Being able to discriminate faces from other objects and to tell apart emotional facial expressions are crucial skills for effective social communication, especially for pre-verbal infants. The overarching aim of this thesis is to aid the current understanding of how face and facial expression processing develops between the first and second semester of life. To do this, I measured infant brain activity using event-related potentials (ERPs) and functional near-infrared spectroscopy (fNIRS) while infants observed houses as well as faces with different expressions. The first part of this dissertation focused on improving the definition of robust neural measures of infant face processing. In one study, I aimed at providing guidelines for developmental fNIRS researchers on how to process infant data. I mainly focus on the most optimal strategy to deal with motion artifacts, which are a major source of noise in infant fNIRS studies. The results suggest the use of a novel combination of methods (Wavelet and Spline) for infant data, in order to obtain reliable brain responses. In another study, I tested which ERP component consistently marks face processing across the first year of life. The results indicate that two components related to visual processing stages (N290; P400) prove to be more robust markers of face processing than a component linked to mechanisms of attention and preference (Nc). The second part of this thesis investigated the development of face and facial expression processing between five and 10 months of age. My research suggests that face processing recruits parts of the adult face-cortical network already at five months of age (fNIRS study), and remains equally active between the first and second semester of life (ERP study). Turning to facial expression processing, this thesis further demonstrates that at five months of age, the infant brain does not discriminate between emotional facial expressions while at 10 months expression discrimination is present. At this later stage of development, I described that infants mainly rely on facial details to discriminate between emotional facial expressions. Crucially, my research further indicated that at 10 months, infants with a higher likelihood to develop Autism Spectrum Disorder (ASD) do not discriminate between emotional facial expressions when either detailed or global information is displayed. Combining these results with findings from previous research we suggest a developmental delay in the use of specific visual information to process facial expressions in ASD. Even though this thesis contributes to the understanding of face and facial expression processing in five to 10-month-olds, the development of such skills across infancy remains not fully clear. To chart development we need large longitudinal studies combining different measures.