The Role of Maternal Milk Feeding in Neurodevelopmental Outcomes

In neonatal intensive care units across the world, mothers of babies born too early are routinely encouraged to express breast milk to be fed to their infants by tube — sometimes for weeks or months before the infant can suckle directly. The clinical justification is partly nutritional, partly immunological, and partly cognitive: a substantial body of research suggests that babies who receive more of their mother’s own milk during this critical period have better cognitive and academic outcomes years later. A 2022 study by Belfort and colleagues in JAMA Network Open provides the most recent dose-response evidence: among 586 Australian preterm infants followed to age 7, every additional 25 mL/kg/day of maternal milk during NICU hospitalization was associated with measurable improvements in performance IQ, reading, math, and ADHD symptom scores. The effect was strongest for the most premature infants and showed a striking domain-specificity: full-scale IQ, verbal IQ, executive function, and behavior were not affected.

What the Belfort 2022 study actually found

The 2022 paper followed 586 infants born at less than 33 weeks’ gestation across five Australian perinatal centers, originally enrolled in the Docosahexaenoic Acid for Improvement of Neurodevelopmental Outcomes (DOMInO) study between 2001 and 2005. Maternal milk intake was measured precisely as volume per kilogram per day during NICU hospitalization, plus the total duration in months across the post-discharge period.

The dose-response findings, after adjustment for confounders, are unusually specific:

• Performance IQ: +0.67 points per additional 25 mL/kg daily (95% CI 0.10–1.23).
• Reading: +1.14 points per additional 25 mL/kg daily (95% CI 0.39–1.89).
• Math: +0.76 points per additional 25 mL/kg daily (95% CI 0.14–1.37).
• ADHD symptoms (Conners ADHD Index): −1.08 points per additional 25 mL/kg daily (95% CI −1.96 to −0.20).

For duration of feeding after NICU discharge:

• Reading: +0.33 points per additional month (95% CI 0.03–0.63).
• Spelling: +0.31 points per month (95% CI 0.01–0.62).
• Math: +0.30 points per month (95% CI 0.03–0.58).

What the study did not find is equally important. Maternal milk intake showed no association with full-scale IQ, verbal IQ, executive function (BRIEF), or behavior (Strengths and Difficulties Questionnaire). The effects are domain-specific. They concentrate on performance IQ (the visuospatial and processing-speed components of intelligence), academic achievement, and attention regulation. They do not extend to verbal abilities, broad executive function, or general behavioral problems.

The longer history of this evidence

The 2022 Belfort findings sit on top of three decades of converging research:

Lucas et al. (1992) in The Lancet conducted the original landmark demonstration. Comparing preterm infants who had received maternal milk in the first weeks of life to those who had not, the authors found an IQ advantage of 8.3 points at age 7.5–8 years that survived adjustment for maternal education and social class. The Lucas study established the basic phenomenon and made the clinical case for promoting maternal milk feeding in NICUs.

Belfort et al. (2016) in The Journal of Pediatrics extended the evidence with an Australian cohort of 180 infants born at less than 30 weeks’ gestation. They added brain imaging at term-equivalent age — predominant breast milk feeding (>50%) for additional days during the first 28 days was associated with larger deep nuclear gray matter volume (0.15 cc per additional day, 95% CI 0.05–0.25). This provided a structural-brain link between early maternal milk exposure and the cognitive outcomes observed at age 7. IQ, math, and working memory each gained roughly 0.5 points per additional day of predominant breast milk feeding.

Kramer et al. (2008) in Archives of General Psychiatry provided the strongest causal evidence available. The Promotion of Breastfeeding Intervention Trial (PROBIT) was a cluster-randomized trial across 31 Belarussian maternity hospitals. Hospitals (not infants) were randomly assigned to a Baby-Friendly Hospital Initiative breastfeeding-promotion intervention or standard care. The intervention raised exclusive breastfeeding at 3 months from 6.4% in the control group to 43.3% in the intervention group — and produced verbal IQ advantages at age 6.5 years in the children of intervention-group mothers. PROBIT remains the only large RCT-like evidence on breastfeeding and child cognition; the population was healthy term infants, not preterm.

The mechanism question: nutrition, infection, or relationship?

A persistent challenge in this literature has been distinguishing among three candidate mechanisms:

• Direct nutritional effects. Breast milk contains long-chain polyunsaturated fatty acids (DHA, ARA), oligosaccharides, hormones, growth factors, and other components that may support neural development directly.
• Indirect protective effects. Breast milk reduces the incidence of late-onset sepsis, necrotising enterocolitis (NEC), and other serious neonatal complications that themselves cause neurodevelopmental damage.
• Maternal-infant interaction effects. Mothers who provide maternal milk may differ in other behaviors — pumping schedules, NICU presence, post-discharge interaction patterns — that themselves shape cognitive development.

Lapidaire, Lucas, Clayden, Clark, and Fewtrell’s 2021 paper in Pediatric Research explicitly tested the second mechanism. Using a long-running preterm cohort followed to age 30, the authors found:

• Higher maternal and banked breast milk intake was associated with lower risk of neonatal infection and NEC.
• Neonatal infection and NEC were associated with lower full-scale and performance IQ at ages 7 and 30 years.
• The relationship between higher maternal milk intake and performance IQ at age 7 was partly mediated by reduced infection and NEC.
• The cognitive effects were stable from childhood into adulthood — they did not fade or amplify between ages 7 and 30.

The implication is that the cognitive benefits of maternal milk in preterm infants are not entirely explained by direct nutritional effects on neural development. A meaningful portion of the benefit operates through reduced acute medical complications during the NICU stay. This has practical relevance: it means even when full direct breastfeeding is impossible, providing maternal expressed milk during the NICU period continues to confer some cognitive benefit by reducing inflammatory and infectious insults to the developing brain.

Why preterm and term contexts differ

Most popular breastfeeding-and-IQ content draws on full-term breastfeeding observational data. The preterm context is meaningfully different:

• Maternal milk in NICU is delivered through expression and tube feeding, not direct breastfeeding. The maternal-infant interaction component of breastfeeding is largely absent during NICU stay.
• The brain is still actively developing during the NICU period. A 28-week-gestation infant in NICU is at a developmental stage where, in utero, the cortex is still rapidly differentiating. Nutritional and protective factors during this window have especially large potential consequences.
• The medical risks are concentrated. Preterm infants face concrete, measurable risks (sepsis, NEC, intraventricular hemorrhage) that maternal milk reduces. Term infants face few of these.
• Effect sizes are larger. Full-term observational studies typically report 1–4 IQ points associated with breastfeeding, often substantially attenuated after adjustment for confounders. Preterm studies report effects in the same range or larger and find them more robust to adjustment.

These differences matter for parents and clinicians: the maternal-milk-and-cognition story is more empirically secure in the preterm population than in the term population.

What practical advice this supports

Several practical implications emerge:

• For NICU staff and lactation consultants: Supporting mothers to express milk and feed it to their preterm infant produces measurable cognitive benefits with effect sizes on academic performance several times larger than most school-based interventions can claim. The case for institutional support of NICU lactation programs is strong.
• For parents of preterm infants: The dose-response findings mean partial supply matters. A mother who can produce only some maternal milk through expression still confers measurable benefits on her infant’s later cognitive trajectory. Maternal milk intake is not all-or-nothing.
• The benefit extends past the NICU. Belfort et al.’s duration findings — academic-achievement gains per additional month of feeding after discharge — argue against early termination of maternal milk feeding once the infant has come home.
• The benefits are domain-specific. A child whose preterm period was fed predominantly by maternal milk should be expected to show advantages in performance IQ, reading, math, and attention regulation — not in verbal IQ or general behavior. Setting realistic expectations matters for parents.
• Donor milk vs. maternal milk distinction. Lapidaire et al.’s findings, along with much of the broader literature, suggest banked donor milk reduces NEC/infection risk but may not be equivalent to maternal milk for some cognitive outcomes. Donor milk is an important fallback when maternal supply is insufficient; it is not a full substitute.

What the evidence does not establish

Several limits to interpretation:

• Most evidence is observational. The Lucas, Belfort, and Lapidaire studies all involve adjustment for measured confounders but cannot fully eliminate residual confounding by maternal characteristics, family resources, or post-discharge environment. PROBIT is the only randomized evidence and is in term infants.
• Mechanism remains incompletely identified. The relative contributions of direct nutritional effects, infection/NEC reduction, and unmeasured maternal-infant interaction differences are not fully resolved.
• The effects are real but bounded. A 1-point IQ gain per 25 mL/kg/day, integrated over a typical NICU stay, produces meaningful but not transformative effects. The literature does not support claims that maternal milk feeding makes the difference between average and gifted cognitive outcomes.
• Generalization across populations. Most evidence comes from high-resource healthcare settings with strong NICU programs. Generalization to preterm infants in lower-resource settings is plausible but not directly demonstrated.
• Why some domains and not others? The Belfort 2022 finding that maternal milk affects performance IQ but not verbal IQ, and academic skills but not executive function, is striking but not yet fully understood. Whether this reflects specific neural-developmental mechanisms or measurement-related artifacts is an open question.

Frequently Asked Questions

Does breast milk really raise IQ in preterm babies?

The evidence is consistent and longstanding. Lucas et al. (1992) found 8.3 IQ points; Belfort et al. (2016, 2022) found dose-response effects on multiple cognitive and academic measures. The effects are most robust for performance IQ, math, reading, and attention regulation, less so for verbal IQ and executive function.

How much milk is enough?

The 2022 Belfort study found dose-response effects per 25 mL/kg/day. There is no clear threshold below which benefits disappear; more is associated with more benefit, and per-month duration after discharge continues to add value.

What about donor breast milk?

Donor milk reduces neonatal infection and NEC (one of the mechanisms by which milk affects later cognition) but may not produce identical cognitive benefits to maternal milk. The Lapidaire et al. analysis suggests donor milk works partly through the infection-reduction pathway. For mothers unable to express enough, donor milk is the next best option but does not fully substitute.

Why are the benefits stronger for the most premature infants?

Two probable reasons. First, the most premature infants face the highest risk of NEC and serious infection — the conditions maternal milk most clearly prevents. Second, their brains are at the most sensitive developmental stage when the nutritional and protective inputs of maternal milk arrive.

Do these effects last?

Lapidaire et al. tracked the effects to age 30 and found them stable. The cognitive advantages associated with maternal milk feeding in the preterm period do not appear to fade with development.

Is the maternal-milk-and-cognition link causal?

For preterm infants, the evidence base is observational with strong consistency across studies and decades, plus mechanistic plausibility through NEC/infection reduction. PROBIT provides randomized evidence in term infants. A clean causal claim requires randomized intervention data on maternal milk in preterm infants specifically, which for ethical and feasibility reasons is difficult to obtain.

What about formula-fed preterm babies?

The studies generally compare relative milk volumes, not maternal-milk vs. zero-milk groups. A formula-fed preterm infant whose mother could not produce milk is not categorically destined for poor outcomes; the effects are statistical and population-scale, and many factors beyond nutrition shape an individual child’s trajectory.