Evidence for a workspace model of working memory from semantic implicit processing in neglect

Three experiments tested the hypothesis that activation of semantic memory from perceptual input does not require initial retention of the perceptual material in working memory as assumed by a widely held view of information processing. In Expt 1, two brain-damaged patients with left-sided unilateral spatial neglect were tested. They were asked to listen to and read a series of familiar (British) and unfamiliar (foreign) proverbs and to choose which proverb was the best match to a depicted figure shown with the target object(s) on the left (neglected side) of the patients' visual field. Expt 2 simulated the testing conditions for the neglect patients with healthy participants using subliminal presentation of one half of each picture. Using different materials, Expt 3 replicated the outcomes of Expts 1 and 2 with a third neglect patient and a new group of controls. In all three experiments, participants appeared to be unaware of target features; however they selected familiar, but not unfamiliar, target proverbs above chance. The results showing implicit processing of semantic material can be explained by a model in which working memory is a separate system that deals with activated contents of semantic memory, and in which there is direct activation of semantic memory from perception without intermediate stages of processing in working memory.