A new study published in Current Biology sheds light on how the brain divides the continuous flow of information in our daily lives into discrete, meaningful events. The findings reveal that these divisions—similar to starting a new chapter in a book—aren’t solely prompted by changes in our surroundings. Instead, they can be influenced by internal scripts based on past experiences and current goals. This active process of segmentation reflects how the brain prioritizes information to fit what matters most to us at any given moment.
Scientists have long been intrigued by how we process the endless stream of events in our daily lives and make sense of them as distinct moments. This segmentation process is essential for understanding the world, updating mental models, and forming lasting memories. But what prompts the brain to mark the boundary between one event and the next?
One possibility is that physical changes in the environment, like moving from an outdoor space into a building, signal the start of a new “chapter.” Another hypothesis suggests that these boundaries are shaped by internal “scripts”—mental templates formed through experience. These scripts allow the brain to filter and prioritize information based on what aligns with current objectives. For example, when dining at a restaurant, an experienced diner might focus on events such as ordering food and its arrival, while ignoring other background changes.
The researchers sought to test whether the brain actively constructs these boundaries by prioritizing internal scripts over environmental cues.
“We wanted to see how the brain processed realistic situations, with complex events that unfolded over multiple minutes,” said study author Chris Baldassano, an associate professor of psychology at Columbia University and director of the Dynamic Perception and Memory Lab.
“Recent work from my lab and others has found that there are parts of the brain (including a group of regions called the default mode network, the DMN) that usually turn off during a typical neuroscience study in which we flash unrelated pictures or words in a sequence, but that are actually highly engaged when we present people with a meaningful experience like a movie or story. We were hoping to understand what information about the narrative is being tracked in the DMN (and especially in the medial prefrontal cortex), and whether a story is processed differently depending on what you are currently paying attention to.”
To explore this, the researchers designed 16 audio narratives that each combined two overlapping “scripts.” One script was location-based (e.g., a restaurant, airport, grocery store, or lecture hall), while the other focused on social interactions (e.g., a breakup, a marriage proposal, a business deal, or a chance romantic encounter). Each story progressed through four distinct events for each script, interwoven in such a way that the boundaries for one script did not coincide with those of the other. This setup allowed researchers to test how attention to different scripts influenced participants’ perceptions of event boundaries and the corresponding brain activity.
“The stories were created by Alexandra De Soares, the member of my lab who led this study, such that each one was a combination of two ‘scripts’ (i.e. common event sequences),” Baldassano explained. “For example, a story might be about a marriage proposal in a restaurant, and proceed through the stages of eating at a restaurant (being seated, ordering, …) while the events of the proposal are happening at the same time (getting out the ring, proposing, reaction of the partner, …).”
In the first part of the study, conducted in an fMRI scanner, 36 participants listened to these narratives while their brain activity was monitored. Before listening to each story, participants were “primed” by being given a set of questions related to one of the scripts, such as location-specific details (e.g., “What does each person order?” in a restaurant) or social-specific details (e.g., “What is the initial reason stated for the breakup?”). This priming encouraged participants to focus on one script over the other, effectively shaping their mental priorities during the task. While listening to the story, participants’ brain activity was measured, particularly in regions of the DMN, which are known to process abstract and meaningful information.
In the second part of the study, conducted online, over 300 participants listened to the same narratives. However, instead of listening continuously, they heard the stories sentence by sentence and were asked to decide after each sentence whether it marked the beginning of a new event. Like in the fMRI study, participants were primed to focus on either the location script or the social script, or they were not primed at all. After listening, participants completed a memory test where they answered questions about story details, both for the script they were primed to focus on and for the other, unprimed script.
The findings revealed that participants’ segmentation of events and their brain activity patterns were strongly influenced by the script they were primed to prioritize. In the online experiment, participants primed with the location script were more likely to identify new events at moments corresponding to location-based transitions, such as entering a restaurant or moving through security at an airport. Similarly, participants primed with the social script were more attuned to changes in social interactions, though this effect was less pronounced than for location-based events.
“I was surprised to find that telling participants to attend to social-related scripts (such as the marriage proposal) didn’t have much of an effect on how chapters were created (compared to the baseline condition in which participants just listened to the stories without specific instructions),” Baldassano told PsyPost. “Instructing participants to focus on location-related scripts (such as the restaurant) caused much larger changes. In retrospect, this makes some sense: by default, listeners are more likely to focus on the characters and their social interactions, so we only saw big changes when we asked them to approach the story from an unusual perspective.”
Priming also improved memory recall: participants remembered details relevant to their primed script more accurately than those from the unprimed script. This demonstrated that focusing attention on specific types of information not only shaped perception but also enhanced memory encoding.
In the fMRI study, brain activity in the DMN, particularly in the medial prefrontal cortex, aligned with the boundaries relevant to the primed script. For example, when participants were primed to focus on location details, neural transitions in this region coincided with location-based event boundaries in the narratives. These findings suggest that the brain actively constructs event boundaries based on internal priorities rather than passively responding to environmental cues.
“For decades, researchers in cognitive psychology have been interested in how our experiences (which unfold continuously over time) get divided into individual chapters, a cognitive process referred to as ‘event segmentation,'” Baldassano said. “The major new finding in this study is that these chapters don’t just come from the way the story is written. People are not just listening for the author of a story to tell them when a new event is starting; the brain is making active choices about when to start a new chapter that depend on your current frame of mind and goals.”
The study’s design provided a controlled look at how internal scripts influence event segmentation, but it had limitations. The stories used followed the scripts closely, with no major deviations, but real-life experiences are often less predictable.
“In this study, the events of the stories perfectly matched the scripts that people had: for example, in the restaurant stories people always got the food that they ordered,” Baldassano noted. “My lab and others are currently trying to understand what happens when there is a mismatch between your expectations and an actual experience, such as if the waiter threw a glass of water in a guest’s face. What impact does this have on how these experiences would be organized in memory?”
The researchers are also analyzing data on how participants’ perspectives during listening influenced their memory recall. This ongoing work aims to deepen our understanding of how real-world experiences are divided into distinct events and stored in memory.
“We are still working to get a complete picture of the ways that brain regions work together to create memories for realistic, familiar events,” Baldassano said. “Specifically, we are interested in how and when DMN regions like the medial prefrontal cortex send information to and from the long-term memory system in the hippocampus. For example, my lab is currently running studies with patients that have implanted electrodes (as part of a surgical procedure for epilepsy) so that we can directly measure neural activity in the memory system.”
The study, “Top-down attention shifts behavioral and neural event boundaries in narratives with overlapping event scripts,” was authored by Alexandra De Soares, Tony Kim, Franck Mugisho, Elen Zhu, Allison Lin, Chen Zheng, and Christopher Baldassano.
