Can fMRIs Determine if a Defendant is Lying?

Review Paper

Can the Brain Imaging Technique fMRI be used to Identify if a Defendant is Lying?

S. Ludlow ^a

Abstract

Functional Magnetic Resonance Imaging (fMRI), a rapidly developing neuroimaging technology, is frequently a focus of research to determine whether the brain images obtained from fMRI scanning can be used to reliably identify whether an individual is lying. The research provides strong evidence that direct involvement of the prefrontal cortex, including the ventrolateral and dorsolateral prefrontal cortices, produce deceptive responses, alongside aspects of executive control. However, a number of other brain regions, such as the anterior cingulate cortex, have also been revealed as playing a significant role in deception. The employment of countermeasures, the spontaneity of lies vs memorised lies, and the influence of mental disorders on deceit are key factors influencing deceptive responses. These are all yet to be successfully transferred from a laboratory environment and reproduced in a real world, criminal scenario. Thus, in its current state, fMRI brain imaging cannot be used accurately and reliably to conclude if a defendant is lying, and subsequently until further research has been undertaken, it should not be implemented into a legal setting.[S1]

Keywords: Deception, fMRI, Prefrontal Cortex, Neuroscience, Lying

1. Introduction

Deception is a complex psychological process involving one individual deliberately convincing another of something known to be a lie, in order to protect themselves or others (Kozel et al., 2004a; Ito et al., 2012). A lie is successful when the recipient is misled and genuinely believes that what they have been told is the truth. The lying process incorporates communication of a coherent lie, contextual knowledge of the lie, modification of the liar’s behaviour to convince the receiver of the lie and, possibly the most significant process, suppression of the truth (Kozel et al., 2005). Langleben et al., 2002 supports the latter, with results indicating that an essential component of intentional deception may be inhibiting the truth.

Several pioneering studies have explored the use of functional magnetic resonance imaging (fMRI) in defining the neuroscience underlying deception, and whether this brain imaging technique can be used to determine the mechanisms and processes involved during lying and truth telling in individuals. Despite wide-spread research and existing knowledge about the complex brain, recent studies report that the basic neural mechanisms of preparing and creating deception are still unclear and yet to be explored further (Ito et al., 2012; Kireev et al., 2013). This gap in knowledge could be due to external validity and the difficulty in producing an adequate laboratory experiment that reflects real life circumstances (Kireev et al., 2013), and Abe 2009 supports this, claiming that the neurological basis for non-experimental, genuine lies has yet to be established and is an avenue for future research. In addition, fMRI studies are generally not carried out on an idiographic (individual) level, as would be the case for a single defendant, but rather on a nomothetic (group) level, with brain activity averaged and analysed, consequently making it difficult to argue that fMRI lie detection should be used in court.

1.1 The Prefrontal Cortex

Accumulating neuroimaging evidence from multiple laboratory experiments has identified various brain regions activated during deception. Abe 2009 summarised that the prefrontal cortex (PFC) was identified in numerous studies as playing a predominant role during deception (Spence et al., 2001, 2008a; Langleben et al., 2002; Lee et al., 2002; Ganis et al., 2003; Kozel et al., 2004a, 2004b; Luan-Phan et al., 2005). The PFC, located at the front of the brain in the rostral end of the frontal lobe (Bear et al., 2007), is split into four regions: lateral prefrontal cortex (Brodmann area’s 9/45/46/47); frontal pole (BA 9/10); medial frontal cortex, which includes the cingulate cortex (BA 24/25/32); and the orbitofrontal cortex (OFC) (BA 10/11/47) (Gazzaniga et al., 2008). Two sub-regions of the PFC include the dorsolateral prefrontal cortex (DLPFC – BA 8/9/46), implicated in the maintenance of information, and the ventrolateral prefrontal cortex (VLPFC – BA 44/45/47), implicated in the manipulation of information (Christ et al., 2009). The location of these regions can be observed in figure 1.

Fig. 1 – A lateral view of the brain, highlighting regions that make up the prefrontal cortex. The PFC, located at the rostral end of the frontal lobe, the ventrolateral prefrontal cortex (VLPFC) – BA 44/45/lateral 47, and dorsolateral prefrontal cortex (DLFPC) – BA 8/9/46 are frequently seen to be activated during deception.

1.2 Neuroimaging studies of Deception

Researchers have carried out studies to elucidate the various forms of deception in different scenarios. The results of a facial recognition study by Bhatt et al., 2008 disclosed areas of activation in the inferior frontal gyrus, right middle frontal gyrus, superior frontal gyrus (SFG) (including the anterior cingulate cortex (ACC[S2])), red nucleus, supramarginal gyrus, DLPFC, and bilateral precuneus during deception; thus the authors reported no singular brain areas activated during deceptive responses, but rather the prefrontal and parietal lobes mediate activation of a neural network. These two lobes are responsible for cognitive processes such as suppression of the truth and working memory. Another more recent study by Ito et al., 2012, revealed the role which the DLPFC played when distinguishing between preparation and execution of deception. Noticeable activation was seen in the left DLPFC, amongst other regions, during the truth-cue vs uncertain-cue condition and the lie-cue vs uncertain-cue condition for the preparation phase of deception and truth telling. The DLPFC was also clearly activated during the certain/lie and uncertain/lie conditions when compared to the certain/truth and uncertain/truth conditions during the execution phase. The opposite comparison revealed no significant activation during the execution of truth-telling, indicating that the left DLPFC is involved in the preparation of both truths and lies, but only involved in the execution of lying. This provided further evidence to their 2011 research reporting that the DLPFC was associated with deception (Ito et al., 2011). Five fMRI studies revealed that the ACC, parietal cortex and the dorsolateral and ventromedial PFC are active when an individual is lying (see Mameli et al., 2010), and this is supportive of the PFC playing an integral role in deception.

Analysis from a pilot study by Kozel et al., 2004a on eight men who lied or told the truth about the object which money was hidden under revealed significant activation in the OFC and ACC during deception. A year later they conducted another study where participants were instructed to ‘steal’ a watch or ring from a room, and lie or tell the truth about the item which they took under different experimental conditions. This resulted in activation in the majority of subjects in the right ACC, right orbitofrontal, inferior frontal, and middle frontal cortices, which overlapped with their previous study (Kozel et al., 2005). Another group, Langleben et al., 2002, conducted a study using playing cards and a Guilty Knowledge paradigm, where participants had to successfully lie about the card in their pocket in order to keep the reward money. This study established significant activation in the left ACC spread to the right SFG, and the left anterior parietal, motor and premotor cortices during deception. Finally, the bilateral VLPFC, DLPFC, medial PFC (mPFC) and left inferior PFC showed significant activation during an experiment where participants recalled recent activities and then lied or told the truth about the activities according to coloured prompts (Spence et al., 2001). Conversely, this study had transference limitations to a real-life situation as participants did not have free choice to lie. This, and limitations in other studies will be critically assessed in terms of the reliability of study’s results.

Evidently, the PFC and its encompassed areas are frequently activated in deception studies; however other regions have also been identified. It needs to be evaluated how reliable these activated regions are in conveying whether an individual is definitely lying and why there is such variance between studies. Possible reasons for these differences include: sample sizes; memorised or spontaneous lies; incentive to provide a convincing lie; group studies and realism to a real-life individual situation; and the medical history of the participants in the study. Here, the effectiveness of studies and whether fMRI lie detection is reliable and credible beyond a controlled laboratory setting is discussed.

This review paper will explore the extensive research surrounding fMRI and deception studies, highlighting areas of the brain identified as being active during deception and critically assessing their viability to be transferred into criminal courts to detect whether a defendant is lying.

2. fMRI as a Brain Imaging Technique

Measuring brain activity using fMRI imaging works by detecting changes in blood oxygenation and flow in an individual engaged in cognitive tasks. During neural activity, the brain consumes more oxygen, and blood flow increases at the active area(s) (Simpson 2008). Imaging focuses on the magnetic properties of haemoglobin in the blood, with deoxygenated haemoglobin being paramagnetic and therefore more sensitive to the magnetic field, and oxygenated haemoglobin being diamagnetic, thus not attracting to the magnetic field of the MRI [S3]scanner. The decrease of deoxygenated haemoglobin at areas of neural activity results in magnetic susceptibility, hence an increase in MR signal as the diamagnetic blood interferes less with this signal (Bear et al., 2007). This form of MRI is known as Blood-Oxygen Level Dependent (BOLD) imaging, and reliability of these BOLD signals is dependent upon participant compliance, and accuracy in following instructions (Rusconi et al., 2013).

3. The Deceptive Brain

The brain comprises multiple, complex regions involved in various functions, with some regions responsible for multiple functions and some functions requiring numerous brain areas, hence the challenges faced during scientific studies into the sections of the brain involved in deception. Brain regions vary amongst individuals and could explain the significant activation seen in certain brain areas for some studies, but not others. Neuroimaging studies frequently reveal ACC and VLPFC activation during deception; however, these two regions are not solely associated with the cognitive processes involved in lying (Monteleone et al., 2009).

3.1 Executive Control

Executive control, encompassing working memory, task switching, and inhibitory control, has been implicated as playing a role in deceptive responses (Spence 2004; Christ et al., 2009). This is based on deception involving preparing deceptive responses while keeping the truth in mind (working memory), switching between deceptive and truthful responses (task switching) and suppressing truthful responses (inhibitory control) (Christ et al., 2009). Neuroimaging studies support this executive control hypothesis, with results showing promising consistencies in the activation of two PFC regions during deception; the VLPFC, activated during response/truth suppression (Bhatt et al., 2008; Spence et al., 2008a) and the DLPFC, linked with working memory/preparing deceptive responses (Christ et al., 2009; Ito et al., 2012). The PFC as a whole plays an important role during the control of executive functions, and in the manipulation and integration of information (Lee et al., 2002), but it is not solely PFC regions involved in executive control as the supramarginal gyrus has been indicated as playing an active role in working memory (see Ito et al., 2012). The notion that the PFC and working memory mediate deception to generate a lie and conceal the truth is seen in the results of Bhatt et al., 2008. However it cannot be assumed that only brain regions associated with executive control are involved in deception, because other regions, such as the parietal lobules (Spence et al., 2001; Langleben et al., 2002; Lee et al., 2002), are activated across other deceptive studies (table 1). Brain regions consist of subdivisions, which appear to have different functions and levels of activation during deception. The inferior parietal lobule (BA 4/7), for example, is not associated with executive control, but may be involved in other neurocognitive functions involved in deception, whereas the superior parietal lobule (BA 13) is linked to executive control (Christ et al., 2009). Furthermore, some activated brain regions may not actually be associated with deception, for example, activation of the red nucleus in Bhatt et al., 2008 and the motor cortex in Ganis et al., 2003 are attributed to the finger movements during motor responses to questions; thus interpretation of activated brain regions should be made with caution.

3.2 Anti-saccades

Similarities have been observed between the involvement of executive control and the DLPFC in inhibition responses, specifically in deception and eye movement (reflexive saccade), providing evidence that unless these brain regions have lesions, they will show significant activation during deceitful responses (Nyffeler et al., 2007). Comprehensive knowledge of the neural correlates associated with reflexive saccade inhibition may aid scientists in understanding the complex neural system that is the brain, and studies based on other brain areas inhibiting functions can be carried out to see if they also inhibit truthful responses.

Anti-saccadic eye movements interlink neurobiologically with truth inhibition. Automatic reflexive saccades refer to the movement of eyes from a stationary focus point to another visual stimulus; consequently anti-saccades require inhibition of this reflexive saccade to avoid looking at the stimulus, and instead direct their gaze in the opposite direction (Nyffeler et al., 2007; Domagalik et al., 2012). Clearly, greater cognitive functioning is required during anti-saccadic tasks than for reflexive saccadic tasks and control of the inhibition has been linked to the DLPFC (Nyffeler et al., 2007). While research is still in its infancy regarding other brain regions activated during anti-saccade movements, investigators recently provided supportive evidence that the DLPFC was involved in inhibiting automatic eye movement responses, alongside the ACC and other brain regions as part of the executive control network, supporting the deceptive research studies detailed previously (Domagalik et al., 2012).

3.3 Study Comparison

It cannot be ignored that there is methodology variation across studies in terms of sample size, group analysis, motivation and incentive, spontaneity of lies, freedom to lie, and vocal and motor responses (table 1), and these could explain inconsistencies in results amongst activated brain areas. Some studies endeavour to add an aspect of realism and emotional salience to their study by offering monetary incentive, although in comparison to the risk of going to prison, failing to receive money for unconvincing lies is not an equal punishment incentive. Luan-Phan et al., 2005 attempted to replicate the GKT study by Langleben et al., 2002, but an overlap of activation was only seen in a couple of brain regions, highlighting the difficulties in consistently replicating results. Unlike a number of other studies, there was no ACC activation, and the authors associated this with the lack of monetary incentive offered in their study compared to the $20 offered to participants by Langleben and colleagues if they correctly concealed the identify of their playing card (Langleben et al., 2002; Luan-Phan et al., 2005). Two other studies support this hypothesis, where no monetary incentive was offered and no ACC activation was detected (Lee et al., 2002; Spence et al., 2008a). On the other hand, Spence and colleagues offered alternative interpretations that ACC activation was due to the different methodologies employed in other studies or that, based on the results by Ganis et al., 2003, the ACC was more likely to be activated during isolated, spontaneous lies expressed through manual (in contrast to verbal) responses, neither of which were used in their study (Spence et al., 2008a). ACC activation has also been connected to working memory, conflict monitoring, inhibition of competing responses, and the cognitive process of confirming answers are not actually the truth (Ganis et al., 2003). Despite ACC activation in a number of studies, there are also experiments where activation is not seen, and consistent interpretation is yet to be made. Subsequently, further investigation into the neuronal function of the ACC is needed, focusing on whether there is a link between ACC activation and manual button pressings, compared to vocal answers.

It should be brought to attention that some studies only report results matching their prior hypothesis and do not record other activated brain regions, whilst others generalise the results and only report the brain regions accounting for the majority of the participant’s activation (Kozel et al., 2005).

[S1] Give overview of key ideas, methods, questions

[S2] What is ACC? function

[S3] Grammarly match