Abstract:

Language is the primary means of human communication. Understanding how it develops and which brain regions control it is a significant concern for psychologists and linguists. In early life, especially during the neonatal stage, infants possess strong perceptual abilities for speech sounds. Newborns can distinguish vowels and consonants from different languages. However, this sensitivity narrows as they are exposed to their native language, making it challenging to perceive non-native phonemes. Studying newborns' perception, discrimination, and learning of speech sounds provides insights into early cognitive mechanisms of language development and aids in understanding neurodevelopmental disorders like autism.

Early studies on newborns and infants often use the “habituation-dishabituation paradigm” with “nipple sucking rate” as an indicator. Newborns exhibit perceptual preferences for speech sounds, their native language, and their mother's voice. Brain observation techniques like electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) are used to study speech perception in infants. EEG focuses on the mismatch response (MMR), while fNIRS reveals specific cortical regions involved in speech processing. Both techniques show left-hemisphere dominance in newborns' language processing, with left temporal and frontal lobes being more activated during language tasks. This leftward bias is evident in both functional and structural aspects of the brain. Understanding early language perception is crucial for developmental psychology and related clinical research.

Newborns possess strong phoneme discrimination abilities, distinguishing vowels and consonants from different languages. They can also recognize syllables and syllable sequences. Studies using the habituation-dishabituation paradigm and brain observation techniques like EEG and fNIRS demonstrate newborns' unique language capabilities. They exhibit perceptual preferences for their native language and can differentiate it from other languages. Behaviorally, newborns can distinguish sentences and perceive accent patterns. EEG and fNIRS studies show newborns' ability to discriminate vowels, consonants, and syllables, with left hemisphere involvement. They also display sensitivity to syllable sequences, showing enhanced brain responses to certain sequence structures. These findings contribute to understanding language processing and statistical learning in early development.

To date, research on newborns' speech perception and phoneme discrimination has primarily employed passive observation of language exposure in utero (e.g., Moon et al., 2013) or single-time-point assessments (excluding Moon's study). Only one study (Partanen et al., 2013) has examined fetal language learning by exposing fetuses to speech sounds prenatally. Few studies have directly observed changes in brain activity before and after speech learning in newborns. Studies using EEG and fNIRS found enhanced neural responses to learned vowels, demonstrating brain plasticity due to learning. Additionally, a study using fNIRS showed that sleep influenced the brain's response to learned vowels. However, the debate remains about whether newborns can differentiate vowels or if their sensitivity is related to prosody rather than phoneme discrimination. Future research should explore the effects of sleep on newborns' language learning and the broader range of phoneme learning in newborns.

Autism is a major neurodevelopmental disorder in early childhood, characterized by social communication difficulties, language impairments, and repetitive behaviors, significantly impacting lifelong social functioning (American Psychiatric Association, 2013). The global prevalence is 1-2% (Hirota & King, 2023; World Health Organization, 2023; Zeidan et al., 2022). Language issues are a primary concern, with affected children showing reduced language abilities and abnormal brain networks (Belteki et al., 2022; Tryfon et al., 2018). Early diagnosis and intervention can improve outcomes, but the link between newborn language development and autism needs further research. This review focuses on infant studies, particularly longitudinal research predicting autism in high-risk infants with genetic and brain risk factors (Hirota & King, 2023). Longitudinal studies suggest language behavioral indicators predict autism after one year, while brain indicators show predictive value as early as three months (Ayoub et al., 2022; Clairmont et al., 2022; Molnar-Szakacs et al., 2021). Three longitudinal studies indicate language brain indicators predict autism in high-risk infants. Language lateralization in the brain also has predictive implications for autism (Lindell, 2020; Herringshaw et al., 2016). Five longitudinal studies on high-risk infants show language lateralization indicators predict autism during infancy. Research on high-risk infants provides valuable insights into the predictive value of early language development in autism. Capturing language processing brain indicators in newborns may offer valuable personalized warning parameters for early autism diagnosis, taking advantage of the brain's greater plasticity at younger ages.

In summary, current research on newborns' speech perception, discrimination, and learning indicates the following: 1) Newborns exhibit speech perception preferences, showing a preference for speech, their native language, and their mother's voice, with a leftward brain lateralization. 2) Newborns possess unique phoneme discrimination abilities, differentiating vowels, consonants of various languages, and complex syllables and sequences. 3) Early language learning in newborns leads to plasticity changes in the brain's language networks. Moreover, several studies suggest that early language development brain indicators have significant predictive value for autism. However, there are three critical issues in the foundational and translational research of newborn language processing. Firstly, the rhythmic features of speech materials have been overlooked, potentially interfering with newborns' phoneme discrimination. Secondly, the cognitive neural mechanisms of newborn speech learning remain unclear, particularly regarding consonant learning and the role of sleep in language learning. Lastly, there is a lack of clinical translational research on newborn language development, necessitating the exploration of early language-related brain markers to predict and warn against neurodevelopmental disorders like autism. Future research should address these gaps by rigorously controlling rhythmic factors in speech materials, investigating the neural mechanisms of consonant learning using EEG and fNIRS techniques, and exploring the role of sleep in memory consolidation during language learning. Additionally, longitudinal studies focusing on high-risk newborns can potentially establish a comprehensive risk assessment system for neurodevelopmental disorders based on multiple brain modalities and clinical evaluations. Addressing these challenges may pave the way for early intervention and prevention of language-related developmental disorders in infancy.

Key words: newborn, native language preference, left lateralization, language development, autism spectrum disorder