Detecting neural state transitions underlying event segmentation

Abstract

Segmenting perceptual experience into meaningful events is a key cognitive process that helps us make sense of what is happening around us in the moment, as well as helping us recall past events. Nevertheless, little is known about the underlying neural mechanisms of the event segmentation process. Recent work has suggested that event segmentation can be linked to regional changes in neural activity patterns. Accurate methods for identifying such activity changes are important to allow further investigation of the neural basis of event segmentation and its link to the temporal processing hierarchy of the brain. In this study, we introduce a new set of elegant and simple methods to study these mechanisms. We introduce a method for identifying the boundaries between neural states in a brain area and a complementary one for identifying the number of neural states. Furthermore, we present the results of a comprehensive set of simulations and analyses of empirical fMRI data to provide guidelines for reliable estimation of neural states and show that our proposed methods outperform the current state-of-the-art in the literature. This methodological innovation will allow researchers to make headway in investigating the neural basis of event segmentation and information processing during naturalistic stimulation. HighlightsO_LIBoundaries between meaningful events are related to neural state transitions. C_LIO_LINeural states are temporarily stable regional brain activity patterns. C_LIO_LIWe introduce novel methods for data-driven detection of neural state boundaries. C_LIO_LIThese methods can identify the location and the number of neural state boundaries. C_LIO_LISimulations and empirical data support the reliability and validity of our methods. C_LI