L2-Minimum-Norm estimations of the location of underlying dipole sources were calculated as described by Hauk and colleagues (2002). The calculation was based on a spherical isotropic volume conductor head model with four shells and 3 (radial, azimuthal, and polar direction) x 197 evenly and spherically distributed dipoles serving as a source model. As a trade-off between depth sensitivity and spatial resolution, a source shell radius of 6 cm was chosen.
The localization of the difference waves comparing the emotional with the neutral gesture categories in the tested time window revealed sources over occipital areas (see Fig. 3.8). For the pleasant gesture, the pattern of source generators was restricted mainly to the medial part of the occipital cortex (see Fig. 3.8, left). In contrast, underlying sources of the unpleasant middle finger category were characterized by higher estimation values and were more widespread over occipital areas, reaching farther into temporo-occipital regions along the ventral visual stream (see Fig. 3.8, right).
Figure 3.8
Right side view of the calculation of the L2-Minimum-Norm estimate in the tested time interval for the difference waves comparing emotional and neutral gestures.
Please note the different scales for each comparison.
Discussion
Extensive research has determined that emotionally salient pictures are selectively processed by the brain. As shown in numerous studies, images of emotional naturalistic scenes, facial expressions and body language elicit increased neuronal activation in extrastriate cortex and affective cortical and sub-cortical structures (e.g., Bradley et al., 2003; de Gelder et al., 2004; Junghöfer et al., 2005;
Junghöfer et al., 2006; Sabatinelli et al., 2004; Sabatinelli et al., 2005; Vuilleumier et al., 2001). In a recent brain imaging study, we extended this knowledge by demonstrating corresponding BOLD activations when passively viewing emotional as compared to affectively neutral expressive hand gestures (Flaisch et al., in preparation).
Providing a temporally more detailed specification of the underlying processes, studies utilizing event-related brain potentials and conventional affective stimulus materials have identified two distinct components associated with emotion processing. The first of those components is observed at around 150 ms post-stimulus over posterior sites (EPN; e.g., Junghöfer et al., 2001), whereas the relatively later one occurs at around 400-600 ms over centro-parietal regions (LPP; e.g., Schupp et al., 2003b). Following up on this research, we tested in the present study whether the processing of emotional hand gestures is also accompanied by the acquainted electro-cortical signature, in this case specifically the EPN.
Utilizing a RSVP paradigm, we presented one pleasant, one unpleasant and one neutral hand gesture to the participants while measuring high-density EEG. In accordance with a priori assumptions, we observed a more negative going deflection of the event-related potential over posterior leads while watching emotional as compared to neutral hand gestures. This differentiation began at around 150 ms after the onset of the stimulus and lasted to the end of the picture epoch. Relating to the results of the previous imaging study (Flaisch et al., in preparation), this negativity was much more pronounced for the insulting middle finger gesture than for the pleasant thumbs-up category. Further correspondences to the found BOLD activation patterns were revealed in terms of the topographic distribution of these effects. First, both thumbs and to an even larger degree middle fingers were associated with a stronger EPN on the right than on the left side. And second, source localization estimates suggest that the generators of the EPN for both emotional categories are
situated in occipital cortex. This pattern of generator sites also corresponds well to reported source locations when watching emotional naturalistic scenes (Junghöfer et al., 2001; Schupp et al., 2006). Moreover, replicating earlier behavioral results regarding the valence evaluation (Flaisch et al., in preparation), subjects rated the thumbs-up gesture as significantly more pleasant than the other categories, whereas the insulting middle finger gesture was rated as maximally unpleasant. Fully confirming previous findings with respect to the gestures’ respective arousal (Flaisch et al., in preparation), only the unpleasant middle-finger category was rated as more arousing than the others.
The interpretation of the observed effects as a result of motivational relevance gains credibility considering the pictures’ respective valence and arousal ratings. Whereas all categories were clearly differentiated from each other in terms of valence, only the insulting middle-finger gesture was perceived as being more arousing. In good correspondence with these findings, the unpleasant gesture was associated with the largest EPN amplitude. This relates both to the previous brain imaging study also showing maximal BOLD activation specifically for this gesture (Flaisch et al., in preparation), as well as to studies using naturalistic scenes demonstrating maximally pronounced neuronal responses as a function of the examined stimuli’s arousal level (Bradley et al., 2003; Junghöfer et al., 2001;
Junghöfer et al., 2005).
The observation of maximal neural responsivity to the insulting gesture is furthermore in line with several studies examining facial affect consistently reporting maximal modulation for angry faces, also signaling potential threat and immediate danger (Pourtois et al., 2006; Schupp et al., 2004c). Moreover, utilizing psychophysical measures it was shown by several studies that anger is the emotion that is most reliably decoded from stimuli depicting dance or gesture (Boone &
Cunningham, 1998; Dittrich et al., 1996; Pollick et al., 2001; Pollick et al., 2002).
Finally, studies examining emotional body language consistently report maximal neuronal activation in response to fearful body postures (de Gelder et al., 2004;
Hadjikhani & de Gelder, 2003). Therefore, it is suggested by existing evidence that the detection of angry or threatening non-verbal signals may have a special role in interpersonal interaction.
In the past, the question has been raised as to what degree observed emotion effects may not be a function of the emotional meaning of the used stimuli, but rather be a result of systematic differences between the various picture categories in terms of physical features such as complexity or brightness of the images. In an earlier study, Junghöfer and colleagues (2001) met these concerns by statistically controlling for a large number of candidate physical characteristics. In their analyses, they did not find any physical parameter accounting for the found ERP modulations. In contrast, the pictures’ value on the arousal dimension proved to be the most reliable predictor of the associated EPN. The present results may provide further corroboration of this interpretation. Even though not finally conclusive in this respect, the great physical similarity between the different gesture categories may be regarded as highly suggestive towards this end. In terms of complexity, all picture categories depicted the back of a hand with a single extended finger. Moreover, all categories showed the hands of the same models, in each case on top of the identical background color.
However, since the highly arousing middle finger gesture depicted a vertically oriented shape whereas both other categories were characterized by horizontal orientations, emotional significance may be confounded with orientation and occurrence frequency. Therefore, further research is needed to control for possible effects of these parameters.
Together with the previous fMRI study (Flaisch et al., in preparation), the present results provide converging evidence that highly symbolic hand gestures bearing emotional meaning are not only processed differently by the brain compared with neutral ones but rather that this differential processing taps on the same neurophysiological mechanisms as conventional emotional picture stimuli. However, in future studies it needs to be determined whether this notion is expendable to all aspects of emotion processing. It still remains an open question e.g., whether later neuronal processes indexed by the LPP are also sensitive to the affective quality of this stimulus class. Also, it is imperative to examine in the behavioral domain whether emotional hand gestures indeed automatically draw attention and whether this increased attention allocation directed by affectively salient gestures is reflected by response advantages as frequently reported in studies utilizing affective facial expressions (e.g., Mogg & Bradley, 1999; Öhman et al., 2001a; Öhman et al., 2001b;
Pessoa et al., 2002b; Vuilleumier, 2002).
In summary, the present results demonstrate the rapid discrimination of emotional hand gestures in early visual cortex. They suggest that the neural processes sensitive to the inherent emotional significance of environmental stimuli is not exclusively dependent of the evolutionary preparedness of these cues. Rather, they indicate that the eliciting emotional significance may get attached to stimuli whose meaning is of a highly symbolic, yet evolutionarily unspecific nature.