When word identification fails: ERP correlates of recognition without identification and of word identification failure
RWI, familiarity, retrieval-failure, N300, preconscious, ERP
Recognition without identification (RWI) refers to people's ability to discriminate studied from unstudied items when the items themselves fail to be identified, as when people fail to identify words from fragments. We sought to identify the ERP correlates of word fragment RWI in an effort to better understand its underlying mechanisms; in so doing, we also examined the ERP correlates of word identification failure vs. success. We found the ERP correlate of the RWI effect to be the N300; greater negativity was shown for unidentified fragments of studied words than for unidentified fragments of unstudied words between 300–325 ms post test fragment onset. We further separated the ERPs according to whether subjects showed the behavioral RWI effect or not; the N300 effect emerged only among those subjects who showed the behavioral effect, suggesting that the N300 is related to the behavioral effect itself. With regard to the ERP correlates of word identification failure vs. success, we found very early indicators of later word identification success vs. failure (starting at 125 ms) that were independent of priming. These early effects may be preconscious markers of downstream word identification success vs. failure. We also found a later persistent negativity associated with successfully identified words that we propose to be associated with executive function and possibly the successful suppression of irrelevant words that might initially come to mind when attempting to complete a unique word fragment; word fragment identification failure may sometimes be due to a failure to suppress irrelevant or incorrect words.
Ryals, Anthony J., Carly A. Yadon, Jason S. Nomi, and Anne M. Cleary. "When word identification fails: ERP correlates of recognition without identification and of word identification failure." Neuropsychologia 49, no. 12 (2011): 3224-3237.
DOI for the article