Language and social-emotional atypicalities may act as important early markers of potential developmental problems in children. Advances in psychophysiological tools have led to an understanding of the neural correlates that underpin infant voice and vocal emotion processing as the precursors to language and social-emotional brain development. However, there is little evidence for the longitudinal developmental trajectory of infant voice and vocal emotion processing. In addition, the influence of early experience, especially maternal caregiving behaviour, on the infant voice and vocal emotion processing is less clear. To address research gaps relating to the early infant language and social-emotional neural development, this PhD study involved developing an acceptable and reliable method to observe changing voice and vocal emotion processing in the infant brain in the first year of life. The study aimed (a) to develop and pilot a functional Near-Infrared Spectroscopy (fNIRS) protocol that could be acceptably and feasibly applied in a clinical setting in order potentially to monitor and evaluate the neurological underpinnings of early language and social-emotional development; (b) to track longitudinal changes in neural correlates of human voice and emotional vocalisation processing in typically developing infants at the age of 6, 9 and 12 months; (c) to explore the role of early maternal caregiving behaviour as a possible mediator in the healthy development of voice and vocal emotion processing in infants. Chapter 2 describes methodological considerations in the longitudinal assessment of infant voice and vocal emotion neural processing. Chapter 3 compared fNIRS data analysis methods between standardised analysis and Independent Component Analysis (ICA) implanted emotion-sensitive components identification procedures (paper 1). Chapter 4 (paper 2) reports the acceptability and feasibility of piloting the fNIRS paradigm in a non-research, community-based clinical setting in typically developing infants at their ages of 6, 9 and 12 months. Chapter 5 to 7 (paper 3 to 5) report the observation of infant neural development in processing emotional vocalisations, human voice and non-vocal sounds over time, and the influence of maternal caregiving behaviour on the infantâs neural response to vocal emotion. Mother-infant pairs successfully completed the longitudinal study with high retention and satisfaction rate. We found consistent temporal cortical activations to auditory stimuli, particularly to happy vocalisations and non-vocal sounds, but the activation locations within temporal cortices were inconsistent over time. Right temporal cortical responses to angry vocalisations were significantly increased with age. Six-month-old infantâs neural responses to vocal anger were significantly associated with maternal directiveness. We also report a lack of significantly enhanced neural responses to voice than to non-vocal stimuli. Findings suggest that the fNIRS paradigm has a promising outlook for the clinical use in infants and children. The infant neural responses to human voice stimuli are continuously being âfine-tunedâ in temporal cortices. We also report a possible interaction with the quality of maternal caregiving. Infants may have distinct neural developmental trajectories of processing positive and negative emotional vocalisations. Future research will benefit from exploring links between neural responses to human vocalisations and language and social-emotional development in typically and atypically developing infants and children. Future study is also needed to optimise infant fNIRS data analysis, especially with the implementation of ICA.