The pre-surgical procedures in patients with antiepileptic drugs (DAEs) refractory epilepsy has as objective the precise location of the epileptogenic zone (ZE), which must be removed to rid the patient of the crises. For this, combine information obtained by various technical of imaging, as the Magnetic Resonance Imaging and the Single Photon Emission Computed Tomography. Co-registration, the name given to the process of the combination of images, presented in recent years increasing evolution and application with the emergence of new accurate and robust techniques. However, the manual method is still taken as the gold standard. The use of similarity voxel measures allow the completely automatic co-registration of intra and inter-modalities medical images through the use of cost functions cost, which may be determined through different mathematics approaches. Mutual Information is the most popular and the most efficient technique in this context and, traditionally, using Shannon's entropy to quantify the information. However, in a previous study, using computational simulations with the application of Mutual Information co-registration technique, finds strong evidence that the use of Tsallis entropy in co-registration may provide important contributions in the location of epileptogenic focus in real situations. Based in these findings emerged the importance of research in detail the possible contribution of this mathematical model in the evaluation of clinical cases. The proposal of the work presented here is to investigate the real relevancy of the use of Tsallis entropy in clinical intra and inter-modalities neuroimagens co-registration procedures through the evaluation of its performance on the location of ZE in retrospective exams of patients with refractory epilepsy. For this, the proposed cost function will be analyzed qualitatively and quantitatively and compared to gold standard and to the others methods already consolidated.
News published in Agência FAPESP Newsletter about the scholarship: