Functional Connectivity Associated with Severe Upper Limb Impairment in Resting-State Electroencephalography Among Chronic Stroke Survivors: A Machine Learning Approach

Functional Connectivity Associated with Severe Upper Limb Impairment in Resting-State Electroencephalography Among Chronic Stroke Survivors: A Machine Learning Approach | 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 Functional Connectivity Associated with Severe Upper Limb Impairment in Resting-State Electroencephalography Among Chronic Stroke Survivors: A Machine Learning Approach Ji-Yoon Lee, Miseon Shim, Won Kee Chang, Hee-Mun Cho, Ji Soo Choi, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5976957/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 22 Dec, 2025 Read the published version in Journal of NeuroEngineering and Rehabilitation → Version 1 posted 10 You are reading this latest preprint version Abstract Background Severe upper limb impairment (ULI) presents a significant challenge in the rehabilitation of chronic stroke survivors, affecting their quality of life. Identifying biomarkers and understanding the neural mechanisms associated with severe ULI are essential for evaluating recovery potential and enhancing rehabilitation effectiveness. This study aimed to identify resting-state electroencephalography (EEG) functional connectivity features related to severe ULI in chronic stroke survivors using machine learning (ML) methods. Methods EEG data were collected from 34 chronic stroke survivors. Participants were categorized into two labels based on their Fugl-Meyer assessment for upper extremity (FMA-UE) scores: a mild/moderate ULI (FMA-UE ≥ 30; n = 19) and a severe ULI (FMA-UE < 30; n = 15). We employed ML algorithms to classify severe ULI, including logistic regression with L1, elastic net regularization, stochastic gradient descent, and support vector machines, along with several feature selection methods. Coherence was evaluated across six frequency bands within both the ipsilesional (affected by the lesion) and contralesional (opposite side of the lesion) hemispheres. Results The logistic regression model with L1 and ReliefF feature selection methods was the most effective, achieving a balanced accuracy of 0.91 (sensitivity = 0.93, specificity = 0.90). This approach identified 14 significant features for distinguishing severe ULI from mild to moderate ULI, including delta interhemispheric and intrahemispheric connectivity of the frontal, parietal, and temporal regions. Additionally, interhemispheric and intrahemispheric theta connectivity was observed in the prefrontal, frontal, temporal, and parietal regions. Low-beta intrahemispheric connectivity was also observed in the contralesional parietal regions. Conclusions Our research highlights the association between alterations in connectivity within low-frequency bands and severe ULI across widespread brain regions, including areas outside the sensorimotor cortex and bilateral intrahemispheric and interhemispheric regions. Further research utilizing larger longitudinal datasets from early stroke survivors employing ML approaches could contribute to the development of more accurate predictive models for motor recovery and rehabilitation responses. Stroke upper limb impairment motor recovery Fugl-Meyer Assessment electroencephalography coherence connectivity machine learning Figures Figure 1 Figure 2 Figure 3 Background Stroke is a major health concern owing to its high prevalence and significant impact on disability [ 1 ]. According to the World Health Organization, over 60% of chronic stroke survivors continue to suffer from upper limb impairment (ULI) [ 2 ]. Improving upper limb function is a key rehabilitation goal; however, individuals with severe ULI face substantial challenges in achieving the desired outcomes [ 3 ]. The impact of severe ULI on the survivors’ quality of life is particularly significant [ 4 ]. Previous research emphasizes the importance of identifying biomarkers to develop targeted and efficient interventions for stroke survivors with severe ULI [ 5 , 6 ]. Electroencephalography (EEG) is advantageous for biomarker development because of its noninvasive nature, excellent temporal resolution, and cost-effectiveness among neuroimaging methods [ 7 ]. In a scoping review conducted by Milani, Antonioni [ 8 ], EEG-derived measures proved valuable for evaluating upper limb recovery. However, most studies have concentrated on features related to power spectrum density; a notable gap remains in our understanding of how different brain regions connect, process, and share information in stroke survivors [ 9 , 10 ]. Understanding these connections is crucial because functional connectivity can serve as a valuable marker for monitoring motor recovery. Defined as the statistical interdependencies between physiological time series from different brain regions, functional connectivity plays a crucial role in brain function [ 11 ]. Furthermore, the brain remains active even during rest, displaying spontaneous activity that is not linked to external stimuli or responses but arises internally [ 12 ]. Therefore, resting-state functional connectivity shows promise for developing biomarkers to monitor motor recovery post-stroke [ 6 , 13 , 14 ]. Carter, Shulman [ 15 ] emphasized that stroke recovery depends on how the brain’s cortical and subcortical networks reorganize, not just during active tasks but also in a resting state. Therefore, resting-state connectivity may act as a potential biomarker of motor impairment following stroke [ 16 – 18 ]. Additionally, the neural mechanisms underlying recovery after stroke, which affect the ipsilesional and contralesional brain hemispheres, are not yet fully understood [ 19 ]. Studies have reported that acute stroke survivors exhibit altered EEG inter-hemispheric connectivity and modifications in the balance between local segregation and global integration [ 20 ]. Employing a hemispheric-focused network analysis is crucial for deeply understanding cost-effective neuronal compensatory mechanisms in stroke survivors [ 21 ]. While many studies have concentrated on the connectivity of specific brain regions [ 22 ], understanding of comprehensive inter- and intra-hemispheric connectivity patterns across all brain regions remains limited. EEG coherence, which indicates the functional connectivity between different electrode sites, reflects the degree of signal synchronization and suggests functional interaction [ 23 ]. The complexity of EEG coherence, characterized by the complexity, multichannel, multifrequency, and inherent complexity of brain function and structure, introduces a challenge in comprehensively analyzing resting-state EEG connectivity. Machine learning (ML) could play a crucial role in analyzing EEG data, which incurs significant computational costs and time owing to the vast amount of data and potential redundancy among EEG features [ 24 ]. A recent study by Lassi and Bandini [ 25 ] demonstrated the effective use of ML and feature-selection methods to classify ULI severity in acute stroke survivors. Their best model, based on a support vector machine (SVM), achieved over 85% accuracy, with the brain symmetry index identified as a key feature. High-dimensional data, such as resting-state coherence analyzed across the whole brain, may reduce the model performance and complicate interpretations. However, by employing optimal ML feature selection techniques, we can condense the features, thereby improving model accuracy and simplifying the interpretation of neural patterns associated with severe ULI. Compared with the conventional hypothesis-driven approach, this data-driven ML approach enables a more comprehensive exploration of the role of EEG coherence in motor impairment after stroke [ 16 , 26 ]. This study aimed to identify resting-state EEG connectivity patterns related to ULI severity in chronic stroke survivors, as indicated by the coherence between each channel in the ipsilesional and contralesional hemispheres. First, we attempted to determine the best ML model for classifying the severity of ULI—severe ULI vs. mild/moderate ULI—and then attempted to identify the EEG connectivity features critical for distinguishing ULI severity in the ML model. Methods Participants Thirty-four stroke survivors, aged between 29 and 80, were recruited in the study from two rehabilitation hospitals between June and December 2019. Participants were selected based on the following criteria: 1) diagnosis of first-ever ischemic or hemorrhagic stroke confirmed by brain computed tomography or magnetic resonance imaging, 2) unilateral upper limb weakness, 3) a minimum of 6 months post-stroke, and 4) the ability to provide written informed consent. Participants were excluded if they met any of the following conditions: 1) a history of central nervous system conditions such as traumatic brain injury, brain tumor, or Parkinson's disease; 2) inability to wear an EEG cap; or 3) difficulty in comprehending and adhering to clinical assessment and EEG study protocols. A clinician reviewed the structural images of the participants and classified their stroke locations into three categories: cortical, subcortical, including corona radiata, internal capsule, or basal ganglia, or a combination of cortical and subcortical. Assessment for upper limb motor impairment The Fugl-Meyer assessment for the upper extremity (FMA-UE) was used to assess upper limb motor function [ 27 ]. The FMA-UE is an objective measure used to assess motor recovery outcomes, with scores ranging from 0 to 66, with higher scores indicating better motor function and recovery. In the study by Coscia, Wessel [ 28 ], ULI severity was classified based on FMA-UE scores: >45 for mild, 30–45 for moderate, and < 30 for severe ULI. In our study, we adopted an FMA-UE threshold of 30 to classify participants into two groups with fairly balanced sample sizes: severe ULI (n = 15) and mild/moderate ULI (n = 19). Baseline group differences are presented in Table 1 . Age and FMA-UE were the only variables that reached statistical significance ( p < 0.05). Table 1 Demographic characteristics of participants Mild/moderate ULI (n = 19) Severe ULI (n = 15) p Age (years) 57.58 ± 13.77 66.93 ± 8.51 0.03 FMA-UE 47 (14.50) 9 (12) < .001 Time since stroke onset (days) 315 (185.50) 321 (159.50) 0.56 Sex Male 10 (52.60%) 5 (33.30%) 0.44 Female 9 (47.40%) 10 (66.70%) Hemiplegia side Left 8 (42.10%) 11 (73.30%) 0.14 Right 11 (57.90%) 4 (26.70%) Legion location Cortical 6 (31.60%) 2 (13.30%) 0.46 Subcortical 11 (57.90%) 11 (73.30%) Cortical / subcortical 2 (10.50%) 2 (13.30%) Continuous variables were analyzed using the Mann-Whitney U test or Student’s t-test based on the Shapiro-Wilk normality test. Categorical variables were assessed using the chi-squared test. Continuous data are presented as mean ± standard deviation or median (interquartile range), whereas categorical data are presented as numbers (percentages). Statistical significance was set at p < 0.05. significant. Abbreviations: FMA-UE, the Fugl-Meyer assessment for upper extremity. EEG recording and preprocessing Participants were seated comfortably in an armchair in front of a monitor to facilitate EEG recordings. To minimize eye movement artifacts and ensure the quality of the EEG recordings, the participants were instructed to focus their gaze on a fixation cross displayed on the monitor. EEG data were collected during two resting-state sessions for a total of 4 minutes (one minute with eyes closed, followed by one minute with eyes open, and repeated twice). A movement-related EEG session was conducted between resting-state sessions. EEG data were recorded at a sampling rate of 1000 Hz using 32 Ag/AgCl scalp electrodes according to the extended international 10–10 system (Brain Products, GmbH Ltd., Gilching, Germany) (Fig. 1 A). The ground electrode was placed between Fp1 and Fp2, and the reference electrode at Cz. The EEG data were band-pass filtered from 0.1 to 55 Hz, and eye-related artifacts were removed by using mathematical procedures based on the first principal component analysis via Curry 7 (Compumedics, USA) software. We analyzed two sets of artifact-free eyes-closed data from the first resting-state session to evaluate the intrinsic brain activity. The data were downsampled to 500 Hz and re-referenced using a common average reference. Each eyes-closed set was divided into 4-second epochs. The first 56 s were used, excluding the last four seconds due to data quality issues, resulting in twenty-eight epochs for analysis. Feature extraction For feature extraction, we used EEG electrodes from the left and right hemispheres, excluding the midline electrodes (Fz, Cz, Pz, and Oz) because our focus was on investigating inter- and intrahemispheric brain connectivity. To calculate functional connectivity, the imaginary part of coherency (iCoh) was computed between the 28 electrodes for each EEG epoch and then averaged across all epochs. iCoh is widely used to avoid volume conduction effects caused by the different conductivities of brain layers (e.g., CSF, skull, and scalp), preserving the original signals from the cortical surface in the functional connectivity pattern [ 29 ]. The iCoh features were computed in six frequency bands: delta (1–4 Hz), theta (4–8 Hz), alpha (8–12 Hz), low-beta (12–20 Hz), high-beta (20–30 Hz), and gamma (30–55 Hz). Finally, a total of 2,268 iCoh values ( 28 C 2 × 6 frequency bands) were used for further analyses. The channels for coherence were reorganized to highlight the effects corresponding to the locations of stroke lesions in each hemisphere. The Matlab-based Fieldtrip toolbox was used to calculate the iCoh values [ 30 ]. Feature selection and ML-based classification A total of 2,268 iCoh values were used to classify groups with different levels of ULI severity; however, a high number of features relative to the number of samples (participants) could lead to overfitting. To mitigate this, we adopted feature-selection methods to optimize the features for classifying the two groups. Three feature selection methods are widely used in EEG studies: analysis of variance (ANOVA), SVM-recursive feature elimination (RFE), and ReliefF were used [ 31 , 32 ]. ANOVA employs F-values to select features by considering the relationship between the label and each feature [ 33 ]. SVM-RFE handles many features without requiring statistical corrections and considers individual differences by eliminating irrelevant or extreme data points [ 34 , 35 ]. ReliefF, a category of relief-based feature selection methods, assesses the importance of features by their ability to distinguish between neighboring data points [ 36 , 37 ]. All features were normalized prior to feature selection. Feature selection was performed in each iteration of the cross-validation process to prevent overfitting and reduce potential selection bias [ 38 ]. Preliminary analyses showed that expanding the candidate feature sets beyond 100 did not significantly improve the classification performance. Consequently, we concentrated on a subset of features, selecting between one and 100 features for further analysis. The classification performance was evaluated using four ML algorithms: logistic regression with L1, elastic net regularization, stochastic gradient descent (SGD), and SVM. Hyperparameters for each algorithm are listed in Supplementary Table S1 . Leave-one-out cross-validation (LOOCV) was used to prevent overfitting problems and biases due to the small sample size (n = 34). Important features were identified based on their occurrence rate, which was calculated as the ratio of their appearances to the total number of LOOCV iterations. The ML analysis was conducted using Python 3.11 ( https://www.python.org/ ) and its compatible open-source libraries. A flowchart of the proposed ML approach is shown in Fig. 1 . B. Results Model performance The performances of all possible combinations of the three-feature selection and classification methods are detailed in Table 2 and Fig. 2 . Logistic regression with the L1 model combined with the ReliefF feature selection method showed superior performance in classifying ULI severity in chronic stroke survivors (accuracy = 0.91, sensitivity = 0.93, and specificity = 0.90 when using 14 features). Table 2 Model performance and the number of selected features for each algorithm and feature selection method ANOVA SVM-RFE ReliefF Num Bal. Accu. Sens. Spec. Num Bal. Accu. Sens. Spec. Num Bal. Accu. Sens. Spec. Logistic regression with L1 2 0.82 0.80 0.84 53 0.71 0.80 0.63 14 0.91 0.93 0.90 Logistic regression with elastic net 2 0.79 0.73 0.84 2 0.65 0.67 0.63 2 0.85 0.80 0.90 Logistic regression with SGD 2 0.74 0.67 0.79 2 0.65 0.60 0.68 2 0.79 0.73 0.84 SVM 2 0.82 0.80 0.84 6 0.65 0.53 0.74 11 0.88 0.80 0.95 Bold font indicates the highest model performance across all algorithms and feature selection methods. Abbreviations: ANOVA, analysis of variance; SVM, support vector machine; RFE, recursive feature elimination; Num, the number of features; Bal. Accu., balanced accuracy; Sens, sensitivity; Spec, specificity; SGD, stochastic gradient descent. Important EEG features Logistic regression with L1 and the ReliefF method identified 30 features from 14 input variables within LOOCV, as shown in Supplementary Table S2. Among these, we focused on important features with empirically determined occurrence rates exceeding 0.5. Important features were identified, including three delta, ten theta, and one low-beta coherences whereas alpha and high-beta coherences were not selected (Table 3 and Fig. 3 ). Table 3 Occurrence rate of top fourteen features in the optimal model distinguishing severe from mild/moderate ULI. Optimal model Algorithm Logistic regression with L1 Feature selection method ReliefF Selected features Occurrence rate Delta TP.ipsi - F.con 1.00 Delta CP.ipsi-P.ipsi 1.00 Low-beta P.con-TP.con 1.00 Theta TP.con-FP.con 1.00 Theta F.con-FP.con 1.00 Theta FC.ipsi-CP.ipsi 1.00 Theta CP.ipsi-FP.con 1.00 Theta F.ipsi-P.ipsi 1.00 Theta FC.ipsi-P.ipsi 1.00 Theta F.ipsi-CP.ipsi 1.00 Theta FC.ipsi-P.ipsi 0.91 Delta T.ipsi-CP.ipsi 0.79 Theta FC.ipsi-P = .con 0.79 Theta P.ipsi-FP.con 0.62 The occurrence rate is computed as the number of appearances divided by the total number of cross-validation iterations. Abbreviations: FMA-UE, the Fugl-Meyer assessment for upper extremity; channel number; ipsi, ipsilesional; con, contralesional; FP, prefrontal; F, frontal; FC, frontocentral; FT, frontotemporal; T, temporal; TP, temporoparietal; C, central; CP, centroparietal; P, parietal; O, occipital. We identified significant features, including the delta interhemispheric connectivity of the ipsilesional temporoparietal and contralesional frontal regions. Additionally, delta connectivity was observed with intrahemispheric connectivity, specifically between the ipsilesional temporal and centroparietal regions as well as the ipsilesional centroparietal and parietal areas. Furthermore, theta interhemispheric connectivity was found between the ipsilesional centroparietal and contralesional prefrontal regions, the ipsilesional frontocentral and contralesional parietal regions, and the ipsilesional parietal and contralesional prefrontal regions. Regarding intrahemispheric connectivity, theta coherence was noted between the ipsilesional frontocentral and centroparietal regions, ipsilesional frontal and parietal regions, ipsilesional frontocentral and parietal regions, and ipsilesional frontal and centroparietal regions. In the contralesional hemisphere, theta coherence was observed between the temporoparietal and prefrontal regions, as well as between the frontal and prefrontal regions. Intrahemispheric coherence in the low beta band was also found between the contralesional parietal and temporoparietal regions. Discussion Achieving a balanced accuracy of approximately 91%, the ML approach in this study achieved a higher model performance in classifying severe ULI in chronic stroke survivors, notably by utilizing whole-brain resting-state EEG coherence features from all channels. Fourteen significant coherence features were identified across the delta, theta, and low-beta frequency bands most frequently selected in LOOCV. Most resting-state EEG coherence features that substantially contribute to distinguishing severe ULI are distributed in low-frequency bands, specifically the delta and theta bands. Logistic regression with L1 significantly selected delta interhemispheric connectivity between the ipsilesional temporoparietal and contralesional frontal regions as well as intrahemispheric connectivity within the ipsilesional temporal, centroparietal, and parietal areas. Significant theta inter-hemispheric and intra-hemispheric connectivity was observed in the frontal and parietal regions. These findings align with those of previous studies reporting that worse outcomes after stroke were associated with low-frequency oscillations [ 39 , 40 ]. Low-frequency oscillations are associated with cognitive control, attentional processing, movement, and skilled motor control [ 41 – 44 ]. Brain damage following stroke, brain tumors, or traumatic brain injury exhibits altered activity in the delta and theta frequency bands [ 40 , 45 , 46 ], with more pronounced slowing observed in cases of more severe brain damage [ 47 ]. Alterations in low-frequency bands are also associated with general stroke-related outcomes, such as the modified Rankin Scale (mRS) and the National Institutes of Health Stroke Scale (NIHSS). Higher delta power in the ipsilateral hemisphere [ 48 ], a higher delta power ratio [ 49 ], and increased delta power during the acute stroke phase [ 50 ] were associated with higher NIHSS scores, indicating worse functional outcomes. Furthermore, a higher delta-theta/alpha-beta ratio (DTABR) was linked to worse functional outcomes, as reflected by higher mRS scores [ 51 , 52 ]. Therefore, changes in delta- or theta-band frequencies on EEG are recognized as consistent and important biomarkers of stroke severity and general functional outcomes. Recent studies investigating EEG parameters related to motor impairment have consistently shown that low-frequency bands can serve as significant correlates or predictors of motor deficits, especially in the context of ULI following stroke, as measured by functional assessments such as FMA-UE and Actional Research Arm Test (ARAT) [ 16 , 17 , 26 , 53 – 56 ]. Saes et al. [ 17 , 53 ] demonstrated that the brain symmetry index in the delta and theta frequency bands is predominantly related to FMA-UE in chronic stroke and serves as a prognostic biomarker for FMA-UE six months post-stroke. In stroke patients with poor motor function (ARAT < 10 and no motor-evoked potential in the paretic hand), delta and theta band power in the ipsilesional hemisphere, as well as theta band power in the contralesional hemisphere, were found to be greater [ 56 ]. Additionally, interhemispheric coherence in the theta band and connectivity parameters such as degree centrality in the delta and theta bands in the contralesional hemisphere were higher in patients with poor motor function [ 56 ]. Low node strength of the delta band in the ipsilesional hemisphere is correlated with poor motor function (low FMA-UE score) [ 54 ]. Cassidy, Wodeyar [ 55 ] observed that a higher delta coherence between the ipsilesional primary motor cortex and bilateral areas (with the contralesional side showing more coherence than the ipsilesional side) correlated with worse motor function, as indicated by a lower FMA-UE score. Moreover, decreased delta coherence between the bilateral primary motor cortices is associated with motor recovery [ 55 ]. These findings align with those of our study, which further support the role of low-frequency band activity as a significant biomarker of motor impairment after stroke. Enhanced EEG power and coherence in low-frequency bands may serve as correlates that reveal the compensatory mechanisms following a stroke arising from anatomical or functional disruptions. However, these mechanisms may be insufficient for achieving meaningful motor recovery, rendering them maladaptive [ 55 – 57 ]. In addition, our data-driven approach underscores the importance of low-frequency bands across various regions, not only in motor areas but also in non-motor regions and across different hemispheres (ipsilesional, contralesional, and interhemispheric), as biomarkers reflecting motor impairments after stroke. This finding may reflect various aspects of the injury- or recovery-related process [ 16 ] and aligns with previous studies reporting contributions from widespread network modulation in non-motor regions, such as frontoparietal and attentional control networks [ 58 – 60 ]. The present study found a significant association between low-beta intrahemispheric connectivity in the contralesional parietal and temporoparietal regions and severe ULI, although only one of the 14 significant features was identified. Alterations in the low-beta frequency band during motor tasks following a stroke are well known. Low-beta event-related desynchronization (ERD) observed when moving the hemiparetic hand showed stronger ERD in the contralesional hemisphere than in healthy controls [ 61 ]. In addition, more efficient connectivity parameters of the low beta band during the task in the ipsilesional hemisphere correlated with better upper limb function in chronic stroke [ 21 ]. This suggests that activation in the contralesional hemisphere in the low beta band may serve as a compensatory mechanism following stroke and could potentially be maladaptive. However, few studies have investigated the association between low beta band activity and motor recovery after stroke, and results regarding beta activity in motor recovery have been inconsistent [ 16 ]. Therefore, further research is required to reveal low-beta connectivity and motor recovery. This study has some limitations. First, the sample size was insufficient to represent all chronic stroke survivors, which may limit the generalizability of the results. Thus, further studies with larger samples are needed. Second, individual brain imaging data were not co-registered with each subject's EEG channels, which may have resulted in deviations in the exact location of the brain region from its anatomical position and limited the ability to conduct brain source-level analysis despite our efforts to successfully fit the extended 10–10 system. Finally, we cannot guarantee that the selected EEG coherence features will serve as reliable biomarkers for motor recovery and responses to rehabilitative interventions, as the analysis was based solely on data from chronic stroke survivors [ 13 ]. Therefore, future research using larger longitudinal datasets that include repeated EEG and clinical measurements after early stroke is necessary [ 62 ]. With these larger longitudinal datasets, applying data-driven approaches using ML methods can reveal more comprehensive EEG features related to motor recovery after stroke that hypothesis-driven studies may overlook [ 16 ]. This approach also has the potential to develop predictive models for motor recovery or impairment that can be applied to new subjects outside the initial dataset, thereby enhancing the accuracy of predictions in clinical settings [ 63 ]. Conclusion The present data-driven analysis employing ML methods utilized whole-brain resting-state EEG connectivity features to classify severe ULI from mild-to-moderate ULI in chronic stroke survivors. Our research highlights the association between alterations in connectivity within low-frequency bands (delta and theta) and severe ULI across widespread brain regions, including areas outside the sensorimotor cortex and bilateral intrahemispheric and interhemispheric regions. Further research utilizing larger longitudinal datasets from patients with early strokes and applying data-driven approaches with ML methods could contribute to identifying more comprehensive EEG features related to motor recovery and the development of more accurate predictive models for motor recovery and rehabilitation responses. Abbreviations ULI upper limb impairment EEG electroencephalography ML machine learning SVM support vector machine FMA-UE Fugl-Meyer Assessment of the Upper Extremities iCoh imaginary part of coherency ANOVA analysis of variance RFE recursive feature elimination SGD stochastic gradient descent LOOCV leave-one-out cross-validation mRS modified rankin scale NIHSS National Institutes of Health Stroke Scale DTABR Higher delta theta/alphabeta ratios ARAT Academic research arm test ERD event-related desynchronization Declarations Ethics approval and consent to participate The Bundang Hospital Institutional Review Board (Registration number: B-1809-493-303) approved the study protocol in accordance with the principles of the Declaration of Helsinki. All participants understood the study procedure and provided written informed consent before participation. Consent for publication Written informed consent was obtained for data collection, analysis, and publication from all study participants. Availability of data and materials Pseudonymized data supporting the findings of this study are available from the corresponding author, Prof. Won-Seok Kim, upon reasonable request, subject to approval by the local IRB and upon the completion of a legal data-sharing agreement. Competing interests The authors declare that they have no competing interests. Funding This study was supported by a National Research Foundation of Korea (NRF) grant funded by the Korean government (Ministry of Science and ICT, MSIT) (NRF-2022R1A2C1006046) and a grant from the SNUBH Research Fund (Grant No: 18-2023-0007). It was also supported by the MSIT, Korea, under the Information Technology Research Center (ITRC) support program (IITP-2024-RS-2023-00258971) supervised by the Institute for Information & Communications Technology Planning & Evaluation (IITP). Authors’ contributions JYL analyzed and interpreted the data and wrote the manuscript. MSS contributed by validating the analytical methods used in the study. HJH, MSS, and WSK were involved in making critical revisions to the manuscript. WKC, HMC, JSC, HJK, BWS, and NJP supported interpreting the results and validating the manuscript. HJH and WSK contributed to the conceptualization and design of the study, as well as to securing project funding. All authors contributed critical feedback and gave their approval for the manuscript to be published. Acknowledgments Not applicable. References Gorelick PB. The global burden of stroke: persistent and disabling. Lancet Neurol. 2019;18(5):417–8. https://doi.org/10.1016/S1474-4422(19)30030-4 . Hatem SM, Saussez G, della Faille M, Prist V, Zhang X, Dispa D, et al. Rehabilitation of Motor Function after Stroke: A Multiple Systematic Review Focused on Techniques to Stimulate Upper Extremity Recovery. Front Hum Neurosci. 2016;10. https://doi.org/10.3389/fnhum.2016.00442 . Hayward KS, Kuys SS, Barker RN, Brauer SG. Clinically important improvements in motor function are achievable during inpatient rehabilitation by stroke patients with severe motor disability: A prospective observational study. NeuroRehabilitation. 2014;34:773–9. https://doi.org/10.3233/NRE-141076 . Shin J-H, Bog Park S, Ho Jang S. Effects of game-based virtual reality on health-related quality of life in chronic stroke patients: A randomized, controlled study. Comput Biol Med. 2015;63:92–8. https://doi.org/10.1016/j.compbiomed.2015.03.011 . Hayward KS, Schmidt J, Lohse KR, Peters S, Bernhardt J, Lannin NA, et al. Are we armed with the right data? Pooled individual data review of biomarkers in people with severe upper limb impairment after stroke. Neuroimage Clin. 2017;13:310–9. https://doi.org/10.1016/j.nicl.2016.09.015 . Lanzone J, Motolese F, Ricci L, Tecchio F, Tombini M, Zappasodi F, et al. Quantitative measures of the resting EEG in stroke: a systematic review on clinical correlation and prognostic value. Neurol Sci. 2023;44(12):4247–61. https://doi.org/10.1007/s10072-023-06981-9 . McLoughlin G, Makeig S, Tsuang MT. In search of biomarkers in psychiatry: EEG-based measures of brain function. Am J Med Genet B Neuropsychiatr Genet. 2014;165(2):111–21. https://doi.org/10.1002/ajmg.b.32208 . Milani G, Antonioni A, Baroni A, Malerba P, Straudi S, Relation Between EEG. Measures and Upper Limb Motor Recovery in Stroke Patients: A Scoping Review. Brain Topogr. 2022;35(5):651–66. https://doi.org/10.1007/s10548-022-00915-y . van Mierlo P, Papadopoulou M, Carrette E, Boon P, Vandenberghe S, Vonck K, et al. Functional brain connectivity from EEG in epilepsy: Seizure prediction and epileptogenic focus localization. Prog Neurobiol. 2014;121:19–35. https://doi.org/10.1016/j.pneurobio.2014.06.004 . Asadi B, Cuenca-Zaldivar JN, Nakhostin Ansari N, Ibáñez J, Herrero P, Calvo S. Brain Analysis with a Complex Network Approach in Stroke Patients Based on Electroencephalography: A Systematic Review and Meta-Analysis. Healthcare. 2023;11(5):666. https://doi.org/10.3390/healthcare11050666 . Ismail LE, Karwowski W. A Graph Theory-Based Modeling of Functional Brain Connectivity Based on EEG: A Systematic Review in the Context of Neuroergonomics. IEEE Access. 2020;8:155103–35. https://doi.org/10.1109/ACCESS.2020.3018995 . Rosazza C, Minati L. Resting-state brain networks: literature review and clinical applications. Neurol Sci. 2011;32(5):773–85. https://doi.org/10.1007/s10072-011-0636-y . Boyd LA, Hayward KS, Ward NS, Stinear CM, Rosso C, Fisher RJ, et al. Biomarkers of stroke recovery: Consensus-based core recommendations from the Stroke Recovery and Rehabilitation Roundtable. Int J Stroke. 2017;12(5):480–93. https://doi.org/10.1177/1747493017714176 . Vatinno AA, Simpson A, Ramakrishnan V, Bonilha HS, Bonilha L, Seo NJ. The Prognostic Utility of Electroencephalography in Stroke Recovery: A Systematic Review and Meta-Analysis. Neurorehabil Neural Repair. 2022;36(4–5):255–68. https://doi.org/10.1177/15459683221078294 . Carter AR, Shulman GL, Corbetta M. Why use a connectivity-based approach to study stroke and recovery of function? NeuroImage. 2012;62(4):2271–80. https://doi.org/10.1016/j.neuroimage.2012.02.070 . Cassidy JM, Wodeyar A, Srinivasan R, Cramer SC. Coherent neural oscillations inform early stroke motor recovery. Hum Brain Mapp. 2021;42(17):5636–47. https://doi.org/10.1002/hbm.25643 . Saes M, Meskers CGM, Daffertshofer A, van Wegen EEH, Kwakkel G. Are early measured resting-state EEG parameters predictive for upper limb motor impairment six months poststroke? Clin Neurophysiol. 2021;132(1):56–62. https://doi.org/10.1016/j.clinph.2020.09.031 . Wu J, Quinlan EB, Dodakian L, McKenzie A, Kathuria N, Zhou RJ, et al. Connectivity measures are robust biomarkers of cortical function and plasticity after stroke. Brain. 2015;138(8):2359–69. https://doi.org/10.1093/brain/awv156 . Vecchio F, Pappalettera C, Miraglia F, Deinite G, Manenti R, Judica E, et al. Prognostic Role of Hemispherical Functional Connectivity in Stroke: A Study via Graph Theory Versus Coherence of Electroencephalography Rhythms. Stroke. 2023;54(2):499–508. https://doi.org/10.1161/STROKEAHA.122.040747 . Vecchio F, Caliandro P, Reale G, Miraglia F, Piludu F, Masi G, et al. Acute cerebellar stroke and middle cerebral artery stroke exert distinctive modifications on functional cortical connectivity: A comparative study via EEG graph theory. Clin Neurophysiol. 2019;130(6):997–1007. https://doi.org/10.1016/j.clinph.2019.03.017 . Shim M, Choi G-Y, Paik N-J, Lim C, Hwang H-J, Kim W-S. Altered Functional Networks of Alpha and Low-Beta Bands During Upper Limb Movement and Association with Motor Impairment in Chronic Stroke. Brain Connect. 2021. https://doi.org/10.1089/brain.2021.0070 . Bistriceanu CE, Danciu FA, Cuciureanu DI. Cortical connectivity in stroke using signals from resting-state EEG: a review of current literature. Acta Neurol Belg. 2023;123(2):351–7. https://doi.org/10.1007/s13760-022-02102-z . Locatelli T, Cursi M, Liberati D, Franceschi M, Comi G. EEG coherence in Alzheimer's disease. Electroencephalogr Clin Neurophysiol. 1998;106(3):229–37. https://doi.org/10.1016/S0013-4694(97)00129-6 . Wu X, Zheng W-L, Li Z, Lu B-L. Investigating EEG-based functional connectivity patterns for multimodal emotion recognition. J Neural Eng. 2022;19(1):016012. https://doi.org/10.1088/1741-2552/ac49a7 . Lassi M, Bandini A, Spina V, Azzollini V, Dalise S, Mazzoni A, et al. editors. Classification of upper limb impairment in acute stroke patients using resting-state EEG markers and machine learning. 2023 11th International IEEE/EMBS Conference on Neural Engineering (NER); 2023 24–27 April 2023. Lin PJ, Li W, Zhai X, Li Z, Sun J, Xu Q, et al. Explainable Deep-Learning Prediction for Brain-Computer Interfaces Supported Lower Extremity Motor Gains Based on Multistate Fusion. IEEE Trans Neural Syst Rehabil Eng. 2024;32:1546–55. https://doi.org/10.1109/tnsre.2024.3384498 . Fugl-Meyer AR, Jääskö L, Leyman I, Olsson S, Steglind S. The post-stroke hemiplegic patient. 1. a method for evaluation of physical performance. Scand J Rehabil Med. 1975;7(1):13–31. Coscia M, Wessel MJ, Chaudary U, Millán JR, Micera S, Guggisberg A, et al. Neurotechnology-aided interventions for upper limb motor rehabilitation in severe chronic stroke. Brain. 2019;142(8):2182–97. https://doi.org/10.1093/brain/awz181 . De Vico Fallani F, Pichiorri F, Morone G, Molinari M, Babiloni F, Cincotti F, et al. Multiscale topological properties of functional brain networks during motor imagery after stroke. NeuroImage. 2013;83:438–49. https://doi.org/10.1016/j.neuroimage.2013.06.039 . Oostenveld R, Fries P, Maris E, Schoffelen J-M. FieldTrip: open source software for advanced analysis of MEG, EEG, and invasive electrophysiological data. Comput Intell Neurosci. 2011:2011:156869. https://doi.org/10.1155/2011/156869 Shan H, Xu H, Zhu S, He B. A novel channel selection method for optimal classification in different motor imagery BCI paradigms. Biomed Eng Online. 2015;14(1):93. https://doi.org/10.1186/s12938-015-0087-4 . Mazumder A, Ghosh P, Khasnobish A, Bhattacharyya S, Tibarewala D, editors. Selection of Relevant Features from Cognitive EEG Signals Using ReliefF and MRMR Algorithm. Advancements of Medical Electronics: Proceedings of the First International Conference, ICAME 2015; 2015: Springer. Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O et al. Scikit-learn: Machine Learning in Python. ArXiv. 2011;abs/1201.0490. Guyon I, Weston J, Barnhill S, Vapnik V. Gene Selection for Cancer Classification using Support Vector Machines. Mach Learn. 2002;46(1):389–422. https://doi.org/10.1023/A:1012487302797 . Hidalgo-Muñoz AR, López MM, Santos IM, Pereira AT, Vázquez-Marrufo M, Galvao-Carmona A, et al. Application of SVM-RFE on EEG signals for detecting the most relevant scalp regions linked to affective valence processing. Expert Syst Appl. 2013;40(6):2102–8. https://doi.org/10.1016/j.eswa.2012.10.013 . Urbanowicz RJ, Olson RS, Schmitt P, Meeker M, Moore JH. Benchmarking relief-based feature selection methods for bioinformatics data mining. J Biomed Inf. 2018;85:168–88. https://doi.org/10.1016/j.jbi.2018.07.015 . Zhang J, Chen M, Zhao S, Hu S, Shi Z, Cao Y. ReliefF-Based EEG Sensor Selection Methods for Emotion Recognition. Sensors. 2016;16(10):1558. https://doi.org/10.3390/s16101558 . Shim M, Lee S-H, Hwang H-J. Inflated prediction accuracy of neuropsychiatric biomarkers caused by data leakage in feature selection. Sci Rep. 2021;11(1):7980. https://doi.org/10.1038/s41598-021-87157-3 . Doerrfuss JI, Kilic T, Ahmadi M, Holtkamp M, Weber JE. Quantitative and Qualitative EEG as a Prediction Tool for Outcome and Complications in Acute Stroke Patients. Clin EEG Neurosci. 2020;51(2):121–9. https://doi.org/10.1177/1550059419875916 . Finnigan S, van Putten MJ. EEG in ischaemic stroke: quantitative EEG can uniquely inform (sub-)acute prognoses and clinical management. Clin Neurophysiol. 2013;124(1):10–9. https://doi.org/10.1016/j.clinph.2012.07.003 . Hall TM, de Carvalho F, Jackson AJN. A common structure underlies low-frequency cortical dynamics in movement, sleep, and sedation. Neuron N. 2014;83(5):1185–99. Schroeder CE, Lakatos PJT. Low-frequency neuronal oscillations as instruments of sensory selection. Trends Neurosci. 2009;32(1):9–18. https://doi.org/10.1016/j.tins.2008.09.012 . López ME, Pusil S, Pereda E, Maestú F, Barceló FJN. Dynamic low frequency EEG phase synchronization patterns during proactive control of task switching. NeuroImage. 2019;186:70–82. https://doi.org/10.1016/j.neuroimage.2018.10.068 . Ramanathan DS, Guo L, Gulati T, Davidson G, Hishinuma AK, Won S-J, et al. Low-frequency cortical activity is a neuromodulatory target that tracks recovery after stroke. Nat Med. 2018;24(8):1257–67. https://doi.org/10.1038/s41591-018-0058-y . Andraus ME, Alves-Leon SV. Non-epileptiform EEG abnormalities: an overview. Arq Neuropsiquiatr. 2011;69(5):829–35. https://doi.org/10.1590/s0004-282x2011000600020 . van Putten MJ, Tavy DL. Continuous quantitative EEG monitoring in hemispheric stroke patients using the brain symmetry index. Stroke. 2004;35(11):2489–92. https://doi.org/10.1161/01.STR.0000144649.49861.1d . Britton JW, Frey LC, Hopp JL, Korb P, Koubeissi MZ, Lievens WE, et al. Electroencephalography (EEG): An introductory text and atlas of normal and abnormal findings in adults, children, and infants. Chicago: American Epilepsy Society; 2016. Assenza G, Zappasodi F, Squitti R, Altamura C, Ventriglia M, Ercolani M, et al. Neuronal functionality assessed by magnetoencephalography is related to oxidative stress system in acute ischemic stroke. NeuroImage. 2009;44(4):1267–73. https://doi.org/10.1016/j.neuroimage.2008.09.049 . Finnigan SP, Walsh M, Rose SE, Chalk JB. Quantitative EEG indices of sub-acute ischaemic stroke correlate with clinical outcomes. Clin Neurophysiol. 2007;118(11):2525–32. https://doi.org/10.1016/j.clinph.2007.07.021 . Finnigan SP, Rose SE, Walsh M, Griffin M, Janke AL, McMahon KL, et al. Correlation of quantitative EEG in acute ischemic stroke with 30-day NIHSS score: comparison with diffusion and perfusion MRI. Stroke. 2004;35(4):899–903. https://doi.org/10.1161/01.STR.0000122622.73916.d2 . Bentes C, Peralta AR, Viana P, Martins H, Morgado C, Casimiro C, et al. Quantitative EEG and functional outcome following acute ischemic stroke. Clin Neurophysiol. 2018;129(8):1680–7. https://doi.org/10.1016/j.clinph.2018.05.021 . Sheorajpanday RV, Nagels G, Weeren AJ, van Putten MJ, De Deyn PP. Quantitative EEG in ischemic stroke: correlation with functional status after 6 months. Clin Neurophysiol. 2011;122(5):874–83. https://doi.org/10.1016/j.clinph.2010.07.028 . Saes M, Meskers CGM, Daffertshofer A, de Munck JC, Kwakkel G, van Wegen EEH. How does upper extremity Fugl-Meyer motor score relate to resting-state EEG in chronic stroke? A power spectral density analysis. Clin Neurophysiol. 2019;130(5):856–62. https://doi.org/10.1016/j.clinph.2019.01.007 . Zhang JJ, Bai Z, Fong KNK. Resting-state cortical electroencephalogram rhythms and network in patients after chronic stroke. J Neuroeng Rehabil. 2024;21(1):32. https://doi.org/10.1186/s12984-024-01328-7 . Cassidy JM, Wodeyar A, Wu J, Kaur K, Masuda AK, Srinivasan R, et al. Low-Frequency Oscillations Are a Biomarker of Injury and Recovery After Stroke. Stroke. 2020;51(5):1442–50. https://doi.org/10.1161/STROKEAHA.120.028932 . Ding Q, Chen J, Zhang S, Chen S, Li X, Peng Y, et al. Neurophysiological characterization of stroke recovery: A longitudinal TMS and EEG study. CNS Neurosci Ther. 2024;30(3):e14471. https://doi.org/10.1111/cns.14471 . Assenza G, Zappasodi F, Pasqualetti P, Vernieri F, Tecchio F. A contralesional EEG power increase mediated by interhemispheric disconnection provides negative prognosis in acute stroke. Restor Neurol Neurosci. 2013;31(2):177–88. https://doi.org/10.3233/rnn-120244 . Bönstrup M, Schulz R, Schön G, Cheng B, Feldheim J, Thomalla G, et al. Parietofrontal network upregulation after motor stroke. Neuroimage Clin. 2018;18:720–9. https://doi.org/10.1016/j.nicl.2018.03.006 . Hordacre B, Lotze M, Jenkinson M, Lazari A, Barras CD, Boyd L, et al. Fronto-parietal involvement in chronic stroke motor performance when corticospinal tract integrity is compromised. Neuroimage Clin. 2021;29:102558. https://doi.org/10.1016/j.nicl.2021.102558 . Rinne P, Hassan M, Fernandes C, Han E, Hennessy E, Waldman A, et al. Motor dexterity and strength depend upon integrity of the attention-control system. Proc Natl Acad Sci USA. 2018;115(3):E536–45. https://doi.org/10.1073/pnas.1715617115 . Stępień M, Conradi J, Waterstraat G, Hohlefeld FU, Curio G, Nikulin VV. Event-related desynchronization of sensorimotor EEG rhythms in hemiparetic patients with acute stroke. Neurosci Lett. 2011;488(1):17–21. https://doi.org/10.1016/j.neulet.2010.10.072 . Bernhardt J, Hayward KS, Kwakkel G, Ward NS, Wolf SL, Borschmann K, et al. Agreed definitions and a shared vision for new standards in stroke recovery research: The Stroke Recovery and Rehabilitation Roundtable taskforce. Int J Stroke. 2017;12(5):444–50. https://doi.org/10.1177/1747493017711816 . Ward NS. Restoring brain function after stroke - bridging the gap between animals and humans. Nat Rev Neurol. 2017;13(4):244–55. https://doi.org/10.1038/nrneurol.2017.34 . Additional Declarations No competing interests reported. Supplementary Files ChronicEEGULISupplementary250207.docx Cite Share Download PDF Status: Published Journal Publication published 22 Dec, 2025 Read the published version in Journal of NeuroEngineering and Rehabilitation → Version 1 posted Editorial decision: Revision requested 07 Apr, 2025 Reviews received at journal 02 Mar, 2025 Reviews received at journal 21 Feb, 2025 Reviewers agreed at journal 13 Feb, 2025 Reviewers agreed at journal 13 Feb, 2025 Reviewers agreed at journal 12 Feb, 2025 Reviewers invited by journal 11 Feb, 2025 Editor assigned by journal 07 Feb, 2025 Submission checks completed at journal 07 Feb, 2025 First submitted to journal 06 Feb, 2025 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-5976957","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":422817647,"identity":"7333b8ce-9121-4e63-8901-7a59035687ba","order_by":0,"name":"Ji-Yoon Lee","email":"","orcid":"","institution":"Seoul National University Bundang Hospital","correspondingAuthor":false,"prefix":"","firstName":"Ji-Yoon","middleName":"","lastName":"Lee","suffix":""},{"id":422817648,"identity":"c00c7680-fb60-4460-82e7-2f3faaf43e95","order_by":1,"name":"Miseon Shim","email":"","orcid":"","institution":"Tech University of Korea","correspondingAuthor":false,"prefix":"","firstName":"Miseon","middleName":"","lastName":"Shim","suffix":""},{"id":422817649,"identity":"0ccffb54-824f-44be-a1ec-3d1d479866ac","order_by":2,"name":"Won Kee Chang","email":"","orcid":"","institution":"Seoul National University Bundang Hospital","correspondingAuthor":false,"prefix":"","firstName":"Won","middleName":"Kee","lastName":"Chang","suffix":""},{"id":422817650,"identity":"13a47ef5-a42d-45d8-ae91-8f3782e62777","order_by":3,"name":"Hee-Mun Cho","email":"","orcid":"","institution":"Seoul National University Bundang Hospital","correspondingAuthor":false,"prefix":"","firstName":"Hee-Mun","middleName":"","lastName":"Cho","suffix":""},{"id":422817651,"identity":"76bbb16e-7418-4a87-bbef-b6aae281b807","order_by":4,"name":"Ji Soo Choi","email":"","orcid":"","institution":"Seoul National University Bundang Hospital","correspondingAuthor":false,"prefix":"","firstName":"Ji","middleName":"Soo","lastName":"Choi","suffix":""},{"id":422817652,"identity":"bc11f75b-3b2d-41f2-b8bf-f07599078ce1","order_by":5,"name":"Hyun Ji Kim","email":"","orcid":"","institution":"Seoul National University Bundang Hospital","correspondingAuthor":false,"prefix":"","firstName":"Hyun","middleName":"Ji","lastName":"Kim","suffix":""},{"id":422817653,"identity":"0deb21db-de52-4277-a6ca-496a42ebee56","order_by":6,"name":"Bong Won Suh","email":"","orcid":"","institution":"Seoul National University","correspondingAuthor":false,"prefix":"","firstName":"Bong","middleName":"Won","lastName":"Suh","suffix":""},{"id":422817654,"identity":"f2faa45a-e9b9-4ffa-af83-d97b849a43e3","order_by":7,"name":"Nam-Jong Paik","email":"","orcid":"","institution":"Seoul National University Bundang Hospital","correspondingAuthor":false,"prefix":"","firstName":"Nam-Jong","middleName":"","lastName":"Paik","suffix":""},{"id":422817655,"identity":"98c80b3b-32cf-403b-8f5d-d69e365f7ddf","order_by":8,"name":"Han-Jeong Hwang","email":"","orcid":"","institution":"Korea University","correspondingAuthor":false,"prefix":"","firstName":"Han-Jeong","middleName":"","lastName":"Hwang","suffix":""},{"id":422817656,"identity":"a87ca2c2-40ba-4e01-8786-f49a28f30048","order_by":9,"name":"Won-Seok Kim","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAxElEQVRIiWNgGAWjYJACCYaC//X9IFZCAdFaDJgZZzaAtBiQomXDARCTGC0Gx3sP3vhgwMZsfH514ocHBgzy/GIHCGg5cy7ZcoYBD5vZjbebJYAOM5w5OwG/FrMbOWbSPAYSPGY3zm4AaUkwuE2cFgMJ4xlnN/8gRUuCgQF/7zbibLE/c8YY6JcDCRI3eLdZJBhIEPaLZHuP4Y0PFQcS+PvPbr75o8JGnl+agBYEkACrlCBWOQjwHyBF9SgYBaNgFIwkAADZEUKHcT3jSwAAAABJRU5ErkJggg==","orcid":"","institution":"Seoul National University Bundang Hospital","correspondingAuthor":true,"prefix":"","firstName":"Won-Seok","middleName":"","lastName":"Kim","suffix":""}],"badges":[],"createdAt":"2025-02-07 01:53:20","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5976957/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5976957/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12984-025-01859-7","type":"published","date":"2025-12-22T15:58:02+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":78248736,"identity":"ea613be3-0051-4007-a8ed-c20ab1d19448","added_by":"auto","created_at":"2025-03-11 09:39:45","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":392503,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConfiguration of electrode positions and flowchart of the proposed machine-learning approach, including feature selection\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Configuration of the electrodes used for coherence analysis in this study. (B) Flowchart illustrating the evaluation of the models using various algorithms and feature selection methods with leave-one-out cross-validation.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAbbreviations:\u003c/em\u003e\u003cem\u003e\u003cstrong\u003e \u003c/strong\u003e\u003c/em\u003e\u003cem\u003eEEG, electroencephalography; num, the number of features; n, the number of participants; ANOVA, analysis of variance; SVM, support vector machine; RFE, recursive feature elimination; SGD, stochastic gradient descent.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Fig.1.png","url":"https://assets-eu.researchsquare.com/files/rs-5976957/v1/c7ccf8ee74f16ee723cb8278.png"},{"id":78246742,"identity":"94588bca-948e-46fc-86c7-48e8cfeff5dc","added_by":"auto","created_at":"2025-03-11 09:31:45","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":719947,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eBalanced classification accuracies of the four classification methods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLogistic regression with (A) L1, (B) elastic net, (C) SGD, and (D) SVM for the three feature selection methods (ANOVA, SVM-RFE, and ReliefF).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAbbreviations: SGD, stochastic gradient descent; SVM, support vector machine.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Fig.2.png","url":"https://assets-eu.researchsquare.com/files/rs-5976957/v1/95d348888c9d15b8cd65fdc1.png"},{"id":78248737,"identity":"562e457c-8426-43b4-a64a-9148e2619a13","added_by":"auto","created_at":"2025-03-11 09:39:45","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":542900,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTopographic maps of the top fourteen functional connectivity between the mild/moderate ULI and severe ULI groups\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Delta band, (B) theta band, and (C) low beta band. The intensity of the red line indicates the occurrence rate, which represents the number of features appearing in a cross-validation iteration.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAbbreviations: ipsi, ipsilesional; con, contralesional.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Fig.3.png","url":"https://assets-eu.researchsquare.com/files/rs-5976957/v1/730075ddb795ab62111a3b6c.png"},{"id":99172457,"identity":"43d22b36-7289-4823-9cd2-9094b9dd5884","added_by":"auto","created_at":"2025-12-29 16:09:46","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2253503,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5976957/v1/88c17315-9cb4-4617-a432-e9a6d850568d.pdf"},{"id":78246739,"identity":"9a5e0d73-abfd-4bec-a882-5e9297ff455f","added_by":"auto","created_at":"2025-03-11 09:31:45","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":39027,"visible":true,"origin":"","legend":"","description":"","filename":"ChronicEEGULISupplementary250207.docx","url":"https://assets-eu.researchsquare.com/files/rs-5976957/v1/2d54075f9f75448c1fe9200c.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Functional Connectivity Associated with Severe Upper Limb Impairment in Resting-State Electroencephalography Among Chronic Stroke Survivors: A Machine Learning Approach","fulltext":[{"header":"Background","content":"\u003cp\u003eStroke is a major health concern owing to its high prevalence and significant impact on disability [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. According to the World Health Organization, over 60% of chronic stroke survivors continue to suffer from upper limb impairment (ULI) [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Improving upper limb function is a key rehabilitation goal; however, individuals with severe ULI face substantial challenges in achieving the desired outcomes [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. The impact of severe ULI on the survivors\u0026rsquo; quality of life is particularly significant [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e \u003cp\u003ePrevious research emphasizes the importance of identifying biomarkers to develop targeted and efficient interventions for stroke survivors with severe ULI [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Electroencephalography (EEG) is advantageous for biomarker development because of its noninvasive nature, excellent temporal resolution, and cost-effectiveness among neuroimaging methods [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. In a scoping review conducted by Milani, Antonioni [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e], EEG-derived measures proved valuable for evaluating upper limb recovery. However, most studies have concentrated on features related to power spectrum density; a notable gap remains in our understanding of how different brain regions connect, process, and share information in stroke survivors [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eUnderstanding these connections is crucial because functional connectivity can serve as a valuable marker for monitoring motor recovery. Defined as the statistical interdependencies between physiological time series from different brain regions, functional connectivity plays a crucial role in brain function [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Furthermore, the brain remains active even during rest, displaying spontaneous activity that is not linked to external stimuli or responses but arises internally [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Therefore, resting-state functional connectivity shows promise for developing biomarkers to monitor motor recovery post-stroke [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Carter, Shulman [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e] emphasized that stroke recovery depends on how the brain\u0026rsquo;s cortical and subcortical networks reorganize, not just during active tasks but also in a resting state. Therefore, resting-state connectivity may act as a potential biomarker of motor impairment following stroke [\u003cspan additionalcitationids=\"CR17\" citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eAdditionally, the neural mechanisms underlying recovery after stroke, which affect the ipsilesional and contralesional brain hemispheres, are not yet fully understood [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. Studies have reported that acute stroke survivors exhibit altered EEG inter-hemispheric connectivity and modifications in the balance between local segregation and global integration [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. Employing a hemispheric-focused network analysis is crucial for deeply understanding cost-effective neuronal compensatory mechanisms in stroke survivors [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. While many studies have concentrated on the connectivity of specific brain regions [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e], understanding of comprehensive inter- and intra-hemispheric connectivity patterns across all brain regions remains limited.\u003c/p\u003e \u003cp\u003eEEG coherence, which indicates the functional connectivity between different electrode sites, reflects the degree of signal synchronization and suggests functional interaction [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. The complexity of EEG coherence, characterized by the complexity, multichannel, multifrequency, and inherent complexity of brain function and structure, introduces a challenge in comprehensively analyzing resting-state EEG connectivity. Machine learning (ML) could play a crucial role in analyzing EEG data, which incurs significant computational costs and time owing to the vast amount of data and potential redundancy among EEG features [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. A recent study by Lassi and Bandini [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e] demonstrated the effective use of ML and feature-selection methods to classify ULI severity in acute stroke survivors. Their best model, based on a support vector machine (SVM), achieved over 85% accuracy, with the brain symmetry index identified as a key feature. High-dimensional data, such as resting-state coherence analyzed across the whole brain, may reduce the model performance and complicate interpretations. However, by employing optimal ML feature selection techniques, we can condense the features, thereby improving model accuracy and simplifying the interpretation of neural patterns associated with severe ULI. Compared with the conventional hypothesis-driven approach, this data-driven ML approach enables a more comprehensive exploration of the role of EEG coherence in motor impairment after stroke [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThis study aimed to identify resting-state EEG connectivity patterns related to ULI severity in chronic stroke survivors, as indicated by the coherence between each channel in the ipsilesional and contralesional hemispheres. First, we attempted to determine the best ML model for classifying the severity of ULI\u0026mdash;severe ULI vs. mild/moderate ULI\u0026mdash;and then attempted to identify the EEG connectivity features critical for distinguishing ULI severity in the ML model.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eParticipants\u003c/h2\u003e \u003cp\u003eThirty-four stroke survivors, aged between 29 and 80, were recruited in the study from two rehabilitation hospitals between June and December 2019. Participants were selected based on the following criteria: 1) diagnosis of first-ever ischemic or hemorrhagic stroke confirmed by brain computed tomography or magnetic resonance imaging, 2) unilateral upper limb weakness, 3) a minimum of 6 months post-stroke, and 4) the ability to provide written informed consent. Participants were excluded if they met any of the following conditions: 1) a history of central nervous system conditions such as traumatic brain injury, brain tumor, or Parkinson's disease; 2) inability to wear an EEG cap; or 3) difficulty in comprehending and adhering to clinical assessment and EEG study protocols. A clinician reviewed the structural images of the participants and classified their stroke locations into three categories: cortical, subcortical, including corona radiata, internal capsule, or basal ganglia, or a combination of cortical and subcortical.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eAssessment for upper limb motor impairment\u003c/h3\u003e\n\u003cp\u003eThe Fugl-Meyer assessment for the upper extremity (FMA-UE) was used to assess upper limb motor function [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. The FMA-UE is an objective measure used to assess motor recovery outcomes, with scores ranging from 0 to 66, with higher scores indicating better motor function and recovery. In the study by Coscia, Wessel [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e], ULI severity was classified based on FMA-UE scores: \u0026gt;45 for mild, 30\u0026ndash;45 for moderate, and \u0026lt;\u0026thinsp;30 for severe ULI. In our study, we adopted an FMA-UE threshold of 30 to classify participants into two groups with fairly balanced sample sizes: severe ULI (n\u0026thinsp;=\u0026thinsp;15) and mild/moderate ULI (n\u0026thinsp;=\u0026thinsp;19). Baseline group differences are presented in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. Age and FMA-UE were the only variables that reached statistical significance (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDemographic characteristics of participants\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMild/moderate ULI\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;19)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSevere ULI\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;15)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge (years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e57.58\u0026thinsp;\u0026plusmn;\u0026thinsp;13.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e66.93\u0026thinsp;\u0026plusmn;\u0026thinsp;8.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.03\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFMA-UE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e47 (14.50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9 (12)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTime since stroke onset (days)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e315 (185.50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e321 (159.50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.56\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSex\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e10 (52.60%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5 (33.30%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e0.44\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e9 (47.40%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10 (66.70%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHemiplegia side\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLeft\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8 (42.10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e11 (73.30%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e0.14\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRight\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e11 (57.90%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (26.70%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLegion location\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCortical\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6 (31.60%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (13.30%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e0.46\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSubcortical\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e11 (57.90%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e11 (73.30%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCortical / subcortical\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (10.50%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (13.30%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003eContinuous variables were analyzed using the Mann-Whitney U test or Student\u0026rsquo;s t-test based on the Shapiro-Wilk normality test. Categorical variables were assessed using the chi-squared test. Continuous data are presented as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation or median (interquartile range), whereas categorical data are presented as numbers (percentages). Statistical significance was set at p\u0026thinsp;\u0026lt;\u0026thinsp;0.05. significant.\u003c/td\u003e\u003c/tr\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003e\u003cem\u003eAbbreviations: FMA-UE, the Fugl-Meyer assessment for upper extremity.\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e\n\u003ch3\u003eEEG recording and preprocessing\u003c/h3\u003e\n\u003cp\u003eParticipants were seated comfortably in an armchair in front of a monitor to facilitate EEG recordings. To minimize eye movement artifacts and ensure the quality of the EEG recordings, the participants were instructed to focus their gaze on a fixation cross displayed on the monitor. EEG data were collected during two resting-state sessions for a total of 4 minutes (one minute with eyes closed, followed by one minute with eyes open, and repeated twice). A movement-related EEG session was conducted between resting-state sessions.\u003c/p\u003e \u003cp\u003eEEG data were recorded at a sampling rate of 1000 Hz using 32 Ag/AgCl scalp electrodes according to the extended international 10\u0026ndash;10 system (Brain Products, GmbH Ltd., Gilching, Germany) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). The ground electrode was placed between Fp1 and Fp2, and the reference electrode at Cz. The EEG data were band-pass filtered from 0.1 to 55 Hz, and eye-related artifacts were removed by using mathematical procedures based on the first principal component analysis via Curry 7 (Compumedics, USA) software. We analyzed two sets of artifact-free eyes-closed data from the first resting-state session to evaluate the intrinsic brain activity. The data were downsampled to 500 Hz and re-referenced using a common average reference. Each eyes-closed set was divided into 4-second epochs. The first 56 s were used, excluding the last four seconds due to data quality issues, resulting in twenty-eight epochs for analysis.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eFeature extraction\u003c/h3\u003e\n\u003cp\u003eFor feature extraction, we used EEG electrodes from the left and right hemispheres, excluding the midline electrodes (Fz, Cz, Pz, and Oz) because our focus was on investigating inter- and intrahemispheric brain connectivity. To calculate functional connectivity, the imaginary part of coherency (iCoh) was computed between the 28 electrodes for each EEG epoch and then averaged across all epochs. iCoh is widely used to avoid volume conduction effects caused by the different conductivities of brain layers (e.g., CSF, skull, and scalp), preserving the original signals from the cortical surface in the functional connectivity pattern [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. The iCoh features were computed in six frequency bands: delta (1\u0026ndash;4 Hz), theta (4\u0026ndash;8 Hz), alpha (8\u0026ndash;12 Hz), low-beta (12\u0026ndash;20 Hz), high-beta (20\u0026ndash;30 Hz), and gamma (30\u0026ndash;55 Hz). Finally, a total of 2,268 iCoh values (\u003csub\u003e28\u003c/sub\u003eC\u003csub\u003e2\u003c/sub\u003e \u0026times; 6 frequency bands) were used for further analyses. The channels for coherence were reorganized to highlight the effects corresponding to the locations of stroke lesions in each hemisphere. The Matlab-based Fieldtrip toolbox was used to calculate the iCoh values [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e].\u003c/p\u003e\n\u003ch3\u003eFeature selection and ML-based classification\u003c/h3\u003e\n\u003cp\u003eA total of 2,268 iCoh values were used to classify groups with different levels of ULI severity; however, a high number of features relative to the number of samples (participants) could lead to overfitting. To mitigate this, we adopted feature-selection methods to optimize the features for classifying the two groups. Three feature selection methods are widely used in EEG studies: analysis of variance (ANOVA), SVM-recursive feature elimination (RFE), and ReliefF were used [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. ANOVA employs F-values to select features by considering the relationship between the label and each feature [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. SVM-RFE handles many features without requiring statistical corrections and considers individual differences by eliminating irrelevant or extreme data points [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. ReliefF, a category of relief-based feature selection methods, assesses the importance of features by their ability to distinguish between neighboring data points [\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. All features were normalized prior to feature selection. Feature selection was performed in each iteration of the cross-validation process to prevent overfitting and reduce potential selection bias [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. Preliminary analyses showed that expanding the candidate feature sets beyond 100 did not significantly improve the classification performance. Consequently, we concentrated on a subset of features, selecting between one and 100 features for further analysis.\u003c/p\u003e \u003cp\u003eThe classification performance was evaluated using four ML algorithms: logistic regression with L1, elastic net regularization, stochastic gradient descent (SGD), and SVM. Hyperparameters for each algorithm are listed in Supplementary Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e. Leave-one-out cross-validation (LOOCV) was used to prevent overfitting problems and biases due to the small sample size (n\u0026thinsp;=\u0026thinsp;34). Important features were identified based on their occurrence rate, which was calculated as the ratio of their appearances to the total number of LOOCV iterations. The ML analysis was conducted using Python 3.11 (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.python.org/\u003c/span\u003e\u003cspan address=\"https://www.python.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and its compatible open-source libraries. A flowchart of the proposed ML approach is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. B.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eModel performance\u003c/h2\u003e \u003cp\u003eThe performances of all possible combinations of the three-feature selection and classification methods are detailed in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. Logistic regression with the L1 model combined with the ReliefF feature selection method showed superior performance in classifying ULI severity in chronic stroke survivors (accuracy\u0026thinsp;=\u0026thinsp;0.91, sensitivity\u0026thinsp;=\u0026thinsp;0.93, and specificity\u0026thinsp;=\u0026thinsp;0.90 when using 14 features).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eModel performance and the number of selected features for each algorithm and feature selection method\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"13\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c13\" colnum=\"13\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c5\" namest=\"c2\"\u003e \u003cp\u003eANOVA\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003eSVM-RFE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c13\" namest=\"c10\"\u003e \u003cp\u003eReliefF\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNum\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBal. Accu.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSens.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSpec.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNum\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eBal. Accu.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eSens.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eSpec.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eNum\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003eBal. Accu.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e \u003cp\u003eSens.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c13\"\u003e \u003cp\u003eSpec.\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLogistic regression with L1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.71\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e\u003cb\u003e14\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e\u003cb\u003e0.91\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e\u003cb\u003e0.93\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e\u003cb\u003e0.90\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLogistic regression with elastic net\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e0.90\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLogistic regression with SGD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e0.84\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSVM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.88\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e0.95\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"13\"\u003eBold font indicates the highest model performance across all algorithms and feature selection methods.\u003c/td\u003e\u003c/tr\u003e \u003ctr\u003e\u003ctd colspan=\"13\"\u003e\u003cem\u003eAbbreviations: ANOVA, analysis of variance; SVM, support vector machine; RFE, recursive feature elimination; Num, the number of features; Bal. Accu., balanced accuracy; Sens, sensitivity; Spec, specificity; SGD, stochastic gradient descent.\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eImportant EEG features\u003c/h3\u003e\n\u003cp\u003eLogistic regression with L1 and the ReliefF method identified 30 features from 14 input variables within LOOCV, as shown in Supplementary Table S2. Among these, we focused on important features with empirically determined occurrence rates exceeding 0.5. Important features were identified, including three delta, ten theta, and one low-beta coherences whereas alpha and high-beta coherences were not selected (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003csup\u003e\u003c/sup\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eOccurrence rate of top fourteen features in the optimal model distinguishing severe from mild/moderate ULI.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"2\" morerows=\"1\" nameend=\"c2\" namest=\"c1\" rowspan=\"2\"\u003e \u003cp\u003eOptimal model\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAlgorithm\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003eLogistic regression with L1\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003eFeature selection method\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003eReliefF\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSelected features\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003eOccurrence rate\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c5\" namest=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDelta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eTP.ipsi - F.con\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDelta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eCP.ipsi-P.ipsi\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLow-beta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eP.con-TP.con\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTheta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eTP.con-FP.con\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTheta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eF.con-FP.con\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTheta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eFC.ipsi-CP.ipsi\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTheta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eCP.ipsi-FP.con\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTheta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eF.ipsi-P.ipsi\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTheta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eFC.ipsi-P.ipsi\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTheta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eF.ipsi-CP.ipsi\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTheta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eFC.ipsi-P.ipsi\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e0.91\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDelta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eT.ipsi-CP.ipsi\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e0.79\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTheta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eFC.ipsi-P\u003csub\u003e=\u003c/sub\u003e.con\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e0.79\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTheta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eP.ipsi-FP.con\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003e0.62\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003eThe occurrence rate is computed as the number of appearances divided by the total number of cross-validation iterations.\u003c/td\u003e\u003c/tr\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003e\u003cem\u003eAbbreviations: FMA-UE, the Fugl-Meyer assessment for upper extremity; channel number; ipsi, ipsilesional; con, contralesional; FP, prefrontal; F, frontal; FC, frontocentral; FT, frontotemporal; T, temporal; TP, temporoparietal; C, central; CP, centroparietal; P, parietal; O, occipital.\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWe identified significant features, including the delta interhemispheric connectivity of the ipsilesional temporoparietal and contralesional frontal regions. Additionally, delta connectivity was observed with intrahemispheric connectivity, specifically between the ipsilesional temporal and centroparietal regions as well as the ipsilesional centroparietal and parietal areas. Furthermore, theta interhemispheric connectivity was found between the ipsilesional centroparietal and contralesional prefrontal regions, the ipsilesional frontocentral and contralesional parietal regions, and the ipsilesional parietal and contralesional prefrontal regions. Regarding intrahemispheric connectivity, theta coherence was noted between the ipsilesional frontocentral and centroparietal regions, ipsilesional frontal and parietal regions, ipsilesional frontocentral and parietal regions, and ipsilesional frontal and centroparietal regions. In the contralesional hemisphere, theta coherence was observed between the temporoparietal and prefrontal regions, as well as between the frontal and prefrontal regions. Intrahemispheric coherence in the low beta band was also found between the contralesional parietal and temporoparietal regions.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eAchieving a balanced accuracy of approximately 91%, the ML approach in this study achieved a higher model performance in classifying severe ULI in chronic stroke survivors, notably by utilizing whole-brain resting-state EEG coherence features from all channels. Fourteen significant coherence features were identified across the delta, theta, and low-beta frequency bands most frequently selected in LOOCV.\u003c/p\u003e \u003cp\u003eMost resting-state EEG coherence features that substantially contribute to distinguishing severe ULI are distributed in low-frequency bands, specifically the delta and theta bands. Logistic regression with L1 significantly selected delta interhemispheric connectivity between the ipsilesional temporoparietal and contralesional frontal regions as well as intrahemispheric connectivity within the ipsilesional temporal, centroparietal, and parietal areas. Significant theta inter-hemispheric and intra-hemispheric connectivity was observed in the frontal and parietal regions. These findings align with those of previous studies reporting that worse outcomes after stroke were associated with low-frequency oscillations [\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]. Low-frequency oscillations are associated with cognitive control, attentional processing, movement, and skilled motor control [\u003cspan additionalcitationids=\"CR42 CR43\" citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e]. Brain damage following stroke, brain tumors, or traumatic brain injury exhibits altered activity in the delta and theta frequency bands [\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e], with more pronounced slowing observed in cases of more severe brain damage [\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e]. Alterations in low-frequency bands are also associated with general stroke-related outcomes, such as the modified Rankin Scale (mRS) and the National Institutes of Health Stroke Scale (NIHSS). Higher delta power in the ipsilateral hemisphere [\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e], a higher delta power ratio [\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e], and increased delta power during the acute stroke phase [\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e] were associated with higher NIHSS scores, indicating worse functional outcomes. Furthermore, a higher delta-theta/alpha-beta ratio (DTABR) was linked to worse functional outcomes, as reflected by higher mRS scores [\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e, \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e]. Therefore, changes in delta- or theta-band frequencies on EEG are recognized as consistent and important biomarkers of stroke severity and general functional outcomes.\u003c/p\u003e \u003cp\u003eRecent studies investigating EEG parameters related to motor impairment have consistently shown that low-frequency bands can serve as significant correlates or predictors of motor deficits, especially in the context of ULI following stroke, as measured by functional assessments such as FMA-UE and Actional Research Arm Test (ARAT) [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan additionalcitationids=\"CR54 CR55\" citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e]. Saes \u003cem\u003eet al.\u003c/em\u003e [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e] demonstrated that the brain symmetry index in the delta and theta frequency bands is predominantly related to FMA-UE in chronic stroke and serves as a prognostic biomarker for FMA-UE six months post-stroke. In stroke patients with poor motor function (ARAT\u0026thinsp;\u0026lt;\u0026thinsp;10 and no motor-evoked potential in the paretic hand), delta and theta band power in the ipsilesional hemisphere, as well as theta band power in the contralesional hemisphere, were found to be greater [\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e]. Additionally, interhemispheric coherence in the theta band and connectivity parameters such as degree centrality in the delta and theta bands in the contralesional hemisphere were higher in patients with poor motor function [\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e]. Low node strength of the delta band in the ipsilesional hemisphere is correlated with poor motor function (low FMA-UE score) [\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e]. Cassidy, Wodeyar [\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e] observed that a higher delta coherence between the ipsilesional primary motor cortex and bilateral areas (with the contralesional side showing more coherence than the ipsilesional side) correlated with worse motor function, as indicated by a lower FMA-UE score. Moreover, decreased delta coherence between the bilateral primary motor cortices is associated with motor recovery [\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e]. These findings align with those of our study, which further support the role of low-frequency band activity as a significant biomarker of motor impairment after stroke. Enhanced EEG power and coherence in low-frequency bands may serve as correlates that reveal the compensatory mechanisms following a stroke arising from anatomical or functional disruptions. However, these mechanisms may be insufficient for achieving meaningful motor recovery, rendering them maladaptive [\u003cspan additionalcitationids=\"CR56\" citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e]. In addition, our data-driven approach underscores the importance of low-frequency bands across various regions, not only in motor areas but also in non-motor regions and across different hemispheres (ipsilesional, contralesional, and interhemispheric), as biomarkers reflecting motor impairments after stroke. This finding may reflect various aspects of the injury- or recovery-related process [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e] and aligns with previous studies reporting contributions from widespread network modulation in non-motor regions, such as frontoparietal and attentional control networks [\u003cspan additionalcitationids=\"CR59\" citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe present study found a significant association between low-beta intrahemispheric connectivity in the contralesional parietal and temporoparietal regions and severe ULI, although only one of the 14 significant features was identified. Alterations in the low-beta frequency band during motor tasks following a stroke are well known. Low-beta event-related desynchronization (ERD) observed when moving the hemiparetic hand showed stronger ERD in the contralesional hemisphere than in healthy controls [\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e]. In addition, more efficient connectivity parameters of the low beta band during the task in the ipsilesional hemisphere correlated with better upper limb function in chronic stroke [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. This suggests that activation in the contralesional hemisphere in the low beta band may serve as a compensatory mechanism following stroke and could potentially be maladaptive. However, few studies have investigated the association between low beta band activity and motor recovery after stroke, and results regarding beta activity in motor recovery have been inconsistent [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. Therefore, further research is required to reveal low-beta connectivity and motor recovery.\u003c/p\u003e \u003cp\u003eThis study has some limitations. First, the sample size was insufficient to represent all chronic stroke survivors, which may limit the generalizability of the results. Thus, further studies with larger samples are needed. Second, individual brain imaging data were not co-registered with each subject's EEG channels, which may have resulted in deviations in the exact location of the brain region from its anatomical position and limited the ability to conduct brain source-level analysis despite our efforts to successfully fit the extended 10\u0026ndash;10 system. Finally, we cannot guarantee that the selected EEG coherence features will serve as reliable biomarkers for motor recovery and responses to rehabilitative interventions, as the analysis was based solely on data from chronic stroke survivors [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Therefore, future research using larger longitudinal datasets that include repeated EEG and clinical measurements after early stroke is necessary [\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e]. With these larger longitudinal datasets, applying data-driven approaches using ML methods can reveal more comprehensive EEG features related to motor recovery after stroke that hypothesis-driven studies may overlook [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. This approach also has the potential to develop predictive models for motor recovery or impairment that can be applied to new subjects outside the initial dataset, thereby enhancing the accuracy of predictions in clinical settings [\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e].\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThe present data-driven analysis employing ML methods utilized whole-brain resting-state EEG connectivity features to classify severe ULI from mild-to-moderate ULI in chronic stroke survivors. Our research highlights the association between alterations in connectivity within low-frequency bands (delta and theta) and severe ULI across widespread brain regions, including areas outside the sensorimotor cortex and bilateral intrahemispheric and interhemispheric regions. Further research utilizing larger longitudinal datasets from patients with early strokes and applying data-driven approaches with ML methods could contribute to identifying more comprehensive EEG features related to motor recovery and the development of more accurate predictive models for motor recovery and rehabilitation responses.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eULI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eupper limb impairment\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eEEG\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eelectroencephalography\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eML\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003emachine learning\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eSVM\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003esupport vector machine\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eFMA-UE\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eFugl-Meyer Assessment of the Upper Extremities\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eiCoh\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eimaginary part of coherency\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eANOVA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eanalysis of variance\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eRFE\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003erecursive feature elimination\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eSGD\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003estochastic gradient descent\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eLOOCV\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eleave-one-out cross-validation\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003emRS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003emodified rankin scale\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eNIHSS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eNational Institutes of Health Stroke Scale\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDTABR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eHigher delta theta/alphabeta ratios\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eARAT\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eAcademic research arm test\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eERD\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eevent-related desynchronization\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cem\u003eEthics approval and consent to participate\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe Bundang Hospital Institutional Review Board (Registration number: B-1809-493-303) approved the study protocol in accordance with the principles of the Declaration of Helsinki. All participants understood the study procedure and provided written informed consent before participation.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eConsent for publication\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eWritten informed consent was obtained for data collection, analysis, and publication from all study participants.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAvailability of data and materials\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003ePseudonymized data supporting the findings of this study are available from the corresponding author, Prof. Won-Seok Kim, upon reasonable request, subject to approval by the local IRB and upon the completion of a legal data-sharing agreement.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCompeting interests\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eFunding\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThis study was supported by a National Research Foundation of Korea (NRF) grant funded by the Korean government (Ministry of Science and ICT, MSIT) (NRF-2022R1A2C1006046) and a grant from the SNUBH Research Fund (Grant No: 18-2023-0007). It was also supported by the MSIT, Korea, under the Information Technology Research Center (ITRC) support program (IITP-2024-RS-2023-00258971) supervised by the Institute for Information \u0026amp; Communications Technology Planning \u0026amp; Evaluation (IITP).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAuthors\u0026rsquo; contributions\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eJYL analyzed and interpreted the data and wrote the manuscript. MSS contributed by validating the analytical methods used in the study. HJH, MSS, and WSK were involved in making critical revisions to the manuscript. WKC, HMC, JSC, HJK, BWS, and NJP supported interpreting the results and validating the manuscript. HJH and WSK contributed to the conceptualization and design of the study, as well as to securing project funding. All authors contributed critical feedback and gave their approval for the manuscript to be published.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAcknowledgments\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eGorelick PB. The global burden of stroke: persistent and disabling. Lancet Neurol. 2019;18(5):417\u0026ndash;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S1474-4422(19)30030-4\u003c/span\u003e\u003cspan address=\"10.1016/S1474-4422(19)30030-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHatem SM, Saussez G, della Faille M, Prist V, Zhang X, Dispa D, et al. Rehabilitation of Motor Function after Stroke: A Multiple Systematic Review Focused on Techniques to Stimulate Upper Extremity Recovery. Front Hum Neurosci. 2016;10. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fnhum.2016.00442\u003c/span\u003e\u003cspan address=\"10.3389/fnhum.2016.00442\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHayward KS, Kuys SS, Barker RN, Brauer SG. Clinically important improvements in motor function are achievable during inpatient rehabilitation by stroke patients with severe motor disability: A prospective observational study. NeuroRehabilitation. 2014;34:773\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3233/NRE-141076\u003c/span\u003e\u003cspan address=\"10.3233/NRE-141076\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShin J-H, Bog Park S, Ho Jang S. Effects of game-based virtual reality on health-related quality of life in chronic stroke patients: A randomized, controlled study. Comput Biol Med. 2015;63:92\u0026ndash;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.compbiomed.2015.03.011\u003c/span\u003e\u003cspan address=\"10.1016/j.compbiomed.2015.03.011\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHayward KS, Schmidt J, Lohse KR, Peters S, Bernhardt J, Lannin NA, et al. Are we armed with the right data? Pooled individual data review of biomarkers in people with severe upper limb impairment after stroke. Neuroimage Clin. 2017;13:310\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.nicl.2016.09.015\u003c/span\u003e\u003cspan address=\"10.1016/j.nicl.2016.09.015\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLanzone J, Motolese F, Ricci L, Tecchio F, Tombini M, Zappasodi F, et al. Quantitative measures of the resting EEG in stroke: a systematic review on clinical correlation and prognostic value. Neurol Sci. 2023;44(12):4247\u0026ndash;61. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10072-023-06981-9\u003c/span\u003e\u003cspan address=\"10.1007/s10072-023-06981-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcLoughlin G, Makeig S, Tsuang MT. In search of biomarkers in psychiatry: EEG-based measures of brain function. Am J Med Genet B Neuropsychiatr Genet. 2014;165(2):111\u0026ndash;21. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/ajmg.b.32208\u003c/span\u003e\u003cspan address=\"10.1002/ajmg.b.32208\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMilani G, Antonioni A, Baroni A, Malerba P, Straudi S, Relation Between EEG. Measures and Upper Limb Motor Recovery in Stroke Patients: A Scoping Review. Brain Topogr. 2022;35(5):651\u0026ndash;66. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10548-022-00915-y\u003c/span\u003e\u003cspan address=\"10.1007/s10548-022-00915-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan Mierlo P, Papadopoulou M, Carrette E, Boon P, Vandenberghe S, Vonck K, et al. Functional brain connectivity from EEG in epilepsy: Seizure prediction and epileptogenic focus localization. Prog Neurobiol. 2014;121:19\u0026ndash;35. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.pneurobio.2014.06.004\u003c/span\u003e\u003cspan address=\"10.1016/j.pneurobio.2014.06.004\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAsadi B, Cuenca-Zaldivar JN, Nakhostin Ansari N, Ib\u0026aacute;\u0026ntilde;ez J, Herrero P, Calvo S. Brain Analysis with a Complex Network Approach in Stroke Patients Based on Electroencephalography: A Systematic Review and Meta-Analysis. Healthcare. 2023;11(5):666. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/healthcare11050666\u003c/span\u003e\u003cspan address=\"10.3390/healthcare11050666\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIsmail LE, Karwowski W. A Graph Theory-Based Modeling of Functional Brain Connectivity Based on EEG: A Systematic Review in the Context of Neuroergonomics. IEEE Access. 2020;8:155103\u0026ndash;35. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/ACCESS.2020.3018995\u003c/span\u003e\u003cspan address=\"10.1109/ACCESS.2020.3018995\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRosazza C, Minati L. Resting-state brain networks: literature review and clinical applications. Neurol Sci. 2011;32(5):773\u0026ndash;85. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10072-011-0636-y\u003c/span\u003e\u003cspan address=\"10.1007/s10072-011-0636-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBoyd LA, Hayward KS, Ward NS, Stinear CM, Rosso C, Fisher RJ, et al. Biomarkers of stroke recovery: Consensus-based core recommendations from the Stroke Recovery and Rehabilitation Roundtable. Int J Stroke. 2017;12(5):480\u0026ndash;93. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/1747493017714176\u003c/span\u003e\u003cspan address=\"10.1177/1747493017714176\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVatinno AA, Simpson A, Ramakrishnan V, Bonilha HS, Bonilha L, Seo NJ. The Prognostic Utility of Electroencephalography in Stroke Recovery: A Systematic Review and Meta-Analysis. Neurorehabil Neural Repair. 2022;36(4\u0026ndash;5):255\u0026ndash;68. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/15459683221078294\u003c/span\u003e\u003cspan address=\"10.1177/15459683221078294\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarter AR, Shulman GL, Corbetta M. Why use a connectivity-based approach to study stroke and recovery of function? NeuroImage. 2012;62(4):2271\u0026ndash;80. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.neuroimage.2012.02.070\u003c/span\u003e\u003cspan address=\"10.1016/j.neuroimage.2012.02.070\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCassidy JM, Wodeyar A, Srinivasan R, Cramer SC. Coherent neural oscillations inform early stroke motor recovery. Hum Brain Mapp. 2021;42(17):5636\u0026ndash;47. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/hbm.25643\u003c/span\u003e\u003cspan address=\"10.1002/hbm.25643\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSaes M, Meskers CGM, Daffertshofer A, van Wegen EEH, Kwakkel G. Are early measured resting-state EEG parameters predictive for upper limb motor impairment six months poststroke? Clin Neurophysiol. 2021;132(1):56\u0026ndash;62. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.clinph.2020.09.031\u003c/span\u003e\u003cspan address=\"10.1016/j.clinph.2020.09.031\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu J, Quinlan EB, Dodakian L, McKenzie A, Kathuria N, Zhou RJ, et al. Connectivity measures are robust biomarkers of cortical function and plasticity after stroke. Brain. 2015;138(8):2359\u0026ndash;69. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/brain/awv156\u003c/span\u003e\u003cspan address=\"10.1093/brain/awv156\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVecchio F, Pappalettera C, Miraglia F, Deinite G, Manenti R, Judica E, et al. Prognostic Role of Hemispherical Functional Connectivity in Stroke: A Study via Graph Theory Versus Coherence of Electroencephalography Rhythms. Stroke. 2023;54(2):499\u0026ndash;508. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1161/STROKEAHA.122.040747\u003c/span\u003e\u003cspan address=\"10.1161/STROKEAHA.122.040747\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVecchio F, Caliandro P, Reale G, Miraglia F, Piludu F, Masi G, et al. Acute cerebellar stroke and middle cerebral artery stroke exert distinctive modifications on functional cortical connectivity: A comparative study via EEG graph theory. Clin Neurophysiol. 2019;130(6):997\u0026ndash;1007. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.clinph.2019.03.017\u003c/span\u003e\u003cspan address=\"10.1016/j.clinph.2019.03.017\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShim M, Choi G-Y, Paik N-J, Lim C, Hwang H-J, Kim W-S. Altered Functional Networks of Alpha and Low-Beta Bands During Upper Limb Movement and Association with Motor Impairment in Chronic Stroke. Brain Connect. 2021. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1089/brain.2021.0070\u003c/span\u003e\u003cspan address=\"10.1089/brain.2021.0070\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBistriceanu CE, Danciu FA, Cuciureanu DI. Cortical connectivity in stroke using signals from resting-state EEG: a review of current literature. Acta Neurol Belg. 2023;123(2):351\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s13760-022-02102-z\u003c/span\u003e\u003cspan address=\"10.1007/s13760-022-02102-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLocatelli T, Cursi M, Liberati D, Franceschi M, Comi G. EEG coherence in Alzheimer's disease. Electroencephalogr Clin Neurophysiol. 1998;106(3):229\u0026ndash;37. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S0013-4694(97)00129-6\u003c/span\u003e\u003cspan address=\"10.1016/S0013-4694(97)00129-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu X, Zheng W-L, Li Z, Lu B-L. Investigating EEG-based functional connectivity patterns for multimodal emotion recognition. J Neural Eng. 2022;19(1):016012. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1088/1741-2552/ac49a7\u003c/span\u003e\u003cspan address=\"10.1088/1741-2552/ac49a7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLassi M, Bandini A, Spina V, Azzollini V, Dalise S, Mazzoni A, et al. editors. Classification of upper limb impairment in acute stroke patients using resting-state EEG markers and machine learning. 2023 11th International IEEE/EMBS Conference on Neural Engineering (NER); 2023 24\u0026ndash;27 April 2023.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLin PJ, Li W, Zhai X, Li Z, Sun J, Xu Q, et al. Explainable Deep-Learning Prediction for Brain-Computer Interfaces Supported Lower Extremity Motor Gains Based on Multistate Fusion. IEEE Trans Neural Syst Rehabil Eng. 2024;32:1546\u0026ndash;55. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/tnsre.2024.3384498\u003c/span\u003e\u003cspan address=\"10.1109/tnsre.2024.3384498\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFugl-Meyer AR, J\u0026auml;\u0026auml;sk\u0026ouml; L, Leyman I, Olsson S, Steglind S. The post-stroke hemiplegic patient. 1. a method for evaluation of physical performance. Scand J Rehabil Med. 1975;7(1):13\u0026ndash;31.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCoscia M, Wessel MJ, Chaudary U, Mill\u0026aacute;n JR, Micera S, Guggisberg A, et al. Neurotechnology-aided interventions for upper limb motor rehabilitation in severe chronic stroke. Brain. 2019;142(8):2182\u0026ndash;97. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/brain/awz181\u003c/span\u003e\u003cspan address=\"10.1093/brain/awz181\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDe Vico Fallani F, Pichiorri F, Morone G, Molinari M, Babiloni F, Cincotti F, et al. Multiscale topological properties of functional brain networks during motor imagery after stroke. NeuroImage. 2013;83:438\u0026ndash;49. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.neuroimage.2013.06.039\u003c/span\u003e\u003cspan address=\"10.1016/j.neuroimage.2013.06.039\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOostenveld R, Fries P, Maris E, Schoffelen J-M. FieldTrip: open source software for advanced analysis of MEG, EEG, and invasive electrophysiological data. Comput Intell Neurosci. 2011:2011:156869. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1155/2011/156869\u003c/span\u003e\u003cspan address=\"10.1155/2011/156869\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShan H, Xu H, Zhu S, He B. A novel channel selection method for optimal classification in different motor imagery BCI paradigms. Biomed Eng Online. 2015;14(1):93. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s12938-015-0087-4\u003c/span\u003e\u003cspan address=\"10.1186/s12938-015-0087-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMazumder A, Ghosh P, Khasnobish A, Bhattacharyya S, Tibarewala D, editors. Selection of Relevant Features from Cognitive EEG Signals Using ReliefF and MRMR Algorithm. Advancements of Medical Electronics: Proceedings of the First International Conference, ICAME 2015; 2015: Springer.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O et al. Scikit-learn: Machine Learning in Python. ArXiv. 2011;abs/1201.0490.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuyon I, Weston J, Barnhill S, Vapnik V. Gene Selection for Cancer Classification using Support Vector Machines. Mach Learn. 2002;46(1):389\u0026ndash;422. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1023/A:1012487302797\u003c/span\u003e\u003cspan address=\"10.1023/A:1012487302797\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHidalgo-Mu\u0026ntilde;oz AR, L\u0026oacute;pez MM, Santos IM, Pereira AT, V\u0026aacute;zquez-Marrufo M, Galvao-Carmona A, et al. Application of SVM-RFE on EEG signals for detecting the most relevant scalp regions linked to affective valence processing. Expert Syst Appl. 2013;40(6):2102\u0026ndash;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.eswa.2012.10.013\u003c/span\u003e\u003cspan address=\"10.1016/j.eswa.2012.10.013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUrbanowicz RJ, Olson RS, Schmitt P, Meeker M, Moore JH. Benchmarking relief-based feature selection methods for bioinformatics data mining. J Biomed Inf. 2018;85:168\u0026ndash;88. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jbi.2018.07.015\u003c/span\u003e\u003cspan address=\"10.1016/j.jbi.2018.07.015\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang J, Chen M, Zhao S, Hu S, Shi Z, Cao Y. ReliefF-Based EEG Sensor Selection Methods for Emotion Recognition. Sensors. 2016;16(10):1558. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/s16101558\u003c/span\u003e\u003cspan address=\"10.3390/s16101558\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShim M, Lee S-H, Hwang H-J. Inflated prediction accuracy of neuropsychiatric biomarkers caused by data leakage in feature selection. Sci Rep. 2021;11(1):7980. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41598-021-87157-3\u003c/span\u003e\u003cspan address=\"10.1038/s41598-021-87157-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDoerrfuss JI, Kilic T, Ahmadi M, Holtkamp M, Weber JE. Quantitative and Qualitative EEG as a Prediction Tool for Outcome and Complications in Acute Stroke Patients. Clin EEG Neurosci. 2020;51(2):121\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/1550059419875916\u003c/span\u003e\u003cspan address=\"10.1177/1550059419875916\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFinnigan S, van Putten MJ. EEG in ischaemic stroke: quantitative EEG can uniquely inform (sub-)acute prognoses and clinical management. Clin Neurophysiol. 2013;124(1):10\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.clinph.2012.07.003\u003c/span\u003e\u003cspan address=\"10.1016/j.clinph.2012.07.003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHall TM, de Carvalho F, Jackson AJN. A common structure underlies low-frequency cortical dynamics in movement, sleep, and sedation. Neuron N. 2014;83(5):1185\u0026ndash;99.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchroeder CE, Lakatos PJT. Low-frequency neuronal oscillations as instruments of sensory selection. Trends Neurosci. 2009;32(1):9\u0026ndash;18. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.tins.2008.09.012\u003c/span\u003e\u003cspan address=\"10.1016/j.tins.2008.09.012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eL\u0026oacute;pez ME, Pusil S, Pereda E, Maest\u0026uacute; F, Barcel\u0026oacute; FJN. Dynamic low frequency EEG phase synchronization patterns during proactive control of task switching. NeuroImage. 2019;186:70\u0026ndash;82. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.neuroimage.2018.10.068\u003c/span\u003e\u003cspan address=\"10.1016/j.neuroimage.2018.10.068\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRamanathan DS, Guo L, Gulati T, Davidson G, Hishinuma AK, Won S-J, et al. Low-frequency cortical activity is a neuromodulatory target that tracks recovery after stroke. Nat Med. 2018;24(8):1257\u0026ndash;67. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41591-018-0058-y\u003c/span\u003e\u003cspan address=\"10.1038/s41591-018-0058-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAndraus ME, Alves-Leon SV. Non-epileptiform EEG abnormalities: an overview. Arq Neuropsiquiatr. 2011;69(5):829\u0026ndash;35. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1590/s0004-282x2011000600020\u003c/span\u003e\u003cspan address=\"10.1590/s0004-282x2011000600020\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan Putten MJ, Tavy DL. Continuous quantitative EEG monitoring in hemispheric stroke patients using the brain symmetry index. Stroke. 2004;35(11):2489\u0026ndash;92. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1161/01.STR.0000144649.49861.1d\u003c/span\u003e\u003cspan address=\"10.1161/01.STR.0000144649.49861.1d\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBritton JW, Frey LC, Hopp JL, Korb P, Koubeissi MZ, Lievens WE, et al. Electroencephalography (EEG): An introductory text and atlas of normal and abnormal findings in adults, children, and infants. Chicago: American Epilepsy Society; 2016.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAssenza G, Zappasodi F, Squitti R, Altamura C, Ventriglia M, Ercolani M, et al. Neuronal functionality assessed by magnetoencephalography is related to oxidative stress system in acute ischemic stroke. NeuroImage. 2009;44(4):1267\u0026ndash;73. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.neuroimage.2008.09.049\u003c/span\u003e\u003cspan address=\"10.1016/j.neuroimage.2008.09.049\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFinnigan SP, Walsh M, Rose SE, Chalk JB. Quantitative EEG indices of sub-acute ischaemic stroke correlate with clinical outcomes. Clin Neurophysiol. 2007;118(11):2525\u0026ndash;32. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.clinph.2007.07.021\u003c/span\u003e\u003cspan address=\"10.1016/j.clinph.2007.07.021\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFinnigan SP, Rose SE, Walsh M, Griffin M, Janke AL, McMahon KL, et al. Correlation of quantitative EEG in acute ischemic stroke with 30-day NIHSS score: comparison with diffusion and perfusion MRI. Stroke. 2004;35(4):899\u0026ndash;903. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1161/01.STR.0000122622.73916.d2\u003c/span\u003e\u003cspan address=\"10.1161/01.STR.0000122622.73916.d2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBentes C, Peralta AR, Viana P, Martins H, Morgado C, Casimiro C, et al. Quantitative EEG and functional outcome following acute ischemic stroke. Clin Neurophysiol. 2018;129(8):1680\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.clinph.2018.05.021\u003c/span\u003e\u003cspan address=\"10.1016/j.clinph.2018.05.021\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSheorajpanday RV, Nagels G, Weeren AJ, van Putten MJ, De Deyn PP. Quantitative EEG in ischemic stroke: correlation with functional status after 6 months. Clin Neurophysiol. 2011;122(5):874\u0026ndash;83. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.clinph.2010.07.028\u003c/span\u003e\u003cspan address=\"10.1016/j.clinph.2010.07.028\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSaes M, Meskers CGM, Daffertshofer A, de Munck JC, Kwakkel G, van Wegen EEH. How does upper extremity Fugl-Meyer motor score relate to resting-state EEG in chronic stroke? A power spectral density analysis. Clin Neurophysiol. 2019;130(5):856\u0026ndash;62. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.clinph.2019.01.007\u003c/span\u003e\u003cspan address=\"10.1016/j.clinph.2019.01.007\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang JJ, Bai Z, Fong KNK. Resting-state cortical electroencephalogram rhythms and network in patients after chronic stroke. J Neuroeng Rehabil. 2024;21(1):32. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s12984-024-01328-7\u003c/span\u003e\u003cspan address=\"10.1186/s12984-024-01328-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCassidy JM, Wodeyar A, Wu J, Kaur K, Masuda AK, Srinivasan R, et al. Low-Frequency Oscillations Are a Biomarker of Injury and Recovery After Stroke. Stroke. 2020;51(5):1442\u0026ndash;50. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1161/STROKEAHA.120.028932\u003c/span\u003e\u003cspan address=\"10.1161/STROKEAHA.120.028932\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDing Q, Chen J, Zhang S, Chen S, Li X, Peng Y, et al. Neurophysiological characterization of stroke recovery: A longitudinal TMS and EEG study. CNS Neurosci Ther. 2024;30(3):e14471. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/cns.14471\u003c/span\u003e\u003cspan address=\"10.1111/cns.14471\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAssenza G, Zappasodi F, Pasqualetti P, Vernieri F, Tecchio F. A contralesional EEG power increase mediated by interhemispheric disconnection provides negative prognosis in acute stroke. Restor Neurol Neurosci. 2013;31(2):177\u0026ndash;88. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3233/rnn-120244\u003c/span\u003e\u003cspan address=\"10.3233/rnn-120244\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eB\u0026ouml;nstrup M, Schulz R, Sch\u0026ouml;n G, Cheng B, Feldheim J, Thomalla G, et al. Parietofrontal network upregulation after motor stroke. Neuroimage Clin. 2018;18:720\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.nicl.2018.03.006\u003c/span\u003e\u003cspan address=\"10.1016/j.nicl.2018.03.006\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHordacre B, Lotze M, Jenkinson M, Lazari A, Barras CD, Boyd L, et al. Fronto-parietal involvement in chronic stroke motor performance when corticospinal tract integrity is compromised. Neuroimage Clin. 2021;29:102558. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.nicl.2021.102558\u003c/span\u003e\u003cspan address=\"10.1016/j.nicl.2021.102558\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRinne P, Hassan M, Fernandes C, Han E, Hennessy E, Waldman A, et al. Motor dexterity and strength depend upon integrity of the attention-control system. Proc Natl Acad Sci USA. 2018;115(3):E536\u0026ndash;45. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1073/pnas.1715617115\u003c/span\u003e\u003cspan address=\"10.1073/pnas.1715617115\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStępień M, Conradi J, Waterstraat G, Hohlefeld FU, Curio G, Nikulin VV. Event-related desynchronization of sensorimotor EEG rhythms in hemiparetic patients with acute stroke. Neurosci Lett. 2011;488(1):17\u0026ndash;21. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.neulet.2010.10.072\u003c/span\u003e\u003cspan address=\"10.1016/j.neulet.2010.10.072\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBernhardt J, Hayward KS, Kwakkel G, Ward NS, Wolf SL, Borschmann K, et al. Agreed definitions and a shared vision for new standards in stroke recovery research: The Stroke Recovery and Rehabilitation Roundtable taskforce. Int J Stroke. 2017;12(5):444\u0026ndash;50. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/1747493017711816\u003c/span\u003e\u003cspan address=\"10.1177/1747493017711816\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWard NS. Restoring brain function after stroke - bridging the gap between animals and humans. Nat Rev Neurol. 2017;13(4):244\u0026ndash;55. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/nrneurol.2017.34\u003c/span\u003e\u003cspan address=\"10.1038/nrneurol.2017.34\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"journal-of-neuroengineering-and-rehabilitation","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jner","sideBox":"Learn more about [Journal of NeuroEngineering and Rehabilitation](http://jneuroengrehab.biomedcentral.com/)","snPcode":"12984","submissionUrl":"https://submission.nature.com/new-submission/12984/3","title":"Journal of NeuroEngineering and Rehabilitation","twitterHandle":"@BioMedCentral","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Stroke, upper limb impairment, motor recovery, Fugl-Meyer Assessment, electroencephalography, coherence, connectivity, machine learning","lastPublishedDoi":"10.21203/rs.3.rs-5976957/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5976957/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eSevere upper limb impairment (ULI) presents a significant challenge in the rehabilitation of chronic stroke survivors, affecting their quality of life. Identifying biomarkers and understanding the neural mechanisms associated with severe ULI are essential for evaluating recovery potential and enhancing rehabilitation effectiveness. This study aimed to identify resting-state electroencephalography (EEG) functional connectivity features related to severe ULI in chronic stroke survivors using machine learning (ML) methods.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eEEG data were collected from 34 chronic stroke survivors. Participants were categorized into two labels based on their Fugl-Meyer assessment for upper extremity (FMA-UE) scores: a mild/moderate ULI (FMA-UE\u0026thinsp;\u0026ge;\u0026thinsp;30; n\u0026thinsp;=\u0026thinsp;19) and a severe ULI (FMA-UE\u0026thinsp;\u0026lt;\u0026thinsp;30; n\u0026thinsp;=\u0026thinsp;15). We employed ML algorithms to classify severe ULI, including logistic regression with L1, elastic net regularization, stochastic gradient descent, and support vector machines, along with several feature selection methods. Coherence was evaluated across six frequency bands within both the ipsilesional (affected by the lesion) and contralesional (opposite side of the lesion) hemispheres.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eThe logistic regression model with L1 and ReliefF feature selection methods was the most effective, achieving a balanced accuracy of 0.91 (sensitivity\u0026thinsp;=\u0026thinsp;0.93, specificity\u0026thinsp;=\u0026thinsp;0.90). This approach identified 14 significant features for distinguishing severe ULI from mild to moderate ULI, including delta interhemispheric and intrahemispheric connectivity of the frontal, parietal, and temporal regions. Additionally, interhemispheric and intrahemispheric theta connectivity was observed in the prefrontal, frontal, temporal, and parietal regions. Low-beta intrahemispheric connectivity was also observed in the contralesional parietal regions.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eOur research highlights the association between alterations in connectivity within low-frequency bands and severe ULI across widespread brain regions, including areas outside the sensorimotor cortex and bilateral intrahemispheric and interhemispheric regions. Further research utilizing larger longitudinal datasets from early stroke survivors employing ML approaches could contribute to the development of more accurate predictive models for motor recovery and rehabilitation responses.\u003c/p\u003e","manuscriptTitle":"Functional Connectivity Associated with Severe Upper Limb Impairment in Resting-State Electroencephalography Among Chronic Stroke Survivors: A Machine Learning Approach","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-03-11 09:31:40","doi":"10.21203/rs.3.rs-5976957/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-04-07T06:49:01+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-03-02T05:23:47+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-02-21T19:19:05+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"19845475810502157698731874492912542066","date":"2025-02-13T23:55:53+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"147423351435151900310176154786177121884","date":"2025-02-13T16:02:44+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"325512052260736610141794729917577471567","date":"2025-02-12T14:29:22+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-02-11T07:29:13+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-02-07T12:47:26+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-02-07T12:44:35+00:00","index":"","fulltext":""},{"type":"submitted","content":"Journal of NeuroEngineering and Rehabilitation","date":"2025-02-07T01:47:25+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"journal-of-neuroengineering-and-rehabilitation","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jner","sideBox":"Learn more about [Journal of NeuroEngineering and Rehabilitation](http://jneuroengrehab.biomedcentral.com/)","snPcode":"12984","submissionUrl":"https://submission.nature.com/new-submission/12984/3","title":"Journal of NeuroEngineering and Rehabilitation","twitterHandle":"@BioMedCentral","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"62c7700f-dbae-49ed-99fd-b57c6bc397c1","owner":[],"postedDate":"March 11th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-12-29T16:05:19+00:00","versionOfRecord":{"articleIdentity":"rs-5976957","link":"https://doi.org/10.1186/s12984-025-01859-7","journal":{"identity":"journal-of-neuroengineering-and-rehabilitation","isVorOnly":false,"title":"Journal of NeuroEngineering and Rehabilitation"},"publishedOn":"2025-12-22 15:58:02","publishedOnDateReadable":"December 22nd, 2025"},"versionCreatedAt":"2025-03-11 09:31:40","video":"","vorDoi":"10.1186/s12984-025-01859-7","vorDoiUrl":"https://doi.org/10.1186/s12984-025-01859-7","workflowStages":[]},"version":"v1","identity":"rs-5976957","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5976957","identity":"rs-5976957","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}