Practical considerations on the design, execution and analysis of developmental neurotoxicity studies to be published in Neurotoxicology and Teratology.

The present article is an attempt to use the "ILSI Research Foundation/Risk Science Institute Reports from the Expert Working Group on Neurodevelopmental Endpoints" (2008) to help improve the quality of the manuscripts submitted to Neurotoxicology and Teratology, as well as the quality of their review. The points discussed in the present paper have been encountered during the peer-review process. A number of recommendations are proposed on the basis of general principles (clarity, full disclosure, and evidence-based interpretation) to help authors and reviewers. Clarity of language is a prerequisite, but clarity of purpose is essential. The methodology and statistical analysis for each dependent variable should be unambiguously presented and justified. Full disclosure encompasses a range of topics, such as defining the sample size for each experiment, clearly distinguishing between hypothesis-testing and hypothesis-generating (e.g., a priori vs. a posteriori analyses), clearly defining the statistical model appropriate to the study design and questions (e.g., repeated-measure approach), recognizing and addressing the multiplicity problem (e.g., conceptual unit for the error rate), identifying the appropriate unit for statistical analysis (e.g., litter), addressing the results of all analyses (e.g., "negative" results are important). Data interpretation should be evidence-based and not exceed the limits of the findings.