METHODOLOGY: This was a prospective cohort study. Shortly after birth, cranial ultrasonography was carried out via the anterior fontanelles of 70 normal control infants and 104 asphyxiated infants with a history of fetal distress and Apgar scores of less than 6 at 1 and 5 min of life, or requiring endotracheal intubation and manual intermittent positive pressure ventilation for at least 5 min after birth. Neurodevelopmental assessment was carried out on the survivors at 1 year of age.
RESULTS: Abnormal cranial ultrasound changes were detected in a significantly higher proportion (79.8%, or n = 83) of asphyxiated infants than controls (39.5%, or n = 30) (P < 0.0001). However, logistic regression analysis showed that only three factors were significantly associated with adverse outcome at 1 year of life among the asphyxiated infants. These were: (i) decreasing birthweight (for every additional gram of increase in birthweight, adjusted odds ratio (OR) = 0.999, 95% confidence interval (CI) 0.998, 1.000; P = 0.047); (ii) a history of receiving ventilatory support during the neonatal period (adjusted OR = 8.3; 95%CI 2.4, 28.9; P = 0.0009); and (iii) hypoxic-ischaemic encephalopathy stage 2 or 3 (adjusted OR = 5.8; 95%CI 1.8, 18.6; P = 0.003). None of the early cranial ultrasound changes was a significant predictor.
CONCLUSIONS: Early cranial ultrasound findings, although common in asphyxiated infants, were not significant predictors of adverse outcome during the first year of life in asphyxiated term infants.
METHODS: Cry signals from 2 different databases were utilized. First database contains 507 cry samples of normal (N), 340 cry samples of asphyxia (A), 879 cry samples of deaf (D), 350 cry samples of hungry (H) and 192 cry samples of pain (P). Second database contains 513 cry samples of jaundice (J), 531 samples of premature (Prem) and 45 samples of normal (N). Wavelet packet transform based energy and non-linear entropies (496 features), Linear Predictive Coding (LPC) based cepstral features (56 features), Mel-frequency Cepstral Coefficients (MFCCs) were extracted (16 features). The combined feature set consists of 568 features. To overcome the curse of dimensionality issue, improved binary dragonfly optimization algorithm (IBDFO) was proposed to select the most salient attributes or features. Finally, Extreme Learning Machine (ELM) kernel classifier was used to classify the different types of infant cry signals using all the features and highly informative features as well.
RESULTS: Several experiments of two-class and multi-class classification of cry signals were conducted. In binary or two-class experiments, maximum accuracy of 90.18% for H Vs P, 100% for A Vs N, 100% for D Vs N and 97.61% J Vs Prem was achieved using the features selected (only 204 features out of 568) by IBDFO. For the classification of multiple cry signals (multi-class problem), the selected features could differentiate between three classes (N, A & D) with the accuracy of 100% and seven classes with the accuracy of 97.62%.
CONCLUSION: The experimental results indicated that the proposed combination of feature extraction and selection method offers suitable classification accuracy and may be employed to detect the subtle changes in the cry signals.