Why Age Two Is the Real Screen-Time Line, Not One Hour

Your child’s grandmother says “cartoon dekh raha hai, we watched Doordarshan too, nothing happened to us.” Your paediatrician says one hour, max. Somewhere between those two positions, actual neuroscience got published in 2025 — and it says something more specific than either of them.

What brain scans actually found this year

A team from A*STAR’s Institute for Human Development and Potential and the NUS Yong Loo Lin School of Medicine followed 168 children from Singapore’s GUSTO cohort, scanning their brains at ages 4.5, 6, and 7.5. Children with higher infant screen time showed accelerated maturation of brain networks responsible for visual processing and cognitive control, and this was linked to slower decision-making and increased anxiety by their teenage years.

Children with these altered brain networks took longer to make decisions during a cognitive task at age 8.5, and those with slower decision-making reported higher anxiety symptoms at age 13.

Here’s the detail that actually matters for what you do at home: screen exposure before age two, but not at ages three or four, predicted these long-term brain changes. That’s a real finding, not a scare headline — the researchers found a specific sensitive window, not a blanket “screens are bad” conclusion.

The same group had already published a related finding in 2024: infant screen time is associated with alterations in brain networks that govern emotional regulation, but parent-child reading could counteract some of these changes — among children whose parents read to them frequently at age three, the link between infant screen time and altered brain development was significantly weakened.

That’s the headline most coverage missed. It’s not “screens before 2 doom your child.” It’s “screens before 2 shift the odds, and reading together shifts them back.”

A second, independent study — same age, same story

A 2025 bioRxiv preprint from a Canadian team scanned 70 toddlers’ brains at age 2, with a subset rescanned at age 3. Higher screen time at age 2 significantly predicted smaller right pars triangularis volumes and poorer 2-year language outcomes. Pars triangularis sits inside Broca’s area — the part of the brain most tied to producing language.

Importantly, the researchers didn’t find that the brain change caused the language problem in a simple chain. The volumes did not predict language outcomes, indicating that the effects of screen time on brain development and language appeared to be occurring in parallel rather than mediated. There’s a second twist worth knowing: children with smaller volumes and weaker language skills at age 2 became more likely to rely on screens at age 3, suggesting that early vulnerabilities amplify later screen use. In plain terms — a toddler who’s already struggling with language may get parked in front of a screen more, which is a parenting-pressure problem, not just a brain problem.

The large population study that’s easy to miss

Brain scans get the headlines, but a Danish study using data from over 52,000 children found something at a much bigger scale. This wasn’t a lab study — it used cross-sectional baseline data from the Danish longitudinal survey TRACES, comprising data from 52,010 children , and found mobile device screen time was associated with poorer language development among toddlers. The scale here matters — this isn’t 70 or 168 children, it’s a national sample, and the direction of the finding matches the smaller brain-imaging studies.

What’s established vs. what’s still contested

Established, with real convergent evidence:

• Under-2 exposure looks different, biologically, from exposure at 3-4. Two independent teams (Singapore, Canada) landed on the same window using different methods — MRI and behavioural testing.
• Solo, passive viewing is consistently worse than the same content watched with an adult who talks about it. A meta-analysis of 100 studies found that watching alone is linked with poorer cognitive outcomes, while co-viewing with an engaged adult is linked with better outcomes.
• Young children genuinely learn less from a screen than from the same interaction face-to-face. Young children learn much less from screen-based media than from the same kind of real-life interaction, and this gap may last until around school age.

Genuinely contested or too early to call:

• Whether the brain changes cause worse outcomes, or whether both are driven by something else (a fussier baby who’s harder to engage gets more screen time AND has different brain development). The Canadian preprint explicitly found the brain and language effects ran in parallel, not one causing the other — this is not settled science.
• Whether the effect is dose-dependent (an hour vs. two hours) or threshold-dependent (any regular exposure before 2 vs. none). The current studies can’t cleanly separate these.
• Video calls with grandparents. Video calling is an increasingly common activity for young children and can help maintain and strengthen family bonds — but there’s genuinely no strong evidence yet on whether it carries the same risks as passive viewing, because it’s interactive and relationship-driven rather than a program playing at the child.
• AI toys and chatbots for under-5s. A UK expert panel reviewing the evidence in early 2026 was blunt about this: the panel found no convincing evidence on the impact or safety of AI chatbots or AI-enabled toys for children under five. If a toy or app is marketed as “AI-powered learning” for your toddler, treat that claim as unverified, not proven safe or proven harmful.

The variable that matters more than the clock: what’s on screen, who’s next to your child

Content type keeps showing up as more predictive than duration alone. In one older but still-cited toddler study, children who watched over 2 hours of television per day had increased odds of low communication scores, and while child-directed media was associated with low language scores, adult-directed media was not. Translation: a toddler watching a slow, dialogue-heavy adult show with you in the room is a different exposure than the same toddler alone with a fast-cut toddler app.

A same-day decision table

Use this instead of trying to track total minutes.

Situation	What the evidence says	Do this
Under 12 months, screen just “on” in background	WHO explicitly flags this as screen time even if the baby isn’t watching	Turn off the TV during play and meals, even if no one’s “watching”
12–24 months, any screen use	Sensitive window per two independent 2025 studies — least certainty, most caution warranted	Default to zero; if travelling or unwell, use video calls with family over passive video
2–4 years, solo tablet time	Meta-analytic evidence favours co-viewing over solo use	Sit with them for at least the first 5 minutes; ask “what do you think happens next?“
2–4 years, video call with grandparents	Interactive, relationship-based — different category from passive viewing	Treat this separately from “screen time” — it’s closer to a phone call than to TV
Any age, reading time in the daily routine

Parent-child reading weakened the link between infant screen time and altered brain development, among children whose parents read to them frequently at age three | 10–15 minutes of shared reading isn’t a “nice extra” — treat it as protective, on the same list as sleep and food |

The daily flashpoints — meals, screens, and advice from family — come up here too: How to Handle Mother-in-Law Parenting Advice Calmly · How to Teach Kids to Name Their Feelings at Dinner · The Picky Eater Fix: What Actually Works (Backed by Science)

The one thing to say tonight

In my house the rule runs both ways. When I’m scrolling reels, my son calls it out — “Papa, no reels” — and the deal is that the moment he says it, the phone goes away. No negotiating. It sounds small, but modelling the boundary myself did more than any limit I set for him.

So if a grandparent or co-parent hands your under-2 a phone to “keep them quiet for five minutes,” you don’t need to lecture. Say: “Let’s do this on video call with Dadi instead — same phone, but she’ll talk back.” That single swap — passive to interactive — is the difference the research keeps pointing to, not the device itself.