Screen Time and Preterm Cognitive Outcomes

Children born extremely preterm — at less than 28 weeks of gestation — start school with measurable cognitive disadvantages relative to term-born peers. A meta-analysis of post-1990 cohorts found a roughly 12-point full-scale IQ gap, with similar deficits in working memory and executive function (Twilhaar et al., 2018). Against this baseline, the question of how screen time affects this group is qualitatively different from how it affects typically-developing children: the relevant comparison is not whether screens cause problems in the general population, but whether they amplify or partially overlap with vulnerabilities that are already present.

The largest direct evidence comes from a 414-child cohort drawn from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Neonatal Research Network. Vohr and colleagues (2021) reported that more than two hours of screen time per day was associated with lower full-scale IQ, more executive-function problems, and higher behavioral symptoms at school age — over and above the deficits expected from extreme prematurity itself.

What the Vohr study actually found

The cohort enrolled 414 children born at less than 28 weeks gestation between 2005 and 2009 and assessed them at ages 6 years 4 months to 7 years 2 months. Of the children evaluated:

• 57% (238 children) had high screen time, defined as more than 2 hours per day.
• 64% (266 children) had a television or computer in their bedroom.

After adjustment for study center, child sex, gestational age, severe retinopathy of prematurity, and a panel of social-determinants variables (parental education, household income, primary caregiver), the high screen-time group showed:

• Significantly lower full-scale IQ on the Wechsler Intelligence Scale for Children (WISC).
• Higher scores on the Behavior Rating Inventory of Executive Function (BRIEF) Metacognition Index, Global Executive Composite, and inhibition subtests of the NEPSY.
• Higher inattention and impulsivity ratings on parent-reported behavior measures.

The bedroom-screen variable mattered independently. Children with a TV or computer in their bedroom showed elevated hyperactivity and impulsivity scores even after controlling for total screen time, suggesting that the environmental presence of screens — sleep displacement, late-evening exposure, the loss of a quiet room — adds to the daily-hours measure rather than being absorbed by it.

Why does the EPT context matter?

The findings would be unsurprising in a typical-development cohort, but the EPT context adds three layers of interpretation.

Baseline vulnerability is already substantial. Twilhaar et al. (2018), in a meta-analysis of cohorts born at less than 32 weeks since 1990, found a pooled standardized mean difference of -0.86 SD on full-scale IQ versus term-born peers — roughly 13 IQ points. The deficit on processing speed was even larger, and the gap on executive-function measures was also pronounced. EPT children are not on the same starting curve as the children studied in general-population screen-time work.

The deficits persist across childhood and into adolescence. Lee and colleagues (2024), in a longitudinal analysis of very-preterm children followed to age 17, found persistent executive-function deficits across working memory, planning, and cognitive flexibility, with the largest gaps in those born before 28 weeks. The associations were largely explained by neonatal medical complications and white matter abnormalities — meaning the deficits track underlying neuroanatomy, not simply early environmental adversity.

Screen time may target the same domains that prematurity already affects. The Vohr findings — full-scale IQ, executive function, inattention — overlap heavily with the documented prematurity-associated deficits. Whether the effect is additive (screen time worsens an independent dimension), multiplicative (vulnerable children are disproportionately affected), or partly confounded (children with worse cognitive trajectories receive more screen time as a coping or babysitting strategy) is not yet settled.

The general-population evidence anchors the mechanism

Outside the EPT context, screen time has been studied at scale, and several findings establish biological and behavioral plausibility for the Vohr observations.

Hutton et al. (2020) used diffusion tensor imaging to examine white matter integrity in 47 preschool-aged children whose parents completed the ScreenQ measure of screen-based media use. After adjustment for age and household income, higher screen-use scores were associated with lower fractional anisotropy and higher radial diffusivity in tracts supporting language and emergent literacy, including the inferior longitudinal fasciculus and the superior longitudinal fasciculus. Effect sizes were modest, the design was cross-sectional, and the sample small — but the same children also performed worse on language and literacy assessments, providing convergent behavioral evidence.

Madigan et al. (2020) conducted a systematic review and meta-analysis of 42 studies (n > 18,000 children) examining screen time and child language skills. The meta-analytic findings were nuanced rather than uniformly negative:

• Quantity: greater screen time was associated with poorer language skills (small to moderate effect).
• Background television (TV on but not actively watched) was negatively associated with language outcomes.
• Educational programming showed positive associations with language when content was age-appropriate.
• Co-viewing with a caregiver, in which the adult discussed and contextualized the content, was protective and reversed the negative effects observed in passive solo viewing.
• Earlier age of screen onset was associated with worse language outcomes.

The general-population picture is therefore not “screens are bad” but “passive, solo, background, early-onset screen exposure is associated with worse language and brain-structural outcomes; co-viewed, age-appropriate, contextualized exposure is not, and may be neutral or modestly positive.” The Vohr 2021 measurement captured total daily hours, which conflates these context dimensions, so the EPT finding is best read as evidence that the typical pattern of high-volume screen use in this cohort skews toward the unfavorable end of the context spectrum.

What the Vohr findings do and do not support

The 414-child cohort is large for an EPT-specific study, and the analytic adjustments are appropriate. Three caveats matter for interpretation.

First, the design is cross-sectional. Children with cognitive or behavioral difficulties may be both more likely to receive heavy screen exposure (parental coping, fewer engaging alternatives) and more likely to score lower on outcome measures — a reverse-causation pathway that adjustment for baseline severity cannot fully eliminate in a single-time-point assessment.

Second, screen-time measurement was parent-reported recall, with the well-known biases of self-reported media use. Differential recall between high- and low-functioning children’s parents is plausible.

Third, the cohort was enrolled in the late 2000s and assessed in the mid-2010s, before the era of widespread tablet and smartphone use among young children. Whether the same effect sizes hold for current-era exposure patterns — shorter sessions, more interactive content, more tablet vs. TV — has not been directly replicated in the EPT population.

What the evidence does support is straightforward. Children born extremely preterm represent a high-vulnerability population in which dose, content, and context of screen exposure all plausibly matter. The Vohr finding aligns with the expected direction from general-population evidence and is large enough that it would survive substantial methodological correction.

Practical implications for families and clinicians

The actionable takeaways for families of EPT children are continuous with general pediatric guidance, with the dose recommendations weighted toward the conservative end of the range.

• Treat 2 hours/day as a meaningful threshold for school-age EPT children. Vohr 2021 used this cut-point and found a discrete effect; the AAP general recommendation for older children is similar.
• Avoid bedroom screens. The independent association with hyperactivity and impulsivity in Vohr 2021, the sleep-displacement evidence in general pediatric populations, and the loss of a screen-free environment all converge on this point.
• Prefer co-viewed and age-appropriate content over passive or background exposure. Madigan et al. (2020) provides the clearest evidence that context modulates the effect, and the protective effect of co-viewing is clinically actionable.
• Build language and literacy alternatives into daily routines. The mechanism that links screen exposure to language and white-matter outcomes is opportunity cost as much as direct effect — replacement activities matter.
• Screen for attention and executive-function difficulties separately. EPT children carry baseline risk on these dimensions regardless of media exposure, and identifying intervention needs early supports trajectory.

For clinicians, an annual conversation about screen time as part of EPT follow-up is well-supported. The evidence does not support framing screens as the dominant cause of EPT cognitive outcomes — neonatal medical complications and white matter integrity carry far more variance — but it does support framing screen exposure as a modifiable contributor that interacts with an already-vulnerable trajectory.

What remains uncertain

Several questions matter for refining the guidance and have not yet been answered.

The first is whether content type modifies the EPT-specific effect in the same way it modifies the general-population effect. Vohr 2021 measured total hours and bedroom presence but not content category. If interactive educational content is genuinely neutral or beneficial, the dose recommendation could be relaxed for that subset.

The second is whether the association is mediated by sleep, by sedentary behavior, by language-environment displacement, or by direct neurocognitive effects of attention-fragmenting media. These pathways suggest different interventions and have not been disentangled in the EPT population.

The third is whether early intervention — limiting screen exposure during specific developmental windows — produces measurable cognitive benefits in EPT children, or whether the screen-time association is a marker of broader environmental factors rather than a discrete causal pathway. Randomized evidence would resolve this question; observational follow-ups cannot.

The takeaway

Vohr et al. (2021) showed that high screen time (more than 2 hours per day) and bedroom screens are associated with lower full-scale IQ and more executive-function and behavioral problems in 6-to-7-year-old children born extremely preterm, after adjustment for the obvious confounders. The finding is consistent with the broader screen-time literature on language and brain structure, fits with the documented baseline vulnerability of EPT children that persists into adolescence, and supports practical recommendations of conservative dose limits, no bedroom screens, and a preference for co-viewed age-appropriate content. Screen time is not the largest determinant of EPT cognitive outcomes — neonatal medical course is — but it is the most modifiable.