174 |
Georg W. Alpers and Antje B.M. Gerdes |
|
|
4.2. Dominance of Emotional Facial Expressions
In order to further pursue the question of preferential perception of emotional stimuli with the aid of binocular rivalry, and stimuli which are better controlled for physical differences, we conducted an experiment with pictures of different facial expressions (Alpers and Gerdes, 2007). Presenting emotional faces is especially useful in emotion research because these stimuli are evolutionary relevant (Dimberg, et al., 2000), largely independent from culture (Ekman, Sorenson, and Friesen, 1969) and because they provoke emotional reactions in every day life (Dimberg, 1982).
Emotional facial expressions are processed very rapidly (Jeffreys, 1996) and holistically (Farah, Wilson, Drain, and Tanaka, 1998) via a specialized subcortical route (Johnson, 2005). Emotional faces attract attention as can be seen in visual search and Dot-Probe paradigms (Mogg and Bradley, 2002). They elicit subcortical activation as well as peripheral physiological and behavioral reactions (Whalen, Kagan, Cook, Davis, Kim, Polis et al., 2004; Whalen, Rauch, Etcoff, McInerney, Lee, and Jenike, 1998) even when they are masked and thus not consciously perceived.
For our study on binocular rivalry we chose pictures of eight actresses, each of them showing angry, frightened, happy, surprised and neutral facial expressions, from a standardized set of pictures (Karolinska Directed Emotional Faces, KDEF, Lundqvist, Flykt, and Öhman, 1998). These pictures are very well standardized regarding physical characteristics such as background and brightness. Pair wise presentation of one neutral and one emotional facial expression of the same actress made it possible for us to adequately control for inter-individual differences of the actresses.
During the stereoscopic presentation the 30 participants continuously coded their perception of emotional, neutral or mixed pictures by button presses. Again, there was clear evidence for predominance of emotional stimuli compared to neutral stimuli for both, the cumulative duration with which each percept was seen as well as the initial percept seen during each trial (see Figure 4).
Emotional faces were consciously perceived by the participants significantly longer throughout the trial and they were significantly more often perceived as the first clear percept of a trial. Different from our expectation, there were no differences between positive and negative facial expressions. These findings support our earlier findings of predominance of emotional pictures over neutral pictures in binocular rivalry. In this study we were able to take into account potential limitations of the first study. Predominance of emotional faces is equally
Temporarily Blind in One Eye |
175 |
|
|
strong as that of emotional IAPS pictures, although the latter normally induce a higher arousal than emotional faces (see Adolph, Alpers, and Pauli, 2006).
Figure 4. Results from the experiment with emotional faces (Alpers and Gerdes, 2007; with permission from APA). Left panel: Cumulative duration of dominant perception (in seconds) and standard errors of the mean, for emotional, neutral and mixed pictures. Right panel: Average frequency of initial perception and standard errors of emotional, neutral and mixed pictures.
As participants did not have to verbalize what they perceived and the categorical classification of perceived facial expressions was easy (emotional vs. neutral), the likelihood of response biases was clearly reduced in this study. Nonetheless, coding of participants' perception was still based on self-report. Thus, biases can not be completely ruled out.
4.3. Inter-Individual Differences: Phobic Stimuli
After having demonstrated that emotional pictures are in fact dominant over neutral pictures, we addressed another early hypothesis of the binocular rivalry literature: Are inter-individual differences reflected in what people perceive in binocular rivalry?
With the help of a further improved experimental design, we investigated whether fear relevant stimuli are perceived as more dominant by individuals who
176 |
Georg W. Alpers and Antje B.M. Gerdes |
|
|
are characterized by high levels of high fear. We recruited patients with a specific phobia of spiders because they are especially well suited for this investigation. The presentation of phobia-related stimuli elicits strong emotional reactions (Globisch, Hamm, Esteves, and Öhman, 1999). That phobic material activates subcortical networks which may in turn prime or boost the activity of the visual cortex, has been documented in a stronger amygdala activation in response to phobic cues (e.g., Alpers, Gerdes, Lagarie, Tabbert, Vaitl, and Stark, submitted). Remarkably, possible influences of mere physical differences of pictures (emotional versus neutral) are not problematic in this endeavor because they should affect phobic and non-phobic participants in equal measure.
Different degrees of dominance between patient and control participants would also support the theory that phobia-related cues are processed more intensely in phobic participants. A group of 23 patients who met diagnostic criteria for spider phobia (DSM-IV, American Psychiatric Association, 1994) and 20 non-phobic control participants were recruited for this investigation (Alpers and Gerdes, 2006).
Twenty different pictures of spiders and flowers were presented stereoscopically. Different from previous studies all of these pictures were paired with an abstract pattern,. We hoped that this would minimize the problems related to response biases and to interindividual differences in decision criteria. These picture-pattern pairs were presented for 8 sec each (see Figure 5).
Participants were asked to continuously code their perceptual impression by pressing one of three different buttons. There was one button each for dominant perceptual impression of a picture (spider and flower), one for the abstract pattern, and and one for mixed percepts. The advantages of this approach were that the specific content of a picture did not have to be identified and that no decision between two semantic pictures was needed.
Figure 5. Stimulus material for the experiment with phobic patients (Alpers and Gerdes, 2006): examples of a spider picture, the abstract pattern, and a flower.
Temporarily Blind in One Eye |
177 |
|
|
Figure 6. Results from the experiment with phobic patients; Left panel: mean duration of dominant perception (with standard errors) of pictures of spiders, patterns and mixed pictures (crossed bars: phobic; empty bars: control participants). Right panel: mean duration of dominant perception (with standard errors) of pictures of flowers, patterns and mixed pictures (crossed bars: phobic; empty bars: control participants).
Spider phobic participants perceived pictures of spiders as dominant for longer periods of time during the trials and they also reported that they perceived phobic pictures as the first clear percept of a trial more often than control participants. At the same time, there are no group differences for pictures of flowers versus the pattern; groups did not differ in the duration with which they perceived pictures of flowers as dominant or in how often they reported seeing flowers as the first percept in a trial (see Figure 6).
This study replicates our previous findings in showing that emotional pictures modulate competition in binocular rivalry. In addition, we were able to demonstrate that personal relevance is reflected in dominance of specific phobiarelated cues. We were also able to support the theoretically founded assumption that phobia-related cues are preferentially processed by individuals with spider phobia.
When we compared negatively valenced and positively valenced pictures in earlier studies, no differences were apparent. With respect to interindividual differences, we have no data concerning personal relevance of positive pictures at this point.