Exploring the Loss of Semantic Memory in Semantic Dementia: Evidence From a Primed Monitoring Study

An on-line primed monitoring study was used in an exploration of the nature of semantic memory loss in a patient (P.P.) with semantic dementia who showed a profound semantic impairment on a range of off-line tasks. Priming for pairs of words taken from a common category (e.g., cat-dog, spade-rake, ruby-emerald) was contrasted with that for word pairs from different categories that were related functionally (e.g., shampoo-hair, broom-floor, theater-play). Control participants showed robust priming for both types of semantic relation. P.P., in contrast, showed a normal priming effect for the functionally related conditions but no priming for category coordinates. This result suggests that P.P.'s semantic memory loss cannot be explained solely as a loss of stored representations or as a problem with deliberate controlled access to that information. Rather, elements of both explanations apply for different kinds of semantic information.