TwinEDA: a sustainable deep-learning approach for limb-position estimation in preterm infants' depth images.

Early diagnosis of neurodevelopmental impairments in preterm infants is currently based on the visual analysis of newborns' motion patterns by trained operators. To help automatize this time-consuming and qualitative procedure, we propose a sustainable deep-learning algorithm for accurate limb-pose estimation from depth images. The algorithm consists of a convolutional neural network (TwinEDA) relying on architectural blocks that require limited computation while ensuring high performance in prediction. To ascertain its low computational costs and assess its application in on-the-edge computing, TwinEDA was additionally deployed on a cost-effective single-board computer. The network was validated on a dataset of 27,000 depth video frames collected during the actual clinical practice from 27 preterm infants. When compared to the main state-of-the-art competitor, TwinEDA is twice as fast to predict a single depth frame and four times as light in terms of memory, while performing similarly in terms of Dice similarity coefficient (0.88). This result suggests that the pursuit of efficiency does not imply the detriment of performance. This work is among the first to propose an automatic and sustainable limb-position estimation approach for preterm infants. This represents a significant step towards the development of broadly accessible clinical monitoring applications.

Contactless Monitoring of Breathing Pattern and Thoracoabdominal Asynchronies in Preterm Infants Using Depth Cameras: A Feasibility Study.

OBJECTIVE: Monitoring infants' breathing activity is crucial in research and clinical applications but remains a challenge. This study aims to develop a contactless method to monitor breathing patterns and thoracoabdominal asynchronies in infants inside the incubator, using depth cameras. METHODS: We proposed an algorithm to extract the 3D displacements of the ribcage and abdomen from the analysis of depth images. We evaluated the accuracy of the system in-vitro vs. a reference motion capture analyzer. We also conducted a feasibility study on 12 patients receiving non-invasive respiratory support to estimate the mean and the variability of the chest wall displacements in preterm infants and evaluate the suitability of the proposed system in the clinical setting. RESULTS: In-vitro, the mean (95% CI) error in the measurement of amplitude, frequency and phase shift between compartmental displacements was -0.14 (-0.57, 0.28) mm, 0.02 (-0.99, 1.03) bpm, and -0.40 (-1.76, 0.95)°, respectively. In-vivo, the mean (95% CI) amplitude of the ribcage and abdomen displacements were 0.99 (0.34, 2.67) mm and 1.20 (0.40, 2.15) mm, respectively. CONCLUSIONS: The developed system proved accurate in-vitro and was suitable for the clinical environment. CLINICAL IMPACT: The proposed method has value for evaluating infants' breathing patterns in research applications and, after further development, may represent a simple monitoring tool for infants' respiratory activity inside the incubator.

Simultaneous monitoring of vocal doses and breathing patterns in professional singers.

BACKGROUND: When learning and improving singing, the quantitative characterization of artists' performances based only on vocal parameters does not provide enough information to identify strategies for improvement. Simultaneous monitoring of sound production and breathing patterns in professional singers can allow the exploration of the mechanisms that promote effective singing modalities through association with respiratory efforts. METHODS: We developed and tested a novel portable device that simultaneously monitors vocal activity and breathing patterns without interfering with natural singing. The device utilizes a combination of a miniature accelerometer to measure vocal doses from skin vibrations on the neck and two respiratory inductive plethysmography (RIP) bands to estimate the breathing pattern by measuring changes in the thoracoabdominal cross-sectional area. RESULTS: The device was tested on 13 professional baroque-style singers and 14 untrained individuals during the execution of singing tasks. It was accurate compared with reference systems (R2 = 0.99 and R2 = 0.91 for the accelerometer and RIP, respectively) and showed differences between the two groups (p ≤ 0.001). CONCLUSIONS: By allowing a quantitative characterization of the effectiveness of singing technique, this novel device can help provide a better understanding of the physiology of singing and may be a tool for support learning and training optimization for professional singers.

Oscillatory mechanics at birth for identifying infants requiring surfactant: a prospective, observational trial.

BACKGROUND: Current criteria for surfactant administration assume that hypoxia is a direct marker of lung-volume de-recruitment. We first introduced an early, non-invasive assessment of lung mechanics by the Forced Oscillation Technique (FOT) and evaluated its role in predicting the need for surfactant therapy. OBJECTIVES: To evaluate whether lung reactance (Xrs) assessment by FOT within 2 h of birth identifies infants who would need surfactant within 24 h; to eventually determine Xrs performance and a cut-off value for early detection of infants requiring surfactant. METHODS: We conducted a prospective, observational, non-randomized study in our tertiary NICU in Milan. Eligible infants were born between 27+0 and 34+6 weeks' gestation, presenting respiratory distress after birth. EXCLUSION CRITERIA: endotracheal intubation at birth, major malformations participation in other interventional trials, parental consent denied. We assessed Xrs during nasal CPAP at 5 cmH2O at 10 Hz within 2 h of life, recording flow and pressure tracing through a Fabian Ventilator for off-line analysis. Clinicians were blinded to FOT results. RESULTS: We enrolled 61 infants, with a median [IQR] gestational age of 31.9 [30.3; 32.9] weeks and birth weight 1490 [1230; 1816] g; 2 infants were excluded from the analysis for set-up malfunctioning. 14/59 infants received surfactant within 24 h. Xrs predicted surfactant need with a cut-off - 33.4 cmH2O*s/L and AUC-ROC = 0.86 (0.76-0.96), with sensitivity 0.85 and specificity 0.83. An Xrs cut-off value of - 23.3 cmH2O*s/L identified infants needing surfactant or respiratory support > 28 days with AUC-ROC = 0.89 (0.81-0.97), sensitivity 0.86 and specificity 0.77. Interestingly, 12 infants with Xrs < - 23.3 cmH2O*s/L (i.e. de-recruited lungs) did not receive surfactant and subsequently required prolonged respiratory support. CONCLUSION: Xrs assessed within 2 h of life predicts surfactant need and respiratory support duration in preterm infants. The possible role of Xrs in improving the individualization of respiratory management in preterm infants deserves further investigation.

An Automated Approach for General Movement Assessment: A Pilot Study.

Objective: The objective of the study was to develop an automatic quantitative approach to identify infants with abnormal movements of the limbs at term equivalent age (TEA) compared with general movement assessment (GMA). Methods: GMA was performed at TEA by a trained operator in neonates with neurological risk. GMs were classified as normal (N) or abnormal (Ab), which included poor repertoire and cramped synchronized movements. The signals from four micro-accelerometers placed on all limbs were recorded for 10 min simultaneously. A global index (KC_index), quantifying the characteristics of individual limb movements and the coordination among the limbs, was obtained by adding normalized kurtosis of the distribution of the first principal component of the acceleration signals to the cross-correlation of the jerk for the upper and lower limbs. Results: Sixty-eight infants were studied. A KC_index cut-off of 201.5 (95% CI: 199.9-205.0) provided specificity = 0.86 and sensitivity = 0.88 in identifying infants with Ab movements. Conclusions: KC_index provides an automatic and quantitative measure that may allow the identification of infants who require further neurological evaluation.

Accuracy of volume and pressure delivery by mechanical ventilators in use in neonatal intensive care units: A quality control study.

OBJECTIVE: Despite technical specifications of neonatal mechanical ventilators (MVs) guarantee clinically irrelevant discrepancies between the set and the delivered values of ventilation parameters, previous studies reported large deviations. Most studies characterized performances of a given model/brand by studying a single device, disregarding possible intramodel differences, and leaving the accuracy of the ventilation parameters effectively delivered in clinical settings unknown. The aim of this study was to evaluate the real-life accuracy of pressure and volume parameters delivered by neonatal ventilators ready to be used on patients in neonatal intensive care units (NICUs). STUDY DESIGN: In vitro study. SUBJECTS SELECTION: Neonatal ventilators (n = 33 of 8 different models) available in four European NICUs. METHODOLOGY: The MVs were connected to a test lung (resistance = 50 cmH2 O*s/L, compliance = 0.35 mL/cmH2 O) provided with pressure and flow sensors. MVs were tested over two different ventilation modes randomly: (a) pressure controlled (PC) with a peak inspiratory pressure (PIP) of 22 cmH2 O, and (b) PC with volume targeted ventilation (VTV) with a tidal volume (VT ) of 6 mL. In all tests, positive end-expiratory pressure (PEEP) was set to 6 cmH2 O, respiratory rate to 45 breaths/min, inspiratory time to 0.33 seconds, and oxygen fraction to 0.3. RESULTS: During PC the median (min-max) values delivered were: PEEP = 5.84(4.95-6.48) cmH2 O, PIP = 21.63(20.04-22.62) cmH2 O. During VTV, VT was 5.94(4.63-8.01) mL. VT was considerably variable, ranging from -22% to +33% of the set and displayed values. Differences in accuracy among devices of the same model were comparable to those found among different models. CONCLUSIONS: Our findings suggest that loss of accuracy in ventilation variables is likely related to daily use of the devices rather than weakness in the design or manufacturing process, urging the improvement of maintenance and quality control procedures to preserve the performances of neonatal MVs during their entire lifespan.