Observing Action in Uncertainty: The Role of Belief-tracking in Action Observation

Recent research provides compelling evidence that our own motor system plays an important role in predicting the outcome of observed actions by others in our environment through a “mirroring system”. It has been argued that this ability to mirror the actions of our peers is the key to understanding others’ minds. Furthermore, an increasing number of studies have shown that humans spontaneously and unconsciously keep track of the mental states of other people. Recent theoretical work hypothesised that in implicit mental state computations the mirroring system and the mentalising system could interlock to predict future actions of agents.

In this study we propose a new paradigm that allows us to study both predictions based on motor cues and on mental state tracking simultaneously. Specifically, we looked at anticipatory looking while an agent reached for either a large or a small object, matching finger aperture to object size. Crucially, the belief of the agent on the location of this object could be true or false. Results of Experiment 1 show that participants used motor cues from the agent to predict where the agent would reach for the object. Moreover, we found evidence suggesting that participants also kept into account the beliefs of the agent on the location of the object, modulating the predictions resulting from motor cues. In Experiment 2, we used motor restriction to tap into the role of motor representations in these predictions. While results from Experiment 1 were partially replicated, we found no effect of motor restriction. In Experiment 3 we investigated the relationship between overt and covert measures of mentalising as well as the impact of additional task load on both processes. While overt and covert measures were highly correlated, both measures proved robust to secondary task load. The overall results deviate from previous experiments and these differences are potentially attributable to individual differences.

Overall, the results provide tentative evidence for belief-based modulation of action anticipation based on motor cues. However, several confounds were identified and discussed. Finally, it is discussed how this paradigm can be refined and tested for future implementations, potentially paving the way for new research on the role of motor processes in social cognition.