Abstract:

Social interactions require individuals to take the perspectives of others and recognize that the views and beliefs of others are sometimes different from their own, these usually happen implicitly, concurrently with the processing of the self-perspective. Previous studies suggested that individuals obtain the mental state of others by imagining that they are in the body of others. This embodied experience when taking others’ perspective arises from the neural mechanisms of shared self and other representations. An interesting but unanswered question is, if self-processing shares neural representations with other-processing, how individuals simultaneously process the self and others’ perspectives in a social interaction. A possible hypothesis is that the brain processes these two types of information alternately and periodically, and when the alternation is fast enough, the two processes can appear to be simultaneous, and formulate behavioral oscillations, similar to that found in other cognitive processes such as attention and priming. Therefore, with three experiments, the current study explored whether there is a behavioral oscillation when individuals process information from the self-perspective and another person’s perspective.

This study adopted a novel face-orientation judgement task to detect spontaneous visual perspective taking (VPT) and its behavioral oscillations. In this task, A top view image of an agent was presented, followed shortly by a central face of different orientations, briefly presented also on the computer screen. Participants were asked to make a quick judgement on the orientation of the central face presented. The accuracy of the judgment was used as the indicator of the participants’ performance. Experiment 1 tested whether participants recognize better the upright faces under the perspective of the agent (i.e. another person), due to spontaneous VPT. Experiment 1 used a 2 (agent orientation) × 3 (face orientation) within-subject design. The agent orientation had two levels: face to the central face and back to the central face. The face orientation had three levels: an upright face under the self-perspective (US), an upright face under another’s perspective (UO), and a non-upright face under both the self- and other- perspectives (NB). Figure 1 is an illustration of the procedure of Experiment 1. Experiment 2 used two face orientations (US vs. UO) with the agent-face SOA, the elapsed time between the appearance of the agent and the face, ranging 0.2~1.2 s across 30 equidistant conditions to explore whether the participants’ face recognition showed behavioral oscillations over time. Figure 2 is an illustration of the procedure of Experiment 2. Experiment 3 extended the agent-face SOA (0.2~2.2 s) to observe the behavioral oscillation over a longer period time.

In Experiment 1, the main effect of agent orientation was not significant, F (1, 22) = 3.84, p = 0.063, the main effect of face orientation was significant, F (2, 21) = 49.50, p < 0.001, η_p² = 0.83, participants’ judgement was the best for the US faces (M = 0.98, SE = 0.01), followed by the UO faces (M = 0.80, SE = 0.02), and then the NB faces (M = 0.76, SE = 0.02), all paired comparisons were significant (US vs. UO: p < 0.001; US vs. NB: p < 0.001; UO vs. NB: p = 0.012). More important, the interaction between the agent orientation and the face orientation was significant, F (2, 21) = 8.30, p = 0.002, η_p² = 0.44. When the agent faced to the central face, participants’ judgement was better for the UO faces (M = 0.83, SE = 0.02) compared to the NB faces (M = 0.76, SE = 0.02), SE = 0.02, p < 0.001; but when the agent was back to the central face, there was no significant difference between UO faces (M = 0.77, SE = 0.02) and NB faces (M = 0.77, SE = 0.03), SE = 0.02, p = 1. The results of Experiment 1 were shown in Figure 3. In Experiment 2, the Fourier transform and permutation tests showed that the accuracy rate of UO face as a function of agent-face SOA had an oscillation frequency of about 1 Hz. Paired t-tests on normalized curves found that the judgement of US faces was better than UO faces in the SOA of 0.2~0.5 s, but a revered pattern was found for the 0.7~1 s interval. The results of Experiment 2 were shown in Figure 4. In Experiment 3, by using the same statistical method as in Experiment 2, the results showed that both the accuracy rate of US faces and UO faces had an oscillation frequency of about 1 Hz. The distribution of the phase difference between the two curves of each participant in the range of 0.5~1.5 Hz did not follow a uniform distribution (using Rayleigh test, p = 0.040), but concentrated around 180 degree (M = 186.96, SD = 73.98), indicating that the enhanced processing in another’s perspective is accompanied by the weakened processing in the self-perspective, and vice versa. The results of Experiment 3 were shown in Figure 5.

We found that individuals showed behavioral oscillations with a frequency of about 1 Hz when processing information in the self-perspective and another’s perspective. These results help us understand how individuals process information about themselves and others in social interactions: individuals will focus their cognitive resources only on one perspective at a time, rather than allocating between the two perspectives at the same time. Through the rapid switching between the two perspectives, the simultaneous processing of the two can be achieved, and the possible confusion caused by the shared-representation mechanism can be effectively avoided. To our knowledge, these findings demonstrate for the first-time evidence of behavioral oscillations in social cognition, further illustrating the pervasiveness of behavioral oscillations in cognitive processing.

Key words: implicit social cognition, visual perspective taking, self-perspective, others’ perspective, behavioral oscillation