It may not be "we smile because we're happy," but "we're happy because we smile" — RIKEN clarifies the brain mechanisms involved in how facial expressions shape emotions

A joint research group led by Team Director Wataru Sato of the Psychological Process Research Team, Guardian Robot Project, Information R&D and Strategy Headquarters, RIKEN, used functional magnetic resonance imaging (fMRI) to demonstrate that "facial expressions" and "subjective emotional experiences (subjective experiences)" associated with emotions such as joy, sadness, anger, and anxiety are each supported by distinct brain networks, and that they interact in a way whereby facial expressions give rise to subjective experiences. The findings are expected to contribute to the development of new treatments for emotional dysregulation disorders and to the development of AI and robots capable of understanding human emotions. The results were published in the January 21 online issue of Communications Biology.

In the study, brain activity was measured using fMRI while participants watched emotionally evocative film clips, and facial movements (facial responses: lip corner pulling actions) and subjective experiences (ranging from unpleasant to pleasant) were simultaneously recorded and analyzed.

The results showed that facial responses were associated with activity in the limbic regions and somatosensory motor cortices, while subjective experiences were associated with activity in the medial parietal and lateral temporoparietal cortices.

It was also found that the medial parietal and lateral temporoparietal regions, which are active in association with subjective experiences, are connected to form a subjective experience network.

Dynamic causal modeling further supported a model in which the sensory processing network first interacts with the facial response network, which subsequently influences the subjective experience network.

These findings are considered to provide evidence for a theory proposed in psychology over 100 years ago—that bodily responses occur first and subjective experiences are formed through them (such as the case of feeling happy because one smiles).

Sato commented: "We wanted to understand the brain mechanisms that give rise to emotions. In this neuroimaging study, we carefully measured human facial movements and subjective feelings and examined the corresponding brain regions and the relationships between them. Contrary to the common assumption that 'feelings arise first and are then reflected in expressions,' the results show that facial expressions are produced first through sensory processing, and that subjective emotions are subsequently shaped on the basis of that information. I found it fascinating that this aligns with the idea intuitively proposed by psychologist William James more than 100 years ago. Going forward, I would like to explore whether we can recreate these mechanisms artificially and thereby realize an android that has emotions."

Journal Information
Publication: Communications Biology
Title: Neural network dynamics associated with facial and subjective emotional responses
DOI: 10.1038/s42003-025-09361-5