An analysis of the kangaroo care intervention using neonatal EEG complexity: a preliminary study.

OBJECTIVE: Skin-to-skin contact (SSC) promotes physiological stability and interaction between parents and infants. Temporal analyses of predictability in EEG-sleep time series can elucidate functional brain maturation between SSC and non-SSC cohorts at similar post-menstrual ages (PMAs). METHODS: Sixteen EEG-sleep studies were performed on eight preterm infants who received 8 weeks of SSC, and compared with two non-SSC cohorts at term (N=126) that include a preterm group corrected to term age and a full term group. Two time series measures of predictability were used for comparisons. RESULTS: The SSC premature neonate group had increased complexity when compared to the non-SSC premature neonate group at the same PMA. Discriminant analysis shows that SSC neonates at 40 weeks PMA are closer to the full term neonate non-SSC group than to the premature non-SSC group at the same PMA; suggesting that the KC intervention accelerates neurophysiological maturation of premature neonates. CONCLUSIONS: Based on the hypothesis that EEG-derived complexity increases with neurophysiological maturation as supported by previously published research, SSC accelerates brain maturation in healthy preterm infants as quantified by time series measures of predictability when compared to a similar non-SSC group. SIGNIFICANCE: Times series methods that quantify predictability of EEG sleep in neonates can provide useful information about altered neural development after developmental care interventions such as SSC. Analyses of this type may be helpful in assessing other neuroprotection strategies.

A watermarking algorithm for polysomnography data.

A blind watermarking algorithm for polysomnography (PSG) data in European Data Format (EDF) has been developed for the identification and attribution of shared data. This is accomplished by hiding a unique identifier in the phase spectrum of each PSG epoch using an undisclosed key so that a third party cannot retrieve the watermark without knowledge of the key. A pattern discovery algorithm is developed to find the watermark pattern even though the data may have been altered. The method is evaluated using 25 PSG studies from the Sleep Heart Health Study database. The integrity of the signal data was determined using time series measures of both the original and watermarked signals, and by determining its effect on scoring sleep stages from the PSG data. The results of the analysis indicate that the proposed watermarking method for PSG data is an effective and efficient way to identify shared data without compromising its intended use.