Source ROCs are (typically) curvilinear: comment on Yonelinas (1999).

On the basis of his assumption that recollection is a threshold process, A. P. Yonelinas (1999) predicted linear source-identification receiver operating characteristics (ROCs) and recently reported data that were consistent with this prediction. In this article, the authors present data showing curvilinear source-identification ROCs across various encoding and test conditions. On the basis of the source-monitoring framework (e.g., M. K. Johnson, S. Hashtroudi, & D. S. Lindsay, 1993), the authors argue that curvilinearity of source-identification ROCs is a result of differences in the qualitative characteristics of memories rather than simply the influence of undifferentiated familiarity as the dual-process model might suggest.

Aging and reflective processes of working memory: binding and test load deficits.

It was hypothesized that age-related deficits in episodic memory for feature combinations (e.g., B. L. Chalfonte & M. K. Johnson, 1996) signal, in part, decrements in the efficacy of reflective component processes (e.g., M. K. Johnson, 1992) that support the short-term maintenance and manipulation of information during encoding (e.g., F. 1. M. Craik. R. G. Morris. & M. L. Gick, 1990; T. A. Salthouse, 1990). Consistent with this, age-related binding deficits in a working memory task were found in 2 experiments. Evidence for an age-related test load deficit was also found: Older adults had greater difficulty than young adults when tested on 2 features rather than 1, even when binding was not required. Thus, disruption of source memory in older adults may involve deficits in both encoding processes (binding deficits) and monitoring processes (difficulty accessing multiple features, evaluating them, or both).

fMRI evidence of age-related hippocampal dysfunction in feature binding in working memory.

Richly detailed memories for particular events depend on processes that bind individual features of experience together. Previous cognitive behavioral research indicates that older adults have more difficulty than young adults in conditions requiring feature binding. We used functional magnetic resonance imaging (fMRI) during a working memory task to identify neural substrates of this age-related deficit in feature binding. For young, but not older, adults there was greater activation in left anterior hippocampus on combination trials (remember objects together with their locations) than on trials in which participants were told to remember only which objects or only which locations occurred. The results provide neuroimaging evidence for an age-related hippocampal dysfunction in feature binding in working memory.

False memories and confabulation.

Memory distortions range from the benign (thinking you mailed a check that you only thought about mailing), to the serious (confusing what you heard after a crime with what you actually saw), to the fantastic (claiming you piloted a spaceship). We review theoretical ideas and empirical evidence about the source monitoring processes underlying both true and false memories. Neuropsychological studies show that certain forms of brain damage (such as combined frontal and medial-temporal lesions) might result in profound source confusions, called confabulations. Neuroimaging techniques provide new evidence regarding more specific links between underlying brain mechanisms and the normal cognitive processes involved in evaluating memories. One hypothesis is that the right prefrontal cortex (PFC) subserves heuristic judgments based on easily assessed qualities (such as familiarity or perceptual detail) and the left PFC (or the right and left PFC together) subserves more systematic judgments requiring more careful analysis of memorial qualities or retrieval and evaluation of additional supporting or disconfirming information. Such heuristic and systematic processes can be disrupted not only by brain damage but also, for example, by hypnosis, social demands and motivational factors, suggesting caution in the methods used by `memory exploring' professions (therapists, police officers, lawyers, etc.) in order to avoid inducing false memories.

The role of prefrontal cortex during tests of episodic memory.

Recent studies of episodic memory using functional neuroimaging techniques indicate that right prefrontal cortex (PFC) is activated while people remember events. Our review suggests that left PFC is also activated during remembering, depending on the reflective demands of the task. As more, or more complex, reflective processes are required (e.g. when criteria for evaluation have to be established and maintained, when the complexity of the evaluation required increases, and when retrieval of additional information is required beyond that activated by an initial cue), left PFC activity is more likely to occur. Our `cortical asymmetry of reflective activity' (CARA) hypothesis summarizes available findings and suggests directions for future research.

Phenomenal characteristics of memories for perceived and imagined autobiographical events.

Two studies explored potential bases for reality monitoring (Johnson & Raye, 1981) of naturally occurring autobiographical events. In Study 1, subjects rated phenomenal characteristics of recent and childhood memories. Compared with imagined events, perceived events were given higher ratings on several characteristics, including perceptual information, contextual information, and supporting memories. This was especially true for recent memories. In Study 2, subjects described how they knew autobiographical events had (or had not) happened. For perceived events, subjects were likely to mention perceptual and contextual details of the memory and to refer to other supporting memories. For imagined events, subjects were likely to engage in reasoning based on prior knowledge. The results are consistent with the idea that reality monitoring draws on differences in qualitative characteristics of memories for perceived and imagined events (Johnson & Raye, 1981) and augment findings from more controlled laboratory studies of complex events (Johnson & Suengas, in press; Suengas & Johnson, 1988).

Is event frequency encoded automatically? The case of alcohol intoxication.

The four experiments reported show that sensitivity to event frequency is diminished by alcohol intoxication. In two experiments on confusion between presented and generated words, the function relating estimated to actual frequency was steeper for sober than intoxicated subjects. In two experiments on word-frequency estimates after a word-pronunciation task, the influence of alcohol intoxication was identified at the input as opposed to the test stage. We conclude that these findings are inconsistent with the idea that frequency-of-occurrence information is automatically encoded.

Dreams and reality monitoring.

Experiment 1 tested the counterintuitive prediction that memories for one's own dreams should not be particularly easy to discriminate from memories for someone else's dreams. Pairs of people reported dreams to each other that they had either dreamed, read, or made up the night before. On a test requiring subjects to discriminate events they had reported from those reported by their partner, subjects had more difficulty with real dreams than with dreams they read or made up. Experiment 2 provided evidence that real dreams do not simply produce overall weaker memories; the deficit for dreams was eliminated with more time to respond and with more detailed cues. In addition, subjects' ratings of various characteristics of their memories (e.g., vividness, personal relevance) indicated that dreams were not generally weaker or impoverished. The results are interpreted within the framework for reality monitoring described by Johnson and Raye (1981): Memories for real dreams are proposed to be deficient in conscious cognitive operations that help identify the origin of information generated in a waking state. At the same time, real dreams are embedded in a network of supporting memories that can be drawn on for reality monitoring decisions under appropriate circumstances. Finally, a comparison of recognition and recall indicated that dreams may leave persisting memories that are difficult to access via free recall.

Age-related changes in confusion between memories for thoughts and memories for speech.

The present experiments compared people's abilities to make decisions about the origin of their memories. Experiment 1 demonstrated that 6-year-olds were as good as 17-year-olds in discriminating memories originating from what they said earlier (self-generations) from memories of what another person said earlier (external presentations). However, in both experiments 1 and 2, 6-year-olds were not as good at discriminating what they had said earlier from what they had only thought. The possibility that younger children simply have more difficulty distinguishing between memories originating from the same class, internal or external, was ruled out because 6-year-olds performed as well as 9-year-olds when differentiating between memories from 2 external sources (experiment 2). Nor could their difficulty be attributed to a general problem in distinguishing memories for their thoughts from any other class of memories because they were at no disadvantage in discriminating their earlier thoughts (words they imagined themselves saying) from words someone else said (experiment 2). Our findings suggest that some distinctions, self versus other, emerge as cues in memory sooner than other distinctions, thoughts versus actions.

Fact and fantasy: the roles of accuracy and variability in confusing imaginations with perceptual experiences.

In Experiment 1, pictures were presented to subjects two, five, or eight times, and subjects were asked to imagine each pciture two, five, or eight times. Subsequently, subjects estimated the number of times each picture had been presented. Their estimates of the frequency of these external events were influenced by imagination trials; this effect was greater for good imagers than for poor imagers. Experiment 2 involved a similar design in which subjects were asked either to imagine the same referent for a word or to imagine a different referent for a word on successive imagination trials. Consistency (same referent) did not increase the influence of imaginations on immediate judgments of external frequency. Thus, the results of Experiment 1 were attributed to the greater accuracy (as opposed to greater consistency) of good imagers' internal generations of the stimuli. Furthermore, variation (imagining different referents), like greater accuracy, increased the effects of imagination trials on immediate but not on delayed judgments of frequency. Possible mechanisms underlying these effects are discussed. In general, the two studies show that qualitative characteristics of completely covert generations influence their impact on estimates of the frequency of external events.

Fact and fantasy: The effects of internally generated events on the apparent frequency of externally generated events.

As John Locke pointed out, there are at least two sources of the contents of the mind: events that occur in the world and events that originate in the mind. Memory, as a record of experience, should contain information from both sources. The present studies investigated memory for the frequency of externally and internally generated events. Individual items were presented to subjects varying numbers of times and tested varying numbers of times. Later, subjects were asked to estimate the frequency of both types of events. Experiment 1 showed that internally generated events influenced the judged frequency of externally generated events and vice versa. The first of these was called the IFE effect and the second the IFI effect. Experiment 2 indicated that the IFE effect was greater when tests consisted of overt, as compared to covert, recall trials. The results were discussed in terms of a model for storing and using occurrence information which would account for both our ability to discriminate between and our tendency to confuse internally and externally generated memory representations.