Page 57 - Demo

Attentional Biases to Facial Emotions552Limitations & Future DirectionsWhile some previous research did not report an attentional bias toward (certain) emotional facial expressions (Puls & Rothermund, 2018; Valk et al., 2015), we found an attentional bias toward emotions for all emotion categories. This bias did not differ significantly between the expressions and ranged around 10 ms (6.37 ms %u2013 13.30 ms) which is comparable with previous research (e.g., Monk et al., 2010). Thus, emotional facial expressions seem to automatically receive prioritised attention, highlighting their suggested communicative function as salient signals for conspecifics. Yet, caution has to be taken in the interpretation of results on attentional biases, as their appearance/significance might depend on various additional factors. One important factor that has been raised in the debate on the validity of attentional biases is the impact of low-level features (e.g., de Cesarei & Codispoti, 2013). Differences between stimulus categories regarding features such as spatial frequency are a general problem in interpreting results from dotprobe studies. We did not control our stimuli for these features to keep them as natural as possible and therefore cannot exclude this as a potential influence. Nevertheless, we compared our emotion categories with regard to specific lowlevel features (Stuit et al., 2021, see Methods section) and could not identify any significant differences. While there were no systematic differences between emotion categories, this comparison does not rule out the existence of differences in, for example, spatial frequency or colour distribution between distinct stimuli, which might have added noise to the data. Future studies investigating attentional processes in the field of emotion should try to account for those. Another limitation of our study is that we did not collect data in clinical populations and could, therefore, only investigate the influence of trait levels associated with ASD and SAD on attentional biases. This complicates the direct comparison to previous research with clinical samples as well as the formulations of potential implications for clinical practice. Yet, the view of a %u201ccontinuum of impairment%u201d has become more popular regarding both social anxiety (Rapee & Spence, 2004) and autism (Robinson et al., 2011), with disorders lying on the extremes of clinical traits in the general population. While clinical relevance of symptoms is eventually determined by difficulties in daily life, alterations in information processing, as well as their underlying mechanisms, may not be qualitatively different along the trait dimension. Thus, our findings in heightened trait levels may also be informative for clinical populations.