Common and distinct neural correlates of self-serving and prosocial dishonesty

Abstract

People often anticipate certain benefits when making dishonest decisions. In this article, we aim to dissociate the neural-cognitive processes of (1) dishonest decisions that focus on overall benefits of being dishonest (regardless of whether the benefits are self-serving or prosocial) from (2) those that distinguish between self-serving and prosocial benefits. Thirty-one participants had the opportunity to maximize their monetary benefits by voluntarily making dishonest decisions while undergoing functional magnetic resonance imaging (fMRI). In each trial, the monetary benefit of being dishonest was either self-serving or prosocial. Behaviorally, we found dissociable patterns of dishonest decisions: some participants were dishonest for overall benefits, while others were primarily dishonest for self-serving (compared with prosocial) benefits. When provided an opportunity to be dishonest for either self-serving or prosocial benefits, participants with a stronger overall tendency to be dishonest had stronger vmPFC activity, as well as stronger functional connectivity between the vmPFC and dlPFC. Furthermore, vmPFC activity was associated with decisions to be dishonest both when the benefits of being dishonest were self-serving and prosocial. Conversely, high self-serving-biased participants had stronger striatum activity and stronger functional connectivity between the striatum and middle-mPFC when they had a chance to be dishonest for self-serving (compared with prosocial) benefits. Altogether, we showed that activity in (and functional connectivity between) regions in the valuation (e.g., vmPFC and Str) and executive control (e.g., dlPFC and mmPFC) systems play a key role in registering the social-related goal of dishonest decisions.

Keywords: deception; dishonest decision making; executive control; prosocial; self-serving.

Figure 1

Task structure of the Coin‐Guessing task (translated from Chinese). In each trial, participants predicted the outcome of a coin flip. Correct predictions corresponded to winning CNY 10, while incorrect predictions corresponded to losing CNY 10. During the No‐Opportunity trials (signified by the word, “RECORD”), participants had to enter their prediction as either “Heads” or “Tails” by pressing either the “H” or the “T” key. During the Opportunity trials (signified by the word, “RANDOM”), participants had to randomly press one of the two “R” keys to control for motor activity. Earnings during the Self trials (signified by the word, “SELF”) would go to the participants themselves, while earnings during Donation trials (signified by the name of a chosen charitable organization, written down here as “CHARITY”) would be donated to a chosen charitable organization

Figure 2

Behavioral results of the Coin‐Guessing task. Figure 2a shows individual differences in Overall Dishonesty, defined by total self‐reported % accuracy across both Opportunity‐Self and Opportunity‐Donation trials. If participants honestly reported their prediction, their overall dishonesty should be around 50%. Sixteen participants (ID 16–31; represented by gray bars) reported improbably high levels of accuracy at the individual level, as revealed by a one‐tailed binomial test, p <.001 (Greene & Paxton, 2009). Figure 2b shows individual differences in Self‐Serving Dishonesty, defined by self‐reported % accuracy during Opportunity‐Self trials minus % accuracy during Opportunity‐Donation trials. Figure 2c shows a scatter plot between Overall Dishonesty and Self‐serving Dishonesty. This plot indicates a nonsignificant relationship between the two indices (r(29) = −.15, p =.42). The gray shaded area in the scatterplot represents 95% CIs around the linear regression line. Figure 2d shows self‐reported % accuracy as a function of Opportunity and Self‐Serving conditions. Error bars represent within‐subject 95% CIs (Morey, 2008)

Figure 3

Neural activity of the Opportunity vs. No‐Opportunity effects as a function of ranked Overall Dishonesty, Opportunity‐Self Dishonesty and Opportunity‐Donation Dishonesty. Figure 3a shows neural activity of the Opportunity vs. No‐Opportunity effects across both Self and Donation trials as a function of ranked Overall Dishonesty. Overall Dishonesty is defined by total self‐reported % accuracy across both Opportunity‐Self and Opportunity‐Donation trials. The top section shows a positive relationship between Overall Dishonesty and neural activity in the ventromedial prefrontal cortex (vmPFC) when participants had an opportunity to over‐report accuracy [Opportunity > No‐Opportunity]. The bottom section shows a positive relationship between Overall Dishonesty and the functional connectivity between the vmPFC and bilateral dorsolateral prefrontal cortex (dlPFC). The functional‐connectivity analysis was conducted using a PPI between Opportunity and No‐Opportunity conditions collapsing across Self and Donation trials with the vmPFC [3 57 −6] as a seed. Figure 3b shows neural activity of the Opportunity vs. No‐Opportunity effects separately for Self and Donation trials as a function of ranked Opportunity‐Self Dishonesty and Opportunity‐Donation Dishonesty, respectively. Opportunity‐Self Dishonesty and Opportunity‐Donation Dishonesty are defined by self‐reported % accuracy in Opportunity‐Self and Opportunity‐Donation trials, respectively. This figure shows positive relationships between self‐reported % accuracy and neural activity in the ventromedial prefrontal cortex (vmPFC) when participants had an opportunity to over‐report accuracy in both Self [Opportunity‐Self Dishonesty and Opportunity‐Self > No‐Opportunity‐Self contrasts] and Donation [Opportunity‐Donation Dishonesty and Opportunity‐Donation > No‐Opportunity‐Donation contrasts] trials. The images were based on whole‐brain regression analyses [Cluster‐forming threshold at p <.005, cluster‐wise corrected (pFWE <.05)]. The pink shaded area in the rank‐transformed scatterplot (higher value = higher rank) represents bootstrapped 95% CIs around the linear regression line (Pernet, Wilcox, and Rousselet, 2013) [Color figure can be viewed at http://wileyonlinelibrary.com]

Figure 4

Neural activity of the Self vs. Donation effects as a function of ranked Self‐Serving Dishonesty. Self‐Serving Dishonesty is defined by self‐reported % accuracy during Opportunity‐Self trials minus % accuracy during Opportunity‐Donation trials. The top section shows a positive relationship between Self‐Serving Dishonesty and neural activity in the Striatum (Str) when participants evaluated the coin‐flip outcome for themselves compared with for donation [Self > Donation]. The bottom section shows that this relationship also applied to situations when people had a chance to make dis/honest decisions [Opportunity‐Self > Opportunity‐Donation]. Note that there was no significant relationship between Self‐Serving Dishonesty and neural activity in the Str when there was no chance [No‐Opportunity‐Self > No‐Opportunity‐Donation]. The bottom section also displays a positive relationship between Self‐Serving Dishonesty and the functional connectivity between the Str and middle‐medial prefrontal cortex (mmPFC). The functional‐connectivity analysis was conducted using a PPI between Opportunity‐Self and Opportunity‐Donation conditions with the dorsal part of the Str [−15 27 12] as a seed. The images were based on whole‐brain regression analyses [Cluster‐forming threshold at p <.005, cluster‐wise corrected (pFWE <.05)]. The pink shaded area in the rank‐transformed scatterplot (higher value = higher rank) represents bootstrapped 95% CIs around the linear regression line (Pernet et al., 2013) [Color figure can be viewed at http://wileyonlinelibrary.com]

Figure 5

Neural activity during the coin‐guessing task across participants. The images were based on whole‐brain one‐sample t‐test analyses [Cluster‐forming threshold at p <.005, cluster‐wise corrected (pFWE <.05)]. IPL, inferior parietal lobe; L‐dlPFC, left dorsolateral prefrontal cortex; dmPFC, dorsomedial prefrontal cortex; mmPFC, middle‐medial prefrontal cortex; vmPFC, ventromedial prefrontal cortex; ACC, anterior cingulate cortex; AI, anterior insula; Str, striatum; PPC, posterior cingulate cortex [Color figure can be viewed at http://wileyonlinelibrary.com]