The recognition potential reflects an intermediate level of visual representation

Abstract

A well-known event-related potential (ERP) component—the recognition potential (RP) is a response to recognizable stimuli, peaking at around 200–250 ms. The aim of the present study is to put to the test the contrasting perceptual and semantic accounts of the RP, with manipulation of word meaning and presentation format of stimuli. The hypothesis was that RP latencies would be observed to increase with the difficulty of identification in presentation formats in accord with both accounts and the present study would replicate the semantic effect independently of the perceptual manipulations. The stimuli were Chinese characters and presented to native speakers. The data yielded a negative component (RP) at around 250 ms for all formats of real characters and pseudocharacters at parietooccipital electrodes. The RP for real characters was larger in normal formats than in mirror formats and the RP for pseudocharacters was greater in normal-upright format than other formats. However, a significant RP latency difference was observed between real characters and pseudocharacters. It seems that our data could not be taken as evidence for semantic processing account. We suggest that the RP is a reflection of intermediate visual representation which integrates information from various pathways and then achieve word form analysis. When word form analysis becomes difficult, the categorical feature of word form contributes to word recognition.