Microstructural analyses of human ingestive patterns: from description to mechanistic hypotheses.

Continuous automated weighing of food while subjects ate was used to test the hypothesis that failure to slow eating rate during a meal indicated a deficient response to satiety signals in obese patients. Cumulative intake curves were fitted to a quadratic equation. The physical form of the food and its palatability were a greater influence on the equation's parameters than the subjects' body weights, and the hypothesis was abandoned for several years (1984-1993). The hypothesis was revived with modifications when we discovered disturbances in eating behavior in patients with bulimia nervosa. The new hypothesis was that overeating was attributable to subjects' inability to detect or respond to satiety-related signals after eating large amounts of food. Patients with eating disorders showed lower ratings of satiety after eating the same amounts of food as controls, but only after eating more than normal. In conclusion, microstructural examination of eating behavior may be more useful for tests of specific hypotheses about the control of eating than as a description of clinical disturbance.