Importance: Prognosticating functional independence after an acute stroke is critical for anticipatory guidance and rehabilitation planning. Here we demonstrate that poor brain health at the time of incident stroke is linked to worse functional outcomes for women compared to men. Objective: To determine if brain health at time of stroke presentation has a differential effect on functional outcomes between men and women. Design: Retrospective cross-sectional study. Setting: Analysis conducted in 2025 with multi-center patient data that included participants from two large acute ischemic stroke cohorts; local (GASROS) and multinational (MRI-GENIE) between the years 2003 and 2011. Participants: Clinical data collected for enrolled study participants included demographic data, medical history of hypertension, diabetes mellitus, hyperlipidemia, smoking status, acute stroke severity as measured by National Institutes of Health Stroke Scale (NIHSS), stroke etiology, and modified Rankin Scale (mRS) score at 90 days post-stroke. Brain health was quantified as effective reserve derived from acute neuroimaging data. Exposure(s): designated sex, retrieved from registration records. Main Outcome: Functional outcome was measured by mRS scores at 90 days post-stroke, in men and women with poor, moderate, or good brain health at time of stroke injury. Results: A total of 1039 patients were included in the analysis, 37.8 % women, median age 67 [interquartile range 56-77]. Women with poor brain health (i.e. lowest quartile of effective reserve) had worse functional outcomes at 90 days (55.6% with mRS>2) compared to men with poor brain health (31.2% with mRS>2: p < 0.001) . This difference between men and women was not observed in categories of moderate or good brain health. There was no observed significant difference in stroke severity, volume of acute lesion, burden of white matter hyperintensities, or stroke etiology between men and women with poor brain health. Conclusions and Relevance: Brain health at the time of incident stroke has a differential effect on functional outcomes at 90 days between men and women. Women with poor brain health endure disproportionately worse outcomes compared to men. This highlights an important step in understanding sex-specific vulnerability in early recovery post-stroke, and can inform disposition, rehabilitation services, and resource allocation planning. ### Competing Interest Statement The authors have declared no competing interest. ### Funding Statement This study did not receive any funding. ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: Institutional review board (IRB) approval at Massachusetts General Hospital was obtained for all subjects enrolled locally at MGH (2001P001186), as was approval from local ethics committee/IRB for each site that enrolled patients in the MRI-GENIE [MRI-GENetics Interface Exploration] study. Datasets were de-identified prior to use in this study. I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes Data will be made available without investigator support for purposes of replicating the results, after approval of a submitted data request form and obtaining the relevant local IRBs approval, and with a signed data-use agreement.

OBJECTIVE:Brain-predicted age difference (BrainAGE) is a neuroimaging biomarker reflecting brain health, with potential implications for post-stroke recovery. However, training robust BrainAGE models requires large, diverse datasets, often restricted by privacy and regulatory concerns. This study evaluates the performance of federated learning (FL) for BrainAGE estimation in ischemic stroke patients treated with mechanical thrombectomy, and investigates its association with clinical phenotypes and functional outcome. METHODS:We used pre-treatment FLAIR brain images from 1674 stroke patients across 16 hospital centers. We implemented standard machine learning and deep learning models for BrainAGE estimates under three data management strategies: centralized learning (pooled data), FL (local training at each site), and single-site learning. We reported prediction errors and examined associations between BrainAGE and vascular risk factors (e.g., diabetes mellitus, hypertension, smoking), as well as functional outcome at three months post-stroke. Logistic regression evaluated BrainAGE's predictive value for this outcome, adjusting for age, sex, vascular risk factors, stroke severity, time between MRI and arterial puncture, prior intravenous thrombolysis, and recanalization outcome. RESULTS:While centralized learning yielded the most accurate predictions, FL consistently outperformed single-site models. BrainAGE was significantly higher in patients with diabetes mellitus across all models. Comparisons between patients with good and poor functional outcome, and multivariate predictions of these outcome showed the significance of the association between BrainAGE and post-stroke recovery. CONCLUSION:FL enables accurate age predictions without data centralization. The strong association between BrainAGE, vascular risk factors, and post-stroke recovery highlights its potential for personalized prognostic modeling in stroke care.

BACKGROUND:Endovascular thrombectomy (EVT) dramatically improves clinical outcomes, but the final infarct volume (FIV) on magnetic resonance imaging only accounts for a minority of the treatment effect. An imaging biomarker that more strongly correlates with post-EVT functional outcome would be helpful for clinical prognosis and serve as a surrogate outcome measure in trials of EVT-adjuvant therapies. Here, we aimed to validate a novel magnetic resonance imaging-based metric, infarct density, which leverages post-EVT apparent diffusion coefficient as a marker of infarct severity. METHODS:A retrospective cohort was derived from a single-center prospective EVT registry. Consecutive patients treated with EVT for anterior circulation large vessel occlusion were included from 2018 to 2019 who achieved successful reperfusion (modified Thrombolysis in Cerebral Infarction ≥2b). Magnetic resonance imaging was performed 12 to 48 hours post-EVT and processed via RAPID to quantify FIV using the apparent diffusion coefficient <620 threshold. Lesion volume was also collected using an apparent diffusion coefficient <470 threshold, and infarct density was calculated as: (volume <470/volume <620)×100%. Good outcome was defined as ≤2 on the 90-day modified Rankin Scale. Multivariable logistic regression models quantified the association between clinical/imaging variables and outcome. Receiver operating characteristic analysis quantified model classification performance. RESULTS:Of 319 patients treated with EVT, 272 met inclusion criteria. The mean age was 69±13 years, 41% were female, and 62% achieved a good outcome. After adjusting for clinical and radiographic factors, FIV (adjusted odds ratio, 0.99 per 1 mL [95% CI, 0.98-1.00]; P=0.03) and infarct density (adjusted odds ratio, 0.95 per 1% [95% CI, 0.94-0.97]; P<0.001) were both independently inversely associated with good outcome. The final model incorporating both FIV and infarct density achieved excellent classification performance (area under the curve, 0.87 [95% CI, 0.83-0.91]). Removing infarct density from the model diminished its performance (area under the curve, 0.83 [95% CI, 0.78-0.88]; P=0.01). CONCLUSIONS:Apparent diffusion coefficient-based infarct density after EVT is independently associated with long-term outcome and provides greater prognostic information than FIV alone. Post-EVT infarct density may be useful in clinical care and as a surrogate outcome measure in trials of EVT-adjuvant therapies.

Background: Endovascular thrombectomy (EVT) dramatically improves clinical outcomes, but the final infarct volume (FIV) on MRI only accounts for a minority of the treatment effect. An imaging biomarker that more strongly correlates with post-EVT functional outcome would be helpful for clinical prognosis and serve as a surrogate outcome measure in trials of EVT-adjuvant therapies. Here, we aimed to validate a novel MRI-based metric, infarct density, which leverages post-EVT apparent diffusion coefficient (ADC) as a marker of infarct severity. Methods: A retrospective cohort was derived from a single-center prospective EVT registry. Consecutive anterior circulation EVT patients were included from 2018-2019 who achieved successful reperfusion (mTICI ≥2b). MRI was performed 12-48 hours post-EVT and processed via RAPID to quantify FIV using the ADC <620 threshold. Lesion volume was also collected using ADC <470 threshold, and infarct density was calculated as: (volume <470/volume <620)x100%. Good outcome was defined as ≤2 on the 90-day modified Rankin Scale. Multivariable logistic regression models quantified the association between clinical/imaging variables and outcome. ROC analysis quantified model classification performance. Results: Of 319 EVT patients, 272 met inclusion criteria. The mean age was 69 ±13 years, 41% were female, and 62% achieved a good outcome. After adjusting for clinical and radiographic factors, FIV (aOR 0.99 per 1mL; 95%CI: 0.98-1.00; p=0.03) and infarct density (aOR 0.95 per 1%; 95%CI: 0.94-0.97; p<0.001) were both independently inversely associated with good outcome. The final model incorporating both FIV and infarct density achieved excellent classification performance (AUC 0.87; 95%CI: 0.83-0.91). Removing infarct density from the model diminished its performance (AUC 0.83; 95%CI: 0.78-0.88; p=0.01). Conclusion: ADC-based infarct density after EVT is independently associated with long‐term outcome and provides greater prognostic information than FIV alone. Post-EVT infarct density may be useful in clinical care and as a surrogate outcome measure in trials of EVT-adjuvant therapies.

BACKGROUND AND PURPOSE:Parenchymal hematomas (PHs) represent an important complication in ischemic stroke after endovascular thrombectomy (EVT), but the risk factors are incompletely understood. Neuroimaging data preintervention, such as infarct topography, may help elucidate predisposing factors and inform more nuanced patient care intra- and postprocedurally. METHODS:Large vessel occlusion patients with pre-EVT MRI were included from a single quaternary center. Diffusion-weighted imaging (DWI) lesions underwent manual segmentation and registration onto a standard brain space for topographical mapping. The presence of PH postintervention was determined. Associations between infarct topography, clinical characteristics, and PH were evaluated. RESULTS:A total of 165 patients (median age: 69; 56% female) were identified. Intravenous alteplase was administered to 52%, 70% achieved thrombolysis in cerebral infarction 2b-3 reperfusion, and 8% had PH postintervention. The preintervention DWI lesions were 48% (38%-60%) white matter, 23% (6%-47%) cortex, and 15% (4%-28%) basal ganglia. Basal ganglia infarct volume was independently associated with PH (adjusted odds ratio = 1.342, 95% confidence interval 1.002-1.797, p = 0.049), accounting for white matter and cortex infarct volume, among other key factors. Basal ganglia infarct volume was associated with susceptibility-weighted imaging vessel sign (betaadjusted = 0.233, p = 0.006) and the National Institutes of Health Stroke Scale (betaadjusted = 0.220, p = 0.012), controlling for other factors. CONCLUSIONS:Preintervention basal ganglia infarct volume may provide important insights into the risk of PH after intervention. Improved understanding of the biology of basal ganglia infarction and hemorrhagic transformation has implications for the management of patients undergoing EVT and may represent a future therapeutic target for neuroprotective strategies.

Once taken into consideration, sex differences in neurological diseases emerge in abundance: (i) Stroke severity is significantly higher in females than in males, (ii) Alzheimer's disease (AD) pathology is more pronounced in females, and (iii) conspicuous links with hormonal cycles led to female-specific diagnoses, such as catamenial migraines and epilepsy. While these differences receive increasing attention in isolation, they likely link to similar processes in the brain. Hence, this review aims to present an overview of the influences of sex chromosomes, hormones, and aging on male and female brains across health and disease, with a particular focus on AD and stroke. The focus here on advancements across several fields holds promise to fuel future research and to lead to an enriched understanding of the brain and more effective personalized neurologic care for all.

Background One third of all patients with acute ischemic strokes have a pre-existing disability. Patients with pre-existing disabilities have historically been excluded from landmark clinical trials of acute stroke interventions, leading to ongoing controversy about the risks and benefits of acute stroke interventions such as endovascular thrombectomy (EVT). To address this controversy, we compared long-term outcomes and end-of-life care in large vessel occlusion (LVO) patients with moderate-to-severe baseline disability treated with EVT versus medical management alone. Methods Patients who presented with an LVO to our comprehensive stroke center between January 2017 and December 2020 were retrospectively identified from a prospectively maintained database. Moderate-to-severe baseline disability was defined as a pre-stroke modified Rankin Scale (mRS) of 3-5. Delta mRS was defined as the difference between the 90-day and baseline mRS. Logistic and ordinal regressions were performed to evaluate the relationships between EVT and outcomes. An analysis of rates and reasons for transitions to comfort care was also performed, where applicable. Results A total of 175/1008 (17%) LVO patients with moderate-to-severe baseline disability were identified. The median age was 82 (IQR 70-89), and 59% were female. Thirty-two patients (18%) with moderate-to-severe baseline disability were treated with EVT. EVT was independently associated with improved delta mRS (B=-1.048; 95%CI=-1.777,-0.318; p=0.005) accounting for age and NIHSS. However, EVT did not reduce the odds of transitioning to comfort care (aOR=0.794; 95%CI=0.347,1.818; p=0.585) accounting for age and NIHSS. Seventy-six (43%) patients were transitioned to comfort care during their hospitalization. Of the 99 who were not transitioned to comfort care, 18 were treated with EVT, and EVT was independently associated with improved delta mRS (B=-2.794; 95%CI=-4.002,-1.586; p<0.0001). The median time from presentation to transition to comfort care was 2 days (IQR 1-7) in the non-EVT group, compared to 7 (IQR 4-11) in the EVT group (H(1)=5.46, p=0.019). The primary reasons for transitions to comfort care were poor perceived prognosis and medical complications. Conclusions Among patients with moderate-to-severe baseline disability, EVT is associated with less post-stroke accumulated disability without limiting transitions to comfort care. EVT is compatible with goal-concordant care and should not be routinely withheld because of baseline disability alone.

Importance:Brain health may facilitate resilience to detrimental consequences from neurological diseases. Infarct volume is associated with poor functional outcome after acute ischemic stroke (AIS), but potential mediating effects through stroke-related brain health loss have not been investigated. Objective:To determine whether stroke-related brain health loss, quantified by change in MRI derived effective Reserve (eR), mediates the effect of acute infarct volume on functional outcome after AIS. Design:Observational multicenter cohort study. Setting:We analyzed data from the GASROS (n=488) and MRI-GENIE (n=560) cohorts, collected 2003-2011. Participants:Adult patients consecutively diagnosed with AIS, with available admission MRI. Exposure:At admission, white matter hyperintensity (WMH) and normal-appearing brain volumes were assessed on T2-FLAIR, and acute infarct volume on diffusion weighted imaging. WMH was normalized by brain volume, creating WMH load. We quantified brain health using eR, a latent variable incorporating age, WMH load, and normal-appearing brain volume. ΔeR reflected the change in eR when acute infarct volume was included, representing stroke-related brain health decline. Mediation analysis was used to determine if ΔeR mediates the effect of infarct volume on functional outcome (modified Rankin Scale [mRS] at 90 days). Main Outcome Measure:Proportion of mediating effect. Results:We included 1,048 patients (median age 67y, 38% females). At baseline, median NIHSS score was 3 (IQR 1-7), median infarct volume 3.1mL (IQR 0.9-15.5). At 90 days, median mRS score was 1 (IQR 1-3) and 51 (5%) patients had died. In mediation analysis, ΔeR significantly mediated 36% (95% CI 16-56%) of the total effect of infarct volume on functional outcome (direct effect (ß=0.15 [95% CI 0.09-0.22], p<0.001; indirect effect mediated through ΔeR: ß=0.09 [95% CI 0.04 to 0.14], p=0.001). In subgroup-analyses, the mediative effect was apparent among female but not male, and among patients aged >67y but not ≤67y. Conclusions and Relevance:Stroke-related structural brain health loss mediates about one third of the effect of acute infarct volume on functional outcome after ischemic stroke, with important sex and age differences. Brain health significantly influences outcome and recovery potential, and may be considered a key biomarker when modeling outcome after AIS.

Objectives:To compare quantitative MRI markers of brain health in their ability to predict functional outcome after acute ischemic stroke (AIS). Methods:We included AIS survivors from the international MRI-GENIE study (multicenter; 2003-2011) with acute T2-FLAIR imaging. Automated pipelines estimated white matter hyperintensity volume (WMHv), brain volume, and intracranial volume (ICV). Assessed brain health markers included: brain parenchymal fraction (brain volume relative to ICV); radiomics derived brain age; brain reserve (normal appearing brain volume relative to ICV), and effective Reserve (eR, latent variable based on age, WMH load and brain volume). We added the markers to a clinical reference model, comparing model performances between separate multivariable regression models in their prediction of unfavorable outcome (90-day modified Rankin Scale score 3-5), using Bayesian Information Criterion (BIC). Results:We analyzed 2,223 patients (median age 67 years, 45% female, 24% unfavorable outcome). All models using brain health markers outperformed the clinical reference model (ΔBIC > 10). The eR model showed the lowest BIC value (BIC=2171.8), providing strong statistical evidence to outperform the brain age model (BIC=2179.5, ΔBIC > 6), and very strong (ΔBIC > 10) statistical evidence to outperform all other models. Discussion:Quantitative MRI markers of brain health, especially eR, enhance personalized outcome prognostication after AIS.

OBJECTIVE:To quantify brain health using a measure of reserve that incorporates pre-existing pathology. METHODS:We analyzed 2 retrospective ischemic stroke cohorts (GASROS and SALVO) with neuroimaging and 90-day modified Rankin Scores (mRS) available. White matter hyperintensity (WMHv), brain, and intracranial volumes were automatically extracted and brain parenchymal fraction (BPF) calculated. The latent variable effective reserve (eR) was modeled using age, WMHv, and BPF or brain volume in GASROS. Models were compared using Bayes Information Criterion (BIC). The best model's eR estimates were categorized into quartiles and evaluated in SALVO. RESULTS:GASROS included 476 (median age: 65.8; 65.3% male) and SALVO included 43 (median age: 69.2; 62.8% male) patients. Inverse associations between eR and mRS was seen in both models, with brain volume outperforming BPF (path coefficients: -0.67, -0.48, respectively; p < 0.001; |ΔBIC| = 362). Quartile-based eR stratification in both studies showed a similar inverse trend, with worse outcomes in the low reserve group (mRS ≤2 - highest vs lowest quartile: 85/90% vs 59/45% for GASROS/SALVO). DISCUSSION:Expanding the concept of eR, highlights its clinical translational potential. The strong link between higher eR and better outcomes underscores its value as a protective brain health metric.

Introduction: Perinatal stroke refers to brain injury sustained between 28 weeks of gestational age (GA) and 28 days of life. Stroke risk in the newborn period is significantly higher relative to other life stages, and over half of all survivors endure lifelong neurological sequalae. Language deficits are observed in many patients, can emerge at any stage in development, and are difficult to predict. Lesion laterality and location are not sufficient to explain language outcomes in a rapidly developing brain with capacities for reassignment of functions to non-injured structures. Structural connectivity studies describe a “rich club” organization of the brain with 6 bilateral highly connected anatomical structures that form the information backbone for all networks. This organization has been found to exist as early as 30 weeks GA. The effects of injury to this rich club have been studied in adult stroke patients with respect to stroke severity and outcome. Given that language is a network function, we hypothesized that injury to rich club regions - a measure of structural network disruption - is associated with poor language outcomes in perinatal stroke. Methods: Retrospective study that identified perinatal stroke patients in a pediatric stroke clinic database. Clinical and imaging data retrieved included brain MRI scans and longitudinal scores of standardized neuropsychological tests. Multiple linear regression models were created for the following variables: age, sex, Autism Spectrum Disorder diagnosis, lesion volume (cm 3 ), and number of rich clubs affected by the stroke. All analyses were conducted using the computing environment R. Significance was set at p < 0.05. Results: Cohort of 75 patients was identified. Median time to follow up was 7 years, and median time between the first and last test was 30 months. Lesion volume significantly influenced receptive language outcomes at younger ages (beta = -0.446), however the number of rich clubs affected by the injury was the only significant variable associated with a poor expressive language outcome at follow-up (beta = -0.699). Discussion: Number of rich clubs affected by injury, as a global measure of structural network disruption in the neonatal period, can be used to inform language outcome prognostication.

Objectives: To determine the relationship between patient-reported outcome measures (PROMs) and volumetric biomarkers assessed on clinical imaging in acute ischemic stroke (AIS). Background: AIS is a leading cause of long-term disability. White matter hyperintensity (V_WMH), brain (V_Brain) and stroke lesion volume (V_Lesion) have been linked as potential determinants of functional outcomes after stroke. Recently, there has been surging attention to PROMs, which allow for direct, patient-centered health assessments of stroke victims. However, the link between neuroimaging biomarkers and PROMs is not well studied. Methods: Patients presenting to the Emergency Department at Massachusetts General Hospital between February 2017 and February 2020 with a confirmed AIS on MRI were eligible and underwent a follow-up telephone interview, including PROM-10 questionnaires. V_WMH and V_Brain were automatically determined from clinical MRI. V_Lesion was manually segmented. Regression analyses were performed to identify associations of brain volumetrics and clinical variables with PROM-10 subscores and global mental and physical summary scores. Results: Utilizing data from 150 patients (mean age: 64.7; 41.9% female), higher V_WMH was associated with worse global mental (β = -0.65), global physical (β = -0.60), social activities (OR = 0.59), physical health (OR = 0.68), fatigue (OR = 0.69) and social satisfaction (OR = 0.66) scores. V_Lesions were associated with poorer global mental (β = -0.79), social (OR = 0.55) and physical (OR = 0.66) activities, mental health (OR = 0.68) and emotional distress (OR = 0.68) scores. Higher V_Brain was linked to better global mental (β = 0.93), global physical (β = 0.79), physical activities (OR = 1.72) and mental health (OR = 1.54) scores. Conclusions: Imaging biomarkers were significantly associated with PROMs. The inclusion of these markers with PROMs in routine post-stroke assessment can enhance our understanding of recovery.

Brain parenchymal fraction (BPF) has been used as a surrogate measure of global brain atrophy, and as a biomarker of brain reserve in studies evaluating clinical outcomes after brain injury. Total brain volume at the time of injury has recently been shown to influence functional outcomes, where larger brain volumes are associated with better outcomes. Here, we assess if brain volume at the time of ischemic stroke injury is a better biomarker of functional outcome than BPF. Acute ischemic stroke cases at a single center between 2003 and 2011, with MR neuroimaging obtained within 48 hours from presentation were eligible. Functional outcomes represented by the modified Rankin Score (mRS) at 90 days post admission (mRS<3 deemed a favorable outcome) were obtained via patient interview or per chart review. Deep learning enabled automated segmentation pipelines were used to calculate brain volume, intracranial volume (ICV), and BPF on the acute neuroimaging data. Patient outcomes were modeled through logistic regressions, and model comparison was conducted using the Bayes Information Criterion (BIC). 467 patients with arterial ischemic stroke were included in the analysis. Median age was 65.8 years, and 65.3% were male. In both models, age and a larger stroke lesion volume were associated with worse functional outcomes. Higher BPF and a larger brain volume were both associated with favorable functional outcomes, however, comparison of both models suggested that the brain volume model (BIC=501) explains the data better compared to the BPF model (BIC=511). The extent of global brain atrophy has been regarded as an important biomarker of post-stroke functional outcomes and resilience to acute injury. Here, we demonstrate that a higher global brain volume at the time of injury better explains favorable functional outcomes, which can be directly clinically assessed.

Coma is an unresponsive state of disordered consciousness characterized by impaired arousal and awareness. The epidemiology and pathophysiology of coma in ischemic stroke has been underexplored. We sought to characterize the incidence and clinical features of coma as a presentation of large vessel occlusion (LVO) stroke. Individuals who presented with LVO were retrospectively identified from July 2018 to December 2020. Coma was defined as an unresponsive state of impaired arousal and awareness, operationalized as a score of 3 on NIHSS item 1a. 28/637 (4.4

AbstractObjectiveTo systematically evaluate which lesion‐based imaging features and methods allow for the best statistical prediction of poststroke deficits across independent datasets.MethodsWe utilized imaging and clinical data from three independent datasets of patients experiencing acute stroke (N1 = 109, N2 = 638, N3 = 794) to statistically predict acute stroke severity (NIHSS) based on lesion volume, lesion location, and structural and functional disconnection with the lesion location using normative connectomes.ResultsWe found that prediction models trained on small single‐center datasets could perform well using within‐dataset cross‐validation, but results did not generalize to independent datasets (median R2N1 = 0.2%). Performance across independent datasets improved using large single‐center training data (R2N2 = 15.8%) and improved further using multicenter training data (R2N3 = 24.4%). These results were consistent across lesion attributes and prediction models. Including either structural or functional disconnection in the models outperformed prediction based on volume or location alone (P < 0.001, FDR‐corrected).InterpretationWe conclude that (1) prediction performance in independent datasets of patients with acute stroke cannot be inferred from cross‐validated results within a dataset, as performance results obtained via these two methods differed consistently, (2) prediction performance can be improved by training on large and, importantly, multicenter datasets, and (3) structural and functional disconnection allow for improved prediction of acute stroke severity.