To study how early gross motor development links to concurrent prelinguistic and social development. We recruited a population-based longitudinal sample of 107 infants between 6 21 months age. Gross performance was quantified using novel wearable technology for at-home recordings infants' spontaneous activity. The assessed in parallel with standardized parental questionnaire (Infant Toddler Checklist). developmental trajectories motor, prelinguistic, were inspected longitudinally at...

This study presents a new glottal inverse filtering (GIF) technique based on closed phase analysis over multiple fundamental periods. The proposed quasi (QCP) method utilizes weighted linear prediction (WLP) with specific attenuated main excitation (AME) weight function that attenuates the contribution of source in model optimization. enables use autocorrelation criterion contrast to covariance used conventional analysis. QCP was compared previously developed methods by using synthetic...

Abstract Infants’ spontaneous and voluntary movements mirror developmental integrity of brain networks since they require coordinated activation multiple sites in the central nervous system. Accordingly, early detection infants with atypical motor development holds promise for recognizing those who are at risk a wide range neurodevelopmental disorders (e.g., cerebral palsy, autism spectrum disorders). Previously, novel wearable technology has shown offering efficient, scalable automated...

Abstract Background Early neurodevelopmental care needs better, effective and objective solutions for assessing infants’ motor abilities. Novel wearable technology opens possibilities characterizing spontaneous movement behavior. This work seeks to construct validate a generalizable, scalable, method measure abilities across all milestones from lying supine fluent walking. Methods A multi-sensor infant was constructed, 59 infants (age 5–19 months) were recorded during their play. novel gross...

This paper proposes a method for generating speech from filterbank mel frequency cepstral coefficients (MFCC), which are widely used in applications, such as ASR, but generally considered unusable synthesis. First, we predict fundamental and voicing information MFCCs with an autoregressive recurrent neural net. Second, the spectral envelope contained is converted to all-pole filters, pitch-synchronous excitation model matched these filters trained. Finally, introduce generative adversarial...

Abstract Assessing infant carrying and holding (C/H), or physical infant-caregiver interaction, is important for a wide range of contexts in development research. An automated detection quantification C/H particularly needed long term at-home studies where infants’ neurobehavior measured using wearable devices. Here, we first developed phenomenological categorization interactions to support five different definitions behaviors. Then, trained assessed deep learning-based classifiers their...

A vocoder is used to express a speech waveform with controllable parametric representation that can be converted back into waveform. Vocoders representing their main categories (mixed excitation, glottal, and sinusoidal vocoders) were compared in this study formal crowd-sourced listening tests. The quality was measured within the context of analysis-synthesis as well text-to-speech (TTS) synthesis modern statistical framework. Furthermore, TTS experiments divided vocoder-specific features...

Recent speech technology research has seen a growing interest in using WaveNets as statistical vocoders, i.e., generating waveforms from acoustic features.These models have been shown to improve the generated quality over classical vocoders many tasks, such text-to-speech synthesis and voice conversion.Furthermore, conditioning with features allows sharing waveform generator model across multiple speakers without additional speaker codes.However, multi-speaker WaveNet require large amounts...

BackgroundElectroencephalogram (EEG) monitoring is recommended as routine in newborn neurocritical care to facilitate early therapeutic decisions and outcome predictions. EEG's larger-scale implementation is, however, hindered by the shortage of expertise needed for interpretation spontaneous cortical activity, EEG background. We developed an automated algorithm that transforms recordings quantified interpretations background provides simple intuitive visualisations patient...

Abstract To investigate how a high risk for infant neurological impairment affects the quality of verbal interactions, and in particular properties infant-directed speech, spontaneous interactions between 14 mothers their 4.5-month-old infants at disorders (7 female) were recorded acoustically compared with those dyads typically developing (8 female). Mothers at-risk had proportionally less voicing, proportion voicing decreased increasing severity infants’ long-term outcome. Follow-up...

Abstract Background Tracking of early motor development is essential for all neurodevelopmental assessments. A multisensor wearable system, MAIJU (Motor Assessment Infants with a JUmpsuit), was recently developed an objective and scalable measurement developing skills in out-of-hospital settings. Here, we assessed its feasibility remote low-resource Methods We recruited 44 infants repeated at-home measurements (total N = 121) the rural Malawi. (i) technical quality measured data, (ii)...

Early development of gross motor skills is foundational for the upcoming neurocognitive performance. Here, we studied whether at-home wearable measurements performed by parents could be used to quantify and track infants' developing abilities. Unsupervised spontaneous activity were made repeatedly using a multisensor suit (altogether 620 from 134 infants at age 4-22 months). Machine learning-based algorithms developed detect reaching milestones (GMM), measure times spent in key postures,...

GlottHMM is a previously developed vocoder that has been successfully used in HMM-based synthesis by parameterizing speech into two parts (glottal flow, vocal tract) according to the functioning of real human voice production mechanism. In this study, new glottal vocoding method, GlottDNN, proposed. The GlottDNN built on principles its predecessor, GlottHMM, but introduces three main improvements: (1) takes advantage new, more accurate inverse filtering (2) uses method deep neural network...

BackgroundEarly neurodevelopmental care and research are in urgent need of practical methods for quantitative assessment early motor development. Here, performance a wearable system was validated compared to developmental tracking physical growth charts.MethodsAltogether 1358 h spontaneous movement during 226 recording sessions 116 infants (age 4–19 months) were analysed using multisensor system. A deep learning-based automatic pipeline quantified categories infants' postures movements at...

Developing objective and quantitative methods of early gross motor assessment is essential to better understand neurodevelopment support therapeutic interventions. Here, we present a method quantify performance using multisensor wearable, MAIJU (Motility Assessment Infants with JUmpsuit), which offers an automated, scalable, quantitative, fully automated cloud-based pipeline. This wearable suit equipped four movement sensors that record synchronized data mobile phone utilizing low-energy...

Achieving high quality and naturalness in statistical parametric synthesis of female voices remains to be difficult despite recent advances the study area. Vocoding is one such key element all speech synthesizers that known affect naturalness. The present focuses on a special type vocoding, glottal vocoders, which aim parameterize based modelling real excitation (voiced) speech, flow. More specifically, we compare three different vocoders by aiming at improved voices. Two are previously...

Neonatal brain monitoring in the neonatal intensive care units (NICU) requires a continuous review of spontaneous cortical activity, i.e., electroencephalograph (EEG) background activity. This needs development bedside methods for an automated assessment EEG In this paper, we present key components classifier, starting from visual scoring to classifier design, and finally possible visualization results. A dataset with 13,200 5-minute epochs (8–16 channels) 27 infants birth asphyxia was used...

In this study, the acoustic properties of shouted speech are analyzed in relation to normal speech, and various synthesis techniques for shouting investigated. The analysis shows large differences between two styles, which induces difficulties synthesis. Analysis-synthesis experiments show that use spectral estimation methods not biased by sparse harmonics is beneficial. performed through adaptation voice conversion. Subjective evaluation reveals that, despite quality degradation, impression...

Recently, a quasi-closed phase (QCP) analysis of speech signals for accurate glottal inverse filtering was proposed. However, the QCP which belongs to family temporally weighted linear prediction (WLP) methods uses conventional forward type sample prediction. This may not be best choice especially in computing WLP models with hard-limiting weighting function. A selective minimization error reduces effective number samples available within given window frame. To counter this problem, modified...