Does memory constrain utilization of top-down information in spoken word recognition? Evidence from normal aging.

Young and elderly adults heard recorded words that had been computer-edited from connected speech so as to be heard in isolation from their linguistic surround. Word identification was tested for words in isolation and when heard with increasing amounts of linguistic context that had either preceded or followed them in their original utterances. Although the elderly subjects were poorer in identifying the words in isolation compared to young adults, both age groups showed similar increases in correct word identification as increasing amounts of prior context were presented. By contrast, context that followed the target words was less effective for the elderly subjects than it was for the young. It is argued that a memory trace of the unclear stimulus must be maintained for effective utilization of following context in a retrospective analysis. The elderly subjects' relative inability to utilize following context implicates an age-related memory deficit operating at the sentence level.