Meeting Banner
Abstract #2866

Dotslashen: a Bayesian decision support system for fMRI clinical investigation

Vncio R, Amaro Jr. E, Taub A, Arajo C, Sanches L, Garrido G, Felix M, Pereira C
Hospital Israelita Albert Einstein, BIOINFO-USP Ncleo de Pesquisa em Bioinformtica

Functional magnetic resonance imaging (fMRI) technique is gaining applicability in brain clinical investigation and one of the main challenges is the determination of what should be a normal fMRI exam. We are developing a normative fMRI database based on epidemiological sampling criteria from a population. We developed a Decision Support System to assist clinicians to decide whether a given brain activation, or lack of activation, is expected from a healthy subject or not. Thirty seven normal volunteers subdivided according to gender and age, participated of these study. Each participant performed fMRI tasks involving visual, motor, somatosensory, language and semantic paradigms comprising 16 conditions. The system make interpretable probabilistic statements such as It is very likely that the majority of subjects have these regions activated or It is likely that almost no subject has these regions activated. This is achieved comparing the findings from the individual patient's fMRI exam with the normati