difficult to incorporate into large-scale projects like the Argus studies, which were designed to assess multiple heath outcomes. Furthermore, they are not required in EPA’s Developmental Neurotoxicity Testing Guidelines (EPA 1998a). However, without additional cognitive assessments that use tests that focus on the specific cognitive domains most likely to be affected by subtle to moderate reductions in thyroid hormone concentrations, it is impossible to rule out an effect of perchlorate on cognition.

Another useful approach would be a transgenic mouse model with a phenotype characterized by enlarged corpus callosum axons (Seeger et al. 2003). Characterization of the behavioral profile in such animals would provide valuable information that would help to correlate behavioral changes with morphologic effects of perchlorate exposure.

Locomotor-activity testing was conducted at the ages of 14, 18, 22, and 61 days in the 1998 Argus study, and at the ages of 14, 18, and 22 days by Bekkedal et al. (2000). The design of the Bekkedal et al. (2000) study was similar to that of the Argus study. Female Sprague-Dawley rats were treated with ammonium perchlorate at 0, 0.1, 1.0, 3.0, or 10.0 mg/kg per day in the drinking water beginning 2 weeks before gestation and continuing until postnatal day 10. The rats in both studies were tested in an open-field activity monitor transected by infrared photobeams. The screening tool was appropriate because treatment with PTU or PCBs has been shown to suppress locomotor activity during preweaning development and increase locomotor activity at later ages (Goldey et al. 1995a, 1995b). However, locomotor activity can be variable because it is influenced by many parameters. Therefore, it is not a very sensitive measure of exposure-related effects. No significant effects were noted in the Bekkedal et al. (2000) study. The only significant effect noted in the Argus study was an increase in activity on postnatal day 14 in male pups in the highest-exposure perchlorate group (10.0 mg/kg). The effect was manifested as a failure to habituate over the 1.5-hour test session. Argus’s statistical analyses indicated a significant effect when male and female pups were evaluated separately, but a reanalysis in which male and female data were combined found no significant effects.

In November 2001, David Dunson, of the National Institute of Environmental Health Sciences (NIEHS), was asked to provide an analysis of Argus (1998) and of Bekkedal et al. (2000) because the original contractor was unable to find statistically significant effects despite large increases in several outcome variables. EPA neurotoxicologists believed that the increases were of concern from a biologic perspective. Because of the high correlation between different types of motor-activity measurements, Dunson focused on the number of ambulatory movements, which was viewed as