Through the looking-glass: mirror reading.

At early stages of object identification we process correctly oriented and mirrored versions of an object similarly. However, in letter and word perception, such tolerance to mirror reversals is harmful for efficient reading. Do readers successfully develop blindness mechanisms for mirror-letters and words? We conducted two masked priming experiments while recording participants' electrophysiological brain responses to briefly presented primes including mirror-letters (Experiment 1) or to shortly presented mirror-words (Experiment 2). Results showed that the human visual word recognition system is not totally blind to mirror-letters and mirror-words, since the early stages of processing mirror-letters and mirror-words produced effects on target word recognition that were highly similar to the effects produced by identical primes (N250 component). In a posterior stage of processing (N400 epoch), the effect of mirror-letters and mirror-words was different from the effect of identical primes, even though reversed primes still elicited N400 priming effects different from unrelated primes. These results demonstrate that readers perceive mirror-letters and words as correct at initial stages of word recognition, and that the visual word recognition system's neural representation is grounded on basic principles that govern object perception.