Linda E. Kapuniai

Change detection model for serially correlated multivariate data.

SUMMARY A method is presented for detecting change in biological multivariate stationary processes of a single subject after stimulation. A finite multivariate autoregression is fitted to pre-stimulus data using a step-wise procedure with tests of significance. The fit of the model is also checked by comparing the estimated spectra, phase, and coherence with fitted curves. The statistic which tests for change at a given time point is a quadratic form involving the one-step prediction error vector ancd the inverse of the one-step prediction error covariance matrix. Under the hypothesis of no change, and for a Gaussian process, these statistics have independent chi-square distributions. The technilque has been applied to the detection of change in the brain waves of two human newborn infants following stimulation.

Infant polysomnography: reliability and validity of infant arousal assessment.

Summary Infant arousal scoring based on the Atlas Task Force definition of transient EEG arousal was evaluated to determine (1) whether transient arousals can be identified and assessed reliably in infants and (2) whether arousal and no-arousal epochs scored previously by trained raters can be validated reliably by independent sleep experts. Phase I for inter- and intrarater reliability scoring was based on two datasets of sleep epochs selected randomly from nocturnal polysomnograms of healthy full-term, preterm, idiopathic apparent life-threatening event cases, and siblings of Sudden Infant Death Syndrome infants of 35 to 64 weeks postconceptional age. After training, test set 1 reliability was assessed and discrepancies identified. After retraining, test set 2 was scored by the same raters to determine interrater reliability. Later, three raters from the trained group rescored test set 2 to assess inter- and intrarater reliabilities. Interrater and intrarater reliability &kgr;’s, with 95% confidence intervals, ranged from substantial to almost perfect levels of agreement. Interrater reliabilities for spontaneous arousals were initially moderate and then substantial. During the validation phase, 315 previously scored epochs were presented to four sleep experts to rate as containing arousal or no-arousal events. Interrater expert agreements were diverse and considered as noninterpretable. Concordance in sleep experts’ agreements, based on identification of the previously sampled arousal and no-arousal epochs, was used as a secondary evaluative technique. Results showed agreement by two or more experts on 86% of the Collaborative Home Infant Monitoring Evaluation Study arousal scored events. Conversely, only 1% of the Collaborative Home Infant Monitoring Evaluation Study-scored no-arousal epochs were rated as an arousal. In summary, this study presents an empirically tested model with procedures and criteria for attaining improved reliability in transient EEG arousal assessments in infants using the modified Atlas Task Force standards. With training based on specific criteria, substantial inter- and intrarater agreement in identifying infant arousals was demonstrated. Corroborative validation results were too disparate for meaningful interpretation. Alternate evaluation based on concordance agreements supports reliance on infant EEG criteria for assessment. Results mandate additional confirmatory validation studies with specific training on infant EEG arousal assessment criteria.

Ontogeny of arousal.

Ontogeny of arousal data constitute a vital supplement to the sparse literature on spontaneous neuronal activity. These data demonstrate that measurable infant spontaneous arousals (SAs) with an inherent oscillatory entrainment occur six times more in active sleep than in quiet sleep of the same duration and are identifiable as a human neurobiologic function. These SAs are not significantly associated with race or ethnicity, gender, total hours spent sleeping, percent time spent in active or quiet sleep, preterm status, history of a life-threatening event, having had a sibling who died of sudden infant death syndrome (SIDS), or having had a mother who smoked during this pregnancy. As measurable neurophysiologic events, SAs establish parameters for research at molecular and molar levels focusing on several critical areas: (1) the neuronal control of SA related to neurotransmitters, (2) as a significant antecedent factor in clinical cardiorespiratory events occurring in infants at high epidemiologic risk for SIDS; (3) as a regulatory biologic factor underlying temperament and executive cognitive functioning, and (4) morbidity and mortality effects possibly related to therapeutic interventions that alter SA levels.

Insomnia. Rational diagnosis and treatment.

PreviewBecause insomnia is “a symptom and not a unitary diagnostic entity,” its treatment should vary with its cause. Yet, overmedication with hypnotics appears to be the rule rather than the exception. According to the authors of this article, effective therapy lies between behavioral and limited pharmacologic measures, with choice of appropriate medication based on a clear knowledge of what it will and will not do.