دوره زمانی تنظیم احساسات اتوماتیک در طول یک کار چهره

Abstract Neuroimaging research has determined that the neural correlates of automatic emotion regulation (AER) include the anterior cingulate cortex. However, the corresponding time course remains unknown. In the current study, we collected event-related potentials (ERPs) from 20 healthy volunteers during a judgment of the gender of emotional faces in a cued Go/Nogo task. The results indicate that Go-N2 amplitudes and latencies following positive and negative faces decreased more than those following neutral faces; Nogo-N2 amplitudes and latencies did not vary with valence. Moreover, positive and negative faces prompted higher P3 amplitudes and shorter P3 latencies than neutral faces in both Go and Nogo trials. These observations suggest that in the executive processes, Go-N2 reflects top-down attention toward emotions, while Go-P3 reflects only motivated attention; in the inhibitory processes, Nogo-N2 reflects cognitive conflict monitoring, while Nogo-P3 overlaps with the automatic response inhibition of emotions. These observations imply that AER can modulate early ERP components, and both Go-N2 and Nogo-P3 can be used as electrophysiological indices of AER.

Introduction Automatic emotion regulation (AER), which is pervasive in daily life and important for mental health (Bargh and Williams, 2006 and Mauss et al., 2007a), is attracting considerable attention from many researchers (Koole and Rothermund, 2011). As an unconscious, implicit, or impulsive process (Williams et al., 2009), AER can be defined as the changes in any aspect of an individual's emotional responses without conscious intent, without engaging in any form of deliberate emotion regulation, and without paying attention to the process of emotional regulation (Mauss et al., 2007a and Mauss et al., 2007b). In fact, AER can be achieved through a number of processes and can modify the subjective experience, peripheral physiology (e.g., cardiac reactivity), or expressive behavior of an emotional response (Mauss et al., 2006, Mauss et al., 2007b, Mocaiber et al., 2011 and Williams et al., 2009). Recent neuroimaging studies using the dot probe, emotional Stroop, and emotional Go/Nogo tasks have determined that the anterior cingulate cortex (ACC) is one of the neural correlates of AER (see Phillips et al., 2008a and Phillips et al., 2008b for a review). However, these investigations cannot reveal the time course of AER due to poor temporal resolution. In contrast, event-related potentials (ERPs) can capture rapid neural responses related to emotions. In the current study, we employed ERP to explore the time course of AER. Emotion regulation, whether deliberate or automatic, involves changes at all levels of the emotion-generation process, including attention deployment, appraisal, and regulation of emotional behaviors (Gross and Thompson, 2007). The late positive potential (LPP), one of the most extensively studied ERP components, is sensitive to deliberate emotional regulation strategies such as reappraisal (DeCiccoa et al., 2011, Foti and Hajcak, 2008, Hajcak and Nieuwenhuis, 2006 and MacNamara et al., 2011), directed attention (Ferrari et al., 2008 and Hajcak et al., 2009), distraction (Thiruchselvam et al., 2011), and directions to increase and decrease subjective emotional responses (Moser et al., 2006, Moser et al., 2009 and Krompinger et al., 2008). Several studies in children and adolescents have determined that the frontal N2 (often dubbed the “inhibitory N2”) is a neurophysiological marker of inhibitory control in Go/Nogo tasks (Lewis et al., 2006a, Lewis et al., 2006b, Lewis et al., 2007, Lewis et al., 2008 and Lamm et al., 2011). To our knowledge, however, only one ERP study has explored the time course of AER, indicating that the implicit reappraisal strategy modulates the LPP associated with affective picture viewing (Mocaiber et al., 2010). Therefore, it remains unclear whether AER modulates early ERP components. As one of the classical approaches to assess ACC function, the Go/Nogo paradigm with non-emotional stimuli has reliably identified two ERP components related to response inhibition, the frontal Nogo-N2 and Nogo-P3, whose amplitudes are larger than Go-N2 and Go-P3, respectively (Albert et al., 2010 and Kiefer et al., 1998). Nogo-N2 is a negative-going component that peaks around 200–400 ms following Nogo stimuli. Subsequent to N2 in Nogo trials, Nogo-P3 is a positive-going component that peaks around 300–700 ms. The Nogo-N2 and Nogo-P3 components reflect different sub-processes of response inhibition. Many researchers suggest that the Nogo-P3 directly reflects the inhibitory process itself (Albert et al., 2010, Bruin and Wijers, 2002, Smith et al., 2008 and Spronk et al., 2008). However, there is still some debate on whether Nogo-N2 reflects conflict monitoring (Donkers and Van Boxtel, 2004, Kenemans et al., 2005 and Nieuwenhuis et al., 2003) or response inhibition (Bartholow et al., 2005 and van Veen and Carter, 2002). Thus, Nogo-N2 may mirror a wide range of cognitive control processes, including attention control, conflict monitoring, inhibition itself, and/or error evaluation (Weissman et al., 2004, Spronk et al., 2008 and Groen et al., 2007). Incorporating affectively salient stimuli into Go/Nogo paradigms provides new insights into the processes related to emotion-modulated execution and inhibition of motor responses (Albert et al., 2010, Albert et al., 2011, Schulz et al., 2007 and Wang et al., 2011). First, affective stimuli draw more attentional resources than neutral stimuli (Lang et al., 1997), and the emotional valence biases behavioral responses when approaching or avoiding stimuli (Hare et al., 2005). Due to limited attentional resources, participants must consciously or unconsciously exert attentional control over emotional stimuli to ensure that they execute goal-related behaviors (Blair et al., 2007 and Taylor and Fragopanagos, 2005). Thus, in the executive process (Go task), a process exists that relates to the attentional control of emotions. For example, there is evidence indicating that Go-N2 signals the extent to which attentional control is required (Nieuwenhuis et al., 2003 and Dennis and Chen, 2007). Second, the emotional valence is sufficient to evoke prepotent response tendencies (Chiu et al., 2008 and Hare et al., 2005). Thus, in the inhibitory process (Nogo task), emotional- and motor-response inhibition coexist and coactivate some brain areas, including the ACC, that are associated with the interaction between emotional processing and motor inhibition (Goldstein et al., 2007, Berkman et al., 2009 and Albert et al., 2011); this interaction is observed in the P3 (but not in the N2) time range (Albert et al., 2011). Moreover, participants use brain responses elicited by emotional stimuli during explicit or implicit tasks in different ways because emotional context can alter behavioral and biological responses (Albert et al., 2010, Carretié et al., 2006 and Hare et al., 2005). In explicit tasks, participants actively use emotional stimuli as perceptual cues to guide them in accomplishing tasks (Dolan, 2002). In implicit tasks, however, participants unconsciously engage in top-down control of the ACC (or the right inferior frontal cortex) over limbic responses elicited by emotional stimuli (Albert et al., 2011, Etkin et al., 2011, Berkman et al., 2009 and Goldstein et al., 2007). Therefore, there may be automatically regulated emotional processes in early conflict monitoring or in the late stages of response inhibition (Wang et al., 2011). Thus, by selecting an implicit emotional task (a cognitive task), tracing the rapid unfolding of emotional responses over time and comparing waveforms between emotional and neutral stimuli, researchers can extract the ERP components that reflect the AER from the executive or inhibitive processes. However, few investigations to date have used the Go/Nogo paradigm with affective stimuli. The present study aimed to analyze the AER-related processes in a cued Go/Nogo paradigm. To this end, we recorded the ERPs of 20 undergraduate volunteers while they performed implicit emotional tasks in which they made only a gender judgment of emotional faces from the Chinese Face Affective Picture System (CFAPS; Wang and Luo, 2005). As noted earlier, if, in the executive process, Go-N2 reflects top-down attention of the ACC toward emotions (Nieuwenhuis et al., 2003 and Dennis and Chen, 2007), then the emotional faces in the Go trials should elicit smaller N2 amplitudes and shorter latencies than neutral faces because emotionally salient faces should automatically activate the neural network for the unconscious focusing of attention to facilitate task completion (Bayle and Taylor, 2010 and Eimer and Holmes, 2007). If in the inhibitory process Nogo-P3 overlaps with automatic response inhibition, then emotional faces in the Nogo trials should elicit larger P3 amplitudes and shorter latencies than neutral faces because emotional- and motor-response inhibitions interact with and activate common brain areas to cause stronger response inhibition (Goldstein et al., 2007, Berkman et al., 2009 and Albert et al., 2011).

. Results 3.1. Behavioral data 3.1.1. Response time Two-way repeated measures ANOVAs revealed a significant main effect of emotional valence on mean response time [F (2,16) = 5.59, p < 0.05, η2 = 0.10], indicating that the mean response times for positive and negative faces were faster than those for neutral faces, but no significant difference between positive and negative faces was observed ( Fig. 1a). The main effect of facial gender and the interaction between emotional valence and facial gender did not reach significance [F (1,16) = 2.17, p > 0.05, η2 = 0.05; F (2,16) = 1.87, p > 0.05, η2 = 0.04, respectively]. Emotional modulation of response time (RT) and accuracy in correct Go trials of ... Fig. 1. Emotional modulation of response time (RT) and accuracy in correct Go trials of the cued Go/Nogo task. (a) The mean RTs for positive and negative faces were lower than the RT for neutral faces when the subjects performed an implicit emotional task (judging the gender of emotional faces), regardless of the cues (male or female) seen prior to the emotional faces. (b) The mean accuracy of recognizing emotional faces was higher than the accuracy of recognizing neutral faces, regardless of the cues. Error bars indicate the standard error. Figure options 3.1.2. Accuracy Three-way repeated measures ANOVAs revealed a significant main effect of trial type on mean accuracy [F (1,16) = 6.35, p < 0.05, η2 = 0.09], indicating a higher accuracy for the Nogo trials (i.e., no responses; M = 0.966, SD = 0.034) than for the Go trials (i.e., button presses; M = 0.950, SD = 0.031). The interaction between trial type and emotional valence was significant [F (2,16) = 10.75, p < 0.001, η2 = 0.15]. Post hoc tests indicated that in the Go trials, the mean accuracies for positive and negative faces were higher than those for neutral faces, although there was no significant difference between positive and negative faces ( Fig. 1b); no significant difference in Emotional valence was observed in the Nogo trials. No other main or interaction effects reached statistical significance. 3.2. ERP data 3.2.1. N2 3.2.1.1. N2 amplitude Three-way repeated measures ANOVAs yielded significant main effects of trial type [F (1,16) = 10.90, p < 0.01, η2 = 0.13] and electrode sites [F (8,152) = 27.28, p < 0.001, η2 = 0.18] as well as a significant three-way interaction between trial type, emotional valence, and electrode sites [F (16,304) = 3.38, p < 0.01, η2 = 0.09]. The N2 amplitudes were largest at the frontocentral sites (e.g., FCz), while the N2 amplitudes in the NoGo trials were more negative than those in the Go trials ( Fig. 2a). The analysis of the three-way interaction showed a significant trial type × emotional valence interaction at frontal [F (2,38) = 5.04, p < 0.05, η2 = 0.08] and frontal-central [F (2,38) = 5.49, p < 0.05, η2 = 0.09] sites, but not at central sites [F (2,38) = 1.93, p > 0.05, η2 = 0.04]. Post hoc tests indicated that positive and negative faces elicited smaller N2 amplitudes than neutral faces in the Go trials ( Fig. 2b); no differences in N2 amplitudes to the three emotional valences in the Nogo trials were significant ( Fig. 2c). Other main and interaction effects did not reach significance. Event-related brain potentials related to automatic emotion regulation at site ... Fig. 2. Event-related brain potentials related to automatic emotion regulation at site FCz. (a) Both N2 and P3 amplitudes were higher in the Nogo trials than in the Go trials. (b) In the Go trials, the N2 amplitudes following emotional faces were lower than the amplitudes following neutral faces, while the P3 amplitudes following emotional faces were higher than the amplitudes following neutral faces. (c) In the Nogo trials, the P3 amplitudes for positive and negative faces were higher than those for neutral faces, and the N2 amplitudes did not vary with emotional valences. Figure options 3.2.1.2. N2 latency Three-way repeated measures ANOVAs assessing N2 latency revealed a significant main effect of trial type [F (1,16) = 7.25, p < 0.05, η2 = 0.10] and a significant interaction between trial type and valence [F (2,16) = 7.89, p < 0.01, partial η2 = 0.16]. Post hoc tests indicated that the N2 latencies in the Nogo trials were longer than those in the Go trials. In the Go trials, positive (M = 276 ms, SD = 12 ms) and negative faces (M = 273 ms, SD = 11 ms) elicited shorter N2 latencies than neutral faces (M = 289 ms, SD = 13 ms), although there was no significant difference between positive and negative faces; in the Nogo trials, no significant differences between positive (M = 295 ms, SD = 10 ms), negative (M = 292 ms, SD = 11 ms), and neutral (M = 297 ms, SD = 12 ms) pictures were observed. Other main and interaction effects did not reach significance. 3.2.2. P3 3.2.2.1. P3 amplitude There was a significant main effect of Trial type on the P3 amplitude [F (1,16) = 8.59, p < 0.01, η2 = 0.10]. Data analyses revealed larger P3 amplitudes in the Nogo trials than in the Go tasks ( Fig. 2a). A significant main effect of emotional valence [F (2,16) = 11.40, p < 0.001, η2 = 0.14] revealed larger P3 amplitudes for positive and negative faces than for neutral faces. Additionally, a significant main effect of electrode site showed that the largest P3 amplitudes were at frontocentral sites [F (8,152) = 23.03, p < 0.01, η2 = 0.12]. No interaction effects reached statistical significance. One-way repeated measures ANOVAs were performed separately for the Go-P3 and Nogo-P3 amplitudes using emotional valence as a factor to further determine whether Go-P3 and Nogo-P3 amplitudes are modulated by emotion. The results indicated that positive and negative faces elicited larger P3 amplitudes than neutral faces in the Go trials [F (2,16) = 8.38, p < 0.01, η2 = 0.12; Fig. 2b] and in the Nogo trials [F (2,16) = 8.67, p < 0.01, η2 = 0.13; Fig. 2c]. 3.2.2.2. P3 latency Three-way repeated measures ANOVAs assessing P3 latency revealed a significant main effect of trial type [F (1,16) = 6.35, p < 0.05, η2 = 0.11]. Data analyses indicated that P3 latencies in the Go trials were longer than those in the Nogo trials. In addition, a significant main effect of emotional valence [F (2,16) = 4.78, p < 0.05, η2 = 0.07] was observed, indicating that both positive (M = 457 ms, SD = 11 ms) and negative (M = 463 ms, SD = 10 ms) faces elicited shorter response latencies than neutral faces (M = 481 ms, SD = 13 ms); the difference between positive and negative faces was not significant. Other main and interaction effects did not achieve statistical significance. Similar to the analysis on the P3 amplitudes, one-way repeated measures ANOVAs were performed separately on the Go-P3 and Nogo-P3 latencies using emotional valence as a factor. The results indicated that positive and negative faces elicited shorter P3 latencies than neutral faces in the Go trials [F (2,16) = 4.81, p < 0.05, η2 = 0.08; Fig. 2b] and in the Nogo trials [F (2,16) = 4.95, p < 0.05, η2 = 0.09; Fig. 2c]. 3.3. Correlation analysis Pearson's correlation analyses demonstrated that Go-P3 amplitudes were negatively correlated with the mean response time in the Go trials (r = −0.43, p < 0.05; Fig. 3a), while the Nogo-P3 amplitudes were positively correlated with Nogo accuracy (r = 0.37, p < 0.05; Fig. 3b). No other significant relationships between ERPs and behavioral data were detected. Two-tailed correlations between Go-P3 amplitudes and response times in correct ... Fig. 3. Two-tailed correlations between Go-P3 amplitudes and response times in correct Go trials (a) (r = −0.43, p < 0.05) and between Nogo-P3 amplitudes and Nogo accuracy (b) (r = 0.37, p < 0.05) across emotional valences.