic indirect projections. The premotor cortex, a key area in the frontal lobe of the brain responsible for motor planning, coordination, and sensory-motor integration, is located anterior to the primary motor cortex (M1) and is an important component of the motor control network. The precentral anterior gyrus anterior inferior cortex (area 6v) is part of the premotor cortex, and area 6v exhibits an abnormally active state when the stroke patient is controlling hand manipulation of an object to occur movement (e.g., grasping and lifting an object), or is involved in performing a specific motor task based on visual cues \u003csup\u003e[28]\u003c/sup\u003e. Area PEF belongs to the same premotor area and is located at the junction of the precentral sulcus and the inferior frontal sulcus. This area is associated with reflexive eye movements, such as reflexive sweeps. Area 1 is the posterior part of the primary sensory area, located at the postcentral gyrus, which extends to the midline, and its white matter often projects to the pyramidal fasciculus, thalamocortex, and parietal lobe, among others.Area 1 is adjacent to area 3b and participates in the processing of tactile stimuli (manifested as area 1 being the second activation point of area 3b after receiving tactile activation); area 1 may also collaborate with area 2 of the somatosensory cortex in the reception of tactile stimulus information in both hands. Area 3b is the anterior half of the primary sensory area, which is the initial area of cortical activation related to tactile stimulation, especially specific information such as tactile stimulation or injurious stimulation of the finger skin. In this study, the increased connections within the premotor cortex (L_55b-L_6v) and between the premotor cortex and several regions such as primary sensory areas or eye movement-related areas (e.g., L_PEF-L_3b, L_6v-L_1, and L_6d-L_8BM) may reflect the close association between the involvement of somatosensory and eye movement behaviors during the rehabilitation of the upper limb motor function in stroke patients. Among them, the FC between the prefrontal side of the left precentral gyrus (L_PEF) and the anterior parietal side of the left postcentral gyrus (L_3b), and between the anterior subcortex of the left precentral gyrus (L_6v) and the left postcentral gyrus (L_3b, L_1) showed a moderate positive correlation with the FMA-UE scores (P\u0026lt; 0.05). This reflected a significant increase in the activation of the sensorimotor network during the recovery of the stroke patient's limb motor function, which was manifested by the occurrence of unification of the premotor cortex with the sensory system.\u003c/p\u003e\n\u003cp\u003eThe left precentral gyrus anterolateral (L_55b) is located in the premotor cortex area and plays an important role in language processing in addition to coordinating motor tasks \u003csup\u003e[28]\u003c/sup\u003e. The present study showed that FC between area 55b and the anterior inferior cortex of the precentral gyrus (area 6v) was positively correlated with FMA-UE scores (p\u0026lt; 0.05).\u003c/p\u003e\n\u003cp\u003eThe left lateral frontal lobe (L_p9-46v), as a key node of the ventral attentional network, has an important role in goal-directed higher-order cognitive processes as well as conscious active control of planned behavior \u003csup\u003e[29]\u003c/sup\u003e. The present study showed that the FC between the left precentral gyrus prefrontalis (L_PEF) and area p9-46v showed a significant positive correlation with FMA-UE scores (p\u0026lt; 0.05). In a multicenter observational study \u003csup\u003e[30]\u003c/sup\u003e, researchers found that the recovery of motor function after stroke was closely related to higher cognitive function, especially in stroke patients with subacute stroke. In this clinical trial, stroke patients needed to filter key information among a large number of information prompts, master the motor imagery nodes, and focus on completing the MI, a process that suggests that, in addition to focusing on the recovery of stroke patients' motor function during the rehabilitation process, attention should also be paid to the enhancement of cognitive function in order to promote the comprehensive rehabilitation of functional impairment in the future.\u003c/p\u003e\n\u003cp\u003e3.4 EEG signal correlation analysis\u003c/p\u003e\n\u003cp\u003eEEG signals are the most commonly employed means of neural recording when exploring MI. Due to its non-invasive characteristics and millisecond temporal resolution, the use of EEG signals in MI studies has been widely promoted in many fields, including neuroscience \u003csup\u003e[31]\u003c/sup\u003e. In particular, changes in the frequency band 8-30 HZ are closely related to MI or motor execution.\u003c/p\u003e\n\u003cp\u003eIn previous studies using EEG recordings of motor imagery (MI), when a subject performs an MI task, the energy corresponding to 8-30 HZ in the sensorimotor cortex of the contralateral brain undergoes a decrease, a phenomenon known as the ERD phenomenon. On the contrary, when the MI activity is terminated, the corresponding energy within this frequency band in the same cortical region increases significantly, and this phenomenon of increased energy is known as Event-relatedsynchronization (ERS) \u003csup\u003e[32]\u003c/sup\u003e. In this study, we utilized the interchannel connectivity relationship, i.e., the energy change between channels in a certain EEG frequency band in a group of stroke patients, to indirectly reflect the attenuation or enhancement of the ERD/ERS phenomenon between channels.\u003c/p\u003e\n\u003cp\u003eThe results showed that when stroke patients in group B performed the MI task with the affected hand, their EEG frequency band was significantly reduced in most of the inter-channel connectivity within 8-30 Hz (α and β bands), such as Pz-C3 in the α band, Pz-F3 in the α + β bands, P3-Fp1 in the α + β bands, and so on. This connectivity was mostly concentrated near the C1 channel in the left cerebral hemisphere, mostly involving channels P3, PO7, C3, F3, etc. Considering the EEG functional data processed by mirror image inversion, all the affected hands were considered as the right hand, and after a period of BCI training, the affected brain of stroke patients in group B would show enhanced ERD phenomenon when performing MI, which also verified the results of PSD analysis and brain topography, reflecting the enhancement of their MI ability. The enhancement of the connectivity of the C4-T7 in α-band suggests that ERS phenomenon is enhanced in the healthy brain of the stroke patient, and the stroke patient The interhemispheric EEG balance was altered and positively correlated with the FMA-LE score (P\u0026lt; 0.05).\u003c/p\u003e\n\u003cp\u003eThe following limitations exist in this study: (1) only the MRI and EEG levels were analyzed in Group B stroke patients in this study, and changes in neural remodeling brought about by routine rehabilitation training and their own recovery could not be excluded, which need to be verified by further controlled studies; (2) this study only focused on the FC enhancement aspect of the brain network, and the meaning of FC attenuation has not been explored; (3) the site of cerebral hemorrhage, the extent of the hemorrhage, and the degree of fiber bundle damage may be different in this study, which may affect the results of the analysis of the functional brain network characteristics; (4) some studies have shown that long-term BCI training can bring more objective benefits. In this study, we only recorded the changes in the rehabilitation of stroke patients with subacute cerebral hemorrhage for 1 month, which may underestimate the benefit of BCI training and produce ambiguous results in the analysis of the functional recovery mechanism, so it is still necessary to expand the sample size, and to develop a more reasonable training cycle and follow-up time for in-depth study.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eBCI rehabilitation training can effectively promote the recovery of limb motor function in stroke patients with cerebral hemorrhage in the basal ganglia region in the subacute stage, but it has little effect on the enhancement of some indexes (e.g., cognitive function scores, modified Barthel index, and brunnstrom's staging).The probable mechanism of BCI rehabilitation training lies in the enhancement of the stroke patient's motor imagery ability as well as the promotion of the brain network activity.\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eEthics approval and consent to participate\u003c/p\u003e\n\u003cp\u003eThe study was approved by the Ethics Committee of Shanxi Provincial People's Hospital under the approval number (2022) Provincial Medical Science Lun Audit No. 95, and all patients participating in the trial signed an informed consent form.\u003c/p\u003e\n\u003cp\u003eConsent for publication\u003c/p\u003e\n\u003cp\u003eAll authors have read and approved the final manuscript. We confirm that this work is original and has not been published elsewhere.\u003c/p\u003e\n\u003cp\u003eAvailability of data and materials\u003c/p\u003e\n\u003cp\u003eAll data supporting this study are included in the article.\u003c/p\u003e\n\u003cp\u003eCompeting interests\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003eFunding\u003c/p\u003e\n\u003cp\u003eThis study was supported by the upper-level project of Natural Science Foundation of Shanxi Provincial Department of Education (No. 2022L184) and the Science and Technology Innovation Project for Higher Schools of Shanxi Provincial Department of Science and Technology (No. 202203021211060).\u003c/p\u003e\n\u003cp\u003eAuthors' contributions\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(I) Conception and design: Peili Cao, Hong Ling, Rui Cheng; (II) Administrative support: Rui Cheng, Gangli Zhang; (III) Provision of study materials or patients: JunHao Wang, Hao Guo; (IV) Collection and assembly of data: Xiang Zan, FuLong Zhang; (V) Data analysis and interpretation: Peili Cao, Hong Ling, Qihang Jin; (VI) Manuscript writing: All authors; (VII) Final approval of manuscript: All authors.\u003c/p\u003e\n\u003cp\u003eAcknowledgements\u003c/p\u003e\n\u003cp\u003eThe authors would like to thank all volunteers who volunteered for the study.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eMeng G, Huang Y, Yu Q, Ding Y, Wild D, Zhao Y, Liu X, Song M. Adopting Text Mining on Rehabilitation Therapy Repositioning for Stroke[J]. Front Neuroinformatics. 2019;13:17.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSoleimani M, Ghazisaeedi M, Heydari S. The efficacy of virtual reality for upper limb rehabilitation in stroke stroke patients: a systematic review and meta-analysis[J]. BMC Med Inf Decis Mak. 2024;24(1):135.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu Q, Ge Y, Ma D, Pang X, Cao Y, Zhang X, Pan Y, Zhang T, Dou W. Analysis of Prognostic Risk Factors Determining Poor Functional Recovery After Comprehensive Rehabilitation Including Motor-Imagery Brain-Computer Interface Training in Stroke Stroke patients: A Prospective Study[J]. Front Neurol. 2021;12:661816.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang Y, Yin J, Zhao B, Zhang Y, Peng T, Zhuang W, Wang S, Huang S, Zhong M, Zhang Y, Tang G, Shen B, Ou H, Zheng Y, Lin Q. Motor Imagery Brain-Computer Interface in Rehabilitation of Upper Limb Motor Dysfunction After Stroke[J]. J Visualized Experiments: JoVE, 2023(199).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang S, Li R, Li H, Xu K, Shi Y, Wang Q, Yang T, Sun X. Exploring the Use of Brain-Computer Interfaces in Stroke Neurorehabilitation[J]. BioMed Research International, 2021, 2021: 9967348.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFan C, Yang B, Li X, Zan P. Temporal-frequency-phase feature classification using 3D-convolutional neural networks for motor imagery and movement[J]. Front NeuroSci. 2023;17:1250991.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLuu TP, Brantley JA, Nakagome S, Zhu F, Contreras-Vidal JL. Electrocortical correlates of human level-ground, slope, and stair walking[J]. PLoS ONE. 2017;12(11):e0188500.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePichiorri F, Morone G, Petti M, Toppi J, Pisotta I, Molinari M, Paolucci S, Inghilleri M, Astolfi L, Cincotti F, Mattia D. Brain-computer interface boosts motor imagery practice during stroke recovery[J]. Ann Neurol. 2015;77(5):851\u0026ndash;65.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNunes JD, Vourvopoulos A, Blanco-Mora DA, Jorge C, Fernandes JC, Bermudez I, Badia S, Figueiredo P. Brain activation by a VR-based motor imagery and observation task: An fMRI study[J]. PLoS ONE. 2023;18(9):e0291528.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang S, Xie L. Improving Attention Mechanism in Graph Neural Networks via Cardinality Preservation[J]. IJCAI: proceedings of the conference, 2020, 2020: 1395\u0026ndash;1402.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGanepola T, Nagy Z, Ghosh A, Papadopoulo T, Alexander DC, Sereno MI. Using diffusion MRI to discriminate areas of cortical grey matter[J]. NeuroImage. 2018;182:456\u0026ndash;68.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarria P, Pino A, Tovar N, Gomez-Vargas D, Baleta K, D\u0026iacute;az CAR, M\u0026uacute;nera M, Cifuentes CA. BCI-Based Control for Ankle Exoskeleton T-FLEX: Comparison of Visual and Haptic Stimuli with Stroke Survivors[J]. Sensors. 2021;21(19):6431.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiaw N, Liebeskind D. Emerging therapies in acute ischemic stroke[J]. F1000Research, 2020, 9: F1000 Faculty Rev-546.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhi JF, Liao QH, He YB, Xu WW, Zhu DW, Shao LH. Superior treatment efficacy of neuromodulation rehabilitation for upper limb recovery after stroke: a meta-analysis[J]. Expert Rev Neurother. 2022;22(10):875\u0026ndash;88.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhan G, Chen S, Ji Y, Xu Y, Song Z, Wang J, Niu L, Bin J, Kang X, Jia J. EEG-Based Brain Network Analysis of Chronic Stroke Stroke patients After BCI Rehabilitation Training[J]. Front Hum Neurosci. 2022;16:909610.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNojima I, Sugata H, Takeuchi H, Mima T. Brain-Computer Interface Training Based on Brain Activity Can Induce Motor Recovery in Stroke patients With Stroke: A Meta-Analysis[J]. Neurorehabilit Neural Repair. 2022;36(2):83\u0026ndash;96.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNan W, Dias APB, Rosa AC. Neurofeedback Training for Cognitive and Motor Function Rehabilitation in Chronic Stroke: Two Case Reports[J]. Front Neurol. 2019;10:800.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYuan Z, Peng Y, Wang L, Song S, Chen S, Yang L, Liu H, Wang H, Shi G, Han C, Cammon JA, Zhang Y, Qiao J, Wang G. Effect of BCI-Controlled Pedaling Training System With Multiple Modalities of Feedback on Motor and Cognitive Function Rehabilitation of Early Subacute Stroke Stroke patients[J]. IEEE Trans neural Syst rehabilitation engineering: publication IEEE Eng Med Biology Soc. 2021;29:2569\u0026ndash;77.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu Y, Huang S, Xu W, Wang Z, Ming D. An fMRI study on the generalization of motor learning after brain actuated supernumerary robot training[J]. NPJ Sci Learn. 2024;9(1):80.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTam WK, Tong Kyu, Meng F, Gao S. A Minimal Set of Electrodes for Motor Imagery BCI to Control an Assistive Device in Chronic Stroke Subjects: A Multi-Session Study[J]. IEEE Trans Neural Syst Rehabil Eng. 2011;19(6):617\u0026ndash;27.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMiao Y, Chen S, Zhang X, Jin J, Xu R, Daly I, Jia J, Wang X, Cichocki A, Jung TP. BCI-Based Rehabilitation on the Stroke in Sequela Stage[J]. Neural Plasticity, 2020, 2020: 1\u0026ndash;10.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePham LT, Chu SD, Hoang HV. Characteristic of 24-Hour Blood Pressure Dipping Patterns in Hypertensive Stroke Stroke patients[J]. Vasc Health Risk Manag. 2024;20:501\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang J, Zhang J, Ye Y, Xu Q, Li Y, Feng S, Xiong X, Jian Z, Gu L. Peripheral Organ Injury After Stroke[J]. Front Immunol. 2022;13:901209.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRuan X, Song Z, Zhang J, Yu T, Chen J, Zhou T. Alterations of brain activity in stroke patients with alcohol use disorder: a resting-state fMRI study[J]. BMC Psychiatry. 2023;23(1):894.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePhunruangsakao C, Achanccaray D, Bhattacharyya S, Izumi SI, Hayashibe M. Effects of visual-electrotactile stimulation feedback on brain functional connectivity during motor imagery practice[J]. Sci Rep. 2023;13(1):17752.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu Q, Yue Z, Ge Y, Ma D, Yin H, Zhao H, Liu G, Wang J, Dou W, Pan Y. Brain Functional Networks Study of Subacute Stroke Stroke patients With Upper Limb Dysfunction After Comprehensive Rehabilitation Including BCI Training[J]. Front Neurol. 2019;10:1419.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKruschwitz JD, Waller L, List D, Wisniewski D, Ludwig VU, Korb F, Wolfensteller U, Goschke T, Walter H. Anticipating the good and the bad: A study on the neural correlates of bivalent emotion anticipation and their malleability via attentional deployment[J]. NeuroImage. 2018;183:553\u0026ndash;64.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaker CM, Burks JD, Briggs RG, Sheets JR, Conner AK, Glenn CA, Sali G, McCoy TM, Battiste JD, O\u0026rsquo;Donoghue DL, Sughrue ME. A Connectomic Atlas of the Human Cerebrum-Chap. 3: The Motor, Premotor, and Sensory Cortices[J]. Operative Neurosurgery (Hagerstown, Md.), 2018, 15(suppl_1): S75-S121.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaker CM, Burks JD, Briggs RG, Conner AK, Glenn CA, Morgan JP, Stafford J, Sali G, McCoy TM, Battiste JD, O\u0026rsquo;Donoghue DL, Sughrue ME. A Connectomic Atlas of the Human Cerebrum-Chap. 2: The Lateral Frontal Lobe[J]. Operative Neurosurgery (Hagerstown, Md.), 2018, 15(suppl_1): S10-S74.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMancuso M, Iosa M, Abbruzzese L, Matano A, Coccia M, Baudo S, Benedetti A, Gambarelli C, Spaccavento S, Ambiveri G, Megna M, Tognetti P, Maietti A, Rinaldesi ML, Gamberini G, Varalta V, Morone G, Ciancarelli I, CogniReMo Study Group. The impact of cognitive function deficits and their recovery on functional outcome in subjects affected by ischemic subacute stroke: results from the Italian multicenter longitudinal study CogniReMo[J]. Eur J Phys Rehabil Med. 2023;59(3):284\u0026ndash;93.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWei Z, Li H, Ma L, Li H. Emotion recognition based on microstate analysis from temporal and spatial patterns of electroencephalogram[J]. Front NeuroSci. 2024;18:1355512.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRimbert S, Lelarge J, Guerci P, Bidgoli SJ, Meistelman C, Cheron G, Cebolla Alvarez AM, Schmartz D. Detection of Motor Cerebral Activity After Median Nerve Stimulation During General Anesthesia (STIM-MOTANA): Protocol for a Prospective Interventional Study[J]. Volume 12. JMIR research protocols; 2023. p. e43870.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Basal ganglia intracerebral hemorrhage, Brain computer interface, Connectomics, Neuroplasticity, recovered","lastPublishedDoi":"10.21203/rs.3.rs-6791097/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6791097/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eObjective: \u003c/strong\u003eObjective to explore the degree of recovery of limb motor function and potential rehabilitation mechanism of stroke patients with subacute basal ganglia cerebral hemorrhage after receiving rehabilitation training based on brain computer interface (BCI) of motor imagination.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods: \u003c/strong\u003eDuring hospitalization from December 2022 to October 2024, stroke patients with subacute basal ganglia intracerebral hemorrhage accompanied with limb movement disorder were randomly divided into conventional rehabilitation treatment group (Group C) and BCI group (group B), who completed traditional rehabilitation training, BCI training combined with traditional rehabilitation training, and the training time was 5 weeks. Before and after the training, the rehabilitation status of stroke patients was evaluated by Fugl Meyer Assessment Scale (FMA), simple mental state examination scale, modified Barthel index, online classification accuracy (CA), Brunnstrom staging, etc. Then resting state fMRI scanning was performed to compare the functional connectivity changes between the regions of interest and the whole brain. The imaginary part coherence analysis method was used to calculate the connectivity changes between channels in the 8-30hz frequency band of EEG, and the changes of power spectral density (PSD) and brain topography in the two stages before and after training.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults: \u003c/strong\u003eA total of 44 stroke patients were included in this study, 22 in group C and 22 in group B. There was no significant difference in clinical related scores between the two groups before training (\u003cem\u003eP\u003c/em\u003e\u0026gt;0.05); After training, compared with group C, the FMA score of group B stroke patients was significantly improved (\u003cem\u003eP\u003c/em\u003e\u0026lt;0.05). The CA values of most stroke patients in group B showed fluctuations and a downward trend, and only 6 stroke patients could maintain relatively high Ca values (more than 70% threshold standard).In group B, the functional connectivity between sensorimotor network and default mode, ventral attention, language network and other networks was significantly enhanced in 19 stroke patients (\u003cem\u003eP\u003c/em\u003e\u0026lt;0.05). Among them, the connection between the prefrontal side of the left precentral gyrus (L_PEF) and the anterior parietal side of the left postcentral gyrus (L_3b), and the connection between the anterior inferior cortex of the left precentral gyrus (L_6v) and the lateral side of the left postcentral gyrus (L_1) were positively correlated with clinical scores (\u003cem\u003eP\u003c/em\u003e\u0026lt;0.05). The connectivity between multiple channels within 4-30hz (αandβfrequency bands) of EEG was significantly weakened, and the enhanced connectivity between C4-T7 and T8-T7 in theαfrequency band during the execution of motor imagination by the affected hand was positively correlated with the change of lower limb clinical score. After training, the PSD curve in the range of 8-30hz in the central motor area of the affected side of the brain tended to be smoother than before, and the energy value decreased to a certain extent; Through the changes of brain topographic map before and after, it was found that during the repeated training of the affected hands of many stroke patients, the ipsilateral hemisphere cortex had obvious personal leave related desynchronization phenomenon or this phenomenon was further enhanced than before, and tended to converge to the central region of the brain and the surrounding areas.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion：\u003c/strong\u003eBCI rehabilitation training can effectively promote the recovery of limb motor function in stroke patients with subacute basal ganglia cerebral hemorrhage. The possible mechanism of rehabilitation lies in enhancing the motor imagination ability of stroke patients and promoting the improvement of brain network activity.\u003c/p\u003e","manuscriptTitle":"Effect of brain computer interface on limb motor function after intracerebral hemorrhage in basal ganglia and its rehabilitation mechanism","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-02 06:14:20","doi":"10.21203/rs.3.rs-6791097/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"825a97d1-3890-492b-b5a7-1c8182fb9087","owner":[],"postedDate":"July 2nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-11-15T02:08:32+00:00","versionOfRecord":[],"versionCreatedAt":"2025-07-02 06:14:20","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6791097","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6791097","identity":"rs-6791097","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}