Short Communication
The time course of face processing: startle eyeblink response modulation by face gender and expression

https://doi.org/10.1016/j.ijpsycho.2013.08.006Get rights and content

Abstract

The purpose of this study was to investigate the effects of both facial expression and face gender on startle eyeblink response patterns at varying lead intervals (300, 800, and 3500 ms) indicative of attentional and emotional processes. We aimed to determine whether responses to affective faces map onto the Defense Cascade Model (Lang et al., 1997) to better understand the stages of processing during affective face viewing. At 300 ms, there was an interaction between face expression and face gender with female happy and neutral faces and male angry faces producing inhibited startle. At 3500 ms, there was a trend for facilitated startle during angry compared to neutral faces. These findings suggest that affective expressions are perceived differently in male and female faces, especially at short lead intervals. Future studies investigating face processing should take both face gender and expression into account.

Introduction

Affective faces can have profound effects on cognition and emotion, drawing more attention than neutral faces (e.g. Eimer et al., 2003) and inducing emotional states (e.g. Anokhin and Golosheykin, 2010). The Defense Cascade Model (Lang et al., 1997) has been developed around the effects of emotional visual scenes, but has not yet been examined with affective faces. According to this model, an early orienting response occurs between 30 and 1000 ms following picture onset and engages attention to facilitate stimulus processing, helping the organism gather important sensory information. Then defensive or appetitive activation takes over between 1000 and 6000 ms, resulting in a more sustained emotional state that may lead to goal-directed behavior.

It has been postulated that a variety of stimulus characteristics, including attention and emotional qualities, affect processing at different time points, which can be mapped onto the defense cascade model. Multiple studies (Bradley et al., 1993, Bradley et al., 2006, Cuthbert et al., 1996) eliciting startle at varying lead intervals found that startle responses elicited at short lead intervals (50–1000 ms) were primarily modulated by stimulus-induced arousal (smaller responses elicited by more emotionally arousing stimuli, compared to neutral stimuli), while startle responses elicited at long lead intervals (1000–6000 ms) were primarily modulated by stimulus valence (larger responses when presented with negative stimuli compared to neutral and positive stimuli). During the short lead interval time window (50–1000 ms), stimuli that are made more salient through increased attention produce greater startle response inhibition compared to those that are ignored (DelPezzo and Hoffman, 1980). This phenomenon has been termed Prepulse Inhibition (PPI), a well established measure of sensorimotor gating and attention (Blumenthal, 1999), whereby the early processing of the prepulse (in this case, the affective stimulus) is protected at the expense of fully processing the startle eliciting stimulus, resulting in startle response inhibition. Studies reporting greater PPI during emotional compared to neutral scenes suggest that more attention was allocated to the emotional scenes compared to the neutral scenes (e.g. Bradley et al., 1993).

Startle responses elicited during the long lead interval time window (1000–6000 ms) are largest while viewing negative pictures compared to neutral and positive pictures (Bradley et al., 1993, Bradley et al., 2006, Cuthbert et al., 1996, Dichter et al., 2002). This phenomenon has been termed Affective Modulation of Startle (AMS), and can be explained within the context of the Motivational Priming Hypothesis (Lang et al., 1998). This hypothesis suggests that the pairing of emotional states with emotionally valenced reflexes related to approach (e.g. salivation) or avoidance (e.g. startle reflex) can facilitate or inhibit the reflex. Specific to AMS, eliciting an avoidance based reflex (startle) during an induced negative emotional state via affective images, will result in facilitation of the startle response.

While the pattern of AMS induced by emotional scenes is reliable, the pattern of AMS induced by affective faces is less consistent. Many studies report the typical pattern of AMS (Anokhin and Golosheykin, 2010, Balaban, 1995, Dunning et al., 2010), although some observed facilitation only in response to anger but not fear (Springer et al., 2007), while others reported facilitation in response to fearful faces in threatening contexts (Grillon and Charney, 2011). Moreover, some studies find no AMS in response to affective faces (Spangler et al., 2001). One source of inconsistency in the literature examining startle response modulation by affective faces lies in the lead intervals at which startle responses are elicited. Previous studies probed only during the long lead interval time window when the startle response is typically modulated by emotional valence, and did not examine startle response modulation at short lead intervals typically associated with attentional processes. Additional sources of variability in studies using affective faces are face gender (Hess et al., 2007) and participant gender (Anokhin & Golosheykin, 2010), variables very few studies take into account. Previous studies found AMS in response to male but not female faces (Hess et al., 2007), and stronger startle response modulation in female compared to male participants (Anokhin & Golosheykin, 2010). These findings are consistent with studies reporting that participants associate male faces with negative emotions and female faces with positive emotions (Becker et al., 2007).

The present study had two aims. First, we aimed to address the variability between studies examining startle modulation by affective faces, using the acoustic startle response to investigate both attentional and emotional reactions to faces expressing negative (angry), positive (happy) and neutral affect. To our knowledge, this study is the first to use the startle response to simultaneously measure startle response modulation by affective faces during both short and long lead intervals. Second, we aimed to investigate how face gender may interact with facial expression to further modulate attentional and emotional reactions.

We predicted that affective faces would demand greater attentional resources than neutral faces and thus, show significant PPI at the short (300 ms) lead interval. In line with previous findings (e.g. Bradley et al., 2006), we predicted that by 800 ms there would be no startle response modulation. We predicted that at the long (3500 ms) lead interval, startle responses would reflect the expected pattern of AMS, showing facilitation during angry compared to happy and neutral faces. In addition, we expected face gender to interact with facial expression, impacting the patterns of attention and emotion modulation observed at the 300 and 3500 ms lead intervals, respectively. Given that very few studies have examined the effect of face gender on startle response modulation, this portion of the analyses was exploratory in nature.

Section snippets

Participants

Thirty-one undergraduate students ranging in age from 18 to 26 years (M = 19.87, SD = 2.46) completed the study. Nine (29%) were male and 22 (71%) were female. Eighteen (58.1%) self-identified as Caucasian, 4 (12.9%) as African American, 3 (9.7%) as Asian, 4 (12.9%) as Hispanic, and 2 (6.5%) as other.

Stimuli

Face stimuli were drawn from the Japanese and Caucasian Facial Expressions of Emotion and Neutral Faces (Matsumoto and Ekman, 1988), containing 56 photos that have been tested in the United States,

Valence ratings

A 3 (Facial Expression: Happy, Neutral, Angry) × 2 (Face Gender: Male, Female) repeated measures ANOVA4 on self reported valence ratings yielded a significant main effect of Face Gender, F(1, 22) = 12.06, p < .01, η2p = .35, with female faces rated as more positive overall (M = 5.3) than male faces (M = 4.9). As expected, there was also a main

Discussion

The aim of the present study was to examine the effects of Facial Expression and Face Gender on early and late stages of processing, using startle eyeblink responses at lead intervals indicative of attention allocation and emotional state. Concordant with the Defense Cascade Model (Lang et al., 1997), affective faces modulated startle eyeblink response during both early and late stages of processing, and Face Gender interacted with face expression at short lead intervals.

Valence ratings

Acknowledgments

The authors thank Tamera Murdock, Ph.D. for her assistance with data analysis, Cary Savage, Ph.D. for methodological guidance, and the undergraduate research participants. This study was supported in part by the University of Missouri–Kansas City Chancellor's Doctoral Fellowship (ERD).

References (22)

  • M.M. Bradley et al.

    Pictures as prepulse: attention and emotion in startle modification

    Psychophysiology

    (1993)
  • Cited by (13)

    • Attention allocation to subliminally presented affective faces in high and low social anxiety

      2020, International Journal of Psychophysiology
      Citation Excerpt :

      When startle responses were elicited at 800 ms, we predicted no differences in PPI with respect to target face expression, indicating attention disengagement from all faces. To determine sample size, a priori power analysis was conducted using data previously collected in our lab (Duval et al., 2013b), using startle eyeblink measures to assess response to affective faces. The effect size for the face gender by face expression interaction in healthy participants was η2p = .17.

    • Startle eye-blink modulation by facial self-resemblance and current mood

      2015, International Journal of Psychophysiology
      Citation Excerpt :

      It was also chosen because this latency corresponds roughly to the time course of face processing as described in previous studies (with the N250 being the first electrophysiological index of recognizing familiar faces; e.g., Schweinberger et al., 2004). Moreover, to exclude the possibility that rapid processing of emotional cues (such as facial expressions) would confound startle responses both at long and short onset delays (Duval et al., 2013), special care was taken only to include stimulus materials as affectively neutral as possible (i.e., neutral facial expression, no hair or clothing, etc.). For exploratory reasons, we also tested the hypothesis that exteroceptive perception of self-referential stimuli might be facilitated by concurrent interoceptive signals, because previous research had indicated a close interplay of external and internal self-referential processing, for example, in the case of multisensory (visuo-tactile) stimulation potentially inducing alterations in the threshold for visual self-recognition (Tsakiris, 2008; Sforza et al., 2010).

    View all citing articles on Scopus
    View full text