An event-related potential study of two kinds of source judgment errors

Abstract

The neural correlates of source judgment errors for confusable and nonconfusable source–item pairings were investigated. At study, subjects were presented with a series of lists of word pairs. Word pairs belonging to the same list were formed by pairing one of two associated words with different associates of an unstudied theme word (e.g., ship–frost, captain–chilly, ship–freeze, captain–frigid, etc., in one list; wife–pane, husband–glass, wife–curtain, husband–sill, etc., in another list). At test, subjects discriminated ‘Old pairs’ (e.g., ship–frost) from ‘Intra’ rearranged pairs (e.g., ship–chilly), ‘Inter’ rearranged pairs (e.g., ship–pane), and ‘Old–New’ pairs (e.g., ship–bulb). Event-related potentials (ERPs) were recorded for each class of test item. Similar proportions of ‘yes’ responses to Old and Intra rearranged pairs were observed, but the false alarm rate for Inter rearranged pairs was markedly lower. ERPs elicited by correctly classified Old pairs and false alarms to Intra rearranged pairs were indistinguishable and exhibited ‘left parietal’ and ‘right frontal’ effects, characteristic of recollection-based recognition. ERPs for false alarms to Inter rearranged pairs did not show these effects. It is concluded that source judgment errors for confusable source–item pairings are mediated by the same processes that support correct source judgments. Such errors reflect recollection of gist information which is nondiagnostic with respect to source.