The invention relates to a fully automated screening, diagnosis and classification technique for Alzheimer's disease using magnetic resonance imaging signals from the ventricular zone contour of the brain to get a fundamental index of brain deterioration comprising the steps: —obtaining a gray scale MRI image of the brain region; —applying a contour edge-detecting algorithm to the image; —employing a grid covering method for calculating a first order metric index of ventricular zone contour; —superimposing metric square grids of increasing edge length to the binary contour image and counting the metric grid squares; —plotting the logs of metric grid squares against the logs of edge lengths, wherein the gradient of the plot is the linear topological metric index.