The study seeks to determine the efficacy of non-nutritive suck (NNS) training using a pacifier-activated device (PAM) with mothers' voice to condition suck-strength and rhythmicity, in improving the feeding and developmental outcomes of infants at high-risk for CP.

Poor neuromotor and sensory function of the aerodigestive system in children with CP often originates in the neonatal period, when they are still classified as "high-risk for CP". Characteristic neuroimaging abnormalities including severe intraventricular (IVH) hydrocephalus and periventricular leukomalacia (PVL), stroke or ischemia with lesions affect the posterior limb of the internal capsule are strong Indicators of high-risk for CP, especially when combined with abnormal General Movements Assessment (Guidelines for Early Detection of CP; Stockholm, 2016). Early intervention, when plasticity is greatest has the largest impact on functional recovery in CP. While intervening in infancy involves treating some infants who will not develop CP, the goal is to establish new neuronal connections and functional patterns before less efficient adaptations can occur. However, no current interventions target the oral-motor dysfunction of infants at high-risk for CP, before their discharge from the NICU.

Evidence for behavioral interventions in feeding disorders for children with CP ranges from insufficient to moderate, with a clear need for rigorous studies. In healthy preterm and late-preterm infants, oromotor practice opportunities such as non-nutritive suck (NNS) are safe and promote effective suck-swallow-breathe patterns, with decreased time to achieving oral feeds. While NNS opportunities are frequent in most NICUs, they must be adapted for effectiveness in infants at high-risk for CP. Motor learning in these infants must incorporate repetitive, self-initiated and task-directed activities. Learning is optimized when contingent on feedback, such as positive reinforcement.

NNS training (rather than simple exposure) has been implemented successfully using rhythmic sound of mother's voice singing contingent upon suck strength and pattern, as detected by a pacifier-sensor device (Pacifier-activated music; PAM) in an cohort of predominantly healthy preterm infants. An RCT demonstrated that NNS-trained infants had feeding tubes removed one week earlier than controls, with fewer aspiration events and feeding difficulties in infancy. The intervention was promising in the dozen infants with significant neural injury. Following this preliminary data, this study seeks to further determining the efficacy of non-nutritive suck (NNS) training using a pacifier-activated device (PAM) with mothers' voice to condition suck-strength and rhythmicity, in improving the feeding and developmental outcomes of infants at high-risk for CP.

Inclusion Criteria:

• PMA at enrollment >32 6/7 and <42 AND
• Diagnoses of Grade III/IV IVH, hydrocephalus or PVL, neonatal encephalopathy or thrombosis/infarct with involvement of the posterior limb of the internal capsule OR
• Abnormal GMA by certified study staff and masked gold-standard reading

Exclusion Criteria: Infants cannot be

• on assisted ventilation
• receiving more than 50% of their feeds orally averaged over the 72 hours prior to study start
• receiving less than 50% of their feeds enterally
• receiving enteral feed for greater than 60 min.
• considered medically unsafe to feed orally by the medical or feeding team as defined by unit/clinical protocol