Printer Friendly

The Initial Emotional Output Cannot be Modified: The Premier Expression.

Decades of research have provided empirical evidence that emotional responses are unique, specific to the aroused emotion, and have low visibility (e.g., Davidson, Ekman, Saron, Senulis, & Friesen, 1990; Levenson, Ekman, & Friesen, 1990; Scherer & Wallbott, 1994). However, facial expression is a much more visible index of emotions than internal measures (i.e., blood pressure); indeed, the face has attracted the most attention as a behavioral index of emotion. Several researchers have also been interested in slight facial reactions, particularly unconscious facial responses to the facial expressions of others (Dimberg, 1982; 2000; Dimberg & Thunberg, 1998). These responses occur roughly 300ms after stimulus onset. Although Dimberg (1990) did report differential facial responses via electromyography (EMG) in response to emotional images, his study neglected a central feature of investigations of emotion responses--namely, how they change across time (Mauss, Levenson, McCater, Wilhelm, & Gross, 2005). While the use of posed facial expressions to reveal emotions (e.g. Ekman & Friesen, 1986; Matsumoto et. al., 2000; Rosenberg & Ekman, 1995) has been popular in the past, spontaneous facial actions might be more useful for understanding the mechanism underlying emotional responses and as well as determining which emotion is actually felt. In other words, the facial expressions that appear just after affective stimulus onset--in particular the very first facial response--could most accurately reflect the genuine felt emotion.

Because faces are the "primary site of affects" (Tomkins & McCarter, 1964, p.121), they might be considered as a broader source of emotions that are superior to any visceral or outer skeletal responses, particularly in terms of speed, precision, and complexity. The very first facial reaction--namely, the primary affective output--has been termed the "premier expression" in the context of inducing surprise (Zhu & Suzuki, 2016). Specifically, Zhu and Suzuki (2016) found particular facial activity occurring about 150ms after surprise stimulus onset that was not affected by conscious suppression, whereas activity subsequent to that (from about 300 ms to 500 ms following onset, depending on the muscle region) was. Based on the assumption that the face is the location for the initial expression of emotion, the premier expression was interpreted as a facial component that genuinely reflected a felt emotion and does not appear to be influenced by either conscious manipulation or social context (Zhu & Suzuki, 2016).

The premier expression is based on the above-stated notion that the face is a superior means for measuring emotions, and researchers interested in exploring it must naturally consider the affective expressive process from a moment-to-moment perspective. Despite chronological changes is the most essential character of the short-lived affective responses (e.g., Kettunen, Ravaja, & Keltikangas-Jarvinen, 2000), most prior studies examined affective expression from a gross perspective, which may have led to missing key initial facial components that require a high temporal resolution approach to discover. Given that the premier expression (Zhu & Suzuki, 2016) is defined as a rapidly occurring primary facial action in response to felt emotions, and is therefore automatic and reflexive, it cannot easily be suppressed (although it is expected to vary positively with the intensity of emotion and arousal level). However, the expressions following the premier expression are considered slower to occur and can be easily concealed or modified via conscious manipulation (Zhu & Suzuki, 2016).

So far, the premier expression has only been observed in response to surprise stimuli. Thus, it would be important to determine whether it occurs for other emotions in order to confirm its generality and features as the first output of emotion. This was the purpose of the present study.

In the present study, facial EMG (fEMG), which has high temporal resolution, was used to capture micro-facial activities. Following Zhu and Suzuki (2016), we measured reactions of the corrugator supercilii muscle and zygomaticus major muscle. Facial reactions located in the corrugator supercilii muscle have proven sensitive to negative emotions (e.g. Caccioppo, Petty, Losch, & Kim, 1986; Dimberg, 1997; Dimberg & Thunberg, 1998; Lundquist & Dimberg, 1995), whereas those in the zygomaticus major muscle are reactive to both extremes of valence, showing a quadratic curve (Larsen, Norris, & Cacioppo, 2003; Vrana, 1993; Yartz & Hawk, 2002).

To test our hypothesis that the premier expression is the very first facial reaction in response to an affective stimulus that expresses the evoked emotion, we also recorded event-related potentials (ERPs) using electroencephalography (EEG). This is because the primary processing of the stimulus' visual input should occur well ahead of the premier expression, if it is indeed an emotional reaction. If not, the response might merely reflect a type of orientating reflex. ERPs include multiple positive and negative components, with each reflecting a distinct stage of processing. A single ERP component reflects one or more aspects of complex informative processing (e.g. Donchin & Coles, 1988; Gehring, Gratton, Coles, & Donchin, 1992). For example, the P300 component is well known to appear in response to emotional stimuli, and shows greater amplitudes in response to negative stimuli than to positive ones (e.g., Chapman, McCrary, Chapman, & Martin, 1980; Conroy & Polich, 2007; Cuthbert, Schupp, Bradley, Birbaumer, & Lang, 2000; Delplanque, Silvert, Hot, Rigoulot, & Sequeira 2006a; Delplanque, Silvert, Hot, & Sequeira, 2006b; Ito, Larsen, Smith, & Cacioppo, 1998; Keil, Bradley, Hauk, Rockstroh, Elbert, & Lang, 2002; Naumann, Bartussek, Diedrich, & Laufer, 1992). In the current study, we examined the relationships between the premier expression and the preceding brain activity in relation to emotional processing, so we decided to consider ERPs that occur even earlier than the P300. Specifically, we focused on the P1 or N1, which are observed during the initial phase of emotional processing. Olofssonet al. (2008) reported that N1 amplitudes were higher in response to emotional visual stimuli compared to neutral stimuli. A similar difference was found for P1 and P2 amplitudes measured over the Pz (Smith, Cacioppo, Larsen, & Chartrand, 2003; Herbert, Kissler, Junghofer, Peyk, & Rockstroh, 2006). ERPs reflecting primary processing of visual emotional stimuli are anticipated correlate with the premier expression and occur before it.

The present study was designed to determine the features of the premier expression. In this experiment, after stimulus onset, the premier expression was assumed to be the first measurable emotional response and was considered to be unaffected by conscious or unconscious manipulation, although the responses occurring subsequent to it were.

Furthermore, we hypothesized that the premier expression was moderated by valence and arousal level. To investigate the premier expression, nine categories of affective stimuli were used in this experiment in order to observe that whether it varies with valence or arousal level. We hypothesized that the corrugator supercilii muscles will respond to unpleasant stimuli, whereas the zygomaticus major muscle would only respond to pleasant stimuli. In addition, the premier expression should only occur after the primary emotional processing of stimuli in the brain.



Participants were 62 undergraduate students (20 males, 42 females) with a mean age of 19.97 years (SD = 1.37). Participation was compensated with partial class credit. Participants were randomly assigned to the control or suppression condition, which resulted in 32 participants for the control condition and 30 participants for the suppression condition. Data from 8 participants were removed for the following reasons: five participants fell asleep during the stimulus presentation; 3 participants showed artifacts in their EEG data due to simultaneously recorded eye movements. The analysis was conducted on the remaining 54 participants. The experiment was approved by Doshisha University's ethics committee.


Eighteen stimuli were used for the presentation, which were selected using a pilot study of 67 undergraduate students (20 males, 47 females) from Doshisha University. Their average age was 20.13 years old (SD=1.09). They also received extra course credit for their participation. Three of these participants were removed because of they did not understand the meaning of arousal. The pictures were selected from the International Affective Picture System (IAPS) according to their overall ratings of valence and arousal (Lang, Bradley, & Cuthbert, 1999). Since the IAPS contains no pictures with a combination of moderate arousal and neutral valence, we employed random geometric patterns as neutral stimuli in this study. Stimuli were selected such that they fell into 9 groups of differing combinations of valence (pleasant, neutral, and unpleasant) and arousal1 (high, middle, and low), with each category containing 12 stimuli. In total, there were 108 pictures, which the 67 undergraduates were asked to rate on a 9-point scale ranging from 1(unpleasant/low arousal) to 9 (pleasant/high arousal). We ultimately selected two sets of 9 pictures (two each in each of the above nine categories) for presentation. The ratings for valence and arousal are shown in Table 1.

Data collection

The fEMG, EEG, and vertical electrooculography (EOG) activities were recorded using NEC 1253 Amplifiers (time constants of 0.01s, 0.3s, and 10s, respectively) and digitized using a BIOPACMP150 data acquisition system. All received data were stored on a personal computer.

The fEMG activity was measured according to the guidelines from Fridlund and Cacioppo (1986). Ag/AgCl (NT-611U NIHON KOHDEN) miniature electrodes were filled with electrode paste and attached to the left corrugator supercilii and zygomaticus major muscle regions, recording at a sampling frequency of 1000Hz.

EEG data were collected from 3 active Ag/AgCl (NIHON KOHDEN) electrodes placed at Fz, Cz, and Pz according to the International 10-20 System. The data were referenced to two mastoid electrodes, recording at a 500Hz sampling frequency. For artifact checking, we recorded vertical eye movements from the electrodes above and beneath the right orbicularis oculi muscle at a sampling frequency of 250 Hz. Impedance of each electrode was kept below 5 k[OMEGA].


Participants all visited the lab individually and first received an explanation of the experiment, after which they signed an informed consent form. Then, the participants took a seat in a comfortable chair in a sound-attenuated shield room (7.38 feet x 7.71 feet, height 6.73 feet). A 33-inch monitor was placed about 3.28 feet in front of the chair, on which the affective pictures were displayed. The fEMG, EEG, and EOG electrodes were then attached to the participants, and they were told to avoid excessive eye blinking during the picture presentation, and to avoid head or body movements. Additionally, participants were informed that if they felt any discomfort during participation, they could quit anytime by pressing an escape button placed on the right side of the chair. Furthermore, they were allowed to practice not blinking with a presentation of 20 pictures.

For the control condition, participants were instructed to just look at the pictures (e.g., Gianotti, 2008), since this condition captures the spontaneous facial responses to affect pictures without interfering with any kind of cognitive load. While in the suppression condition, the participants were told to keep their face neutral as much as possible no matter how they feel. To test our hypothesis, the suppression condition designed to examine the premier expression could not be modified even if the participants tried to.

Each participant viewed one set of the 9 pictures (one picture for each valence and arousal combination). The stimuli were presented in a random sequence of 270 trials in 30 blocks. Each trial began with a 0.5-s focus point (a "+" shown in the middle of the screen), followed by a picture for 2s and then a white screen for 0.9 to 1.4s.


Data processing

All fEMG signals were first high-pass filtered (60 Hz) and then integrated using Acknowledge 4.2 offline. The fEMG data were integrated in 50 ms intervals during the first 1000 ms after stimulus onset. The facial responses were extracted and averaged only from blocks 3-7, since the repeated presentations were intended for ERP averaging, and because the repetition might have led to an attenuation of the affective response. Facial reactions were defined as any changes in activity (in microvolts) from the pre-stimulus level, which was defined as the mean activity in the 500 ms before stimulus onset. Additionally, we calculated integrated fEMG responses for intervals of 0-300ms, 300-600ms, and 600-900ms after stimulus onset as three periods for further analysis.

EEG signals were initially filtered from 0.01 to 30 Hz, after which they were processed using EEGLAB. After visually checking for artifacts and eye blinking, epochs of 1200ms (200ms pre- and 1000 ms post onset) were extracted, and ERPs were averaged for each category. No correction was used for the artifact. We examined the time windows (50-150 ms after stimulus onset) for the P1 and Nlcomponents in this analysis.

fEMG Responses

Overall, fEMG responses for the corrugator supercilii muscle were available for 51 participants (14 males, 37females), of which 24 were in the control condition and 27 were in the suppression condition. For the zygomatuc major muscle, 52 participants (14 males, 38 females) were available, of which 25 were in the control condition and 27 were in the suppression condition.

A 3 x 3 x 3 x 2 (valence x arousal x period x condition) mixeddesign analysis of variance (ANOVA) was conducted separately for each muscle region. Since we postulated premier expression as the initial emotional output, it was necessary to conducted a detailed checking on how the muscle activities changed during the first 300ms after stimuli onset. Another 3 x 3 x 7 x 2 (valence x arousal x time-series x condition) mixed-design ANOVA was also conducted for the first 300ms after stimulus onset. The time series variable reflected the seven 50-ms intervals during the first 300 ms. When Mauchly's test of sphericity was significant, we applied Greenhouse-Geisser corrections where appropriate, and Bonferroni corrections for posthoc comparisons.

Corrugator supercilii muscle

The 300-ms interval grand mean and standard error are illustrated in Figure1.

The main effects of valence (F(1.83, 89.72) = 12.67, p = .000, [[eta].sup.2] = .039) and arousal (F(2, 98) = 3.75, p = .027, [[eta].sup.2] = .011) were significant. Furthermore, the valence x period interaction (F(2.11, 103.45) = 4.11, p = .017, [[eta].sup.2] = .005), arousal x period interaction (F(2.28, 111.52) = 4.21, p = .014, [[eta].sup.2] = 004) and conditionx valence x period interaction(F(2.11, 103.45) = 3.99, p = .020, [[eta].sup.2] = .005) were also significant. To test how the valence and arousal affected the muscle activity in the corrugator supercilii muscle, further comparisons of the simple effects for the valencex period interaction were then executed. At 0-300 ms (F(2, 98) = 5.44, p = .006), 300-600ms (F(1.71, 83.98) = 8.98, p = .001) and 600-900ms (F(1.61, 79.08) = 12.36, p = .000), the simple effect of valence was significant, which indicated that across all 3 periods, the muscle activity was affected by valence. Furthermore, for the neutral (F(1.82, 89.33) = 6.77, p = .003) and pleasant stimuli (F(1.82, 88.94) = 3.28, p = .047), the simple effects of period were significant. Post hoc tests showed that, during all 3 periods, activity in the corrugator supercilii muscle for unpleasant stimuli were greater than activities in responding to neutral (0-300 ms, p = .003; 300-600 ms, p = .001; 600-900 ms, p = .000) or pleasant (0-300 ms, p = .011; 300-600 ms, p = .002; 600-900 ms, p = .001) stimuli. This indicated that the muscle activity collected from the corrugator supercilii muscle varied with valence across time. For the neutral stimuli, activity in the corrugator supercilii muscle at 0-300 ms was greater than that during the 300-600 ms (p = .002) or 600-900 ms(p = .010) period; and for pleasant stimuli, activity during 0-300 ms was greater than that during 300-600 ms (p = .015).These results showed that as we anticipated, the corrugator supecilii muscle was an appropriate index for negative affect.

We then performed further comparisons of the simple effects for the arousalx period interaction. We found a significant simple effect of arousal during the 600-900 ms period (F(1.69, 83.03) =5.96, p= .006), and a simple effect of period for low arousal (F(1.76, 86.01)=8.49, p= .001). Post hoc tests indicated that, during the 600-900 ms period, the corrugator supercilii muscle showed greater activity in response to high arousal stimuli than in response to low arousal stimuli (p=.003). For high arousal stimuli, activity in the corrugator supercilii muscle region was greatest at 0-300 ms (0-300 ms> 300-600 ms, p = .001; 0-300 ms> 600-900 ms, p = .003).

The multiple comparisons for the condition x valencex period interaction confirmed that the first affective facial response--the premier expression--was not affected by suppression, but the subsequent responses were. The condition x valence interaction was found significant for 600-900 ms (F(1.61, 79.08) = 5.16, p = .012). For unpleasant stimuli, activities from the corrugator supercilii muscle under the control condition were greater than that under the suppression condition, which indicated the suppression was successful when performed during 600-900 ms after stimuli onset.

Extra one-sample Ltests were applied to the activity of the corrugator supercilii muscle during the 0-300 ms period, confirming that the increase of muscle activities was statistically significant (Table 2).

The activities in the corrugator supercilii muscle during the first 300 ms after unpleasant stimuli onset were found significant (marginally significant for Middle Arousal-Unpleasant stimuli under control condition) for both conditions.

Finally, to examine the detailed activity in the corrugators supercilii muscle, another 3x 3 x 7x 2 (valence x arousal x time series x condition) mixed-design ANOVA was also conducted for only the first 300ms in 50 ms intervals after stimulus onset (Figure2).

The main effect for valence (F(2, 98) = 5.84, p = .004, [[eta].sup.2] = .012) and valencex time-series interaction was found to be significant (F(9.08, 444.68) = 2.55, p = .007, [[eta].sup.2] = .007).This indicated that muscle activities collected from the corrugator supercilii muscle within 300ms after stimuli onset also did vary with valance and arousal. Further analysis showed that significant simple effect of valence was found across all time points 150 ms after stimuli onset(150ms, F(2, 98) = 5.28, p = .007; 200ms, F(2, 98) = 3.27, p = .042; 250 ms, F(2, 98) = 7.52, p = .001; 300 ms F(1.81, 88.56) = 4.17, p = .022). Specifically, at 150ms, 200ms and 250ms, compared to neutral pleasant stimuli, responses were greater for unpleasant (150ms, p = .002; 200ms, p = .009; 250ms,p = .001); at 150ms and 250ms, responses for unpleasant stimuli were also greater than that for pleasant stimuli (150ms, p = .019; 250ms,p = .015). Such increasing of muscle activity was observed in responding to unpleasant stimuli. Further multiple comparisons were executed since significant simple effect was found for time series of unpleasant stimuli (F(4.3, 210.83) = 4.12, p = .002). The results indicated that processing of unpleasant stimuli elicited significant increases of activity in the corrugator supercilii muscle at 150ms (p = .000), 200ms (p = .002), and 250ms (p = .000) after stimulus onset from 0 level.

Zygomaticus major muscle

The same analyses as above were performed for the zygomaticus major muscle activity. The grand mean and standard error of the first 300 ms are illustrated in Figure 3.

The main effect of period was significant (F(1.74, 86.97) = 10.40, p = .000, [[eta].sup.2] = 006). Moreover, a condition x period interaction was also significant (F(1.74, 86.97) = 4.41, p = .019, [[eta].sup.2] = .003). Further analysis indicated that, during the 300-600 ms period, activity in the control condition was significantly greater than in the suppression condition (F(1,50) = 4.28, p = .044), which suggested that in the suppression condition, the muscle activity from the zygomaticus major muscle was successfully suppressed during this period. Furthermore, a significant simple effect of period was found in the control condition (F(1.69,40.51) = 7.93, p = .002). Compared to the 0-300 ms period, activity in the zygomaticus major muscle in the control condition was significantly greater at 300-600ms(p = .015) and 600-900 ms(p = .003).

Further comparisons of the condition x valence x period interaction (F(2.93, 146.59 = 1.49, n.s.) were also conducted, as with the corrugator supercilii muscle. The results showed that activity in response to unpleasant stimuli varied according to period (F(2, 100) = 8.29, p = .001). We also found significant condition x period (F(2, 100 = 4.75, p = .010) and arousal x period interactions (F(4, 200 = 2.95, p = .021) for unpleasant stimuli. At 600-900 ms, activity in the zygomaticus major muscle in the control condition was marginally significantly greater than that in the suppression condition (F (1, 50 = 3.98, p = .052). Furthermore, activity gradually increased over time in the control condition (0-300ms < 300-600 ms, p= .003; 300-600 ms< 600-900 ms, p = .036; 0-300 ms< 600-900 ms, p= .008) when processing unpleasant stimuli.

A 3 x 3 x 7 x 2 (valence x arousal x time series x condition) mixeddesign ANOVA was also conducted for activity in the zygomaticus major muscle in the first 300ms (Figure 4).

Activity was found to decrease over time (F(4.23, 211.62) = 3.74, p= .005, [[eta].sup.2] = .006). At 50ms (p = .001), 150ms (p = .001), and 200ms (p = .001) after stimulus onset, activity was significantly smaller than that before stimulus onset. This analysis suggested that there was no sign of premier expression in the zygomaticus major muscle that we expected increasing muscle activity for positive affect.


EEG data were available for 53 participants (14 males, 39 females). A window of 50-150 ms after stimulus onset was applied in order to investigate the peak-to-peak amplitudes between P1 and N1. This window was used based on Begleiter et al. (1967), and measured the peak-to peak amplitude because it was considered to reflect the primary emotional processing in relation to the premier expression.

A 3 x 3 x 2 (valence x arousal x condition) mixed-design ANOVAs were conducted for Fz, Cz, and Pz. Significant main effects were found for valence(L(2, 102) = 4.68, p = .011, [[eta].sup.2] = .010) and arousal(L(1.78), 90.55) = 4.05, p = .025, [[eta].sup.2] = .009) at Fz, and subsequent multiple comparisons indicated the peak-to-peak amplitude of the P1 and N1 for unpleasant stimuli were significantly greater than for neutral stimuli(p = .005).The significant valence effect suggested that the initial emotional processing was executed at this early stage. Furthermore, we also found significant valence x condition(F (2,102) = 3.29, p = .041, [[eta].sup.2] = .007) and valence x arousal x condition interactions (F (4, 204) = 2.46, p = .046, [[eta].sup.2] = .012). Post hoc tests indicated that while processing high arousal stimuli, peak-to-peak amplitudes were significantly greater for neutral stimuli than for pleasant stimuli (p = .008) in the suppression condition. By contrast, when processing middle arousal stimuli, amplitudes were significantly greater for unpleasant stimuli than for neutral stimuli (p = .004). At Cz or Pz, no significant differences in amplitudes were found.


To test determine whether the premier expression correlated with the preceding emotional processing, cross-correlation analyses between the ERP grand mean waves over Fz and average fEMG activity in the corrugator supercilii muscle were conducted for unpleasant stimuli from 200ms before to 1000ms after stimulus onset. Since the activities in the first 300 ms after stimuli onset from the zygomaticus major muscle showed no sign of the premier expression, we performed this analysis only for the corrugator supercilii muscle. Figure 6 illustrates how fEMG activities in response to unpleasant stimuli were strongly associated with the preceding ERPs.

The maximal correlation coefficients and time lags between the ERP grand mean waves and fEMG activity in the corrugator supercilii muscle are shown in Table 3. The increasing muscle activities were strongly correlated to their preceding processing of unpleasant stimuli.


The current study (a) examines whether the premier expression varies with valence or arousal and is not subject to manipulation and (b) tests whether the premier expression is associated with initial emotion processing. If the premier expression is associated with initial emotional processing, we would be able to specify the lags in brain activity from visual input until the first physiological output of emotion.

The results from fEMG showed that the activity collected form the corrugator supercilii muscle varied with valence, and the muscle activity increased with the arousal level of unpleasant stimuli. The significant activity increasing was observed within 300ms after unpleasant stimuli onset, which supported our hypothesis that premier expression exists and it is an emotional response. Also the subsequent facial activity was found greater for control group indicated that the premier expression, which captured in the corrugator supercilii cannot be modified. However, to the pleasant stimuli, the initial muscle activity was not found in the zygomaticus major muscle. Moreover, the valence effect was found for peak-to-peak amplitudes between P1 and N1 that the amplitudes were greater for unpleasant stimuli, which indicated that the initial emotional processing was conducted within 150 ms after stimuli onset. Cross-correlation was applied between fEMG activity from the corrugator supercilii muscle and the ERPs grand mean wave over Fz for unpleasant stimuli showed the facial response was strongly correlated to the brain activity and about 100ms behind it.

The fEMG activity in the corrugator supercilii muscle was greater for unpleasant stimuli (Figure1). This was consistent with the results of Dimberg (1982, 1990, 1997) and Lundquist and Dimberg (1995), which indicated that the corrugator supercilii muscle was an appropriate index of the processing of unpleasant emotions. Furthermore, for high arousal/unpleasant and middle arousal-unpleasant stimuli, the muscle activity in the control group gradually increased over time, while the activity in the suppression group showed a smaller increase or remained stable. The premier expression is considered to emerge during the first 300ms after stimulus onset and be resistant to suppression--this was indeed found for unpleasant stimuli across all 3 arousal levels, as shown in Figure 1. When looking at the activity across the seven 50-ms segments of the first 300ms, shown in Figure 2, we observed that high arousal-neutral stimuli showed different response patterns compared to high arousal-unpleasant and middle arousal-unpleasant stimuli. Although for high arousal-neutral stimuli, activity in the suppression group increased at 50 ms after stimuli onset which appeared slightly increasing during the 0-300 ms period; this might be indicative of a possible response to the stimulus specifically. Such a response would not interfere with our conclusion that premier expression was found for unpleasant stimuli.

On the other hand, facial muscle activity in the zygomaticus major muscle showed three patterns of responses. First, for high arousal-unpleasant, middle arousal-unpleasant, and high arousal-pleasant stimuli, muscle activity in the control group increased sharply over time, while the activity in the suppression group gradually decreased. However, the activity was relatively lower in this muscle than that in the corrugator supercilii muscle. Overall, these findings suggest that the zygomaticus major muscle responded to the valence dimension in a quadratic manner, which was consistent with the preceding study by Larsen et al. (2003). The second major pattern observed was that there were minor changes in the high arousal-neutral, middle arousal-neutral, middle arousal-pleasant, and low arousal-pleasant over time in both groups. Finally, the last pattern suggested gross decreases in activity from stimulus onset for low arousal/unpleasant and low arousal-neutral stimuli in both groups. Nevertheless, there were no specific responses during 0-300ms after stimulus onset, the time course at which the premier expression would occur, in the zygomaticus major muscle. Rinn (1984) suggested that the muscles of the lower face were relatively easier to control, which suggested that the zygomaticus major muscle might not be an appropriate method of examining emotional expressions in experiments that use suppression as a methodological necessity.

The present study indicates that the premier expression was observed only in response to unpleasant stimuli, in the corrugator supercilii muscle. Thus, the premier expression sometimes occurred in the present experiment and was more sensitive to unpleasant emotions. Furthermore, the activity underlying the premier expression varied with arousal level for unpleasant stimuli only. This accords with findings that the corrugator supercilii muscle is a major location for emotional expression (Lang, Greenwald, Bradley, & Hamm, 1993; Larsen et al., 2003). It is also crucial to note that the activity in the zygomaticus major muscle did not exhibit signs of a premier expression, unlike what was observed for surprising stimuli (Zhu & Suzuki, 2016). This might suggest that the premier expression was stronger in the corrugator supercilii muscle for unpleasant stimuli of mild to high arousal. Nevertheless, it is also possible the sensitivity to unpleasant stimuli is caused by the intensity of pleasant stimuli, which were actually evaluated as intense as the unpleasant ones, but relatively less perceived. If it were the case, activity from the zygomaticus major muscle would show the same response pattern as that from the corrugator supercilii muscle to unpleasant stimuli as to more intensive pleasant stimuli.

The ERP analysis indicated that, during the early stage of processing incoming stimuli, the peak-to-peak amplitudes of the P1 and the subsequent N1 varied with the valence and arousal of the stimuli. The greater amplitudes in response to unpleasant stimuli compared to neutral stimuli were consistent with the notion that the P1 is sensitive to negative stimuli (Carretie et al., 2004a, b; Delplanque et al., 2004; Smith et al., 2003). Furthermore, we extend findings on the "negative bias" proposed by Cacioppo and colleagues (Cacioppo & Berntson, 1994; Cacioppo et al., 1997; Ito & Cacioppo, in press), which refer to how negative stimuli are associated with large late positive potential amplitudes. In our study, this negative bias also appears to be true of ERPs in the early period. Notably, participants were not instructed to attend to either the stimuli's valence or arousal level, which means that their evaluation of the valence during primary processing was automatic. However, we found no differences in activity among the frontal regions between conditions, even for affective stimuli. This is in contrast to Beauregard, Levesque, and Bourgouin, (2001) and Levesque et al. (2003), who reported that the right superior frontal gyrus and right anterior cingulate cortex activation were associated with attempted inhibition of sexual arousal and suppression of sadness. Since the current experiment used pictures from the IAPS to manipulate valence and arousal levels, the emotions generated by the stimuli were perhaps less intense than in the former experiments.

Regarding the final purpose of the present study, we investigated the relationship between ERPs and the premier expression through a cross-correlation analysis. This analysis was only applied to the responses to unpleasant stimuli, since the responses to other categories were mild. From the results, the ERPs and premier expression were highly positively correlated, and the premier expression emerged after the ERPs. This shows that the premier expression is not merely an automatic muscular contraction, but represents the first affective expression after primary emotional stimulus processing.

Although our experiment was designed to use multiple presentations of the same stimuli to reduce this orienting reflex as much as possible, it does not rule out the reflex itself. This might cause some limitation to the conclusion that the premier expression is a completely emotional response. However, for pleasant stimuli, no sign of continuously increasing of activities were found eliminating the possibility that premier expression was an orienting reflex. If that was the case, the increasing of muscle activity should be observed in responding to all stimuli. This suggested that the response for unpleasant stimuli, the initial emotional response, emerged by the evoked unpleasant affect. Furthermore, some would argue that the block design was more sufficient to evoke an emotion. However, the block design might cause unconscious bias since such a design uses continuous presentation of same category stimulus, which may also result in biased responses to the stimuli. Randomizing the presentation of stimuli was considered the best way to eliminate any bias (Efron, 1971).

In sum, the present study reveals that the premier expression is an emotional facial component that can be observed in the first 300ms after stimulus onset, is resistant to voluntary suppression, and is strongly linked with the negative valence. Considering the timing of its occurrence, its speed, and the fact that it occurs subsequent to ERPs, there is a high possibility that it is the first emotional output. From a broader perspective, being able to detect the first output of human emotion would be valuable for devising more accurate systems of detecting and predicting human internal emotions and behavior in artificial intelligence. Note that the present experiment could not discriminate the orientating reflex from emotional expression; in any social context, emotion is often evoked by physical stimuli, such as others' words, gestures, and behavior. In addition, to clarify whether the premier expression is merely a negative response or if it could be observed in responding to intense pleasant stimuli over the zygomaticus major muscle, further research is needed. Also the investigation of other facial muscles is important to confirm if there are specific patterns for all emotions.


Beauregard, M., Levesque, J., &_ Bourgouin, P. (2001). Neural correlates of conscious self-regulation of emotion. The Journal of Neuroscience, 2, RC165.

Begleiter, H., Gross, M.M., Kissin, B. (1967). Evoked cortical responses to affective visual stimuli. Psychophysiology, 3, 336-344.

Cacioppo, J. T., Petty, R. E., Losch, M. E., & Kim, H. S. (1986). Electromyographic activity over facial muscle regions can differentiate the valence and intensity of affective reactions. Journal of Personality and Social Psychology, 50, 260-268.

Cacioppo, J. T., & Berntson, G.G. (1994). Relationship between attitudes and evaluative space: A critical review, with emphasis on the separability of positive and negative substrates. Psychological Bulletin, 115, 401-423.

Cacioppo, J. T., Gardner, W. L., & Berntson, G.G. (1997). Beyond bipolar conceptualizations and measures: The case of attitudes and evaluative space. Personality and Social Psychology Review, 1, 3-25.

Carretie, L., Hinojosa, J.A., Martin-Loeches, M., Mercado, F., & Tapia, M., (2004a). Automatic attention to emotional stimuli: Neural correlates. Human Brain Mapping, 22, 290-299.

Carretie, L., Mercado, F., Hinojosa, J.A., Martin-Loeches, M., &Sotillo, M., (2004b). Valence-related vigilance biases in anxiety studied through event-related potentials. Journal of Affective Disorders, 78, 119-130.

Chapman, R. M., McCrary, J. W., Chapman, J. A., & Martin, J. K. (1980). Behavioral and neural analysis of connotative meaning: Word classes and rating scales. Brain and Language, 11, 319-339.

Cuthbert, B. N., Schupp, H. T., Bradley, M., Birbaumer, N., & Lang, P. J. (2000). Brain potentials in affective picture processing: Covariation with autonomic arousal and affective report. Biological Psychology, 52, 95-111.

Conroy, M.A. & Polich, J. (2007). Affective valence and P300 when stimulus arousal level is controlled. Cognition and Emotion, 21, 891-901.

Davidson, R. J., Ekman, P., Saron, C. D., Senulis, J.A., & Friesen, W. V. (1990). Approach-withdrawal and cerebral asymmetry: Emotional expression and brain physiology. I. Journal of Personality and Social Psychology, 58, 330-341.

Delplanque, S., Silvert, L., Hot, P., Rigoulot, S., & Sequeira, H. (2006a). Arousal and valence effects on event-related P3a and P3b during emotional categorization. International Journal of Psychophysiology, 60, 315-322.

Delplanque, S., Silvert, L., Hot, P., & Sequeira, H. (2006b). Attentional modulation of appraisal in emotion: the case of the P3b. Psychophysiology, 43, S7.

Dimberg, U. (1982). Facial reactions to facial expression. Psychophysiology, 19, 643-647.

Dimberg, U. (1990). Facial electromyography and the emotion reactions. (Award address for distinguished early career contribution to psychophysiology, 1988.). Psychophysiology, 27, 481-494.

Dimberg, U. (1997). Facial reactions: Rapidly evoked emotional responses. Journal of Psychophysiology, 11, 115-123.

Dimberg, U., & Thunberg, M. (1998). Rapid facial reactions to emotional facial expressions. Scandinavian Journal of Psychology, 39, 39-45.

Dimberg, U., Thunberg, M., & Elmehed, K. (2000). Unconscious facial reactions to emotional facial expressions. Psychological Science, 11, 86-89.

Donchin, E., & Coles, M. G. H. (1988). Is the P300 component a manifestation of context updating? Behavioral and Brain Sciences, 11, 357-374.

Efron, B. (1971). Forcing a sequential experiment to be balanced. Biometrika, 58, 403-417.

Ekman, P., & Friesen, W. V. (1986). A new pan-cultural facial expression of emotion. Motivation and Emotion, 10, 159-168.

Fridlund, A. J., & Cacioppo, J. T. (1986). Guidelines for human electromyographic research. Psychophysiology, 23, 567-589.

Gehring, J. W., Gratton, G., Coles, H. G. M., &Donchin, E. (1992). Probability effects on stimulus evaluation and response processes. Journal of Experimental Psychology, 18, 198-216.

Gianotti, L. R., Faber, P. L., Schuler, M., Pascual-Marqui, R. D., Kochi, K., & Lehmann, D. (2008). First valence, then arousal: the temporal dynamics of brain electric activity evoked by emotional stimuli. Brain Topography, 20, 143-156.

Herbert, C., Kissler. J., Junghofer, M., Peyk, P., & Rockstroh, B. (2006). Processing of emotional adjectives: Evidence from startle EMG and ERPs. Psychophysiology, 43, 197-206.

Ito, T. A., & Cacioppo, J. T. (in press). The psychophysiology of utility appraisals. In D. Kahneman, E. Diener, & N. Schwarts(Eds.). Understanding quality of life: Scientific perspectives on enjoyment and suffering. New York: Russell Sage Foundation.

Ito, T. A., Larsen, J. T., Smith, N. K., & Cacioppo, J. T. (1998). Negative information weighs more heavily on the brain: The negativity bias in evaluative categorizations. Journal of Personality and Social Psychology, 75, 887-900.

Keil, A., Bradley, M., Hauk, O., Rockstroh, B., Elbert, T., & Lang, P. J. (2002). Large-scale neural correlates of affective picture processing. Psychophysiology, 39, 641-649.

Kettunen, J., Ravaja, N., &_ Keltikangas-Jarvinen, L. (2000). Smoothing facilitates the detection of coupled responses in psychophysiological time series. Journal of Psychophysiology, 14, 1-10.

Lang, P. J., Greenwald, M. K., Bradley, M. M., & Hamm, A. O. (1993). Looking at pictures: Affective, facial, visceral, and behavioral reactions. Psychophysiology, 30, 261-273.

Lang, P.J., Bradley, M.M., & Cuthbert, B. N. (1999). International affective picture system (IAPS): Technical manual and affective ratings. Gainesville: University of Florida, Center for Research in Psychophysiology.

Larsen, J. T., Norris, C. J., & Cacioppo, J. T. (2003). Effects of positive and negative affect on electromyographic activity over the zygomaticus major and corrugatorsupercilii. Psychophysiology, 40, 776-785.

Levenson, W. R., Ekman, P., & Friesen, W.V. (1990). Voluntary facial action generates emotion-specific autonomic nervous system activity. Psychophysiology, 27, 363-384.

Levesque, J., Eugene, F., Joanette, Y., Paquette, V., Mensour, B., Beaudoin, G., Leroux, J. M., Bourgouin, P., & Beauregard, M. (2003). Neural circuitry underlying voluntary suppression of sadness. Biological Psychiatry, 53, 502-510.

Lundqvist, L. O., &_Dimberg, U. (1995). Facial expressions are contagious. Journal of Psychophysiology, 9(3), 203-211.

Mauss, B. I., Levenson, W. R., McCater, L., Wilhelm, H. F., & Gross, J. J. (2005). The tie that binds? Coherence among emotion experience, behavior and physiology. Emotion, 5, 175-190.

Matsumoto, D., LeRoux, J., Wilson-Cohn, C., Raroque, J., Kooken, K., Ekman, P., & Goh, A. (2000). A new test to measure emotion recognition ability: Matsumoto and Ekman's Japanese and Caucasian Brief Affect Recognition Test (JACBART). Journal of Nonverbal Behavior, 24, 179-209.

Naumann, E., Bartussek, D., Diedrich, O., & Laufer, M. E. (1992). Assessing cognitive and affective information processing functions of the brain by means of the late positive complex of the event-related potential. Journal of Psychophysiology, 6, 285-298.

Olofsson, J. K., Nordin, S., Sequeira, H, & Polish, J. (2008). Affective picture processing: An integrative review of ERP findings. Biological Psychiatry, 77, 247-265.

Rinn, W. E. (1984). The neuropsychology of facial expression: A review of the neurological and psychological mechanisms for producing facial expressions. Psychological Bulletin, 95, 52-77.

Rosenberg, E. L.,_& Ekman, P. (1995). Conceptual and methodological issues in the judgment of facial expressions of emotion. Motivation and Emotion, 19, 111-138.

Scherer, R. K., & Wallbott, G. H. (1994). Evidence for universality and cultural variation of differential emotion response patterning. Journal of Personality and Social Psychology, 66, 310-328.

Smith, N.K., Cacioppo, J. T., Larsen, J.T., & Chartrand, T. L. (2003). May I have your attention please: Electrocortical responses to positive and negative stimuli. Neuropsychologia, 41, 171-183.

Vrana, S. R. (1993). The psychophysiology of disgust: Differentiating negative emotional contexts with facial EMG. Psychophysiology, 30, 279-286.

Tomkins, S. S., & McCarter, R. (1964). What and where are the primary affects? Some evidence for a theory. Perceptual and Motor Skills, 18, 119-158.

Yartz, A. R., & Hawk, L. W. (2002). Addressing the specificity of affective startle modulation: Fear versus disgust. Biological Psychology, 59, 55-68.

Zhu, Y., & Suzuki, N. (2016). Your face is even more surprised than you are: Evidence from facial electromyographic activity exposing the genuine surprise. The North American Journal of Psychology, 18, 45-58.

Endnote (1) The word arousal is translated into Japanese as "[phrase omitted]" with the explanation that high arousal refers to the state that feels bodily or psychologically awaken, and low arousal refers to the state that feels sleepy.

Yinghan Zhu, Naoto Suzuki

Doshisha University

Author info: Correspondence should be sent to: Yinghan Zhu, Graduate School of Psychology, Doshisha University, Room316 Suzuki Office, 1-3 Miyakodani, Tatara, Kyotanabe, Kyoto 6100394, Japan.

Caption: FIGURE 1 EMG Response Change in Activities for the Corrugator Supercilii Muscle Plotted in the Interval of 300 Ms During First Second of Exposure. Error Bars Represent Standard Errors.

Caption: FIGURE 2 EMG Response Change in Activities for the Corrugator Supercilii Muscle Plotted in the Interval of 50 Ms During First 300Ms of Exposure. Error Bars Represent Standard Errors.

Caption: FIGURE 3 EMG Response Change in Activities for the Zygomaticus Major Muscle Plotted in the Interval of 300 Ms During First Second of Exposure. Error Bars Represent Standard Errors.

Caption: FIGURE 4 EMG Response Change in Activities for the Zygomaticus Major Muscle Plotted in the Interval of 50 Ms During First 300Ms of Exposure. Error Bars Represent Standard Errors.

Caption: FIGURE 5 Grand-average Event-related Potential Waveforms for Each Category 600Ms after Pictures Onset from Electrodes Fz.

Caption: FIGURE 6 Cross-correlograms between fEMG & ERPs Grand Mean Wave for 200ms Before & 1000ms After Unpleasant Stimulus in High-, middle-, & Low-arousal Levels.
TABLE 1 Characteristics of Picture Stimuli.

                            LAPS        Valence       Arousal
                            stimuli     rating        rating
                            number      M      SD     M      SD

High Arousal-Unpleasant     1111        1.47   0.85   6.74   2.28
                            1205        1.73   1.29   7.13   2.15
Middle Arousal-Unpleasant   9040        1.86   1.12   5.97   2.40
                            9300        1.35   0.75   5.97   2.51
Low Arousal-Unpleasant      9001        3.11   1.26   3.76   1.98
                            9220        3.43   1.32   4.25   1.94
High Arousal-Neutral        7640        4.23   1.58   6.98   1.77
                            1726        4.62   1.62   7.43   1.53
Middle Arousal-Neutral      pattern 1   4.82   1.61   4.32   1.77
                            pattern 2   4.88   1.45   4.59   1.63
Low Arousal-Neutral         7175        5.62   1.30   3.65   1.74
                            7010        5.15   1.42   3.59   1.63
High arousal-Pleasant       8185        7.12   1.38   7.03   1.74
                            8193        6.95   1.41   7.05   1.48
Middle Arousal-Pleasant     5199        7.65   1.16   4.83   1.94
                            5210        7.77   1.43   5.67   2.45
Low Arousal-Pleasant        1441        8.40   1.02   3.77   2.73
                            5800        7.11   1.67   3.78   2.13

TABLE 2 One sample t-test for Activities from the Corrugator
Supercilii Muscle during 1st 300ms after Stimuli Onset for each

Condition     Valence                     M       SD       t

CONTROL       High Arousal-Unpleasant     56.11   123.44   2.23
              Middle Arousal-Unpleasant   42.97   108.07   1.95
              Low Arousal-Unpleasant      39.42   92.14    2.1
SUPPRESSION   High Arousal- Unpleasant    52.38   77.75    3.5
              Middle Arousal-Unpleasant   40.26   92.76    2.26
              Low Arousal-Unpleasant      13.38   83.86    0.83

Condition     P      LL       UL

CONTROL       .036   3.98     108.23
              .064   -2.66    88.61
              .047   0.51     78.33
SUPPRESSION   .001   21.62    83.13
              .032   3.56     76.95
              .415   -19.79   46.553

Note. CI = confidence interal: LL = lower limit; UL = higher limit.

TABLE 3 Maximal Correlation Coeffecients between ERPs Grand Mean
Waves from Fz & fEMG Activities Over Corrugator Supercilii Muscle
for Unpleasant Stimuli.

Condition     High                        Middle
              arousal-Unpleasant          arousal-Unpleasant
              lag(ms)         r            lag(ms)           r

CONTROL       96             0.80            50             0.79
SUPPRESSION   104            0.85            134            0.82

Condition     Low
              lag(ms)           r

CONTROL       118             0.79
SUPPRESSION   114             0.77

[Please note: Some non-Latin characters were omitted from this article]

Please Note: Illustration(s) are not available due to copy right restrictions.
COPYRIGHT 2017 North American Journal of Psychology
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2017 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Zhu, Yinghan; Suzuki, Naoto
Publication:North American Journal of Psychology
Date:Dec 1, 2017
Previous Article:Development and Validation of a Persian Verbal Analogies Test Using the Rasch Model.
Next Article:"As Seen on TV": The Relationship between Media, Demographics, and Perceptions of Campus Safety.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters