Zhou, X., Ghorbani, F., Roessner, V., Hommel, B., Prochnow, A., & Beste, C. (2025). Metacontrol instructions lead to adult-like event segmentation in adolescents. Developmental Cognitive Neuroscience, 101521.

 

What They Did

The authors showed the movie The Red Balloon to adults and adolescents, instructing them to press the space bar at event transitions. The movie was divided into three sections for the adolescent group and four for the adult group, with participants taking breaks as long as desired between sections. Each individual also had a five-minute practice video to be sure they understood the task.

During the experiment, individuals wore caps with electrodes to record brain activity. Half of the adolescents were told to make their distinctions as fine-grained as possible, while the other half were not given any special instructions. The two groups of adolescents were matched by age and gender, though each group had 8 to 10 participant records excluded because their data was unusable.

The authors found that the adolescents who were told to make fine-grained distinctions marked more transitions than the other adolescents and a similar number to the adults. The patterns of brain region activation, however, were different between the instructed adolescent group and the adults. The authors note that the adult-style behavior involved greater top-down processing and attention control and that the differences in brain activity may indicate that such behavior is more effortful for adolescents.

The hypothesis of increased effort is also consistent with the fact that the instructed adolescents differed most from the adults during periods when the movie had many things happening at a time. The authors remark that their study suggests that adolescents can use more mature cognition if given instruction.

Further Exploration

Once I read that the authors had used The Red Balloon as the segmenting task, I paused my reading and found the video to try the task for myself (see https://www.youtube.com/watch?v=gw4QbQLqd1E).  I didn’t have the training video or any way to mark space bar presses, so it wasn’t an equivalent setup. I paused the video whenever I saw an event transition, and I used a Notepad file to type in the timestamp and a phrase about what was happening. My breaks were formed by daily life events.

I found the exercise pretty tedious, and I continued only to compare my results to the data in the paper. In the end, I averaged about 7 seconds between events, which was on the low end of the adult results. I had also already read that one group of adolescents had been given the “fine-grained” instructions, so having that in mind may have influenced the way I approached the task.

Some of my cues for a new event were changes in speed or direction of movement, interaction with a new person or object, and scene breaks. Nonetheless, I wondered whether scene breaks really counted when they were simple switches between groups engaged in fairly consistent, repetitive activity. If the two sets of activity had been presented separately, I would have counted each one as a single event, but with the switches, I counted six or so. It feels like mechanical segmentation is different from narrative segmentation, but that's a rabbit hole for another day!

Image credit: Erik Drost

https://commons.wikimedia.org/wiki/File:Igers_Cleveland_(16276155143).jpg