For the functional data, slice timing correction was first performed to account for the interleaved acquisition, while geometrical displacements due to head movement were corrected with rigid registration using the AFNI software (https://afni.nimh.nih.gov/). Brain extraction of the fMRI data was performed using FSL's Brain Extraction Tool (BET) \cite{Smith_2002}. The 150 functional images for each patient were, then, affinely registered to the corresponding T1 image, transformed to the MNI template, and underwent mean intensity normalization. Finally, nuisance signal regression was performed for white matter, CSF and global mean, and the functional time series were band-pass filtered (0.01-0.1Hz) and scrubbed for extreme frame displacement (\(>3mm\)). The structural and functional preprocessing steps are summarized in Fig.\ref{113850}.

Ischemic lesion outlines and rich club nodes characterisation

A recent study on structural connectivity subdivided the brain into 68 cortical and 14 subcortical regions \cite{Desikan_2006} and identified those belonging to the rich club \cite{van_den_Heuvel_2011}. These nodes are characterized by high connection strength, high betweenness centrality and low path length. Betweenness centrality is an indicator of a node's centrality within a network with respect to its influence on the transfer of information, while path length is an indicator of efficient information transmission. These regions are not only individually central, but also more densely interconnected than expected by chance, a fact that renders them critical for normal brain function. This set comprises 6 bilateral regions, including superior frontal and parietal cortex and the precuneus, along with subcortical regions including putamen, hippocampus and thalamus. These regions are illustrated in Fig.\ref{378227}. An expert neurologist (M.R.E.) manually outlined the acute infarct lesions on the DWI image and identified the number of rich club nodes affected by the lesion.