[Electrophysiologic dissociations in memory encoding of abstract and concrete words]

This study examined memory encoding of auditorily presented abstract and concrete nouns. 22 subjects performed various blocks of a free recall memory task in which lists of 22 either abstract or concrete words had to be memorized. Consistent with a large variety of memory studies, recall performance was better for concrete than for abstract words. When the event-related potentials (ERPs) recorded during study were selectively averaged for those words that were subsequently recalled and those subsequently not recalled, the ERPs were more positive going for words that were subsequently recalled. These Dm effects (Difference due to memory) started around 500 ms post-stimulus and differed in timing and scalp topography for both types of words: For abstract words, they were present in an early (i.e., 600 to 1100 ms) time interval at parieto-occipital electrodes only. In contrast, for concrete words, Dm effects were obtained with a broad topographic distribution in the 600 to 1000 ms time range and were also present in a late time interval (1100 to 1600 ms) at fronto-central recording sites. The topographical dissociations of the Dm effects in the early time interval are taken to reflect the larger distinctiveness of concrete words during encoding, whereas the late effects presumably play a functional role in elaborative processing of concrete words. The results do not agree with models of word concreteness that propose separate processing systems for the two types of words, and rather support those models that propose quantitative differences in the processing of abstract and concrete words.