Abstract

Section 1

Section 2

Section 3

Section 4

Section 5

Discussion
Board

The use of hypnosis to enhance recall increases the likelihood of a false positive response (Dywan & Bowers, 1993; McConkey, 1992; Orne, Whitehouse, Dinges & Orne, 1988; Perry, Laurence, D'Eon, & Tallant, 1988). This phenomenon has been attributed to a lowering of report criterion (Klatzky & Erdelyi, 1985). Alternate explanations have not been pursued and so the results of experimental investigations are usually interpreted within this framework (e.g., Dinges et al., 1992; McConkey, 1992; Orne, Soskis, Dinges, & Orne, 1984; Orne et al., 1988; Whitehouse, Dinges, Orne, & Orne, 1988). However, from the perspective of signal detection theory (Green & Swets, 1966), lowering report criteria should increase output but at reduced levels of confidence. This is not what happens. Increases in confidence are as consistently a part of hypnotic recall as are increases in false positive information. This is evident in case study reports (e.g., Orne, Soskis, Dinges, Orne, & Tonry, 1985) as well as under experimental conditions (e.g., Dinges et al., 1992; Dywan, 1983; McConkey & Kinoshita, 1988; Nogrady, McConkey, & Perry, 1985; Whitehouse, Dinges, Orne, & Orne, 1988; see also Sheehan, 1988, for a review).

A better explanation of the hypnotic memory phenomenon would be that hypnotized individuals do not necessarily lower the criterion they use for memory report, but that more of the items generated during hypnotic recall reach and surpass that criterion (Dywan, 1983; Dywan & Bowers, 1983; Jacoby, Kelley, & Dywan, 1989; Dywan, 1995, 1998). According to this perspective, the highly focused, imagistic involvement of the hypnotized subject during retrieval attempts enhances the vividness and ease with which self-generated items are experienced. This "perceptual fluency" would result in an illusion of familiarity which can be interpreted as true recollection. This distinction has important practical as well as theoretical implications. If it is the experience of remembering itself that is altered by hypnosis, then exhorting individuals to avoid false positive errors by attending to confidence levels would be of little value.

Even in nonhypnotic situations, misattributions about the original source of events have been related to the vividness with which they can be imaged (Johnson, Raye, Wang, & Taylor, 1979) and the apparent ease with which images can be generated (Durso & Johnson, 1979). Subsequent research has consistently demonstrated that perceptual fluency can induce the illusion of familiarity irrespective of the mnemonic status of an event (Johnston, Dark, & Jacoby, 1985; Kelley, Jacoby, & Hollingshead, 1989; Whittlesea, 1993).

The Role of Attention in Mnemonic Decisions

Perceptual fluency is more likely to result in misattributions about the source of remembered events under conditions of reduced attentional capacity. Jacoby and his colleagues have shown that young adults are typically very good at distinguishing between names that are famous and names that are familiar because they have been read in the experimental context. However, when distracted by a dual-task, the ability to make this discrimination breaks down. Under dual-task conditions, young adults mistake the familiarity of the previously read nonfamous names for real-world fame (Jacoby, Woloshyn, & Kelley, 1989). This 'false fame' effect is also observed in populations whose attentional capacity has been reduced either through aging (Dywan & Jacoby, 1990; Dywan, Segalowitz, & Williamson, 1992) or as a result of traumatic brain injury (Dywan, Segalowitz, Henderson, & Jacoby, 1993). In both cases, those with reduced attentional capacity were less able than controls to ignore the perceptual fluency of recently presented nonfamous names and were more likely to attribute this fluency to the names being famous.

Event-related potentials (ERPs) are ideally suited to provide insights into the processes that accompany ongoing cognitive operations because of their ability to reveal momentary changes in the pattern of brain activation (Johnson, 1995; Pribram & McGuinness, 1992). It is also well established that ERPs are sensitive to the previous occurrence of information (e.g., Rugg, 1985) and this is manifest when that portion of the waveform occurring approximately 300 to 800 milliseconds after the onset of a stimulus shows an increase in amplitude for the second and subsequent presentations of that stimulus (Segalowitz, Van Roon, & Dywan, 1997).

Using a strategy similar to that employed in the famous name studies, Dywan, Segalowitz, and Webster (1998) asked participants to carefully read a list of words out loud (study words) because they may be required to recognize the words later. After a short break, participants were presented with a list of test words which were made up of the study words interspersed with a larger number of new words (foils). The experimental manipulation consisted of having some of these foils repeat after a lag of 6 items (lag words). These lag words became very effective lures because they would seem salient to the participant for reasons other than having been in the study list. Participants were directed to press "yes" whenever they saw a study word and "no" for any new words. They were also warned that some of the new words would repeat but to ignore them, pressing the "yes" button only for words from the study list. We hypothesized that inhibiting a response to salient but nontarget items would require full attentional capacity. To test this, we reduced the attentional capacity of one group by testing them under dual-task conditions.

As expected, the dual-task group made significantly more false positive responses to familiar but nontarget information than did those in the full attention condition. The ERP results are are also consistent with expectations as can be seen in Figure 1.

Fig. 1: Grand average ERPs generated by full-attention and dual-task groups to study words, lag words, and foils.

Fz, Cz and Pz refer to the frontal, central and parietal midline electrode sites. As can be seen, the two groups process all categories of stimuli in the same way up to about 300 msec post stimulus onset. However, by about 400 msec (and especially by 600 msec) those in the full attention group showed markedly increased positivity to target stimuli but seemed able to inhibit the late positivity to both the lag and foil items. It appears that through top-down, controlled cognitive processes, these participants were able to attend to one category of familiar information and very quickly abort their response to familiar but nontarget items. The outcome was different for those in the dual-task group. They were as likely as the full-attention group to inhibit their response to foils but were far less successful at inhibiting their response to the familiar but nontarget lag words.

Thus, when attentional capacity was reduced, the ERP response discriminated study items from lag items less and this corresponded with more false positive behavioural responses. We assume that the heightened reactivity to lag words relative to study words in the dual task condition occurred because the lag words were encountered more recently than the study words, thus increasing their perceptual salience. It is important to note, as well, that the ERP wave forms were averaged over correct trials only. We have to assume, therefore, that the high amplitude late positivities associated with the lag words were not the result of having made a false positive response. Rather, they represented a physiological reactivity that influenced response tendencies.

It would seem that participants in the dual task condition were more influenced by bottom-up, perceptually-based aspects of the stimuli. In this case it is the recency of the information, in another it might be the vividness or the emotional salience of the information. In fact, a recent experiment in our lab (McNeely, Dywan, & Segalowitz, 1998) has demonstrated this to be the case for emotional words. Participants were given a list of study words followed by a test of recognition. The test list consisted of words from the study list and foils. In this case, however, we did not manipulate the salience of foils through repetition. We simply included some foils that were likely to elicit a greater emotional response than others. Behaviourally, emotional foils were more likely than neutral foils to be selected as study words. Electrophysiolgically, the emotional foils produced a heightened late positivity relative to the neutral foils. We assume that the neural reactivity elicited by the emotional words was misinterpreted by participants as familiarity thus producing the mnemonic misattributions.

In the neuropsychological literature, source error effects are typically linked to an impairment in frontal lobe function and/or those cognitive processes associated with the frontal lobes (Moscovitch, 1994; Shimamura, 1994). Both electrophysiological and functional imaging studies have generally confirmed that the frontal lobes are specifically involved context-specific recollection. Typically, when stimuli are presented for a second time they elicit a greater amplitude ERP late positivity that is maximal at those electrode sites over temporal-parietal areas of the scalp and larger over the left than the right hemisphere (e.g., Paller, Kutas, & McIsaac, 1992). However, when the recognition task requires context-specific recollection one finds, in addition, an ERP response that is maximal over the frontal electrode sites, with greater activation noted over the right frontal lobe (Wilding & Rugg, 1996; 1997). Similarly, functional neuroimaging has shown that the right dorsolateral prefrontal cortex is selectively activated during tasks requiring episodic retrieval (Buckner & Tulving, 1995). Nyberg, McIntosh, Cabeza, Habib, Houle and Tulving (1996) suggest that this part of the right frontal cortex is involved in maintaining attentional focus on the temporal features of the acquisition episode.

Attention and Recollection Accuracy in Hypnotic Contexts

In the hypnotic context individuals become passive with respect to the retrieval process. Indeed, the standard hypnotic induction exhorts participants to relax and to focus their attention as directed by the hypnotist. Suggestions for improving recollection typically require participants to attend closely to the sensory and perceptual aspects of the to-be-remembered experience. It is also suggested that images be allowed to emerge freely as they attempt to recreate some past event.

Even though hypnotized, individuals are not engaged in a dual task, it is possible that such instruction (whether given in hypnosis or in a waking interrogation) can have the same effect as a dual task in that it would induce the individual to disengage from active attentional control, a process associated with frontal lobe function (e.g., Fuster, 1995; Goldman-Rakic, 1995). While it may take attentional control to immerse oneself in the rich sensory/perceptual experience of an imagistic recollective processes, once in this state, the normal 'gating' of sensory input may not occur. Knight and his colleagues have demonstrated that posteriorly generated middle latency auditory (Knight, Scabini, & Woods, 1989) and somatosensory (Yamaguchi & Knight, 1990) evoked potentials show marked increases in amplitude in individuals with frontal lobe damage. The same heightened reactivity in posterior cortical regions can be seen in cats submitted to temporary cooling of the prefrontal cortex (Skinner & Yingling, 1977). These data are consistent with the view that one aspect of frontally-based attentional control involves selective inhibition of cortical and subcortical structures. Thus, it may be that hypnotic induction techniques, when successful, reduce frontally-mediated gating influences and enhance perceptually-driven aspects of experience.

A more detailed account of how functional shifts in brain activation may relate to hypnotic experience has been presented by Crawford and Gruzelier (1992). Their focus, however, was primarily on right versus left hemisphere activation. Woody and Bowers (1994) have proposed an account of hypnotic experience based on reduced supervisory control which in turn is based on reduced frontal activation. The model described above would be consistent with the Woody and Bowers perspective. The conundrum, however, is that while attentional control appears to be reduced in the hypnotic experience, it may be essential for anyone wishing to embark on that experience in the first place and to maintain the hypnotic focus once one is there (e.g., Crawford, 1994). Thus, achieving some level of hypnotic experience may require active frontally-based attentional control with perhaps a greater role for right frontal systems. However, with respect to the recollective process, once attention is fully focussed on the sensory and perceptual aspects of the to-be-retrieved information, there may be a suspension of the normal gating function of the frontal cortex. This would allow for heightened reactivity in posterior sensory-perceptual regions and leave the individual more susceptible to the illusion of familiarity.

Back to the top.

Back to the top.

Back to the top.

Back to the top.

Back to the top.