Identifying moments of peak audience engagement from brain responses during story listening

ABSTRACT

Stories in general, and peak moments within a single story in particular, can evoke strong responses across recipients. Between the content of a story and these shared audience responses lies an explanatory gap that neuroimaging can help close. Accordingly, this study examined how the brains of an audience responded during a story. We performed two types of analyses: First, we correlated the story’s physical characteristics to brain activity. Second, we reverse-correlated moments of peak brain engagement to story segments. We found that activity peaks in the temporo-parietal junction identify socially engaging points within the story, such as a pie-in-the-face scene, hyperbole, and sexual references. We discussed how these results and reverse correlation neuroimaging more broadly advance communication science.

KEYWORDS:

• Communication neuroscience
• story
• social cognition
• audience engagement
• neuroimaging

Acknowledgements

We thank the authors of the original study for making the data publicly available. We also thank the creators of the nilearn and BrainIAK packages for neuroimaging data analysis and the developers of pandas, seaborn, and Jupyter software packages. We acknowledge the support of the ICER high-performance computing cluster at Michigan State University. Three anonymous reviewers deserve credit for providing invaluable advice to improve and clarify the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Critically, we provide macro-level overview rather than a microscopic account of these subsystems. There are other relevant systems that we skip over here, particularly the complex topic of neurolinguistics. The supplementary materials provide analyses of hundreds of additional regions. Also, this is a somewhat simplified theoretical account of how the brain transforms information from simple sounds into complex meaning, emphasizing modularity and serial processing.

2 Please also see the extensive argument about the distinction between reverse correlation and reverse inference in the Discussion.

3 Laughter provides an objective criterion for strong audience impact and one that is clearly codeable. The decision to use laughter as an index of socially engaging content as opposed to other operationalized indicators has to do with the fact that only one particular story is used in this study, which naturally limits the number of peaks and story segments that are available for reverse correlation, including the contents that are covered by the story. In the future, we envision that larger datasets will enable to reverse-correlate from corpora containing hours of story content with thousands of brain activation peaks, enabling analyses that go beyond laughter. For this study, however, we use laughter as an objective criterion, but we encourage readers to examine the sound trailers and text segments.