The amygdala, a key player in emotional processing, undergoes rapid growth and development prenatally and in infancy. Environmental factors, such as prenatal maternal depressive symptoms, can alter the structure of the amygdala. The altered structure of the amygdala can influence neural and functional development, contributing to negative outcomes in offspring. However, identifying limbic system structures, such as the amygdala and hippocampus, in infants' brain magnetic resonance (MR) images is challenging. This difficulty arises from the small size of these structures, the lower resolution of the scans, and the lack of accurate segmentation methods. This thesis encompasses three studies involving mothers and infants from the FinnBrain Birth Cohort Study. In study I, a manual segmentation protocol tailored to neonatal MR images was established, focusing on the amygdala and hippocampus in healthy infants. In study II, we investigated the links between prenatal maternal depressive symptoms (PMDS) at different gestational stages and infant amygdala microstructure using diffusion tensor imaging (DTI). Sex-specific responses to PMDS were revealed, with boys showing a stronger positive association between PMDS and the left amygdala DTI parameter (mean diffusivity). In study III, we explored the neural basis of emotional perception in infancy, correlating amygdala microstructure with infants' recognition of emotional faces using eye-tracking. Increased right amygdala mean diffusivity was associated with reduced disengagement from fearful faces, while sex differences revealed that girls with elevated amygdala mean diffusivity exhibited greater disengagement from scrambled non-face control pictures. These studies advance our understanding of early neurodevelopment by examining the development of the amygdala, a crucial element in emotional processing and regulation.