Prenatal methamphetamine exposure is associated with functional and neurostructural alterations, but neuroimaging investigations of these effects in infants are almost non-existent. Studies in neonates permit a degree of separation of drug exposure effects from potential confounders in the postnatal environment. Magnetic resonance imaging (MRI) was used to investigate the neurostructural effects of prenatal methamphetamine exposure on neonates recruited from a Cape Town community. Mothers were recruited during pregnancy and interviewed regarding methamphetamine use. Women in the exposure group used methamphetamine at least twice per month during pregnancy, while control mothers did not use methamphetamine. MRI scans were acquired within the first postnatal month. Anatomical images were processed using FreeSurfer and subcortical and cerebellar structures manually segmented with Freeview. Volumes were regressed with methamphetamine exposure (days/month of pregnancy) and related confounding variables, including total brain volume, gestational age at scan, exposure to cigarette smoking and infant sex. Diffusion data were processed with FSL, and diffusion tensors and tensor parameters determined using AFNI. Probabilistic tractography defined white matter connections between target regions. For the first analysis, five major white matter networks (commissural, and bilateral projection and association networks) were defined between spherical targets. For the second analysis, regions traced in the anatomical study were used as targets. Averaged DTI parameters were then calculated for each connection, and multiple regression analysis determined associations between DTI parameters and methamphetamine exposure at network level and in the individual connections. Methamphetamine exposure was associated with reduced caudate nucleus volume bilaterally, and in the right caudate following adjustment for confounders. Exposure was associated with reduced fractional anisotropy in all major white matter networks, and in individual connections within the limbic meso-cortico-striatal circuit. Exposure was associated with increased radial diffusivity in a subset of these. These results support findings in older children of methamphetamine-induced neurostructural damage, and demonstrate that such effects are already measurable in neonates. Corticostriatal circuit changes may underlie the impaired executive function observed in prenatally exposed children, and suggest a specific mechanism of damage in dopaminergic-related circuits that is consistent with the neurotoxic actions of methamphetamine.