Alison M. Harris¹, Chandlyr M. Denaro¹, Eric J. Moody², Catherine L. Reed¹; ¹Claremont McKenna College, ²University of Wyoming

The human ability to rapidly infer others’ mental states from their body movements has been hypothesized to depend on motoric representations, whereby observed actions are internally simulated within our own sensorimotor circuits. Supporting this idea, studies using electroencephalography (EEG) have reported reductions in the mu (8-14 Hz) and beta (15-20 Hz) rhythms over sensorimotor cortex both when executing one’s own movements and when observing the actions of others. Yet, this idea suggests that actions closer to one’s own motor repertoire should be easier to simulate, particularly for more idiosyncratic expressive movements that convey emotional information. To test this prediction, we measured behavioral performance and brain activity using high-density 128-channel EEG in participants (N =56) while they viewed point-light display (PLD) videos of their own movements vs. those of unfamiliar controls. Every participant first completed a motion capture session modeling three examples each of expressive movements (happy, angry, sad) and non-expressive neutral movements (marching, sidestep, trunk twist). Individual PLDs were then created using 3D animation software, removing any visual form cues that could be used to distinguish identity (e.g., height, body shape). In a subsequent testing session, participants were able to behaviorally discriminate PLDs of their own vs. others’ movements, but only for expressive movements (d’ >1). However, EEG analysis found no significant differences in the central mu rhythm for observation of expressive movements as a function of self vs. other. Instead, significant reductions were seen at frontocentral sensors in the beta band (15-17 Hz). These results support the idea that observers are differentially sensitive to their own motor repertoires, even without other identifying visual information. However, this self-other discrimination appears to occur beyond previously-reported central sites and alpha-band frequencies, suggesting the involvement of broader networks in the analysis of motor representations during action observation.

Acknowledgements: This research was supported by NSF BCS #1923178