Predicting outcome and recovery after stroke with lesions extracted from MRI images

Tissue Outcome Prediction in Patients with Proximal Vessel Occlusion and Mechanical Thrombectomy Using Logistic Models

Perfusion CT is established to aid selection of patients with proximal intracranial vessel occlusion for thrombectomy in the extended time window. Selection is mostly based on simple thresholding of perfusion parameter maps, which, however, does not exploit the full information hidden in the high-dimensional perfusion data. We implemented a multiparametric mass-univariate logistic model to predict tissue outcome based on data from 405 stroke patients with acute proximal vessel occlusion in the anterior circulation who underwent mechanical thrombectomy. Input parameters were acute multimodal CT imaging (perfusion, angiography, and non-contrast) as well as basic demographic and clinical parameters. The model was trained with the knowledge of recanalization status and final infarct localization. We found that perfusion parameter maps (CBF, CBV, and Tmax) were sufficient for tissue outcome prediction. Compared with single-parameter thresholding-based models, our logistic model had comparable volumetric accuracy, but was superior with respect to topographical accuracy (AUC of receiver operating characteristic). We also found higher spatial accuracy (Dice index) in an independent internal but not external cross-validation. Our results highlight the value of perfusion data compared with non-contrast CT, CT angiography and clinical information for tissue outcome-prediction. Multiparametric logistic prediction has high potential to outperform the single-parameter thresholding-based approach. In the future, the combination of tissue and functional outcome prediction might provide an individual biomarker for the benefit from mechanical thrombectomy in acute stroke care.

Predicting recovery following stroke: Deep learning, multimodal data and feature selection using explainable AI

Machine learning offers great potential for automated prediction of post-stroke symptoms and their response to rehabilitation. Major challenges for this endeavour include the very high dimensionality of neuroimaging data, the relatively small size of the datasets available for learning and interpreting the predictive features, as well as, how to effectively combine neuroimaging and tabular data (e.g. demographic information and clinical characteristics). This paper evaluates several solutions based on two strategies. The first is to use 2D images that summarise MRI scans. The second is to select key features that improve classification accuracy. Additionally, we introduce the novel approach of training a convolutional neural network (CNN) on images that combine regions-of-interests (ROIs) extracted from MRIs, with symbolic representations of tabular data. We evaluate a series of CNN architectures (both 2D and a 3D) that are trained on different representations of MRI and tabular data, to predict whether a composite measure of post-stroke spoken picture description ability is in the aphasic or non-aphasic range. MRI and tabular data were acquired from 758 English speaking stroke survivors who participated in the PLORAS study. Each participant was assigned to one of five different groups that were matched for initial severity of symptoms, recovery time, left lesion size and the months or years post-stroke that spoken description scores were collected. Training and validation were carried out on the first four groups. The fifth (lock-box/test set) group was used to test how well model accuracy generalises to new (unseen) data. The classification accuracy for a baseline logistic regression was 0.678 based on lesion size alone, rising to 0.757 and 0.813 when initial symptom severity and recovery time were successively added. The highest classification accuracy (0.854), area under the curve (0.899) and F1 score (0.901) were observed when 8 regions of interest were extracted from each MRI scan and combined with lesion size, initial severity and recovery time in a 2D Residual Neural Network (ResNet). This was also the best model when data were limited to the 286 participants with moderate or severe initial aphasia (with area under curve = 0.865), a group that would be considered more difficult to classify. Our findings demonstrate how imaging and tabular data can be combined to achieve high post-stroke classification accuracy, even when the dataset is small in machine learning terms. We conclude by proposing how the current models could be improved to achieve even higher levels of accuracy using images from hospital scanners.

Predicting Language Function Post-Stroke: A Model-Based Structural Connectivity Approach

BACKGROUND

The prediction of post-stroke language function is essential for the development of individualized treatment plans based on the personal recovery potential of aphasic stroke patients.

OBJECTIVE

To establish a framework for integrating information on connectivity disruption of the language network based on routinely collected clinical magnetic resonance (MR) images into Random Forest modeling to predict post-stroke language function.

METHODS

Language function was assessed in 76 stroke patients from the Non-Invasive Repeated Therapeutic Stimulation for Aphasia Recovery trial, using the Token Test (TT), Boston Naming Test (BNT), and Semantic Verbal Fluency (sVF) Test as primary outcome measures. Individual infarct masks were superimposed onto a diffusion tensor imaging tractogram reference set to calculate Change in Connectivity scores of language-relevant gray matter regions as estimates of structural connectivity disruption. Multivariable Random Forest models were derived to predict language function.

RESULTS

Random Forest models explained moderate to high amount of variance at baseline and follow-up for the TT (62.7% and 76.2%), BNT (47.0% and 84.3%), and sVF (52.2% and 61.1%). Initial language function and non-verbal cognitive ability were the most important variables to predict language function. Connectivity disruption explained additional variance, resulting in a prediction error increase of up to 12.8% with variable omission. Left middle temporal gyrus (12.8%) and supramarginal gyrus (9.8%) were identified as among the most important network nodes.

CONCLUSION

Connectivity disruption of the language network adds predictive value beyond lesion volume, initial language function, and non-verbal cognitive ability. Obtaining information on connectivity disruption based on routine clinical MR images constitutes a significant advancement toward practical clinical application.

Disentangling neuroplasticity mechanisms in post-stroke language recovery

A major objective in post-stroke aphasia research is to gain a deeper understanding of neuroplastic mechanisms that drive language recovery, with the ultimate goal of enhancing treatment outcomes. Subsequent to recent advances in neuroimaging techniques, we now have the ability to examine more closely how neural activity patterns change after a stroke. However, the way these neural activity changes relate to language impairments and language recovery is still debated. The aim of this review is to provide a theoretical framework to better investigate and interpret neuroplasticity mechanisms underlying language recovery in post-stroke aphasia. We detail two sets of neuroplasticity mechanisms observed at the synaptic level that may explain functional neuroimaging findings in post-stroke aphasia recovery at the network level: feedback-based homeostatic plasticity and associative Hebbian plasticity. In conjunction with these plasticity mechanisms, higher-order cognitive control processes dynamically modulate neural activity in other regions to meet communication demands, despite reduced neural resources. This work provides a network-level neurobiological framework for understanding neural changes observed in post-stroke aphasia and can be used to define guidelines for personalized treatment development.

Computer-Aided Diagnosis of Complications After Liver Transplantation Based on Transfer Learning

Liver transplantation is one of the most effective treatments for acute liver failure, cirrhosis, and even liver cancer. The prediction of postoperative complications is of great significance for liver transplantation. However, the existing prediction methods based on traditional machine learning are often unavailable or unreliable due to the insufficient amount of real liver transplantation data. Therefore, we propose a new framework to increase the accuracy of computer-aided diagnosis of complications after liver transplantation with transfer learning, which can handle small-scale but high-dimensional data problems. Furthermore, since data samples are often high dimensional in the real world, capturing key features that influence postoperative complications can help make the correct diagnosis for patients. So, we also introduce the SHapley Additive exPlanation (SHAP) method into our framework for exploring the key features of postoperative complications. We used data obtained from 425 patients with 456 features in our experiments. Experimental results show that our approach outperforms all compared baseline methods in predicting postoperative complications. In our work, the average precision, the mean recall, and the mean F1 score reach 91.22%, 91.70%, and 91.18%, respectively.

Diagnosis of cognitive and motor disorders levels in stroke patients through explainable machine learning based on MRI

BACKGROUND

Globally, stroke is the third most significant cause of disability. A stroke may produce motor, sensory, perceptual, or cognitive disorders that result in disability and affect the likelihood of recovery, affecting a person's ability to function. Evaluation post-stroke is critical for optimal stroke care.

PURPOSE

Traditional methods for classifying the clinical disorders of cognitive and motor in stroke patients use assessment and interrogative measures, which are time-consuming, complex, and labor-intensive. In response to the current situation, this study develops an algorithm to automatically classify motor and cognitive disorders in stroke patients by 3D brain MRI to assist physicians in diagnosis.

METHODS

First, radiomics and fusion features are extracted from the OAx T2 Propeller of 3D brain MRI. Then, we use 14 machine learning models and one model ensemble method to predict Fugl-Meyer and MMSE levels of stroke patients. Next, we evaluate the models using accuracy, recall, f1-score, and area under the curve (AUC). Finally, we employ SHAP to explain the output of the model.

RESULTS

The best predictive models come from Random Forest (RF) Classifier with fusion features in cognitive classification and Linear Discriminant Analysis (LDA) with radiomics features in motor classification. The highest accuracies are 92.0 and 82.5% for cognitive and motor disorders.

CONCLUSIONS

MRI brain maps can classify the cognitive and motor disorders of stroke patients. Radiomics features demonstrate its merits. The proposed algorithms with MRI images can efficiently assist physicians in diagnosing the cognitive and motor disorders of stroke patients in clinical practice. Additionally, this lessens labor costs, improves diagnostic effectiveness, and avoids the subjective difference that comes with manual assessment.

Multivariate lesion symptom mapping for predicting trajectories of recovery from aphasia

Individuals with post-stroke aphasia tend to recover their language to some extent; however, it remains challenging to reliably predict the nature and extent of recovery that will occur in the long term. The aim of this study was to quantitatively predict language outcomes in the first year of recovery from aphasia across multiple domains of language and at multiple timepoints post-stroke. We recruited 217 patients with aphasia following acute left hemisphere ischaemic or haemorrhagic stroke and evaluated their speech and language function using the Quick Aphasia Battery acutely and then acquired longitudinal follow-up data at up to three timepoints post-stroke: 1 month (n = 102), 3 months (n = 98) and 1 year (n = 74). We used support vector regression to predict language outcomes at each timepoint using acute clinical imaging data, demographic variables and initial aphasia severity as input. We found that ∼60% of the variance in long-term (1 year) aphasia severity could be predicted using these models, with detailed information about lesion location importantly contributing to these predictions. Predictions at the 1- and 3-month timepoints were somewhat less accurate based on lesion location alone, but reached comparable accuracy to predictions at the 1-year timepoint when initial aphasia severity was included in the models. Specific subdomains of language besides overall severity were predicted with varying but often similar degrees of accuracy. Our findings demonstrate the feasibility of using support vector regression models with leave-one-out cross-validation to make personalized predictions about long-term recovery from aphasia and provide a valuable neuroanatomical baseline upon which to build future models incorporating information beyond neuroanatomical and demographic predictors.

Progressive lesion necrosis is related to increasing aphasia severity in chronic stroke

BACKGROUND

Volumetric investigations of cortical damage resulting from stroke indicate that lesion size and shape continue to change even in the chronic stage of recovery. However, the potential clinical relevance of continued lesion growth has yet to be examined. In the present study, we investigated the prevalence of lesion expansion and the relationship between expansion and changes in aphasia severity in a large sample of individuals in the chronic stage of aphasia recovery.

METHODS

Retrospective structural MRI scans from 104 S survivors with at least 2 observations (k = 301 observations; mean time between scans = 31 months) were included. Lesion demarcation was performed using an automated lesion segmentation software and lesion volumes at each timepoint were subsequently calculated. A linear mixed effects model was conducted to investigate the effect of days between scan on lesion expansion. Finally, we investigated the association between lesion expansion and changes on the Western Aphasia Battery (WAB) in a group of participants assessed and scanned at 2 timepoints (N = 54) using a GLM.

RESULTS

Most participants (81 %) showed evidence of lesion expansion. The mixed effects model revealed lesion volumes significantly increase, on average, by 0.02 cc each day (7.3 cc per year) following a scan (p < 0.0001). Change on language performance was significantly associated with change in lesion volume (p = 0.025) and age at stroke (p = 0.031). The results suggest that with every 10 cc increase in lesion size, language performance decreases by 0.9 points, and for every 10-year increase in age at stroke, language performance decreases by 1.9 points.

CONCLUSIONS

The present study confirms and extends prior reports that lesion expansion occurs well into the chronic stage of stroke. For the first time, we present evidence that expansion is predictive of longitudinal changes in language performance in individuals with aphasia. Future research should focus on the potential mechanisms that may lead to necrosis in areas surrounding the chronic stroke lesion.

Global trends and research hotspots of stroke and magnetic resonance imaging: A bibliometric analysis

BACKGROUND

In this study, we used CiteSpace and VOSviewer to create a bibliometric visualization of research papers relating to stroke and magnetic resonance imaging (MRI) between 2000 and 2022. To fully understand the trends and hotspots in MRI and stroke research and provide new perspectives for future studies.

METHODS

The Web of Science Core Collection was selected as the source of data for this paper. Using CiteSpace and VOSviewer, publications were analyzed for authors, countries, institutions, journals, references, and keywords.

RESULTS

We found 1423 papers after searching and removing duplicates, which indicated an upward trend over the previous 23 years. Fiebach J.B. is the most published author (21 publications), Hacke W. is the most cited author (213 citations), and the United States (449 publications) and Harvard University (86 publications) are the most prolific nations and institutions. Stroke is the journal with the most co-citations (1275) and the most papers (171) published. The most representative reference was the 1995 article by Marler et al, which received 115 citations and had the top 3 co-occurring keywords: stroke, magnetic resonance imaging, and MRI. The article by Nogueria et al showed the strongest citation burst at the end of 2022 (strength = 17.32). High-frequency keywords in recent years are time, association, functional connectivity, thrombectomy, and rehabilitation.

CONCLUSION

This study provides a scientific perspective on stroke and MRI research, provides valuable information for researchers to understand the current status of research, hotspots, and trends, and guides future research directions.

Machine learning in the detection of dental cyst, tumor, and abscess lesions

BACKGROUND AND OBJECTIVE

Dental panoramic radiographs are utilized in computer-aided image analysis, which detects abnormal tissue masses by analyzing the produced image capacity to recognize patterns of intensity fluctuations. This is done to reduce the need for invasive biopsies for arriving to a diagnosis. The aim of the current study was to examine and compare the accuracy of several texture analysis techniques, such as Grey Level Run Length Matrix (GLRLM), Grey Level Co-occurrence Matrix (GLCM), and wavelet analysis in recognizing dental cyst, tumor, and abscess lesions.

MATERIALS & METHODS

The current retrospective study retrieved a total of 172 dental panoramic radiographs with lesion including dental cysts, tumors, or abscess. Radiographs that failed to meet technical criteria for diagnostic quality (such as significant overlap of teeth, a diffuse image, or distortion) were excluded from the sample. The methodology adopted in the study comprised of five stages. At first, the radiographs are improved, and the area of interest was segmented manually. A variety of feature extraction techniques, such GLCM, GLRLM, and the wavelet analysis were used to gather information from the area of interest. Later, the lesions were classified as a cyst, tumor, abscess, or using a support vector machine (SVM) classifier. Eventually, the data was transferred into a Microsoft Excel spreadsheet and statistical package for social sciences (SPSS) (version 21) was used to conduct the statistical analysis. Initially descriptive statistics were computed. For inferential analysis, statistical significance was determined by a p value < 0.05. The sensitivity, specificity, and accuracy were used to find the significant difference between assessed and actual diagnosis.

RESULTS

The findings demonstrate that 98% accuracy was achieved using GLCM, 91% accuracy using Wavelet analysis & 95% accuracy using GLRLM in distinguishing between dental cyst, tumor, and abscess lesions. The area under curve (AUC) number indicates that GLCM achieves a high degree of accuracy. The results achieved excellent accuracy (98%) using GLCM.

CONCLUSION

The GLCM features can be used for further research. After improving the performance and training, it can support routine histological diagnosis and can assist the clinicians in arriving at accurate and spontaneous treatment plans.

The Use of Artificial Intelligence to Predict the Prognosis of Patients Undergoing Central Nervous System Rehabilitation: A Narrative Review

Applications of machine learning in the healthcare field have become increasingly diverse. In this review, we investigated the integration of artificial intelligence (AI) in predicting the prognosis of patients with central nervous system disorders such as stroke, traumatic brain injury, and spinal cord injury. AI algorithms have shown promise in prognostic assessment, but challenges remain in achieving a higher prediction accuracy for practical clinical use. We suggest that accumulating more diverse data, including medical imaging and collaborative efforts among hospitals, can enhance the predictive capabilities of AI. As healthcare professionals become more familiar with AI, its role in central nervous system rehabilitation is expected to advance significantly, revolutionizing patient care.

Stroke lesion size - Still a useful biomarker for stroke severity and outcome in times of high-dimensional models

BACKGROUND

The volumetric size of a brain lesion is a frequently used stroke biomarker. It stands out among most imaging biomarkers for being a one-dimensional variable that is applicable in simple statistical models. In times of machine learning algorithms, the question arises of whether such a simple variable is still useful, or whether high-dimensional models on spatial lesion information are superior.

METHODS

We included 753 first-ever anterior circulation ischemic stroke patients (age 68.4±15.2 years; NIHSS at 24 h 4.4±5.1; modified Rankin Scale (mRS) at 3-months median[IQR] 1[0.75;3]) and traced lesions on diffusion-weighted MRI. In an out-of-sample model validation scheme, we predicted stroke severity as measured by NIHSS 24 h and functional stroke outcome as measured by mRS at 3 months either from spatial lesion features or lesion size.

RESULTS

For stroke severity, the best regression model based on lesion size performed significantly above chance (p < 0.0001) with R2 = 0.322, but models with spatial lesion features performed significantly better with R2 = 0.363 (t(752) = 2.889; p = 0.004). For stroke outcome, the best classification model based on lesion size again performed significantly above chance (p < 0.0001) with an accuracy of 62.8%, which was not different from the best model with spatial lesion features (62.6%, p = 0.80). With smaller training data sets of only 150 or 50 patients, the performance of high-dimensional models with spatial lesion features decreased up to the point of being equivalent or even inferior to models trained on lesion size. The combination of lesion size and spatial lesion features in one model did not improve predictions.

CONCLUSIONS

Lesion size is a decent biomarker for stroke outcome and severity that is slightly inferior to spatial lesion features but is particularly suited in studies with small samples. When low-dimensional models are desired, lesion size provides a viable proxy biomarker for spatial lesion features, whereas high-precision prediction models in personalised prognostic medicine should operate with high-dimensional spatial imaging features in large samples.

Therapies and Challenges in the Post-Stroke Aphasia Rehabilitation Arena: Current and Future Prospects

Aphasia is a serious consequence of stroke that results in a breakdown in communication. The course of aphasia recovery differs between afflicted individuals, and responsiveness to treatment cannot be predicted. Aphasiologists continue to investigate numerous behavioral treatment protocols that have shifted their focus to complimentary rehabilitation strategies. The aim of this study is threefold. First, to summarize the different categories of aphasia interventions post-stroke, considering their respective protocols, and present available evidence on the effectiveness of those protocols. Second, to document the challenges regarding the prediction of aphasia treatment response post-stroke in individual patients. Third, to report the challenges faced by researchers in recruiting people with aphasia (PWA) for treatment studies, and provide recommendations on how to increase participant recruitment and retention. This study provides up-to-date information on (i) effective therapies and aphasia recovery processes, and (ii) research recruitment hurdles together with potential strategies for overcoming them.

Variation in the rate of recovery in motor function between the upper and lower limbs in patients with stroke: some proposed hypotheses and their implications for research and practice

Background

Stroke results in impairment of motor function of both the upper and lower limbs. However, although it is debatable, motor function of the lower limb is believed to recover faster than that of the upper limb. The aim of this paper is to propose some hypotheses to explain the reasons for that, and discuss their implications for research and practice.

Method

We searched PubMED, Web of Science, Scopus, Embase and CENTRAL using the key words, stroke, cerebrovascular accident, upper extremity, lower extremity, and motor recovery for relevant literature.

Result

The search generated a total of 2,551 hits. However, out of this number, 51 duplicates were removed. Following review of the relevant literature, we proposed four hypotheses: natural instinct for walking hypothesis, bipedal locomotion hypothesis, central pattern generators (CPGs) hypothesis and role of spasticity hypothesis on the subject matter.

Conclusion

We opine that, what may eventually account for the difference, is the frequency of use of the affected limb or intensity of the rehabilitation intervention. This is because, from the above hypotheses, the lower limb seems to be used more frequently. When limbs are used frequently, this will result in use-dependent plasticity and eventual recovery. Thus, rehabilitation techniques that involve high repetitive tasks practice such as robotic rehabilitation, Wii gaming and constraint induced movement therapy should be used during upper limb rehabilitation.

Characteristics of post-stroke dysphagia: A retrospective study based on FEES

OBJECTIVE

This study aims to examine the characteristics of dysphagia in stroke patients with different lesion sites and explore the factors that impact the duration of nasogastric tube after post-stroke dysphagia (PSD).

METHODS

Patients with PSD were screened for analysis. Stroke types and lesion sites were confirmed using MRI or CT scans. Included patients were categorized into two groups: supratentorial stroke group (including lobar and deep intracerebral stroke subgroups) and infratentorial stroke group (including brainstem and cerebellar stroke subgroups). Fiberoptic endoscopic evaluation of swallowing (FEES), Penetration-aspiration scale (PAS), Yale pharyngeal residue (PR) severity rating scale, Functional oral intake scale (FOIS), Murray secretion severity rating scale (MSS), laryngopharyngeal sensation, and vocal fold mobility were investigated to assess the swallowing function.

RESULTS

A total of 94 patients were included in the final analysis. Significant differences were found in PR scores (p < .001), PAS scores (p < .05), MSS scores (p < .05), and vocal fold mobility (p < .001) between infratentorial and supratentorial stroke groups. Moreover, lobar stroke showed significantly higher PR scores compared to the deep intracerebral stroke group (p < .05). Kaplan-Meier survival analysis indicated significant differences in the duration of nasogastric tube placement among the following groups: infratentorial versus supratentorial stroke, PAS ≤ 5 versus PAS > 5, PR ≥ 3 versus PR < 3, and normal vocal fold mobility versus vocal fold motion impairment group (p < .05).

CONCLUSIONS

The infratentorial stroke may lead to worse swallowing function as compared to a supratentorial stroke. Additionally, patients with infratentorial stroke, PAS > 5, PR ≥ 3, or vocal fold motion impairment may contribute to a longer duration of nasogastric tube placement.

Mapping spoken language and cognitive deficits in post-stroke aphasia

Aphasia is an acquired disorder caused by damage, most commonly due to stroke, to brain regions involved in speech and language. While language impairment is the defining symptom of aphasia, the co-occurrence of non-language cognitive deficits and their importance in predicting rehabilitation and recovery outcomes is well documented. However, people with aphasia (PWA) are rarely tested on higher-order cognitive functions, making it difficult for studies to associate these functions with a consistent lesion correlate. Broca's area is a particular brain region of interest that has long been implicated in speech and language production. Contrary to classic models of speech and language, cumulative evidence shows that Broca's area and surrounding regions in the left inferior frontal cortex (LIFC) are involved in, but not specific to, speech production. In this study we aimed to explore the brain-behaviour relationships between tests of cognitive skill and language abilities in thirty-six adults with long-term speech production deficits caused by post-stroke aphasia. Our findings suggest that non-linguistic cognitive functions, namely executive functions and verbal working memory, explain more of the behavioural variance in PWA than classical language models imply. Additionally, lesions to the LIFC, including Broca's area, were associated with non-linguistic executive (dys)function, suggesting that lesions to this area are associated with non-language-specific higher-order cognitive deficits in aphasia. Whether executive (dys)function - and its neural correlate in Broca's area - contributes directly to PWA's language production deficits or simply co-occurs with it, adding to communication difficulties, remains unclear. These findings support contemporary models of speech production that place language processing within the context of domain-general perception, action and conceptual knowledge. An understanding of the covariance between language and non-language deficits and their underlying neural correlates will inform better targeted aphasia treatment and outcomes.

Interpreting and validating complexity and causality in lesion-symptom prognoses

This paper considers the steps needed to generate pragmatic and interpretable lesion-symptom mappings that can be used for clinically reliable prognoses. The novel contributions are 3-fold. We first define and inter-relate five neurobiological and five methodological constraints that need to be accounted for when interpreting lesion-symptom associations and generating synthetic lesion data. The first implication is that, because of these constraints, lesion-symptom mapping needs to focus on probabilistic relationships between Lesion and Symptom, with Lesion as a multivariate spatial pattern, Symptom as a time-dependent behavioural profile and evidence that Lesion raises the probability of Symptom. The second implication is that in order to assess the strength of probabilistic causality, we need to distinguish between causal lesion sites, incidental lesion sites, spared but dysfunctional sites and intact sites, all of which might affect the accuracy of the predictions and prognoses generated. We then formulate lesion-symptom mappings in logical notations, including combinatorial rules, that are then used to evaluate and better understand complex brain-behaviour relationships. The logical and theoretical framework presented applies to any type of neurological disorder but is primarily discussed in relationship to stroke damage. Accommodating the identified constraints, we discuss how the 1965 Bradford Hill criteria for inferring probabilistic causality, post hoc, from observed correlations in epidemiology-can be applied to lesion-symptom mapping in stroke survivors. Finally, we propose that rather than rely on post hoc evaluation of how well the causality criteria have been met, the neurobiological and methodological constraints should be addressed, a priori, by changing the experimental design of lesion-symptom mappings and setting up an open platform to share and validate the discovery of reliable and accurate lesion rules that are clinically useful.

Improvement of predictive accuracies of functional outcomes after subacute stroke inpatient rehabilitation by machine learning models

OBJECTIVES

Stepwise linear regression (SLR) is the most common approach to predicting activities of daily living at discharge with the Functional Independence Measure (FIM) in stroke patients, but noisy nonlinear clinical data decrease the predictive accuracies of SLR. Machine learning is gaining attention in the medical field for such nonlinear data. Previous studies reported that machine learning models, regression tree (RT), ensemble learning (EL), artificial neural networks (ANNs), support vector regression (SVR), and Gaussian process regression (GPR), are robust to such data and increase predictive accuracies. This study aimed to compare the predictive accuracies of SLR and these machine learning models for FIM scores in stroke patients.

METHODS

Subacute stroke patients (N = 1,046) who underwent inpatient rehabilitation participated in this study. Only patients' background characteristics and FIM scores at admission were used to build each predictive model of SLR, RT, EL, ANN, SVR, and GPR with 10-fold cross-validation. The coefficient of determination (R2) and root mean square error (RMSE) values were compared between the actual and predicted discharge FIM scores and FIM gain.

RESULTS

Machine learning models (R2 of RT = 0.75, EL = 0.78, ANN = 0.81, SVR = 0.80, GPR = 0.81) outperformed SLR (0.70) to predict discharge FIM motor scores. The predictive accuracies of machine learning methods for FIM total gain (R2 of RT = 0.48, EL = 0.51, ANN = 0.50, SVR = 0.51, GPR = 0.54) were also better than of SLR (0.22).

CONCLUSIONS

This study suggested that the machine learning models outperformed SLR for predicting FIM prognosis. The machine learning models used only patients' background characteristics and FIM scores at admission and more accurately predicted FIM gain than previous studies. ANN, SVR, and GPR outperformed RT and EL. GPR could have the best predictive accuracy for FIM prognosis.

Using in vivo functional and structural connectivity to predict chronic stroke aphasia deficits

Focal brain damage caused by stroke can result in aphasia and advances in cognitive neuroscience suggest that impairment may be associated with network-level disorder rather than just circumscribed cortical damage. Several studies have shown meaningful relationships between brain-behaviour using lesions; however, only a handful of studies have incorporated in vivo structural and functional connectivity. Patients with chronic post-stroke aphasia were assessed with structural (n = 68) and functional (n = 39) MRI to assess whether predicting performance can be improved with multiple modalities and if additional variance can be explained compared to lesion models alone. These neural measurements were used to construct models to predict four key language-cognitive factors: (i) phonology; (ii) semantics; (iii) executive function; and (iv) fluency. Our results showed that each factor (except executive ability) could be significantly related to each neural measurement alone; however, structural and functional connectivity models did not explain additional variance above the lesion models. We did find evidence that the structural and functional predictors may be linked to the core lesion sites. First, the predictive functional connectivity features were found to be located within functional resting-state networks identified in healthy controls, suggesting that the result might reflect functionally specific reorganization (damage to a node within a network can result in disruption to the entire network). Second, predictive structural connectivity features were located within core lesion sites, suggesting that multimodal information may be redundant in prediction modelling. In addition, we observed that the optimum sparsity within the regularized regression models differed for each behavioural component and across different imaging features, suggesting that future studies should consider optimizing hyperparameters related to sparsity per target. Together, the results indicate that the observed network-level disruption was predicted by the lesion alone and does not significantly improve model performance in predicting the profile of language impairment.

DEEP LEARNING PREDICTION OF MOTOR PERFORMANCE IN STROKE INDIVIDUALS USING NEUROIMAGING DATA

The degree of motor impairment and profile of recovery after stroke are difficult to predict for each individual. Measures obtained from clinical assessments, as well as neurophysiological and neuroimaging techniques have been used as potential biomarkers of motor recovery, with limited accuracy up to date. To address this, the present study aimed to develop a deep learning model based on structural brain images obtained from stroke participants and healthy volunteers. The following inputs were used in a multi-channel 3D convolutional neural network (CNN) model: fractional anisotropy, mean diffusivity, radial diffusivity, and axial diffusivity maps obtained from Diffusion Tensor Imaging (DTI) images, white and gray matter intensity values obtained from Magnetic Resonance Imaging, as well as demographic data (e.g., age, gender). Upper limb motor function was classified into "Poor" and "Good" categories. To assess the performance of the DL model, we compared it to more standard machine learning (ML) classifiers including k-nearest neighbor, support vector machines (SVM), Decision Trees, Random Forests, Ada Boosting, and Naïve Bayes, whereby the inputs of these classifiers were the features taken from the fully connected layer of the CNN model. The highest accuracy and area under the curve values were 0.92 and 0.92 for the 3D-CNN and 0.91 and 0.91 for the SVM, respectively. The multi-channel 3D-CNN with residual blocks and SVM supported by DL was more accurate than traditional ML methods to classify upper limb motor impairment in the stroke population. These results suggest that combining volumetric DTI maps and measures of white and gray matter integrity can improve the prediction of the degree of motor impairment after stroke. Identifying the potential of recovery early on after a stroke could promote the allocation of resources to optimize the functional independence of these individuals and their quality of life.

The elusive metric of lesion load

One of the widely used metrics in lesion-symptom mapping is lesion load that codes the amount of damage to a given brain region of interest. Lesion load aims to reduce the complex 3D lesion information into a feature that can reflect both site of damage, defined by the location of the region of interest, and size of damage within that region of interest. Basically, the process of estimation of lesion load converts a voxel-based lesion map into a region-based lesion map, with regions defined as atlas-based or data-driven spatial patterns. Here, after examining current definitions of lesion load, four methodological issues are discussed: (1) lesion load is agnostic to the location of damage within the region of interest, and it disregards damage outside the region of interest, (2) lesion load estimates are prone to errors introduced by the uncertainty in lesion delineation, spatial warping of the lesion/region, and binarization of the lesion/region, (3) lesion load calculation depends on brain parcellation selection, and (4) lesion load does not necessarily reflect a white matter disconnection. Overall, lesion load, when calculated in a robust way, can serve as a clinically-useful feature for explaining and predicting post-stroke outcome and recovery.

Recovery from aphasia in the first year after stroke

Most individuals who experience aphasia after a stroke recover to some extent, with the majority of gains taking place in the first year. The nature and time course of this recovery process is only partially understood, especially its dependence on lesion location and extent, which are the most important determinants of outcome. The aim of this study was to provide a comprehensive description of patterns of recovery from aphasia in the first year after stroke. We recruited 334 patients with acute left hemisphere supratentorial ischaemic or haemorrhagic stroke and evaluated their speech and language function within 5 days using the Quick Aphasia Battery (QAB). At this initial time point, 218 patients presented with aphasia. Individuals with aphasia were followed longitudinally, with follow-up evaluations of speech and language at 1 month, 3 months, and 1 year post-stroke, wherever possible. Lesions were manually delineated based on acute clinical MRI or CT imaging. Patients with and without aphasia were divided into 13 groups of individuals with similar, commonly occurring patterns of brain damage. Trajectories of recovery were then investigated as a function of group (i.e. lesion location and extent) and speech/language domain (overall language function, word comprehension, sentence comprehension, word finding, grammatical construction, phonological encoding, speech motor programming, speech motor execution, and reading). We found that aphasia is dynamic, multidimensional, and gradated, with little explanatory role for aphasia subtypes or binary concepts such as fluency. Patients with circumscribed frontal lesions recovered well, consistent with some previous observations. More surprisingly, most patients with larger frontal lesions extending into the parietal or temporal lobes also recovered well, as did patients with relatively circumscribed temporal, temporoparietal, or parietal lesions. Persistent moderate or severe deficits were common only in patients with extensive damage throughout the middle cerebral artery distribution or extensive temporoparietal damage. There were striking differences between speech/language domains in their rates of recovery and relationships to overall language function, suggesting that specific domains differ in the extent to which they are redundantly represented throughout the language network, as opposed to depending on specialized cortical substrates. Our findings have an immediate clinical application in that they will enable clinicians to estimate the likely course of recovery for individual patients, as well as the uncertainty of these predictions, based on acutely observable neurological factors.

Treatment-induced neural reorganization in aphasia is language-domain specific: Evidence from a large-scale fMRI study

Studies investigating the effects of language intervention on the re-organization of language networks in chronic aphasia have resulted in mixed findings, likely related to-among other factors-the language function targeted during treatment. The present study investigated the effects of the type of treatment provided on neural reorganization. Seventy individuals with chronic stroke-induced aphasia, recruited from three research laboratories and meeting criteria for agrammatism, anomia or dysgraphia were assigned to either treatment (N = 51) or control (N = 19) groups. Participants in the treatment group received 12-weeks of language intervention targeting sentence comprehension/production, naming, or spelling. At baseline and post-testing, all participants performed an fMRI story comprehension task, with blocks of auditorily-presented stories alternated with blocks of reversed speech. Participants in the treatment, but not control, group significantly improved in the treated language domain. FMRI region-of-interest (ROI) analyses, conducted within regions that were either active (or homologous to active) regions in a group of 22 healthy participants on the story comprehension task, revealed a significant increase in activation from pre-to post-treatment in right-hemisphere homologues of these regions for participants in the sentence and spelling, but not naming, treatment groups, not predicted by left-hemisphere lesion size. For the sentence (but not the spelling) treatment group, activation changes within right-hemisphere homologues of language regions were positively associated with changes in measures of verb and sentence comprehension. These findings support previous research pointing to recruitment of right hemisphere tissue as a viable route for language recovery and suggest that sentence-level treatment may promote greater neuroplasticity on naturalistic, language comprehension tasks, compared to word-level treatment.

Neural basis of dysphagia in stroke: A systematic review and meta-analysis

Objectives

Dysphagia is a major cause of stroke infection and death, and identification of structural and functional brain area changes associated with post-stroke dysphagia (PSD) can help in early screening and clinical intervention. Studies on PSD have reported numerous structural lesions and functional abnormalities in brain regions, and a systematic review is lacking. We aimed to integrate several neuroimaging studies to summarize the empirical evidence of neurological changes leading to PSD.

Methods

We conducted a systematic review of studies that used structural neuroimaging and functional neuroimaging approaches to explore structural and functional brain regions associated with swallowing after stroke, with additional evidence using a live activation likelihood estimation (ALE) approach.

Results

A total of 35 studies were included, including 20 studies with structural neuroimaging analysis, 14 studies with functional neuroimaging analysis and one study reporting results for both. The overall results suggest that structural lesions and functional abnormalities in the sensorimotor cortex, insula, cerebellum, cingulate gyrus, thalamus, basal ganglia, and associated white matter connections in individuals with stroke may contribute to dysphagia, and the ALE analysis provides additional evidence for structural lesions in the right lentiform nucleus and right thalamus and functional abnormalities in the left thalamus.

Conclusion

Our findings suggest that PSD is associated with neurological changes in brain regions such as sensorimotor cortex, insula, cerebellum, cingulate gyrus, thalamus, basal ganglia, and associated white matter connections. Adequate understanding of the mechanisms of neural changes in the post-stroke swallowing network may assist in clinical diagnosis and provide ideas for the development of new interventions in clinical practice.

Predicting aging trajectories of decline in brain volume, cortical thickness and fractional anisotropy in schizophrenia

Brain-age prediction is a novel approach to assessing deviated brain aging trajectories in different diseases. However, most studies have used an average brain age gap (BAG) of individuals with schizophrenia of different illness durations for comparison with healthy participants. Therefore, this study investigated whether declined brain structures as reflected by BAGs may be present in schizophrenia in terms of brain volume, cortical thickness, and fractional anisotropy across different illness durations. We used brain volume, cortical thickness, and fractional anisotropy as features to train three models from the training dataset. Three models were applied to predict brain ages in the hold-out test and schizophrenia datasets and calculate BAGs. We divided the schizophrenia dataset into multiple groups based on the illness duration using a sliding time window approach for ANCOVA analysis. The brain volume and cortical thickness models revealed that, in comparison with healthy controls, individuals with schizophrenia had larger BAGs across different illness durations, whereas the BAG in terms of fractional anisotropy did not differ from that of healthy controls after disease onset. Moreover, the BAG at the initial stage of schizophrenia was the largest in the cortical thickness model. In contrast, the BAG from approximately two decades after disease onset was the largest in the brain volume model. Our findings suggest that schizophrenia differentially affects the decline of different brain structures during the disease course. Moreover, different trends of decline in thickness and volume-based measures suggest a differential decline in dimensions of brain structure throughout the course of schizophrenia.

Machine learning segmentation of core and penumbra from acute stroke CT perfusion data

Introduction

Computed tomography perfusion (CTP) imaging is widely used in cases of suspected acute ischemic stroke to positively identify ischemia and assess suitability for treatment through identification of reversible and irreversible tissue injury. Traditionally, this has been done via setting single perfusion thresholds on two or four CTP parameter maps. We present an alternative model for the estimation of tissue fate using multiple perfusion measures simultaneously.

Methods

We used machine learning (ML) models based on four different algorithms, combining four CTP measures (cerebral blood flow, cerebral blood volume, mean transit time and delay time) plus 3D-neighborhood (patch) analysis to predict the acute ischemic core and perfusion lesion volumes. The model was developed using 86 patient images, and then tested further on 22 images.

Results

XGBoost was the highest-performing algorithm. With standard threshold-based core and penumbra measures as the reference, the model demonstrated moderate agreement in segmenting core and penumbra on test images. Dice similarity coefficients for core and penumbra were 0.38 ± 0.26 and 0.50 ± 0.21, respectively, demonstrating moderate agreement. Skull-related image artefacts contributed to lower accuracy.

Discussion

Further development may enable us to move beyond the current overly simplistic core and penumbra definitions using single thresholds where a single error or artefact may lead to substantial error.

The role of the hippocampus in statistical learning and language recovery in persons with post stroke aphasia

Although several studies have aimed for accurate predictions of language recovery in post stroke aphasia, individual language outcomes remain hard to predict. Large-scale prediction models are built using data from patients mainly in the chronic phase after stroke, although it is clinically more relevant to consider data from the acute phase. Previous research has mainly focused on deficits, i.e., behavioral deficits or specific brain damage, rather than compensatory mechanisms, i.e., intact cognitive skills or undamaged brain regions. One such unexplored brain region that might support language (re)learning in aphasia is the hippocampus, a region that has commonly been associated with an individual's learning potential, including statistical learning. This refers to a set of mechanisms upon which we rely heavily in daily life to learn a range of regularities across cognitive domains. Against this background, thirty-three patients with aphasia (22 males and 11 females, M = 69.76 years, SD = 10.57 years) were followed for 1 year in the acute (1-2 weeks), subacute (3-6 months) and chronic phase (9-12 months) post stroke. We evaluated the unique predictive value of early structural hippocampal measures for short-term and long-term language outcomes (measured by the ANELT). In addition, we investigated whether statistical learning abilities were intact in patients with aphasia using three different tasks: an auditory-linguistic and visual task based on the computation of transitional probabilities and a visuomotor serial reaction time task. Finally, we examined the association of individuals' statistical learning potential with acute measures of hippocampal gray and white matter. Using Bayesian statistics, we found moderate evidence for the contribution of left hippocampal gray matter in the acute phase to the prediction of long-term language outcomes, over and above information on the lesion and the initial language deficit (measured by the ScreeLing). Non-linguistic statistical learning in patients with aphasia, measured in the subacute phase, was intact at the group level compared to 23 healthy older controls (8 males and 15 females, M = 74.09 years, SD = 6.76 years). Visuomotor statistical learning correlated with acute hippocampal gray and white matter. These findings reveal that particularly left hippocampal gray matter in the acute phase is a potential marker of language recovery after stroke, possibly through its statistical learning ability.

Anterior connectivity critical for recovery of connected speech after stroke

Connected speech recovers to different degrees across people after left hemisphere stroke, but white matter predictors of differential recovery from the acute stage of stroke are unknown. We assessed changes in lexical-syntactic aspects of connected speech in a longitudinal analysis of 40 individuals (18 females) from the acute stage of left hemisphere stroke (within an average of 4 days post-stroke) to subacute (within 2 months) and chronic stages (early: 6 months, late: 1 year) while measuring the extent of acute lesions on white matter tracts to identify tracts predictive of recovery. We found that acute damage to the frontal aslant tract led to a decreased recovery of the fluency and structural complexity of connected speech during the year following left hemisphere stroke. The results were independent of baseline performance, overall lesion volume and the proportion of damage to tract-adjacent grey matter. This longitudinal analysis from acute to chronic stroke provides the first evidence that recovery of fluent and structurally complex spontaneous connected speech requires intact left frontal connectivity via the frontal aslant tract. That the frontal aslant tract was critical for recovery at early as well as later stages of stroke demonstrates that anterior connectivity plays a lasting and important role for the reorganization of function related to the successful production of connected speech.

Structural disconnections associated with language impairments in chronic post-stroke aphasia using disconnectome maps

Inconsistent findings have been reported about the impact of structural disconnections on language function in post-stroke aphasia. This study investigated patterns of structural disconnections associated with chronic language impairments using disconnectome maps. Seventy-six individuals with post-stroke aphasia underwent a battery of language assessments and a structural MRI scan. Support-vector regression disconnectome-symptom mapping analyses were performed to examine the correlations between disconnectome maps, representing the probability of disconnection at each white matter voxel and different language scores. To further understand whether significant disconnections were primarily representing focal damage or a more extended network of seemingly preserved but disconnected areas beyond the lesion site, results were qualitatively compared to support-vector regression lesion-symptom mapping analyses. Part of the left white matter perisylvian network was similarly disconnected in 90% of the individuals with aphasia. Surrounding this common left perisylvian disconnectome, specific structural disconnections in the left fronto-temporo-parietal network were significantly associated with aphasia severity and with lower performance in auditory comprehension, syntactic comprehension, syntactic production, repetition and naming tasks. Auditory comprehension, repetition and syntactic processing deficits were related to disconnections in areas that overlapped with and extended beyond lesion sites significant in SVR-LSM analyses. In contrast, overall language abilities as measured by aphasia severity and naming seemed to be mostly explained by focal damage at the level of the insular and central opercular cortices, given the high overlap between SVR-DSM and SVR-LSM results for these scores. While focal damage seems to be sufficient to explain broad measures of language performance, the structural disconnections between language areas provide additional information on the neural basis of specific and persistent language impairments at the chronic stage beyond lesion volume. Leveraging routinely available clinical data, disconnectome mapping furthers our understanding of anatomical connectivity constraints that may limit the recovery of some language abilities in chronic post-stroke aphasia.

Efficient and effective assessment of deficits and their neural bases in stroke aphasia

OBJECTIVE

Multi-assessment batteries are necessary for diagnosing and quantifying the multifaceted deficits observed post-stroke. Extensive batteries are thorough but impractically long for clinical settings or large-scale research studies. Clinically-targeted "shallow" batteries superficially cover a wide range of language skills relatively quickly but can struggle to identify mild deficits or quantify the impairment level. Our aim was to compare these batteries across a large group of chronic stroke aphasia and to test a novel data-driven reduced version of an extensive battery that maintained sensitivity to mild impairment, ability to grade deficits and the underlying component structure.

METHODS

We tested 75 chronic left-sided stroke participants, spanning global to mild aphasia. The underlying structure of these three batteries was analysed using cross-validation and principal component analysis, in addition to univariate and multivariate lesion-symptom mapping.

RESULTS

This revealed a four-factor solution for the extensive and data-reduced batteries, identifying phonology, semantic skills, fluency and executive function in contrast to a two-factor solution using the shallow battery (language severity and cognitive severity). Lesion symptom mapping using participants' factor scores identified convergent neural structures for phonology (superior temporal gyrus), semantics (inferior temporal gyrus), speech fluency (precentral gyrus) and executive function (lateral occipitotemporal cortex). The two shallow battery components converged with the phonology and executive function clusters. In addition, we show that multivariate models could predict the component scores using neural data, however not for every component.

CONCLUSIONS

Overall, the data-driven battery appears to be an effective way to save time yet retain maintained sensitivity to mild impairment, ability to grade deficits and the underlying component structure observed in post-stroke aphasia.

Better long-term speech outcomes in stroke survivors who received early clinical speech and language therapy: What's driving recovery?

Establishing whether speech and language therapy after stroke has beneficial effects on speaking ability is challenging because of the need to control for multiple non-therapy factors known to influence recovery. We investigated how speaking ability at three time points post-stroke differed in patients who received varying amounts of clinical therapy in the first month post-stroke. In contrast to prior studies, we factored out variance from: initial severity of speaking impairment, amount of later therapy, and left and right hemisphere lesion size and site. We found that speaking ability at one month post-stroke was significantly better in patients who received early therapy (n = 79), versus those who did not (n = 64), and the number of hours of early therapy was positively related to recovery at one year post-stroke. We offer two non-mutually exclusive interpretations of these data: (1) patients may benefit from the early provision of self-management strategies; (2) therapy is more likely to be provided to patients who have a better chance of recovery (e.g., poor physical and/or mental health may impact suitability for therapy and chance of recovery). Both interpretations have implications for future studies aiming to predict individual patients' speech outcomes after stroke, and their response to therapy.

Oscillatory beta/alpha band modulations: A potential biomarker of functional language and motor recovery in chronic stroke?

Stroke remains one of the leading causes of various disabilities, including debilitating motor and language impairments. Though various treatments exist, post-stroke impairments frequently become chronic, dramatically reducing daily life quality, and requiring specific rehabilitation. A critical goal of chronic stroke rehabilitation is to induce, usually through behavioral training, experience-dependent plasticity processes in order to promote functional recovery. However, the efficiency of such interventions is typically modest, and very little is known regarding the neural dynamics underpinning recovery processes and possible biomarkers of their efficiency. Some studies have emphasized specific alterations of excitatory–inhibitory balance within distributed neural networks as an important recovery correlate. Neural processes sensitive to these alterations, such as task-dependent oscillatory activity in beta as well as alpha bands, may be candidate biomarkers of chronic stroke functional recovery. In this review, we discuss the results of studies on motor and language recovery with a focus on oscillatory processes centered around the beta band and their modulations during functional recovery in chronic stroke. The discussion is based on a framework where task-dependent modulations of beta and alpha oscillatory activity, generated by the deep cortical excitatory–inhibitory microcircuits, serve as a neural mechanism of domain-general top-down control processes. We discuss the findings, their limitations, and possible directions for future research.

Vascular Cognitive Impairment After Mild Stroke: Connectomic Insights, Neuroimaging, and Knowledge Translation

Contemporary stroke assessment protocols have a limited ability to detect vascular cognitive impairment (VCI), especially among those with subtle deficits. This lesser-involved categorization, termed mild stroke (MiS), can manifest compromised processing speed that negatively impacts cognition. From a neurorehabilitation perspective, research spanning neuroimaging, neuroinformatics, and cognitive neuroscience supports that processing speed is a valuable proxy for complex neurocognitive operations, insofar as inefficient neural network computation significantly affects daily task performance. This impact is particularly evident when high cognitive loads compromise network efficiency by challenging task speed, complexity, and duration. Screening for VCI using processing speed metrics can be more sensitive and specific. Further, they can inform rehabilitation approaches that enhance patient recovery, clarify the construct of MiS, support clinician-researcher symbiosis, and further clarify the occupational therapy role in targeting functional cognition. To this end, we review relationships between insult-derived connectome alterations and VCI, and discuss novel clinical approaches for identifying disruptions of neural networks and white matter connectivity. Furthermore, we will frame knowledge translation efforts to leverage insights from cutting-edge structural and functional connectomics research. Lastly, we highlight how occupational therapists can provide expertise as knowledge brokers acting within their established scope of practice to drive substantive clinical innovation.

Artificially-reconstructed brain images with stroke lesions from non-imaging data: modeling in categorized patients based on lesion occurrence and sparsity

Brain imaging is necessary for understanding disease symptoms, including stroke. However, frequent imaging procedures encounter practical limitations. Estimating the brain information (e.g., lesions) without imaging sessions is beneficial for this scenario. Prospective estimating variables are non-imaging data collected from standard tests. Therefore, the current study aims to examine the variable feasibility for modelling lesion locations. Heterogeneous variables were employed in the multivariate logistic regression. Furthermore, patients were categorized (i.e., unsupervised clustering through k-means method) by the charasteristics of lesion occurrence (i.e., ratio between the lesioned and total regions) and sparsity (i.e., density measure of lesion occurrences across regions). Considering those charasteristics in models improved estimation performances. Lesions (116 regions in Automated Anatomical Labeling) were adequately predicted (sensitivity: 80.0-87.5% in median). We confirmed that the usability of models was extendable to different resolution levels in the brain region of interest (e.g., lobes, hemispheres). Patients' charateristics (i.e., occurrence and sparsity) might also be explained by the non-imaging data as well. Advantages of the current approach can be experienced by any patients (i.e., with or without imaging sessions) in any clinical facilities (i.e., with or without imaging instrumentation).

Psychological Stress Management and Stress Reduction Strategies for Stroke Survivors: A Scoping Review

BACKGROUND

Stroke can be a life-changing event, with survivors frequently experiencing some level of disability, reduced independence, and an abrupt lifestyle change. Not surprisingly, many stroke survivors report elevated levels of stress during the recovery process, which has been associated with worse outcomes.

PURPOSE

Given the multiple roles of stress in the etiology of stroke recovery outcomes, we aimed to scope the existing literature on stress management interventions that have been trialed in stroke survivors.

METHODS

We performed a database search for intervention studies conducted in stroke survivors which reported the effects on stress, resilience, or coping outcome. Medline (OVID), Embase (OVID), CINAHL (EBSCO), Cochrane Library, and PsycInfo (OVID) were searched from database inception until March 11, 2019, and updated on September 1, 2020.

RESULTS

Twenty-four studies met the inclusion criteria. There was significant variation in the range of trialed interventions, as well as the outcome measures used to assess stress. Overall, just over half (13/24) of the included studies reported a benefit in terms of stress reduction. Acceptability and feasibility were considered in 71% (17/24) and costs were considered in 17% (4/24) of studies. The management of stress was rarely linked to the prevention of symptoms of stress-related disorders. The overall evidence base of included studies is weak. However, an increase in the number of studies over time suggests a growing interest in this subject.

CONCLUSIONS

Further research is required to identify optimum stress management interventions in stroke survivors, including whether the management of stress can ameliorate the negative impacts of stress on health.

Syntax Acquisition in Healthy Adults and Post-Stroke Individuals: The Intriguing Role of Grammatical Preference, Statistical Learning, and Education

Previous work has provided contrasting evidence on syntax acquisition. Syntax-internal factors, i.e., instinctive knowledge of the universals of grammar (UG) for finite-state grammar (FSG) and phrase-structure grammar (PSG) but also syntax-external factors such as language competence, working memory (WM) and demographic factors may affect syntax acquisition. This study employed an artificial grammar paradigm to identify which factors predicted syntax acquisition. Thirty-seven healthy individuals and forty-nine left-hemispheric stroke patients (fourteen with aphasia) read syllable sequences adhering to or violating FSG and PSG. They performed preference classifications followed by grammatical classifications (after training). Results showed the best classification accuracy for sequences adhering to UG, with performance predicted by syntactic competence and spatial WM. Classification of ungrammatical sequences improved after training and was predicted by verbal WM. Although accuracy on FSG was better than on PSG, generalization was fully possible only for PSG. Education was the best predictor of syntax acquisition, while aphasia and lesion volume were not predictors. This study shows a clear preference for UG, which is influenced by spatial and linguistic knowledge, but not by the presence of aphasia. Verbal WM supported the identification of rule violations. Moreover, the acquisition of FSG and PSG was related to partially different mechanisms, but both depended on education.

Language systems from lesion-symptom mapping in aphasia: A meta-analysis of voxel-based lesion mapping studies

•

Meta-analysis of 2,007 individuals with aphasia from 25 voxel-based lesion mapping studies.

•

Distinctive patterns of lesions in aphasia are associated with different language functions.

•

The patterns of lesion in aphasia support the dual pathway model of language processing.

Background

Aphasia is one of the most common causes of post-stroke disabilities. As the symptoms and impact of post-stroke aphasia are heterogeneous, it is important to understand how topographical lesion heterogeneity in patients with aphasia is associated with different domains of language impairments. Here, we aim to provide a comprehensive overview of neuroanatomical basis in post-stroke aphasia through coordinate based meta-analysis of voxel-based lesion-symptom mapping studies.

Methods

We performed a meta-analysis of lesion-symptom mapping studies in post-stroke aphasia. We obtained coordinate-based structural neuroimaging data for 2,007 individuals with aphasia from 25 studies that met predefined inclusion criteria.

Results

Overall, our results revealed that the distinctive patterns of lesions in aphasia are associated with different language functions and tasks. Damage to the insular-motor areas impaired speech with preserved comprehension and a similar pattern was observed when the lesion covered the insular-motor and inferior parietal lobule. Lesions in the frontal area severely impaired speaking with relatively good comprehension. The repetition-selective deficits only arise from lesions involving the posterior superior temporal gyrus. Damage in the anterior-to-posterior temporal cortex was associated with semantic deficits.

Conclusion

The association patterns of lesion topography and specific language deficits provide key insights into the specific underlying language pathways. Our meta-analysis results strongly support the dual pathway model of language processing, capturing the link between the different symptom complexes of aphasias and the different underlying location of damage.

Predictors of Therapy Response in Chronic Aphasia: Building a Foundation for Personalized Aphasia Therapy

Chronic aphasia, a devastating impairment of language, affects up to a third of stroke survivors. Speech and language therapy has consistently been shown to improve language function in prior clinical trials, but few clinicially applicable predictors of individual therapy response have been identified to date. Consequently, clinicians struggle substantially with prognostication in the clinical management of aphasia. A rising prevalence of aphasia, in particular in younger populations, has emphasized the increasing demand for a personalized approach to aphasia therapy, that is, therapy aimed at maximizing language recovery of each individual with reference to evidence-based clinical recommendations. In this narrative review, we discuss the current state of the literature with respect to commonly studied predictors of therapy response in aphasia. In particular, we focus our discussion on biographical, neuropsychological, and neurobiological predictors, and emphasize limitations of the literature, summarize consistent findings, and consider how the research field can better support the development of personalized aphasia therapy. In conclusion, a review of the literature indicates that future research efforts should aim to recruit larger samples of people with aphasia, including by establishing multisite aphasia research centers.

Stroke population–specific neuroanatomical CT-MRI brain atlas

Purpose

Development of a freely available stroke population–specific anatomical CT/MRI atlas with a reliable normalisation pipeline for clinical CT.

Methods

By reviewing CT scans in suspected stroke patients and filtering the AIBL MRI database, respectively, we collected 50 normal-for-age CT and MRI scans to build a standard-resolution CT template and a high-resolution MRI template. The latter was manually segmented into anatomical brain regions. We then developed and validated a MRI to CT registration pipeline to align the MRI atlas onto the CT template. Finally, we developed a CT-to-CT-normalisation pipeline and tested its reliability by calculating Dice coefficient (Dice) and Average Hausdorff Distance (AHD) for predefined areas in 100 CT scans from ischaemic stroke patients.

Results

The resulting CT/MRI templates were age and sex matched to a general stroke population (median age 71.9 years (62.1–80.2), 60% male). Specifically, this accounts for relevant structural changes related to aging, which may affect registration. Applying the validated MRI to CT alignment (Dice > 0.78, Average Hausdorff Distance < 0.59 mm) resulted in our final CT-MRI atlas. The atlas has 52 manually segmented regions and covers the whole brain. The alignment of four cortical and subcortical brain regions with our CT-normalisation pipeline was reliable for small/medium/large infarct lesions (Dice coefficient > 0.5).

Conclusion

The newly created CT-MRI brain atlas has the potential to standardise stroke lesion segmentation. Together with the automated normalisation pipeline, it allows analysis of existing and new datasets to improve prediction tools for stroke patients (free download at https://forms.office.com/r/v4t3sWfbKs).

Supplementary Information

The online version contains supplementary material available at 10.1007/s00234-021-02875-9.

Functional MRI evidence for reorganization of language networks after stroke

In this chapter, we review fMRI evidence for language reorganization in individuals with poststroke aphasia. Several studies in the current literature have utilized fMRI as a tool to understand patterns of functional reorganization in poststroke aphasia. Consistent with previous models that have been proposed to explain the trajectory of language recovery, differential patterns of language processing and language recovery have been identified across individuals with poststroke aphasia in different stages of recovery. Overall, a global network breakdown typically occurs in the early stages of aphasia recovery, followed by normalization in "traditional" left hemisphere language networks. Depending on individual characteristics, right hemisphere regions and bilateral domain-general regions may be further recruited. The main takeaway of this chapter is that poststroke aphasia recovery does not depend on individual neural regions, but rather involves a complex interaction among regions in larger networks. Many of the unresolved issues and contrastive findings in the literature warrant further research with larger groups of participants and standard protocols of fMRI implementation.

Multimodal Neural and Behavioral Data Predict Response to Rehabilitation in Chronic Poststroke Aphasia

BACKGROUND

Poststroke recovery depends on multiple factors and varies greatly across individuals. Using machine learning models, this study investigated the independent and complementary prognostic role of different patient-related factors in predicting response to language rehabilitation after a stroke.

METHODS

Fifty-five individuals with chronic poststroke aphasia underwent a battery of standardized assessments and structural and functional magnetic resonance imaging scans, and received 12 weeks of language treatment. Support vector machine and random forest models were constructed to predict responsiveness to treatment using pretreatment behavioral, demographic, and structural and functional neuroimaging data.

RESULTS

The best prediction performance was achieved by a support vector machine model trained on aphasia severity, demographics, measures of anatomic integrity and resting-state functional connectivity (F1=0.94). This model resulted in a significantly superior prediction performance compared with support vector machine models trained on all feature sets (F1=0.82, P<0.001) or a single feature set (F1 range=0.68-0.84, P<0.001). Across random forest models, training on resting-state functional magnetic resonance imaging connectivity data yielded the best F1 score (F1=0.87).

CONCLUSIONS

While behavioral, multimodal neuroimaging data and demographic information carry complementary information in predicting response to rehabilitation in chronic poststroke aphasia, functional connectivity of the brain at rest after stroke is a particularly important predictor of responsiveness to treatment, both alone and combined with other patient-related factors.

The Effectiveness of Transcranial Magnetic Stimulation (TMS) Paradigms as Treatment Options for Recovery of Language Deficits in Chronic Poststroke Aphasia

Background

In an effort to boost aphasia recovery, modern rehabilitation, in addition to speech and language therapy (SALT), is increasingly incorporating noninvasive methods of brain stimulation. The present study is aimed at investigating the effectiveness of two paradigms of neuronavigated repetitive transcranial magnetic stimulation (rTMS): (i) 1 Hz rTMS and (ii) continuous theta burst stimulation (cTBS) each as a standalone treatment for chronic aphasia poststroke.

Methods

A single subject experimental design (SSED) trial was carried out in which six people with aphasia (PWA) were recruited, following a single left hemispheric stroke more than six months prior to the study. Three individuals were treated with 1 Hz rTMS, and the remaining three were treated with cTBS. In all cases, TMS was applied over the right pars triangularis (pTr). Language assessment, with standardized and functional measures, and cognitive evaluations were carried out at four time points: twice prior to treatment (baseline), one day immediately posttreatment, and at follow-up two months after treatment was terminated. Quality of life (QoL) was also assessed at baseline and two months posttreatment. In addition, one of the participants with severe global aphasia was followed up again one and two years posttherapy.

Results

For all participants, both rTMS paradigms (1 Hz rTMS and cTBS) generated trends towards improvement in several language skills (i.e., verbal receptive language, expressive language, and naming and reading) one day after treatment and/or two months after therapy. Rated QoL remained stable in three individuals, but for the other three, the communication scores of the QoL were reduced, while two of them also showed a decline in the psychological scores. The participant that was treated with cTBS and followed for up to two years showed that the significant improvement she had initially exhibited in comprehension and reading skills two months after TMS (1^st follow-up) was sustained for at least up to two years.

Conclusion

From the current findings, it is suggested that inhibitory TMS over the right pTr has the potential to drive neuroplastic changes as a standalone treatment that facilitates language recovery in poststroke aphasia.

Machine Learning in Action: Stroke Diagnosis and Outcome Prediction

The application of machine learning has rapidly evolved in medicine over the past decade. In stroke, commercially available machine learning algorithms have already been incorporated into clinical application for rapid diagnosis. The creation and advancement of deep learning techniques have greatly improved clinical utilization of machine learning tools and new algorithms continue to emerge with improved accuracy in stroke diagnosis and outcome prediction. Although imaging-based feature recognition and segmentation have significantly facilitated rapid stroke diagnosis and triaging, stroke prognostication is dependent on a multitude of patient specific as well as clinical factors and hence accurate outcome prediction remains challenging. Despite its vital role in stroke diagnosis and prognostication, it is important to recognize that machine learning output is only as good as the input data and the appropriateness of algorithm applied to any specific data set. Additionally, many studies on machine learning tend to be limited by small sample size and hence concerted efforts to collate data could improve evaluation of future machine learning tools in stroke. In the present state, machine learning technology serves as a helpful and efficient tool for rapid clinical decision making while oversight from clinical experts is still required to address specific aspects not accounted for in an automated algorithm. This article provides an overview of machine learning technology and a tabulated review of pertinent machine learning studies related to stroke diagnosis and outcome prediction.

Natural Language Processing Enhances Prediction of Functional Outcome After Acute Ischemic Stroke

Background

Conventional prognostic scores usually require predefined clinical variables to predict outcome. The advancement of natural language processing has made it feasible to derive meaning from unstructured data. We aimed to test whether using unstructured text in electronic health records can improve the prediction of functional outcome after acute ischemic stroke.

Methods and Results

Patients hospitalized for acute ischemic stroke were identified from 2 hospital stroke registries (3847 and 2668 patients, respectively). Prediction models developed using the first cohort were externally validated using the second cohort, and vice versa. Free text in the history of present illness and computed tomography reports was used to build machine learning models using natural language processing to predict poor functional outcome at 90 days poststroke. Four conventional prognostic models were used as baseline models. The area under the receiver operating characteristic curves of the model using history of present illness in the internal and external validation sets were 0.820 and 0.792, respectively, which were comparable to the National Institutes of Health Stroke Scale score (0.811 and 0.807). The model using computed tomography reports achieved area under the receiver operating characteristic curves of 0.758 and 0.658. Adding information from clinical text significantly improved the predictive performance of each baseline model in terms of area under the receiver operating characteristic curves, net reclassification improvement, and integrated discrimination improvement indices (all P<0.001). Swapping the study cohorts led to similar results.

Conclusions

By using natural language processing, unstructured text in electronic health records can provide an alternative tool for stroke prognostication, and even enhance the performance of existing prognostic scores.

The strange role of brain lesion size in cognitive neuropsychology

The size of brain lesions is a variable that is frequently considered in cognitive neuropsychology. In particular, lesion-deficit inference studies often control for lesion size, and the association of lesion size with post-stroke cognitive deficits and its predictive value are studied. In the present article, the role of lesion size in cognitive deficits and its computational or design-wise consideration is discussed and questioned. First, I argue that the commonly discussed role or effect of lesion size in cognitive deficits eludes us. A generally valid understanding of the causal relation of lesion size, lesion location, and cognitive deficits is unachievable. Second, founded on the theory of causal inference, I argue that lesion size control is no generally appropriate covariate control. Instead, it is identified as a procedure with only situational benefits, which is supported by empirical data. This theoretical background is used to suggest possible research practices in lesion-deficit inference, post-stroke outcome prediction, and behavioural studies. Last, control for lesion size is put into a bigger historical context - it is identified to relate to a long-known association problem in neuropsychology, which was previously discussed from the perspectives of a mislocalisation in lesion-deficit mapping and the symptom complex approach.

Lesion site and therapy time predict responses to a therapy for anomia after stroke: a prognostic model development study

Stroke is a leading cause of disability, and language impairments (aphasia) after stroke are both common and particularly feared. Most stroke survivors with aphasia exhibit anomia (difficulties with naming common objects), but while many therapeutic interventions for anomia have been proposed, treatment effects are typically much larger in some patients than others. Here, we asked whether that variation might be more systematic, and even predictable, than previously thought. 18 patients, each at least 6 months after left hemisphere stroke, engaged in a computerised treatment for their anomia over a 6-week period. Using only: (a) the patients' initial accuracy when naming (to-be) trained items; (b) the hours of therapy that they devoted to the therapy; and (c) whole-brain lesion location data, derived from structural MRI; we developed Partial Least Squares regression models to predict the patients' improvements on treated items, and tested them in cross-validation. Somewhat surprisingly, the best model included only lesion location data and the hours of therapy undertaken. In cross-validation, this model significantly out-performed the null model, in which the prediction for each patient was simply the mean treatment effect of the group. This model also made promisingly accurate predictions in absolute terms: the correlation between empirical and predicted treatment response was 0.62 (95% CI 0.27, 0.95). Our results indicate that individuals' variation in response to anomia treatment are, at least somewhat, systematic and predictable, from the interaction between where and how much lesion damage they have suffered, and the time they devoted to the therapy.

Factors predicting long-term recovery from post-stroke aphasia

BACKGROUND

It remains widely accepted that spontaneous recovery from aphasia is largely limited to the first related factors. This has direct implications for acute and chronic interventions for aphasia. few months following stroke. A few recent studies challenge this view, revealing that some individuals' language abilities improve even during the chronic stage.

AIMS

To identify prognostic indicators of long-term aphasia recovery.

METHODS & PROCEDURES

Eighteen people with aphasia initially evaluated in the chronic stage were retested at least one year later. The Western Aphasia Battery-Revised (WAB-R) Aphasia Quotient (AQ) was used to quantify changes in language impairment. Prognostic factors included those related to the patient (demographic, psychosocial), stroke (lesion volume and location), and treatment (medical, rehabilitative).

OUTCOMES & RESULTS

Twelve participants improved and 6 remained stable or declined. Linear regression analysis revealed that lesion volume predicted long-term language gains, with smaller lesions yielding greater improvements. Individuals who did not improve were more likely to have lesions encompassing critical frontal and temporoparietal cortical regions and interconnecting white matter pathways. Exploratory regression analysis of psychosocial and treatment-related factors revealed a positive relationship between improvement and satisfaction with life participation, and a negative relationship between improvement and perceived impairment severity. Critically, psychosocial and treatment-related factors significantly improved model fit over lesion volume, suggesting that these factors add predictive value to determining long-term aphasia prognosis.

CONCLUSIONS

Long-term aphasia recovery is multidetermined by a combination of stroke-, psychosocial-, and treatment-related factors. This has direct implications for acute and chronic interventions for aphasia.

Strategies for feature extraction from structural brain imaging in lesion-deficit modelling

High‐dimensional modelling of post‐stroke deficits from structural brain imaging is highly relevant to basic cognitive neuroscience and bears the potential to be translationally used to guide individual rehabilitation measures. One strategy to optimise model performance is well‐informed feature selection and representation. However, different feature representation strategies were so far used, and it is not known what strategy is best for modelling purposes. The present study compared the three common main strategies: voxel‐wise representation, lesion‐anatomical componential feature reduction and region‐wise atlas‐based feature representation. We used multivariate, machine‐learning‐based lesion‐deficit models to predict post‐stroke deficits based on structural lesion data. Support vector regression was tuned by nested cross‐validation techniques and tested on held‐out validation data to estimate model performance. While we consistently found the numerically best models for lower‐dimensional, featurised data and almost always for principal components extracted from lesion maps, our results indicate only minor, non‐significant differences between different feature representation styles. Hence, our findings demonstrate the general suitability of all three commonly applied feature representations in lesion‐deficit modelling. Likewise, model performance between qualitatively different popular brain atlases was not significantly different. Our findings also highlight potential minor benefits in individual fine‐tuning of feature representations and the challenge posed by the high, multifaceted complexity of lesion data, where lesion‐anatomical and functional criteria might suggest opposing solutions to feature reduction.

A Bayesian optimization approach for rapidly mapping residual network function in stroke

Post-stroke cognitive and linguistic impairments are debilitating conditions, with limited therapeutic options. Domain-general brain networks play an important role in stroke recovery and characterizing their residual function with functional MRI has the potential to yield biomarkers capable of guiding patient-specific rehabilitation. However, this is challenging as such detailed characterization requires testing patients on multitudes of cognitive tasks in the scanner, rendering experimental sessions unfeasibly lengthy. Thus, the current status quo in clinical neuroimaging research involves testing patients on a very limited number of tasks, in the hope that it will reveal a useful neuroimaging biomarker for the whole cohort. Given the great heterogeneity among stroke patients and the volume of possible tasks this approach is unsustainable. Advancing task-based functional MRI biomarker discovery requires a paradigm shift in order to be able to swiftly characterize residual network activity in individual patients using a diverse range of cognitive tasks. Here, we overcome this problem by leveraging neuroadaptive Bayesian optimization, an approach combining real-time functional MRI with machine-learning, by intelligently searching across many tasks, this approach rapidly maps out patient-specific profiles of residual domain-general network function. We used this technique in a cross-sectional study with 11 left-hemispheric stroke patients with chronic aphasia (four female, age ± standard deviation: 59 ± 10.9 years) and 14 healthy, age-matched control subjects (eight female, age ± standard deviation: 55.6 ± 6.8 years). To assess intra-subject reliability of the functional profiles obtained, we conducted two independent runs per subject, for which the algorithm was entirely reinitialized. Our results demonstrate that this technique is both feasible and robust, yielding reliable patient-specific functional profiles. Moreover, we show that group-level results are not representative of patient-specific results. Whereas controls have highly similar profiles, patients show idiosyncratic profiles of network abnormalities that are associated with behavioural performance. In summary, our study highlights the importance of moving beyond traditional 'one-size-fits-all' approaches where patients are treated as one group and single tasks are used. Our approach can be extended to diverse brain networks and combined with brain stimulation or other therapeutics, thereby opening new avenues for precision medicine targeting a diverse range of neurological and psychiatric conditions.

Individualized response to semantic versus phonological aphasia therapies in stroke

Attempts to personalize aphasia treatment to the extent where it is possible to reliably predict individual response to a particular treatment have yielded inconclusive results. The current study aimed to (i) compare the effects of phonologically versus semantically focussed naming treatment and (ii) examine biographical and neuropsychological baseline factors predictive of response to each treatment. One hundred and four individuals with chronic post-stroke aphasia underwent 3 weeks of phonologically focussed treatment and 3 weeks of semantically focussed treatment in an unblinded cross-over design. A linear mixed-effects model was used to compare the effects of treatment type on proportional change in correct naming across groups. Correlational analysis and stepwise regression models were used to examine biographical and neuropsychological predictors of response to phonological and semantic treatment across all participants. Last, chi-square tests were used to explore the association between treatment response and phonological and semantic deficit profiles. Semantically focussed treatment was found to be more effective at the group-level, independently of treatment order (P = 0.041). Overall, milder speech and language impairment predicted good response to semantic treatment (r range: 0.256-0.373) across neuropsychological tasks. The Western Aphasia Battery-Revised Spontaneous Speech score emerged as the strongest predictor of semantic treatment response (R 2 = 0.188). Severity of stroke symptoms emerged as the strongest predictor of phonological treatment response (R 2 = 0.103). Participants who showed a good response to semantic treatment were more likely to present with fluent speech compared to poor responders (P = 0.005), whereas participants who showed a good response to phonological treatment were more likely to present with apraxia of speech (P = 0.020). These results suggest that semantic treatment may be more beneficial to the improvement of naming performance in aphasia than phonological treatment, at the group-level. In terms of personalized predictors, participants with relatively mild impairments and fluent speech responded better to semantic treatment, while phonological treatment benefitted participants with more severe impairments and apraxia of speech.