Distinguishing False from True in Human Memory

studied in Schacter et al.'s well designed and controlled experiment and the type of behaviors often referred to National Institute of Neurological Disorders and Stroke as " recovered memories ". The latter, usually discussed National Institutes of Health in clinical or legal settings, are often descriptions of Bethesda, Maryland 20892 events that are recently remembered but are purported to have occurred years or even decades previously (see The PET study of memory by Dan Schacter and his Schacter, 1996). Not only is there often no independent colleagues in this issue of Neuron (Schacter et al., confirmation of events described in reports of recovered 1996b) has generated much advance publicity and is memories, but there is no way of determining how many eagerly awaited by many readers. It is the first study to times and in what variety of contexts the information examine human brain activity under conditions in which might have been presented and rehearsed in the ensuing subjects falsely " recognize " materials that in fact were retention interval since the original episode. This is very never presented for study during the experiment. The different from the situation in the present study in which experiment compared two critical memory conditions. the experimenters tested recognition at a relatively short In the true target recognition condition, subjects made delay and carefully controlled such variables as study yes/no recognition judgments to words that had been time, rate of presentation, modality of presentation, re-studied earlier. In the false target recognition condition, cency of study, retrieval cues, memory task demands, the test words were never actually presented for study. and even the items themselves across study and re-Rather, subjects had previously heard lists of close se-trieval conditions. mantic associates of these target items (e.g., " taste, Memory for Emotionally Charged Events bitter, candy, chocolate, cake, " etc. were studied asso-As described in detail in their paper, Schacter et al. ciates of the false target " sweet "). Subjects gave compa-examined recognition memory for lists of typical English rable levels of recognition endorsements to words in words and their semantic associates. At no time did the true and false target conditions. The interesting PET they expose subjects to events that had any particular findings were that left hippocampus, which has long emotional content, either positive or negative. This is been thought to be a critical substrate of verbal long-important to remember in light of clinical …