English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Machine learning offers great potential for automated prediction ofpost-stroke symptoms and their response to rehabilitation. Major challenges forthis endeavour include the very high dimensionality of neuroimaging data, therelatively small size of the datasets available for learning, and how toeffectively combine neuroimaging and tabular data (e.g. demographic informationand clinical characteristics). This paper evaluates several solutions based ontwo strategies. The first is to use 2D images that summarise MRI scans. Thesecond is to select key features that improve classification accuracy.Additionally, we introduce the novel approach of training a convolutionalneural network (CNN) on images that combine regions-of-interest extracted fromMRIs, with symbolic representations of tabular data. We evaluate a series ofCNN architectures (both 2D and a 3D) that are trained on differentrepresentations of MRI and tabular data, to predict whether a composite measureof post-stroke spoken picture description ability is in the aphasic ornon-aphasic range. MRI and tabular data were acquired from 758 English speakingstroke survivors who participated in the PLORAS study. The classificationaccuracy for a baseline logistic regression was 0.678 for lesion size alone,rising to 0.757 and 0.813 when initial symptom severity and recovery time weresuccessively added. The highest classification accuracy 0.854 was observed when8 regions-of-interest was extracted from each MRI scan and combined with lesionsize, initial severity and recovery time in a 2D Residual Neural Network.Ourfindings demonstrate how imaging and tabular data can be combined for highpost-stroke classification accuracy, even when the dataset is small in machinelearning terms. We conclude by proposing how the current models could beimproved to achieve even higher levels of accuracy using images from hospitalscanners.

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