Is anterior N2 enhancement a reliable electrophysiological index of concealed information?

Abstract

Concealed information tests (CITs) are used to determine whether an individual possesses information about an item of interest. Event-related potential (ERP) measures in CITs have focused almost exclusively on the P3b component, showing that this component is larger when lying about the item of interest (probe) than telling the truth about control items (irrelevants). Recent studies have begun to examine other ERP components, such as the anterior N2, with mixed results. A seminal CIT study found that visual probes elicit a larger anterior N2 than irrelevants (Gamer and Berti, 2010) and suggested that this component indexes cognitive control processes engaged when lying about probes. However, this study did not control for potential intrinsic differences among the stimuli: the same probe and irrelevants were used for all participants, and there was no control condition composed of uninformed participants. Here, first we show that the N2 effect found in the study by Gamer and Berti (2010) was in large part due to stimulus differences, as the effect observed in a concealed information condition was comparable to that found in two matched control conditions without any concealed information (Experiments 1 and 2). Next, we addressed the issue of the generality of the N2 findings by counterbalancing a new set of stimuli across participants and by using a control condition with uninformed participants (Experiment 3). Results show that the probe did not elicit a larger anterior N2 than the irrelevants under these controlled conditions. These findings suggest that caution should be taken in using the N2 as an index of concealed information in CITs. Furthermore, they are a reminder that results of CIT studies (not only with ERPs) performed without stimulus counterbalancing and suitable control conditions may be confounded by differential intrinsic properties of the stimuli employed.

The effect of mental countermeasures on neuroimaging‐based concealed information tests

“…Although this study provides new evidence on the effect of mental countermeasures in fMRI‐based CIT paradigms, some limitations should be mentioned. First, the paradigm and stimuli were intentionally minimalistic in order to reduce potential perceptual difference between stimuli that can be problematic for the interpretation of CIT results (Ganis et al, ). This approach was justified here because: (a) employing numbers as stimuli could be used in various forensic situations, for example, where a certain sum of money was stolen, where a number was part of the crime scene (e.g., the number of the apartment where a crime took place), or where somebody may be in possession of certain secret information (e.g., a bank account number); in that respect, numbers are not different from other types of crime relevant items and they may even be more generalizable than using specific objects (e.g., a watch), (b) the main goal of the study was to determine whether the accuracy of 3‐Stimulus CIT paradigms is affected by mental countermeasures, regardless of absolute accuracy rates in the field, and (c) even elaborate mock crime scenarios are generally far from ecologically valid situations because participants still know that the entire situation is fictitious, they are just following instructions (and so, the probe is not a crime item, nothing unlawful is done by “stealing” it, for example), and the items employed in the scenarios do not usually have much of an intrinsic value to the participants.…”

“…The stimuli were followed by a black screen with a fixation dot lasting between 1,000 and 9,000 ms (2000 ms on average), according to a pseudo-random sequence (Dale, 1999). These stimuli were used because they were very similar to each other visually and they had already been successfully used in prior work by this group (Ganis, Bridges, Hsu, & Schendan, 2016).…”

“…Although this study provides new evidence on the effect of mental countermeasures in fMRI-based CIT paradigms, some limitations should be mentioned. First, the paradigm and stimuli were intentionally minimalistic in order to reduce potential perceptual difference between stimuli that can be problematic for the interpretation of CIT results (Ganis et al, 2016). This approach was justified here because:…”

P300 amplitude at Pz and N200/N300 latency at F3 differ between participants simulating suspect versus witness roles in a mock crime

“…Also, Figure 3 of van Gaal et al (2008) shows FCZ negative peaks at either 180 versus 290 ms depending upon whether a go trial versus a no-go trial-requiring inhibition-is evoking the potential. In our lab, we have seen such N300 components (that, typically, immediately precede P300, see Ganis et al, 2016) in association with our P300-based CITs (e.g., Hu, Pornpattananangkul, & Rosenfeld, 2012). Based on Falkenstein et al (1999) and related papers cited above, and in conjunction with our pilot observations (see below), we thus chose to monitor and analyze poststimulus negativity at F3 (left frontal, chosen rather than Fz, which we also recorded and which may have served us just as well) as a possible CNS sign of inhibition, as conceptualized in klein Selle et al (2015).…”

“…To select a putative ERP sign of inhibition for the present study, we reviewed the large and complex literature relating frontal (left, central, and right) cortical activity (e.g., N200s, N300s, BOLD responses) to various kinds of inhibitory processes in various paradigms (e.g., Armilio, 2002;Aron, Monsell, Sahakian, & Robbins, 2004;Debruille, Brodeur, & Franco Porras, 2012;Fahrenfort, Scholte, & Lamme, 2008;Falkenstein, Hoormann, & Hohnsbein, 1999, Folstein & Van Petten, 2008Ganis, Bridges, Chun-Wei, & Schendan, 2016;Ganis & Schendan, 2013;Holroyd, 2004;Picton et al, 2007;Swick, Ashley, & Turken, 2008;van Gaal, Ridderinkhof, Fehrenfort, Scholte, & Lamme, 2008;Wang, Lin, Liang, Wang, & Zhang, 2011;for review, see Huster, Enriquez-Geppert, Lavallee, Falkenstein, & Herrmann, 2013). The negative component processes seen in some of these references tend to peak around 300 ms even though they were often called N200 (e.g., Falkenstein et al, 1999).…”