Monitoring the Time Course of Perception without Awareness: A Comparison of Mirror Masked Words and Nonwords

Mirror masked words are embedded into a context that makes them appear as senseless patterns or as strings of unfamiliar letters. Thus, mirror masked words can be shown for several hundreds of milliseconds without being recognised as words. We sought to further investigate effects of nonsconscious reading by monitoring event-related brain potentials (ERPs) while participants observed mirror masked letter strings. ERPs were recorded while participants observed mirror masked words and nonwords. Data of 15 participants was segmented into periods of quasi-stable field topography (microstates). Microstates for masked words and nonwords were compared using randomization tests, statistical parametric scalp maps and Low Resolution Electromagnetic Tomography (LORETA). ERPs to masked words and nonwords showed significant topographic differences between 136 and 256 ms, indicating that stimuli were nonconsciously discriminated. A LORETA model localised sources of activation discriminating between masked words and nonwords in left operculum, the right superior parietal lobe and right superior temporal gyrus indicating higher current density for nonwords than for words in these areas. ERPs of mirror masked stimuli can indicate unconscious discrimination even in cases where behavioural priming is unreliable. This approach might be useful for investigating differences in early, nonconscious stages of word perception.