Infant Cry Research Articles

Frustration responses of single and partnered mothers to prolonged infant crying.

Prolonged infant crying can be a trigger for maternal frustration and can even predict intrusive infant-related thoughts of harm. In this study, we compared frustration responses to prolonged infant crying between single and partnered mothers and attempted to identify variables that mediated any difference between the two groups. We also identified acoustic characteristics of infant cries that were related to higher levels of reported maternal frustration. Twenty-five single and 25 partnered mothers with infants under the age of 6 months completed several mental health questionnaires, and then rated their frustration level after listening to each of 50 consecutive 15s infant cry videos from 50 different infants. As expected, greater maternal perceived stress was associated with higher frustration ratings in response to infant crying, and this was mediated by increased maternal negative affect. Also as expected, both financial strain and low social support were associated with greater perceived stress. However, our sample of single mothers did not experience more stress than our sample of partnered mothers. Nor did they find infant crying to be more frustrating, perhaps due to a recruitment bias toward higher functioning single mothers. Finally, several cry acoustic characteristics were associated with increased maternal frustration, including higher fundamental frequency, air energy, shimmer and longer duration of expiratory phonations, as well as a longer cumulative duration of crying. Our results suggest that maternal frustration in response to infant crying may be decreased by lowering maternal stress levels, and this may be achieved by increasing social support and decreasing financial strain. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Premature Infant Cry Classification via Deep Convolutional Recurrent Neural Network Based on Multi-class Features

Premature Infant Cry Classification via Deep Convolutional Recurrent Neural Network Based on Multi-class Features

Maternal perceived stress and infant behavior during the COVID-19 pandemic

BackgroundMaternal stress has negative consequences on infant behavioral development, and COVID-19 presented uniquely stressful situations to mothers of infants born during the pandemic. We hypothesized that mothers with higher levels of perceived stress during the pandemic would report higher levels of infant regulatory problems including crying and interrupted sleep patterns.MethodsAs part 6 sites of a longitudinal study, mothers of infants born during the pandemic completed the Perceived Stress Scale, the Brief Infant Sleep Questionnaire, and an Infant Crying survey at 6 (n = 433) and 12 (n = 344) months of infant age.ResultsMaternal perceived stress, which remained consistent at 6 and 12 months of infant age, was significantly positively correlated with time taken to settle infants. Although maternal perceived stress was not correlated with uninterrupted sleep length, time taken to put the infant to sleep was correlated. Perceived stress was also correlated with the amount of infant crying and fussiness reported at 6 months.ConclusionsMothers who reported higher levels of perceived stress during the pandemic reported higher levels of regulatory problems, specifically at 6 months. Examining how varying levels of maternal stress and infant behaviors relate to overall infant developmental status over time is an important next step.ImpactWomen giving birth during the COVID-19 pandemic who reported higher levels of stress on the Perceived Stress Scale also reported higher levels of infant fussiness and crying at 6 months old, and more disruptive sleep patterns in their infants at 6 months and 12 months old. Sleeping problems and excessive crying in infancy are two regulatory problems that are known risk factors for emotional and behavioral issues in later childhood. This paper is one of the first studies highlighting the associations between maternal stress and infant behaviors during the COVID-19 pandemic.

Sistem Kendali Ayunan Bayi Berbasis Mikrokontroller dan Android

This study aims to design a prototype of a microcontroller and arduino-based baby swing control system that will be very useful for housewives who have babies. This system will automatically work to move the swing when the sound of a baby and water is detected. This system is designed using a sound sensor that functions to detect the baby's voice, a water sensor to detect water when a baby is urinating, a DC motor that functions to move the swing automatically and a mobile application that functions to monitor swings. The test results show that the tool functions as desired, where when the tool detects the sound of a baby's cry, the tool automatically moves the swing, when the tool detects the presence of water, the swing will send a notification to the smartphone, this tool also successfully displays information to the smartphone in the form of a state baby in the form of video and sensor conditions on the swing. The communication distance between the smartphone and the device is 20 meters.

Infant cry classification by MFCC feature extraction with MLP and CNN structures

In this study, Dunstan’s infant cry data set is pre-processed with the feature vector approach, including MFCC (19 features) and energy (one feature). By using extracted features and Support Vector Machine (SVM), MultilayerPerceptron (MLP), and Convolutional Neural Network (CNN) classifiers, five classes of infant cry (“Neh” = hungry; “Eh” = need to burp; “Owh” = tired; “Eairh” = stomach cramp; “Heh” = physical discomfort) are distinguished. The proposed MLP and CNN structures are analyzed according to the loss and the accuracy based on the epoch; moreover, to evaluate the performance of classifiers AUC-ROC, Confusion matrix, accuracy, f1_score, recall, and precision have been used. All three classifiers are analyzed, and their results show that the CNN-designed model has the best performance. Results show that the performance will improve by increasing the complexity of the model. With this approach, classifiers are run 10 times, and the average accuracy for SVM for SMOTE and non-SMOTE data are obtained with tolerance 0.823 ± 0.02, 0.861 ± 0.02, respectively. These accuracies for MLP are 0.876 ± 0.01, 0.892 ± 0.01, and finally, for CNN, are 0.921 ± 0.005, 0.911 ± 0.005. At the best condition, an accuracy of 92.1 % is obtained for five classes of infant cries by the proposed CNN structure.

A Machine Learning Approach to Classify Biomedical Acoustic Features for Baby Cries

Communication is imperative for living beings for exchanging information. But for newborns, the only way of communicating with the world is through crying, and it is the only medium through which caregivers can know about the needs of their children. Timely addressing baby cries is very important so that the child is relieved at the earliest. It has been a challenge, especially for new parents. The literature says newborn babies use The Dustan Baby Language to communicate. According to this language, there are five words to understand a baby’s needs, which are “Neh” (hungry), “Eh” (burp is needed), “Owh/Oah” (fatigue), “Eair/Eargghh” (cramps), “Heh” (feel hot or wet, physical discomfort). This research aims to develop a model for recognizing baby cries and distinguishing between different kinds of baby cries. Here we more broadly focus on whether the infant is in pain due to hunger or discomfort. The study proposes a comparative approach using four classification models: random forest, support vector machine, logistic regression, and decision tree. These algorithms learn from the spectral features: chroma_stft, spectral_centroid, bandwidth, spectral_rolloff, mel-frequency cepstral coefficients, linear predictive coding, res, zero_crossing_rate extracted from the infant cry. The support vector machine model outperforms other classifiers for correctly classifying infant cries.

Infant Cry Signal Diagnostic System Using Deep Learning and Fused Features.

Early diagnosis of medical conditions in infants is crucial for ensuring timely and effective treatment. However, infants are unable to verbalize their symptoms, making it difficult for healthcare professionals to accurately diagnose their conditions. Crying is often the only way for infants to communicate their needs and discomfort. In this paper, we propose a medical diagnostic system for interpreting infants' cry audio signals (CAS) using a combination of different audio domain features and deep learning (DL) algorithms. The proposed system utilizes a dataset of labeled audio signals from infants with specific pathologies. The dataset includes two infant pathologies with high mortality rates, neonatal respiratory distress syndrome (RDS), sepsis, and crying. The system employed the harmonic ratio (HR) as a prosodic feature, the Gammatone frequency cepstral coefficients (GFCCs) as a cepstral feature, and image-based features through the spectrogram which are extracted using a convolution neural network (CNN) pretrained model and fused with the other features to benefit multiple domains in improving the classification rate and the accuracy of the model. The different combination of the fused features is then fed into multiple machine learning algorithms including random forest (RF), support vector machine (SVM), and deep neural network (DNN) models. The evaluation of the system using the accuracy, precision, recall, F1-score, confusion matrix, and receiver operating characteristic (ROC) curve, showed promising results for the early diagnosis of medical conditions in infants based on the crying signals only, where the system achieved the highest accuracy of 97.50% using the combination of the spectrogram, HR, and GFCC through the deep learning process. The finding demonstrated the importance of fusing different audio features, especially the spectrogram, through the learning process rather than a simple concatenation and the use of deep learning algorithms in extracting sparsely represented features that can be used later on in the classification problem, which improves the separation between different infants' pathologies. The results outperformed the published benchmark paper by improving the classification problem to be multiclassification (RDS, sepsis, and healthy), investigating a new type of feature, which is the spectrogram, using a new feature fusion technique, which is fusion, through the learning process using the deep learning model.

A Scientometric Review of Infant Cry and Caregiver Responsiveness: Literature Trends and Research Gaps over 60 Years of Developmental Study.

Infant cry is an adaptive signal of distress that elicits timely and mostly appropriate caring behaviors. Caregivers are typically able to decode the meaning of the cry and respond appropriately, but maladaptive caregiver responses are common and, in the worst cases, can lead to harmful events. To tackle the importance of studying cry patterns and caregivers' responses, this review aims to identify key documents and thematic trends in the literature as well as existing research gaps. To do so, we conducted a scientometric review of 723 documents downloaded from Scopus and performed a document co-citation analysis. The most impactful publication was authored by Barr in 1990, which describes typical developmental patterns of infant cry. Six major research thematic clusters emerged from the analysis of the literature. Clusters were renamed "Neonatal Pain Analyzer" (average year of publication = 2002), "Abusive Head Trauma" (average year of publication = 2007), "Oxytocin" (average year of publication = 2009), "Antecedents of Maternal Sensitivity" (average year of publication = 2010), "Neurobiology of Parental Responses" (average year of publication = 2011), and "Hormonal Changes & Cry Responsiveness" (average year of publication = 2016). Research clusters are discussed on the basis of a qualitative inspection of the manuscripts. Current trends in research focus on the neurobiology of caregiver responses and the identification of factors promoting maternal sensitivity. Recent studies have also developed evidence-based strategies for calming crying babies and preventing caregivers' maladaptive responses. From the clusters, two topics conspicuously call for future research: fathers' responsiveness to infant cry and the impact of caregiver relationship quality on cry responsiveness.

Intelligent Baby Monitoring System

Abstract: IOT leads to a wide use a mobile phones with with broadband internet. One important concept of the same is the use of mobile phones by working parents to watch the activities of baby while babysitting. This paper presents the design of intelligent baby monitoring which supports such video monitoring. This system automatically on detection of baby cry sound. It switch on buzzer and sends a notification if baby any continues till specific time. That means system cannot handle baby and require personal attention and also if the bed in the cradle is damp.

Infant Posture Analyzer and Wet Diaper System

Abstract: Monitoring the well-being of infants is of utmost importance for ensuring their safety and providing timely care. This paper presents a novel approach to enhance traditional baby monitors by integrating intelligent classification of infant cries and real-time monitoring of sleep positions. The proposed smart baby monitor employs advanced signal processing techniques, machine learning algorithms, and computer vision methods to analyse audio and visual cues, providing valuable insights into a baby's needs and sleep patterns. This research paper presents the development of a Smart Baby Monitoring System equipped with one machine learning module and a wet diaper detection model. The system aims to enhance infant safety and improve caregiving practices. The first module employs YOLO and a Convolutional Neural Network (CNN) for position detection, identifying whether the baby is on their belly, back, or side. The position detection model achieves an average accuracy of 90%, with precision and recall rates 85%, ensuring reliable identification of the baby's needs Second module is a wet diaper detection module is incorporated to detect if the baby's diaper needs changing. The integration of these machine learning modules in the Smart Baby Monitoring System offers realtime monitoring, timely alerts, and improved infant care.

Sleep SAAF Responsive Parenting Intervention for Black Mothers Impacts Response to Infant Crying: A Randomized Clinical Trial

ObjectiveMany parents use food to soothe their infant, regardless of infant hunger, which can increase risk for rapid weight gain. Interventions promoting alternative soothing strategies may help parents respond more appropriately to crying. This secondary analysis aimed to examine effects of the Sleep SAAF (Strong African American Families) responsive parenting (RP) intervention on maternal responses to infant crying and to explore moderating effects of infant negativity. MethodsPrimiparous Black mothers (n = 212) were randomized to an RP or safety control intervention, delivered during home visits at 3 and 8 weeks postpartum. Parents were empowered to first use non-food soothing strategies (eg, white noise, swaddling) when responding to crying. Mothers completed the Babies Need Soothing questionnaire at 8 and 16 weeks, and the Infant Behavior Questionnaire at 16 weeks. Data were analyzed using linear or logistic regression. ResultsRP mothers were significantly more likely than controls to use shushing/white noise to soothe their infant at 8 (OR = 4.9, 95% CI: 2.2–10.6) and 16 weeks (OR = 4.8, 95% CI: 2.2–10.5), to go for a walk in stroller/ride in car at 8 weeks (OR = 2.3, 95% CI: 1.2–4.6), and to swing/rock/bounce their infant at 16 weeks (OR = 5.5, 95% CI: 1.2–25.7). RP mothers also reported significantly more frequent use of deep breathing, exercising, and bathing/showering than controls when frustrated with crying. Infant negativity moderated some intervention effects such that the RP intervention was more effective at increasing use of some soothing practices among mothers with less negative infants. ConclusionsAn RP intervention positively impacted first-time Black mothers’ responses to infant crying.

Pathways between Risk/Protective Factors and Maternal Postnatal Depressive Symptoms: The ELFE Cohort.

The risk factors for postnatal depressive symptoms (PNDS) are numerous, but little is known about the protective factors or the interactions between different exposures. The present study explored the pathways between maternal, infant and parenthood vulnerabilities or risk/protective factors and PNDS at 2 months postpartum (PP) in a large sample of women from the general population. We used data from the French ELFE cohort, a nationally representative cohort of children followed-up from birth. The available information about vulnerabilities or risk/protective factors for PNDS was collected during the maternity ward stay (mother or medical records) and at 2 months PP (mother by phone). PNDS were evaluated with the Edinburgh Postnatal Depression Scale (EPDS) at 2 months. A measurement model was built based on the psychosocial model for PNDS of Milgrom and colleagues using exploratory factor analysis. The Structural Equation Model was used to investigate the pathways between vulnerability, risk/protective factors and PNDS at 2 months PP. In the study sample (n = 11,583), a lack of a partner's perceived antenatal emotional support, consultation with a mental health specialist before pregnancy, family financial difficulties, prenatal psychological distress and a difficult pregnancy experience were directly associated with the severity of maternal PNDS at 2 months PP, as well as lack of perceived postnatal support. Family financial difficulties and consultation with a mental health specialist before pregnancy were also indirectly associated with the intensity of PNDS through a lack of perceived antenatal emotional support, a difficult pregnancy experience, prenatal psychological distress and a lack of perceived postnatal support. Regarding infant and parenthood characteristics, infant self-regulation difficulties, maternal difficulty in understanding infant crying and infant hospitalisation were directly associated with PNDS severity at 2 months PP, while maternal difficulty in understanding an infant's cries was also indirectly associated with infant self-regulation difficulties. Perinatal professional support should begin antenatally and target the couple's prenatal functioning, with particular attention to women presenting a history of psychiatric disorders, especially those of low socioeconomic status. After delivery, addressing infant and parenthood characteristics is also recommended.

Ecological Predictors of Maternal Nurturance to Distress and Beliefs About Infant Crying: Examining the Roles of Household and Neighborhood Resources

SYNOPSIS Objective. Maternal nurturance to infant distress is associated with positive child developmental outcomes including greater attachment security, emotion regulation skills, and social and behavioral competencies. However, factors at multiple levels of parents’ environments may impede parents’ ability to respond sensitively to their infants’ distress. This study examined whether household-level burden (low maternal education, financial need) and neighborhood-level resources (distribution of educational, health/environmental, and socio/economic resources in a community) are associated with parents’ beliefs about infant crying and observed maternal nurturance to infant distress. Design. Ninety-nine mothers (M age = 28.99 years, SD = 5.41) of 6- to 12- month-old infants (42.4% female) completed questionnaires assessing household-level burden as well as the Infant Crying Questionnaire (ICQ) during a home visit. Mother-infant interactions were also filmed to assess maternal sensitivity to infant distress. Results. In separate models, higher household burden and reduced neighborhood resources were associated with increased maladaptive beliefs about infant crying and reduced nurturance to distress. When considered together, household burden was uniquely associated with maternal nurturance. Conclusions. Implications for intervention include considering efforts at both the household and neighborhood levels to address multi-systemic disparities that families experience in efforts to promote greater maternal nurturance.

Infant crying predicts real-time fluctuations in maternal mental health in ecologically valid home settings.

Exposure to infant crying is a well-established predictor of mothers' mental health. However, this association may reflect many potential mechanisms. Capturing dynamic fluctuations in mothers' states simultaneously with caregiving experiences is necessary to identify the real-time processes influencing mental health. In this study, we leveraged ecological momentary assessments (EMAs) and infant-worn audio recorders to capture variability in mothers' mental health symptoms and their exposure to infant crying over one week in a racially and socio-economically diverse urban North-American sample (N = 53). We use multilevel modeling to characterize within- and between-person effects of crying on maternal negative affect and symptoms of depression and anxiety. Within participants, when infants cried more than average in the 10 min, 1 hr, and 8 hr prior to an EMA report, mothers' negative affect subsequently increased, controlling for mean levels of infant crying. In contrast to findings from laboratory studies, in everyday settings crying exposure did not immediately increase feelings of depression. Only when crying was above average for 8 hr prior to EMA did mothers report increases in subsequent depression symptoms, suggesting that the effects of crying on maternal mental health take hours to unfold in ecologically valid home settings. Between participants, mothers of infants who cried more on average did not report higher negative affect or symptoms of depression or anxiety. Overall, our findings reveal that crying exposure dynamically influences maternal negative affect and depression but not anxiety in ecologically valid real-world settings. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Adverse Childhood Experiences: Associations with a Blunted Cortisol Stress Response During Pregnancy.

Adverse Childhood Experiences (ACEs) are associated with a range of negative physical and mental health outcomes, yet there is limited research focused on the effect of ACEs on stress responses during pregnancy. Expectant mothers experience an increase in cortisol levels as pregnancy progresses, with this increase having important implications for fetal and early infant development. Little is known about the impact of ACEs on maternal cortisol levels. This study explored the relationship between maternal ACEs and cortisol response among expectant mothers nearing or in the third trimester of pregnancy. 39 expectant mothers were exposed to a Baby Cry Protocol via an infant simulator, with salivary cortisol collected at five points in time (N = 181). Stepwise, multilevel model creation resulted in a random intercept and random slope model with an interaction term for total number of ACEs and week of pregnancy. The repeated measures data showed that cortisol levels decreased across collection times, from arrival at the lab, through the Baby Cry Protocol, to recovery. Predictive margins for the interaction term showed that while exposure to a greater number of ACEs was associated with higher cortisol levels early in the third trimester, the expected increase in cortisol late in pregnancy was blunted for expectant mothers who were exposed to a greater number of ACEs. These findings findings suggest the importance of ACEs screening and intervention efforts as part of prenatal care.

The Lice, and: Scythe, and: Double Occupancy, and: Devotional

The Lice, and: Scythe, and: Double Occupancy, and: Devotional Marlo Starr (bio) The Lice The first event is known to have been an expulsion, and the last is hoped to be a reconciliation and return. —Marilynne Robinson, Housekeeping Our story begins with an expulsion. Mom swivelsto count and recount the milling brood. Us,sticky and slick with shampoo. See, I'm just below,on the back porch, crouched between her knees,my bright scalp bared for the biteof the comb. The boys launch cicadas onto the roof,the girls mock-scold, their heads lacquered whiteto drown the new colonies. Except for J who propsthe Children's Bible against the freezer boxto study its pictures: a sword bars the gardenwith orange flames, fig-leafed Adam and Eveclutch each other at the gates. J drops the tomeon the cement slab where cloudy drops pool.I try to make out the teeming multitude.Mom holds the comb for me to see:"You needed proof?" [End Page 13] Scythe There is a diagram imprinted in me,a first-tooth memory, my early phantomroot. I'd like to hold myself againstits hard edge, to arrange the giveand sag of me into its stony relief,but the order of the world is lesstaut, a false equation of hieroglyphand brassy machine. Herethe remains of a marbling tulipinfection. All those clawed headsmake a garish heap on the floor.In the clean language of want,a fixed lexicon for love:even the epitaph lies. The silenceis not still. They are mowingand mowing your grave [End Page 14] Double Occupancy My father purchases the burial plots years before she dies. A doubleoccupancy cement bunker, the bottom for my mother. She sits sunk in the reclining chair with her eyes closed, listening or not. Sometimes she slathers Vaseline on her cracked heels and covers them with socks. I move houses and give away the round table, the plants. I move houses again and give away the bed I shared for a time with someone and then with someone else. A month before her funeral, she asks my father to spend the extra money for a plot where they will lie side by side. I sit outside in the cold sun with my sister, the baby crying. "Mom" in the call record makes me think I've missed her again. There are still dishes in the sink when I leave for the airport. Someone in the flat dark is waiting for me to return across the distance. I press my temple to the glass. [End Page 15] Devotional I read words but don't hear God in them. —Susan Howe Friends send words on grieving. Instead, I cometo rely on the mystery of facts—the composition of tears, more mineral than fluid, a different saltarchitecture for relief or prayer. The book of photos magnifies their scale: landscapes repeatin miniature, binding crystal, the substance of stars or stone. On the bus, the crumbling paperwhitesgather in my lap, fallen from the nest of pages, the dusting of petals and pollen—how didn't I notice? Lost in the private room of a sentence. Under the microscope, spires of elationburst in fireworks of glass, while onion tears bloom in the dense symmetry of fern. [End Page 16] The bus rumbles across the city's grid.Through the windows, the stream of red brick material as myth, continuous as a voicemoving over the face of the waters. [End Page 17] Marlo Starr marlo starr is a writer and assistant professor of English at Wittenberg University. Her poems have appeared in The Threepenny Review, Ghost City Review, Napkin Poetry Review, and elsewhere. Copyright © 2023 Marlo Starr

Infant cry classification by using different deep neural network models and hand-crafted features

Crying is the way babies communicate with the outside world. These cries may be related to the needs of the baby or maybe an expression of a medical disorder. For this reason, infant cries are examined to support inexperienced parents and to make an early diagnosis if there is a medical disorder. Infant cry signals are classified using signal processing methods such as hand-crafted features or image processing methods based on the spectral image of the cry. In this study, we investigate the effect of using hand-crafted features and spectral images individually and hybrid in the classification of infant cries. In this context, experiments were conducted with the 1D CNN model, transfer learning, texture analysis methods, hand-crafted features, and their combination. In addition, the number of classes used in most of the studies in the literature is two or three, whereas in this study 5-classes in the dataset are used. Classification with hand-crafted and hybrid features was performed with SVM, RNN, and PNN. In addition, hand-crafted features were also classified with 1D CNN. GoogLeNet, ShuffleNet, ResNet-18 were used for transfer learning in image-based classification. The results show that texture analysis methods are insufficient and that hand-crafted feature sets and spectrogram and scalogram images provide high success. In addition, the 1D CNN model showed lower success than traditional classifiers and transfer learning models. Especially on DB2, the lowest result was obtained with 1D CNN in all experiments. The results were compared on two data commonly used in the literature, and a success rate of 97.6% and 95.2% was achieved. These results show that both signal processing methods and spectrogram and scalogram images can be used successfully in infant cry classification studies.

Maternal and infant touching behaviours during perturbed interactions: Associations with maternal depressive symptomatology and infant crying

Touch is an important means through which mothers and infants co-regulate during periods of stress or perturbation. The present study examined the synchrony of maternal and infant touching behaviours among 41 mother-infant dyads, some of whom were deemed at-risk due to maternal depressive symptomatology. Mothers and their 4-month-old infants participated in the Still-Face (maternal emotional unavailability; SF) and Separation (maternal physical unavailability; SP) procedures. Infant crying was examined across procedures and investigated as a brief period of perturbation. Results revealed that mothers and infants displayed a positive pattern of tactile synchrony (coordinated, analogous changes in touch) during infant crying episodes. However, dyads in the high depression group displayed significantly less affectionate touch during instances of infant crying. Furthermore, more depressive symptoms were associated with less maternal and infant touch and lower rates of infant crying. This group of dyads may be less expressive via touch, be less affected by disruptions in their interactions, have impaired regulatory abilities, or simply require minimal amounts of touch to mutually regulate following social stressors and during brief perturbation periods. These findings enrich our limited knowledge about the dynamic interplay of maternal and infant touch and inform preventative intervention programs for at-risk groups.

Acoustic Cry Characteristics in Preterm Infants and Developmental and Behavioral Outcomes at 2 Years of Age

Acoustic cry characteristics have been associated with severe medical problems in newborns. However, little is known about the utility of neonatal acoustic cry characteristics in the prediction of long-term outcomes of very preterm infants. To evaluate whether acoustic characteristics of infant cry at neonatal intensive care unit (NICU) discharge are associated with behavioral and developmental outcomes at age 2 years in infants born very preterm. Infants born less than 30 weeks postmenstrual age (PMA) were enrolled from April 2014 through June 2016 as part of a multicenter (9 US university affiliated NICUs) cohort study and followed to adjusted age 2 years. Reported analyses began on September 2021. Data were analyzed from September 2021 to September 2022. The primary exposure was premature birth (<30 weeks PMA). Cries were recorded during a neurobehavioral examination administered during the week of NICU discharge. Cry episodes were analyzed using a previously published computerized system to characterize cry acoustics. Year-2 outcomes included the Bayley-III Composite scores, Child Behavior Checklist (CBCL) and the Modified Checklist for Autism in Toddlers (M-CHAT R/F), dichotomized using clinically significant cutoffs (<85 on Bayley Language, Cognitive and/or Motor Composite scores, T-score >63 on the CBCL Internalizing, Externalizing and/or Total Problem Scales and total M-CHAT R/F score >2). Analyzed infants (363 participants) were primarily male (202 participants [55.65%]) and had a mean [SD] gestational age of 27.08 [1.95] weeks). Cross-validated random forest models revealed that cry acoustics were associated with 2-year outcomes. Tests of diagnostic odds ratios (DOR) revealed that infants who exhibited total problem behavior CBCL scores greater than 63 at age 2 years were 3.3 times more likely (95% CI, 1.44-7.49) to be identified as so by random forest model estimates relative to other infants (scores ≤63); this association was robust to adjustment for family-wise type-I error rates and covariate measures. Similar associations were observed for internalizing (DOR, 2.39; 95% CI, 1.04-5.47) and externalizing (DOR, 2.25; 95% CI, 1.12-4.54) scores on the CBCL, clinically significant language (DOR, 1.71; 95% CI, 1.10-2.67) and cognitive (DOR, 1.70; 95% CI, 1.00-2.88) scores on the Bayley-III, and a positive autism screen on the M-CHAT (DOR, 1.91; 95% CI, 1.05-3.44). In this cohort study of preterm infants, findings pointed to the potential use of acoustic cry characteristics in the early identification of risk for long-term developmental and behavioral deficits.

Ensemble of multimodal deep learning autoencoder for infant cry and pain detection

Background: Babies cannot communicate their pain properly. Several pain scores are developed, but they are subjective and have high variability inter-observer agreement. The aim of this study was to construct models that use both facial expression and infant voice in classifying pain levels and cry detection. Methods: The study included a total of 23 infants below 12-months who were treated at Dr Soetomo General Hospital. The the Face Leg Activity Cry and Consolability (FLACC) pain scale and recordings of the baby's cries were taken in the video format. A machine-learning-based system was created to detect infant cries and pain levels. Spectrograms with the Short-Time Fourier Transform were used to convert the audio data into a time-frequency representation. Facial features combined with voice features extracted by using the Deep Learning Autoencoders was used for the classification of infant pain levels. Two types of autoencoders: Convolutional Autoencoder and Variational Autoencoder were used for both faces and voices. Result: The goal of the autoencoder was to produce a latent-vector with much smaller dimensions that was still able to recreate the data with minor losses. From the latent-vectors, a multimodal data representation for Convolutional Neural Network (CNN) was used for producing a relatively high F1 score, higher than single data modal such as the voice or facial expressions alone. Two major parts of the experiment were: 1. Building the three autoencoder models, which were autoencoder for the infant’s face, amplitude spectrogram, and dB-scaled spectrogram of infant’s voices. 2. Utilising the latent-vector result from the autoencoders to build the cry detection and pain classification models. Conclusion: In this paper, four pain classifier models with a relatively good F1 score was developed. These models were combined by using ensemble methods to improve performance, which resulted in a better F1 score.