Abstract We present a patient, who, following a right posterior ischemia, showed a selective deficit in visually recognising pictures, objects and faces. She was able to read and comprehend any kind of written material and could recognise letters and numbers. Her inability to recognise pictures did not arise from a deficit at the structural description level and/or from a poor semantic knowledge of the stimuli. We argue that her recognition deficit arose from an inability in combining the different elements of the visual stimuli in an unitary percept. Results are discussed in terms of dissociations between local versus global processing, as well as bottom-up versus top-down mechanisms.