Latent disconnectome prediction of long-term cognitive-behavioural symptoms in stroke

Abstract

Stroke significantly impacts the quality of life. However, the long-term cognitive evolution in stroke is poorly predictable at the individual level. There is an urgent need to better predict long-term symptoms based on acute clinical neuroimaging data. Previous works have demonstrated a strong relationship between the location of white matter disconnections and clinical symptoms. However, rendering the entire space of possible disconnection-deficit associations optimally surveyable will allow for a systematic association between brain disconnections and cognitive-behavioural measures at the individual level. Here we present the most comprehensive framework, a composite morphospace of white matter disconnections (disconnectome) to predict neuropsychological scores 1 year after stroke. Linking the latent disconnectome morphospace to neuropsychological outcomes yields biological insights that are available as the first comprehensive atlas of disconnectome-deficit relations across 86 scores—a Neuropsychological White Matter Atlas. Our novel predictive framework, the Disconnectome Symptoms Discoverer, achieved better predictivity performances than six other models, including functional disconnection, lesion topology and volume modelling. Out-of-sample prediction derived from this atlas presented a mean absolute error below 20% and allowed personalize neuropsychological predictions. Prediction on an external cohort achieved an R2 = 0.201 for semantic fluency. In addition, training and testing were replicated on two external cohorts achieving an R2 = 0.18 for visuospatial performance. This framework is available as an interactive web application (http://disconnectomestudio.bcblab.com) to provide the foundations for a new and practical approach to modelling cognition in stroke. We hope our atlas and web application will help to reduce the burden of cognitive deficits on patients, their families and wider society while also helping to tailor future personalized treatment programmes and discover new targets for treatments. We expect our framework’s range of assessments and predictive power to increase even further through future crowdsourcing.

Predicting individual long-term prognosis of spatial neglect based on acute stroke patient data

“…Contrary, Talozzi and colleagues (2023) observed the opposite, as their structural disconnection model outperformed lesion-based models. In contrast to models investigated by Talozzi and colleagues (2023), our best predictive model included the acute behavior as an important predictor, in addition to lesion-based features. However, it is important to note that the exclusion of acute behavior in the DSD model may only partially contribute to the superior performance of our best model, because another model of the current study achieved a superior model performance of R² = 0.55 by using only lesion-based features.…”

“…This performance was achieved for the DSD when predicting individual performances on the Bells cancellation test using a sophisticated model validation approach, including training and validation on different external cohorts. However, as model performance within the training phase did also not exceed 30% explained variance (Talozzi et al, 2023), we can speculate whether structural disconnection data might be less suited for long-term prediction in comparison to lesion anatomy. Contrary, Talozzi and colleagues (2023) observed the opposite, as their structural disconnection model outperformed lesion-based models.…”

“…However, as model performance within the training phase did also not exceed 30% explained variance (Talozzi et al, 2023), we can speculate whether structural disconnection data might be less suited for long-term prediction in comparison to lesion anatomy. Contrary, Talozzi and colleagues (2023) observed the opposite, as their structural disconnection model outperformed lesion-based models. In contrast to models investigated by Talozzi and colleagues (2023), our best predictive model included the acute behavior as an important predictor, in addition to lesion-based features.…”

Preserved Corticospinal Tract Revealed by Acute Perfusion Imaging Relates to Better Outcome After Thrombectomy in Stroke

“…Finally, there is growing evidence supporting the assumption, that the individual disruption of white matter tracts by the subacute and chronic stroke lesion, that is, disconnectivity, highly determines the outcome in the motor 39 or cognitive domain. 40,41 Furthermore, the affection and preservation of white matter pathways following intravenous thrombolysis are associated with the recovery beyond the effect of lesion growth alone 42 ; thus, the measurement of salvaged structural connectivity contains information on treatment effect in addition to lesion volume.…”

Topographic Axes of Wiring Space Converge to Genetic Topography in Shaping Human Cortical Layout

“…The interareal connectivity variation has further revealed the spatially heterogeneous patterns of cortical organization, such as the regional controllability related to cognitive dynamics ( 12–14 ). In contrast, the deficit of interareal connectivity has demonstrated the specificity to the behavioral manifestations of the brain lesion ( 15, 16 ). Although the spatial heterogeneity has been separately found to be embedded in genetic architecture and interareal connectivity variation, it is still largely unknown if there exists a unified organizing principle underlying their indication of the spatial heterogeneity.…”