Newborn girls in a Cambridge hospital spent a larger share of their viewing time on a moving human face than newborn boys did, and boys spent a larger share on a swinging mobile of metal balls. The 130 babies, 67 boys and 63 girls, were tested at a mean age of 33 hours, and the size of the gap was d = 0.40, a small-to-moderate effect by conventional benchmarks.
Yumnah Khan, a PhD student at the University of Cambridge’s Autism Research Centre, led the work under Professor Simon Baron-Cohen. It appeared in Biology of Sex Differences on September 21, 2026.
Face Against Newton’s Cradle
Babies were recruited at the Rosie Hospital in Cambridge, part of Cambridge University Hospitals NHS Foundation Trust. Each was shown two videos side by side, according to Cambridge’s research news page: a human face moving naturally, which counts as the social stimulus, and a Newton’s cradle, which counts as non-social. Eye movements were recorded, and the video coders did not know the baby’s sex or which stimulus the baby was watching at any moment.
Both sexes leaned toward the face overall. The sex difference sits on top of that shared preference: girls, on average, gave the face a greater proportion of their looking time, and boys gave the cradle a greater proportion. The abstract in Cambridge’s repository record puts the female excess on the face at d = 0.40.
The project behind the experiment is the Cambridge Baby Study, which the Autism Research Centre describes on its project page as a test of whether the earlier finding, that female newborns tend to look longer at faces and male newborns at objects, can be replicated. The page says the team assessed the 130 babies within their first two days after birth, consulted recently delivered mothers beforehand to make the procedures easier for families, and plans to follow the same infants up at 18 months.
Because the original 2000 experiment was itself a Cambridge study, the new work is a replication within Cambridge and not by an outside team. The repository abstract states that earlier neonatal studies raised methodological concerns and have not been replicated, which is the gap this experiment was designed to close, and which an outside team could close further.
A Replication, With Better Controls
The experiment revisits a 2000 Cambridge study that reported the same direction in newborns. The release says the new version used more controlled testing conditions, a more reliable way of tracking infant gaze, and blinded coding, which answers the methodological criticisms that followed the original. The authors say prior research on neonatal sex differences is extremely limited, has produced mixed results, and has often gone unreplicated.
Baron-Cohen framed the result in terms of interaction. “It’s nature and nurture, not nature or nurture,” he said, arguing that prenatal factors such as sex hormones or genes work together with experience after birth. Khan said her earlier work found average sex differences in brain structure at birth, and that the new study extends the picture to attention and behaviour.
The Authors’ Own Limits
The caveats come from the researchers themselves. Carrie Allison of the Autism Research Centre stressed that the differences are averages and “don’t apply to all male or all female infants,” with large variation inside each sex and large overlap between them. A d of 0.40 leaves the two distributions overlapping heavily.
The prenatal explanation is an inference. The babies were tested at roughly a day and a half old, which rules out much of postnatal learning but not every route by which experience could act, and the paper calls the suggestion that sex hormones contribute speculative. It also raises alternative accounts: the differences may reflect general maturational differences between male and female infants, or low-level visual properties of the two videos rather than their social content, such as motion or contrast.
Alex Tsompanidis, another member of the Cambridge group, told the university that early differences are useful starting points but do not fully determine later learning and development. Khan told Newsweek that early prenatal factors might have contributed, and that further research would be needed to test it. A longitudinal follow-up funded by the K. Lisa Yang Centre for Autism Research at Cambridge is under way to see whether the newborn pattern predicts later differences, and Khan’s comments to Newsweek say planned future work will look at language acquisition.
The work was funded by the Wellcome Trust, the Simon Foundation Autism Research Initiative, and a Trinity College and Cambridge Trust studentship. News-Medical’s account notes that the autism context in the release, with diagnosis about twice as common in boys, is background and not a finding of this experiment. All 130 babies came from one hospital, and nothing in the paper shows whether a 0.40 effect measured at 33 hours of age persists to the 18-month follow-up.
This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.
More from Morning Overview
- Hybrids have 15% fewer problems than gas cars, while EVs and plug-in hybrids have about 80% more, Consumer Reports finds
- Regulators cleared the first U.S. small modular reactor, 4 months early
- Lake Powell sank to a record 3,517 feet, nearing the level that stops Glen Canyon Dam’s turbines
- Seven of Earth’s nine planetary boundaries are breached, and all seven are worsening
