Visual object categorization in infancy.
Level 4 - case-series / case-control
Cross-sectional developmental observational study (graded by design analogy for basic cognitive neuroscience).
PubMed 35169072 · doi:10.1073/pnas.2105866119
What was done
Eye tracking was used to measure differential looking times in 4-, 10-, and 19-month-old infants viewing pairs of images across eight animate and inanimate categories (human versus nonhuman, faces versus bodies, big versus small real-world size, and natural versus artificial). Looking times were used to construct representational similarity spaces for each age group, which were then compared against computational hypothesis-based models and adult functional MRI models of visual cortex object categorization.
What was found
The abstract reports no numerical values, test statistics, or effect sizes. Directionally, as infants matured from 4 to 19 months, their looking patterns matched neural representations across progressively larger portions of the adult visual cortex. Categorization showed two distinct shifts: between 4 and 10 months, visual exploration transitioned from low-level saliency toward an animate versus inanimate distinction; between 10 and 19 months, a category spurt shifted representations toward mature, adult-like organization.
Why it matters
This work outlines an incremental timeline for how spontaneous visual categorization emerges in early life, linking behavioral exploration in infants directly to representational maps in the adult visual cortex.
Limits
The abstract does not disclose sample sizes, exclusion rates, or quantitative effect sizes. The study infers neural development in infants by correlating infant looking behavior with adult fMRI data rather than measuring infant brain activity directly.
Cited by
- context By age six, human children can learn tens of thousands of distinct object categories.