Behavioral similarity likely drives interbrain neural synchrony

Hyperscanning methods where brain imaging data are acquired from two or more people simultaneously are increasingly being used in the field of social neuroscience. However, the interpretation of these data can be controversial. A recent paper (1) makes some bold parallels between interpersonal neural synchrony and quantum entanglement, applying entanglement analyses to a dataset where two participants conduct a memory task together. However, there are important behavioral features that likely drive the observed interpersonal neural synchrony (INS), but have been overlooked in the interpretation of results.

In any hyperscanning study, the effects observed depend strongly on the task participants engage in. Ecologically valid paradigms allow participants to communicate freely via speech, gaze, and movement thus capturing the delicate processes of mutual prediction (2). In contrast, minimal paradigms use computer-mediated tasks where any coordination between brains is likely to be related to the strategic and cognitive requirements of the task. Yang’s experiment represents an extreme of minimalism. On each trial, participants see a noun and press a button to classify it as “animate/inanimate.” On collaborative trials (Collab), they must coordinate their reaction time with the other participant; feedback after each trial allows them to converge. Using this feedback, participants adopt a strategy of “fast and consistent” reaction times in the Collab trials, and these trials are faster than trials where they respond individually and without coordination (Indep). Similar strategies are seen in other minimal social tasks (3).

This strategy is likely to have major implications for the neural responses contributing to the observed INS. Fast and consistent behavior on the Collab trials is likely linked to tightly aligned event-related desynchronization (ERD) time-series. In contrast, the Indep trials are both slower and more variable between participants, likely evoking temporally misaligned ERD. This has been established in previous work: behavioral synchrony drives neural synchrony (4, 5) and when participants coordinate the timing of button presses, intersubject neural alignment of alpha and beta ERD signals is observed across all participants producing this behavior, irrespective of whether they interact with each other or not (6). Critically, Yang’s paper does not include an analysis of pseudopairs (participants tested on different days that do not directly interact with each other), or a stricter analysis of RT-matched pseudotrials. If motor behavior drives the observed effects, then INS in real trials will not differ from pseudotrials. This would mean the INS does not reflect an interpersonal mechanism.

More broadly, this paper highlights the challenges of trying to study social interaction with minimal paradigms. By imposing tight experimental control and minimizing the transmission of information between participants, this experimental design may eliminate the very interactivity that drives many INS effects. We suggest it would be more helpful for the field to recognize that INS is closely related to shared sensory, motor, semantic, and task processing between participants and to take this shared processing itself as an object of study relevant to the understanding of real-world social interaction.

Data, Materials, and Software Availability

There are no data underlying this work.