Item 3
In 2021, we published a paper in Neurology titled « Prediction of long-term cognitive function after minor stroke using functional connectivity » (10.1212/wnl.0000000000011452).
In a previous study, also published in Neurology as part of Clement Bournonville’s PhD thesis, we characterized a brain functional network associated with post-stroke cognitive impairment at 6 months (10.1212/WNL.0000000000005553 ). The goal of the second study was to determine whether the brain functional network, measured using MRI at 6 months, can predict long-term cognitive function 36 months after a minor stroke.
Based on the functional network, a machine learning model successfully predicted memory, attention, visuospatial functions, and language functions 36 months post-stroke. Specific patterns were identified for the four cognitive domains, with the left superior frontal cortex playing a role in memory, attention, and visuospatial functions.
Through collaboration with the Institute for Stroke and Dementia Research at LMU Munich University Hospital, we replicated our results on an external dataset.
![[Translate to English:] Figure scintifique](/fileadmin/_processed_/2/0/csm_machine_learn_ing_c6ab58f6c2.jpg)
This multidisciplinary study involved researchers from various fields within our team, including neurologists, psychologist, radiologist, and engineers. Brain connectivity analysis and machine learning are two strong research themes within our team (10.1002/mds.28888 ; 10.1212/wnl.0000000000201596).
