A current issue in alcohol research is whether a "neurobehavioral profile" can be identified for prenatal alcohol exposure, even when dysmorphic features are not present, or whether comparable neurobehavioral deficits are detected when damage is incurred by numerous neurotoxicants to which the fetus is exposed during a common developmental period. Failure to detect such differences may, in part, be an artifact of the global developmental tests used to assess outcome. Cognitive effects of prenatal exposure to three different teratogens [polychlorinated biphenyls (PCBs), alcohol, and cocaine] are examined to determine whether exposure to each substance results in a common or different pattern of impairment on the same set of newer, more narrow band infant tests. Comparison of findings from three independent cohorts indicate that PCB exposure was related to poorer recognition memory on the Fagan Test of Infant Intelligence (FTII) in Michigan infants exposed prenatally to PCB-contaminated fish, whereas prenatal alcohol exposure was unrelated to recognition memory but to slower processing speed on a new FTII measure and slower reaction time on Haith's Visual Expectancy Paradigm (VExP) in our Detroit alcohol-exposed infants. Preliminary findings from a new study of infants recently born to Taiwanese women accidentally contaminated with sizable amounts of PCBs indicate recognition memory deficits, confirming our Michigan findings, but no processing speed effects on the FTII. Recent findings from our Detroit cohort suggest that heavy prenatal cocaine exposure is related to poorer recognition memory on the FTII, but faster reaction times on the VExP, a pattern different from that seen for either PCBs or alcohol.