Sounds in the NICU: Protecting Your Baby from Harmful Noise and Giving Them the Voices They Need

What four new research studies say about noise, silence, and the sound of a parent's voice for babies born early

Hospital neonatal intensive care units (NICUs) are loud places, and that noise can be hard on a baby whose hearing and brain are still developing. But new research also suggests that babies born early need to hear human voices, especially their parents', and that a quiet room is not automatically a better room. Four recent studies help explain how to lower the sounds that cause stress while adding the sounds that help.

Why sound matters so much for a baby who arrives early

Before birth, a baby hears a muffled, steady world. The mother's body and the fluid around the baby soften the sounds from outside, so what gets through is mostly low and rhythmic: her heartbeat, her breathing, and her voice. Researchers estimate that this soundscape rarely goes above about 30 decibels, roughly as quiet as a whisper [1]. A baby's hearing system starts working during the middle months of pregnancy, and the connections between the ear and the brain are largely in place by about 30 weeks [1][2].

A baby born early is suddenly in a very different place. The NICU is full of lifesaving equipment: heart monitors, oxygen machines, ventilators, and alarms, along with the voices and footsteps of a busy care team. Studies going back many years linked loud sound in the NICU to changes in babies' heart rate, breathing, and sleep [3]. For that reason, the American Academy of Pediatrics suggests that NICU sound stay at or below about 45 decibels by day and 35 at night, about as loud as a quiet library or a quiet bedroom [1].

How we came to worry about too much noise, and then about too little voice

For many years, the main goal was to make NICUs quieter. Units began "quiet times," trained staff to speak softly, and built private rooms for babies and families. Those steps made good sense.

Later, researchers noticed a different problem. Babies born early hear fewer words than babies still in the womb [4], and studies of NICUs found that language exposure there is often limited [5]. In some units, babies in private rooms heard fewer words than babies in open rooms with more activity [6][7]. Researchers also found that when mothers' voices were played or when parents talked to their babies, babies' development seemed to benefit [8][9]. So scientists began asking a more careful question: how can we remove harmful noise without also removing the human sounds that babies need?

What the research found about how loud NICUs still are

A research team in Vienna and colleagues combined the results of 113 earlier studies from around the world that had measured NICU noise with sound meters [1]. In nearly all of them, the sound was louder than the recommended 45 decibels, and some brief sounds topped 100 decibels, about as loud as a power tool. Alarms, staff talking, incubator motors, and opening or closing incubator doors all added to the noise. The highest sound often came from inside the incubator itself, from breathing-support machines, which means the incubator does not fully shield a baby from noise [1].

The team also noted limits. The studies measured sound in many different ways, so the results could not be combined into a single number, most of them measured sound without following babies' health later, and the search included only English-language studies [1].

A second study looked at one hospital room in Damascus, Syria, where several incubators sat close together [10]. Researchers placed four sound meters in the corners of the room and recorded sound over 20 days, noting what was going on at the time. Even the room's quietest moments were about 49 to 51 decibels, above the recommended level. Alarms and ventilator alarms were among the loudest sounds, and busy staff movement around noon was the noisiest time of day. Researchers also showed that a computer program could tell what type of event was happening, such as a door or an alarm, from the sound alone, about 85 times out of 100. They say this is not meant to replace alarms, but may help units find and fix their noisiest habits [10]. This was one unusually crowded room, so other units may differ.

What sound exposure looked like for the smallest babies

A team in the United States followed 64 babies born at 28 weeks or earlier in a large NICU connected to Washington University School of Medicine [2]. A small recorder placed within two feet of each baby's ear, on the incubator or crib, collected 16 hours of sound at four times between birth and the time the baby would have been due. Near the baby's due date, a trained examiner checked how each baby responded to sights and sounds, how relaxed or stressed the baby was, and how the baby's muscles felt. Forty-four babies had this exam; ten of the original 64 babies sadly died before going home [2].

Babies who had been exposed to higher average sound levels were less able to focus on sights and sounds, and those exposed to higher peak levels had stiffer muscles. Babies exposed to more silence also had stiffer muscles, while babies who heard more adult words around them looked less stressed [2]. These are connections, not proof of cause. Sicker babies are around more machines, and that alone could explain some of the findings. The group is also small, from one hospital [2]. Even so, the results fit what many families and nurses feel instinctively: loud, harsh sounds are not helpful, but a silent bedside is not ideal either.

Can a recording of a parent's voice help?

Many parents cannot sit at the bedside all day. They may have other children, work, or recovery from birth, or live far away. Recordings of a parent's voice could help fill the gap [8]. In earlier research, babies appeared to learn most when the voice recording played in response to something the baby did, instead of just playing in the background [11][12][13].

A research team in Atlanta, Georgia, tested this directly in two hospitals' NICUs, randomly assigning babies to one of two approaches [14]. They included babies born before 35 weeks who had reached about 32 to 36 weeks of pregnancy-equivalent age. Caregivers recorded their own voices, including lullabies, nursery rhymes, and children's books. Half the babies heard their parent's voice only when the baby sucked on a pacifier, so the baby "switched on" the voice. The other half heard the same recording playing continuously. Each session lasted about 20 minutes, two to three times a day, over two to three weeks. The volume was kept low, at or under 55 decibels [14].

Of 214 babies who joined the study, 191 finished it. When the team measured brain activity with an EEG net during short speech sounds, the babies who had turned on the voice themselves showed a stronger brain response to speech than the babies who heard it passively. The difference was statistically detectable but small. Strikingly, babies who "switched on" the voice actually heard it for less total time, which suggests that taking part mattered more than just hearing more [14]. The benefit did not differ meaningfully by crib type, room type, background sound level, or whether a baby needed breathing support [14].

What this means for your family

Families do not need to choose between "quiet" and "talking." The research points toward both. You can help by speaking, reading, and singing softly to your baby during and around care, in your own voice, even if you cannot be there all day. Staff can help by cutting down avoidable loud sounds such as door slams, careless handling of equipment, and loud conversations at the bedside [1][10]. If you are in a private room, you may want to ask the team how the unit makes sure your baby still hears people [6][7].

It is important to be realistic. The recorded-voice study measured how babies' brains respond to speech sounds. It did not yet show that children later learn to talk better, and the team plans to follow up [14]. The device was made by a company one of the study's authors helped found, which the authors openly reported and managed with safeguards [14]. And the noise studies describe patterns; they cannot tell us how much any one baby was harmed by sound [1][2].

If your baby is in the NICU, you can ask simple questions: How loud is the unit, and has anyone measured it? What can we do to keep my baby's space calm? When and how can I talk or sing to my baby? Is there a way for my baby to hear my voice when I am away? Your voice is not noise; babies appear to benefit from hearing familiar voices.

What researchers are working on next

Researchers want to test quieter breathing machines and better-designed alarms, track hearing and language into childhood, and learn how much voice time is best [1][2][14]. They also want to find out whether tools that map sound in a unit can help staff keep rooms calmer while still allowing the conversation and singing that help babies grow [10]. For now, the guiding idea is simple: protect babies from sounds that stress them, and surround them with the voices that comfort them.

References

  1. Giordano V, Mandl S, Bartha-Doering L, et al. What's that noise—tackling sound pollution in the NICU: a systematic review. Front Pediatr. 2026;14:1798036. doi:10.3389/fped.2026.1798036 ↩
  2. Pineda R, Woodward LJ, Vesoulis ZA, et al. Language and sound exposure across neonatal intensive care hospitalization and relationships with early outcome. J Perinatol. 2026;46(8):1482-1488. doi:10.1038/s41372-026-02623-y ↩
  3. Wachman EM, Lahav A. The effects of noise on preterm infants in the NICU. Arch Dis Child Fetal Neonatal Ed. 2011;96(4):F305-F309. doi:10.1136/adc.2009.182014 ↩
  4. Monson BB, Ambrose SE, Gaede C, et al. Language exposure for preterm infants is reduced relative to fetuses. J Pediatr. 2023;262:113344. doi:10.1016/j.jpeds.2022.12.042 ↩
  5. Best K, Bogossian F, New K. Language exposure of preterm infants in the neonatal unit: a systematic review. Neonatology. 2018;114(3):261-276. doi:10.1159/000489600 ↩
  6. Pineda R, Durant P, Mathur A, et al. Auditory exposure in the neonatal intensive care unit: room type and other predictors. J Pediatr. 2017;183:56-66.e3. doi:10.1016/j.jpeds.2016.12.072 ↩
  7. Liszka L, Smith J, Mathur A, et al. Differences in early auditory exposure across neonatal environments. Early Hum Dev. 2019;136:27-32. doi:10.1016/j.earlhumdev.2019.07.001 ↩
  8. Provenzi L, Broso S, Montirosso R. Do mothers sound good? A systematic review of the effects of maternal voice exposure on preterm infants' development. Neurosci Biobehav Rev. 2018;88:42-50. doi:10.1016/j.neubiorev.2018.03.009 ↩
  9. Caskey M, Stephens B, Tucker R, et al. Importance of parent talk on the development of preterm infant vocalizations. Pediatrics. 2011;128(5):910-916. doi:10.1542/peds.2011-0609 ↩
  10. Mukhaiber H, Ali K, Ismaiel E. Decoding and spatial mapping of acoustic noise in the neonatal intensive care unit. Sci Rep. 2025;15(1):32039. doi:10.1038/s41598-025-17707-6 ↩
  11. Richard C, Jeanvoine A, Stark AR, et al. Randomized trial to increase speech sound differentiation in infants born preterm. J Pediatr. 2022;241:103-108.e3. doi:10.1016/j.jpeds.2021.10.035 ↩
  12. Maitre NL, Kjeldsen CP, Jeanvoine A, et al. The MIND randomized controlled trial: an intervention to improve neural speech processing and 2-year language outcomes of infants born preterm. J Pediatr. 2026;296:115187. doi:10.1016/j.jpeds.2026.115187 ↩
  13. Kjeldsen CP, Neel ML, Stark AR, et al. Contingent mother's voice intervention facilitates attention in hospitalized preterm infants with neural insults. Mind Brain Educ. 2025;19(1):37-46. doi:10.1111/mbe.70000 ↩
  14. Kjeldsen CP, Moran M, Jeanvoine A, et al. Recorded contingent caregiver voice independently improves neural speech processing in hospitalized preterm infants: a multisite randomized controlled trial. J Clin Med. 2026;15(17):6880. doi:10.3390/jcm15176880 ↩