Drugs used to treat affective disorders exert their effects largely through their actions on various neurotransmitter systems. Neurotransmitters are important regulators of neural development. Thus, even limited periods of drug use/exposure during fetal- or postnatal brain ontogeny, which continues to the end of the second decade, can, potentially, cause enduring, functionally significant changes in neuronal circuitry. The duration and timing of drug treatment are critical in determining long-term effects (Popa et al, 2008), owing to developmental changes in the neural substrates on which the drugs act. Stress, which is associated with affective disorders, adversely impacts neural development during fetal life and postnatally, and is ameliorated by therapies that improve affect. Thus, potential, adverse, long-term effects of exposing the developing brain to medication must be weighed against the deleterious impact of stress elevation as a result of interrupted pharmacotherapy of pregnant women, or not treating children or adolescents. Human studies have revealed few effects of fetal antidepressant drug (ADD) exposure that endure beyond infancy. By contrast, adult rodents exposed to ADDs only during stages corresponding to fetal life, childhood, or adolescence in humans exhibit a spectrum of abnormalities, many of which are, surprisingly, components of depression. Similar, long-term, ‘mirror image' effects (that is, drug-induced increases in disease components that are reduced by pharmacological treatment of adults) occur after early-life anxiolytic treatment (Depino et al, 2008). Early exposure to atypical antipsychotic drugs, used to treat bipolar disorder in children and adolescents, induces significant, long-lasting cognitive deficits (Zuo et al, 2008) and morphological abnormalities (Frost et al, 2009). Why do rodent and human data seem discrepant? In rodents, the behavioral syndromes induced by early ADD (Ansorge et al, 2008) or anxiolytic (Depino et al, 2008) treatment are progressive and emerge fully only in early adulthood. This is because early-life brain insults induce a cascade of effects over the course of development. Human studies of the effects on progeny of maternal ADD treatment during pregnancy so far reach only up to 7–10 years of age—too early to characterize long-term outcomes. Early-life experience and environmental factors are emerging as additional, important modulators of the effects of psychotropic medications. We recently showed that, in rats, some of the enduring behavioral consequences of fetal exposure to fluoxetine do not occur if exposed infants are subjected to behavioral testing, which, by its nature, provides supplemental sensory stimulation and exercise (R Gibb, DO Frost and B Kolb, unpublished data). A provocative, parallel finding has recently been reported in humans: In infants who are not breast fed, maternal SSRI use during pregnancy is associated with altered hypothalamo–pituitary–adrenal (HPA) stress responses, whereas breast feeding abolishes this effect of fetal SSRI exposure by 3 months of age (Oberlander et al, 2008). The mode of feeding seems to act epigenetically: breast-fed infants have lower methylation (and probably higher expression) of the glucocorticoid receptor gene NR3C1 than nonbreast-fed infants (Oberlander et al, 2008). Animal and human studies are consistent in demonstrating that environment and experience during fetal and postnatal brain maturation interact with early-life psychotropic drug exposure to shape the ontogeny of behavior and neural circuitry. Epigenetic modulation of gene expression appears to be one important mediator of these effects. These findings suggest possibilities for designing new therapeutic strategies that mitigate the effects of early-life ADD exposure.