Temporal dynamics of visual representations in the infant brain.

Area of Science:

• Computational Neuroscience
• Developmental Neuroscience
• Cognitive Neuroscience

Background:

• Computational neuroscience tools enable studying neural representations.
• Limited understanding of early-life representational content in the brain.
• Investigating early visual processing in infants is crucial for developmental insights.

Purpose of the Study:

• To determine if neural activity patterns from complex visual stimuli can be decoded in infants.
• To compare infant and adult neural representations of visual stimuli using EEG data.
• To explore the developmental trajectory of visual representations from infancy to adulthood.

Main Methods:

• Electroencephalography (EEG) data collected from 12-15-month-old infants and adult controls.
• Analysis of pairwise classification accuracy of neural activity patterns over time post-stimulus onset.
• Assessment of linear separability of neural representations across different visual domains (animals, human body).

Main Results:

• Classification accuracies for visual stimuli exceeded chance levels within 500 ms in both infants and adults.
• Infant neural representations were not linearly separable across visual domains, unlike adults.
• Neural representations showed similarity within age groups but not across them.

Conclusions:

• Infant visual representations undergo significant reorganization between infancy and adulthood.
• Decoding EEG data within-subject is a feasible method for studying infant brain object representation.
• Findings provide a proof-of-concept for using neuroimaging to understand early visual development.