Evaluation of computational methods for estimating the cortical stroke lesion evolution from magnetic resonance images

On brain MR images, stroke lesions coalesce with other manifestations of white matter disease and inflammation producing the same signal change. Dedicated special MR sequences only allow to distinguish 60% of the strokes during a short period of time after they occur. This makes difficult to locate and delineate stroke lesions and consequently leads to erroneous prognosis, incorrect estimation about the extent of the damage caused by the stroke and may lead to false diagnosis. On a previous project we developed an atlas of brain vascular territories to aid on the identification of the stroke lesion boundary, but further research points to a wider and generalised influence. With this project we aim to apply different criteria to identify, on MR images of stroke patients, the extent of the damage directly influenced by the stroke, evaluate their aplicability to clinical research on stroke and suggest modifications and alternatives for further developments.