Emotion regulation is essential for individual’s social life (Gross, 2015), and it refers to the attempts to influence which emotions to have, when to have them, and how to experience or express them (Gross, 1998). According to the process model of emotion regulation (Gross, 1998), there are five essential emotion regulation strategies: situation selection, situation modification, attentional deployment, cognitive change and response modulation. Suppression (inhibition) has been attributed as one of the best forms of response modulation during emotional regulation, which refers to the efforts to inhibit individual’s emotion-expressive behavior (Gross, 2015). Response inhibition to affectively salient stimuli can provide more insights into the emotion-modulated inhibition control processes (Schulz et al., 2007; Chiu et al., 2008; Albert et al., 2010, 2012). It has also been found that emotional processing on the erotic and painful information can impair individual’s following inhibitory control processes (Yu et al., 2012, 2015), and the threatening information can improve inhibitory performances due to the enhancement of perceptual and cognitive processes (Senderecka, 2016). Moreover, emotional conflict control is regarded as playing an important role in monitoring affective conflict situations and modulating the influence of emotional and social distractors on behavior (Botvinick et al., 2001), and it relates to attentional deployment during emotion regulation by directing selective attention with the goal of affecting individual’s emotional response (Gross et al., 2011a,b; Etkin et al., 2015; Gross, 2015; Sheppes et al., 2015). The main aim of current study was to investigate how perceived emotional information modulated the interaction between response inhibition and conflict control processes.

Response inhibition is broadly investigated by the Go/Nogo paradigm with the requirement of executing responses to one type of stimuli (Go response) and withholding responses to the other type (Nogo response), and response inhibition on affective information might relate to suppression on improper emotions during emotion regulation (Ladouceur et al., 2006; Schacht et al., 2009; Frischen et al., 2012; Yu et al., 2014). In the Go/Nogo task, a negative event-related potentials (ERPs) component, the Nogo-N2 with fronto-central neural generation is elicited in 250–350 ms time window after the occurrence of Nogo stimuli; the N2 component reflects the detection of conflict between inhibition requirement and response execution (Nieuwenhuis et al., 2003; Donkers and van Boxtel, 2004; Lo et al., 2013; Ocklenburg et al., 2013). It has also been found that frontal beta power relates to the neurodevelopment of inhibitory control during early childhood (Lo et al., 2013). Some studies have shown that the amplitude of Nogo-N2 cannot be affected by the emotional valence of information (Chiu et al., 2008; Todd et al., 2008; Yu et al., 2009; Albert et al., 2010; Zhang and Lu, 2012), however, some other studies reported that arousing negative information or highly unpleasant information would induce more negative Nogo-N2 responses compared with neutral and pleasant information (Albert et al., 2012; Yuan et al., 2012).

Another essential ERP component during response inhibition is the Nogo-P3 occurring in 300–600 ms time window after the onset of the Nogo stimuli over fronto-central brain areas (Kiefer et al., 1998), and it is associated with successful motor suppression and the evaluation of the inhibition outcome (Bruin et al., 2001). Relative to the Nogo-P3, the Go-P3 is in response to Go stimuli and distributed mainly over parietal areas, and it relates to execution rather than inhibition (Tekok-Kilic et al., 2001). In the emotional Go/Nogo tasks, Nogo-P3 responses to emotional information were larger and faster than Go emotional responses (Zhang and Lu, 2012). Varied findings have been reported about the modulation of emotional valence on the Nogo-P3 responses. Albert et al. (2010, 2012) reported that Nogo-P3 is greater to pleasant information compared to unpleasant information, which was not observed in Yuan et al.’s (2012) study. Zhang and Lu (2012) observed that the Nogo-P3 responses to both negative and positive expressions were greater than that to neutral expressions, however, Chiu et al. (2008) found that the Nogo-P3 could not be modulated by the emotional valence of the stimuli.

Conflict control, which is widely measured by the flanker paradigm (Eriksen and Eriksen, 1974) relates to the processes of monitoring the conflicts in perceptual inputs or between required responses and individual’s preferred responses (Botvinick et al., 2001), and conflict control on affective information relates to the monitoring and resolution processes on affective conflicts (Albert et al., 2010). Several electrophysiological and brain imaging studies have manifested that conflict N2 is sensitive to conflict monitoring with the neural generators of frontal regions, such as anterior cingulated cortex (ACC) and prefrontal cortex (van Veen and Carter, 2002a,b; Ullsperger et al., 2005; Folstein and Van Petten, 2008), and N2 amplitudes are more negative for incongruent trials compared to congruent trials (van Veen and Carter, 2002a,b). The P3 responses are associated with conflict resolution and allocation of attentional control (Hillman et al., 2009a,b; Clayson and Larson, 2011), and the incongruent trials induce larger P3 than the congruent trials (Hillman et al., 2009a,b). Previous behavioral and electrophysiological studies showed that response speed was faster in emotional congruent trials than emotional incongruent trials, and happy faces with sad distracters induced more negative N2 amplitudes than happy faces flanked by identical faces during conflict monitoring stage (Fenske and Eastwood, 2003; Liu et al., 2013); while during the conflict resolution stage, happy faces in the incongruent trials elicited slower P3 responses compared to that in the congruent trials, and the attentional control on sad faces in the incongruent trials induced larger P3 responses than that in the congruent trials (Fenske and Eastwood, 2003; Liu et al., 2013). In the same line, Ochsner et al. (2009) adopted functional magnetic resonance imaging (fMRI) technology to investigated conflict control on affective words and reported that bilateral dorsal ACC were strongly activated in the affective incongruent condition.

As introduced above, conflict control and response inhibition are two vital emotion regulation strategy processes (Zhang and Lu, 2012), and recent studies have investigated how these two processes interplayed in the non-emotional context (Bunge et al., 2002; Huizinga et al., 2006; Brydges et al., 2012, 2013). By using a Go/Nogo flanker task, Brydges et al. (2012) reported that maximum N2 responses distributed on midline electrodes over central areas for conflict control and midline electrodes over frontal areas for response inhibition, and they also found that N2 latencies were longer for conflict control than for responses inhibition. Two developmental studies also demonstrated the distinction of conflict control and response inhibition with the Go/Nogo flanker tasks, and it has been reported that these two processes induced distinct brain areas in children and adults (Bunge et al., 2002; Brydges et al., 2013). Brydges et al. (2013) further illustrated that N2 responses gradually frontally distributed with age for response inhibition, and the N2 latencies and amplitudes became shortened or decreased with age development; however, with regard to conflict control, the significant N2 effect was only shown in adults but not in children.

Moreover, emotional stimuli, such as facial expressions, might carry essential social-emotional information, and the accurate and proper detection, perception, management and regulation on facial expressions were essential for individual’s emotion regulation (Calder and Young, 2005). However, it is less known about the interaction and/or distinction between response inhibition and conflict control in emotional contexts. Hence, the main aim of current study was to investigate the interaction between response inhibition and conflict control on facial expressions. It was hypothesized that the situation that required both Nogo responses and conflict control would require greater N2 and P3 activation compared to other conditions, and response inhibition and conflict control in emotional context induced different brain activities.

The current study investigated the interaction between response inhibition and conflict control on facial expressions. It was found that the identification of happy faces was more accurate in the congruent trials than incongruent trials. During neural processing of monitoring process, Nogo responses induced faster and stronger detection processing compared to Go responses; for Nogo responses, incongruent trials elicited greater detection process than congruent trials. During the neural processes of conflict resolution/inhibition control, the valence of the expressions could modulate the frontal P3 responses in response inhibition but not in conflict control. Taken together with the difference waveform analyses, the current findings showed that response inhibition and conflict control on emotional information relied on distinct processes.

Conflict control on emotional faces mainly focuses on the detection and resolution on affective conflicts (Albert et al., 2010; Bayle and Taylor, 2010; Etkin et al., 2011). In line with prior studies, it was currently observed that the response speed was faster to identify happy faces than fearful faces (Leppänen et al., 2003; Leppänen and Hietanen, 2004; Schulz et al., 2007). In addition, participants performed with higher accuracies when identifying happy faces in congruent trials (HHHHH) than in incongruent trials (FFHFF), which was consistent with previous findings of conflict control on emotional information (Eastwood et al., 2001; Fenske and Eastwood, 2003; Rowe et al., 2007; Ochsner et al., 2009; Liu et al., 2013). Fenske and Eastwood’s (2003) behavioral findings manifested that the flanker-compatibility effect was smaller for negative target faces compared with positive target faces, which illustrated that the constriction of attention could be influenced by the valence of emotional faces. Previous studies have found that cognitive control could be influenced by emotional valence of facial expressions (Fenske and Eastwood, 2003; Liu et al., 2013) and also levels of motivation intensity (Harmon-Jones et al., 2013). Harmon-Jones et al. (2013) demonstrated that the affects of low motivational intensity could broaden the attentional scope, however, the affects of high motivational intensity narrowed the scope. The regulation of emotion when viewing sport video could induce the changes of skin conductance responses (SCRs), and the SRCs were increased in the emotion regulation condition compared with the control condition (Morawetz et al., 2016). In the current study, the emotional expressions were not as arousing as the sport film clips which might only require participants’ recognition of expressions and might only further induce their emotional state changes during the judgment of the expressions. In addition, response inhibition on emotional faces can be attributed as inhibition processes to certain emotion or affective situation, which activates frontal cortex to engage in deliberate top-down attentional control upon emotional context (Nieuwenhuis et al., 2003; Shafritz et al., 2006; Dennis and Chen, 2007; Chiu et al., 2008; Kiss et al., 2008; Zhang and Lu, 2012). Besides emotional faces, Yu et al. (2012) adopted erotic images or painful video clips before a classic stop signal task, and intended to find out whether participants’ cognitive inhibitory control was affected by the emotional processes on the emotional information. The authors observed that emotional processing on the emotional stimuli impaired male participants’ inhibitory control with slower stop signal RT after viewing emotional stimuli compared to the neutral stimuli. Interestingly, when participants were told that their expressions were recorded by a webcam during the experiment, the impairing effect of inhibitory control was eliminated; these significant findings suggested that the interaction between inhibition control and emotional processing might be affected individual’s state of self-consciousness and cognitive load (Yu et al., 2015).

Priori electrophysiological studies have shown that both conflict control and response inhibition processes could elicit frontal N2 responses, and Nogo-N2 in response inhibition related to the detection of required inhibitory response (Falkenstein et al., 1999) and N2 in conflict control was associated with conflict monitoring processing (van Veen and Carter, 2002a; Carter and van Veen, 2007; Folstein and Van Petten, 2008). Similar to Nogo-N2 responses to non-emotional information, the current Nogo-N2 responses to affective faces were also found to be slower than Go-N2 responses, which illustrated that it took more neural processes to detect Nogo signals comparative to Go signals (Albert et al., 2010). Nogo-N2 responses did not vary with the affective valences of faces (Chiu et al., 2008; Albert et al., 2010, 2012; Zhang and Lu, 2012), which further supported that the interaction between response inhibition and emotional process did not occur in the detection stage (N2 responses; Albert et al., 2012). Neuroimaging and brain injury studies showed that the frontal areas (such as the right inferior frontal cortex [rIFC], and ACC) were the fundamental brain regions for general response inhibition (Aron et al., 2004; Hampshire et al., 2010) and also affective response inhibition on emotional faces (Ochsner et al., 2004; Shafritz et al., 2006; Albert et al., 2010). The neural development study using Time-Frequency analysis also found that frontal beta power highly related to the development of inhibition control in 5–6 years old children (Lo et al., 2013). Moreover, Morawetz et al. (2016) investigated the functional interrelationships among several brain regions during emotion regulation by presenting short videos to participants and checking the manipulation with SCRs, and it was found that dorsolateral prefrontal cortex (DLPFC) was the key node of the prefrontal emotion regulation network and strongly connected with the inferior frontal gyrus (IFG).

Neuroimaging studies have shown that bilateral dorsal ACC, posterior medial frontal cortex, and DLPFC were strongly activated to execute affective conflict control (Ochsner et al., 2009), and DLPFC was also activated during threat-related learning (Wheelock et al., 2014). It was also currently shown that the conflict detection of happy faces with fearful distractors induced more negative N2 responses relative to that of fearful faces with happy distractors, and in the same vein Fenske and Eastwood’s (2003) behavioral findings manifested that the flanker-compatibility effect was smaller for negative target faces compared with positive target faces, which illustrated that the constriction of attention could be influenced by the valence of emotional faces (Fenske and Eastwood, 2003). Furthermore, it was currently observed that the monitoring of target happy face induced greater N2 activation in incongruent trials than congruent trials, which replicated the previous findings (Liu et al., 2013). It was further demonstrated that individual’s visual view field was widened by positive information; therefore, it required incremental interference control on happy targets with fearful distractors (Eastwood et al., 2001; Rowe et al., 2007). Unexpectedly, the reverse effect on N2 amplitudes of fearful faces between congruent and incongruent trials was observed compared to that of happy faces. Some relevant ERP and behavioral studies did not observe the significant differences between congruent trials and incongruent trials in response to fearful faces as that to happy faces (Liu et al., 2013; Yang et al., 2016). Yang et al. (2016) regarded the disappearance of emotional conflict effect for negative faces were due to the timely resolution on the negative conflicts. The current results might further reveal that the conflict monitoring was enhanced when the stimuli only contained fearful faces, and the negative fearful information might narrow one’s view field (Fenske and Eastwood, 2003). Taken together, these findings suggested that when the processes required both conflict control and response inhibition, and the congruency effect would be weaken. Compared to conflict control processes, response inhibition processes were likely modulated by emotion.

With regard to P3 responses, P3 activation in response inhibition relates to executive/inhibitory control (Bokura et al., 2001), and P3 in conflict control reflects conflict resolution (Hillman et al., 2009a,b; Clayson and Larson, 2011). Source location analyses showed that prefrontal cortex and parietal areas might be the neural generators of P3, which were responsible for inhibitory control, conflict resolution and outcome valuation processes (Bokura et al., 2001). It was currently found that Go responses to happy faces induced faster frontal P3 than that to fearful faces and frontal P3 Nogo response to fearful faces was faster than that to happy faces. These findings might reveal that it was easier to execute response on happy information than fearful information; however, on the contrary, it was easier to inhibit on fearful information than happy information. Nogo responses induced greater frontal P3 than Go responses, which illustrated that frontal cortex was related with inhibition control processes. In addition, inhibition control on fearful faces elicited greater P3 activation relative to executive control, and Nogo-P3 responses were greater for fearful faces compared to happy faces, which was similar with the findings that response inhibition on negative stimuli required more action inhibitory process (Yu et al., 2014). Moreover, Senderecka (2016) also investigated the interaction between inhibition control and emotional processing on arousal stimuli via a emotional stop signal task, and it was observed that emotional processing on threatening stimuli improved participants’ inhibition control with increasing inhibitory rate and decreasing stop signal RT. Taken together with Yu et al.’s (2012) study, it can be deduced that whether emotional processing impairs or improves the subsequent inhibition control might also be influenced by the valence of the emotional stimuli.

The neural mechanisms of parietal P3 responses were similar with the late positive potential (LPP) during emotion regulation processes which was sensitive to emotional attentional deployment, such as, distraction (Thiruchselvam et al., 2011), attention modulation (Hajcak et al., 2009), and directions to regulate subjective affective responses (Krompinger et al., 2008; Moser et al., 2006, 2009). Moreover, fearful faces with happy distractors required greater conflict resolution processing (longer and greater P3 responses) compared to fearful faces in congruent trials, which was consistent with previous findings (Liu et al., 2013) and might reveal that affective incongruent trials required more attentional control process on the conflicts compared to affective congruent trials (Liotti et al., 2000; Frühholz et al., 2011).

Importantly, the difference analyses revealed that the multiple processes and response inhibition induced stronger N2 and P3 responses than conflict control processing, and this finding manifested that the multiple emotion regulation processes required more neural efforts than the comparatively simple conflict control processing. Furthermore, these current findings might indicate that response inhibition and conflict control on emotional information were two distinct processes, which was consistent the findings in the non-emotional contexts (Bunge et al., 2002; Brydges et al., 2012, 2013). The separation between interference control and response inhibition has been investigated via the non-emotional Go/Nogo flanker task (Brydges et al., 2012, 2013), and participants were instructed to respond according to the color and direction of the stimuli (fishes). It is reported that the incongruent flanker condition induced a larger and later N2 responses compared with the Nogo condition, indicating the separation of interference control and response inhibition. In the same vein, Bunge et al. (2002) adopted the arrow flanker and Go/Nogo task in children, and it was observed that children experienced difficulty in suppressing inappropriate responses and were more susceptible to interference. Children’s effective interference control was related with prefrontal activation and their effective response inhibition was associated with activation of posterior brain regions. Therefore, from the neurodevelopment perspective, the conflict control and response inhibition processes required varied brain activities since the childhood.

There were contrary opinions on whether dissociable brain mechanisms were involved in emotional and non-emotional cognitive control. Some studies demonstrated the domain-general control network for both emotional and non-emotional control, which observed the similar activation of ACC and DLPFC during conflict control on both emotional and non-emotional stimuli in the revised AX Continuous Performance Task (AX-CPT; Chiew and Braver, 2011). However, some studies held the opposite opinion that brain activated distinct brain networks for emotional and non-emotional control on conflicts, and the dissociation also occurred on the brain function level (Egner et al., 2008; Soutschek and Schubert, 2013; Mian and Eskandar, 2016; Xue et al., 2016). By using emotional and non-emotional stroop tasks, Egner et al. (2008) illustrated that cognitive control on non-emotional conflicts activated dACC and DLPFC for conflict monitoring and attentional control, respectively; while, for emotional conflict control processing, dACC and amygdala were enhanced for conflict monitoring, and rostral ACC instead of DLPFC was increasingly activated for attention control on conflicts. Xue et al. (2016) also reported the differences between emotional and non-emotional cognitive control processes with larger conflict effect (RT in the incongruent trials minus RT in the congruent trials) in non-emotional than emotional stroop tasks. Besides, Ochsner et al. (2009) using emotional and non-emotional flanker tasks to compare the neural activation differences, and it was observed that DLPFC, dorsal ACC and posterior medial frontal cortex were activated during conflict control on both emotional and non-emotional conflicts; however, rostral medial PFC and left ventrolateral PFC were separately activated during affective and cognitive control processes. Therefore, Ochsner et al. (2009) further indicated that conflict control on emotional and non-emotional conflicts depended on shared and distinct brain systems. With regard to inhibition control on affective information, Depue et al. (2006) revealed that inhibitory effects were larger for emotional information than non-emotional information when investigating suppression in memory. As for the interaction processes between responses inhibition and conflict control, it required further exploration to identify the differences in emotional and non-emotional contexts in the future study. In addition, it has been reported that individuals showed perceptual expertise for own-related faces (Wiese, 2013) and they could remember own-gender faces more accurately than other-gender faces (Wolff et al., 2014). However, we did not adopt large amount of trials to investigate the own-gender bias in the present study, and in our future work, we would increase the numbers of trials and participants to further explore own-related bias when investigating the interaction between response inhibition and conflict control processes.

In summary, the current study investigated the interactions between responses inhibition and conflict control on affective information with the emotional Go/Nogo flanker task. The electrophysiological findings showed that the interaction between these two emotion regulation processes mainly occurred during monitoring/detection stage, and the multiple processes and response inhibition on emotional information required more neural efforts than conflict control processing. The current study also suggested that response inhibition and conflict control required distinct neural network activities in the emotional context.

TL and JS designed the experiment. TL and TX collected and analyzed the data. TL, TX and JS wrote the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.