Automatic Feature Discrimination of Non-Nutritive Suck Dynamics Among Extremely Preterm Infants.

Prematurity is the leading cause of death in the first month of life. Neonates are forced to engage the extrauterine environment with an underdeveloped nervous systems, with limited connectivity between orofacial sensorimotor anatomy, brainstem pattern-generating circuits, and sensory-driven thalamocortical inputs to layer IV of the neocortex. This is correlated with an inability to produce the complex neuromotor behaviors of oral feeding. The NTrainer System is an assessment and treatment tool that promotes ororhythmic motor patterning in preterm infants though pulsed orocutaneous stimulation to improve NNS. This report represents an interim analysis of NNS development among 42 extremely preterm infants enrolled in an ongoing randomized control trial. Participants, diagnosed with RDS or BPD, were randomized to receive the pulsed orocutaneous or non-pulsed treatment. Digitized records of NNS dynamics were automatically processed using a new Python based software program, NeoNNS, that extracts frequency domain features of NNS performance. Linear mixed modeling was used to examine the effect of treatment type among several dependent variables for 817 NNS files. Significant treatment effects were shown for NNS cycles/min, NNS cycles/burst, max NNS cycles/burst, and NNS amplitude.Infants’ NNS performance increased for some measures when the SHAM was utilized, while the number of bursts per minute and compression cycle amplitude increased the NTrainer therapy. A greater increase in variable measures was observed in RDS infants who received the SHAM. Infants’ diagnosis in combination with treatment as well as sex resulted in no significant effects across all parameters. All measures were highly dependent on PMA. Advisor: Steven M. Barlo