Open Access
Add to BibSonomyAbstract
In this study, functional near-infrared spectroscopy (fNIRS) was adopted to investigate the prefrontal cortical responses to deception under different motivations. By using a feigned memory impairment paradigm, 19 healthy adults were asked to deceive under the two different motivations: to obtain rewards and to avoid punishments. Results indicated that when deceiving for obtaining rewards, there was greater neural activation in the right inferior frontal gyrus (IFG) than the control condition. When deceiving for avoiding punishments, there was greater activation in the right inferior frontal gyrus (IFG) and the left middle frontal gyrus (MFG) than the control condition. In addition, deceiving for avoiding punishments led to greater neural activation in the left MFG than when deceiving for obtaining rewards. Furthermore, the results showed a moderate hit rate in detecting deception under either motivation. These results demonstrated that deception with different motivations led to distinct responses in the prefrontal cortex. fNIRS could provide a useful technique for the detection of deception with strategy of feigning memory impairment under different motivations.
© 2015 Optical Society of America
1. Introduction
Functional near-infrared spectroscopy (fNIRS), which can assess the relative changes of oxy hemoglobin (HbO), deoxy hemoglobin (Hb) and total hemoglobin (HbT) in the cortical regions of the brain [1, 2 ], is an emerging brain-imaging technique [3]. fNIRS has excellent time resolution and acceptable spatial resolution [4]. Due to its non-invasiveness, low cost and portability [5], fNIRS has been increasingly applied to various areas, such as the detection of deception.
In the past, polygraphs were widely used to detect deception. The principle of polygraphs is to detect the changes of peripheral physiological responses (such as blood pressure, pulse, heart rate and respiration) aroused by deception [6, 7 ]. However, these changes are not considered to represent the process of deception itself, as they just reflect the emotional arousal caused by deception [8]. As a method of directly measuring the brain responses linked to the cognitive processes [8], brain-imaging technique is believed to be able to examine the process of deception itself. Recent efforts have focused on finding some brain signatures that would allow for reliable detection of deception with various brain-imaging techniques [9]. As a result, some fMRI and fNIRS studies have consistently reported that the prefrontal cortex was involved in deceiving [10–13 ]. During the process of deception, the prefrontal cortex plays a significant role in executive functions, including planning, application of strategies, and suppressing the truth [14–16 ]. Thus the prefrontal cortex is recognized as the primary activated brain area during deception [17]. Among the regions of the prefrontal cortex, the right inferior frontal gyrus (IFG) and the left middle frontal gyrus (MFG) were recognized as two important regions for deception [12, 18 ]. Both regions were associated with inhibition, which is the fundamental function of deceiving [19]. Moreover, previous studies revealed that, by using the prefrontal cortex as the detecting index, the fNIRS technique could provide a moderate accuracy of detecting deception [20, 21 ].
However, few studies on deception have considered the effect of motivation. As deceiving is an intentional behavior, understanding the motivation for deceiving is one of the main interests of studying deception [22]. In real life, people often try to deceive others based on two different motivations: obtaining rewards or avoiding punishments [23]. For example, the victims of traffic accidents may deceive to obtain more compensation, while the wreckers may deceive to protect themselves from accountability. The principle of negativity bias indicates that the subjective value of a loss weighs more heavily than a gain [24], suggesting punishments would exert more effects than rewards to the individuals. According to the view of attention model [25], compared to gains, losses would invoke more attention to tasks. Therefore, it is reasonable to assume that the motivation of avoiding punishments would result in additional effort for deceiving than that of obtaining rewards. In fact, some studies have provided evidences that different motivations led to different activation patterns of prefrontal cortex during cognitive tasks [26, 27 ], but the principle of negativity bias has not been fully confirmed in these brain-imaging studies. We inferred that greater prefrontal cortical responses would arise when deceiving for avoiding punishments than when deceiving for obtaining rewards.
In the present study, we firstly aimed to explore whether different motivations (obtaining rewards and avoiding punishments) would lead to different prefrontal cortical responses during deception. Secondly, we aimed to examine the hit rates of detecting deception under different motivations at the individual level. Being the first study to examine the motivational differences of deception by a brain-imaging method, the present study could deepen the understanding for the influence of motivation on the neural responses of deception and test the feasibility of fNIRS to detect deception under different motivations. For the experimental method, we adopted the paradigm of feigned memory impairment, which is a kind of deceptive strategy. When people pretend to have memory impairment, they often intentionally report that they cannot remember the things that they have witnessed, or report that they can remember things that they have not witnessed [28]. Furthermore, to make the memory impairments real, people will make both untruthful statements and truthful ones during the process of deceptions [14, 29 ]. Feigned memory impairment is very common in everyday life and forensic evaluation [14], thus studying the prefrontal cortical responses to deception by feigned memory impairment would improve the ecological validity of researches on the detection of deception.
2 Methods
2.1 Participants and protocol
23 undergraduate and graduate students of South China Normal University participated in this study. The data from 4 participants were excluded as they did not understand our instructions for the experiment. Eventually, 19 valid participants were enrolled (9 males and 10 females, aged from 18 to 26, with mean age 22.05 ± 2.59). All of them were right-handed and with normal or corrected-to-normal visions. None of them had a history of neurological or psychiatric diseases. They were all given informed consents before this experiment. This study was approved by South China Normal University.
Each participant was required to perform face recognition tasks (judging whether they had seen the persons in the photos before) under three different conditions: (1) the control condition, (2) the condition of deceiving for obtaining rewards and (3) the condition of deceiving for avoiding punishments. For the detection of deception, the more activated brain areas compared to the baseline are considered as the useful detection regions. Thus the control condition was set to measure the baseline neural activities of the participants. Each participant needed to complete all the three conditions. In the beginning, each participant was told that they would get 15 RMB for payment after the experiment. In the control condition, participants were required to tell the truth all the time without any other instruction (Telling the truth without any other motivation was consistently perceived as the baseline response [9]). In the condition of deceiving for obtaining rewards, participants were told that the computer was recording their data, and they should deceive the computer that they had memory impairments with some strategies: When they performed the face recognition tasks, they should tell both lies and truths rather than always telling lies. They could tell lies and truths spontaneously. Participants were instructed that in the end of the experiment, the computer would evaluate whether they had memory impairments only based on their performance under this condition. If they successfully deceived the computer, they would get another 10 RMB for rewards. In the condition of deceiving for avoiding punishments, they were given the same instruction except that they were told if they did not successfully deceive the computer, they would lose 10 RMB for punishments. In fact, every participant was told they had successfully deceived the computer in each condition and got 25 RMB after the experiment.
The face recognition tasks adopted 32 photos (16 celebrity photos and 16 non-celebrity photos) as the stimuli. These photos were chosen from 50 photos rated by 60 students who did not participate in this experiment. If the person in one photo had been seen before by 80% of the participants, we defined it as the celebrity photo. If the person in one photo had not been seen before by 80% participants, we defined it as a non-celebrity photo. These photos had the same size, brightness and background colors. The persons in these photos all had neutral facial expressions and postures. The 32 photos were presented in a random order. In each trial (see Fig. 1 ), a fixation “+” was presented for 0.7s. Then a photo was presented for 2.1s, followed by a prompt for 2.8s. The promptings asked the participants that if they had seen the person in the photo before, then they should respond with the keyboard. If they had seen the person in the photo, then they should pressed “Q”, if not, pressed “P”. Participants were told not to judge until the prompts appeared. At last, a black screen would appear for 7.7s. The duration of each step was set to a multiple of the temporal resolution (0.07s) of the measurement of fNIRS.
2.2 Experimental setup
In this study, 42 channels of an fNIRS system (FOIRE-3000, Shimadzu Corporation, Kyoto, Japan) were used to measure the prefrontal cortex of the brain. This fNIRS system contains three wavelengths (780nm, 805nm, and 830nm) [1, 30 ]. The concentration changes of oxygenated hemoglobin (HBO), deoxygenated hemoglobin (HB) and total hemoglobin (HbT) were measured. The parameters were determined with the modified Beer-Lambert law [1, 30 ]. The international 10-10 system [31] was used to localize the whole prefrontal cortex (PFC). The locations of the channels are shown in Fig. 2 . According to the 10-10 system, channels 8, 16, 25, and 33 were mostly located in the right IFG, channels 1, 10, 18, and 27 were mostly located in the left IFG, channels 15, 23, 24, 32, 40, and 41 were mostly located in the right MFG, and channels 11, 19, 20, 28, 36, and 37 were mostly located in the left MFG.
2.3 Data analysis
We used the hemodynamic response function (HRF) filter and a wavelet-MDL (minimum description length) detrending algorithm to remove physical noise and artifacts for each participant [32, 33 ]. Then the data were subtracted by their baseline values (the mean values within 2.03s before the first condition appeared) in each channel for each participant. In this study, only the HbO data was adopted for further analysis.
In order to select the behavioral and cortical responses into further analysis, several steps were processed. Firstly, the trials where the behavioral responses were missing were excluded from further analysis. Secondly, if celebrity photos were judged as “had not been seen before” or non-celebrity photos were judged as “had been seen before” in the control condition, these photos and their corresponding behavioral and fNIRS data would be excluded in all the three conditions. We defined two trial types (the truth-telling trials and the lying trials) in this analysis. “Truth-telling trials” were defined as the trials of the celebrity photos which were judged as “had been seen before”, or the trials of non-celebrity photos which were judged as “had not been seen before” in all the three conditions. In addition, “lying trials” were defined as the celebrity photos which were judged as “had not been seen before”, or the trials of the non-celebrity photos which were judged as “had been seen before” in the two deceiving conditions.
In this study, we mainly focused on the inhibition function of the brain during deception [12, 18 ], and thus we selected one channel in the right IFG and one channel in left MFG as two regions of interest (ROIs) respectively according to topographic images.
When we analyzed the differences of deception under different motivations, we performed the following two steps. Firstly, we processed the individual data. For each participant, we respectively averaged the data of truth-telling trials and the lying trials in each condition channel by channel. Secondly, we performed the group analysis. In the present study, we aimed to separately examine the changes in HbO in the lying trials and the truth-telling trials of the deceiving conditions. We used 2 (ROIs) * 3 (conditions) repeated measurement analysis of variance (ANOVA) at the group level. The comparison among the two ROIs was one channel in the right IFG vs. one channel in the left MFG. The comparisons among the three conditions were the truth-telling trials in the control condition vs. the lying trials of deceiving for obtaining rewards condition vs. the lying trials of deceiving for avoiding punishments condition (the first ANOVA), and the truth-telling trials in the control condition vs. the truth-telling trials of deceiving for obtaining rewards condition vs. the truth-telling trials of deceiving for avoiding punishments condition (the second ANOVA).
When we examined the hit rates of detecting deception under the two different motivations at the individual level, we performed the following steps. Firstly, we selected some detection regions based on the results of group analysis. The effective detection regions were derived from the ROIs mentioned above. Secondly, we adopted a one-way analysis of variance (F test) in each detection region for each participant (the experimental trials which needed to be compared were selected according to the results of group analysis), and then executed the post hoc tests. The standards to identify the deceptions were described later.
3. Results
3.1 Behavioral data
3.1.1 The number of truth-telling trails and lying trials in each deceiving condition
In the condition of deceiving for obtaining rewards, participants had averaged 11 (SD = 5.29) effective lying trials and 14 (SD = 5.10) effective truth-telling trials respectively. In the condition of deceiving for avoiding punishments, participants took part in 9 (SD = 4.43) effective lying trials and 15 (SD = 5.67) effective truth-telling trials. Chi-square tests indicated that there was no significant difference between the number of lying trials and the number of truth-telling trials both in the condition of deceiving for obtaining rewards (χ2(1) = 0.36, p = 0.549) and in the condition of deceiving for avoiding punishments(χ2(1) = 1.50, p = 0.221).
3.1.2 The truth-telling trials in the control condition vs. the lying trials of deceiving for obtaining rewards condition vs. the lying trials of deceiving for avoiding punishments condition
The results of ANOVA showed a significant effect of condition on reaction time (RT) (F(2,36) = 14.527, p = 0.00007, ηp 2 = 0.447). Post hoc tests revealed that the RTs of the lying trials of both deceiving conditions were significantly longer than the RTs of the truth-telling trials of the control condition (MD = 319.572, p = 0.000004; MD = 218.825, p = 0.001, respectively). In addition, there was no significant difference between the RTs of the lying trials of deceiving for obtaining rewards condition and the lying trials of deceiving for avoiding punishments condition (MD = 100.746, p = 0.189). (See Table 1 )
Table 1. The mean and standard error of reaction time (ms) are shown for truth-telling trials in the control condition, and both lying trials and truth-telling trials in the two deceiving conditions (mean ± standard error).
3.1.3 The truth-telling trials in the control condition vs. the truth-telling trials of deceiving for obtaining rewards condition vs. the truth-telling trials of deceiving for avoiding punishments condition
The results of ANOVA showed a significant effect of condition on reaction time (RT) (F(2,36) = 9.294, p = 0.001, ηp 2 = 0.341). Post hoc tests revealed that the RTs of the truth-telling trials of both deceiving conditions were significantly longer than the RTs of the telling-truth trials of the control condition (MD = 273.177, p = 0.0004; MD = 185.819, p = 0.012, respectively). In addition, there was no significant difference between the RTs of the truth-telling trials of deceiving for obtaining rewards condition and the truth-telling trials of deceiving for avoiding punishments condition (MD = 87.358, p = 0.199). (See Table 1)
3.2 fNIRS data
3.2.1 The topographic images of the PFC
As shown in Fig. 3 , deceiving for obtaining rewards and for avoiding punishments both led to increased HbO than the control in the right IFG. Moreover, the lying trails and the truth telling trails in each deceiving condition gave the similar patterns of HbO maps. In contrast to the control, deceiving for obtaining rewards caused more activation in the right IFG (Fig. 3(A), 3(C)), and deceiving for avoiding punishments led to more activation in the bilateral IFG and the left MFG (Fig. 3(B), 3(D)). However, because the left IFG is not the main activation region for deceptions in the previous studies, regarding it as a candidate for detection region at individual analysis is not appropriate. Also, it was difficult to interpret the activation of this region. As mentioned above, we only focused on the right IFG and the left MFG because they were two significant regions in deception. Therefore, we just selected Channel 8 (in the right IFG) and Channel 19 (in the left MFG) as the two ROIs because they were two relatively obvious activation channels in the above two regions according to the topographic images.”
Fig. 3 The topographic images of PFC: (A) The change in HbO from the “lying trials of deceiving for obtain rewards condition” minus “the truth-telling trials of the control condition; (B) The change in HbO from “the lying trials of deceiving for avoiding punishments condition” minus “the truth-telling trials of the control condition”; (C) The change in HbO from “the truth-telling trials of deceiving for obtaining rewards condition” minus “the truth-telling trials of the control condition”; (D) The change in HbO from “the truth-telling trials of deceiving for avoiding punishments condition” minus “the truth-telling trials of the control condition”. The locations of the channels are the same as in Fig. 2.
3.2.2 The comparison among “the truth-telling trials in the control condition vs. the lying trials of deceiving for obtaining rewards condition vs. the lying trials of deceiving for avoiding punishments condition”
The results of descriptive analysis were shown in Table 2 and Fig. 4(A) and 4(B) . We performed a two-way ANOVA to test the effects of condition and ROI on changes in HbO in the truth-telling trials in the control condition and the lying trials in two deceiving conditions. The results showed that there was a significant main effect of condition on changes in HbO (F(2,36) = 12.370, p = 0.00008, ηp 2 = 0.407). The interaction effect of ROI and condition was marginally significant (F(2,36) = 3.201, p = 0.053, ηp 2 = 0.151).
Table 2. The mean and standard error of change in HbO are shown for truth-telling trials in the control condition, and lying trials in each deceiving condition (mean ± standard error).
Fig. 4 Neural activities in the three conditions: (A)The change in HbO from truth-telling trials of the control condition, lying trials of deceiving for obtaining rewards condition and lying trials of deceiving for avoiding punishments condition in Channel 8 (the right IFG); (B) The change in HbO from the truth-telling trials of the control condition, lying trials of deceiving for obtaining rewards condition and lying trials of deceiving for avoiding punishments condition in Channel 19 (the left MFG); (C) The change in HbO from truth-telling trials of the control condition, truth-telling trials of deceiving for obtaining rewards condition and truth-telling trials of deceiving for avoiding punishments condition in Channel 8 (the right IFG); (D)The change in HbO from truth-telling trials of the control condition, truth-telling trials of deceiving for obtaining rewards condition and truth-telling trails of deceiving for avoiding punishments condition in Channel 19 (the left MFG).
Post hoc tests indicated that in Channel 8, lying trials of deceiving for obtaining rewards condition showed significantly larger increase in HbO compared to the truth-telling trials of the control condition (MD = 0.024, p = 0.006), and lying trials of deceiving for avoiding punishments condition showed significantly larger increase in HbO compared to the truth-telling trials of the control condition (MD = 0.034, p = 0.001).
In Channel 19, the lying trials of deceiving for avoiding punishments condition exhibited significantly larger increase in HbO compared to the truth-telling trials of the control condition (MD = 0.018, p = 0.00002). In addition, lying trials of deceiving for avoiding punishments condition showed marginally significant larger increase in HbO compared to lying trials of deceiving for obtaining rewards condition (MD = 0.009, p = 0.056).
3.2.3 The comparison among “the truth-telling trials in the control condition vs. the truth-telling trials of deceiving for obtaining rewards condition vs. the truth-telling trials of deceiving for avoiding punishments condition”
The results of descriptive analysis were shown in Table 3 and Fig. 4(C) and 4(D). We performed a two-way ANOVA to test the effects of condition and ROI on changes in HbO in the truth-telling trials in the control condition and the two deceiving conditions. The results showed that there was a significant main effect of condition on changes in HbO (F(2,36) = 10.793, p = 0.0002, ηp 2 = 0.375). The interaction effect of ROI and condition was marginally significant (F(2,36) = 3.087, p = 0.058, ηp 2 = 0.146).
Table 3. The mean and standard error of change in HbO are shown for truth-telling trials in the control condition, and truth-telling trials in each deceiving condition (mean ± standard error).
Post hoc tests indicated that in Channel 8, truth-telling trials of deceiving for obtaining rewards condition showed significantly larger increase in HbO compared to the truth-telling trials of the control condition (MD = 0.023, p = 0.007), and truth-telling trials of deceiving for avoiding punishments condition showed significantly larger increase in HbO compared to the truth-telling trials of the control condition (MD = 0.034, p = 0.002).
In Channel 19, the truth-telling trials of deceiving for avoiding punishments condition exhibited significantly larger increase in HbO compared to the truth-telling trials of the control condition (MD = 0.018, p = 0.00002). In addition, truth-telling trials of deceiving for avoiding punishments condition showed marginally significant larger increase in HbO compared to truth-telling trials of deceiving for obtaining rewards condition (MD = 0.010, p = 0.055).
3.3 The hit rates of detecting deception under different motivations
As the group analysis of behavioral data and fNIRS data both indicated that whether lying trials or truth-telling trials in the two deceiving conditions showed significantly different responses than truth-telling trials in the control condition, these two trial types were both adopted in the following analysis. Several standards were employed for detecting deception at the individual level: for the behavioral data of one participant, if the p value of F test was significant, and post hoc tests exhibited that effective trials (including the truth-telling trial and lying trial) of deceiving condition for obtaining rewards or avoiding punishments showed significantly longer RTs than the effective trials of the control condition, we decided that this participant could be successfully detected using corresponding detection index; for the fNIRS data of one participant, if the p value of F test was significant, and the post hoc tests exhibited that effective trials (including the truth-telling trial and lying trial) of the deceiving conditions for obtaining rewards or avoiding punishments led to significantly larger increase in HbO compared to the effective trials of the control condition, we decided that this participant could be successfully detected using corresponding detection index. We selected Channel 8 as the detection region of deceiving for obtaining rewards and Channel 8, 19 as the detection regions of deceiving for avoiding punishment, because these brain regions were more activated when participants deceived under the corresponding motivation.
The results showed that, when we detected deceiving for obtaining rewards, 6 out of 19 participants were successfully detected using RT data (the hit rate was 31.6%), and 13 out of 19 participants were successfully detected using fNIRS data in Channel 8 (the hit rate was 68.4%). When we detected deceiving for avoiding punishments, 8 out of 19 participants were successfully detected using RT data (the hit rate was 42.1%). In addition, 16 out of 19 participants were successfully detected using fNIRS data in Channel 8 (The hit rate was 84.2%), and 17 out of 19 participants were successfully detected using fNIRS data in Channel 19 (the hit rate was 89.5%). (See Table 4 ).
Table 4. The results of using different methods (RT data and fNIRS data) to detect deception under the motivations of obtaining rewards and of avoiding punishments
“Rewards” is deceiving for obtaining rewards; “Punishment” is deceiving for avoiding punishment. “Yes” indicates that a particular participant could be successfully detected; “no” indicates that a particular participant could not be successfully detected.
4. Discussion
In this study, we utilized a functional near-infrared spectroscopy technique to explore the prefrontal cortical responses to deception under different motivations. With the facial recognition tasks, we measured the neural activation of lying trials and truth-telling trials of the two deceiving conditions. We examined the different neural activation patterns in the PFC between deceiving for obtaining rewards and deceiving for avoiding punishments, and the hit rates of detecting deception under these two motivations.
Our findings indicated that, the lying trials both in deceiving for obtaining rewards and deceiving for avoiding punishments conditions led to significantly larger increase in HbO in the PFC than truth-telling trials in the control condition, which was similar to previous brain imaging studies [12, 20 ]. Numerous studies have revealed the important role of the PFC in deception [10–13 ], which mainly reflected various executive functions of the brain [12, 19 ]. Specially, both motivations would activate the right IFG during deception. Among the PFC, the right IFG is most related to suppressing the pre-potent truth response [20]. Our results suggested that, like other forms of deception, when people attempt to deceive by feigning memory impairments, they need more efforts to inhibit their pre-potent truth responses. An interesting finding was that truth-telling trials in the two deceiving conditions both showed similar activation patterns to the lying trials in these conditions. These results indicated that when participants made truthful responses among the lying responses, they still attempted to achieve deceitful goals [34]. As a cunning strategy [34], telling some truth during the process of deceptions did not need to inhibit the truthful response, but it still contributed to pretending memory impairments. That is, it was a disguised way of suppressing the truth. In general, these results suggested that the neural signal of deception with strategy of feigning memory impairment could be detected.
Furthermore, our results showed different neural activation patterns when deceiving for two different motivations. The lying and the truth-telling trials of deceiving for obtaining rewards condition were associated with greater activation in the right IFG than the truth-telling trials in the control condition, while these trials of deceiving for avoiding punishments condition were linked to greater activation in the right IFG and left MFG. In addition, these trials of deceiving for avoiding punishments were associated with greater activation in the left MFG than that of deceiving for obtaining more rewards. These results supported the perspective that punishments would exert more effects than rewards to the individuals [24, 25 ]. With more care about the negative consequences, participants paid more attention to the deceiving task when facing punishments, and this led to more effort to inhibit the truth. These effects embodied in additional more activation of the left MFG compared to the baseline when deceiving for avoiding punishments, as well as greater neural activation of the left MFG when deceiving for avoiding punishments than that for obtaining rewards. Although the right IFG and the left MFG are both functioned as inhibition [12, 18, 35 ], some differences still exist. Generally speaking, the right IFG is the most crucial brain region for inhibiting the truth, which is even recognized as a neural marker for successful deceiving [35]. Nevertheless, the left MFG is also linked to other functions such as working memory and generating responses [12]. Therefore, it suggests that when deceiving for avoiding punishments, participants need to make some additional effort to maintain relevant information in the brain and give a misleading impression while answering the questions. These cognitive operations are helpful to inhibit the truth better. That is, the left MFG plays an assistant role in inhibition.
Finally, we examined the hit rates of detecting deception using both behavioral data and fNIRS data at the individual level. When attempting to detect deception for obtaining rewards, 31.6% participants could be successfully detected (hit rate) using behavioral data (RT). However, using fNIRS data (measuring change in Channel 8), this hit rate was 68.4%. Similarly, when we used behavioral data to detect deception for avoiding punishments, 42.1% participants could be successfully detected. Correspondingly, 84.2% participants (measuring change in HbO in Channel 8) and 89.5% participants (measuring change in HbO in Channel 19) were successfully detected using different fNIRS data. Because it was difficult to distinguish the neural activation of lying trials and truth-telling trials under the two deceptive conditions, we adopted both trials for individual analysis. Hence, the above results reflected the hit rates of detecting deception with strategy of feigning memory impairment rather than detecting specific deceptive response. These results showed the hit rates of detecting deception under the two motivations were both moderate, which were similar to the previous fNIRS studies [20, 21 ]. It reflected that fNIRS was sensitive to the neural activities of cortical regions. However, due to the absent of data from “innocent” populations (e.g., patients with memory impairments in this study), the specificity could not be calculated [36]. Even though the hit rates of fNIRS data were higher than the behavioral data in this study, it could not be sufficient to address that fNIRS data was more reliable to detect deception than behavioral data. We could not know whether the increased hit rates would go on hand in hand with increased false positive rates. Future studies should examine the sensitivity as well as the specificity by including the data from innocent people. But at least, the existing results revealed that fNIRS could provide a useful tool to detect deception with strategy of feigning memory impairment under different motivations.
5. Conclusion
In summary, our study has indicated that deceiving for obtaining rewards and for avoiding punishments both exhibited greater neural activation in PFC than the control condition. In addition, deceiving for avoiding punishments was linked to bigger activated regions than deceiving for obtaining rewards, and deceiving for avoiding punishments led to greater neural activation in the left MFG than deceiving for obtaining rewards. In addition, we have shown a moderate hit rate of detecting deception under either motivation. Our study suggests that fNIRS could provide an effective tool to detect deception with strategy of feigning memory impairment under different motivations.
Acknowledgment
This work was supported partially by Guangdong Innovative Research Team Program (No. 201001D0104799318), the Natural Science Foundation of Guangdong province (2014A030310502) and SOARD. We thanked all the participants for the experiment. We thanked Prof. Xue Zheng of SCNU for discussion.
Reference and links
1. H. Bortfeld, E. Wruck, and D. A. Boas, “Assessing infants’ cortical response to speech using near-infrared spectroscopy,” Neuroimage 34(1), 407–415 (2007). [CrossRef] [PubMed]
2. A. Villringer and B. Chance, “Non-invasive optical spectroscopy and imaging of human brain function,” Trends Neurosci. 20(10), 435–442 (1997). [CrossRef] [PubMed]
3. L. Sai, X. Zhou, X. P. Ding, G. Fu, and B. Sang, “Detecting concealed information using functional near-infrared spectroscopy,” Brain Topogr. 27(5), 652–662 (2014). [CrossRef] [PubMed]
4. H. Zhu, J. Li, Y. Fan, X. Li, D. Huang, and S. He, “Atypical prefrontal cortical responses to joint/non-joint attention in children with autism spectrum disorder (ASD): a functional near-infrared spectroscopy study,” Biomed. Opt. Express 6(3), 690–701 (2015). [CrossRef] [PubMed]
5. M. Ferrari and V. Quaresima, “A brief review on the history of human functional near-infrared spectroscopy (fNIRS) development and fields of application,” Neuroimage 63(2), 921–935 (2012). [CrossRef] [PubMed]
6. A. Ginton, “Good intentions that fail to cope with the main point in CQT: a comment on Palmatier and Rovner (2015),” Int. J. Psychophysiol. 95(1), 25–28 (2015). [CrossRef] [PubMed]
7. J. P. Rosenfeld, “Alternative views of Bashore and Rapp’s (1993) alternatives to traditional polygraphy: A critique,” Psychol. Bull. 117(1), 159–166 (1995). [CrossRef]
8. M. J. Farah, J. B. Hutchinson, E. A. Phelps, and A. D. Wagner, “Functional MRI-based lie detection: scientific and societal challenges,” Nat. Rev. Neurosci. 15(2), 123–131 (2014). [CrossRef] [PubMed]
9. M. Gamer, “Mind reading using neuroimaging: Is this the future of deception detection?” Eur. Psychol. 19(3), 172–183 (2014). [CrossRef]
10. S. Henry and D. Plemmons, “Neuroscience, neuropolitics and neuroethics: the complex case of crime, deception and FMRI,” Sci. Eng. Ethics 18(3), 573–591 (2012). [CrossRef] [PubMed]
11. X. P. Ding, X. Du, D. Lei, C. S. Hu, G. Fu, and G. Chen, “The neural correlates of identity faking and concealment: an FMRI study,” PLoS One 7(11), e48639 (2012). [CrossRef] [PubMed]
12. A. Marchewka, K. Jednorog, M. Falkiewicz, W. Szeszkowski, A. Grabowska, and I. Szatkowska, “Sex, Lies and fMRI-Gender Differences in Neural Basis of Deception,” PLoS One 7(8), e43076 (2012). [CrossRef] [PubMed]
13. N. Lisofsky, P. Kazzer, H. R. Heekeren, and K. Prehn, “Investigating socio-cognitive processes in deception: a quantitative meta-analysis of neuroimaging studies,” Neuropsychologia 61, 113–122 (2014). [CrossRef] [PubMed]
14. T. M. Lee, H. L. Liu, C. C. H. Chan, Y. B. Ng, P. T. Fox, and J. H. Gao, “Neural correlates of feigned memory impairment,” Neuroimage 28(2), 305–313 (2005). [CrossRef] [PubMed]
15. S. M. Carlson, L. J. Moses, and H. R. Hix, “The role of inhibitory processes in young children’s difficulties with deception and false belief,” Child Dev. 69(3), 672–691 (1998). [CrossRef] [PubMed]
16. E. K. Miller and J. D. Cohen, “An integrative theory of prefrontal cortex function,” Annu. Rev. Neurosci. 24(1), 167–202 (2001). [CrossRef] [PubMed]
17. S. E. Christ, D. C. Van Essen, J. M. Watson, L. E. Brubaker, and K. B. McDermott, “The contributions of prefrontal cortex and executive control to deception: evidence from activation likelihood estimate meta-analyses,” Cereb. Cortex 19(7), 1557–1566 (2009). [CrossRef] [PubMed]
18. D. Sun, T. M. C. Lee, and C. C. H. Chan, “Unfolding the spatial and temporal neural processing of lying about face familiarity,” Cereb. Cortex 25(4), 927–936 (2015). [CrossRef] [PubMed]
19. B. Verschuere, T. Schuhmann, and A. T. Sack, “Does the inferior frontal sulcus play a functional role in deception? a neuronavigated theta-burst transcranial magnetic stimulation study,” Front. Hum. Neurosci. 6, 284 (2012). [CrossRef] [PubMed]
20. F. Tian, V. Sharma, F. A. Kozel, and H. Liu, “Functional near-infrared spectroscopy to investigate hemodynamic responses to deception in the prefrontal cortex,” Brain Res. 1303(25), 120–130 (2009). [CrossRef] [PubMed]
21. X. S. Hu, K. S. Hong, and S. S. Ge, “fNIRS-based online deception decoding,” J. Neural Eng. 9(2), 026012 (2012). [CrossRef] [PubMed]
22. K. E. Sip, J. C. Skewes, J. L. Marchant, W. B. McGregor, A. Roepstorff, and C. D. Frith, “What if I get busted? Deception, choice and decision-making in social interaction,” Front. Neurosci. 6, 58 (2012). [CrossRef] [PubMed]
23. A. Ito, N. Abe, T. Fujii, A. Hayashi, A. Ueno, S. Mugikura, S. Takahashi, and E. Mori, “The contribution of the dorsolateral prefrontal cortex to the preparation for deception and truth-telling,” Brain Res. 1464(29), 43–52 (2012). [CrossRef] [PubMed]
24. R. F. Baumeister, E. Bratslavsky, C. Finkenauer, and K. D. Vohs, “Bad is stronger than good,” Rev. Gen. Psychol. 5(4), 323–370 (2001). [CrossRef]
25. E. Yechiam and G. Hochman, “Losses as modulators of attention: review and analysis of the unique effects of losses over gains,” Psychol. Bull. 139(2), 497–518 (2013). [CrossRef] [PubMed]
26. S. M. Tom, C. R. Fox, C. Trepel, and R. A. Poldrack, “The neural basis of loss aversion in decision-making under risk,” Science 315(5811), 515–518 (2007). [CrossRef] [PubMed]
27. J. M. Spielberg, G. A. Miller, S. L. Warren, A. S. Engels, L. D. Crocker, B. P. Sutton, and W. Heller, “Trait motivation moderates neural activation associated with goal pursuit,” Cogn. Affect. Behav. Neurosci. 12(2), 308–322 (2012). [CrossRef] [PubMed]
28. J. N. Browndyke, J. Paskavitz, L. H. Sweet, R. A. Cohen, K. A. Tucker, K. A. Welsh-Bohmer, J. R. Burke, and D. E. Schmechel, “Neuroanatomical correlates of malingered memory impairment: event-related fMRI of deception on a recognition memory task,” Brain Inj. 22(6), 481–489 (2008). [CrossRef] [PubMed]
29. E. G. Mak and T. M. Lee, “Detection of feigned memory impairments using a Chinese word task,” Psychol. Rep. 98(3), 779–788 (2006). [CrossRef] [PubMed]
30. J. Li and L. Qiu, “Temporal correlation of spontaneous hemodynamic activity in language areas measured with functional near-infrared spectroscopy,” Biomed. Opt. Express 5(2), 587–595 (2014). [CrossRef] [PubMed]
31. L. Koessler, L. Maillard, A. Benhadid, J. P. Vignal, J. Felblinger, H. Vespignani, and M. Braun, “Automated cortical projection of EEG sensors: anatomical correlation via the international 10-10 system,” Neuroimage 46(1), 64–72 (2009). [CrossRef] [PubMed]
32. S. Tak and J. C. Ye, “Statistical analysis of fNIRS data: a comprehensive review,” Neuroimage 85(Pt 1), 72–91 (2014). [CrossRef] [PubMed]
33. K. E. Jang, S. Tak, J. Jung, J. Jang, Y. Jeong, and J. C. Ye, “Wavelet minimum description length detrending for near-infrared spectroscopy,” J. Biomed. Opt. 14(3), 034004 (2009). [CrossRef] [PubMed]
34. X. P. Ding, L. Sai, G. Fu, J. Liu, and K. Lee, “Neural correlates of second-order verbal deception: A functional near-infrared spectroscopy (fNIRS) study,” Neuroimage 87(15), 505–514 (2014). [CrossRef] [PubMed]
35. O. Vartanian, P. J. Kwantes, D. R. Mandel, F. Bouak, A. Nakashima, I. Smith, and Q. Lam, “Right inferior frontal gyrus activation as a neural marker of successful lying,” Front. Hum. Neurosci. 7, 616 (2013). [CrossRef] [PubMed]
36. I. D. Hill, “What are the sensitivity and specificity of serologic tests for celiac disease? Do sensitivity and specificity vary in different populations?” Gastroenterology 128(4Suppl 1), S25–S32 (2005). [CrossRef] [PubMed]
