may become manifest as disease in the elderly. On the basis of the available evidence, the committee hypothesizes that an as-yet-unspecified fraction of human neurologic and psychiatric disease is attributable to chemical agents in the environment.

The committee concludes that a major obstacle to assessing the extent to which chemicals in the environment cause nervous system diseases and dysfunction is that little qualitative or quantitative information is available on possible adverse effects of most environmental chemicals on the nervous system. Some chemicals in commerce are known to have neurotoxic potential, but most commercial chemicals have not been assessed for neurotoxicity. Although resources are not readily available to undertake across-the-board testing of all chemicals already in commerce, prudent public policy dictates that all chemicals, both old and new, be subject to at least basic screening for neurotoxicity when use and exposure warrant. There is a particular lack of data on chronic and long-latency neurotoxic effects. Structure-activity relationships, now the most widely used approach to assessment of toxicity, provide a poor basis for predicting neurotoxic potential; however, greater understanding of chemical structural correlations and the underlying mechanisms of toxicity can be expected to lead to the discovery of more useful applications of structure-activity relationships.

The committee recommends that a more accurate estimate of the extent of the problem of neurologic and psychiatric dysfunction attributable to chemical agents in the environment be made.

The estimate must be based on a combination of clinical, epidemiologic, and toxicologic studies coupled with the techniques of quantitative risk assessment.

The committee concludes that additional biologic markers for the assessment of subclinical neurotoxic effects are needed. Such markers can be biochemical, structural, or functional. They can be developed either through in vitro analyses, through animal studies, or during observational studies in human populations exposed to environmental neurotoxicants. Although associations between biologic markers and disease are usually established initially in cross-sectional studies, a particular need exists to validate putative biologic markers in prospective studies. Only in longitudinal prospective studies can the ability of biologic markers to predict the occurrence of disease be accurately assessed.

The committee recommends that putative biologic markers in animal species be evaluated and validated in in vivo and in vitro systems. The committee further recommends that biologic markers be regularly incorporated into epidemiologic and clinical studies of neurologic disease, particularly prospective studies.

The primary goal of the incorporation of biologic markers into such studies should be to validate their predictive accuracy and to test hypothesized quantitative relationships between specific markers that are likely to be on or close to causal pathways and neurotoxic outcomes.

The legislated authorities of the federal regulatory agencies (the Occupational Safety and Health Administration, the Environmental Protection Agency, the Consumer Product Safety Commission, and the Food and Drug Administration) to regulate or to require neurotoxicity testing might be too narrow to encourage investigation of neurotoxicity. The potential authority of the existing mandates of those agencies to control toxic substances, including neurotoxicants, has not been fully used.

The committee concludes that tests are available to construct a tiered approach to neurotoxicity testing. The first tier, or screen, is intended for hazard identification. The results of the screen and a chemical's

Front Matter (R1-R12)

Executive Summary (1-8)

1. Introduction: Defining the Problem of Neurotoxicity (9-20)

2. Biologic Basis of Neurotoxicity (21-42)

3. Biologic Markers in Neurotoxicology (43-52)

4. Testing for Neurotoxicity (53-94)

5. Surveillance to Prevent Neurotoxicity in Humans (95-110)

6. Risk Assessment (111-122)

7. Conclusions and Recommendations (123-128)

References (129-148)

Index (149-154)