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.

Sensitivity and specificity of the Neonatal Visual Assessment to predict motor and cognitive outcomes in infants born very preterm.

BACKGROUND: Very preterm infants are at increased risk of neurodevelopmental impairments. The Neonatal Visual Assessment (NVA) assesses visual function and outcomes and has been used to assess early neurodevelopmental outcomes. This study aimed to compare NVA results of very preterm and term-born infants and to calculate the sensitivity and specificity of the NVA at term equivalent age (TEA) and three months corrected age (CA) to predict motor and cognitive outcomes at 12 months CA in very preterm infants. METHODS: This prospective observational cohort study recruited infants born before 31 weeks gestation and a healthy term-born control group. The NVA was assessed at TEA and three months CA, and neurodevelopmental outcomes (Bayley Scales of Infant and Toddler Development, Third Edition; Neurosensory Motor Developmental Assessment; Alberta Infant Motor Scale) were performed at 12 months CA. The sensitivity and specificity of the NVA to predict outcomes were calculated based on a previously published optimality score. RESULTS: 248 preterm (54 % male) and 46 term-born infants (48 % male) were analysed. The mean NVA scores of preterm and term-born infants were significantly different at TEA (preterm 3.1±2.1; term-born 1.2±1.7, p < 0.001). The NVA had moderate sensitivity (59-78 %) and low specificity (25-27 %) at TEA, and low sensitivity (21-28 %) and high specificity (86-87 %) at three months CA for the prediction of preterm infants' outcomes at 12 months CA. CONCLUSION: The NVA at TEA and three months CA was not a strong predictor of motor and cognitive impairments in this contemporary cohort of very preterm infants.

Correlation between Early Visual Functions and Cognitive Outcome in Infants at Risk for Cerebral Palsy or Other Neurodevelopmental Disorders: A Systematic Review.

Early key visual skills, such as tracking objects, sustaining gaze, and shifting attention, rapidly develop within the first 6 months of infant life. These abilities play a significant role in the development of cognitive functions but are frequently compromised in infants at risk of developing neurodevelopmental disorders. This systematic review evaluates the potential of early vision function in the prediction of cognition at or above 12 months. Five databases were searched for relevant articles, and their quality was assessed with the Quality Assessment of Diagnostic Accuracy Studies tool. Eight studies were suitable, including 521 preterm-born infants at varying risk of developing Cerebral Palsy (CP). Each study showed a significant correlation between vision and cognitive outcome. Predictive analysis including sensitivity and specificity was possible for three studies. Methodological quality was variable. Sensitivity ranged between 57 and 100% in the vision function assessments items, while specificity ranged from 59 to 100%. In conclusion, early vision showed strong correlation with cognition ≥ 12 months. While no single vision assessment was found to be superior, evaluation of specific functions, namely fixation and following, both at term age and between 3 and 6 months, demonstrated strong predictive validity.

Implementation of Listening Visits with Parents of Preterm Infants in an Italian Neonatal Intensive Care Unit.

PURPOSE: To assess the feasibility of implementing Listening Visits (LV) in an Italian neonatal intensive care unit (NICU). STUDY DESIGN AND METHODS: This feasibility implementation of LV included empathic listening and problem-solving sessions provided by a psychologist to 26 parents of hospitalized preterm newborns. Using the RE-AIM implementation framework, three facets of feasibility were assessed: reach, adoption, and implementation. RESULTS: It is feasible to integrate LV into the NICU: 76% of families were willing to try LV (reach). Listening Visits recipients reported high satisfaction. Twelve of the 16 families (75%) received six or more LV sessions (adoption), with mothers attending more sessions. Implementation fidelity, defined here as the percentage of LV recipients that received at least four sessions, was 94% among mothers and 30% among fathers. CLINICAL IMPLICATIONS: The LV intervention for parental support during the NICU stay is feasible and deemed helpful by parents. Parents were motivated to participate even though their levels of depression, stress, and anxiety were not high. In addition to the use of standardized screening questionnaires, parental requests and clinical team indications should be included in the decision-making for the provision of parental support services.