Personalized Music Therapy for Preterm Infants

Background and rationale. Preterm infants in the Neonatal Intensive Care Unit (NICU) are exposed to noise, light, and frequent handling, which can affect physiology and development and increase parental distress. Family-centered music therapy using the parental voice is a low-risk intervention that may modulate cerebral oxygenation and autonomic parameters during exposure. Neonatal NIRS studies report small-to-moderate average changes in cerebral oxygenation with parental voice or music and heterogeneous individual responses, supporting standardized procedures and within-participant designs (van Bel 2008; Alderliesten 2016; Meder 2021; Yue 2021).

Ethics and consent. Enrollment proceeds under Ethics Committee approval on file; written informed consent is obtained from the parent(s) or legal guardian(s) per local policy. Consent from both parents is sought when applicable; a single signature is acceptable in single-parent or court-sanctioned guardianship cases according to site policy.

Design. Single-center, randomized 2 by 2 crossover pilot. Infants are randomized to Sequence AB (Period 1 = PMT, washout 2 days, Period 2 = Quiet Rest) or Sequence BA (Period 1 = Quiet Rest, washout 2 days, Period 2 = PMT).

Pre-intervention voice preparation (outside study periods). A trained music therapist records the parent's singing or voice and performs basic audio processing (smoothing peaks, removing artifacts, setting a target intensity). The audio files are stored securely for audio playback during PMT. No outcomes are collected during preparation.

Conduct of each period (50 minutes, 8:00 to 9:00 PM local time, low-stimulation NICU room). A bedside speaker is positioned approximately 30 cm from the infant's head; audio playback is set to approximately 45 dBA (plus or minus 3) at the infant's ear level. Each period follows the same structure:

• Phase 1 (0 to 9 minutes): Acclimation Rest (no stimulation)
• Phase 2 (10 to 40 minutes): Treatment Exposure (PMT or Quiet Rest, per randomized sequence)
• Phase 3 (40 to 50 minutes): Post-Exposure Rest (no stimulation) Measurements within each period. TOI (NIRS), HR, RR, and SpO2 are summarized as 5-minute window means at minutes 5 to 9 (Phase 1), minutes 21 to 25 (Phase 2), and minutes 46 to 50 (Phase 3). A 2-day washout separates periods to mitigate carryover.

Stopping criteria and resumption. Stop immediately if any of the following occurs: SpO2 below 85 percent for 30 seconds or more, or below 88 percent persistent; HR below 100 beats per minute for 30 seconds or more, or above 200 beats per minute persistent; apnea 20 seconds or longer, or RR 80 breaths per minute or higher persistent; cyanosis, marked pallor, hypotonia, or inconsolable irritability or crying; urgent procedures or non-deferrable handling; intolerance to the acoustic stimulus (for example, marked agitation or avoidance); parent request, if present. Resumption occurs only after clinical reassessment by the attending neonatologist or covering clinician and stabilization (parameters within safety ranges for 5 minutes or more). Nurses and physicians are authorized to stop according to criteria.

Parental stress. At discharge, parents complete the PSS:NICU as a self-report questionnaire assisted by trained NICU staff (for example, a nurse or psychologist).

Sample size. This 2 by 2 crossover pilot focuses on feasibility and estimation of the within-participant variance and effect magnitude on TOI. Based on NICU NIRS literature-small average changes in cerebral oxygenation with parental voice or music and non-trivial within-participant variability unless procedures are standardized-a clinically relevant within-participant mean difference (delta) of about 3 TOI percentage points (PMT minus Quiet Rest during minutes 21 to 25) and an expected within-participant standard deviation of approximately 4 to 5 points are assumed under standardized procedures (fixed time windows, low-stimulation room, consistent sensor placement, artifact rules). With 20 infants, the approximate 95 percent confidence-interval half-width is about 1.96 divided by the square root of 20 times the standard deviation (about 0.44 times SD), yielding roughly plus or minus 1.8 to 2.2 percentage points if SD is 4 to 5. This precision is acceptable for a pilot focused on estimation and feasibility and will inform sizing of a subsequent confirmatory trial (van Bel 2008; Alderliesten 2016; Meder 2021; Yue 2021).

Statistical analysis. The primary estimand is the within-participant mean difference (PMT minus Quiet Rest) in TOI during minutes 21 to 25 (5-minute window). The primary analysis uses a linear mixed-effects model with fixed effects for treatment (PMT versus Quiet Rest), period (1 versus 2), and sequence (AB versus BA), and a random intercept for participant. First-order carryover is assessed; if present, a sensitivity analysis restricted to Period 1 is reported. Secondary endpoints are exploratory and are analyzed analogously; multiplicity adjustments are not applied unless specified in the Statistical Analysis Plan.