Automated detection of abnormal general movements from pressure and positional information in hospitalized infants.

BACKGROUND: Prechtl's general movements assessment (GMA) allows visual recognition of movement patterns that, when abnormal (cramped synchronized, or CS), have very high sensitivity in predicting later neuromotor disorders; however, training requirements and subjective perceptions from some clinicians may hinder universal adoption of the GMA in the newborn period. METHODS: To address this, we used a three-phased approach to design a preliminary and clinically-oriented approach to automated CS GMA detection. 335 hospitalized infants were dually recorded on video and a pressure-sensor mat that collected time, spatial, and pressure data. Video recordings were scored by advanced GMA readers. We then conducted a series of unsupervised machine learning and supervised classification modeling with features extracted from clinician- and mat-driven datasets. Finally, the resulting algorithm was converted to a software interface. RESULTS: A classification model combining normalization, clustering, and decision tree modeling resulted in the highest sensitivity for CS movements (100%). Results were delivered via the software interface within 20 min of data recording. CONCLUSION: The combination of clinical research, machine learning, and repurposing of existing sensor mat technology produced a feasible preliminary approach to automatically detect abnormal GMA in infants while still in the NICU. Further refinements of software and algorithms are needed. IMPACT STATEMENT: Machine learning can differentiate cramped synchronized general movement patterns in the neonatal intensive care unit with good sensitivity and specificity. Increasing access to the GMA through automated detection methods may allow for earlier identification of a greater number of children at high risk for movement delay. Large studies leveraging new artificial intelligence approaches could increase the impact of such detection.

Vibration-based mitigation of noxious-evoked responses to skin puncture in neonates and infants: a randomised controlled trial.

OBJECTIVE: To assess the effect of a non-noxious vibratory stimulus on noxious-evoked cortical responses to skin puncture and to determine whether the presence of certain behavioural components may be used to predict such cortical responses. DESIGN: Randomised controlled trial. SETTING: Level IV neonatal intensive care unit at a stand-alone children's hospital. PATIENTS: 134 hospitalised infants between 36 and 52 weeks' postmenstrual age and ordered to receive a clinically required laboratory draw. INTERVENTIONS: Infants randomised to receive the intervention, a vibratory stimulus at the site of skin puncture beginning 10 s prior to a heel stick, or the control, no vibration. MAIN OUTCOME MEASURES: Electroencephalography and video recording time-locked to the deployment of the lancet for the skin puncture. Noxious-evoked cortical responses were measured by the area under the curve in the somatosensory region contralateral to the skin puncture. Behavioural responses were coded through video analysis. RESULTS: Noxious-evoked cortical responses were significantly reduced in participants receiving the vibratory stimulus compared with the control (frontal, p<0.0001; central, p=0.0088; central-parietal, p=0.0111). There were no significant differences in behavioural responses between groups (all p>0.05). CONCLUSIONS: A non-noxious vibratory stimulus presented prior to and continuing simultaneously with skin puncture significantly mitigates nociception in hospitalised infants. The presence or absence of facial expression components is inadequate to reliably predict pain signalling in the brain. TRIAL REGISTRATION NUMBER: NCT04050384.

Contingent Mother's Voice Intervention Targeting Feeding in Hospitalized Infants with Critical Congenital Heart Defects.

Infants with critical congenital heart defects (CCHD) are at high risk for feeding challenges and neurodevelopmental delays; however, few interventions promoting the neurodevelopmental progression of feeding have been studied with this population. Contingent mother's voice has been successfully used as positive reinforcement for non-nutritive suck (NNS) in studies with preterm infants, leading to improved weight gain and more rapid cessation of tube feedings; however, this type of intervention has not been studied in infants with CCHD. This study aimed to determine whether an NNS-training protocol using the mother's voice as positive reinforcement and validated in preterm infants could improve oral feeding outcomes in hospitalized infants with CCHD undergoing cardiac surgical procedures. Infants were randomized to receive the contingent mother's voice intervention before or after cardiac surgery, with a control comparison group receiving passive exposure to the mother's voice after surgery. There were no significant differences in discharge weight, PO intake, length of stay, time to full feeds, or feeding status at 1-month post-discharge between infants who received contingent mother's voice compared to those who did not. There were significant differences in PO intake and time to full feeds following surgery based on infants' pre-enrollment PO status and severity of illness. At 1-month post-discharge, parents of infants in the intervention group expressed a higher rate of positive feelings and fewer concerns regarding their infant's feeding compared to parents of infants in the control group. While the current protocol of 5 sessions was not associated with improved feeding outcomes in infants with CCHD, it empowered parents to contribute to their infant's care and demonstrated the feasibility of using the mother's voice as positive reinforcement for infants with CCHD. Further study of timing, intensity, and duration of interventions leveraging the mother's voice in this population is needed. ClinicalTrials.gov Identifier: NCT03035552.