The social setting is an extremely complex environment in which humans are embedded. In order to deal with its complexity, humans have developed numerous cognitive and perceptual mechanisms that facilitate processing of relevant information, while ignoring what is irrelevant for particular situations at hand. Establishing a common social context with an interaction partner is one of the mechanisms that allow for narrowing the environment to the setting that is relevant for current interaction. One fundamental mechanism for establishing a common social context is joint attention – directing attention to where interaction partner is attending. Joint attention can be achieved in numerous ways, such as gaze direction, pointing gestures or verbal communication, with gaze direction being perhaps the most rudimentary mechanism (see, e.g., Baron Cohen, 1995). The mechanism of gaze-guided attention has been investigated with the use of a gaze-cueing paradigm (e.g., Friesen & Kingstone, 1998), in which participants observe a face gazing in some direction, and then are asked to detect/discriminate a stimulus presented either in the gazed-at (cued) location or elsewhere. Typically, participants exhibit better performance related to stimuli presented in gazed-at location, relative to other locations.
Attending to gaze direction has been traditionally assumed to be a bottom-up process which is used by the brain as a social signal to make inferences about another's intentions and mental states (for a review: Nummenmaa & Calder, 2009). However, Teufel, Fletcher and Davis (2010) questioned this assumption by proposing that even the most basic perceptual processing of social stimuli is top-down modulated by mental state attribution. We investigated the influence of mental state attribution on social attention by manipulating participants’ beliefs about the mental states of the observed gaze cue providers (Wiese, Wykowska, Zwickel, & Müller, in prep.). In particular, we presented human and robot faces in a gaze cueing paradigm and manipulated the instruction participants received. When participants attributed mental states to the observed faces, the gaze cueing effect was significantly larger as compared to conditions when mental state attribution was less likely. Interestingly, this effect was independent of whether a human or a robot face was presented, and depended only on whether participants believed that the stimuli had mental states.
Using the same gaze cueing paradigm and the EEG/ERP methodology, we investigated whether this differential effect occurs at the stage of perceptual/attentional processing or is rather due to response biases (Wykowska, Wiese, & Müller, in prep.). The ERP results showed a modulation of P1 (an early attention-related component) by the type of cue provider. That is, human faces elicited more enhanced P1 for validly cued trials as compared to invalid trials. No such effect was observed for the robot faces in the P1 time window. Interestingly, the response-related LRP component did not show any modulation by cue type or cue validity. Taken together, the present results suggest that already early attentional stages of processing might be influenced by such high-level mechanisms like mental state attribution.
Modulation of the early P1 component by mental state attribution. Validity effect was observed for the human faces but not for the robot faces (Wykowska, Wiese, & Müller, in prep.)
References:
