RESUMO
The optimal design of behavioral experiments includes measures to minimize observer bias while maximizing researcher efficiency. To this end, we have developed a reliable, autonomous monitor of animal feeding behavior. Previous studies of caterpillar (Manduca sexta) feeding have relied on human observation for visual evaluation of food consumption at specific time points over the course of several hours. This method is eliminated by our new behavior rig, which collects data automatically. Individual animals are monitored via CCD cameras activated in sequence by a computer. Images are obtained at preset time intervals using a frame grabber to capture still pictures from the cameras. Subsequently, images are analyzed using software written in MatLab to determine food selection and quantify consumption.
