True But Not False Memories Produce a Sensory ...

True But Not False Memories Produce a Sensory Signature in Human Lateralized Brain Potentials Monica Fabiani, Michael A. Stadler, and Peter M. Wessels University of Missouri

Abstract & False memories (e.g., recognition of events that did not occur) are considered behaviorally and subjectively indistinguishable from true memories. We report that brain activity differs when true and false memories are retrieved. Strongly associated lists of words were presented to one or the other cerebral hemisphere at study. This led to lateralized brain activity for these words during a centrally presented recognition test, reflecting their lateralized encoding. This activity was absent for nonstudied but strongly associated words falsely

INTRODUCTION False memories occur when we believe we remember events that did not actually happen (Roediger, 1996). They are one of many types of illusion that can give us clues about the nature of cognitive processes (Ramachandran, 1998). Information about the genesis of false memories may also provide insights into ‘‘recoveredmemory’’ phenomena, and therefore contribute to the current debate about the validity of such memories (e.g., Shobe & Kihlstrom, 1997). In this article, we demonstrate the existence of differential brain activity elicited by false and true recognition in a laboratory setting in which these memories are typically indistinguishable on the basis of behavioral and subjective indices (Stadler, Roediger, & McDermott, 1999; Duzel, Yonelinas, Mangun, Heinze, & Tulving, 1997; Gallo, Roberts, & Seamon, 1997; Johnson et al., 1997; Payne, Elie, Blackwell, & Neuschatz, 1996; Payne, Neuschatz, Lampinen, & Lynn, 1997; Schacter, 1996a; Roediger & McDermott, 1995). Recent studies have used a simple laboratory technique for producing false memories (Deese, Roediger, & McDermott’s paradigm, or DRM; Duzel et al., 1997; Johnson et al., 1997; Schacter, Verfaellie, & Pradere, 1996b; Roediger & McDermott, 1995; see also Deese, 1959). Subjects are presented with a list of words all associatively related to a nonpresented critical target. For example, the words DOOR, GLASS, PANE, SHADE, LEDGE, SILL, HOUSE, OPEN, CURTAIN, FRAME, VIEW, BREEZE, SASH, SCREEN, SHUTTER are all related to © 2000 Massachusetts Institute of Technology

recognized as studied items. These results indicate that studied words leave sensory signatures of study experiences that are absent for false memories. In addition, hemifield effects emerged, including a slower reaction time (RT) for false recognition of nonstudied words whose associated lists were presented to the left hemifield (i.e., right hemisphere). These false recognition responses were accompanied by frontal slow wave activity, which may reflect a differential ability of the two hemispheres with respect to semantic processing. &

WINDOW. When subjects are asked to recognize words from such a list on a later test, they erroneously recognize the nonpresented false target (WINDOW) as having occurred on the study list. From a behavioral standpoint, it has proven difficult to discriminate between veridical and false memories using this paradigm. In fact, the probability of false recognition of the critical targets is approximately the same as the probability of correctly recognizing words that were actually on the list (Stadler et al., 1999; Payne et al., 1996; Roediger & McDermott, 1995). Subjects are equally confident in making true and false recognition judgments (see Miller & Wolford, 1999 and Roediger & McDermott, 1999 for a discussion) and typically report similar conscious experiences in the two cases (Payne et al., 1997; Roediger & McDermott, 1995; but see Mather, Henkel, & Johnson, 1997 and Norman & Schacter, 1997 for evidence of differences in specific aspects of conscious experience for true vs. false recognition). Even when instructed about the likelihood of making false recognition errors, subjects have difficulty suppressing them (McDermott & Roediger, 1998; Gallo et al., 1997; Roediger & McDermott, 1995; but see Schacter, Israel, & Racine, 1999). Because of the impressive behavioral and subjective similarities between true and false memories, researchers have attempted to find brain indices that would discriminate between them, with the assumption that memory retrieval involves, in part, reactivation of senJournal of Cognitive Neuroscience 12:6, pp. 941–949

sory information that was present during processing of the original event. Since false memories involve events that did not occur, there can be no reactivation of sensory experience, hence, there should be brain activity (i.e., sensory signatures) that differentiates true from false memories in regions of the brain associated with sensory experience. Schacter et al. (1996a) used positron emission tomography (PET) with a blocked design (i.e., with blocks of test trials consisting of words all in the same experimental condition) and reported increased neural activity in the left temporo-parietal region for true memories but not for false memories. However, these findings were not replicated in subsequent studies with randomized word presentation, in which brain activity was recorded by means of functional magnetic resonance imaging (fMRI; Schacter, Buckner, Koutstaal, Dale, & Rosen, 1997) and event-related brain potentials (ERPs; Duzel et al., 1997; Johnson et al., 1997). A problem with these studies is that sensory signatures, like the original sensation, may only be seen in transient brain activity and may therefore not be readily visible in the slower hemodynamic response measures obtained with fMRI and PET. These signatures may also correspond to a relatively small portion of the neural processing of verbal material. Therefore, special controls may be required to make them more apparent, even when measures with higher temporal resolution such as ERPs are used. Using nonverbalizable stimuli, Gratton, Corballis, and Jain (1997) showed that the ERPs elicited by correctly recognized line patterns during test were systematically lateralized according to their side of presentation during study. Although recognition performance was largely dependent on whether the study and test hemisphere matched, subjects did not have a conscious recollection of the side at which the item had been studied. This suggests that the lateralized brain activity was related to the operation of a hemispherically organized sensory memory whose functioning may be independent of the subject’s conscious control (see also Fabiani, Gratton, & Ho, submitted; Gratton, Fabiani, Goodman-Wood, & DeSoto, 1998). This contralateral-control procedure may be particularly useful for demonstrating the presence of sensory signatures in the false memory paradigm. In fact, the lateralization effect visible at retrieval is related to a physical property of the stimulus at encoding (i.e., the hemifield of presentation), which can only exist if the stimulus was actually presented. However, in the Gratton et al. (1997) study, the lateralization effect was elicited using nonverbal stimuli. It was an open question whether a similar phenomenon would also occur in a paradigm in which highly associated words are used. In this experiment, subjects studied words from associated lists (Stadler et al., 1999) presented randomly to 942

Journal of Cognitive Neuroscience

the left or right of fixation (with the constraint that words from the same associative list were all displayed in the same hemifield; see Figure 1A). The study phase was followed by a test phase in which words were presented at the center of the screen, and ERPs elicited by each word were recorded (Figure 1B). During test, subjects were asked to indicate, via a choice button-press, whether each word was ‘‘old’’ (i.e., studied) or ‘‘new’’ (recognition task). Studied words (i.e., true targets) were randomly intermixed with nonpresented but highly associated words (i.e., false targets) and other nonpresented words from similar but nonstudied lists (control words). We predicted that true target (studied) words would leave sensory signatures on the ERPs elicited at test, which would be lateralized consistently with the side of presentation at study (Gratton et al., 1997). Brain responses to related but nonpresented

...

S tu d y P h as e

A)

table + sit + + nap

C h a ir L is t

S lee p L is t

desk + + snore

...

Tes t P h ase

B)

“NEW”

...

+ door “OLD ”

T RU E TA R G E T C O N T R O L L is t N o t S tu d ie d

+ nap “NEW ”

Correct Response

Condition

T RU E TA R G E T L is t S tu d ie d R ig h t

+ window “N E W ”

FA L S E TA R G E T C O N T R O L L is t N o t S tu d ie d

+ chair

...

FA L S E TA R G E T L is t S tu d ie d Le ft

Figure 1. (A) Schematic representation of the study phase. (B) Schematic representation of the test phase.

Volume 12, Number 6

words (false targets) that are falsely recognized should not be lateralized, thus distinguishing true from false recognition.

RESULTS Reaction Time and Accuracy The behavioral results (Figure 2A and B) showed a robust false memory effect: The percentage of ‘‘old’’ responses correctly given to true targets did not differ from that erroneously given to false targets (t(13) < 1), whereas the rate of old responses erroneously given to control words was significantly lower than those given to false targets (t(13) = 9.67, p < .001). Reaction times (RTs) were longer for control words that were erroneously classified as old (896 msec, SD = 137) than for the other two groups given ‘‘old’’ responses (752 msec, SD = 111 for true targets, t(13) = 5.45, p < .001; and 782 msec, SD = 115 for false targets, t(13) = 4.31, p < .001). The prolonged RTs to false targets with respect to true targets (t(13) = 2.31, p < .05) were due to the RTs to false targets whose associated lists were presented to the left hemifield (and were therefore first processed by the right hemisphere). This finding suggests that there may be a difference in the processing of the false targets depending on the hemisphere to which associated words were first presented. ERPs The main hypothesis of this study concerned the occurrence of encoding-related lateralizations in response to true but not false targets, which was tested using a double-subtraction method (Gratton, 1998; see also Methods).

Figure 2. (A) A false memory effect is evident in the percentage of ‘‘old’’ responses given to true targets (white bar, 72.60%, SD = 7.93), and false targets (hatched bar, 71.19%, SD = 14.05), and is coupled with a low overall false alarm rate to other nonpresented control words (black bar, 20.07%, SD = 12.96). (B) Mean RT for each response type.

As predicted, the ERPs recorded at test showed lateralized brain activity for the true targets (t(13) = 2.43, p < .05 two-tailed, central and posterior electrode sites, 210–700 msec time window) but not for the false targets (t(13) = – 0.21, ns). This activity can be visualized in both the original data (Figure 3A) and in the data derived by means of contralateral-control subtractions (Figures 3B and 4). The interaction between these two effects (i.e., the difference between the lateralizations for true and false targets) was tested with a nonparametric sign test, because of the difference in variability for the two groups of trials, due to the difference in the number of trials included in the averages. This interaction was significant for the early part of the waveform (210–400 msec, p < .05) and only marginally significant when the entire interval (210–700 msec) was considered (p < .10). Note that differential lateralized activity cannot be due to stimulus or response requirements at test, because all stimuli were words presented at fixation, and all were judged to be ‘‘old’’ by the subjects. Therefore, this lateralized activity can be interpreted as an expression of the memory trace formed at study. The absence of this activity for the false targets correctly indicates that their memory traces did not include a sensory signature. Because the paradigm involved repeated study–test blocks, it could be hypothesized that exposure to recognition test trials could alter subsequent encoding, and help the subjects develop strategies that may lead to the critical results. If this were the case, one would expect to see larger lateralization effects for the studied words in the second-half of the experiment than in the first. In fact, there was no significant difference between the first- and second-half of the study (t < 1.2 for both latency ranges). Figure 3A shows that in addition to a difference in lateralization, there were also differences in the amplitude of the ERP response to true and false targets. These differences can be seen more clearly in Figure 5, which shows true targets, false targets, and new trials at all the recording sites. From this figure, it is evident that the positivity elicited by false targets was reduced in amplitude with respect to the activity elicited by true targets (midline electrode sites, F(1,13) = 4.86, p < .05). However, when this effect was examined more closely, it was apparent that the reduced positivity was due to those false targets whose associated lists were first processed by the right hemisphere (F(2,26) = 5.23, p < .05), as can be seen more clearly in Figure 6. This finding possibly suggests that the semantic processing underlying the DRM effect may be carried out predominantly by the left hemisphere (Metcalfe, Funnell, & Gazzaniga, 1995; Phelps & Gazzaniga, 1992), and therefore, may be more readily elicited by words presented in the right visual field. Fabiani, Stadler, and Wessels

943

Figure 3. (A) Grand average ERPs at test are displayed at electrodes T5 (left hemisphere) and T6 (right hemisphere) for true targets (top row) and false targets (bottom row) for lists studied on the right hemifield (thin line, LH = left hemisphere) and left hemifield (thick line, RH = right hemisphere). Amplitude in microvolts is on the ordinate, and time in milliseconds is on the abscissa. Time ‘‘0’’ indicates stimulus onset. The lateralization for true targets is indicated by the fact that the ERP amplitude in response to words studied on the right (left hemisphere) is larger at electrode T5, whereas the opposite is true for words studied on the left (right hemisphere). (B) Grand average of the lateralization waveforms across middle and posterior electrodes. The shaded area indicates the period in which there is a significant lateralization for the true targets (thick line). No lateralization is evident for false targets (thin line).

Results obtained with PET (Schacter et al., 1996a) and fMRI (Schacter et al., 1997) have pointed to the anterior prefrontal cortex as an area involved in true and false

recognition. Data obtained in the present study suggest that there may be a difference in both RT and ERP activity for false targets whose associated lists were first

Figure 4. Grand average field map of the memory-related lateralizations for true targets (left) and false targets (right) obtained with the contralateralcontrol procedure at a latency of 500–600 msec after stimulus. Only one hemisphere is shown because, given the analytic procedure, the two hemispheres will display effects that are identical in magnitude but of opposite sign.

944


Volume 12, Number 6

Figure 5. Grand average of ERPs at test for new, true targets and false targets at all electrode sites used in the study. Amplitude in microvolts is on the ordinate, and time in milliseconds is on the abscissa. Time ‘‘0’’ indicates stimulus onset.

presented to the left hemifield (i.e., right hemisphere). Therefore, we tested whether true and false targets from lists studied in the left or right visual fields differed with respect to the activity they elicited at prefrontal elec-

trode sites (Fp1 and Fp2) at test. These results are shown in Figure 7. In a latency window between 700 and 1,500 msec, the false targets associated with lists presented to the left hemifield (right hemisphere) at

Figure 6. Grand average at the Pz electrode for test ERPs elicited by true targets studied to the left and right of fixation, and by false targets whose associated lists were studied to the left or right of fixation. Amplitude in microvolts is on the ordinate, and time in milliseconds is on the abscissa. Time ‘‘0’’ indicates stimulus onset.

Fabiani, Stadler, and Wessels

945

Figure 7. Grand average at frontal sites (Fp1 and Fp2) for test ERPs elicited by true targets studied to the left and right of fixation, and by false targets whose associated lists were studied to the left or right of fixation. Amplitude in microvolts is on the ordinate, and time in milliseconds is on the abscissa. Time ‘‘0’’ indicates stimulus onset.

study differed significantly from the other true and false targets (F(1,13) = 6.54, p < .05).

DISCUSSION This study demonstrates that with contralateral-control procedures, it is possible to isolate brain activity representing sensory signatures associated with memory for verbal material. We cannot yet localize the area within the brain that generates these signatures, but the involvement of the ventral stream of visual processing is suggested by the position of the electrodes showing the lateralization effects, by source localization algorithms applied to these data, and by converging evidence from optical imaging recordings from the visual cortex using the same contralateral-control procedure in a different paradigm (Fabiani et al., submitted; Gratton et al., 1998). At this time, we also cannot identify the exact nature of the processing that generates these signatures. Perhaps further research employing different orienting tasks that may encourage or discourage the use of semantic or orthographic information could distinguish these possibilities. Recently, research has shown that making studied items more distinctive helps subjects avoid false memory errors (Schacter et al., 1999). Memory traces are assumed to be composed of a number of features, including information about the memory’s source, that is, the context in which it was formed (Johnson, Hashtroudi, & Lindsay, 1993). These features, if consciously accessible, may lead to the use of strategies or heuristics that allow 946


subjects to better distinguish false from true memories (Schacter et al., 1999; Johnson et al., 1997; Mather et al., 1997). The effects of the lateralization manipulation used here may be interpreted in terms of distinctiveness. True memories may leave a sensory signature that makes each memory trace distinctive; false memories presumably lack such a distinctive feature. Of course, in this study, the distinctive sensory feature did not influence subjects’ recognition judgments, probably because this feature was not consciously accessible (Gratton et al., 1997; Mather et al., 1997). Additional effects were observed in this study suggesting that false targets whose associates were first presented to the right hemisphere differed from true targets, and from false targets associated with words studied on the right (left hemisphere). Namely, false recognition of these words was accompanied by longer RT, smaller P300 amplitude at posterior sites (latency 400–700 msec), and larger positive slow wave activity at frontal sites later in the waveform (700–1,500 msec). This latter finding is consistent with results obtained by Wilding and Rugg (1996; see also Trott, Friedman, Ritter, Fabiani, & Snodgrass, 1999) showing sustained prefrontal activity in source-monitoring retrieval tasks. It also converges with ERP results suggesting that frontal positive slow waves may be associated with elaboration (Fabiani, Karis, & Donchin, 1990; Karis, Fabiani, & Donchin, 1984), and with the idea that the left hemisphere may act as an ‘‘interpreter’’ (Phelps & Gazzaniga, 1992). Finally, although it is not possible to draw direct inferences between the location of scalp Volume 12, Number 6

electrodes and the underlying brain structures, taken at face value, these results are also consistent with the involvement of the prefrontal cortex in the DRM task as reported by previous studies (Schacter 1996a; Schacter et al., 1997). These differential effects between true and false targets were not observed in the Johnson et al. (1997) ERP study of false memory. However, in that study, words were encoded and tested at fixation, therefore it is possible that the results reflected predominantly the influence of left-hemisphere processing. In fact, in the present experiment, false targets whose associated lists were presented to the right hemifield (left hemisphere) did not differ from true targets either behaviorally or with respect to the ERP activity they elicited. Presumably, it was our addition of the contralateral-control procedure that accounts for the differences in findings between the two studies. Note that the lateralization effects are visible for true recognition of words encoded at both the left and right of fixation. However, the differences in P300 and slow wave activity mostly distinguish the false targets associated with words processed by the right hemisphere from the other words. Thus, these may represent functionally distinct phenomena, the first of which (lateralization) may represent an actual sensory signature, whereas the second (P300 and slow wave effects) may reflect the interaction between the encoding manipulation and the hemispheric specialization for semantic processing. In conclusion, we have demonstrated differences in brain activity at the time of retrieval for true and false memories. Studied words elicited a sensory signature, whereas words erroneously identified as part of the studied set did not. Thus, these results suggest that even when subjects cannot consciously discriminate between true and false memories, their brain activity may reveal information relevant to this discrimination. These results were obtained using a contralateralcontrol procedure (Gratton, 1998), which may be ideal for revealing transient differences related to a small portion of the processing of verbal material. They also confirmed previous evidence that visual memory traces are hemispherically organized, even in the case of verbal stimuli (Gratton et al., 1997, 1998), and that the left hemisphere may have a prominent role in the semantic processes that accompany false memory effects.

METHODS Subjects Fourteen right-handed adults (10 females, age range 20– 28), with normal or corrected-to-normal vision, signed informed consent and were paid for their participation in this study. Some additional subjects whose perfor-

mance in the recognition task was at chance were not included in further analyses. Stimuli and Procedures The stimuli were 36, 15-word associative lists (Stadler et al., 1999). Subjects were presented with six study–test blocks. During each study phase, words from four lists were displayed in random order for 200 msec each, 1.58 to either the left or the right of a fixation cross, with a 1,500-msec interstimulus interval (ISI), and with the constraint that words from the same associative list were all presented in the same hemifield. The 200-msec stimulus duration used during the study was chosen to prevent subjects from performing saccadic eye movements toward the laterally presented stimuli. Even if subjects were to move their eyes toward a stimulus, this should only add to the experimental noise and decrease the probability of observing memoryrelated lateralizations, rather than bias the results. Each study list was repeated twice to improve word reading, with the same words presented in the same hemifield on both occasions. Subjects were instructed to fixate on the central cross. During every test phase, 24 words were displayed in random order for 200 msec each, with a 2,000-msec ISI between words. Twelve of the test words were true targets (i.e., words that were actually studied, from positions 1, 8, and 10 of the associative lists) and were randomly intermixed with four false targets (i.e., the nonpresented lures associated to the studied words), six true target controls (words from positions 1, 8, and 10 of the 12 lists that were not studied) and two false target controls (i.e., the critical lures associated with the nonstudied lists). Subjects were asked to indicate, via button presses, whether or not each word was part of the studied set. The response assignments were counterbalanced across subjects. ERP Recording and Analysis ERPs elicited by test words were recorded from the full 10–20 system montage (19 scalp electrodes) by means of an electrode cap (ElectroCap International). The left mastoid was used as reference, and an average mastoid reference was computed off-line. The data were digitized at 100 Hz, and were filtered on-line using a 0.01–30-Hz band-pass. Vertical and horizontal eye movements were recorded and corrected off-line (Gratton, Coles, & Donchin, 1983). Each recording epoch started 100 msec before stimulus presentation and lasted 1,600 msec. ERP data recorded at test were averaged separately for each subject, electrode, and experimental condition. Average data files were digitally filtered with a band-pass of 0–15 Hz before measurement. Mean amplitude measures were derived for each subject, condition, and electrode for two time windows (210–700 msec and 710–1,500 poststimulus). These extended measurement Fabiani, Stadler, and Wessels

947

time windows, encompassing several hundred milliseconds of recording, were chosen to avoid use of multiple comparisons. The choice of the specific interval encompassed by the window used for the lateralization analysis was based on two criteria. The first criterion was the a priori knowledge derived from previous studies: Gratton et al. (1997) found sustained lateralization effects with a peak latency of approximately 450 msec, whereas optical imaging studies (Fabiani et al., submitted; Gratton et al., 1998) showed lateralization effects at earlier latencies in the visual cortex. The second criterion was the component structure of the average waveform in this study, which is typical of recognition paradigms, and includes a P200 with a peak latency of approximately 250 msec and a P300 with a peak latency exceeding 500 msec. In studies examining priming effects versus aspects of the recollective experience (e.g., Paller & Kutas, 1992), priming effects are evident at short latencies whereas recollective processes are often riding on the P300 or other late components (e.g., Fabiani et al., 1986, 1990). This latter distinction was the basis for examining the interaction of the lateralization effects for two separate intervals (200–400 and later). The Contralateral-Control Method The contralateral-control method makes it possible to isolate brain activity that systematically occurs in one or the other hemisphere depending on the experimental condition (in this case, presentation of the word to the left or right hemifield during study; Gratton, 1998). The computation of lateralized waveforms is achieved by means of a two-step procedure. First, the ERPs recorded at homologous electrode locations over the left and right hemispheres are subtracted from each other, with the activity from the electrode ipsilateral to the manipulation (i.e., ipsilateral to the hemifield of word presentation during study) subtracted from the activity recorded at the contralateral electrode. This step eliminates activity that occurs symmetrically in both hemispheres. The second step is to average the lateralization waveforms obtained for the left- and right-hemisphere conditions. This step eliminates the influence of electrical activity that is not lateralized or whose lateralization is independent of the experimental manipulation. Acknowledgments This work was supported by a McDonnell-Pew grant to Monica Fabiani and Michael A. Stadler. We thank Nelson Cowan, Gabriele Gratton, Steve Hackley, and Jonathan King for comments on an earlier version of this manuscript. Reprint requests should be sent to Monica Fabiani, University of Missouri, Department of Psychology, 210 McAlester Hall, Columbia, MO 65211, USA; e-mail: [email protected]. 948


REFERENCES Deese, J. (1959). On the prediction of occurrence of particular verbal intrusions in immediate recall. Journal of Experimental Psychology, 58, 17–22. Duzel, E., Yonelinas, A. P., Mangun, G. R., Heinze, H. J., & Tulving, E. (1997). Event-related brain potential correlates of two states of conscious awareness in memory. Proceedings of the National Academy of Sciences, U.S.A., 94, 5973–5978. Fabiani, M., Gratton, G., & Ho, J. (submitted). Multiple visual memory phenomena in a memory search task. Fabiani, M., Karis, D., & Donchin, E. (1986). P300 and recall in an incidental memory paradigm. Psychophysiology, 23, 298– 308. Fabiani, M., Karis, D., & Donchin, E. (1990). Effects of mnemonic strategy manipulation in a Von Restorff paradigm. Electroencephalography and Clinical Neurophysiology, 75, 22–35. Gallo, D. A., Roberts, M. J., & Seamon, J. G. (1997). Remembering words not presented in lists: Can we avoid creating false memories? Psychonomic Bulletin and Review, 4, 271–276. Gratton, G. (1998). The contralateral organization of visual memory: A theoretical concept and a research tool. Psychophysiology, 35, 638–647. Gratton, G., Coles, M. G., & Donchin, E. (1983). A new method for off-line removal of ocular artifact. Electroencephalography and Clinical Neurophysiology, 55, 468–484. Gratton, G., Corballis, P. M., & Jain, S. (1997). Hemispheric organization of visual memories. Journal of Cognitive Neuroscience, 9, 92–104. Gratton, G., Fabiani, M., Goodman-Wood, M. R., & DeSoto, M. C. (1998). Memory-driven processing in human medial occipital cortex: An event-related optical signal (EROS) study. Psychophysiology, 35, 348–351. Johnson, M. K., Hashtroudi, S., & Lindsay, D. S. (1993). Source monitoring. Psycholological Bulletin, 114, 3–28. Johnson, M. K., Nolde, S. F., Mather, M., Kounios, J., Schacter, D. L., & Curran, T. (1997). The similarity of brain activity associated with true and false recognition memory depends on test format. Psychological Science, 8, 250–257. Karis, D., Fabiani, M., & Donchin, E. (1984). ‘‘P300’’ and memory: Individual differences in the von Restorff effect. Cognitive Psychology, 16, 177–216. Mather, M., Henkel, L. A., & Johnson, M. K. (1997). Evaluating characteristics of false memories: Remember/know judgments and memory characteristics questionnaire compared. Memory and Cognition, 25, 826–837. McDermott, K. B., & Roediger, H. L. III (1998). Attempting to avoid illusory memories: Robust false recognition of associates persists under conditions of explicit warnings and immediate testing. Journal of Memory and Language, 39, 508– 520. Metcalfe, J., Funnell, M., & Gazzaniga, M. S. (1995). Righthemisphere memory superiority: Studies of a split-brain patient. Psychological Science, 6, 157–164. Miller, M. B., & Wolford, G. L. (1999). Theoretical commentary: The role of criterion shift in false memory. Psychological Review, 106, 398–405. Norman, K. A., & Schacter, D. L. (1997). False recognition in younger and older adults: Exploring the characteristics of illusory memories. Memory and Cognition, 25, 838–848. Paller, K. A., & Kutas, M. (1992). Brain potentials during memory retrieval provide neurophysiological support for the distinction between conscious recollection and priming. Journal of Cognitive Neuroscience, 4, 375–391. Payne, D. G., Elie, C. J., Blackwell, J. M., & Neuschatz, J. S. (1996). Memory illusions: Recalling, recognizing, and recollecting events that never occurred. Journal of Memory and Language, 35, 261–285. Volume 12, Number 6

Payne, D. G., Neuschatz, J. S., Lampinen, J. M., & Lynn, S. J. (1997). Compelling memory illusions: The qualitative characteristics of false memories. Current Directions in Psychological Science, 6, 56–60. Phelps, E. A., & Gazzaniga, M. S. (1992). Hemispheric differences in mnemonic processing: The effects of left hemisphere interpretation. Neuropsychologia, 30, 293–297. Ramachandran, V. S. (1998). Consciousness and body image: Lessons from phantom limbs, Capgras syndrome and pain asymbolia. Philosophical Transactions of the Royal Society of London, Series B: Biological Sciences, 353, 1851–1859. Roediger, H. L. III (1996). Memory illusions. Journal of Memory and Language, 35, 76–100. Roediger, H. L. III, & McDermott, K. B. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21, 803–814. Roediger, H. L. III, & McDermott, K. B. (1999). False alarms and false memories. Psychological Review, 106, 406–410. Schacter, D. L. (1996). Illusory memories: A cognitive neuroscience analysis. Proceedings of the National Academy of Sciences, U.S.A., 93, 13527–13533. Schacter, D. L., Buckner, R. L., Koutstaal, W., Dale, A. M., & Rosen, B. R. (1997). Late onset of anterior prefrontal activity during true and false recognition: an event-related fMRI study. Neuroimage, 6, 259–269.

Schacter, D. L., Israel, L., & Racine, C. (1999). Suppressing false recognition in younger and older adults: The distinctiveness heuristic. Journal of Memory and Language, 40, 1–24. Schacter, D. L., Reiman, E., Curran, T., Yun, L. S., Bandy, D., McDermott, K. B., & Roediger, H. L. III (1996a). Neuroanatomical correlates of veridical and illusory recognition memory: Evidence from positron emission tomography. Neuron, 17, 267–274. Schacter, D. L., Verfaellie, M., & Pradere, D. (1996b). The neuropsychology of memory illusions: False recall and recognition in amnesic patients. Journal of Memory and Language, 35, 319–334. Shobe, K. K., & Kihlstrom, J. F. (1997). Is traumatic memory special? Current Directions in Psychological Science, 6, 70– 74. Stadler, M. A., Roediger, H. L. III, & McDermott, K. B. (1999). Norms for word lists that create false memories. Memory and Cognition, 27, 494–500. Trott, C. T., Friedman, D., Ritter, W., Fabiani, M., & Snodgrass, J. G. (1999). Episodic priming and memory for temporal source: Event-related potentials reveal age-related differences in prefrontal functioning. Psychology and Aging, 14, 390–413. Wilding, E. L., & Rugg, M. D. (1996). An event-related potential study of recognition memory with and without retrieval of source. Brain, 119, 889–905.

Fabiani, Stadler, and Wessels

949