Event-related potential correlates of implicit priming and explicit memory tasks.

The difference between implicit and explicit memory was examined using event-related potentials (ERP). 16 college students read a study list of 100 words (50 high-emotion and 50 low-emotion) from a video monitor. They then performed a perceptual identification (P.I.) task in which they attempted to identify each of a series of 100 words (including 50 from the study list) presented at recognition threshold. Finally, the subjects performed a yes-no recognition task in which they were presented with the other 50 words from the first list and 50 new words (suprathreshold), and they had to identify which ones had been seen in the study list. Subjects were more likely to identify studied than nonstudied words in the perceptual identification task; there was no effect of emotion. In the yes-no recognition task, more high than low-emotion words were responded to correctly for the old words, whereas more low than high-emotion words were responded to correctly for the new words. Data from 11 of the subjects contained enough artifact-free trials for ERP analysis. Analysis revealed that in both the perceptual identification task and the yes-no recognition task, P3 amplitude was larger for old than new words, and emotionality had no effect. Analysis of ERPs in the perceptual identification task averaged according to behavioral response, as well as according to prior study effect showed that P3 amplitude was larger for the studied words than non-studied words, regardless of behavioral response. This effect is electrophysiological evidence that a specific event occurs in the brain in response to the presentation of studied words, whether or not those words are consciously perceived. P3 latency differed as a function of previous study in the yes-no task but not the perceptual identification task.