Computerized cry analysis in infants affected by severe protein energy malnutrition

Video and audio processing in paediatrics: a review

OBJECTIVE

Video and sound acquisition and processing technologies have seen great improvements in recent decades, with many applications in the biomedical area. The aim of this paper is to review the overall state of the art of advances within these topics in paediatrics and to evaluate their potential application for monitoring in the neonatal intensive care unit (NICU).

APPROACH

For this purpose, more than 150 papers dealing with video and audio processing were reviewed. For both topics, clinical applications are described according to the considered cohorts-full-term newborns, infants and toddlers or preterm newborns. Then, processing methods are presented, in terms of data acquisition, feature extraction and characterization.

MAIN RESULTS

The paper first focuses on the exploitation of video recordings; these began to be automatically processed in the 2000s and we show that they have mainly been used to characterize infant motion. Other applications, including respiration and heart rate estimation and facial analysis, are also presented. Audio processing is then reviewed, with a focus on the analysis of crying. The first studies in this field focused on induced-pain cries and the newest ones deal with spontaneous cries; the analyses are mainly based on frequency features. Then, some papers dealing with non-cry signals are also discussed.

SIGNIFICANCE

Finally, we show that even if recent improvements in digital video and signal processing allow for increased automation of processing, the context of the NICU makes a fully automated analysis of long recordings problematic. A few proposals for overcoming some of the limitations are given.

The effect of positioning on infant cries: implications for sudden infant death syndrome

OBJECTIVE

A definitive cause for sudden infant death syndrome (SIDS) has not yet been identified, but some theories point to laryngeal or respiratory causes, in addition to theories of reduced arousal or reduced autonomic response. The occurrence of SIDS has dropped since the movement to place newborns to sleep in the supine position; however, some research has found a respiratory disadvantage for infants in this position. The current paper studied acoustic characteristics of infant pain cries to determine the potential differences related to prone versus supine positioning.

METHODS

Fifty-one newborn infant cries were recorded during and following a blood draw screening procedure, with infants placed either in the supine or prone position. All infants were healthy, full-term infants. Complete crying episodes were audio-recorded, and results were based on compositional analysis and long-time average spectrum analysis across each crying episode.

RESULTS

Spectral analysis revealed acoustic differences related to infant positioning, and acoustic analysis also revealed that there were no respiratory differences between supine-positioned and prone-positioned infants. Overall, the acoustic differences suggest decreased arousal and/or a decreased response to pain for healthy infants recorded in the prone position.

CONCLUSIONS

As decreased arousal and prone positioning have been seen as possible causative factors for SIDS, the current results are seen as a successful step in evaluating the possibility of using acoustic analysis of infant cries as a means of evaluating SIDS risk for healthy infants.

Ultrasonic call characteristics of rat pups are altered following prenatal malnutrition

The male offspring of rats provided with a protein deficient diet (6% casein) for 5 weeks prior to mating and throughout pregnancy were subjected to a brief period of isolation and cooling at postnatal Days (P)7, 9, and 11, and their ultrasonic vocalizations were compared with those of well-nourished pups. Calls were categorized into 12 different types based upon their sonographic patterns. Although call rates were equal, the call characteristics of the prenatally malnourished pups differed significantly from those of well-nourished controls. At P7, their mean peak sound frequency (irrespective of call type) was significantly higher, and constant frequency calls were of both higher frequency and longer duration. Over the age range studied, prenatally malnourished pups emitted a smaller variety of calls, with significantly fewer ascending frequency vocalizations while producing either significantly fewer (P9) or greater (P11) descending frequency calls. Altered crying patterns have been related to brain damage in human babies, with more abnormal cries being associated with more severe neurological impairment. Therefore, the present results most likely reflect altered central nervous system development and function. Ultrasonic vocalization characteristics in rat pups may provide a useful early marker of the severity of disturbance to the development of the central nervous system following an insult, and offer the potential for predicting the degree of functional and behavioral deficits later in life.

Brainstem auditory and visual evoked potentials in children with protein-energy malnutrition

BACKGROUND

To determine the effects of malnutrition on the developing brain with brainstem auditory evoked potentials (BAEP) and flash visual evoked potentials (fVEP).

METHODS

The BAEP and fVEP of 11 kwashiorkor (8 +/- 1.56 months) and 10 marasmus (7.9 +/- 1.27 months) patients and 10 healthy control subjects (7.65 +/- 0.82 months) were recorded and the measurements were compared with each other in relation with plasma total protein and albumin levels.

RESULTS

There were no differences between the mean latencies of the waves I, II, III and IV and mean interpeak latencies (IPL) of the waves I-III of the BAEP and the wave IV (N2) of the fVEP between the three groups. Mean latency of the wave V and mean IPL of the waves I-V and the waves III-V were significantly different between the three groups. The kwashiorkor group had significantly longer mean latency of the wave V than the marasmus group on the right ear and the control group on the both of the ears. The kwashiorkor group had significantly longer mean IPL of the waves I-V than the marasmus group on the right ear and than the control group on the left ear. The kwashiorkor group had also significantly longer mean IPL of the waves III-V than the control group on the left.

CONCLUSIONS

The BAEP and fVEP are non-invasive electrophysiologic methods reflecting the integrity or disruption of the central neurologic pathways. The present results confirm the disruption of the central nervous system with the BAEP in children with protein-energy malnutrition, especially in kwashiorkor patients.

Acoustic examination of preterm and full-term infant cries: the long-time average spectrum

The acoustic characteristics of crying behavior displayed in 2 groups of newborn infants are reported. The crying episodes of 10 full-term and 10 preterm infants were audio recorded and analyzed with regard to the long-time average spectrum (LTAS) characteristics. An LTAS display was created for each infant's non-partitioned crying episode, as well as for 3 equidurational partitions of the crying episode. Measures of first spectral peak, mean spectral energy, and spectral tilt were revealing of differences between full-term and preterm infants' non-partitioned crying episodes. In addition, the full-term infants demonstrated significant changes in their crying behavior across partitions, whereas the preterm infants changed little across the crying episode. Discussion focuses on possible differences between full-term and preterm infants in their neurophysiological maturity, and the subsequent impact on their speech development. The importance of examining entire crying episodes when evaluating the crying behavior of infants is also discussed.

Acoustic analysis of newborn infant cry signals

This paper aims at estimating the fundamental frequency (pitch) and the vocal tract resonant frequencies (formants) from newborn infant cry signals. Such parameters are of interest in exploring brain function at early stages of child development, for the timely diagnosis of neonatal disease and malformation. The paper compares a spectral parametric technique and the cepstrum approach, extending previous results. The parametric technique is based on autoregressive models whose order is adaptively estimated on subsequent signal frames by means of a new method. This allows the correct tracking of pitch and formant variations with time. The traditional cepstrum approach is modified in order to follow signal variability. In particular, the cepstrum spectral resolution is improved by applying the chirp Z-transform (CZT) and by adaptively varying the 'lifter' length. The two methods are tested on simulated data, as far as robustness to noise and spectral resolution are concerned, and are then applied to real baby cry data.

Parametric and non-parametric estimation of speech formants: application to infant cry

The present paper addresses the issue of correctly estimating the peaks in the speech envelope (formants) occurring in newborn infant cry. Clinical studies have shown that the analysis of such spectral characteristics is a helpful noninvasive diagnostic tool. In fact it can be applied to explore brain function at very early stage of child development, for a timely diagnosis of neonatal disease and malformation. The paper focuses on the performance comparison between some classical parametric and non-parametric estimation techniques particularly well suited for the present application, specifically the LP, ARX and cepstrum approaches. It is shown that, if the model order is correctly chosen, parametric methods are in general more reliable and robust against noise, but exhibit a less uniform behaviour than cepstrum. The methods are compared also in terms of tracking capability, since the signals under study are nonstationary. Both simulated and real signals are used in order to outline the relevant features of the proposed approaches.