In the NICU, It Is Fathers’ Singing That Wakes the Preterm Brain

When a baby arrives weeks or months before its due date, the neonatal intensive care unit takes over an environment that evolution spent millions of years engineering. Instead of the muffled, rhythmic world of the womb, dominated by the low-frequency wash of the mother’s voice and the steady thump of her heartbeat, the preterm infant meets the clatter of pumps, the shrill of alarms and the hum of monitors. Singing has long been offered as a way of handing some of that lost soundscape back, and a growing body of evidence suggests it can calm infants, stabilize physiology and support early brain development. Yet almost every singing intervention in neonatal care has been built around a single, rarely questioned premise: that the mother’s voice is the one that matters most. A new registered trial, published in Pediatric Research, put that premise to a direct experimental test, and the results point in a direction almost nobody in the field expected.

The study, led by Efthymios Papatzikis of the Canadian University Dubai together with colleagues working at the Panagiotis and Aglaia Kyriakou Children’s Hospital in Athens, was designed to disentangle two factors that are usually tangled together in real life: kinship and voice type. When a mother sings to her preterm infant, her baby is hearing both a familiar, genetically related person and a female voice with a characteristic fundamental frequency, the acoustic parameter known as F0. Previous work could not say which of these signals the developing cortex was actually responding to. The research team therefore recruited preterm infants and recorded their electroencephalographic activity while each child heard the same block of stimuli delivered by four different facilitators: the infant’s own mother, the infant’s own father, a female music therapist and a male music therapist. Crossing kinship with pitch in this way meant that any difference in the brain’s response could be traced to whether the singer was a parent, whether the singer was male or female, or some interaction between the two.

The team pre-registered the trial, protocol NCT06398912, and committed in advance to a primary outcome: delta-band power in the quantitative EEG. Delta activity, the slowest and highest-amplitude rhythm visible in the newborn electroencephalogram, is a well-established marker of cortical engagement and maturation in preterm infants, and quantitative EEG features of this kind have been linked in systematic reviews to long-term neurodevelopmental outcomes. The pre-registered hypothesis was straightforward and reflected decades of literature on fetal learning and newborn voice recognition: the maternal voice would dominate. Infants famously prefer their mother’s low-pass filtered voice within hours of birth, and studies of fetal learning have shown that the prenatal soundscape leaves measurable traces in the newborn brain. If any voice was going to light up the preterm cortex, the reasoning went, it would be the mother’s.

The hypothesis was not supported. When the researchers analyzed the recordings, the mother’s apparent lead dissolved once each session was referenced to its own baseline of silence, a methodological control that matters enormously in an environment as acoustically noisy as an intensive care unit. Without that per-session baseline, any facilitator who happened to sing during a calmer stretch of the infant’s state could appear more engaging than one who sang during a restless one. Referencing each condition to its own preceding silence stripped that confounder away, and what remained was a pattern that ran in the opposite direction from the prediction. During the lullaby, the most complex stimulus in the block and, because of the fixed presentation order, the one that arrived last, when infant engagement was already waning, the father’s voice drew the highest delta power of any facilitator. Taken together, the two male voices outdrew the female voices across the cohort.

The finding survived the team’s statistical scrutiny. The researchers used linear mixed-effects modeling to handle the hierarchical structure of the data, with repeated measurements nested within individual infants, and confirmed the key results under bootstrap resampling, a technique that estimates the stability of an effect by repeatedly re-estimating it from random subsamples of the data. In other words, the male-voice advantage was not an artifact of a few unusually responsive infants or of a single session that happened to go well. The pattern held across the cohort. To the authors’ knowledge, this is the first full-cohort quantitative EEG evidence that a singer’s biological sex, rather than maternal kinship, tracks the preterm cortical response to song.

Why would a low, spectrally full male voice engage the preterm cortex more powerfully than a higher female voice? The authors point to the physics and physiology of the human vocal apparatus. Adult male vocal tracts are longer and vocal folds longer and heavier than those of adult females, producing a lower fundamental frequency and a richer spectral envelope, with more acoustic energy distributed across the lower frequency bands. Titze’s classic acoustic analyses of male and female voices documented these differences in detail, and Fant’s work on vocal tract size factors explains how tract length reshapes the entire frequency spectrum of speech and song. A preterm auditory system, still calibrating itself to sound after an abrupt exit from the womb, may find this dense, low-frequency signal easier to encode, or simply more salient, than a brighter, higher-pitched one. Intriguingly, the fetus in the third trimester hears the maternal voice mainly in low-pass filtered form, with high frequencies attenuated by tissue and fluid, so the intrauterine template of a voice is itself dominated by low-frequency energy.

The study is not without caveats, and the authors are careful about what their design can and cannot claim. The lullaby condition, where the father’s advantage was most visible, was also the last stimulus in a fixed order, delivered against declining infant engagement, which means the result could partly reflect a differential ability of the male voice to re-capture attention in a tired or drowsy infant rather than a simple superiority across all states. The delta-band measure captures short-term cortical response, not long-term developmental benefit, and the qEEG data themselves are held under privacy restrictions typical of infant clinical research, available from the corresponding author on reasonable request. Replication in independent cohorts, ideally with randomized stimulus orders and larger samples, will be needed before the finding can reshape clinical practice. The trial was approved by the hospital’s institutional review board and conducted with written informed parental consent, and the authors report no competing interests.

Even with those qualifications, the implications are striking, because they land on one of the most overlooked figures in modern neonatal medicine. Family-centered care models, including the large multinational cluster-randomized trial of Family Integrated Care published in The Lancet Child and Adolescent Health, have demonstrated real benefits when parents are woven into the daily care of their hospitalized infants, yet the operational reality of most NICUs remains built largely around the mother: kangaroo care protocols, lactation support, visiting schedules and, above all, music and singing programs staffed and delivered with maternal bonding as the explicit goal. Fathers, by contrast, are frequently sidelined, their role in early preterm care under-specified and under-studied. If the preterm cortex responds preferentially to the low, spectrally full male voice, then fathers may hold a defined, testable, biologically grounded role in their own infant’s care, one that goes far beyond being a supportive presence at the incubator side.

The broader lesson may be about how science handles its most comfortable assumptions. The maternal-dominance hypothesis in this trial was not a fringe idea; it was the reasonable synthesis of forty years of research on fetal learning, newborn voice preference and early auditory plasticity, including landmark studies showing that mother’s voice and heartbeat sounds elicit auditory plasticity before full gestation and that infants learn speech melodies in the womb. The field’s singing interventions inherited that framework wholesale. What the Athens team did was insist on the control that could falsify it: reference every condition to its own silence, cross kinship against voice type, pre-register the prediction, and let the delta rhythm speak. The preterm brain answered, and it answered for the fathers. For the hundreds of thousands of families who pass through neonatal intensive care each year, the suggestion that a father’s lullaby is not merely permitted but may be neurologically potent is a message worth singing about, and a hypothesis that neonatal science now has every tool to test.

Subject of Research: Quantitative EEG study of how maternal versus male singing voices shape cortical responses in preterm infants in neonatal intensive care.

Article Title: Fathers, not mothers: the male singing voice and the preterm brain’s response to song in the NICU

Article References: Papatzikis, E., Dimitropoulos, K., Tataropoulou, K., Pasoudi, E., & Nika, A. (2026). Fathers, not mothers: the male singing voice and the preterm brain’s response to song in the NICU. Pediatric Research. https://doi.org/10.1038/s41390-026-05551-x

Image Credits: AI Generated

DOI: 10.1038/s41390-026-05551-x

Keywords: preterm infants, NICU, EEG, delta-band power, singing, father's voice, maternal voice, music therapy, fundamental frequency, quantitative EEG, family-centered care, Pediatric Research