Detection of visual change: mismatch or rareness?

How do we detect changes in our visual environment? By continuously comparing visual inputs to templates of experiences in the immediate past? Or by determining their rareness, how infrequently a visual event occurred previously? Recent results from event-related potentials have been interpreted in favour of the first hypothesis, as in the case of the auditory mismatch negativity. Here we demonstrate that rareness, rather than mismatch with a template, underlies visual change detection. Such rareness is detected through a dedicated mechanism in human visual cortex about 100 ms after the rare event occurs, reflected in the rareness-related negativity (RRN).