Two-month-old babies categorize objects earlier than thought

Neuroscientists at Trinity College Dublin have found that babies as young as two months old can already sort visual information into categories like animals and toys. Using brain scans and AI, the study reveals early foundations of perception. This challenges previous assumptions about infant cognition.

New research published in Nature Neuroscience demonstrates that infants begin organizing the world around them much sooner than expected. Led by Dr. Cliona O'Doherty at Trinity College Dublin's Cusack Lab, the study involved 130 two-month-old babies from Dublin's Coombe and Rotunda Hospitals. Each infant underwent functional MRI (fMRI) scans while lying on a soft beanbag, wearing sound-cancelling headphones, and viewing colorful images from 12 categories, including cats, birds, rubber ducks, shopping carts, and trees. The sessions lasted 15-20 minutes to maintain attention.

Researchers combined these brain activity patterns with artificial intelligence models to decode how infants represent and categorize visuals. "Although at two months, infants' communication is limited by a lack of language and fine motor control, their minds were already not only representing how things look, but figuring out to which category they belonged," Dr. O'Doherty explained. This marks the largest longitudinal fMRI study of awake infants to date.

Team leader Rhodri Cusack, Thomas Mitchell Professor of Cognitive Neuroscience at Trinity, highlighted the broader applications: "This study... opens up a whole new way to measure what babies are thinking at a very early age. It also highlights the potential for neuroimaging and computational models to be used as a diagnostic tool in very young infants." Co-author Anna Truzzi, now at Queen's University Belfast, noted how recent advances in AI and neuroimaging enabled these insights into the first year of rapid brain development.

The findings could guide early-years education, support for neurodevelopmental conditions, and more efficient AI designs. Neonatologist Eleanor Molloy emphasized the need to understand disorders affecting early brain growth, where awake fMRI shows promise. The work, detailed in a 2026 Nature Neuroscience paper (DOI: 10.1038/s41593-025-02187-8), underscores the richness of infant brain function from the outset.

相关文章

Split image of two brains illustrating high and low connectivity patterns in autism from a scientific study.
AI 生成的图像

Brain scans and mouse studies point to two autism connectivity subtypes

由 AI 报道 AI 生成的图像 事实核查

A large study combining human brain imaging with data from genetically engineered mouse models has identified two recurring patterns of brain connectivity in autism—one marked by higher-than-typical connectivity and another marked by lower connectivity—each tied to different biological pathways, researchers report.

A new analysis of the best-preserved Neanderthal infant skeleton shows that these ancient babies developed bones and brains at a pace matching modern humans aged 12 to 14 months, despite being only about six months old. The findings, based on the Amud 7 infant from Israel, suggest Neanderthals grew rapidly in early childhood as an adaptation to harsh environments. Researchers observed similar patterns in other young Neanderthal remains.

由 AI 报道

A new theory from scientists at Cold Spring Harbor Laboratory suggests that brain cells use their cellular family tree to organize into a complex organ. The model explains how a single starting cell can form a brain with roughly 170 billion precisely positioned neurons.

A new survey by Autism BrainNet reveals strong public support for studying the autistic brain but low awareness of brain donation's role. While 92 percent of Americans view such research as very important, 70 percent have never heard of brain donation. The findings highlight confusion between brain and organ donation processes.

此网站使用 cookie

我们使用 cookie 进行分析以改进我们的网站。阅读我们的 隐私政策 以获取更多信息。
拒绝