Elsevier

NeuroImage

Volume 55, Issue 1, 1 March 2011, Pages 312-319
NeuroImage

Lying in the scanner: Covert countermeasures disrupt deception detection by functional magnetic resonance imaging

https://doi.org/10.1016/j.neuroimage.2010.11.025Get rights and content

Abstract

Functional magnetic resonance imaging (fMRI) studies have documented differences between deceptive and honest responses. Capitalizing on this research, companies marketing fMRI-based lie detection services have been founded, generating methodological and ethical concerns in scientific and legal communities. Critically, no fMRI study has examined directly the effect of countermeasures, methods used by prevaricators to defeat deception detection procedures. An fMRI study was conducted to fill this research gap using a concealed information paradigm in which participants were trained to use countermeasures. Robust group fMRI differences between deceptive and honest responses were found without, but not with countermeasures. Furthermore, in single participants, deception detection accuracy was 100% without countermeasures, using activation in ventrolateral and medial prefrontal cortices, but fell to 33% with countermeasures. These findings show that fMRI-based deception detection measures can be vulnerable to countermeasures, calling for caution before applying these methods to real-world situations.

Research Highlights

► Activation in prefrontal cortex distinguishes lies from truth in single subjects. ► Countermeasures can disrupt single subject FMRI-based deception detection. ► Caution needs to be used in applying these methods in real-world settings.

Introduction

Deception is a pervasive behavior that can serve useful social purposes (DePaulo et al., 1996) but can also have enormous negative consequences, which is why societies have long sought reliable methods for determining when people lie (Vrij, 2008). Methods have included observing behavioral and peripheral physiology (Vrij, 2008). To improve upon these methods (National Research Council, 2003), researchers recently began monitoring brain activity with event-related brain potentials (ERPs) and, lately, with functional magnetic resonance imaging (fMRI). FMRI laboratory studies have shown that deceptive and honest responses can be differentiated in group data and intraindividual analyses have revealed deception detection accuracies around 90% (e.g., Abe et al., 2008, Bhatt et al., 2008, Davatzikos et al., 2005, Gamer et al., 2007, Gamer et al., 2009, Ganis et al., 2003, Ganis et al., 2009, Kozel et al., 2004, Kozel et al., 2005, Langleben et al., 2002, Langleben et al., 2005, Lee et al., 2005, Lee et al., 2009, Mohamed et al., 2006, Monteleone et al., 2008, Nose et al., 2009, Nunez et al., 2005, Spence et al., 2001, Spence et al., 2008). Capitalizing on this research, companies have begun marketing fMRI-based “lie detection” services, capturing the imagination of the popular media, but generating methodological and ethical concerns in scientific and legal communities (Greely and Illes, 2007, Nature_Neuroscience_Editorial, 2008). One such concern is that the accuracy of current fMRI-based methods for real world applications may be overestimated by the public because neuroimaging data are typically perceived by non-experts as being more compelling than other types of data (Weisberg et al., 2008). A virtually unexplored aspect of this concern is whether countermeasures, methods prevaricators employ to confuse deception detection procedures, could defeat fMRI deception tests. This issue is critical because countermeasures are known to degrade the accuracy of deception detection using peripheral physiological and ERP measurements (Honts et al., 1996, Rosenfeld et al., 2004).

To address this issue, we conducted an fMRI study using a modified concealed information test (CIT, also referred to as “guilty knowledge test”) in which participants were trained to use a covert countermeasure while lying about knowing their birth date. Methodologically, CIT paradigms are the gold standard in laboratory research to determine if a person is lying about possessing knowledge of an item of interest (or “probe”) (Ben-Shakhar and Elaad, 2003). CIT paradigms rely on the finding that a salient stimulus, such as an infrequent and meaningful item presented within a series of nonsalient items, produces an orienting response that directs attention to potentially important changes in the environment (Lykken, 1974). This oddball response is greater if it is the only one associated with deception. By using appropriate nonsalient comparison items (or “irrelevants”), this response can be used to infer that a person possesses knowledge about a probe but deceptively reports no such knowledge. Individuals with no knowledge about the probe, and who truthfully claim so, will show a much smaller response.

The CIT protocol employed here has been called the “3-stimulus” protocol in the ERP literature (Winograd and Rosenfeld, 2010) since it contains on any trial either a probe, irrelevant or an attention holding “target” (an irrelevant item to which participants are assigned a unique response, as articulated in more detail in the Design and procedure section). This protocol has been used both in fMRI (e.g., Nose et al., 2009, Gamer et al., 2007) and ERP (e.g., Rosenfeld et al., 2004) work. FMRI studies using variants of this protocol have reported stronger activation to probes than irrelevants in regions including the lateral and medial prefrontal cortex (e.g., Langleben et al., 2002, Nose et al., 2009Phan et al., 2005, Gamer et al., 2007). Such activations have been attributed to memory-related and executive control processes (Christ et al., 2009) that are likely to be engaged more strongly by probes (especially when they require a deceptive response) than by irrelevants. Critically, ERP studies using the 3-stimulus protocol and focusing on the P300 potential have shown that this protocol is vulnerable to countermeasures in which participants covertly assign meaning to the nonsalient comparison stimuli in order to reduce the relative salience of the probe (Rosenfeld et al., 2004, also replicated in recent work of ours to be published elsewhere). The key question examined here is whether these same countermeasures can also decrease the accuracy of a 3-stimulus CIT paradigm using fMRI measures of brain activation. This is an important question because fMRI responses to probe items may not index the same brain activity underlying the P300 to these same items, and so it is possible that an fMRI-based CIT protocol using the 3-stimulus protocol might not be susceptible to the countermeasures used in the ERP studies.

Note that Rosenfeld et al. (2008) recently developed a new ERP-based “Complex Trial” protocol that is more resistant to countermeasures and that might have been used here for fMRI. However, this protocol cannot be easily adapted to fMRI, given the slow hemodynamic nature of the signals it measures, because it requires the presentation of stimuli in rapid succession during each trial. Therefore, we chose to start with the simpler 3-stimulus protocol here so as to answer the basic empirical question of its vulnerability to countermeasures using fMRI signals.

Section snippets

Subjects

Twenty-six Harvard University undergraduates (14 females; mean age = 20.1 years) participated. All gave written informed consent following protocols approved by the Massachusetts General Hospital and Harvard University Institutional Review Boards. Twelve participants were employed in the main study (main group). Fourteen others were included in a second group (ROI group) to obtain independent ROIs of which two did not complete the study, due to technical problems, and their data were not used. All

Behavioral results

To confirm that participants performed the tasks as instructed, first we used one-sample t-tests on error data for all item types and conditions. This analysis showed that accuracy was well above the 50% chance level, all ts(11) > 11.5, all ps < 0.0001, with the lowest accuracy at 86.5% for targets in the concealed knowledge condition. Next, planned t-tests on the RT and error data showed slower responses for probes (M = 748 ms, SE = 39 ms) than irrelevants (M = 688 ms, SE = 38 ms) in the concealed knowledge

Discussion

This study shows that hemodynamic signals from lateral and medial prefrontal cortices could differentiate deceptive and honest responses but that such differential activation becomes much smaller when participants use a simple covert countermeasure. Critically, single subject classification accuracy, required for any deception test, is substantially reduced by the covert countermeasure, even in this controlled laboratory situation and with highly salient personal information. These effects are

Competing interests statement

The authors declare that they have no competing financial interests.

Acknowledgment

This research was supported in part by the National Science Foundation (BCS0322611).

References (47)

  • J.P. Rosenfeld et al.

    P300-based detection of concealed autobiographical versus incidentally acquired information in target and non-target paradigms

    Int. J. Psychophysiol.

    (2006)
  • J.P. Rosenfeld et al.

    Single versus multiple probe blocks of P300-based concealed information tests for self-referring versus incidentally obtained information

    Biol. Psychol.

    (2007)
  • S.A. Spence et al.

    Speaking of secrets and lies: the contribution of ventrolateral prefrontal cortex to vocal deception

    Neuroimage

    (2008)
  • N. Abe et al.

    Neural correlates of true memory, false memory, and deception

    Cereb. Cortex

    (2008)
  • H. Alkadhi et al.

    Reproducibility of primary motor cortex somatotopy under controlled conditions

    Am. J. Neuroradiol.

    (2002)
  • G. Ben-Shakhar et al.

    The validity of psychophysiological detection of information with the Guilty Knowledge Test: a meta-analytic review

    J. Appl. Psychol.

    (2003)
  • S. Bhatt et al.

    Lying about facial recognition: an fMRI study

    Brain Cogn.

    (2008)
  • S.E. Christ et al.

    The contributions of prefrontal cortex and executive control to deception: evidence from activation likelihood estimate meta-analyses

    Cereb. Cortex

    (2009)
  • A. Dale

    Optimal experimental design for event-related fMRI

    Hum. Brain Mapp.

    (1999)
  • B.M. DePaulo et al.

    Lying in everyday life

    J. Pers. Soc. Psychol.

    (1996)
  • J. Downar et al.

    A cortical network sensitive to stimulus salience in a neutral behavioral context across multiple sensory modalities

    J. Neurophysiol.

    (2002)
  • B. Efron

    The Jackknife, the Bootstrap, and Other Resampling Plans

    (1982)
  • M. Gamer et al.

    Covariations among fMRI, skin conductance, and behavioral data during processing of concealed information

    Hum. Brain Mapp.

    (2007)
  • Cited by (125)

    • How Vulnerable is the Reaction Time Concealed Information Test to Faking?

      2021, Journal of Applied Research in Memory and Cognition
    • Forensic psychiatry and neurolaw: Description, developments, and debates

      2019, International Journal of Law and Psychiatry
    View all citing articles on Scopus
    View full text