Although both konzo and neurolathyrism are diseases associated with diet, we know surprising little about the diets of the groups at risk. The objective of this paper is to discuss methods for assessing dietary intake in populations at risk for konzo and lathyrism. These methods include weighed food records and interview based techniques like 24-h recalls and food frequency questionnaires (FFQs). Food records have the potential to provide accurate information on food quantities, and are generally the method of choice. Interview based methods provide less precise information on the quantities of foods ingested, and are subject to recall bias, but may be useful in some studies or for surveillance. Sample size needs to be adequate to account for day-to-day and seasonal variability in food intake, and differences between age and sex groups. Adequate data on the composition of foods, as actually consumed, are needed to evaluate the food intake information. This is especially important in the case of cassava and grass pea where the toxins in the diet is a function of processing. Biomarkers for assessing the cyanogen exposure from cassava-based diets are available; biomarkers for the β-ODAP exposure from grass pea diets need development.