Neural event segmentation of continuous experience in human infants

Abstract

How infants experience the world is fundamental to understanding their cognition and development. A key principle of adult experience is that, despite receiving continuous sensory input, we perceive this input as discrete events. Here we investigate such event segmentation in infants and how it differs from adults. Research on event cognition in infants often uses simplified tasks in which (adult) experimenters help solve the segmentation problem for infants by defining event boundaries or presenting discrete actions/vignettes. This presupposes which events are experienced by infants and leaves open questions about the principles governing infant segmentation. We take a different, data-driven approach by studying infant event segmentation of continuous input. We collected whole-brain functional MRI (fMRI) data from awake infants (and adults, for comparison) watching a cartoon and used a hidden Markov model to identify event states in the brain. We quantified the existence, timescale, and organization of multiple-event representations across brain regions. The adult brain exhibited a known hierarchical gradient of event timescales, from shorter events in early visual regions to longer events in later visual and associative regions. In contrast, the infant brain represented only longer events, even in early visual regions, with no timescale hierarchy. The boundaries defining these infant events only partially overlapped with boundaries defined from adult brain activity and behavioral judgments. These findings suggest that events are organized differently in infants, with longer timescales and more stable neural patterns, even in sensory regions. This may indicate greater temporal integration and reduced temporal precision during dynamic, naturalistic perception.

Automated and manual segmentation of the hippocampus in human infants

“…Although other studies have shown the success of automated hippocampal segmentation in adults ( Yushkevich et al, 2015 ), children ( Schlichting et al, 2019 ), and infants ( Guo et al, 2015 , Zhu et al, 2019 ), this study is the first, to our knowledge, to show that automated methods for segmenting infant hippocampus can generalize across tracers and can work for scans collected in awake infants. As task-based, awake infant fMRI becomes more prevalent ( Ellis et al, 2020 , Ellis et al, 2021a , Ellis et al, 2021b , Ellis et al, 2021c , Yates et al, 2022 ), there will be increasing need for protocols that produce accurate and reliable segmentations of the infant hippocampus from noisy data. Automated methods make it possible to accelerate and improve the study of the developing hippocampus.…”

Multi-view manifold learning of human brain state trajectories

“…Event segmentation theory explains that humans automatically divide continuous streams of information into discrete events [45, 47] in order to form, organize, and recollect memories, make decisions, and predict the future [63]. Participants show high consistency in explicitly segmenting continuous stimuli [46] and also in how their brains represent these event boundaries [50, 58]. Event boundaries can be represented, both behaviorally and in brain activity, at different timescales depending on the information being used to draw event boundaries [57].…”

“…The first step of this process is to use the HMM to learn from the data the number of events a brain region represents for a given stimulus. Past studies have used HMMs for event segmentation on multivoxel activity patterns and have validated this approach against behavioral segmentation [50, 51, 58]. This shows that voxel resolution data reflect meaningful event segmentation, so we chose to estimate the optimal number of events ( K ) for each brain region using the voxel resolution data (Fig.…”

Face processing in the infant brain after pandemic lockdown

“…We followed validated procedures and parameters (see Figure 1a ) for collecting fMRI data from awake infants (Ellis et al., 2020 , 2021a , 2021b , 2021c ; Yates, et al., 2022 ). Data were acquired using the bottom half of a 20‐channel head coil on a Siemens Prisma (3T) MRI.…”

“…for collecting fMRI data from awake infants (Ellis et al, 2020(Ellis et al, , 2021a(Ellis et al, , 2021b(Ellis et al, , 2021cYates, et al, 2022). Data were acquired using the bottom half of a 20-channel head coil on a Siemens Prisma (3T) MRI.…”