اثرات جنگ پاسخ و نظارت راهبردی بر روی مولفه های مثبت اواخر و فعالیت مغز مربوط به حافظه اپیزودیک: پاسخ فریبنده

The cognitive processes and neural mechanisms underlying deceptive responses were studied using behavioral responses (RT) and event-related brain potentials (ERPs) while participants made truthful and deceptive responses about perceived and remembered stimuli. Memorized words were presented in a recognition paradigm under three instructional conditions: Consistent Truthful, Consistent Deceptive, Random Deceptive. Responses that conflicted with the truth about both perceived and remembered items produced the same pattern of slower RTs and decreased LPC amplitudes. When long-term response patterns were monitored, RTs became much slower and LPC amplitudes decreased greatly. The different behavioral and ERP changes in the two deception conditions suggested that two dissociable executive control processes, each requiring additional processing resources, can contribute to deceptive responses. The parietal episodic memory (EM) effect, thought to reflect recollection, was unaffected by whether participants responded truthfully or deceptively suggesting that it provides a measure of guilty knowledge.

In the past 10 years, investigators have attempted to demonstrate that event-related brain potentials (ERP) can be used to detect concealed information in humans (for reviews, see Bashore and Rapp, 1993 and Rosenfeld, 1995). These efforts can be characterized as falling into two general categories, those aimed at detecting the presence of guilty knowledge and those aimed at detecting when persons are feigning memory loss (i.e., malingering). The guilty knowledge studies have demonstrated that the late positive component (LPC, also known as the P300) of the ERP can provide a useful index of the presence of concealed memories (e.g., Allen and Iacono, 1997, Farwell and Donchin, 1991 and Johnson and Rosenfeld, 1992). Similarly, because persons feigning amnesia are attempting to conceal particular memories, the LPC also reveals the true memory status of items in such persons (e.g., Allen, 2002, Allen et al., 1992, Allen and Movius, 2000, Rosenfeld et al., 1996, Rosenfeld et al., 1998 and Rosenfeld et al., 1999). All these studies focused on the LPC because the stimuli were presented in an “Oddball” paradigm in order to take advantage of the inverse relation between LPC amplitude and stimulus probability. That is, the items of interest (i.e., the “guilty knowledge” items) are presented infrequently and randomly with two other categories of control stimuli, one that is presented infrequently and the other frequently. If the person has guilty knowledge of the items of interest, the infrequent nature of these items will cause them to elicit a LPC like that for the infrequent control stimuli. However, if the person has no knowledge of the items, they will be perceived as belonging to the frequent stimulus set and thus elicit a LPC like that for the frequent control stimuli. Hence, all these studies are best characterized as “applied” because they took previously known aspects of the ERP and used them, albeit creatively, to devise methods to detect the presence of concealed information in “real-world” situations. Despite this interest in detecting guilty knowledge, no “basic” ERP studies have been done to determine the nature of the cognitive processes involved when persons are deceptive. This is surprising given that the ERP technique has been used to investigate the neural basis of many aspects of cognition. To address this lack of knowledge, we conducted a series of experiments designed to identify the cognitive processes used during deceptive responding. Determining which cognitive processes are used during a behavior as complex and multi-faceted as deception is hampered by a number of factors. One initial problem is that there are many types of deception that vary considerably in nature and complexity (Vrij, 2001). This may account for another impediment for designing empirical studies of deception, that there appears to be no widely accepted definition of deception (for reviews, see Mitchell, 1986 and Vrij, 2001). A further complication is the likelihood that the cognitive operations that different people use for any given type of deception will vary as a function of a variety of factors, including their personality and personal habits (e.g., how often they lie), and the circumstances surrounding the deception. In addition, non-cognitive processes, such as those related to processing of any emotional components of a deception, are also likely to be involved. To create a conceptual framework for studying the cognition of deception, we attempted to categorize the general types of processes that might be used by a deceptive person. Although there is, to the best of our knowledge, no definition of deception that specifies the cognitive processes involved, it seemed to us that at least some definitions can be seen as implicitly or explicitly dividing deception operations into two broad categories: (1) the cognitive/emotional processes used to formulate such factors as, for example, the rationale, intent and strategies relevant to a deception, and (2) those used in the act of deception (cf., Furedy et al., 1988). Further, we hypothesized that, rather than being a unique process, both the intent and action components of deception likely draw on general purpose processes used in other aspects of cognition. This view of deception is thus compatible with conclusions, based on studies of autonomic nervous system responses, that there is no unique pattern of autonomic responding associated with deception (i.e., “specific lie response”) (e.g., Ben-Shakhar and Furedy, 1990). Regardless of the nature and extent of the cognitive/emotional processes that precede and accompany a decision to deceive, all deceptions require the execution of a response that is incompatible with the truth. Further, in contrast to the great variability in the nature and amount of processing that precedes a deceptive act, we thought that the executive control processes used in the output stages would be easier to specify and study. The control of response selection and execution processes is a fundamental aspect of one’s ability to successfully interact with the environment in all situations. To do this, executive processes are thought to play an important role in the control of actions by monitoring and resolving response conflicts whenever interference arises from competing response tendencies (e.g., Luu et al., 2000). Hence, we reasoned that the unusual circumstances surrounding deceptive responses would mean that such executive processes are likely play an even more important role than during truthful responding. One reason for this is that the deceptive person must not only perform all the usual task and response-related processes necessary to identify the truthful response, but they must also perform the extra processes required to inhibit the tendency to make this pre-potent response. In addition, falsifying a response means that the deceptive person then must select a response that is incompatible (i.e., conflicts) with the truth and execute this deceptive response. To be deceptive, therefore, a person must make greater use of the executive control processes that affect response inhibition, selection and execution. In addition, they may well rely on these processes more heavily than in most other situations because of the consequences that would follow from being caught in a lie if they do not successfully make the intended deceptive response. Thus, even when these executive processes may comprise a small part of the overall processing used in a deception, they are vital because not performing them correctly carries the risk of voiding all preceding deception-related processing. Note that, in many situations, these control processes may be the primary type of processing performed at the time of the actual deception. That is, it is reasonable to posit that many deceptions are such that a person is likely to do most of the processing related to a pending deception (e.g., work out the details related to ones intent, strategies, etc.) well before the time when one is interrogated and actually has to execute the deceptive responses. In such situations, the need to perform additional deception-related processing in the pre-response category is greatly reduced, leaving the executive control processes as the major remaining cognitive component when the deceptive response is made. The variable nature of the pre-response processing stages across deceptions noted above, coupled with the fact that researchers have already studied the nature of the executive processes used when persons must resolve conflicting response tendencies, led us to focus on the cognitive processes used to execute and monitor deceptive responses. One major unknown in this regard is whether the nature and effects of the conflicting response information that arises in different situations engages the same cognitive processes. For example, it is possible to differentiate deceptive responses on the basis of whether the source of the conflicting response information that must be monitored and overcome is internal or external. That is, one can be deceptive either about perceptual experiences as they occur or about things that one did or said at some time in the past. For example, when one responds falsely “the light was green” when it was actually red, the conflicting response information arises from external stimuli. In contrast, when one falsely states “I didn’t do that” about a past action, the source of the conflicting response information arises from the internal representation of the person’s memory of the event. Hence, in this conceptualization, the processing of conflicting response information and inhibition of truthful responses are necessary but not sufficient components of every deception, regardless of whether the subject of the deception is about perceived or remembered events. Whereas, the effects of conflicting response information have been studied in a variety of situations, all manipulations to date have used perceptual events to convey the conflicting response information. Examples of such perceptually-driven response conflicts are those in which participants must make either compatible or incompatible responses to stimuli such as the words “left” and “right” (Magliero et al., 1984 and McCarthy and Donchin, 1981), color-word stimuli in the Stroop task (Duncan-Johnson and Kopell, 1981), left and right facing arrows (Ragot and Lesevre, 1986 and Doucet and Stelmack, 1999), or left and right spatial locations (Leuthold and Sommer, 1998). These studies demonstrated that, compared to conditions containing no conflicting information, reaction times (RT) increased when the stimuli conveyed conflicting response information. At the same time, however, the concomitant increases in LPC latency were a small fraction of the changes in RT. In contrast, these same studies also showed that this dissociation between the RT and LPC measures of processing time was limited to response conflict situations because the increases in RT and LPC latency were nearly the same for manipulations of stimulus processing time (i.e., task effects). Although LPC amplitude has been quantified less frequently, both Doucet and Stelmack (1999) and Magliero et al. (1984) reported that smaller LPCs were elicited when conflicting response information was introduced into the task. Note that the monitoring processes used to deal with the conflicting response information in all these situations can be characterized as “tactical” because they are limited to those occurring within a trial (i.e., in the interval between the stimulus and the response to that stimulus). Given that all the studies just reviewed used external events to deliver all the conflicting response information, the extent to which the same results would be found for memorially-based response conflicts characteristic of deceptions about one’s past experiences is not known. In other words, it is not known if there is one general purpose processor for conflicting response information or separate mechanisms for different types of response conflicts. No answers are provided by the guilty knowledge studies cited above because they did not isolate and/or quantify conflict-related ERP activity in those studies in which responses were required. Moreover, cognitive experiments generally require participants to make correct responses while keeping errors to a minimum. In contrast, when persons are deceptive, their intent is to make the incorrect response. Thus, the extent to which the processes used to resolve externally and internally based conflicting response information differ as a function of the intent to commit an error is also not known. In addition to the executive control processes used to ensure that each individual response conforms to one’s goals, other processes are presumably involved when it is necessary to monitor how the individual responses fit into a series of responses. Hence, we hypothesized that more complex deceptions, such as those involving multiple responses and/or extended time intervals, probably engage additional control processes. We refer to them as long-term, “strategic” monitoring processes and their function is to ensure that responses are consistent with one another and conform to the person’s longer-term goals. Clearly, the need to engage tactical and strategic monitoring processes during deceptive responding necessitates the use of additional processing resources. Using these extra control processes would, therefore, be equivalent to engaging in a secondary task that the deceptive person must perform in addition to the primary task of responding truthfully. We reasoned, therefore, that deceptive responses may be distinguishable from truthful responses by the fact that they require more controlled processing resources. Previous ERP studies have demonstrated that LPC amplitude can provide a sensitive measure of how processing resources are allocated between two simultaneously performed tasks (for reviews, see Johnson, 1986 and Johnson, 1993). The sensitivity of LPC amplitude to the allocation of processing resources was demonstrated elegantly in a series of studies showing that primary-task LPC amplitudes decreased in a graded manner as attentional resources were systematically diverted away from the primary task toward a secondary task (Isreal et al., 1980a and Isreal et al., 1980b; Kramer et al., 1985 and Wickens et al., 1983). Based on these results, we expected that LPC amplitudes would be reduced for stimuli that elicited deceptive responses compared to those that elicited truthful responses, in proportion to the amount of deception-related processing that was performed.

The present experiment represents a first step at empirically specifying the cognitive processes used when persons are deceptive. To isolate better these processes, deceptive responses were made under restricted conditions involving deceptions that may be described as simulated because both the participants and the experimenter knew the truth and the participants knew that. However, we believe that the fact that the same pattern of results was obtained in both the perceptual and memory tasks, as well as in both deception conditions, supports the idea that the cognitive processes used to inhibit the truthful response and make a conflicting response, basic components of all deceptions, depend on general purpose processors. Although the Random Deceptive condition was intended primarily to create a strategic monitoring task, it also provided a somewhat more realistic simulation because the participants determined, according to their own criteria, whether to respond truthfully or deceptively on any given trial. Despite the seemingly simplistic nature of our deception conditions, they both required the types of response processes that would presumably be used in any deception and it is difficult to see how more realistic scenarios would somehow obviate the need for the executive processes studied here. Whereas the present results provide support for the idea that basic studies of the processes that contribute to deceptions can provide useful information, they represent only a beginning and more realistic scenarios need to be tried. Given that executive processes have been shown to be essential for dealing with conflicting response tendencies in a broad range of circumstances, more realistic deception scenarios are likely to either produce exaggerations of the altered activity described here and/or add additional components (e.g., those related to the much greater salience of deceptive responses in real-world situations). An important addition for future studies would be attempts to introduce an emotional component, a normal aspect of most deceptions. Thus, based on the present results, it is reasonable to expect that studies of more realistic deceptions would reveal additional differences, such as those related to the intent to deceive or the emotional components of deception. Overall, the present results have demonstrated that, despite the complexity of deception, a functional approach can be used advantageously to provide information about the cognitive processes and neural mechanisms involved when persons are deceptive. We identified the neural-behavioral correlates of at least one process that is central to all deceptions, the tactical monitoring processes used to overcome the pre-potent truthful response in order to make a deceptive response. In addition, we found evidence for a second independent, strategic monitoring process that would be invoked in order to keep track of the pattern of one’s past truthful and deceptive responses. Taken together, the results indicate that making deceptive responses about perceived and remembered items engages some apparently general purpose cognitive processes that are used to deal with conflicting response information generated in a variety of circumstances and across a range of task difficulty. The LPC results, combined with our previous MFN results, suggest that, in a real-world situation, even an attempt to respond with a consistent or well-practiced deception that requires no special planning or monitoring may be detectable.