Automated analysis of child phonetic production using naturalistic recordings

Parentese in infancy predicts 5-year language complexity and conversational turns

Parental input is considered a key predictor of language achievement during the first years of life, yet relatively few studies have assessed its effects on longer-term outcomes. We assess the effects of parental quantity of speech, use of parentese (the acoustically exaggerated, clear, and higher-pitched speech), and turn-taking in infancy, on child language at 5 years. Using a longitudinal dataset of daylong LENA recordings collected with the same group of English-speaking infants (N=44) at 6, 10, 14, 18, 24 months and then again at 5 years, we demonstrate that parents' consistent (defined as stable and high) use of parentese in infancy was a potent predictor of lexical diversity, mean length of utterance, and frequency of conversational turn-taking between children and adults at Kindergarten entry. Together, these findings highlight the potential importance of a high-quality language learning environment in infancy for success at the start of formal schooling.

The Stability and Validity of Automated Indices of Vocal Development in Infants With Autistic and Non-Autistic Siblings

PURPOSE

This study evaluates the extent to which automated indices of vocal development are stable and valid for predicting language in infants at increased familial likelihood for autism and/or language impairment and relatively lower likelihood infants.

METHOD

A group of infants with autistic siblings (Sibs-autism; 20 infants) and a comparison group of infants with non-autistic siblings (Sibs-NA; 20 infants) wore Language ENvironment Analysis (LENA) recording devices for 16 hr on 2 days within a 1-week period. Extant software was used to derive several putative indices of vocal development from these recordings. Stability of these variables was examined across and within groups. Expressive and receptive language aggregates were calculated for each participant. Multiple regression analyses were used to (a) evaluate zero-order correlations for variables derived from LENA recordings with concurrent and future language and (b) test whether those associations were moderated by group status.

RESULTS

Both stability and validity differed by variable and group status. All variables reached acceptable stability in the Sibs-autism group within two to three observations, whereas stability of most variables was attenuated in the Sibs-NA group. No variables were associated with concurrent language in the theoretically motivated direction across groups, but two variables were strongly associated with concurrent expressive language in only the Sibs-NA group. Additionally, two variables were associated with later expressive language, though these correlations were again stronger in the Sibs-NA versus Sibs-autism group.

CONCLUSIONS

Although selected automated indices of vocal development were stable in Sibs-autism and/or valid for predicting expressive language within Sibs-NA, no scores showed strong, theoretically motivated associations with language within the Sibs-autism group. Automated indices of vocal development may, thus, have limited validity or clinical utility for predicting language development in infants at elevated familial likelihood for autism.

SUPPLEMENTAL MATERIAL

https://doi.org/10.23641/asha.24415735.

Investigating Children's Interactions in Preschool Classrooms: An Overview of Research Using Automated Sensing Technologies

New technologies that combine digital sensors with automated processing algorithms are now being deployed to study preschool classrooms. This article provides an overview of these new sensing technologies, focusing on automated speaker classification, the analysis of children's and teachers' speech, and the detection and analysis of their movements over the course of the school day. Findings from recent studies utilizing these technologies are presented to illustrate the contribution of these sensing technologies to our understanding of classroom processes that predict children's language and social development. In particular, the potential to collect extended real-time data on the speech and movement of all children and teachers in a classroom provides a broader window on the variability of individual children's interactions with peers and teachers and their integration into classroom social networks. The article describes current challenges related to the use of sensing technologies in preschool settings, as well as advances that may overcome these challenges and allow for more in-depth investigations of children's early classroom experiences.

Sex differences in infant vocalization and the origin of language

Seeking to discern the earliest sex differences in language-related activities, our focus is vocal activity in the first two years of life, following up on recent research that unexpectedly showed boys produced significantly more speech-like vocalizations (protophones) than girls during the first year of life.We now bring a much larger body of data to bear on the comparison of early sex differences in vocalization, data based on automated analysis of all-day recordings of infants in their homes. The new evidence, like that of the prior study, also suggests boys produce more protophones than girls in the first year and offers additional basis for informed speculation about biological reasons for these differences. More broadly, the work offers a basis for informed speculations about foundations of language that we propose to have evolved in our distant hominin ancestors, foundations also required in early vocal development of modern human infants.

Effects of multilingualism on Australian infants' language environments in early childhood education centers

This study investigates differences in the language environments experienced by multilingual and monolingual infants in early childhood education and care (ECEC) settings. The Language Environment Analysis (LENA) technology was used to collect day-long audio-recordings from 181 one-year-old infants (age range from 12 to 21 months). We examined whether infants' multilingual status predicts the amount of educators' language input (adult word count, AWC), child vocalizations (CVC) and conversational turns (CTC), as well as interaction effects on AWC, CVC and CTC of infants' multilingual status and other infant, home and ECEC characteristics. Multilevel mixed effects models revealed no main effect of infants' multilingual status on the language environment outcome variables. Instead, infant gender significantly predicted adult word count, with female infants hearing more words from educators than male infants. There was a significant interaction effect between the infants' multilingual status and both their age and length of time in an ECEC setting on child vocalizations. While monolingual infants produced more vocalizations as their age increased, multilingual infants did not show this increase in vocalizations with age. Further, the difference between monolingual and multilingual children's vocalizations decreased as the length of time in ECEC increased. There were no significant predictors of conversational turns. Findings from this study suggest that early childhood educators do not adjust their talk according to the multilingual status of the infants. However, multilingual infants do not increase their vocalizations as their age increases to the same extent as do their monolingual peers. The interaction effect between multilingualism and the length of ECEC attendance also implies that ECEC environments may be particularly beneficial for supporting multilingual infants' vocalizations. This study highlights the need to provide pedagogical support to educators to help them to encourage multilingual infants' vocalizations in ECEC settings.

Objective measurement of vocalizations in the assessment of autism spectrum disorder symptoms in preschool age children

Assessment of autism spectrum disorder (ASD) relies on expert clinician observation and judgment, but objective measurement tools have the potential to provide additional information on ASD symptom severity. Diagnostic evaluations for ASD typically include the autism diagnostic observation schedule (ADOS-2), a semi-structured assessment composed of a series of social presses. The current study examined associations between concurrent objective features of child vocalizations during the ADOS-2 and examiner-rated autism symptom severity. The sample included 66 children (49 male; M = 40 months, SD = 10.58) evaluated in a university-based clinic, 61 of whom received an ASD diagnosis. Research reliable administration of the ADOS-2 provided social affect (SA) and restricted and repetitive behavior (RRB) calibrated severity scores (CSS). Audio was recorded from examiner-worn eyeglasses during the ADOS-2 and child and adult speech were differentiated with LENA SP Hub. PRAAT was used to ascertain acoustic features of the audio signal, specifically the mean fundamental vocal frequency (F0) of LENA-identified child speech-like vocalizations (those with phonemic content), child cry vocalizations, and adult speech. Sphinx-4 was employed to estimate child and adult phonological features indexed by the average consonant and vowel count per vocalization. More than a quarter of the variance in ADOS-2 RRB CSS was predicted by the combination of child phoneme count per vocalization and child vocalization F0. Findings indicate that both acoustic and phonological features of child vocalizations are associated with expert clinician ratings of autism symptom severity. LAY SUMMARY: Determination of the severity of autism spectrum disorder is based in part on expert (but subjective) clinician observations during the ADOS-2. Two characteristics of child vocalizations-a smaller number of speech-like sounds per vocalization and higher pitched vocalizations (including cries)-were associated with greater autism symptom severity. The results suggest that objectively ascertained characteristics of children's vocalizations capture variance in children's restricted and repetitive behaviors that are reflected in clinician severity indices.

Objectively measured teacher and preschooler vocalizations: Phonemic diversity is associated with language abilities

Over half of US children are enrolled in preschools, where the quantity and quality of language input from teachers are likely to affect children's language development. Leveraging repeated objective measurements, we examined the rate per minute and phonemic diversity of child and teacher speech-related vocalizations in preschool classrooms and their association with children's end-of-year receptive and expressive language abilities measured with the Preschool Language Scales (PLS-5). Phonemic diversity was computed as the number of unique consonants and vowels in a speech-related vocalization. We observed three successive cohorts of 2.5-3.5-year-old children enrolled in an oral language classroom that included children with and without hearing loss (N = 29, 16 girls, 14 Hispanic). Vocalization data were collected using child-worn audio recorders over 34 observations spanning three successive school years, yielding 21.53 mean hours of audio recording per child. The rate of teacher vocalizations positively predicted the rate of children's speech-related vocalizations while the phonemic diversity of teacher vocalizations positively predicted the phonemic diversity of children's speech-related vocalizations. The phonemic diversity of children's speech-related vocalizations was a stronger predictor of end-of-year language abilities than the rate of children's speech-related vocalizations. Mediation analyses indicated that the phonemic diversity of teacher vocalizations was associated with children's receptive and expressive language abilities to the extent that it influenced the phonemic diversity of children's own speech-related vocalizations. The results suggest that qualitatively richer language input expands the phonemic diversity of children's speech, which in turn is associated with language abilities.

Computational approaches to understanding interaction and development

Audio-visual recording and location tracking produce enormous quantities of digital data with which researchers can document children's everyday interactions in naturalistic settings and assessment contexts. Machine learning and other computational approaches can produce replicable, automated measurements of these big behavioral data. The economies of scale afforded by repeated automated measurements offer a potent approach to investigating linkages between real-time behavior and developmental change. In our work, automated measurement of audio from child-worn recorders-which quantify the frequency of child and adult speech and index its phonemic complexity-are paired with ultrawide radio tracking of children's location and interpersonal orientation. Applications of objective measurement indicate the influence of adult behavior in both expert ratings of attachment behavior and ratings of autism severity, suggesting the role of dyadic factors in these "child" assessments. In the preschool classroom, location/orientation measures provide data-driven measures of children's social contact, fertile ground for vocal interactions. Both the velocity of children's movement toward one another and their social contact with one another evidence homophily: children with autism spectrum disorder, other developmental disabilities, and typically developing children were more likely to interact with children in the same group even in inclusive preschool classrooms designed to promote interchange between all children. In the vocal domain, the frequency of peer speech and the phonemic complexity of teacher speech predict the frequency and phonemic complexity of children's own speech over multiple timescales. Moreover, children's own speech predicts their assessed language abilities across disability groups, suggesting how everyday interactions facilitate development.

The utility of LENA as an indicator of developmental outcomes for young children with autism

BACKGROUND

There is growing understanding of the potential benefits of a multi-method approach to accurately capture language skills of children on the autism spectrum. Tools such as Language ENvironment Analysis (LENA) provide an efficient means of capturing and analysing early child vocalizations (CVs) and the language learning environment. While developed to capture whole-day recordings of child language in naturalistic settings, there is potential utility in capturing, but little knowledge about, primary LENA metrics-including CVs and conversational turns (CTs)-and novel metrics, such as vocalization ratios (VRs), sampled in clinical practice settings where children are often seen. Moreover, recent research indicates that the novel VR may offer a broad indicator of children's developmental level, beyond just their language abilities, a hypothesis yet to be investigated in a large sample of children for whom the LENA was designed (i.e., pre-schoolers).

AIMS

To explore the extent to which primary and novel LENA metrics collected during brief one-to-one clinical interaction was a useful indicator of developmental outcomes for children on the autism spectrum.

METHODS & PROCEDURES

Participants were recruited as part of an on-going research programme evaluating early intervention outcomes (n = 99; age 14-47 months). Language samples were collected at intake (T1) using the wearable LENA Digital Language Processors during a one-to-one, play-based assessment with a clinician. Direct (Mullen Scales of Early Learning-MSEL) and parent-report (Vineland Adaptive Behavior Scales-VABS) measures of verbal and non-verbal skills were also collected at intake (T1) and again at exit (T2), approximately 12 months later.

OUTCOMES & RESULTS

Few correlations were observed between child measures and CVs, a primary LENA metric. The novel VR metric was associated with concurrent direct assessment (MSEL) (and to a lesser extent parent report; VABS) measures of verbal and non-verbal skills, with moderate positive correlations found between VRs and all directly assessed subscale scores. However, VRs did not uniquely contribute to the prediction of child outcomes when baseline skills were also considered.

CONCLUSIONS & IMPLICATIONS

The novel VR may provide an insight into autistic children's overall development in addition to their language ability, suggesting that even when collected in a short recording format, LENA might be a useful component of a multi-method assessment approach.

WHAT THIS PAPER ADDS

What is already known on the subject To accurately capture language skills of children on the autism spectrum, multi-method approaches, including natural language sampling, are recommended. Tools such as LENA provide an efficient means of capturing and analysing naturalistic child language and the language learning environment. What this paper adds to existing knowledge This study demonstrates the potential benefits, and limitations, of using LENA to augment assessment of young children on the autism spectrum. Specifically, LENA provides a complementary, and low burden, method for capturing child language samples. What are the potential or actual clinical implications of this work? Novel metrics, such as the VR, collected during brief clinical interactions might be a useful component of a multi-method assessment approach for children on the autism spectrum.

Protophones, the precursors to speech, dominate the human infant vocal landscape

Human infant vocalization is viewed as a critical foundation for vocal learning and language. All apes share distress sounds (shrieks and cries) and laughter. Another vocal type, speech-like sounds, common in human infants, is rare but not absent in other apes. These three vocal types form a basis for especially informative cross-species comparisons. To make such comparisons possible we need empirical research documenting the frequency of occurrence of all three. The present work provides a comprehensive portrayal of these three vocal types in the human infant from longitudinal research in various circumstances of recording. Recently, the predominant vocalizations of the human infant have been shown to be speech-like sounds, or 'protophones', including both canonical and non-canonical babbling. The research shows that protophones outnumber cries by a factor of at least five based on data from random-sampling of all-day recordings across the first year. The present work expands on the prior reports, showing the protophones vastly outnumber both cry and laughter in both all-day and laboratory recordings in various circumstances. The data provide new evidence of the predominance of protophones in the infant vocal landscape and illuminate their role in human vocal learning and the origin of language. This article is part of the theme issue 'Vocal learning in animals and humans'.

Analysis of acoustic and voice quality features for the classification of infant and mother vocalizations

Classification of infant and parent vocalizations, particularly emotional vocalizations, is critical to understanding how infants learn to regulate emotions in social dyadic processes. This work is an experimental study of classifiers, features, and data augmentation strategies applied to the task of classifying infant and parent vocalization types. Our data were recorded both in the home and in the laboratory. Infant vocalizations were manually labeled as cry, fus (fuss), lau (laugh), bab (babble) or scr (screech), while parent (mostly mother) vocalizations were labeled as ids (infant-directed speech), ads (adult-directed speech), pla (playful), rhy (rhythmic speech or singing), lau (laugh) or whi (whisper). Linear discriminant analysis (LDA) was selected as a baseline classifier, because it gave the highest accuracy in a previously published study covering part of this corpus. LDA was compared to two neural network architectures: a two-layer fully-connected network (FCN), and a convolutional neural network with self-attention (CNSA). Baseline features extracted using the OpenSMILE toolkit were augmented by extra voice quality, phonetic, and prosodic features, each targeting perceptual features of one or more of the vocalization types. Three web data augmentation and transfer learning methods were tested: pre-training of network weights for a related task (adult emotion classification), augmentation of under-represented classes using data uniformly sampled from other corpora, and augmentation of under-represented classes using data selected by a minimum cross-corpus information difference criterion. Feature selection using Fisher scores and experiments of using weighted and unweighted samplers were also tested. Two datasets were evaluated: a benchmark dataset (CRIED) and our own corpus. In terms of unweighted-average recall of CRIED dataset, the CNSA achieved the best UAR compared with previous studies. In terms of classification accuracy, weighted F1, and macro F1 of our own dataset, the neural networks both significantly outperformed LDA; the FCN slightly (but not significantly) outperformed the CNSA. Cross-examining features selected by different feature selection algorithms permits a type of post-hoc feature analysis, in which the most important acoustic features for each binary type discrimination are listed. Examples of each vocalization type of overlapped features were selected, and their spectrograms are presented, and discussed with respect to the type-discriminative acoustic features selected by various algorithms. MFCC, log Mel Frequency Band Energy, LSP frequency, and F1 are found to be the most important spectral envelope features; F0 is found to be the most important prosodic feature.

Quantifying Voice Characteristics for Detecting Autism

The presence of prosodic anomalies in autistic is recognized by experienced clinicians but their quantitative analysis is a cumbersome task beyond the scope of typical pen and pencil assessment. This paper proposes an automatic approach allowing to tease apart various aspects of prosodic abnormalities and to translate them into fine-grained, automated, and quantifiable measurements. Using a harmonic model (HM) of voiced signal, we isolated the harmonic content of speech and computed a set of quantities related to harmonic content. Employing these measures, along with standard speech measures such as loudness, we successfully trained machine learning models for distinguishing individuals with autism from those with typical development (TD). We evaluated our models empirically on a task of detecting autism on a sample of 118 youth (90 diagnosed with autism and 28 controls; mean age: 10.9 years) and demonstrated that these models perform significantly better than a chance model. Voice and speech analyses could be incorporated as novel outcome measures for treatment research and used for early detection of autism in preverbal infants or toddlers at risk of autism.

Vocal development in a large-scale crosslinguistic corpus

This study evaluates whether early vocalizations develop in similar ways in children across diverse cultural contexts. We analyze data from daylong audio recordings of 49 children (1-36 months) from five different language/cultural backgrounds. Citizen scientists annotated these recordings to determine if child vocalizations contained canonical transitions or not (e.g., "ba" vs. "ee"). Results revealed that the proportion of clips reported to contain canonical transitions increased with age. Furthermore, this proportion exceeded 0.15 by around 7 months, replicating and extending previous findings on canonical vocalization development but using data from the natural environments of a culturally and linguistically diverse sample. This work explores how crowdsourcing can be used to annotate corpora, helping establish developmental milestones relevant to multiple languages and cultures. Lower inter-annotator reliability on the crowdsourcing platform, relative to more traditional in-lab expert annotators, means that a larger number of unique annotators and/or annotations are required, and that crowdsourcing may not be a suitable method for more fine-grained annotation decisions. Audio clips used for this project are compiled into a large-scale infant vocalization corpus that is available for other researchers to use in future work.

How Chatty Are Daddies? An Exploratory Study of Infants' Language Environments

Purpose Fathers play a critical but underresearched role in their children's cognitive and linguistic development. Focusing on two-parent families with a mother and a father, the present longitudinal study explores the amount of paternal input infants hear during the first 2 years of life, how this input changes over time, and how it relates to child volubility. We devote special attention to parentese, a near-universal style of infant-directed speech, distinguished by its higher pitch, slower tempo, and exaggerated intonation. Method We examined the daylong recordings of the same 23 infants at ages 6, 10, 14, 18, and 24 months, given English-speaking families. The infants were recorded in the presence of their parents (mother-father dyads), who were predominantly White and ranged from mid to high socioeconomic status (SES). We analyzed the effects of parent gender and child age on adult word counts and parentese, as well as the effects of maternal and paternal word counts and parentese on child vocalizations. Results On average, the infants were exposed to 46.8% fewer words and 51.9% less parentese from fathers than from mothers, even though paternal parentese grew at a 2.8-times faster rate as the infants aged. An asymmetry emerged where maternal word counts and paternal parentese predicted child vocalizations, but paternal word counts and maternal parentese did not. Conclusions While infants may hear less input from their fathers than their mothers in predominantly White, mid-to-high SES, English-speaking households, paternal parentese still plays a unique role in their linguistic development. Future research on sources of variability in child language outcomes should thus control for parental differences since parents' language can differ substantially and differentially predict child language.

Parent education to improve early language development: A preliminary evaluation of LENA StartTM

Parents play an important role in creating home language environments that promote language development. A nonequivalent group design was used to evaluate the effectiveness of a community-based implementation of LENA Start™, a parent-training program aimed at increasing the quantity of adult words (AWC) and conversational turns (CT). Parent-child dyads participated in LENA Start™ (n = 39) or a generic parent education program (n = 17). Overall, attendance and engagement in the LENA StartTM program were high: 72% of participants met criteria to graduate from the program. Within-subject gains were positive for LENA Start™ families. Comparison families declined on these measures. However, both effects were non-significant. Between-group analyses revealed small to medium-sized effects favoring LENA Start™ and these were significant for child vocalizations (CV) and CT but not AWC. These results provide preliminary evidence that programs like LENA StartTM can be embedded in community-based settings to promote quality parent-child language interactions.

The education word gap emerges by 18 months: findings from an Australian prospective study

BACKGROUND

The idea of the '30 million word gap' suggests families from more socioeconomically advantaged backgrounds engage in more verbal interactions with their child than disadvantaged families. Initial findings from the Language in Little Ones (LiLO) study up to 12 months showed no word gap between maternal education groups.

METHODS

Families with either high or low maternal education were purposively recruited into a five-year prospective study. We report results from the first three waves of LiLO when children were 6, 12 and 18 months old. Day-long audio recordings, obtained using the Language Environment Analysis software, provided counts of adult words spoken to the child, child vocalizations and conversational turns.

RESULTS

By the time children were 18 months old all three measures of talk were 0.5 to 0.7 SD higher among families with more education, but with large variation within education groups. Changes in talk from 6 to 18 months highlighted that families from low educated backgrounds were decreasing the amount they spoke to their children (- 4219.54, 95% CI -6054.13, - 2384.95), compared to families from high educated backgrounds who remained relatively stable across this age period (- 369.13, 95% CI - 2344.57, 1606.30).

CONCLUSIONS

The socioeconomic word gap emerges between 12 and 18 months of age. Interventions to enhance maternal communication, child vocalisations and vocabulary development should begin prior to 18 months.

Exposure to electronic media between 6 and 24 months of age: An exploratory study

This study explores the associations between electronic media exposure, age, and socioeconomic status (SES) in a longitudinal sample of 24 infants from English-speaking families. Leveraging Language ENvironment Analysis (LENA) technology, the study seeks to characterize the relation between electronic media exposure and parental and child vocal activity. We analyzed ecologically valid, daylong audio recordings collected in infants' homes when they were 6, 10, 14, 18, and 24 months old. SES was measured with the Hollingshead Index, and exposure to electronic media and adult and infant vocal activity were measured automatically with LENA. On average, the children in the sample were exposed to 58 min of electronic media daily. We found that electronic media exposure was negatively associated with SES and decreased with child age, but only amongst high-SES families. We also found that electronic media exposure negatively impacted concurrent adult and child vocal activity, irrespective of SES and infant age. The present findings are an important step forward in examining the role of demographic factors in exposure to electronic media and enhance our understanding of the mechanisms through which exposure to electronic media may impact linguistic development in infancy and beyond.

Is silence golden? A pilot study exploring associations between children's language environment and their language skills in Estonian-speaking families

This two-wave longitudinal study explored how Estonian children's language environment relates to their language skills. The Language ENvironment Analysis (LENA) system's automated measures were used as a source of information about children's home language environment. Children's expressive vocabulary was measured via the parent-reported Estonian CDI III (ECDI-III), and language comprehension and production were measured via the examiner-administered New Reynell Developmental Language Scales (NRDLS). The assessments were made 1 year apart at ages 3;0 (years; months) (N = 22) and 4;0 (N = 19). The results revealed wide variability in children's home language environment and language skills. Girls' language production scores were higher; they heard a larger quantity of adult words and spent less time in noisy environments than boys at Wave 2. At Wave 1, children's word count was positively associated with productive language scores, whereas silence was negatively associated with language production and expressive vocabulary. At Wave 2, children who had been more exposed to electronic media scored higher on expressive vocabulary. Distant speech at Wave 1 was positively correlated with language comprehension scores at Wave 2 also when controlling for language comprehension scores at Wave 1. The results, which indicate that distant talk is a positive aspect and silence is a negative aspect of the language environment, highlight the importance of "languagizing" homes also in cultures where silence tends to be more highly valued and talkativeness tends to be less highly valued when compared with English-speaking middle-class families.

A First Step toward the Clinical Application of Landmark-Based Acoustic Analysis in Child Mandarin

As an initial step for the clinical application of landmark-based acoustic analysis in child Mandarin, the study quantified the developmental trajectories of consonants produced by four-to-seven-year-old children who acquired Taiwanese Mandarin as their first language. The results from a total of 80 children (20 in each age group, with gender balanced) indicated that younger age groups produced more +b landmark features than seven-year-olds did, showing that the development of obstruents was not completed by the age of six. A multiple regression showed that the participants' speech intelligibility scores could be predicted by landmark features. Additionally, the +b landmark feature demonstrated the strongest net effect on speech intelligibility scores. The findings indicated that: (a) the landmark feature +b was an essential indicator of speech development in child Mandarin and; (b) the consonantal development in child Mandarin could be predicted by the physiological complexity of the articulatory gestures. Future studies focusing on a wider range of population (e.g., typically developing adults, aging and other clinical groups) with different language backgrounds are encouraged to apply landmark-based acoustic analysis to trace the linguistic development of a particular group.

Naturalistic Language Input is Associated with Resting-State Functional Connectivity in Infancy

The quantity and quality of the language input that infants receive from their caregivers affects their future language abilities; however, it is unclear how variation in this input relates to preverbal brain circuitry. The current study investigated the relation between naturalistic language input and the functional connectivity (FC) of language networks in human infancy using resting-state functional magnetic resonance imaging (rsfMRI). We recorded the naturalistic language environments of five- to eight-month-old male and female infants using the Linguistic ENvironment Analysis (LENA) system and measured the quantity and consistency of their exposure to adult words (AWs) and adult-infant conversational turns (CTs). Infants completed an rsfMRI scan during natural sleep, and we examined FC among regions of interest (ROIs) previously implicated in language comprehension, including the auditory cortex, the left inferior frontal gyrus (IFG), and the bilateral superior temporal gyrus (STG). Consistent with theory of the ontogeny of the cortical language network (Skeide and Friederici, 2016), we identified two subnetworks posited to have distinct developmental trajectories: a posterior temporal network involving connections of the auditory cortex and bilateral STG and a frontotemporal network involving connections of the left IFG. Independent of socioeconomic status (SES), the quantity of CTs was uniquely associated with FC of these networks. Infants who engaged in a larger number of CTs in daily life had lower connectivity in the posterior temporal language network. These results provide evidence for the role of vocal interactions with caregivers, compared with overheard adult speech, in the function of language networks in infancy.

Comparing Automatic and Manual Measures of Parent-Infant Conversational Turns: A Word of Caution

The Language ENvironment Analysis system (LENA) records children’s language environment and provides an automatic estimate of adult–child conversational turn count (CTC). The present study compares LENA’s CTC estimate to manually coded CTC on a sample of 70 English‐speaking infants recorded longitudinally at 6, 10, 14, 18, and 24 months of age. At each age, LENA’s CTC was significantly higher than manually coded CTC (all ps < .001, Cohen’s ds: 0.9–2.05), with the largest discrepancies between the two methods observed at younger ages. The Limits of Agreement Analyses confirm wide disagreements between the two methods, highlighting potential problems with automatic measurement of parent–infant verbal interaction. These findings suggest that future studies should validate LENA’s CTC estimates with manual coding.

Predicting Expressive Language From Early Vocalizations in Young Children With Autism Spectrum Disorder: Which Vocal Measure Is Best?

Purpose This study was designed to test the incremental validity of more expensive vocal development variables relative to less expensive variables for predicting later expressive language in children with autism spectrum disorder (ASD). We devote particular attention to the added value of coding the quality of vocalizations over the quantity of vocalizations because coding quality adds expense to the coding process. We are also interested in the added value of more costly human-coded vocal variables relative to those generated through automated analyses. Method Eighty-seven children with ASD aged 13-30 months at study initiation participated. For quantity of vocalizations, we derived one variable from human coding of brief communication samples and one from an automated process for daylong naturalistic audio samples. For quality of vocalizations, we derived four human-coded variables and one automated variable. A composite expressive language measure was derived at study entry, and 6 and 12 months later. The 12 months-centered intercept of a simple linear growth trajectory was used to quantify later expressive language. Results When statistically controlling for human-coded or automated quantity of vocalization variables, human-coded quality of vocalization variables exhibited incremental validity for predicting later expressive language skills. Human-coded vocal variables also predicted later expressive language skills when controlling for the analogous automated vocal variables. Conclusion In sum, these findings support devoting resources to human coding of the quality of vocalizations from communication samples to predict later expressive language skills in young children with ASD despite the greater costs of deriving these variables. Supplemental Material https://doi.org/10.23641/asha.12276458.

Remote Microphone System Use in the Homes of Children With Hearing Loss: Impact on Caregiver Communication and Child Vocalizations

Purpose This study examined the impact of home use of remote microphone systems (RMSs) on caregiver communication and child vocalizations in families of children with hearing loss. Method We drew on data from a prior study in which Language ENvironmental Analysis recorders were used with 9 families during 2 consecutive weekends-1 that involved using an RMS and 1 that did not. Audio samples from Language ENvironmental Analysis recorders were (a) manually coded to quantify the frequency of verbal repetitions and alert phrases caregivers utilized in communicating to children with hearing loss and (b) automatically analyzed to quantify children's vocalization rate, duration, complexity, and reciprocity when using and not using an RMS. Results When using an RMS at home, caregivers did not repeat or clarify their statements as often as when not using an RMS while communicating with their children with hearing loss. However, no between-condition differences were observed in children's vocal characteristics. Conclusions Results provide further support for home RMS use for children with hearing loss. Specifically, findings lend empirical support to prior parental reports suggesting that RMS use eases caregiver communication in the home setting. Studies exploring RMS use over a longer duration of time might provide further insight into potential long-term effects on children's vocal production.

Parent coaching increases conversational turns and advances infant language development

Parental language input is one of the best predictors of children's language achievement. Parentese, a near-universal speaking style distinguished by higher pitch, slower tempo, and exaggerated intonation, has been documented in speech directed toward young children in many countries. Previous research shows that the use of parentese and parent-child turn-taking are both associated with advances in children's language learning. We conducted a randomized controlled trial to determine whether a parent coaching intervention delivered when the infants are 6, 10, and 14 mo of age can enhance parental language input and whether this, in turn, changes the trajectory of child language development between 6 and 18 mo of age. Families of typically developing 6-mo-old infants (n = 71) were randomly assigned to intervention and control groups. Naturalistic first-person audio recordings of the infants' home language environment and vocalizations were recorded when the infants were 6, 10, 14, and 18 mo of age. After the 6-, 10-, and 14-mo recordings, intervention, but not control parents attended individual coaching appointments to receive linguistic feedback, listen to language input in their own recordings, and discuss age-appropriate activities that promote language growth. Intervention significantly enhanced parental use of parentese and parent-child turn-taking between 6 and 18 mo. Increases in both variables were significantly correlated with children's language growth during the same period, and children's language outcomes at 18 mo. Using parentese, a socially and linguistically enhanced speaking style, improves children's social language turn-taking and language skills. Research-based interventions targeting social aspects of parent-child interactions can enhance language outcomes.

How many words are Australian children hearing in the first year of life?

Background

There is evidence that parents from more socioeconomically disadvantaged backgrounds engage in fewer verbal interactions with their child than more advantaged parents. This leads to the so-called, ‘30 million-word gap’. This study aims to investigate the number of words children hear and the number of vocalizations children produce in their first year of life and examines whether these aspects of the early language home environment differ by maternal education.

Methods

Mothers were recruited into a five-year prospective cohort study and categorized into either high or low maternal education groups. Data was derived from the first two waves of the study, when the children were six and twelve months old. At both waves, children were involved in day-long audio recordings using the Language Environment Analysis software that provided automatic counts of adult words spoken to the child, child vocalizations and conversational turns. Descriptive results are presented by maternal education groups.

Results

There was large variation within each maternal education group, with the number of adult words spoken to the child ranging from 2958 to 39,583 at six months and 4389 to 45,849 at twelve months. There were no meaningful differences between adult words, child vocalizations or conversational turns across maternal education groups at either wave of data collection.

Conclusions

These results show that a word gap related to maternal education is not apparent up to twelve months of age. The large variability among both maternal education groups suggests that universal interventions that encourage all parents to talk more to their child may be more appropriate than interventions targeted towards disadvantaged families during the first year of life.

A modular, extensible approach to massive ecologically valid behavioral data

We explore here the application of modern computer hardware and software to the collection and analysis of behavioral data. We discuss the issues of ecological validity, storage and processing, data permanence, automation, validity, and algorithmic determinism. Taking the modern landscape into account, we demonstrate several varying projects we have recently undertaken as proofs of concept of the viability and utility of this approach. In particular, we describe four research projects, which involve work on child-directed speech; the application of automatic methods to clinical populations, including children with hearing loss; quality control and the assessment of validity; and the sharing of data in a public database. We conclude by pointing out how the methodology described here can be extended to a wide variety of interdisciplinary and detailed projects that are likely to lead to better science and improved outcomes for populations served by the behavioral, social, and health sciences.

Acoustic properties of early vocalizations in infants with fragile X syndrome

Fragile X syndrome (FXS) is a neurogenetic syndrome characterized by cognitive impairments and high rates of autism spectrum disorder (ASD). FXS is often highlighted as a model for exploring pathways of symptom expression in ASD due to the high prevalence of ASD symptoms in this population and the known single-gene cause of FXS. Early vocalization features-including volubility, complexity, duration, and pitch-have shown promise in detecting ASD in idiopathic ASD populations but have yet to be extensively studied in a population with a known genetic cause for ASD such as FXS. Investigating early trajectories of these features in FXS may inform our limited knowledge of potential mechanisms that predict later social communication outcomes. The present study addresses this need by presenting preliminary findings which (a) characterize early vocalization features in FXS relative to low-risk controls (LRC) and (b) test the specificity of associations between these features and language and ASD outcomes. We coded vocalization features during a standardized child-examiner interaction for 39 nine-month-olds (22 FXS, 17 LRC) whose clinical outcomes were assessed at 24 months. Our results provide preliminary evidence that within FXS, associations between vocalization features and 24-month language outcomes may diverge from those observed in LRC, and that vocalization features may be associated with later ASD symptoms. These findings provide a starting point for more research exploring these features as potential early markers of ASD in FXS, which in turn may lead to improved early identification methods, treatment approaches, and overall well-being of individuals with ASD. Autism Res2019. © 2019 International Society for Autism Research, Wiley Periodicals, Inc. LAY SUMMARY: Although vocal features of 9-month-olds with FXS did not differ from those of low-risk controls, several features were associated with later language and ASD outcomes at 24 months in FXS. These preliminary results suggest acoustic data may be related to clinical outcomes in FXS and potentially other high-risk populations. Further characterizing these associations may facilitate understanding of biological mechanisms and risk factors associated with social communication development and ASD.

Validity of Vocal Communication and Vocal Complexity in Young Children with Autism Spectrum Disorder

To identify valid measures of vocal development in young children with autism spectrum disorder in the early stages of language learning, we evaluated the convergent validity, divergent validity, and sensitivity to change (across 12 months) of two measures of vocal communication and two measures of vocal complexity through conventional coding of communication samples. Participants included 87 children with autism spectrum disorder (M = 23.42 months at entry). All four vocal variables demonstrated consistent evidence of convergent validity, divergent validity, and sensitivity to change with large effect sizes for convergent validity and sensitivity to change. The results highlight the value of measuring vocal communication and vocal complexity in future studies.

Use of the LENA Autism Screen with Children who are Deaf or Hard of Hearing

Background and Objectives: This systematic review reports the evidence from the literature concerning the potential for using an automated vocal analysis, the Language ENvironment Analysis (LENA, LENA Research Foundation, Boulder, CO, USA) in the screening process for children at risk for autism spectrum disorder (ASD) and deaf or hard of hearing (D/HH). ASD and D/HH have increased comorbidity, but current behavioral diagnostic and screening tools have limitations. The LENA Language Autism Screen (LLAS) may offer an additional tool to disambiguate ASD from D/HH in young children. Materials and Methods: We examine empirical reports that use automatic vocal analysis methods to differentiate disordered from typically developing children. Results: Consensus across the sampled scientific literature shows support for use of automatic methods for screening and disambiguation of children with ASD and D/HH. There is some evidence of vocal differentiation between ASD, D/HH, and typically-developing children warranting use of the LLAS, but additional empirical evidence is needed to better understand the strengths and weaknesses of the tool. Conclusions: The findings reported here warrant further, more substantive, methodologically-sound research that is fully powered to show a reliable difference. Findings may be useful for both clinicians and researchers in better identification and understanding of communication disorders.

Acoustic Correlates and Adult Perceptions of Distress in Infant Speech-Like Vocalizations and Cries

Prior research has not evaluated acoustic features contributing to perception of human infant vocal distress or lack thereof on a continuum. The present research evaluates perception of infant vocalizations along a continuum ranging from the most prototypical intensely distressful cry sounds ("wails") to the most prototypical of infant sounds that typically express no distress (non-distress "vocants"). Wails are deemed little if at all related to speech while vocants are taken to be clear precursors to speech. We selected prototypical exemplars of utterances representing the whole continuum from 0 and 1 month-olds. In this initial study of the continuum, our goals are to determine (1) listener agreement on level of vocal distress across the continuum, (2) acoustic parameters predicting ratings of distress, (3) the extent to which individual listeners maintain or change their acoustic criteria for distress judgments across the study, (4) the extent to which different listeners use similar or different acoustic criteria to make judgments, and (5) the role of short-term experience among the listeners in judgments of infant vocalization distress. Results indicated that (1) both inter-rater and intra-rater listener agreement on degree of vocal distress was high, (2) the best predictors of vocal distress were number of vibratory regimes within utterances, utterance duration, spectral ratio (spectral concentration) in vibratory regimes within utterances, and mean pitch, (3) individual listeners significantly modified their acoustic criteria for distress judgments across the 10 trial blocks, (4) different listeners, while showing overall similarities in ratings of the 42 stimuli, also showed significant differences in acoustic criteria used in assigning the ratings of vocal distress, and (5) listeners who were both experienced and inexperienced in infant vocalizations coding showed high agreement in rating level of distress, but differed in the extent to which they relied on the different acoustic cues in making the ratings. The study provides clearer characterization of vocal distress expression in infants based on acoustic parameters and a new perspective on active adult perception of infant vocalizations. The results also highlight the importance of vibratory regime segmentation and analysis in acoustically based research on infant vocalizations and their perception.

Correlation and agreement between Language ENvironment Analysis (LENA™) and manual transcription for Dutch natural language recordings

The Language ENvironment Analysis system (LENA™) automatically analyzes the natural sound environments of children. Among other things, it estimates the amounts of adult words (AWC), child vocalizations (CV), conversational turns (CT), and electronic media (TV) that a child is exposed to. To assess LENA's reliability, we compared it to manual transcription. Specifically, we calculated the correlation and agreement between the LENA estimates and manual counts for 48 five-min audio samples. These samples were selected from eight day-long recordings of six Dutch-speaking children (ages 2-5). The correlations were strong for AWC, r =  . 87, and CV, r =  . 77, and comparatively low for CT, r =  . 52, and TV, r =  . 50. However, the agreement analysis revealed a constant bias in AWC counts, and proportional biases for CV and CT (i.e., the bias varied with the values for CV and CT). Agreement for detecting electronic media was poor. Moreover, the limits of agreement were wide for all four metrics. That is, the differences between LENA and the manual transcriptions for individual audio samples varied widely around the mean difference. This variation could indicate that LENA was affected by differences between the samples that did not equally affect the human transcribers. The disagreements and biases cast doubt on the comparability of LENA measurements across families and time, which is crucial for using LENA in research. Our sample is too small to conclude within which limits LENA's measurements are comparable, but it seems advisable to be cautious of factors that could systematically bias LENA's performance and thereby create confounds.

Deviations of acoustic low-level descriptors in speech features of a set of triplets, one with autism

Verbal speech of children diagnosed with ASD is explored in order to identify patterns autism has left in speech, and to model such patterns for implementing automatic diagnostic and screening frameworks. In this study, we identify the deviations of acoustic low-level descriptors (LLDs) in voice of an autistic adolescent from her typically developing triplet siblings. The goal is to identify the atypicality in voice introduced by autism under minimum gender, age, genetic, and language bias and use the gained insights to build a more generalized model by adding more subjects hierarchically. We report the most significant LLDs that describe the deviations of acoustic features due to autism for categories of utterances and feature groups.

Speech-driven mobile games for speech therapy: User experiences and feasibility

Purpose: To assist in remote treatment, speech-language pathologists (SLPs) rely on mobile games, which though entertaining, lack feedback mechanisms. Games integrated with automatic speech recognition (ASR) offer a solution where speech productions control gameplay. We therefore performed a feasibility study to assess children's and SLPs' experiences towards speech-controlled games, game feature preferences and ASR accuracy. Method: Ten children with childhood apraxia of speech (CAS), six typically developing (TD) children and seven SLPs trialled five games and answered questionnaires. Researchers also compared the results of ASR to perceptual judgment. Result: Children and SLPs found speech-controlled games interesting and fun, despite ASR-human disagreements. They preferred games with rewards, challenge and multiple difficulty levels. Automatic speech recognition-human agreement was higher for SLPs than children, similar between TD and CAS and unaffected by CAS severity (77% TD, 75% CAS - incorrect; 51% TD, 47% CAS, 71% SLP - correct). Manual stop recording yielded higher agreement than automatic. Word length did not influence agreement. Conclusion: Children's and SLPs' positive responses towards speech-controlled games suggest that they can engage children in higher intensity practice. Our findings can guide future improvements to the ASR, recording methods and game features to improve the user experience and therapy adherence.

Language Experience in the Second Year of Life and Language Outcomes in Late Childhood

OBJECTIVES

Quantity of talk and interaction in the home during early childhood is correlated with socioeconomic status (SES) and can be used to predict early language and cognitive outcomes. We tested the effectiveness of automated early language environment estimates for children 2 to 36 months old to predict cognitive and language skills 10 years later and examined effects for specific developmental age periods.

METHODS

Daylong audio recordings for 146 infants and toddlers were completed monthly for 6 months, and the total number of daily adult words and adult-child conversational turnswere automatically estimated with Language Environment Analysis software. Follow-up evaluations at 9 to 14 years of age included language and cognitive testing. Language exposure for 3 age groups was assessed: 2 to 17 months, 18 to 24 months, and ≥25 months. Pearson correlations and multiple linear regression analyses were conducted.

RESULTS

Conversational turn counts at 18 to 24 months of age accounted for 14% to 27% of the variance in IQ, verbal comprehension, and receptive and/or expressive vocabulary scores 10 years later after controlling for SES. Adult word counts between 18 and 24 months were correlated with language outcomes but were considerably weakened after controlling for SES.

CONCLUSIONS

These data support the hypothesis that early talk and interaction, particularly during the relatively narrow developmental window of 18 to 24 months of age, can be used to predict school-age language and cognitive outcomes. With these findings, we underscore the need for effective early intervention programs that support parents in creating an optimal early language learning environment in the home.

Clinician vs. Machine: Estimating Vocalizations Rates in Young Children With Developmental Disorders

PURPOSE

The purpose of this study was to investigate the reliability of an automated language analysis system, the Language Environment Analysis (LENA), compared with a human transcriber to determine the rate of child vocalizations during recording sessions that were significantly shorter than recommended for the automated device.

METHOD

Participants were 6 nonverbal male children between the ages of 28 and 46 months. Two children had autism diagnoses, 2 had Down syndrome, 1 had a chromosomal deletion, and 1 had developmental delay. Participants were recorded by the LENA digital language processor during 14 play-based interactions with a responsive adult. Rate of child vocalizations during each of the 84 recordings was determined by both a human transcriber and the LENA software.

RESULTS

A statistically significant difference between the 2 methods was observed for 4 of the 6 participants. Effect sizes were moderate to large. Variation in syllable structure did not explain the difference between the 2 methods. Vocalization rates from the 2 methods were highly correlated for 5 of the 6 participants.

CONCLUSIONS

Estimates of vocalization rates from nonverbal children produced by the LENA system differed from human transcription during sessions that were substantially shorter than the recommended recording length. These results confirm the recommendation of the LENA Foundation to record sessions of at least 1 hr.

Automated Language Environment Analysis: A Research Synthesis

PURPOSE

The Language Environment Analysis (LENA®) represents a breakthrough in automatic speech detection because it makes one's language environment, what adults and children actually hear and say, efficiently measurable. The purpose of this article was to examine (a) current dimensions of LENA research, (b) LENA's sensitivity to differences in populations and language environments, and (c) what has been achieved in closing the Word Gap.

METHOD

From electronic and human searches, 83 peer-reviewed articles using LENA were identified, and 53 met inclusionary criteria and were included in a systematic literature review. Each article reported results of 1 study.

RESULTS

Originally developed to make natural language research more efficient and feasible, systematic review identified a broad landscape of relevant LENA findings focused primarily on the environments and communications of young children but also older adults and teachers. LENA's automated speech indicators (adult input, adult-child interaction, and child production) and the audio environment were shown to meet high validity standards, including accuracy, sensitivity to individual differences, and differences in populations, settings, contexts within settings, speakers, and languages. Researchers' own analyses of LENA audio recordings have extended our knowledge of microlevel processes in adult-child interaction. To date, intervention research using LENA has consisted of small pilot experiments, primarily on the effects of brief parent education plus quantitative linguistic feedback to parents.

CONCLUSION

Evidence showed that automated analysis has made a place in the repertoire of language research and practice. Implications, limitations, and future research are discussed.

Language ENvironment analysis (LENA) system investigation of day long recordings in children: A literature review

The Language ENvironment Analysis (LENA) System is a relatively new recording technology that can be used to investigate typical child language acquisition and populations with language disorders. The purpose of this paper is to familiarize language acquisition researchers and speech-language pathologists with how the LENA System is currently being used in research. The authors outline issues in peer-reviewed research based on the device. Considerations when using the LENA System are discussed.

Babbling development as seen in canonical babbling ratios: A naturalistic evaluation of all-day recordings

Canonical babbling (CB) is critical in forming foundations for speech. Research has shown that the emergence of CB precedes first words, predicts language outcomes, and is delayed in infants with several communicative disorders. We seek a naturalistic portrayal of CB development, using all-day home recordings to evaluate the influences of age, language, and social circumstances on infant CB production. Thus we address the nature of very early language foundations and how they can be modulated. This is the first study to evaluate possible interactions of language and social circumstance in the development of babbling. We examined the effects of age (6 and 11 months), language/culture (English and Chinese), and social circumstances (during infant-directed speech [IDS], during infant overhearing of adult-directed speech [ADS], or when infants were alone) on canonical babbling ratios (CBR = canonical syllables/total syllables). The results showed a three-way interaction of infant age by infant language/culture by social circumstance. The complexity of the results forces us to recognize that a variety of factors can interact in the development of foundations for language, and that both the infant vocal response to the language/culture environment and the language/culture environment of the infant may change across age.

Automated Assessment of Child Vocalization Development Using LENA

PURPOSE

To produce a novel, efficient measure of children's expressive vocal development on the basis of automatic vocalization assessment (AVA), child vocalizations were automatically identified and extracted from audio recordings using Language Environment Analysis (LENA) System technology.

METHOD

Assessment was based on full-day audio recordings collected in a child's unrestricted, natural language environment. AVA estimates were derived using automatic speech recognition modeling techniques to categorize and quantify the sounds in child vocalizations (e.g., protophones and phonemes). These were expressed as phone and biphone frequencies, reduced to principal components, and inputted to age-based multiple linear regression models to predict independently collected criterion-expressive language scores. From these models, we generated vocal development AVA estimates as age-standardized scores and development age estimates.

RESULT

AVA estimates demonstrated strong statistical reliability and validity when compared with standard criterion expressive language assessments.

CONCLUSIONS

Automated analysis of child vocalizations extracted from full-day recordings in natural settings offers a novel and efficient means to assess children's expressive vocal development. More research remains to identify specific mechanisms of operation.

Mapping the Early Language Environment Using All-Day Recordings and Automated Analysis

PURPOSE

This research provided a first-generation standardization of automated language environment estimates, validated these estimates against standard language assessments, and extended on previous research reporting language behavior differences across socioeconomic groups.

METHOD

Typically developing children between 2 to 48 months of age completed monthly, daylong recordings in their natural language environments over a span of approximately 6-38 months. The resulting data set contained 3,213 12-hr recordings automatically analyzed by using the Language Environment Analysis (LENA) System to generate estimates of (a) the number of adult words in the child's environment, (b) the amount of caregiver-child interaction, and (c) the frequency of child vocal output.

RESULTS

Child vocalization frequency and turn-taking increased with age, whereas adult word counts were age independent after early infancy. Child vocalization and conversational turn estimates predicted 7%-16% of the variance observed in child language assessment scores. Lower socioeconomic status (SES) children produced fewer vocalizations, engaged in fewer adult-child interactions, and were exposed to fewer daily adult words compared with their higher socioeconomic status peers, but within-group variability was high.

CONCLUSIONS

The results offer new insight into the landscape of the early language environment, with clinical implications for identification of children at-risk for impoverished language environments.

Fidelity of Automatic Speech Processing for Adult and Child Talker Classifications

Automatic speech processing (ASP) has recently been applied to very large datasets of naturalistically collected, daylong recordings of child speech via an audio recorder worn by young children. The system developed by the LENA Research Foundation analyzes children's speech for research and clinical purposes, with special focus on of identifying and tagging family speech dynamics and the at-home acoustic environment from the auditory perspective of the child. A primary issue for researchers, clinicians, and families using the Language ENvironment Analysis (LENA) system is to what degree the segment labels are valid. This classification study evaluates the performance of the computer ASP output against 23 trained human judges who made about 53,000 judgements of classification of segments tagged by the LENA ASP. Results indicate performance consistent with modern ASP such as those using HMM methods, with acoustic characteristics of fundamental frequency and segment duration most important for both human and machine classifications. Results are likely to be important for interpreting and improving ASP output.

The stability and validity of automated vocal analysis in preverbal preschoolers with autism spectrum disorder

Theory and research suggest that vocal development predicts "useful speech" in preschoolers with autism spectrum disorder (ASD), but conventional methods for measurement of vocal development are costly and time consuming. This longitudinal correlational study examines the reliability and validity of several automated indices of vocalization development relative to an index derived from human coded, conventional communication samples in a sample of preverbal preschoolers with ASD. Automated indices of vocal development were derived using software that is presently "in development" and/or only available for research purposes and using commercially available Language ENvironment Analysis (LENA) software. Indices of vocal development that could be derived using the software available for research purposes: (a) were highly stable with a single day-long audio recording, (b) predicted future spoken vocabulary to a degree that was nonsignificantly different from the index derived from conventional communication samples, and (c) continued to predict future spoken vocabulary even after controlling for concurrent vocabulary in our sample. The score derived from standard LENA software was similarly stable, but was not significantly correlated with future spoken vocabulary. Findings suggest that automated vocal analysis is a valid and reliable alternative to time intensive and expensive conventional communication samples for measurement of vocal development of preverbal preschoolers with ASD in research and clinical practice. Autism Res 2017, 10: 508-519. © 2016 International Society for Autism Research, Wiley Periodicals, Inc.

Subtlety of Ambient-Language Effects in Babbling: A Study of English- and Chinese-Learning Infants at 8, 10, and 12 Months

Prior research on ambient-language effects in babbling has often suggested infants produce language-specific phonological features within the first year. These results have been questioned in research failing to find such effects and challenging the positive findings on methodological grounds. We studied English- and Chinese-learning infants at 8, 10, and 12 months and found listeners could not detect ambient-language effects in the vast majority of infant utterances, but only in items deemed to be words or to contain canonical syllables that may have made them sound like words with language-specific shapes. Thus, the present research suggests the earliest ambient-language effects may be found in emerging lexical items or in utterances influenced by language-specific features of lexical items. Even the ambient-language effects for infant canonical syllables and words were very small compared with ambient-language effects for meaningless but phonotactically well-formed syllable sequences spoken by adult native speakers of English and Chinese.

HomeBank: An Online Repository of Daylong Child-Centered Audio Recordings

HomeBank is introduced here. It is a public, permanent, extensible, online database of daylong audio recorded in naturalistic environments. HomeBank serves two primary purposes. First, it is a repository for raw audio and associated files: one database requires special permissions, and another redacted database allows unrestricted public access. Associated files include metadata such as participant demographics and clinical diagnostics, automated annotations, and human-generated transcriptions and annotations. Many recordings use the child-perspective LENA recorders (LENA Research Foundation, Boulder, Colorado, United States), but various recordings and metadata can be accommodated. The HomeBank database can have both vetted and unvetted recordings, with different levels of accessibility. Additionally, HomeBank is an open repository for processing and analysis tools for HomeBank or similar data sets. HomeBank is flexible for users and contributors, making primary data available to researchers, especially those in child development, linguistics, and audio engineering. HomeBank facilitates researchers' access to large-scale data and tools, linking the acoustic, auditory, and linguistic characteristics of children's environments with a variety of variables including socioeconomic status, family characteristics, language trajectories, and disorders. Automated processing applied to daylong home audio recordings is now becoming widely used in early intervention initiatives, helping parents to provide richer speech input to at-risk children.

Novel methodology to examine cognitive and experiential factors in language development: combining eye-tracking and LENA technology

Developmental systems theory posits that development cannot be segmented by influences acting in isolation, but should be studied through a scientific lens that highlights the complex interactions between these forces over time (Overton, 2013a). This poses a unique challenge for developmental psychologists studying complex processes like language development. In this paper, we advocate for the combining of highly sophisticated data collection technologies in an effort to move toward a more systemic approach to studying language development. We investigate the efficiency and appropriateness of combining eye-tracking technology and the LENA (Language Environment Analysis) system, an automated language analysis tool, in an effort to explore the relation between language processing in early development, and external dynamic influences like parent and educator language input in the home and school environments. Eye-tracking allows us to study language processing via eye movement analysis; these eye movements have been linked to both conscious and unconscious cognitive processing, and thus provide one means of evaluating cognitive processes underlying language development that does not require the use of subjective parent reports or checklists. The LENA system, on the other hand, provides automated language output that describes a child's language-rich environment. In combination, these technologies provide critical information not only about a child's language processing abilities but also about the complexity of the child's language environment. Thus, when used in conjunction these technologies allow researchers to explore the nature of interacting systems involved in language development.